Carbon abundance and silicate mineralogy of anhydrous interplanetary dust particles

Carbon abundance and silicate mineralogy of anhydrous interplanetary dust particles

Gmchimica et Cosmochimica Am Vol. 57. pp. 1551-1566 Copyright 0 1993 Pergamon Press Ltd. Printed in U.S.A. 0016-7037/93/$6.00 + .OLl Carbon abundan...

1MB Sizes 2 Downloads 20 Views

Gmchimica et Cosmochimica Am Vol. 57. pp. 1551-1566 Copyright 0 1993 Pergamon Press Ltd. Printed in U.S.A.

0016-7037/93/$6.00

+ .OLl

Carbon abundance and silicate mineralogy of anhydrous interplanetary dust particles KATHIE L. THOMAS,* GEORGE E. BLANFORD,’ LINDSAY P. KELLER,~ WOLFGANG KL&K,~ and DAVID S. MCKAY 3

‘Lockheed Engineering and Science Co., 2400 Nasa Rd, Houston TX, 77058 USA ‘University of Houston-Clear Lake, 2700 Bay Area Blvd., Houston TX, 77058 USA ‘NASA Johnson Space Center, Mail Code SN14, Houston TX 77058 USA 41nstitut fur Planetologie, Wilhelm-Klemm-Str. 10, 4400 Munster, FRG (Received March 2, 1992; accepted in revisedform September 30, 1992)

Abstract-We have studied nineteen anhydrous chondritic interplanetary dust particles (IDPs) using analytical electron microscopy. We have determined a method for quantitative light element EDX analysis of small particles and have applied these techniques to a group of IDPs. Our results show that some IDPs have significantly higher bulk carbon abundances than do carbonaceous chondrites. We have also identified a relationship between carbon abundance and silicate mineralogy in our set of anhydrous IDPs. In general, these particles are dominated by pyroxene, olivine, or a subequal mixture of olivine and pyroxene. The pyroxenedominated IDPs have a higher carbon abundance than those dominated by olivines. Members of the mixed mineralogy IDPs can be grouped with either the pyroxene- or olivinedominated particles based on their carbon abundance. The high carbon, pyroxene-dominated particles have primitive mineralogies and bulk compositions which show strong similarities to cometary dust particles. We believe that the lower carbon, olivine-dominated IDPs are probably derived from asteroids. Based on carbon abundances, the mixed-mineralogy group represents particles derived from either comets or asteroids. We believe that the high carbon, pyroxene-rich anhydrous IDPs are the best candidates for cometary dust. INTRODUCTION

from the comet Halley encounter suggest that particles containing abundant light elements (especially C) are a major component of cometary dust ( KISSEL and KRUEGER, 1987;

CHONDRITIC INTERPLANETARYDUST PARTICLES( IDPs) are samples of materials that are believed to have formed very early in the history of our solar system. Of particular interest are a subset of particles that are classified as anhydrous IDPs. These particles are considered to be the most pristine samples available for the study of the early solar system because of their primitive mineralogy, chemistry, and isotopic characteristics ( MCKEEGANet al., 1985; ALLAMANDOLA et al., 1987; BRADLEY, 1988). These particles have escaped the thermal processing and aqueous fluid interactions that have altered many of the parent bodies of the chondritic meteorites. In general, anhydrous IDPs have a chondritic composition within a factor of two of CI for most major and minor elements with the exception of carbon which has been reported to be up to -4 times chondritic in some particles ( BLANFORD et al., 1988; THOMAS et al., 1992). The high carbon abundance in some anhydrous IDPs is incompatible with an origin from known chondritic meteorites. The presence of carbonaceous materials in IDPs has broad implications for the processes involved in the formation and evolution of the early solar system. Although carbon-bearing materials have been observed previously in IDPs (e.g., BRADLEYet al., 1984; CHRISTOFFERSENand BUSECK, 1984, 1985; RIETMEIJER and MACKINNON, 1985; WOPENKA, 1988), their nature, abundance, and distribution are still poorly known. A knowledge of the abundance and nature of carbon in IDPs is useful for determining sources of IDPs and for comparisons with other chondritic materials. Estimates of C abundance in anhydrous IDPs suggest that many of these particles have significantly higher C than do the carbonaceous chondrites ( BLANFORD et al., 1988; BRADLEY et al., 1988; SCHRAMM et al., 1989; THOMAS et al., 1992). Data

JESSBERGERet al., 1988).

The most carbon-rich meteorites (the CI chondrites) show a range of carbon abundance from 2.7 l-4.46 wt% C (KERRIDGE, 1985 ). ANDERSand GREVES~E( 1989) report a mean bulk carbon abundance of 3.45 wt% for Orgueil. The only estimate of carbon abundance in comets is from Puma 1 mass spectra from Halley particles which showed enrichments in carbon up to -12 X CI (JESSBERGERet al., 1988). Although carbon abundances vary widely in Halley data, these cometary dust particles are clearly enriched in carbon compared with abundances from the most carbon-rich chondrites. Our previous work ( BLANFORD et al., 1988; THOMAS et al., 1992) has shown that C is an important and significant component of most anhydrous IDPs and that C abundance is apparently correlated with the silicate mineralogy of the particles (see following discussion ). The main objectives of this work were to ( 1) accurately and quantitatively determine bulk major, minor, and light elements in anhydrous IDPs, (2) determine relationships between IDP mineralogy and bulk chemistry, and (3) determine likely sources for anhydrous IDPs. METHODS Nineteen anhydrous IDPs, ranging in size from 6.5 to 20 Frn, were selected from the NASA/JSC Cosmic Dust catalogs (CDC). Two main criteria were used to select IDPs: surface morphology and bulk chemistry. Only IDPs with a porous surface morphology were se&ted, their morphology indicates that the particles are aggregates of submicron-sized grains. The second criterion used was bulk chemistry. We chose only those 1DPs that were classified as C-type in the CDCs. Of the chondritic IDPs available, we further reduced the number by 1551

1552

K. L. Thomas et al.

selecting the IDPs that were not calcium depleted. Generally, hydrated IDPs show a strong calcium depletion compared with anhydrous IDPs ( SCHRAMMet al., 1989). The IDPs were transferred onto beryllium substrates which had been cleaned with acetone and ultrapure freon I 13 and plasma etched to remove any carbon residues. The cleaned surfaces were checked for carbon contamination by examination in the same manner as the IDPs (see following text). If carbon was detected after this procedure, the surfaces were plasma etchqd again until no carbonwas detected. The IDPs were transferred onto the substrates in a class 100 clean room and were rinsed with hexane. The IDPs were examined with a JEOL 35 CF scanning electron microscope operated at I5 kV and equipped with a Princeton Gamma Tech (PGT) X-ray energy dispersive spectrometer. The PGT spectrometer is a windowless detector which was used to determine bulk particle compositions. The windowless detector allows detection of elements withZ > 5. The IDPs were analyzed with a beam that approximated the particle diameter. The particles were imaged using a JEOL 100 CX scanning transmission electron microscope (STEM). One particle was examined at high resolution using a JEOL 2000 FX STEM. After the initial bulk analysis, eleven of the IDPs were embedded in epoxy, thin sectioned using an ultramicrotome, and examined with a JEOL IO0 CX STEM. For these particles we have mineralogical data in addition to the chemical data. The techniques for determining mineralogy and more details about the mineralogy of some of our particles are given by KLoCK et al. ( 1989). One IDP, W7027Hl4, with the highest bulk carbon content, was thin sectioned and examined with a JEOL 2000 FX STEM at 200 kV. ANALYTICAL

TECHNIQUES

Energy dispersive X-ray spectra were processed through the PGT bulk particle data reduction (BSAM ) program; each IDP was analyzed once for 500 seconds (see Reported Errors section). This program is PGT’s version of Frame C ( MYKLEBUST,1979) modified for light elements using tabulated mass absorption coefficients. It performs peak overlap corrections, background subtractions for elements Z > 6, and calculates the ratios of X-ray counts from the unknown to

counts from a pure, flat element standard (k-ratios).Standard counts were stored for computation by analyzing pure element and mineral standards. The BSAM procedure takes into account peak overlap for both K-peaks and L-peaks. It corrects, for example, for partial overlap between the iron L-peak and the oxygen K-peak. Carbon k-ratios were determined for each spectrum manually. After manual background subtraction for C, we recorded the counts in a window defined between 170-340 eV. These counts were divided by the average of three measurements of a cleaved diamond standard to obtain carbon k-ratios. All k-ratios were used as input to the CITPTC (Ver. 2.03 ) matrix and I$(pp) correction procedure (J. T. Armstrong, pers. commun.). The CITPTC program was designed for analyzing small particles using X-rays produced from electron beams. It is upgraded from the program first presented by ARMSTRONGand BUSECK( 1975). The newer program uses tabulated mass absorption coefficients (J. T. Armstrong, pers. commun.). As in the earlier program, the CITPTC program results are normalized to 100%. If normalization is used, it is not necessary to compute the number of electrons which interact with the sample which is highly dependent on size and geometry ( ARMSTRONGand BUSECK, 1975 ) . The CITPTC software computes corrections using the size and several regular geometric shapes. This program has been shown to give good results for heavier elements despite the fact that the unknowns do not have semi-infinite plane geometry (ARMSTRONGand BUSECK, 1975). This result has not been proven for light elements. We have tested this routine on the reported particles using sizes that are larger and smaller than the actual particle sizes and have computed small but significant differences in composition. For example, if a particle is greater than 10 pm, and a size of 10 pm or greater is used as input, then there are no significant differences in composition. If a much smaller size is chosen, for example 5 pm, then significant elemental differences become apparent. Within the CITPTC program there are six geometric shape models which can be selected. Since we found no differences within statistical uncertainties between geometric models, we chose the square-pyramid model for all analyses. We compensated for the non-linearity of the detector at low energy by analyzing standards under the same conditions as unknowns. Consequently, we believe the analytical results for IDPs are valid despite their irregular shapes. The CITPTC program is an iterative routine which uses the last cal-

Table 1 Quantitative EDX Analyeee of Polished, Flat Calcite Standard and 17 Calcite Fragments

Laboratory

17

Standard

Crushed

Crushed

Graine

Grain6

Stoichiometric Value

Accepted

Meaeured

Value

Value

(Mean and lu Error)

17

(Range of Values)

C

12.0

12.02

12.6

12.1 f 1.2

9.5-14.2

0

40.0

47.98

48.6

47.4 f 3.2

42.9-53.7

38.8

40.4 f 4.4

32.6-47.5

100.0

100.0 f 5.6

Mg Ca

0.01 40.0

Mn Total

39.98 0.01

100.0

100.00

Carbon content of IDPs

1553

Table 2 Quantitative EDX Analyses of 19 Anhydrous Interplanetary Dust Particles

uzo15

U2015

D21 20 lun

Size

D22b

U2015

U2034

U2034

E3b

E3

F27

7, 7, 7 jun 7, 12, 12 p

O-P

O-P

C

9.3 f 0.8

9.5 f 3.5

0

35.2 f 0.9

12.5 q

8.8 p U

U

7.6 f 0.4

15.4 f 1.4

7.4 f 0.6

35.1 f 2.2

32.4 f 2.9

28.4 f 0.7

24.9 f 0.6

1.0 f 0.1

1.0 f 0.3

0.9 f 0.3

1.5 f 0.1

1.2 f 0.1

12.1 f 0.6

10.3 f 1.5

9.2 f 1.5

7.4 f 0.4

8.1 f 0.4

Al

1.1 f 0.1

1.4 f 0.1

1.1 f 0.3

1.7 f 0.1

1.1 f 0.1

Si

14.7 f 0.4

15.6 f 1.7

14.9 f 1.6

13.6 f 0.4

10.9 f 0.3

P

0.4 f 0.1

0.3 f 0.2

0.7 f 0.6

0.6 f 0.1

0.4 f 0.1

S

5.3 f 0.2

4.3 f 3.0

5.9 f 3.0

8.6 f 0.3

10.5 f 0.4

Ca

1.0 f 0.1

0.8 f 0.2

1.1 f 0.6

0.6 f 0.1

1.1 f 0.1

Cr

0.3 f 0.1

0.1 f 0.1

0.3 f 0.1

0.1 f 0.1

0.6 f 0.1

Mn

0.2 f 0.1

0.1 f 0.1

0.2 f 0.1

0.1 f 0.1

0.5 f 0.1

Fe

18.6 f 1.3

20.5 f 1.6

24.9 f 4.4

21.0 f 1.4

31.1 f 2.1

Ni

0.8 f 0.1

1.0 f 0.2

0.8 f 0.2

1.1 f 0.1

2.0 f 0.1

TYPea

Na

Total

100.0

ao= (>66

0

100.0

100.0

100.1

99.8

modal% olivine); O-P = (<66 modal% olivine and <66 modal%

pyroxene); P = (>66 modal% pyroxene); U = Unknown. bMean and standard deviations

(la) for particle fragments.

culated composition to calculate the next fit and concludes when the routine has passed a convergence test. Currently, all such routines assume that the elements analyzed are homogeneously distributed in the sample. The implications that this has for carbon analysis are discussed in the following section. CARBON ANALYSIS

PROCEDURES

Quantitative analyses for carbon are difficult and require painstaking procedures in order to obtain reliable results. The most severe problem with carbon analysis results from the strong absorption of carbon X-rays by other elements such that carbon X-rays essentially come from a shallow depth in the sample which is less than the thickness of the particle. Two major issues must be addressed when analyzing for carbon in IDPs: the distribution of carbon in each IDP and the extent ofcontamination by silicone oil in each particle. First, if the carbon is concentrated near the surface of the particle, then the abundance of carbon will be significantly overestimated. Con-

versely, if C is concentrated in the center of the particle, C abundances will be grossly underestimated. Second, if the IDPs retain some of the silicone oil used as a collecting medium, then the IDPs may also be enriched in carbon from this contaminant. Because we want to report accurate carbon abundances in IDPs, we performed a variety of checks on our analytical techniques to assess the reliability of our carbon values. We describe these checks in detail in the following text.

REPORTED

ERRORS

Electron beam analyses of small particles using the CITPTC program do not yield the same precision as analyses of flat, polished samples. ARMSTRONG and BUSECK (1975) analyzed 122 anorthite particles and had relative errors of 4.0%, 4.5%, and 8.5% for Si, Al, and Ca, respectively. Our analyses of

K. L.Thomas et al.

1554

Table 2. (Continued)

w7013 813 Size

7.5 /un

w7013

w7013

B17*

C16*

7, 7.5, 8 p

14, 20, 23 jun

w7013

w7013

DlZ

E9

7 m

14 lun

U

IJ

0

P

O-P

C

9.4 f 0.8

7.6 f 2.9

6.4 f 5.1

11.1 f 1.0

16.0 f 1.4

0

35.6 f 0.9

38.1 f 3.1

33.7 f 3.2

32.4 f 0.8

34.0 f 0.4

1.5 f 0.1

1.2 f 0.2

0.7 f 0.1

0.8 f 0.1

1.5 f 0.1

12.6 f 0.6

14.7 f 2.0

12.4 f 2.9

8.3 f 0.4

10.9 f 0.4

Al

1.7 f 0.1

1.2 f 0.1

0.8 f 0.1

1.4 f 0.1

1.1 f 0.1

Si

18.5 f 0.5

18.5 f 1.7

13.5 f 1.3

13.2 f 0.5

16.1 f 0.5

P

0.3 f 0.1

0.2 f 0.1

0.2 f 0.1

0.5 f 0.1

0.2 f 0.1

S

5.7 f 0.2

1.9 f 0.6

10.1 f 4.4

7.7 f 0.3

5.4 f 0.1

Cd

1.2 f 0.1

0.7 f 0.3

0.6 f 0.1

0.6 f 0.1

1.0 f 0.1

Cr

0.3 f 0.1

0.3 f 0.1

0.1 f 0.1

0.3 f 0.1

0.2 f 0.1

Mn

0.2 i 0.1

0.2 f 0.1

1.1 f 0.1

0.1 f 0.1

Fe

12.6 f 0.8

15.1 f 7.6

20.8 f 6.1

21.8 f 1.5

13.1 f 0.2

Ni

0.5 f 0.1

0.4 f 0.1

0.8 f 0.1

0.8 f 0.1

0.5 f 0.1

TyPea

Na

Total

100.1

100.1

co.1

100.1

100.0

LOO.1

=O = (>66 modal% 01 .vine) ; O-P = (~66 modal% olivine and <66 modal% pyroxene); P = (>66 modal% pyroxene); U = Unknown. b

Mean and standard

deviations

(la)

seventeen calcite particles had relative errors of 9.9%, 6.8%,

and 10.9% for C, 0, and Ca, respectively (Table 1). Average measurement of Ca of ARMSTRONGand BUSECK( 1975) was I .3% higher than the standard and our Ca average was also 1.3% higher than the standard. When we report measurement uncertainties for standards and IDPs we derive them in one of several ways. For standards, element abundance uncertainties are given by the standard deviation ( 1a) of multiple measurements. For IDPs, the element abundance uncertainties have been determined by one of two ways: ( 1) repeated bulk measurements of the same intact particle to determine analytical errors and (2) averaged standard deviations of particles which broke into several fragments prior to analysis. To estimate analytical errors, four particles (W70 13E9, W70 13G6, W70 13H5, and

for

particle

fragments.

W70 13H24) were measured in bulk several times in the same geometrical configuration. The errors reported for these particles in Table 2 are the standard deviations of the repeated measurements. From these measurements we determined that Na, Al, P, Cr, Mn, and Ni have absolute errors of 0.1 wt%. The largest relative errors for C, 0, Mg, Si, S, Ca, and Fe were found to be 8.9%, 2.5%, 4.8%, 2.9%, 4.0%, 10.8%, and 6.8%, respectively. These errors were considered to be representative of our measuring abilities and are reported (as absolute errors) in Table 2 for particles U20 I5D2 1, U2034E3, U2034F27, W7013B13, W7013D12, W7013E17, W7013G1, W7027E6, W7027H14, and W7029*B9. Five particles (U2015D22, U2015E3, W7013B17, W7013C16, and W7029 *A28 ) broke into fragments during sample washing and mounting. The uncertainties reported for these samples

1555

Carbon content of IDPs Table 2. (Continued)

w7013

w7013

w7013

w7013

El7

Gl

G6

Ii5

Size

23 *

20 w

Type'

O-P

12.5

0

jun

w7013 Ii24

10 p

O-P

6.5 jun U

U

c

12.9

f 1.1

5.0

f 0.4

7.7

f 0.1

24.1

f 1.4

22.1

f 0.1

0

35.0

f 0.9

41.9

f 1.0

34.0

f 0.3

27.8

f 0.6

32.6

f 0.8

Na

4.2

f 0.1

0.2

f 0.1

0.4

f 0.1

0.6

f 0.1

1.2

f 0.1

4

7.6

f 0.4

27.9

f 1.3

13.5

f 0.1

8.1

f 0.1

7.0

f 0.1

Al

5.6

f 0.1

0.2

f 0.1

0.9

f 0.1

0.9

f 0.1

1.2

f 0.1

Si

17.5

f 0.4

17.4

f 0.5

18.3

f 0.1

12.8

f 0.4

11.2

f 0.1

P

X0.1

0.1

f 0.1

0.1

f 0.1

0.2

f 0.1

S

5.3

f 0.2

0.4

f 0.1

4.6

f 0.2

8.8

f 0.3

11.0

f 0.4

Ca

1.2

f 0.1

0.3

f 0.1

1.6

f 0.1

1.1

f 0.1

0.4

f 0.1

Cr

0.1

f 0.1

0.3

f 0.1

0.5

f 0.1

0.1

f 0.1

co.1

Mn

0.1

f 0.1

0.9

f 0.1

0.5

f 0.1


Fe

10.0

f 0.7

5.4

f 0.4

16.8

f 0.5

15.0

f 1.0

12.3

f 0.4

Ni

0.4

f 0.1

0.2

f 0.1

0.9

f 0.1

0.6

f 0.1

0.4

f 0.1

=

(~66

pyroxene); b

Mean

100.1

99.9

Total

aO

<0.2

and

modal% P =

olivine); (>66 modal%

etandard

99.8

O-P

= (<66 modal%

pyroxene);

deviations

(Table 2) are the standard deviations for the particle fragments. These uncertainties are in general much larger than the actual analytical errors and reflect the heterogeneity of carbon distribution in the IDPs. They are more representative of the standard deviation of the average of all the particles (Table 3) rather than the analytical errors. ANALYTICAL CHECKS Calcite particles, which were finely crushed and dispersed on a Be surface, were analyzed as an analytical check on the accuracy and precision of the bulk carbon measurements in small irregularly shaped particles with a homogeneous C distribution. The results for C, Ca, and 0 for seventeen calcite grains ranging in size from lo-20 pm are reported in Table


100.0

olivine

99.6

and

~66

modal%

U = Unknown.

(1~) for particle

fragments.

1 and can be compared to a bulk analysis of a flat, polished standard, and the accepted value. The results are generally good with the carbon showing the largest variation approximately 10% relative from particle to particle. The mean carbon value from all the particles, the analysis of the standard, and the accepted value for the standard are within 5% relative error. Particles of finely crushed Smithsonian Allende Reference Sample (SARS) ( JAROSEWICH et al., 1987) were covered with silicone oil and rinsed in the same manner as IDPs to serve as analytical blanks to check our small particle data reduction program for major and minor element abundances and to check for the extent of silicone oil contamination. The SARS sample is a crushed and homogenized aliquot of the Allende CV carbonaceous chondrite with grain sizes somewhat larger

K. L. Thomas et al.

1556

Table 2. (Continued)

W7021

W7027

W7029

W7029

E6

H14

*A28 b

*B9

Size

12 W

12 elm

U

P

Typea

12,

20 /lIn

7W

P

U

C

11.7

f 1.0

22.8

f 2.0

12.7

f 2.7

19.2

f 1.7

0

36.1

f 0.9

29.1

f 0.7

27.4

f 4.1

31.5

f 0.8

1.5

f 0.1

0.6

f 0.1

0.8

f 0.2

1.0

f 0.1

9.4

f 0.4

8.8

f 0.4

6.7

f 1.9

8.4

f. 0.4

Al

1.2

f 0.1

0.9

f 0.1

0.8

f 0.1

1.3

f 0.1

Si

16.9

f 0.5

14.6

f 0.4

1.8

f 1.4

11.3

f 0.3

P

0.4

f 0.1

0.3

f 0.1

0.2

f 0.1

0.4

f 0.1

S

6.9

f 0.3

1.3

f 0.3

11.6

f 2.1

8.1

f 0.3

Ca

0.9

f 0.1

0.7

f 0.1

0.6

f 0.2

1.1

f 0.1

Cr

0.2

f 0.1

0.1

f 0.1

0.2

f 0.1

0.1

f 0.1

Mn

0.1

f 0.1

0.1

f 0.1

0.1

f 0.1

X0.1

Fe

14.0

f 1.0

14.2

f 1.0

30.2

f 8.3

17.2

f 1.2

Ni

0.6

f 0.1

0.5

f 0.1

0.9

f 0.1

0.5

f 0.1

Na

Total

‘0

99.9

100.0

= (>66 modal% olivine);

pyroxene);

100.0

O-P = (<66 modal% olivine

P = (~66 modal% pyroxene);

bMean and standard

100.1

deviations

than that of IDPs. To more closely approximate the finegrained nature of IDPs, we crushed the sample further. The

SARS sample was divided into three groups: ( 1) untreated, (2) sintered, and (3) sintered, then treated with silicone oil, and then rinsed in the same manner as IDPs. The untreated SARS group was dispersed on clean Be foil and forty aggregates from IO-20 pm in size were measured and analyzed by the same methodology used for the IDPs. The average analysis is reported in the second column of Table 4 which can be compared to the accepted value in column 1. The agreement for all major and minor elements is within statistical error except for S, which is too low; the overall agreement of our analyses of individual grains of SARS sample with the bulk is excellent. The second group of SARS particles were heated to sinter the particles partially together. Thirty aggregates from IO-20 pm were analyzed on clean Be foil in the

(la)

and <66 modal%

U = Unknown. for

particle

fragments.

same manner as IDPs and the first SARS group. The results are reported in the third column of Table 4. Again, the agreement for all major and minor elements is within statistical error except for S, which is lower than the unheated SARS group indicating volatilization from heating, and Fe which is slightly above the accepted value. The overall agreement of our analyses of partially sintered SARS grains with the bulk is excellent. SILICONE OIL CONTAMINATION

The problem of contamination of stratospherically collected IDPs by silicone oil has been considered by a number of researchers ( FRAUNDORFand SHIRCK, 1979; SANDFORD, 1986; RIETMEIJER, 1987; BLANFORDet al., 1988; SCHRAMM et al.. 1989). Although it is a persistent problem, the actual

Carbon content of IDPs Table Average

Element

3

Compoeition (in wt.%) for 19 IDPa

C

12.5 f 5.7

0

32.9 f 4.0

Na

1.1 f 0.8

Mg

10.7 f 4.6

Al

1.3 f 1.1

si

14.6 f 2.9

P

0.3 f 0.2

s

6.9 f 2.9

Ca

0.9 f 0.3

Cr

0.2 f 0.1

Mn

0.2 f 0.3

Fe

17.6 f 6.3

Ni

0.7 f 0.4

and lu Errors

1557

(( CHs)zSiO) contains 0 and C ( N 30% by wt.), we are not detecting excess carbon or oxygen. This implies that the hexane washing may be very efficient in removing silicone oil from the surfaces of the particles, beyond the depth of detection for C and 0 X-rays. Therefore, we believe our carbon abundances are probably not significantly affected even though the silicon abundances are overestimated. Third, our findings agree with SCHRAMM et al. ( 1989), who estimated similar Si enhancements using a different methodology and meteorite sample. ORGUEIL Because SARS has very little carbon (0.25 wt% ) , crushed fragments of the Orgueil (CI) chondrite were analyzed to compare the results of our technique with the accepted bulk CI composition ( ANDERS and GREVESSE, 1989). We analyzed thirty fragments from lo-20 pm in size. The results given in Table 5 show good agreement for most major elements, although oxygen is low in the analyzed fragments. Unlike the SARS sample, our Orgueil sample is not a homogenized aliquot; this is clearly seen in the ranges of the thirteen elements including C (Table 5). Carbon abundances range from O7.4 wt% which is a reflection of heterogeneity on the lomicrometer scale. Nevertheless, the average of the individual bulk analyses for the fragments, including that for carbon, shows good agreement with the accepted bulk values.

CARBON DISTRIBUTION IN IDPs

extent of this contamination is poorly understood because it is not known whether or not the physical characteristics of the IDPs, such as the degree of porosity or the abundance of carbon, will influence the amount of silicone oil retention. SCHRAMM et al. ( 1989) suggest that the excess silicon in chondritic IDPs is less than 10% of the measured silicon abundance, yet they do emphasize that the individual physical parameters of particles can affect the amount of silicone oil retention. The third group of SARS samples was used to check the extent of silicone oil contamination of IDPs. These particles were pulverized, partially sintered, saturated with silicone oil, washed in hexane, and analyzed for major elements, carbon, and oxygen. The pulverizing and sinteting steps were necessary to simulate the porosity and binding of mineral grains in chondritic porous IDPs. The result of analyzing twenty six aggregates is shown in column 4 of Table 4. Silicon is - 10% higher in the third group of SARS particles when compared to the second group of SARS particles in column 3 and is - 13% higher when compared to the first group of SARS particles in column 2; carbon and oxygen are not significantly increased. This result leaves us with three conclusions: first, that there is a real contamination problem from the silicone oil which causes an - 10% enhancement in Si. As for individual IDPs, the true extent of this contamination is dependent on a variety of particle parameters and the extent to which the IDP was rinsed, silicone contamination varies in our standard from lo-13%. Second, although silicone oil

As previously mentioned, there is concern that the EDX measurement of carbon will be affected by the heterogeneous distribution of carbon within the IDP. If carbon is primarily distributed on the IDP surface, then the CITPTC program, which assumes a homogeneous distribution, will overestimate the carbon abundance. We performed three checks to determine the distribution of carbon in IDPs. In the first check, U2034F27 was analyzed for its bulk carbon abundance and major element composition, then crushed between two sheets of clean beryllium foil. We obtained analyses from twentyone fragments and compared the average of these analyses to the bulk composition reported in Table 6. By comparing the size of the fragments to the estimated bulk volume, we estimate that we measured - 50% of the volume of the IDP. The large uncertainties in the abundances of the major elements reflect the inhomogeneity of this IDP. Compared with the bulk composition, however, we find excellent agreement for most elements. Iron and sulphur abundances are higher in the individual fragments indicating that FeS grains were preferentially analyzed relative to the carbon-rich fragments. Carbon abundances in the fragments varied from -9 wt% to essentially none. The high carbon fragments appear to be very fine-gmined, whereas the low carbon regions are rather large, individual grains with little fine-grain material. It is apparent that the carbon is distributed throughout the particle and the most important parameter in determining carbon abundance is probably the range of grain sizes within the IDP. A similar result was obtained by SCHRAMM et al.( 1989) who observed that the coarse particles (those with mineral grains larger than 3 pm and comprise more than 50% of the

K. L. Thomas et al.

1558

Table 4

Quantitative EDX Analyses of Smithsonian Allen&

Reference Sample 1319

Weight Percents and la Errors (see text for details)

Accepted Value'

Crushed Only

Sintered Only

Sintered and Treated with Silicone Oil

Number of

30

40

26

Analyses co.1

so.1

37.7 * 2.2

36.5 f 1.8

37.2 f 2.0

0.34 f 0.01

0.5 f 0.3

0.3 f 0.2

0.2 f 0.2

Wg

14.83 f 0.12

13.6 i 1.7

13.3 f 0.7

13.1 f 0.9

Al

1.74 f 0.04

1.8 f 0.6

1.7 f 0.2

2.1 f 1.0

Si

16.02 f 0.10

15.6 f 0.7

16.1 f 0.6

17.7 f 1.1

P

0.10 f 0.00

0.1 i 0.1

0.1 f 0.1

0.2 f 0.1

S

2.10 f 0.03

1.4 * 0.4

1.0 f 0.4

1.0 f 0.2

Ca

1.84 f 0.05

1.6 f 0.5

1.9 f 0.4

1.8 f 0.6

cr

0.36 f 0.01

0.2 f 0.1

0.3 * 0.1

0.2 f 0.1

Fe

23.57 zt0.08

26.2 i 3.7

27.4 f 2.6

25.4 * 2.4

Ni

1.42 * 0.02

1.2 f 1.3

1.4 f 0.5

1.1 f 0.1

C

0.25 f 0.02

0.1 f 0.3

0

36.60 f 0.20

Na

'Jaroeewich et al. (1987)

particle mass) have -2 times less bulk carbon than the anhydrous particles which seem to have more uniform grain size ranges (e.g., not dominated by one or several large grains), In our second check on the distribution of carbon in IDPs, we analyzed twenty-four spots on the surface of one fragment of W70 I 3C 16 with a spot size < 0.5 ,cim (W70 13C16 broke into three fragments). The mean carbon abundances for the spot analyses are compared with the bulk measurement for the fragment and the mean for the three fragments of W70 13C16 in Table 7. Although the mean of the spot analyses is higher than the bulk measurement, the bulk measurement falls within 10 of the spot analyses. The range of the spot analyses (0- 17 wt.% ) demonstrates the unevenness of the carbon distribution: some surface grains are free of carbonaceous material while others have up to 17 wt% C. Obviously, carbon is not present as a smooth coating on exterior of the particle.

Our third check on the distribution of carbon in IDPs involved the direct observation of low-2 material in microtomed thin sections in the TEM. In none of the anhydrous IDP sections we examined, did we observe carbon mantling the outside of any particle. We selected one particle, W7027H 14, for closer examination because of its unusually high carbon abundance ( TabIe 2 ) . W7027H 14 was thin sectioned using an ultramicrotome and examined using a JEOL 2000 FX STEM equipped with a Link spectrometer with a Be window. The mineralogy of this IDP is dominated by pyroxene grains, but olivine and glass grains are also present. Solar flare tracks were observed in a forsteritic olivine indicating that this particle is extraterrestrial. The grains in thin sections range from a few nanometers to - 1 Mm in diameter. Surrounding these grains is amorphous material that optically is distinguishable from the embedding medium. EDX analyses of the amorphous material faII into two distinct compositional groups. In the ftrst group, we fmd major Si and

Carbon content of IDPs Table Mean Compoeition

(wt%)

of

29 Crushed

Compared

to

Ite

1559

5

Fragments

of

the

Orgueil

CI Chondrite

Bulk Composition

Bulk Compoeit iona

Cruehed Fragment a Mean and la Errors

Range

C

3.6

f 2.1

0

0

40.3

f 3.1

29.1

Na

1.0

f 0.4

0.4

-

1.8

0.5

M9

11.9

f 3.0

0.5

- 14.6

9.5

Al

1.6

f 0.4

0.4

-

2.5

0.9

SF

13.7

f 3.4

0.8

- 16.6

10.7

P

0.9

f 3.1

0.1

- 17.5

0.1

s

3.1

f 1.5

0.2

-

8.2

5.2

Ca

2.0

f 7.4

0

- 40.3

0.9

Cr

0.2

f 0.1

0

-

0.4

0.3

Mn

0.1

f 0.1

0

-

0.2

0.2

Fe

19.9

f 8.5

0.5

- 58.5

18.5

Ni

1.7

f 0.5

0

-

“Anders

and Grevesse

-

7.4

3.4

- 46.0

46.4

2.5

1.1

(1989)

Mg and minor Ca and/or K (Fig. lc). We infer from this pyroxene-like composition and the lack of long-range order (as shown from electron diffraction patterns) that this material is probably glass. Grains with major Si and Al and minor Ca and/or K are also observed without long-range order and are also likely to be glass. Analyses from the second group contain only Si and/or S in varying small amounts (Fig. la,e). We infer from the lack of elements other than Si and S and the lack of long-range order that this mat&d is a carbonaceous phase containing adsorbed atmospheric S ( MACKINNON and MOGK, 1985 ) and absorbed silicone oil. We estimated the volume percent carbonaceous material in W7027H14 by point counting a microtomed thin section (Fig. 2). This result shows that the particle is 40 to 50 volume percent carbonaceous material which is in good agreement with an estimate of -40 volume percent assuming a particle diameter of 10 pm, a particle density of 1 g/cm3 (FLYNN and SUTTON, 1991), and -23 wt% C (bulk). Our experimental results indicate that the carbon is distributed throughout IDPs, appearing unevenly on the particle surface and surrounding the internal grains. Therefore, we

conclude not been program, with the for IDPs

that the carbon abundances for these particles have systematically overestimated by the data reduction and considering the degree of accuracy of carbon Orgueil test, we believe that our carbon abundances are within - 10% of the true values. RESULTS AND DISCUSSION

Quantitative EDX analyses of the nineteen anhydrous IDPs are given in Table 2. The element abundances have been normalized to Si and are listed in Table 8 along with the average values and ranges. Our results are compared to the composition of anhydrous IDPs reported by SCHRAMMet al. ( 1989) (Table 8). The mean element /Si ratios are the same for major elements except for Mg/Si, which is lower in our data set; however, the range of values is the same as the SCHRAMM et al. ( 1989) data. The discrepancy for Mg/Si probably results from the relative sizes of the data sets (nineteen compared to ninety) and the fact that the particles are mineralogically heterogeneous at the submicrometer scale. It is worth noting some of the following trends relative to

K. L. Thomas et al.

1560 Table Comparison

of a Broad of U2034F21

Analysis

6

Beamed Analyeie to

reported by other workers for some anhydrous IDPs (BRADLEY et al., 1989). Magnesium shows a wide range of values, but Mg shows no apparent correlation with Fe. On the other hand, Ni is correlated with Fe. The Fe/Ni ratio is 26 + 16 compared to 18 for chondritic meteorites. We have determined C/Si and O/Si ratios for each of the nineteen IDPs. The average C/Si ratio is -2 for our IDPs, but varies considerably from particle to particle (from 0.7 to 4.9) (Table 8). As we have taken considerable effort to ensure that our carbon values are reliable (see Methods), we argue that the variability in C/Si ratios reflects major differences in carbon abundance among our analyzed IDPs. SCHRAMM et al. ( 1989) reported a C/Si ratio of 2.4 for a subset of thirty anhydrous IDPs (no range of values given) and BRADLEYet al. (1989) suggested a mean C/Si ratio of 1.75. Our mean C/Si ratio, the SCHRAMMet al. ( 1989) ratio, and the BRADLEY et al. (1989) ratio indicate an average C/Si ratio for anhydrous IDPs of -2. This ratio is considerably higher than that for the most carbon-rich meteorites, e.g., the CI chondrites (-0.3 C/Si for Orgueil; ANDERS and GREVESSE, 1989).

(wt%)

Mean

the

of 21 Crushed

Fragments

Crushed Fragmentsa

Broad Beam

C

1.4

4.3

0

24.9

20.8

Na

1.2

1.2

f

0.1

KJ

8.1

5.9

f

4.0

Al

1.1

1.2

f

0.6

Si

10.9

10.4

f

7.2

f

0.1

f

4.5

f 11.0

P

0.4

0.5

s

10.5

16.6

K

0.2

0.1

f

0.1

Ca

1.1

0.8

f

1.1

Cr

0.6

0.2

f

0.4

Mn

0.5

0.2

f

0.2

Fe

31.1

35.8

Ni

2.0

1.9

Based on the range of carbon abundances in the Orgueil fragments (Table 5), we have established that a particle with C approximately less than 3 times CI is chondritic. We examined the relationship of carbon abundance relative to silicate mineralogy from our limited data set. Some clear

f 11.1

f 15.9 f

1.8

aMean and lo Errors

the Fe abundance. The IDPs with the highest Fe also have high S, indicating a correlation with sulfur and the presence of FeS (Table 2 ). We have not observed the sulfur depletion Table

trends are apparent: particles with chondritic carbon (C < 3 X CI) are dominated by olivines and those with non-chondritic carbon (C > 3 X CI ) are dominated by low Ca pyroxenes ( Fig. 3 ). Of the nineteen particles, eleven were thin sectioned and classified into one of three groups based on TEM observations, those dominated by olivine (>66 modal R), by pyroxene (>66 modal ?&), or with an approximately equal mixture of olivine and pyroxene. The olivine-dominated IDPs include U20 15E3, W70 13C 16, and W70 13G 1. Their carbon abundances are 7.6, 6.4, and 5.0 wt%, respectively (mean = 6 wt.% C, Table 2). In contrast, the pyroxene-dominated particles include W7027H 14, W70 13D 12. and W7029 *A28, with carbon abundances of 22.8, I I. 1, and 12.7 wt% C. respectively (mean = 16 wt% C). The olivine/ pyroxene group consists of particles W70 13G6, U2015D21, U20 l5D22.

I

Carbon (wt.%) in W7013C16: Spot and Bulk Analyses from One Fragment Compared to the Particle Mean from All Three Fragments

Region(s) Analyzed

24 Spot

Analyses

(1 Fragment)

Analysia

(1 Fragment)

Mean (3 Fragments)

aMean and la Errors

9.0

f 4.0a

0 - 17.0

Range Bulk

Carbon

6.8

f 0.6a

6.4

f 5.1a

1561

Carbon content of IDPs

(RIETMEIJER and MACKINNON, 1987). What is commonly

observed are regions of spongy-textured, poorly crystalline, low-Z material which we conclude contains most of the carbon between crystalline mineral grains (Fig. 2). It serves as a matrix or cement holding the mineral grains together. COMPARISON WITH OTHER EXTRAT’RRRBSI’RIAL MATERIAL

0

IO 0

keV

0

k&I

10 0

keV

k8V

10

10

255

E

a

0

u si

E 1

CUK

S

CUL

0

keV

IO

FIG. 1. X-ray, energy dispersive spectra for individual phases in a thin section of W7027H 14. The spectra are identified with letters AE and these regions are shown in Fig. 3. All spectra were obtained using the same analytical conditions. Copper peaks in the spectra come from the supporting grid. (A) Amorphous, carbon-rich region showing only Si and S in the spectrum. The full-scale count is 2000 indicating that this region is unusually thick. Cl is absent in this spectrum which indicates no intrusion from epoxy. (B) Pyroxene grain with a full scale of 2000 counts. (C) Round glass bead with a pyroxene-like composition. The full scale is 2000 counts. (D) Epoxy spectrum showing Si, S, and Cl peaks. The Cl peak clearly indicates that this region is epoxy which is the embedding medium for thinsectioning. The full scale is 127 counts; most of the counts come from the Cu grid supporting the thin-section. (E) Amorphous C-rich region showing Si and S in the spectrum, but no Cl. The full-scale count is 255 which is considerably lower than spectrum A. This spectrum, with a lower total count, is more typical of the spectra from amorphous C-rich regions.

W70 13E 17, and W70 13E9, with carbon abundances of 7.7, 9.3, 9.5, 12.9, 16.0 wt%, respectively (mean = 11 wt% C). Of the mixed group, the first three particles have C < 3 X CI (chondritic carbon) whereas the last two particles have C > 3 X CI (Fig. 3). The nature of the carbon-bearing phases in anhydrous IDPs is poorly understood. In Table 8, the atomic element/Si ratios are given for our subsets of anhydrous IDPs. Although the C/ Si ratio is enriched in our pyroxene-dominated and some mixed mineralogy anhydrous IDPs, the O/Si ratio is not significantly enriched, indicating that the carbon is not in an oxidized state. We have not observed graphitized carbon (i.e., 0.34 nm, (002) spacings) in any of our particles, nor have we observed carbon in the form of carbonates, although rare occurrences of carbonates in anhydrous IDPs are known

We suggest that the anhydrous pyroxene-dominated, carbon-rich IDPs originate from a cometary rather than an asteroidal source because: ( 1) they have much higher bulk carbon abundance than any known chondritic material, and (2) they consist predominantly of fine-grained Ca-poor pyroxene, unlike any fine-grained chondrite matrix. Atomic element/ Si ratios for major and minor elements in anhydrous IDPs, carbonaceous chondrites, and comet HalIey are listed in Table 8. In general, the carbon and sulfur abundances in pyroxenedominated IDPs are a closer match to comet Halley than to chondrites. We assume throughout this discussion that the Halley data are representative of cometary material. The Mg/ Si ratio for the pyroxene-rich IDPs is higher than that for comet Halley and is only slightly lower than the Mg/ Si range for carbonaceous chondrite matrices. Fe/Si ratios in pyroxene-rich IDPs are higher than those for Halley, but are comparable to that for the carbonaceous chondrites. However, unlike the chondrites, higher Fe in pyroxene-rich IDPs is correlated with higher S as is the case for the comet Halley data (Fig. 4). Further evidence that the C-rich pyroxenedominated IDPs may be derived from cometary sources comes from recent micro-IR spectroscopy measurements of IDP thin sections by BRADLEY et al. (1992). These authors showed that the best match to the IR spectra of cometary dust are C-rich IDPs that contain abundant fine-grained lowCa pyroxenes and glass. We suggest that the low-carbon, olivine-dominated anhydrous IDPs may be derived from anhydrous asteroids. While it is possible that this population could also come from comets, this population of low-carbon, olivine-dominated IDPs chemically resembles carbonaceous chondrite matrices (Table 8 ) . The Mg/ Si of olivine-rich IDPs is higher than for any group, while the Fe/Si ratio falls within the range for chondrites. The sulfur abundance is slightly higher than that of the carbonaceous chondrites. The major elements (Mg, S, Fe) and the lower carbon abundance of olivine-dominated IDPs are better chemical matches with carbonaceous chondrites than with comet Halley dust. Fe/Si ratios in carbonaceous chondrites are generally similar to Fe/Si ratios in anhydrous IDPs, but S is depleted in chondrite matrices relative to anhydrous IDPs (Table 8 and Fig. 4). However, the mineral chemistry in olivine-rich IDPs is not comparable to that of carbonaceous chondrites. While the fine-grained olivine in anhydrous IDPs tends to be. more forsteritic than the olivines in most carbonaceous chondrites, certain unequilibrated ordinary chondrites (e.g., Semarkona) contain finegrained olivine which is compositionally similar to that observed in anhydrous IDPs. Chondrites lack the fine-grained sulfides that are typical of IDPs. Based on our limited data set, we suggest that olivine IDPs generally are derived from asteroidal sources although we do not exclude the possibility

1562

K. L, Thomas et al

W FIG. 2. Three views of particle W7027Hl4. (a) SEM photomicrograph of the particle on a nucleopore filter, (b) TEM photomicrograph of the particle after embedding in epoxy and thin-sectioning. (c) Mineralogical map of the thin section shown in b. Carbonaceous material comprises 40-503’0 of the total area of this particle.

:)

1563

Carbon content of IDPs Table 8 Comparieon of Atomic Abundance Ratios to Silicon for Anhydroue IDPe and Other ExtraterrestrialMaterial0

Anhydroue IDPea

Anhydrous

Semarkona

IDPeb

Matrix=

Carbonaceous Chondrite

Comet Halleye

Matriceed

C

2.4

4.4

4.0

4.0

4.8

Na

0.10

0.05

0.12

0.02-0.04

0.05

Mg

0.8

1.0

0.5

0.9 -1.1

0.5

(0.5 - 1.8)

(0.6 - 2.0)

0.10

0.07

0.12

0.09-0.13

0.04

0.4

0.4

0.05

0.02-0.2

0.4

(0.2 - 1.3)

(0.1 - 1.8)

Ca

0.04

0.05

0.02

0.01-0.1

0.04

Cr

0.008

Mn

0.008

Fe

0.6

0.7

0.6

0.5 -1.0

0.3

(0.2 - 2.0)

(0.2 - 2.0)

0.02

0.02

0.03

0.04-0.06

0.02

2.0 (0.7 - 4.9)

0

Al S

Ni

that some carbon-rich, olivine-dominated particles could be derived from cometary sources. The mixed mineralogy IDPs show a range of carbon abundances from values indistinguishable from meteoritic to values comparable to the high-carbon, pyroxene-rich particles. The two mixed mineralogy IDPs with carbon abundances greater than 10% are probably derived from a cometary source. The remaining three mixed mineralogy particles could be derived from either asteroidal or cometary sources. We do not exclude the possibility that some C-rich IDPs could be derived from asteroids in the outer belt (e.g., P and D asteroids). P and D asteroids have colors and albedos that are comparable to some comet nuclei (e.g., comet Halley, CRUIKSHANK et al., 1985) and inactive comets ( CAMPINS et al., 1987; MILLIS~~al., 1988). Their low albedos and reddish colors in the visible spectrum are believed to result from the presence of C-bearing phases (FRENCH et al., 1989), although

the abundance of carbonaceous material is poorly constrained. In addition, it also appears that P and D asteroids are anhydrous and lack minerals which form in response to aqueous alteration (e.g., clays) ( LEBOF~KY et al., 1990). However, because of the large heliocentric distances involved and the low albedos of these objects, reflection spectroscopy measurements have low signal to noise ratios; thus little is known about the mineralogical composition of these asteroids. At this point in time, we believe it is speculative to propose that the outer asteroids are a major source of C-rich anhydrous IDPs, until more is known about their chemical and mineralogical composition. In summary, we believe that the high-carbon, pyroxenerich anhydrous IDPs are derived from cometary sources and that the low-carbon, olivine-rich anhydrous IDPs are produced from asteroids. We believe that the distribution of C abundances in the mixed-mineralogy particles indicates that

1564

K. L. Thomas etal. Table8.(Continued)

~66 modal% z-66 modal% Olivinef

Olivine ~66 modal%

~66 modal% Pyroxeneh

PyroxeneO

C

1.0

1.6

3.1

0

4.1

3.7

4.4

Na

0.05

0.12

0.08

WI

1.2

0.8

0.8

Al

0.05

0.13

0.09

0.3

0.3

0.6

Ca

0.03

0.05

0.04

CT

0.008

0.008

0.009

Mn

0.018

0.006

0.019

Fe

0.6

0.5

0.9

Ni

0.02

0.02

0.03

S

aThis work.

Nineteen particles.

'Schramm et al. (1989).

Ninety particles; thirty particles for carbon.

'Huss et al. (1981). dMcSween and Richardson (1977). (CI, CM, CO, CV). eJessberger et al. (1988). fThie work.

U2015E3, W7013C16, W7013Gl.

gThie work.

U2015D21, U2015D22, W7013E9, W7013E17, W701366.

hThia work.

W7013D12, 17027814, W7029*A28.

this group represents particles derived from both asteroidal and cometary sources. If true, then the low-carbon anhydrous IDPs are derived from very primitive (mineralogically) source(s) unlike the carbonaceous chondrites but perhaps resembling the fine-grained material in unequilibrated ordinary chondrites (e.g., Semarkona). Another possibility is that these low-carbon anhydrous particles are derived from a unique, dust producing, primitive asteroid(s) that has escaped secondary processing (i.e., aqueous and thermal alteration). While all nineteen particles could be derived from cometary sources, the distribution of carbon abundances indicates that they cannot all be derived from known asteroidal sources.

CONCLUSIONS We have determined the bulk composition of nineteen anhydrous IDPs including the light elements carbon and oxygen, we have studied their mineralogy, and we have identified a relationship between carbon abundance and silicate mineralogy. We find that, in general, the pyroxene-dominated IDPs have higher carbon abundances than do olivine-dominated particles. In fact, carbon in pyroxene-dominated particles is not chondritic and the average carbon abundance is >3 X CI. By our analytical procedure, the carbon abundance in olivine-dominated particles is indistinguishable from chondritic (~3 X CI). Particles with a nearly equal mix of

Carbon content of IDPs

Cl

1565

i

f *

‘5

0 -

-

-

ti

i 10

15

20

25

Carbon (wt%) * Error is the standard deviation of the average of fragments from the same particle FIG. 3. The modal silicate mineralogy of 11 thin-sectioned IDPs plotted against their carbon abundance. Although the particles are grouped by only three broad mineralogy groups, the correlation of these groups with the C abundance isevident. Olivinedominated IDPs: (Al) W7013Gl. (A2) W7013C16. (A3) U2015E3. Mixed mineralogy IDPs: (BI)

W7013G6. (B2) U2015D21. (B3) U2015D22. (B4) W7013E17. (B5) W7013E9. Pyroxenedominated IDPs: (Cl) W7013D12. (C2) W7029*A28. (C3) W7027H14. Points marked with an asterisk have larger uncertainties because the error is the standard deviation of the average of several fragments of the particle. olivine and pyroxene have either chondritic or non-chondritic carbon abundances and can be grouped with either the pyroxene or olivine-dominated groups based upon their carbon abundance. The anhydrous particles with non-chondritic (>3 X CI) carbon abundances have primitive mineralogies and 1.4 ,

1

A

1.3 1.2

PyroxeneDominated

A

IDPs

.

Dominated ICRs

1.0

Mixed IDPs

0.2

I

0.8

Acknowledgments-We thank J. Warren for assistance in sample preparation, E. K. Gibson for providing the Smithsonian Allende Reference Sample, and A. Dodson for preparing Fig. 2c. Constructive

Editorial handling: G. Crozaz REFERENCES

0

0.2

0.4

0.8

0.8

1.0

1.2

1.4

1.6

1.8

2.0

Atomic Fe/Si FIG. 4. Bulk sulfur content of eleven thin-sectioned IDPs plotted against the bulk iron content. The mean IDP value, comet Halley value, and a field of chondritic matrix values are shown for comparison.

etary dust from particles collected in the stratosphere.

reviews of the manuscript were provided by John Bradley, George Flynn, and Phil Fraundoti This work was supported in part by NASA RTOPs 152-17-40-23 and 199-52-11-02.

l OlMne-

1.1

bulk compositions and show strong similarities to COmetary dust particles. We believe that these high-carbon, pyroxenedominated anhydrous IDPs are the best candidates for com-

ALLAMANDOLAL. J., SANDFORDS. A., and WOPENKA B. ( 1987) Interstellar polycyclic hydrocarbons and carbon in interplanetary dust particles and meteorites. Science 237,56-59. ANDERS E. and GREVE~SEN. ( 1989) Abundance of the elements: Meteoritic and solar. Geochim. Cosmochim. Acta 53, 197-2 14. ARMSTRONGJ. T. and BUSECKP. R. ( 1975) Quantitative chemical analysis of individual microparticles using the electron microprobe: Theoretical. Anal. Chem. 47,2 178-2 192. BLANFORDG. E., THOMASK. L., and MCKAY D. S. ( 1988) Microbeam analysis of four chondritic interplanetary dust particles for major elements, carbon and oxygen. Meteoritics 23, 113- 12 1. BRADLEYJ. P. ( I988 ) Analysis of chondritic interplanetary dust thin sections. Geochim. Cosmochim. Acta 52,889-900. BRADLEYJ. P., BROH~NLEE D. E., and FRAUNDORFP. ( 1984) Carbon compounds in interplanetary dust: Evidence for formation by heterogeneous catalysis. Science 223,56-58.

1566

K. L. Thomas et al.

BRADLEYJ. P., BROWNLEED. E., SCHRAMM L. S., and DIETZ N. L. ( 1988) The abundance, distribution and chemical state of carbon in interplanetary dust. (Abstract) Meteoritics23,259-260. BRADLEYJ. P., GERMAN]M. S., and BROWNLEED. E. ( 1989) Automated thin film analyses of anhydrous interplanetary dust particles in the analytical electron microscope. Earth Planet. Sci. Left. 93, l-13. BRADLEYJ. P., HUMECKIH. J., and GERMANIM. S. ( 1992) Combined infrared and analytical electron microscope studies of interplanetary dust particles. Astrophys. .I. 394, 643-65 I. CAMPINSH., A’HEARN M., and MCFADDEN L. A. ( 1987) The bare nucleus of comet P/Neujmin I. Astrophys..I. 316, 847-857. CHRISTOFFXRSEN R. and BUSECKP. R. ( 1984) Mineralogy of platelet grains in carbon-rich CP interplanetary dust particles. (Abstract) Lunar Planet. Sci. 15, 152- 153. CHRISTOFEERSEN R. and BUSECKP. R. ( 1985 ) Epsilon carbide: A low temperature component of interplanetary dust particles. Science 222, l327- 1329. CRUIKSHANKD. P., THOLEN D. J., and HARTMANNW. K. (1985) Color, albedo, and nuclear size of Halley’s comet. Nature 315, 122. FLm G. J. and SUTTON S. R. ( 199 I ) Cosmic dust particle densities: Evidence for two populations of stony meteorites. Proc. Lunar Planet. Sci. Conf 21, 541-547. FRAUNDORF P. and SHIRCK J. ( 1979) Microcharacterization of Brownlee particles: Features which distinguish interplanetary dust from meteorites. Proc. Lunar Planet. Sci. Conf: lOth, 95 1-976. FRENCH L. M., VILAS F., HARTMANNW. K., and THOLEN D. J. ( 1989) Distant asteroids and Chiron. In Aster0id.sII (Ed. R. P. BINZEL,et al.) pp. 468-486. Univ. Arizona Press. Huss G. R., KEIL K., and TAYLOR G. J. ( 198 I ) The matrices of unequilibrated ordinary chondrites: Implications for the origin and history of chondrites. Geochim. Cosmochim. Acta 45, 33-5 I. JAROSEWICHE., CLARKER. S., and BARROWSJ. N., ed. ( 1987) The Allende Meteorite Reference Sample; Smithsonian Contributions IOthe Earth Sciences No. 27. JJXSFJERGER E. K., CHRISTOFORIDIS A., and KISSELJ. ( 1988) Aspects of the major element composition of Halley dust. Nature 332, 691-695. KERRIDCE J. F. ( 1985) Carbon, hydrogen, and nitrogen in carbonaceous chondrites: Abundances and isotopic compositions in bulk samples. Geochim. Cosmochim. Acta 49, 1707- I7 14. KISSELJ. and KRUEGER F. R. ( 1987) The organic component in dust from comet Halley as measured by the PUMA mass spectrometer on board Vega 1. Nature 326, 755-760.

[email protected] W., THOMASK. L., MCKAY D. S., and PALME H. (1989) Unusual olivine and pyroxene compositions in interplanetary dust and unequilibrated ordinary chondrites. Nature 339, 126-128. LEBOFSKY L. A., JONES T. D., OWENSBYP. D., FEIERBERGM. A., and CON~OLMACNOG. J. ( 1990) The nature of low albed asteroids from 3 pm multi-color photometry. Icarus 83, 16-26. MACKINNONI. D. R. and MOGK D. W. ( 1985) Surface sulfur measurements on stratospheric particles. Geophys. Res. Lett. 12, 9396. MCKEEGAN K. D., WALKERR. M., and ZINNER E. ( 1985) Ion microprobe isotopic measurements of individual interplanetary dust particles. Geochim. Cosmochim. Acta 49, 197 I-1987. MCSWEEN H. Y., Jr., and RICHARDSONS. M. (1977) The composition of carbonaceous chondrite matrix. Geochim. Cosmochim. Acta 41, 1145-l 161. MILLIS R., A’HEARN M., and CAMPINSH. ( 1988) An investigation of the nucleus and coma of comet P/Arend-Rigaux. Astrophys.J. 324, 1194-1209. MYKLEBUSTR. L., FIORIC. E., and HEINRICHK. F. J. (1979) Frame C: A Compact Procedurefor QuantitativeEnergy-DispersiveElectron Probe X-ray Analysis; NBS Technical Note 1106. US Dept. of Commerce/NIST. RIETMEIJERF. J. M. ( 1987) Silicone oil: A persistent contaminant in chemical and spectra1 micro-analysis of interplanetary dust particles. (Abstract) Lunar Planet. Sci. XVIII, 836-837. RIETMEIJERF. J. M. and MACKINNONI. D. R. ( 1985) Poorly graphitized carbon as a new cosmothermometer for primitive extraterrestrial materials. Nature 315, 733-736. RIETMEIJERF. J. M. and MACKINNONI. D. R. ( 1987) Metastable carbon in 2 chondritic porous interplanetary dust particles. Nature 326, 162-165. SANDFORDS. A. ( 1986) Acid dissolution experiments: Carbonates and the 6.8-micrometer bands in interplanetary dust particles. Science 231, 1540-1541. SCHRAMML. S., BROWNLEED. E., and WHEELOCKM. M. (1989) Major element composition of stratospheric micrometeorites. Metroritics 24, 99-I 12. THOMAS K. L., KELLER 1~. P., BLANFORDG., KL~CK W., and MCKAY D. S. ( 1992) Carbon in anhydrous interplanetary dust particles: Correlations with silicate mineralogy and sources of anhydrous IDPs. (Abstract) Lunar Planet. Sci. XXIII 1425- 1426. WOPENKA B. ( 1988) Raman observations on individual interplanetary dust particles. Earth Planet. Sci. Lett. 88, 22 l-23 1.