The effects of space weathering on Apollo 17 mare soils: Petrographic and chemical characterization

Transcription

1 $5 +, Meteoritics & Planetary Science 36, (2001) The effects of space weathering on Apollo 17 mare soils: Petrographic and chemical characterization LAWRENCE A. TAYLORl*, CARLE PIETERS2, LINDSAY P. KELLER3, RICHARD V. MORRIS, DAVID S. McKAY3, ALLAN PATCHEN1 AND SUSAN WENTWORTHS I Planetary Geosciences Institute, University of Tennessee, Knoxville, Tennessee 37996, USA 2Geological Sciences, Brown University, Providence, Rhode Island 02912, USA 3Earth Science and Solar System Exploration Division, NASA Johnson Space Center, Houston, Texas 77058, USA 4Lockheed Martin Space Operations Co., Houston, Texas 77058, USA *Correspondence author's address: lataylor@utk.edu (Received 2000 April 26; accepted in revised form 2000 October 23) Abstract-The lunar soil characterization consortium, a group of lunar-sample and remote-sensing scientists, has undertaken the extensive task of characterization of the finest fractions of lunar soils, with respect to their mineralogical and chemical makeup. These compositional data form the basis for integration and modeling with the reflectance spectra of these same soil fractions. This endeavor is aimed at deciphering the effects of space weathering of soils on airless bodies with quantification of the links between remotely sensed reflectance spectra and composition. A beneficial byproduct is an understanding of the complexities involved in the formation of lunar soil. Several significant findings have been documented in the study of the <45 pm size fractions of selected Apollo 17 mare soils. As grain size decreases, the abundance of agglutinitic glass increases, as does the plagioclase, whereas the other minerals decrease. The composition of the agglutinitic glass is relatively constant for all size fractions, being more feldspathic than any of the bulk compositions; notably, is substantially depleted in the agglutinitic glass. However, as grain size decreases, the bulk composition of each size fraction continuously changes, becoming more Al-rich and Fe-poor, and approaches the composition of the agglutinitic glasses. Between the smallest grain sizes (10-20 and <1 Opm), the Is/ values (amount of total iron present as nanophase Feo) increase by greater than 100% (>2x), whereas the abundance of agglutinitic glass increases by only 10-15%. This is evidence for a large contribution from surface-correlated nanophase Feo to the Is/ values, particularly in the <10 pm size fraction. The surface nanophase Feo is present largely as vapordeposited patinas on the surfaces of almost every particle of the mature soils, and to a lesser degree for the immature soils (Keller et al., 1999a). It is reasoned that the vapor-deposited patinas may have far greater effects upon reflectance spectra of mare soils than the agglutinitic Feo. INTRODUCTION Reflectance spectroscopy has proven to be an extremely usell tool for remote compositional and mineralogical analyses of planetary surfaces. The foundation for remote mineralogical analysis lies in optical absorption physics (Bums, 1993) and the linking of spectral properties of materials measured in the laboratory to well understood mineral species and their mixtures. The pioneering works by Bums (1970), Hapke et al. (1970), McCord and Adams (1973), McCord et al. (1981) and others have amply demonstrated the potential of spectral reflectance measurements for lunar materials. For example, reviews by Pieters (1 986, 1993) showed how near-infrared spectroscopy has been used successfully to identify and map a variety of lunar rock types on relatively unweathered lunar surfaces. It is the comminuted regolith, however, that is mostly seen by remote-sensing instruments and not bedrock. The accurate estimation of rock and mineral compositions is complicated by the nature of the pervasive lunar soil (<1 cm fraction of the regolith), which contains the cumulative optical effects of "space weathering," including impact-produced glasses and agglutinates (aggregates of rock and mineral fragments bonded together by glass). As noted by McCord and Adams (1973) and described by Fischer and Pieters (1 994), the three principal optical manifestations of space weathering on lunar materials are: (1) overall reduction of reflectance; (2) general attenuation of diagnostic absorption bands; and (3) development of a redsloped continuum. These effects increase with soil maturity (ie., surface exposure). A principal cause for these variations is believed to be accumulation of nanophase Feo (native Fe of Meteoritical Society, Printed in USA.

2 286 Taylor et al. single-domain size, nm; Housley et al., 1973) particularly dominant in the finest fraction of the lunar soils. The relative abundance of this nanophase-size Feo has traditionally been used as a maturity index for lunar soils (Morris, 1976,1978). The concentration of nanophase Feo, as determined by ferromagnetic resonance (FMR) is expressed as Is. This is normalized to the total iron content of a soil, expressed as, resulting in the value Is/. Actually, it is this Is/ that is the maturity index for lunar soils. The normalization is necessary because the concentration of nanophase Feo is proportional to both the duration of surface exposure and the amount of available for reduction. By implication, the lunar-surface regions mentioned above where reflectance spectra have been most successful in determining mineralogical compositions are those where the effects of space weathering are minimal (Pieters, 1993). The nature of the weathering processes on the Moon, although markedly different from those on Earth, is also representative of those on relatively small, "airless" bodies in the solar system, such as asteroids (e.g., Vesta, Eros) and moons such as Phobos and Deimos, albeit with different dynamics. The energetics may be different at various distances from the Sun, thereby affecting the character and rates of formation of the soil. However, the overall effects of space weathering upon lunar reflectance spectra provides the basis for application of this important remote sensing technique to all airless bodies in general. Rationale for the Present Study Although bulk-soil properties are greatly altered by the effects of space weathering, mature lunar soils retain weak, yet distinct spectral signatures that are due to the inherent mineralogy of the dominant local lithology. However, it is the <45 pm particle-size domains of the lunar soil that are the most similar to and appear to dominate the spectral signature of the bulk soil (Pieters et al., 1993). This is partly because fine particles coat larger particles, and photons that enter large particles are unlikely to escape. Optically, the and pm size fractions are the most similar to the bulk soil (Fischer, 1995). Larger size fractions are not representative of bulk soil properties (Pieters et al., 1993), and the <10 pm fractions are most sensitive to surface-correlated processes (Fischer, 1995; Noble et al., 2000a; Pieters et al., 2000). The detailed petrographic properties of lunar soils, particularly the finer fractions are poorly known. Therefore, modern techniques are required to characterize soil compositions with the accuracy necessary for integration with spectroscopic analyses. To make a more direct and quantitative link between soil mineralogy and chemistry and soil spectral properties requires acquisition of several data sets. It is imperative to accurately measure and characterize the petrography of lunar soils, in terms relevant to remote analyses. This should be coupled with measurement of precise reflectance spectra and testing and use of appropriate analytical tools that identify and characterize the spectral effects of individual mineral and glass components (through spectral deconvolution of a mixture). Such a large-scale integrated operation has recently been established as the lunar soil characterization consortium (LSCC; Taylor et al., 1999a, 2000d). This group consists of lunar-soil and remote-sensing scientists that integrate detailed measurements of soil mineralogy and chemistry of the <45 pm size fractions with the reflectance-spectral characterization of these same fractions. The authors of this paper are the main members of this consortium. Herein, we report some of the new discoveries and understandings of lunar mare soils that have come from our initial efforts with Apollo 17 mare soils. METHODOLOGY The sieving, splitting, and distribution of the soil fractions for use by the several LSCC members is shown in Fig. 1. Notice that only "pristine" lunar soils were used for this major endeavor and the important role played by the planetary materials curatorial personnel at Johnson Space Center (JSC). The logistics of the many handlings of the soil splits, the immense amount of documentation, and the enormous quantities of paper work, as well as the preparation of polished grain mounts of the soils, were conducted in good fashion by this organization. Lunar Samples Four lunar mare soils were selected from the Apollo 17 sample collection based on similarities in chemical compositions and diversity in maturity, as reflected by their I@eO values (Moms, 1976). The mare soils are pristine samples of (immature); (submature); (submature); and (mature). They were obtained from NASA's planetary materials curation facility at Johnson Space Center. Lunar mare soils from Apollo 11, 12, and 15 are also being studied, and preliminary results were presented in Taylor et al. (2000a-d). Sieving All sieving operations were conducted in the Johnson Space Center laboratories of Dr. David McKay. Over the years, the general procedure of using Freon as a fluid medium for the sieving of lunar soils, particularly for fine-grain sizes, was in operation. Later, a fluid-free technique using the application of an ultrasonic sifter was employed instead. However, after considerable study and experimentation, it seems that the energetics of the ultrasonic sifting operation are sufficient to disrupt surfaces of soil particles, effectively "weather" them. These grains comminute and lose portions of their surface coatings to the finer sizes, as confirmed by Is/ values, which were anomalously high for the <10pm fraction (Noble et al., 2000a). After evaluating several sieving methods (Noble et al., 2000a), the JSC group determined that triply distilled water, although not as efficient as Freon for sieving fine fractions

3 The effects of space weathering on Apollo 17 mare soils 287 Lunar Sample Collection (Pristine Lab) McKav / Wentworth 10 mg 1-2 mg SEWTEM FIG. 1. Flow sheet for lunar sample distribution from the lunar sample curatorial facilities (at Johnson Space Center) to the lunar soil characterization consortium. due to increased surface tension, is an effective sieving fluid medium. It leaves no surface residue, particularly if the total wash collected in the bottom pan is largely decanted off. Therefore, the Apollo 17 mare soils were sieved with triply distilled water to obtain a <45pm size fraction. AAer removing material for analyses (Fig. l), the <45 pm fraction was further sieved into three finer size fractions: 20-45, 10-20, and 40pm size fractions. Portions of each size separate were distributed for analyses to each laboratory of the lunar soil characterization consortium. Chemical Analyses The bulk chemistry of the separated size fractions was determined by electron microprobe analyses of fused-glass beads. A 5-mg representative split of each size fraction was fused on a Mo-strip heater, bathed in pure nitrogen, to minimize possible oxidation. The samples are easily melted with annealing times of s; less than 0.5% Mo has been determined to dissolve in this melt. This fused glass was then mounted in epoxy and polished for EMP study, where an average composition was obtained from at least 15 analyses, using a 20 pm spot size. This technique is commonly used for determination of bulk compositions of lunar samples, which are particularly suited to this technique due to their general lack of volatiles (Schuraytz and Ryder, 1990). Modal and Digital-Imaging Analyses Since the return of the first lunar samples, standard operating procedure for soil petrography has been to characterize a lunar soil by particle counting (e.g., Heiken and McKay, 1974; Labotka et al., 1980; Simons et al., 198 1). Such analyses provide detailed information about the abundances ofmineral and lithic fragments, volcanic/impact-produced glasses, and glass-bonded aggregates, called agglutinates. These data are interpreted in terms of in situ soil development, as well as additions from elsewhere. This type of soil petrographic analysis has proven highly successful in studying the petrogenesis and provenance of lunar soils (e.g., Simons etal., 1981). These particle-count data, however, do not provide information on the actual percentages of minerals (modes) locked in lithic fragments and fused-soil particles (e.g., agglutinates). But the actual amounts of the various minerals and glasses in the soil that interact with solar radiation are the important input data for the proposed remote compositional analysis and space weathering study. Particle counting simply

4 288 Taylor et al, involves classifying a soil fragment with a title (e.g., pyroxene, basalt, breccia, agglutinate). Modal analysis, sensu stricto, is defined as the volume percentages (or wt%) of the mineral constituents, not the particle type. In order to develop quantitative links with spectroscopic mineral assessment, it is essential that accurate quantitative modal analyses of the components of lunar soils be obtained. This is accomplished using the techniques detailed and illustrated by Taylor etal. (1 993,1996), Chambers et al. (1994a,b, 1995), and Higgins et al. (1994, 1995, 1996). Dr. Larry Taylor's group at the University of Tennessee has established the software and chemicallshape parameters with which to perform x-ray digital-imaging analyses on grain mounts of lunar soils, thereby producing accurate modal analyses. The details of this technique are given in Taylor et al. (1 996), and the analyses were performed using an Oxford instrument energy dispersive spectrometer (EDS) unit on a Cameca SX-50 electron microprobe. In addition, it is possible to determine the average chemical composition of each phase (e.g., pyroxene, plagioclase, hi-ti glass, hi-a1 agglutinitic glass). Most of the previous studies at the University of Tennessee (e.g., Taylor et al., 1996) have involved the pm size fractions of various lunar soils. Grain mounts of the magnetic splits (ie., agglutinitic-rich portions) from the <45 pm fraction of soil have also been examined using the x-ray digitalimaging technique (Taylor et al., 1996). At such fine grain sizes, many of the agglutinates have lost much of their diagnostic vesicular texture and have been largely broken into their individual mineral and glass components. It is the amount of this impact-produced glass, with its nanophase Feo, which is important for evaluations of space weathering. Optical examination would overlook many of the agglutinitic glass fragments because they do not have certain agglutinate characteristics, such as the presence of vesicles. However, with the fine-tuning of the x-ray imaging parameters, it has been possible to identify agglutinitic glass, with reasonable precision, entirely by its composition. Thus, it is now possible to fully quantify lunar soils with respect to: (1) volume percentages of particle types (e.g., lithic fragments, minerals, agglutinates); (2) volume percentages of different mineral and glass phases (e.g., pyroxene, agglutinitic glass); and (3) average chemical compositions of the different mineral and glass phases. Ferromagnetic Resonance Analyses Measurements of the abundances of nanophase Feo (abbreviated as the intensity of "single-domain" Fe; IS) were performed in the Magnetics Lab of Dr. Richard Morris at Johnson Space Center. The procedures and instrumentation employed for these ferromagnetic magnetic resonance (FMR) measurements are the standard ones that have been applied in obtaining Is values for all measured lunar samples. This laboratory has exclusively performed Is measurements on lunar samples since about 1972 (Morris, 1976). Reflectance Spectroscopy Several approaches to quantifying mineralogical information derived from reflectance spectra have made significant progress, as described in Pieters (1 986, 1993) and Pieters and Taylor (1 988). Several of these approaches are important to lunar science issues and applications: (1) spectral mixing analysis (linear and nonlinear, or Hapke-modeling ofhow spectra ofdiverse components combine in a mixture); (2) modified Gaussian model (MGM) deconvolution (modeling individual mineral absorption bands separately and in a mixture); (3) space-weathering normalization (modeling the optical effects that alter diagnostic properties); and (4) empirical statistical methods that isolate diagnostic properties. These different approaches have been reviewed by Pieters et al. (1 996). For mineralogical assessment of lunar soils, the most important issues are: (a) accurately accommodating the effects of space weathering and (b) linking residual diagnostic absorption features with a quantitative measure of mineral abundance. The modal and chemical data in this present study have been integrated with the Apollo 17 soil reflectance spectra and preliminary results are presented in Noble et al. (2000a,b) and Pieters et al. (2000). MODAL ABUNDANCES The modal percentages of mineral and glass components of the four Apollo 17 mare soils were determined by digital-imaging x-ray analyses of polished grain mounts of each size fraction. The results are shown graphically in Fig. 2, with the details given in Table 1. The bulk-soil maturities (Is/) and and values are also shown in Fig. 2, for general comparison between the four soils. In past years, several authors (e.g., Labotka et al.,1980; Simons et af., 1981; Fischer, 1995) have stated that the abundances of agglutinates decrease as grain size decreases. However, it is the abundance of the agglutinitic glass that is at the crux of this issue, not the overall agglutinates which may lose their large-scale identities, a concept originally put forth by Hu and Taylor (1 977). Modal analyses of the phases in the soils permit us to address the question of the abundances of agglutinitic glass as a fimction of grain size. The overall suite of Apollo 17 mare soils have comparable contents; therefore, the concentration of nanophase Feo increases with overall soil maturity, as does the modal percentage of agglutinitic glass (Fig. 2). However, for individual soils, the most striking result is the increase in impactproduced agglutinitic-glass content with decreasing grain size. In several soils we have examined, there is >loo% increase in the agglutinitic glass contents between the and the 10-20pm size fractions (Taylor et al., 1996, 1997, 1998, 1999qb). The modal abundance of crystalline plagioclase is relatively constant to slightly increasing, with decreasing grain size of the fractions. However, the abundances of all other components (pyroxene, oxides, volcanic glass, and olivine) decrease with particle diameter. This decrease, at least in part, results from closure

5 The effects of space weathering on Apollo 17 mare soils 289 Cumdative Modal Percentage of Apollo 17 Mare Soil Components (<10-45 pm) = = I Others Volcanic GI= Olivine llmenite Plagioclase Pyroxene Agglutinitic Glass <I <I <I <I pm pm pm pm FIG. 2. Modal analyses of the finest size fractions of Apollo 17 Mare Soils, with the least mature soil (i.e., 71061) on the left and the most mature on the right (ie., 79221). The small number after the sample number (e.g., ) is the Isi value ofthe <250pm fraction. Note the general increase in agglutinitic glass and decrease in pyroxene content as maturity increases between soils, as well as the same trends within a given soil with decrease in grain size. TABLE 1. Modal abundance of minerals and glasses in finest size fractions of Apollo 17 mare soils.* pm pm <10pm pm pm <I0 pm Oxide minerals 7.3 Plagioclase 16.9 Pyroxene 13.2 Olivine 4.8 Agglutinitic glass 46.8 Volcanic glass 10.9 Others pm pm <10pm pm 10-20pm <10pm Oxide minerals Plagioclase Pyroxene ivine Agglutinitic glass Volcanic glass Others *Maturity as Is/ of the <250 pm fraction is given directly after the soil number.

6 ~ ~ ~ ~ ~ 290 Taylor et al. (ie., the component sum is loo%), although the decrease for pyroxene is pronounced and real. This decrease in pyroxene with decreasing grain size appears very significant since pyroxene is probably the most optically active of the lunar minerals. Another important observation is that ifthe secondary agglutinitic products are eliminated, the abundance of plagioclase increases relative to the mafic minerals with decreasing particle size. Cintala and Horz (1 992) experimentally verified this observation of selective comminution. As shown in Fig. 3 and detailed in Table 2, it was possible to further dissect the pyroxenes into four chemical groups-opx, Pig, Mg-CPX, and Fe-CPX. This is important since the reflectance spectra for orthopyroxene and clinopyroxene are distinctly En Fs FIG. 3. Abundances offour pyroxene modes in the finest size fractions of mare soil (e.g., Opx, Pigs, Mg-Cpx, and Fe-Cpx). The three numbers above each envelope represents the modal percentages of each pyroxene, from left to right, of the 2W5, 10-20, and <10 pm size fractions. TABLE 2. Modal percentages of four sub-sets of pyroxenes in the finest size fractions of Apollo 17 mare soils.* pm pm <10pm Orthopyroxene Pigeonite Mg-Clinopyroxene Fe-Clinopyroxene pm pm <10pm Orthopyroxene Pigeonite Mg-Clinopyroxene Fe-Clinopyroxene pm pm <10pm Orthopyroxene Pigeonite Mg-Clinopyroxene Fe-Clinopyroxene pm pm <10pm Orthopyroxene Pigeonite Mg-Clinopyroxene Fe-Clinopyroxene *Modal percentages as measured are normalized in the second column. Maturity as Is/ of the <250 p m fraction is given directly after the soil number.

7 The effects of space weathering on Apollo 17 mare soils * = 16.2 TiO2= * * * I0 12 I0 8 6t- R Ti 0 2 I * Is/ of ~ 250 I-Lm W <lo <lo <lo <I prn prn prn pm FIG. 4. Chemistry of the finest size fractions of Apollo 17 mare soils. The soil data are presented from left to right from the immature soil (ie., 71061) to the most mature soil (ie., 79221). The data for a given soil are presented from left to right for decreasing grain size of the soil fractions. With decreasing grain size, note the systematic increase in plagioclase components (,, Na2O) and decrease in ferromagnesian-mineral components (,, ). Also notice the increase in Is/ with soil maturity, as well as the huge increases with decreasing grain size. different, and the positions of absorption features of clinopyroxene spectra are a hction of chemistry. Furthermore, inasmuch as pyroxene in the lunar soils is the most spectrally active of the minerals in the lunar soil, it is important that this subdivision be made in order to be useful in the refinement of the reflectance spectra (see Noble et al., 2000b). SOIL CHEMISTRY The four Apollo 17 mare soils were chosen for study because of similarities in bulk and, contents, allowing for direct comparisons of evolutions of changes in chemical composition between size fractions and different soil maturities. With decreasing grain size,,, and concentrations decrease, and,, and (plagioclase components) increase for all soils (Fig. 4). These chemical variations parallel increases in agglutinitic glass and decrease in oxides (mainly ilmenite), pyroxene, olivine, and volcanic glass (Fig. 2). Changes in soil chemistry with particle size are not caused by distinct changes in individual phase compositions; instead, changes in the chemical compositions of the soil are caused mainly by the abundances of the phases (e.g., pyroxene, agglutinitic glass). In contrast to the systematic chemical variations observed in bulk chemistry for different size fractions (Table 3), the

8 292 Taylor et af. agglutinitic glass compositions (Table 4) remain relatively constant with changes in grain size and soil maturity. The systematic changes in soil composition with decreasing grain size reveals that the bulk fraction of each size fraction becomes more feldspathic with increasing maturity, with the effect being most pronounced in the finest fractions (Fig. 5). Notice that the composition of the agglutinitic glass, however, is relatively invariant and even more feldspathic (ie., rich in Al2O3) than even the <I Opm fraction. Thus, it would appear that the bulk chemical compositions of the soil fractions, with decreasing grain size, approach the composition of the agglutinitic glasses. This relation not only strengthens the "fusion of the finest fraction" (F3) hypothesis (Papike et af., 1981; Walker and Papike, 1981), but also highlights the important role of plagioclase in the formation of agglutinitic glass. The F3 model is based upon the selective crushing of basalts such that the plagioclase feldspar is enriched in the finest fractions of the soil, relative to the olivine/pyroxene components; and it is this finest fraction that is most readily melted by micrometeorite impact (Cintala and Horz, 1992). MINERAL AND GLASS CHEMISTRY The composition of the several mineral and glass phases present in the soil fractions was determined by the x-ray digital analyses of each soil split (Table 4). The most impressive observation from all these analyses is that the composition of the agglutinitic glass does not appear to change as a function TABLE 3. Bulk chemistry and Is/ values of the finest size fractions of Apollo 17 mare soils.* Size range <45 pm 20-45pm 10-20pm <10pm <45 pm 2045pm 10-20pm <IOpm (sample #) (,120) (, 129) (, 134) (, 139) (, 149) (,158) (,I631 (, 168) MnO NiO p205 so2 IS/ 41.7(2) 6.39( 7) 13.5(1) 0.37(4) 10.3( 1) 10.8( 1) 0.2 1(3) 14.0(2) 0.41(1) ( 1) 0.07(3) 0.19(3) I I) 7.38( 11) 11.6(9) 0.40(4) 10.9(5) 10.3(3) 0.22(2) 15.8(9) C (3) ( 1) (2) 0.17(4) (2) 7.21(8) l2.9( 1) 0.40(3) l0.4( I) 10.4(1) 0.20(2) 15.0(2) 0.39( 1) O.lO(1) 0.07(3) 0.19(2) (2) 5.83(5) 15.9( 1) 0.35(3) 9.59(7) 11.7( 1) 0. I7(3) 11.3(2) 0.49( 1) (1) 0.07(4) 0.17(3) (2) 7.57( 12) 12.4(1) 0.42(3) 10.1(1) l0.6( 1) 0.21(3) 15.3(2) 0.39( 1) (2) 0.07(4) 0.17(3) ( 1) 8.11( 13) 11.5( 1) 0.43(4) 10.1(1) 10.3( 1) 0.22(3) I6.0(2) 0.35(2) ( 1) (3) 0.16(3) ( 1) 7.88( 7) 12.7(1) 0.42(3) 9.97(8) 10.4( 1) 0.23(2) 15.5(2) 0.34(2) ( 1) (3) (2) (1) 6.54(8) 15.4( 1) 0.39(3) 9.12(5) 11.5(1) 0.18(2) 12.7(2) 0.46(2) 0.13(1) 0.1 O(4) 0.20(4) Size range <45 pm 204 pm pm <lopm 145 pm 20-45pm 10-20pm <lopm (sample #) (,241) (,250) (255) (260) (, 152) (,161) (, 166) (,171) MnO NiO p205 so2 I@eO 39.7(4) 9.3 l(45) 11.3(2) 0.42(5) 9.73( 19) 10.2( 1) 0.22(3) 16.5(2) 0.38(2) ( I ) 0.07(3) 0.16(3) (11) 0.7(10) 9.94(68) 0.46(6) 9.97(2 1) 9.94(31) 0.24(3) 17.8(8) 0.35(2) 0.07( I ) 0.07(3) 0.17(3) (2) 9.83(22) 11.6(1) 0.45(5) 9.52(8) 10.1(1) 0.23(3) 16.4(2) 0.39(2) ( I ) (2) 0.19(4) ( 1) 8.27(8) 14.5(1) 0.40(3) 8.76( 6) 11.2( 1) 0.19(3) 13.5(2) 0.42(2) 0.1 I ( I) (3) 0.25(3) ( 1) 8.76( 9) 10.5(1) 0.48(8) 1031) 9.90(9) 0.24(2) 17.5(2) 0.41(3) ( 1) (2) (3) (3) 9.48(20) 9.33( 13) 0.48(3) 10.8(2) 9.58( 10) 0.23(2) 18.5(2) 0.34(4) 0.07(2) 0.04(3) 0.17(5) (2) 8.94( 12) 10.8(1) 0.48(3) I0.4( 1) 9.79(9) 0.23(2) 17.5(1) 0.40(3) (2) (3) 0.20(5) (3) 7.89( 10) 13.8(1) 0.44(3) 9.18(8) 10.7( 1) 0.2 O( 2) 14.8(4) 0.46(3) 0.11(1) (4) 0.25(4) *The chemistry was determined by EMP analyses of fused beads of the soil. Values of Is/ are from FMR analyses. Maturity as Is/ of the 1250 pm fraction is given directly after the soil number. Soils are listed from mature (79221) to immature (71061). The number in parentheses represents the la range of variation in the chemistry expressed as the least digit cited.

9 The effects of space weathering on Apollo 17 mare soils 293 TABLE 4. Average compositions of minerals and glasses in the finest fractions of Apollo 17 mare soils.* Plag Ilm Olivine Vol Gls. Agglut. GIs. Opx Pig Mg-Cpx Fe-Cpx Na2O (20-45 pm) Cf203 Na2O K2O (10-20pm) Na2O (<lo pm) I (20-45 pm) Na2O (10-20 pm) c

10 294 Taylor et GI. TABLE 4. Continued* Plag Ilm Olivine Vol GIs. Agglut. Gls. Opx Pig Mg-Cpx Fe-Cpx SiOz (40 pm) k SiOz (20-45 pm) SO oo Mi (10-20pm) ( SiO (<lo pm) c SiOz Na2O (20-45 pm) I

11 ~ The effects of space weathering on Apollo 17 mare soils 295 TABLE 4. Continued* Plag Ilm Olivine Vol GIs. Agglut. GIs. Opx Pig Mg-Cpx Fe-Cpx (10-20 pn) oo (40prn) *Maturity as Is/ ofthe <250pm fraction is given directly after the soil number. Abbreviations: Plag = plagioclase; Ilm = ilmenite; Vol. GIs. = volcanic glass; Agglut. GIs. = agglutinitic glass; Opx = orthopyroxene; Pig = pigeonite; Mg-Cpx = high magnesium clinopyroxene; Fe-Cpx = high iron clinopyroxene. of grain size in the soils, as discussed above and shown in Fig. 5. However, another important observation of the chemistry of agglutinitic glass is important. It was suggested that the composition of agglutinitic glass in a soil mimics, to large extent, the chemistry of the bulk soil, as originally suggested by Hu and Taylor (1977). The compositions of the agglutinitic glasses analyzed here, at least for the grain sizes <45 pm, the glass compositions are more feldspathic than the overall bulk chemistry of the soil. Thus, the composition of the agglutinitic glass is a fair representation of the bulk chemistry plus a feldspathic component. This is basically true for all components, except for one-. For all the various fine size fractions of the four high- mare soils, the compositions of the agglutinitic glass are significantly low in, relative to that of the bulk-soil chemistry (compare Tables 3 and 4). This would seem to indicate that either (1) ilmenite does not readily comminute to the finest grain sizes involved in the formation of the agglutinitic melt, or (2) the ilmenite that is present in the finest fractions does not undergo the complete melting that the silicate minerals do. Since the abundance of ilmenite appears to mirror the abundance of mafic minerals, this anomalous lack of in agglutinitic glass probably relates to differences between the physical properties of oxide and silicate minerals, that retards ilmenite from entering the melt. Is/ VALUES The relative abundance of nanophase Feo, expressed as IS, was determined by ferromagnetic resonance. The value Is/, where is the total iron content of a soil size fraction is the relative proportion of the total iron in a sample that is present as nanophase Feo. Is/ is not necessarily related to the agglutinitic glass content because the concentrations of nanophase Feo in the glass need not be constant. Values of Is/ increase with decreasing grain size, even though the bulk contents are decreasing (Fig. 4). That is, the percentage of the total iron that is present as nanophase Feo has increased substantially in the smaller size fraction. Note that the increase in nanophase Feo in smaller size fractions is significantly greater than the increase in agglutinitic glass content. This would seem to indicate either that at least some of the Feo is surface correlated or that the nanophase Feo concentration in agglutinitic glass increases or both. For the latter, the composition of the agglutinitic glass, regardless of grain size, is relatively constant. Thus, it is unlikely that any other property of this glass systematically changes. However, if it is assumed that the nanophase Feo is entirely surface correlated, then equal masses of 15 and 6 pm spherical soil particles should have -3x as much Feo in the finer fraction due to their difference in surface to volume ratio. The increase in Is/ is about 2x, indicating the probability that most of the nanophase Feo is surface-

12 296 Taylor et al Q) LL I0 - FIG. 5. Comparisons of changes in the bulk chemistry (in wt%) ofthe three finest size fractions of the Apollo 17 soil versus the compositions of their agglutinitic glasses. The soil symbols, presented from top to bottom, are for immature soil (ie., 71061) to the most mature soil (i.e., 79221). The small number after the sample number (e.g., ) is the Is/ value ofthe <250pm fraction. Note the bulk compositions of decreasing size fractions within a given soil converge on the composition of the agglutinitic glass. correlated (e.g., Taylor et al., 1999a). The surface area of size fractions increases greatly with decreasing particle size. For example, the surface-aredvolume ratio for the pm size fraction is over 8-fold greater than for the pm size fraction. The idea that the space-weathering products, which strongly modify the spectral effects, may be surface correlated was also suggested by Pieters et al. (1 993). This was based on the observations that the spectral properties of a fine-fraction separate of a lunar soil are unique; they could not be reproduced by artificially preparing (by grinding) a fine fraction from an agglutinate-rich larger-size fraction of the same soil. VAPOR-DEPOSITED COATINGS There is a significant major, contribution from space weathering products on the surface of many soil particles (Keller and McKay, 1993,1997; Taylor et al., 2000a,b,c). Recent findings have been described of the major role of vapor-deposited, nanophase Feo-containing patinas on most soil particles (Keller and McKay, 1993, 1997; Keller et al., 1998, 1999a,b; Taylor et al., 1999a,b, 2000c; Wentworth et al., 1999). This is a major "break-through" in our understanding of the distribution of between agglutinitic glass and condensed vapor coatings (patinas) upon particle surfaces. It is micrometeorite impacts, and/or sputtering due to impacting of high-energy particles, that cause the volatilization of soil grains with the subsequent condensation ofthis vapor as thin patinas (<1 pm) on soil particles. It should be pointed out, however, that such surface-deposited material was originally predicted by Hapke et al. (1 975). From backscattered electron (BSE) images and x-ray maps of the pm split of mare soil 79921, it is obvious that each grain possesses a thin-rind or patina consisting mainly of Si- and Al-rich glass. But this glass also contains abundant nanophase Feo (Fig. 6). These patinas were mostly formed by deposition of a Fe-, Si-, Al-rich vapor created by the extreme temperatures reached during impacting processes of micrometeorites and commonly display overlapping layers of patinas. Sputter-deposited contributions also occur, although in lesser quantities (Bematowicz et al., 1994). The plagioclase, as shown by the Fe Ka x-ray map (Fig. 6), contains these thin (<1 pm) coatings of glass, but with noticeably higher Fe content than the plagioclase interior, which only has <0.3%. In

13 The effects of space weathering on Apollo 17 mare soils 297 FIG. 6. Backscattered electron (BSE) and x-ray elemental maps of 10-20pm fraction of mature mare soil The BSE map of the left represents the different chemistry of each mineral (the brighter the color, the higher the average atomic number of the phase). The Fe Ka and Al Ka x-ray maps reflect the varying degrees of Fe and Al contents, respectively, in the minerals. Note the Fe-rich rims on the plagioclase (dark interior [<0.3% Fe] with faint light rim) and Al-rich rim on the ilmenite (the interior of which has no Al) as a result of vapor deposition. The small number after the sample number (e.g., ) is the Is/ value of the 1250pm fraction. addition, by reference to the A1 Ka x-ray map (Fig. 6), an Al-rich rim can be seen on the ilmenite grain (FeTi03) that does not contain appreciable A1 (<0.4%). Detailed high-resolution transmission electron microsope (HRTEM) examinations by Keller et al. (1999a,b; 2000) have verified patinas of this composition on almost every soil particle in this mature soil. SUMMARY The collaborative efforts of the lunar soil characterization consortium have determined the chemical and physical properties of the fine-size fractions of four Apollo 17 soils. These studies have important and highly significant ramifications for understanding lunar soil formation processes and remotely obtained reflectance spectra. For each Apollo 17 mare soil: As grain size decreases, the amount of agglutinitic glass increases, as does the plagioclase, whereas the other minerals decrease. The agglutinitic glass composition is relatively constant for all size fractions, being more feldspathic than any of the bulk compositions; TiO2, however, is significantly depleted in agglutinitic glass. As grain size decreases, the bulk composition of each size fraction continuously changes, becoming more Al-rich and Fe-poor, and approaches the composition of the agglutinitic glasses. The Is/ values increase by greater than 100% (>2x) between the smallest grain sizes (10-20 and <10 pm), whereas the abundance of agglutinitic glass increases by only 10-15%. 0 There is evidence for a large contribution from surfacecorrelated nanophase Feo to the Is/ values, particularly in the <10 pm size fraction. 0 The surface nanophase Feo is present as vapor-deposited patinas on the surfaces of almost every particle of the mature soils, and to a lesser degree for the immature soils (Keller et al., 1999a). 0 The vapor-deposited patinas possibly have far greater effects upon reflectance spectra of mare soils than the agglutinitic Feo (Keller et al., 2000). The scientific potential from having finely tuned, highly accurate, compositional and mineralogical data for lunar soils is immense. Requisite towards improving and expanding the interpretations of reflectance spectra of lunar maria is coupling of these data with the highly diagnostic mineral absorption features. The Apollo 17 mare soil studies presented above have revealed several additional complexities in the soils that will be integrated with the reflectance spectroscopy in subsequent studies. Acknowledgements-Dr. Chuck Meyer, Acting Lunar Curator at the time, is acknowledged for his perceptive recognition that, originally, we had been studying "degraded" lunar soils. These samples had been previously allocated to other investigators and subsequently returned and placed in the RSPL. The previous studies had degraded some of their physical and chemical properties, rendering them useless for our studies. In addition, the Lunar Sample Curatorial Staff is thanked for the distribution of the many size fractions and for preparing the numerous grain mounts, which were generally excellent. Dawn Taylor, with her infinite patience and creative powers, produced the fine figures and tables. The original manuscript benefited considerably from the extensive and constructive reviews of B. Ray

14 298 Taylor et al. Hawke and Grant Heiken, as well as from the review and careful editing of Mike Gaffey. Last, but certainly not the least, this research has received support from the NASA grants to each ofthe Lunar Soil Characterization Consortium members, and for this we are collectively grateful. Editorial handling: M. J. Gaffey REFERENCES BERNATOWICZ T. J., NICHOLS R. H., JR. AND HOHENBURG C. M. (1994) Origin of amorphous rims on lunar soil grains (abstract). Lunar Planet. Sci. 25, BURNS R. G. (1970) Mineralogical Application of Crystal Field Theory. Cambridge Univ. Press, London, U.K. 224 pp. BURNS R. G. (1993) Mineralogical Application of Crystal Field Theory, Second Ed. Cambridge Univ. Press, London, U.K. 551 pp. CHAMBERS J. G., TAYLOR L. A., PATCHEN A. AND MCKAY D. S. (1994a) Mineral liberation and beneficiation of lunar high-ti mare basalt, 71055: Digital-imaging analyses. In Engineering, Construction, and Operations in Space ZV (eds. R. G. Galloway and S. Lokaj), pp American Society of Civil Engineers, New York, New York, USA. CHAMBERS J. G., PATCHEN A., TAYLOR L. A., HIGGINS s. J. AND MCKAY D. S. (l994b) Lunar mineral feedstocks from rocks and soil: X-ray digital-imaging in resource evaluation (abstract). Lunar Planet. Sci. 25, CHAMBERS J. G., TAYLOR L. A,, PATCHEN A. AND MCKAY D. S. (1995) Quantitative mineralogical characterization of lunar high- Ti mare basalts and soils for oxygen production. J. Geophys. Res. 10, CINTALA M. J. AND HORZ F. (1992) An experimental evaluation of mineral-specific comminution. Meteoritics 27, FISCHER E. M. (1995) Quantitative Compositional Analysis of the Lunar Surface from Reflectance Spectroscopy: Iron, Aluminum and a Model for Removing the Optical Effects of Space Weathering. Ph.D. thesis, Brown University, Providence, Rhode Island, USA. 195 pp. FISCHER E. M. AND PIETERS C. M. (1994) Remote determination of exposure degree and iron concentration of lunar soils using VIS-NIR spectroscopic methods. kurus 111, HAPKE B. W., COHEN A. J., CASSIDY w. A. AND WELLS E. N. (1970) Solar radiation effects on the optical properties of Apollo 11 samples. Proc. Apollo I1 Lunar Sci. Conf: 3rd, HAPKE B., CASSIDY W. A. AND WELLS E. N. (1975) Effects ofvaporphase deposition processes on the optical, chemical, and magnetic properties of the lunar regolith. Moon 13, HEIKEN G. H. AND MCKAY D. S. (1974) Petrography of Apollo 17 soils. Proc. Lunar Sci. Conf 4th, HIGGINS. J., PATCHEN A., CHAMBERS J. G., TAYLOR L. A. AND MCKAY D. S. (1994) Petrographic characterization of lunar soils: Application of X-ray digital-imaging to quantitative and automated analysis (abstract). Lunar Planet. Sci. 25, HIGGINS. J., TAYLOR L. A,, PATCHEN A., CHAMBERS J. G. AND MCKAY D. S. (1995) X-ray digital-imaging analysis: Requisite data for spectral reflectance studies (abstract). Lunar Planet. Sci. 26, HIGGINS. J., TAYLOR L. A., CHAMBERS J. G., PATCHEN A. AND MCKAY D. S. (1996) X-ray digital-imaging petrography: Technique development for lunar mare soils. Meteorit. Planet. Sci. 31, HOUSLEY R. M., GRANT R. W. AND PATON N. E. (1973) Origin and characteristics of excess Fe metal in lunar glass welded aggregates. Proc. Lunar Sci. Con$ 4th, HU H. N. AND TAYLOR L. A. (1977) Lack of chemical fractionation in major and minor elements during agglutinate formation. Proc. Lunar Planet. Sci. ConJ8th, KELLER L. AND MCKAY D. S. (1993) Discovery of vapor deposits in the lunar regolith. Science 261, KELLER L. P. AND MCKAY D. S. (1997) The nature and origin of rims on lunar soil grains. Geochim. Cosmochim. Acta 61, KELLER L. P., WENTWORTH s. J. AND MCKAY D. S. (1998) Space weathering: Reflectance spectroscopy and TEM analysis of individual lunar soil grains (abstract). Lunar Planet. Sci. 29, #1762, Lunar and Planetary Institute, Houston, Texas, USA (CD-ROM). KELLER L. P., WENTWORTH S., MCKAY D. S., TAYLOR L. A,, PIETERS C. M. AND MORRIS R. V. (1999a) Space weathering in the fine size fraction of lunar soils: Soil maturity effects. In New Views of the Moon ZZ. Flagstaff, Lunar Planet. Inst. 980, KELLER L. P., WENTWORTH S., GEZO J., MCKAY D. S., TAYLOR L. A., PIETERS c. M. AND MORRIS R. V. (1999b) Space weathering alteration of lunar soil grains (abstract). LunarPlanet. Sci. 30, #1820, Lunar and Planetary Institute, Houston, Texas, USA (CD-ROM). KELLER L. P., WENTWORTH S., MCKAY D. S., TAYLOR L. A,, PIETERS c. M. AND MORRIS R. V. (2000) Space weathering in the fine size fraction of lunar soils: Mare/highland differences (abstract). Lunar Planet. Sci. 31, #1655, Lunar and Planetary Institute, Houston, Texas, USA (CD-ROM). LABOTKA T. C., KEMPA M. J., WHITE C., PAPIKE J. J. AND LAUL J. C. (1980) The lunar regolith: Comparative petrology of the Apollo sites. Proc. Lunar Planet. Sci. ConJ llth, MCCORD T. B. AND ADAMS J. B. (I 973) Progress in remote optical analysis of lunar surface composition. Moon 7, MCCORD T. B., CLARK R. N., HAWKE B. R., MCFADDEN L. A., OWENSBY P. D., PIETERS C. M. AND ADAMS J. B. (1981) Moon: Near-infrared spectral reflectance, a first good look. J. Geophys. Res. 86, MORRIS R. V. (1976) Surface exposure indices of lunar soils: A comparative FMR study. Proc. Lunar Planet. Sci. Con$ 7th, MORRIS R. V. (1978) The surface exposure (mature) of lunar soils: Some concepts and Is/ compilation. Proc. Lunar Planet. ConJ9th, NOBLE S. K., PIETERS C. M., TAYLOR, L. A,, Moms R. V., ALLEN C. c., MCKAY D. S. AND KELLER L. P. (2OOOa) Optical properties of the finest fraction of lunar soil: Implications for space weathering environments (abstract). Lunar Planet. Sci. 31, #1880, Lunar and Planetary Institute, Houston, Texas, USA (CD-ROM). NOBLE S. K., PIETERS C. M., HIROl T., TAYLOR L. A., MORRIS R. V., KELLER results of MGM analysis on Apollo 17 soil suite (abstract). Lunar Planet. Sci. 31, #1810, Lunar and Planetary Institute, Houston, Texas, USA (CD-ROM). L. P., MCKAY D. s. AND WENTWORTH s. (2000b) Initial PAPIKE J. J., SIMON S. B., WHITE C. AND LAUL, J. C. (1981) The relationship ofthe lunar regolith <1 Opm fraction and agglutinates. Part I: A model for agglutinate formation and some indirect supportive evidence. Proc. Lunar Planet. Sci. Conf: 12th, PIETERS C. M. (1986) Composition of the lunar highland crust from near-infrared spectroscopy. Geophys. Rev. 24, PIETERS C. M. (1993) Compositional diversity and stratigraphy of the lunar crust derived from reflectance spectroscopy. In Remote Geochemical Analysis: Elemental and Mineralogical Composition (eds. C. M. Pieters and P. A. J. Englert), pp Cambridge University Press, New York, New York, USA. PIETERS C. M. AND TAYLOR G. J. (1988) Millimeter petrology and kilometer mineralogical exploration of the Moon. Proc. Lunar Planet Sci. Conf 19th,

15 The effects of space weathering on Apollo 17 mare soils 299 PIETERS C. M., FISCHER E. M., RODE 0. AND BASU A. (1993) Optical effects of space weathering: The role of the finest fraction. J. Geophys. Res. Planets 98, PIETERS C. M., MUSTARD J. F. AND SUNSHINE J. M. (1996) Quantitative mineral analyses of planetary surfaces using reflectance spectroscopy. In MineralSpectroscopy: A Tribute to Roger G. Burns (eds. M. D. Dyar, C. McCammonand M. W. Schafer), pp The Geochemical Society, Houston, Texas, USA. PIETERS C. M., TAYLOR L. A., NOBLE S. K., KELLER L. P., HAPKE B., MORRIS R. v., ALLEN c., MCKAY D. s. AND WENTWORTH s. (2000) Space weathering on asteroids: Resolving a mystery with lunar samples. Meteorit. Planet. Sci. 35, SCHURAYTZ B. C. AND RYDER G. (1990) An evaluation ofthe reliability and usefulness of microprobe fused bead analyses for petrogenetic interpretations (abstract). Lunar Planet. Sci. 21, SIMONS. B., PAPIKE J. J. AND LAUL J. C. (1981) The lunar regolith: Comparative studies of the Apollo and Luna sites. Petrology of soils from Apollo 17, Lunar 16,20, and 24. Proc. Lunar Planet. Sci. Con$ 7th, TAYLOR L. A., CHAMBERS J. G., PATCHEN A., JERD E. A., MCKAY D. S., GRAF J. AND ODER R. R. (1993) Evaluation of lunar rocks and soils for resource utilization: Detailed image analysis of raw material and beneficiated products (abstract). Lunar Planet. Sci. 24, TAYLOR L. A., PATCHEN A., TAYLOR D-H., CHAMBERS J. G. AND MCKAY D. S. (1996) X-ray digital imaging and petrography of lunar mare soils: Data input for remote sensing calibrations. Icarus 124, TAYLOR L. A., PIETERS c., PATCHEN A., WENTWORTH s. AND MCKAY D. S. (1997) Spectral reflectance versus abundances of minerals and glasses in the 10 to 45 micron size fraction of mare soil (abstract). Lunar Planet. Sci. 28, TAYLOR L. A., PlETERS c. M., PATCHEN A., TAYLOR D-H., WENTWORTH S. AND MCKAY D. S. (1998) Optical properties and abundances of minerals and classes in the 10 to 44 micron size fraction of mare soils: Part I (abstract). Lunar Planet. Sci. 29, #I 160, Lunar and Planetary Institute, Houston, Texas, USA (CD-ROM). TAYLOR L. A. PlETERS C. M., MORRIS R. V., KELLER L. P., MCKAY D. S., PATCHEN A. AND WENTWORTH S. (1999a) Integration of the chemical and mineralogical characteristics of lunar soils with reflectance spectroscopy (abstract). Lunar Planet. Sci. 30, #1859, Lunar and Planetary Institute, Houston, Texas, USA (CD-ROM). TAYLOR L. A,, PATCHEN A., MORRIS R. V., TAYLOR D-H., PIETERS c. M., KELLER L. P., MCKAY D. s. AND WENTWORTH s. (l999b) Chemical and mineralogical characterization of the 44-20,20-10, and <10 micron fractions of lunar mare soils (abstract). Lunar Planet. Sci. 30, #1885, Lunar and Planetary Institute, Houston, Texas, USA (CD-ROM). TAYLOR L. A., MORRIS, R. V., KELLER L. P., PIETERS C. M., PATCHEN A., TAYLOR D-H., WENTWORTH S. AND MCKAY D. s. (2000a) Major contributions to spectral reflectanc opacity by nonagglutinitic, surface-correlated nanophase iron (abstract). Lunar Planet. Sci. 31, #1842, Lunar and Planetary Institute, Houston, Texas, USA (CD-ROM). TAYLOR L. A., MORRIS R. V., PIETERS C. M., PATCHEN A., TAYLOR D-H., KELLER, L. P., WENTWORTH S. AND MCKAY D. s. (2000b) Chemical characterization of lunar mare soils (abstract). Lunar Planet. Sci. 31, #1697, Lunar and Planetary Institute, Houston, Texas, USA (CD-ROM). TAYLOR L. A., PATCHEN A., TAYLOR D-H., MORRIS R. V., PIETERS C. M., KELLER L. P., WENTWORTH s. AND MCKAY D. s. (2000C) Mineralogical characterization of lunar mare soils (abstract). Lunar Planet. Sci. 31, #1706, Lunar and Planetary Institute, Houston, Texas, USA (CD-ROM). TAYLOR L. A., PIETERS C. M., KELLER L. P., MORRIS, R. V., MCKAY D. S., PATCHEN A. AND WENTWORTH S. (2000d) Space weathering of lunar mare soils: New understanding of the effects on reflectance spectroscopy. In Space 2000: Proc. of Space 2000, 7th Int? Conf And Expos. On Engr., Constr., Oper., and Business in Space (eds. S. W. Johnson, K. M. Chua, R. G. Galloway, and P. J. Richter), pp Amer. SOC. Civil Engrs. Press, Virginia, USA. WALKE R.J. AND PAPIKE J. J. (1981) The relationship of the lunar regolith <I0 mm fraction and agglutinates. Part 11: Chemical composition of agglutinate glass as a test of the "fusion of the finest fraction" F3 model. Proc. Lunar. Planet. Sci. 7th, WENTWORTH S. J., KELLER L. P., MCKAY D. S. AND MORRIS R. V. (1999) Space weathering on the Moon: Patina on Apollo 17 samples and Meteorit. Planet. Sci. 34,