correlated morphological, chemical, and isotopic...

Geochimicu o Cwtwchimica Aclu Vol. 52, pp. 2827-2839 Copyright 8 1988 Pcrgamon Press pk. Printed in U.S.A.

0016-7037/88/$3.04 + .C’l

Correlated morphological, chemical, and isotopic characteristics of hibonites from the Murchison carbonaceous chondrite

TREVOR R. IRELAND Research School of Earth Sciences, The Australian National University, Canberra, ACT 260 1, Australia

and Physics Department and McDonnell Center for the Space Sciences,? Washington University, St. Louis, MO 63 130, U.S.A.

(Received December 18, 1987; accepted in revisedform August 11, 1988)

Abstract-Twenty-four h&o&e-dominated refractory grains from the Murchison carbonaceous chondrite have been identified from polished grain-mounts of density separates, and have been morphologically and chemically divided into three groups: (1) 14 colourless PLAty Crystal fragments (PLACs) with less than 2.7% TiOz; (2) 3 pleochroic Blue AGgregatea (BAGS) which are composed of crystal plates and fragments and have Ti02 concentrations between 5.1% and 6.5%; (3) Six Spinel-HIBonite spherules and crystal aggregates (SHIBs), often with spinel, Fe-silicate, and clinopyroxene rims. The maximum Ti02 concentrations of hibonite in the SHIBs ranges from 3.5 to 7.6%, but the TiO, concentration in one grain was heterogeneous and ranged from 0.5% to 5.0%. Hibonite in all types is stoichiometric with Ti and MR oredominantly in a COUDM substitution of M$’ and Ti’+ for 2Al’+. One arain with a rounded hibonite core does not &arly fit into <he three designated groups. -

The hibonites have been analysed by ion microprobe mass spectrometry for their Mg and Ti isotopic compositions. Twelve of the PLACs have (*6A1/2’Al)o ratios less than 1 X 10e5 but the other two PLACs have (26A1/27Al)o of 5.5 X lo-’ and 7.7 X 10m5. The excess 26Mg may reflect initial Mg isotopic heterogeneities in the source region of the PLACs, or be the result of in situ 26A1 decay, in which case the distribution of 26Al was heterogeneous. Three BAGS show substantial positive mass-fractionation of Mg. One grain has excess %Mg while the other two have 26Mgdepletions. The 26Mg depletions indicate the preservation of Mg isotopic anomalies in the BAGs, but the excess %Mg in one grain could be due to in situ “Al decay. The six SHIBs have excess 26Mg that is correlated with 27A1/2’Mg. The Mg isotopic systematics of the SHIBs are consistent with the in situ decay of 26Al with (26A1/2’Alb of ca. 5 X lo-‘. The unclassified hibonite has a clearly-resolved deficit in 26Mg. Titanium isotopic anomalies are common in the hibonites measured here. The largest anomalies are in ‘@Ti with variations in 5’?i/“‘Ti ranging from -50 to + 16’Rw relative to terrestrial. Smaller anomalies are present in ‘vi and “Ti. The Ti isotopic compositions are not directly correlated with morphological type, but the PLACs show the largest variations in isotopic composition, whereas the SHIBs are generally close to terrestrial.

The presence of hibonite populations with distinct morphological, chemical and Mg isotopic systematics suggests that there were several hibonite formation episodes within the early solar system. The hibonites probably formed by the melting of refractory dust aggregates during local transient thermal events.

1. INTRODUCIION

METEORUK HIBONITES [Ca(Al,Mg,Ti)120,9] contain the most anomalous refractory-element isotopic &natures yet measured. Titanium isotopic analyses have shown large “‘Ti anomalies with variations in ‘?i/“Ti from -7% to +lO% relative to terrestrial Ti (HUTCHEON et al., 1983; FAHEY et al., 1985a, 1987a; IRELAND et al., 1985; HINTON m al., 1987). Calcium isotopic measurements have shown anomalies in 48Ca which a~. Iar8e and of the same sign as the qi anomalies in the same grains (ZINNER et al., 1986). Magnesium isotopic compositions have shown excess 26Mg up to 400% (IRELAND and COMPSTON, 1987), however hibonites commonly show little or no excess 26Mg (FAHEY et al., 1987a, and references therein). The oxygen isotopic compositions of seven h&mites, with both positive and negative vi anomalies, were all found to be enriched in “‘0 by 4 to 7% relative to the terrestrial standard SMOW (FAHEY et al., 1987b).

Trace element compositions of hibonites have also been measured in an attempt to correlate chemical and isotopic characteristics of these grains (FAHEY et al., 1985b, 1987a; HINTON et al., 1985, 1987). These authors found that the

t Present address.

2827

largest Ti isotopic anomalies are present in hibonite with Group III rare earth element patterns, whereas no hibonites with Group II patterns had Ti isotopic anomalies. FAHEY et al. (1987a) also found that, except for one inclusion, the 8rains with Ti isotopic anomalies did not exhibit 26Mg excesses. HUTCHEON et al. (1986) noted that refractory inclusions containing hibonite as the only refractory phase rarely contain excess 26Mg, whereas n&actory inclusions containing hibonite associated with spine1 and/or melilite had (2aA1/27Al)o of ap proximately 5 X 10w5. Previous work has therefore suggested that there might be distinct populations of hibonite grains present, but the relatively smaII population of samples analysed has not allowed a clear definition of any correlated effects.

This paper presents morphological descriptions, chemical analyses, as well as Mg and Ti isotopic compositions, of 24 hibonite grains From the Murchison carbonaceous chondrite. Titanium isotopic compositions of eight of these hibonites were presented in IRELAND et al. (1985), and preliminary results of the remaining hibonites were presented in IRELAND (1987a). Mg isotopic compositions were presented in IRE LAND et al. (1986a,b) and IRELAND (1987a), but some of these Mg isotopic analyses were subject to analytical error (IRELAND, 1987b). Rare earth element and trace element abundances have also been determined and will be presented elsewhere (IRELAND et al., 1988).

1. R. Ireland

2. EXPERIMENTAL PROCEDURES

Sumple preparation

The Murchison hibonites examined in this study were separated from 67 g of small meteorite fragments. The fragments were crushed to pass a 400 pm mesh and then separated into 40-75 pm and 75- 400 Frn fractions. The coarse-grained fraction was density-separated in tetrabromoethane (p = 2.9 g cm-‘), followed by methylene iodide (p = 3.3 g cm-‘), to leave a dense fraction dominated by green crys- talline olivine. This fraction was washed in acetone, dried, and passed through a Franz isodynamic magnetic separator. The non-magnetic fraction was mounted in epoxy and polished for examination (mount #7).

The finer grain-size fraction could not be magnetically separated in the same manner and so an alternative magnetic separation method was used which had been developed for separating fine-grained evap orites (MCLEOD et al., 1985). The sample was centrifuged in tetrabromoethane, the heavy fraction was collected, washed in acetone. and dried. A 20 ml burette, modified with an enlarged stopcock, was mounted between the magnetic poles of the separator. The whole apparatus was rotated so that the burette was in a vertical position and the burette was filled with ethanol. The magnetic field was applied and a slurry of the sample mixed with ethanol was gradually passed down the burette. The magnetic fraction adhered to the walls of the burette, while the non-magnetic fraction passed to the bottom. The non-magnetic fraction was collected by opening the stopcock while maintaining the magnetic field and taking care that the ethanol did not fall to such a level as to disturb the magnetic fraction. After collection of the nonmagnetic fraction, the magnetic field was re- moved, the burette was drained, and the magnetic fraction was washed from the burette. This process was repeated five times with increasing fields, and the final non-magnetic fraction was centrifuged in methylene iodide. This high-density, non-magnetic concentrate was hand- picked for the characteristic blue of hibonite and several fractions of this were mounted in epoxy and polished (mount #IO).

Each mount was examined optically to identify grains suitable for further study. Wavelength-dispersive X-ray analyses were obtained with a Camebax Microbeam electron microprobe for the hibonites in mount 10, and energy-dispersive X-ray analyses were made on the hibonites in mount 7. Back-scattered electron images were produced on a JEOL 848 SEM. Mount 10 (fine-grained) had 7 I grains including 8 hibonite, 44 spinel-rich grains (with common inclusions of hibonite, perovskite, and occasionally melilite), I3 olivine, 5 pyroxene, and 1 rutile. Mount 7 (coarse-grained) had 1115 grains, the majority of which were olivine and Fe sulfide. Spine1 grains were again common and 20 hibonite-dominated grains were initially identified. Other refractory mineral species identified as major constituents of grains included corundum, rutile, perovskite, and ilmenite.

Mg isotopic analyses

Magnesium isotopic analyses presented here were carried out on the Cameca IMS-3f at Washington University with the analytical procedures described by FAHEY et al. (1987a). The Mg isotopic compositions are expressed in delta notation relative to the terrestrial values of 2sMg/24Mg = 0.12663, and 26Mg/2’Mg = 0.13932 determined by CATANZARO et a/. (1966). The observed Mg isotopic fractionation, Az5Mg, is given by

A2’Mg = 1000[(2sMg/24Mg),~0.12663 - l] (Y&/amu)

and includes both instrumentally induced fractionation as well as that which is intrinsic to the sample. Positive A”Mg is defined as an enrichment in the heavy isotopes relative to terrestrial Ivlg. No special effort was made to determine the intrinsic mass fractionation, FMg, of the hibonite grains because the presence of sputter holes on the surface could affect the instrumental mass-fractionation contribution. Nevertheless, the intrinsic mass fractionation of the Murchison hibonites was estimated by subtracting the mean A2’Mg of terrestrial standards measured during the course of this work (Table 3). The limit of detection of intrinsic mass fractionation is probably of the order of Sk/amu.

The measured 2nMg/24Mg ratio is corrected by a hnear mass-fractionation law and is expressed as the residual 62”Mg, where

6*‘Mg = 1000((2hMg/2’Mg),~~/~. I3932 - 1] - 2A”Mg (so).

The *7Alt/24Mg+ ion ratios were multiplied by the sensitivity factor of I .37 (FAHEY et ul., 1987a) to obtain 27Al/24Mg. The apparent initial (26Al/27Al),, for the hibonites is modelled by assuming that in situ 26Al decay is responsible for the excess 26Mg, and that the initial Mg isotopic composition is normal. The (26Al/27AI)o is therefore a model-dependent value and is not determined from an isochron in- corporating analyses of several different phases with differing 27Al/24Mg.

Magnesium isotopic compositions of hibonites from mounts 7 and 10 were initially reported by IRELAND et al. (1986a,b) from analyses which utilised the Faraday cup on the Sensitive High Resolution-Ion Microprobe (SHRIMP) at the Australian National Universitv. How- ever, Mg isotopic analyses of the PLACs on SHRIMP were subject to analytical error. The problem involved cross-talk between the Far- aday cup leadthrough and the pre-slit deflection plates which con- trolled peak-centering. This only affected analyses of the low-Mg PLACs with 24Mg+ signals below 2 X 1O-‘2 A. Replicate analyses of the SHIBs made on the Washington University Cameca IMS-3f and SHRIMP agreed within error.

Ti isotopic analysm

Titanium isotopic analyses were carried out on SHRIMP with standard operating procedures (IRELAND et al., 1986a). At a mass- resolving power of 7000 (M/AM, 1% valley) all molecular interferences, including hydrides, were resolved from the Ti isotopes. Only the atomic isobaric interferences from ?a, “V, and “Cr were un- resolved; these were estimated by monitoring 44Ca, 5’V, and 52Cr. Isotopic mass-fractionation of Ca and Cr was found to be relatively low in the phases examined, and so the terrestrial ratios 46Ca/“Ca = 0.00152 (NIEDERER and PAPANASTASSIOU, 1984), 50V/5’V = 0.002503 (FLESCH et al., 1966) and “Cr/“Cr = 0.05 1859 (SHIELDS, 1966) were used. Isobaric corrections for atomic isobaric interferences were generally less than 10%.

Sufficiently flat-topped peaks were maintained to ensure that the contribution of the isobaric interferences was not compromised when the 46Ti and “Ti peaks were centered in the collector slit. 48Ca was fully separated from 48Ti when the latter was centered in the collector slit. In addition, because the intensity of 48Ca is low relative to “Ti in the hibonites, there is no observable contribution (<IO-*) to the 48Ti count due to scattered 48Ca ions. The mass-resolving power is the same as that of FAHEY et ul. (1985a, 1987a) which was defined therein as 13,000 M/AM.

During data acquisition the peaks were automatically centered into the collector slit by electrostatic deflection coupled with a magnetic field trim. The signal intensities were generally measured on the ion counter; however, measurements were also carried out with the Faraday cup when the 4sTi+ signal exceeded lo-” A. This was the case for minerals with high Ti02 concentrations such as rutile, brookite, ilmenite, and sphene.

The relative abundances of the Ti isotopes measured on SHRIMP from terrestrial mineral phases differ from the accepted isotopic abundances because of isotopic mass-fractionation. Normalisation of the isotopic ratios is designed to remove the effects of mass-fractionation and produce more readily comparable ratios. HEYDEGGER ef al. (1979) used 49Ti as the denominator mass with mass fmction- ation corrected from the observed 4’Ti/@Ti ratio. In ion microprobe mass spectrometry this has the advantage of having the reference isotopes free from isobaric interferences and variations in dead time affect only the 48Ti/“9Ti ratio to a significant degree_However, the ion microprobe measurements of FAHEY et al. (1985a) indicated that the two most anomalous isotopes in the h&mite samples they examined were ‘@Ti and 4%. Hence the hibonite analyses of IRELAND et al. ( 1985). which used the “‘?i normahsation. resulted in correlated effects on all isotopic ratios and anomalies being artificially induced on 846Ti and b4’Ti. Therefore the Ti isotopic analyses presented here have been normalised to 46Ti/‘8Ti to correct for mass fractionation. This scheme has the advantage of having two shielded isotopes as

Hibonite from Murchison 2829

references (NIEDERER et al., 1981), and the most abundant isotope as the denominator. However, in the ion counting mode, it does have the disadvantage of making all ratios reliant on the determination of counting system dead time.

The data for the Murchison hibonites and perovskkes were collected at a ‘*Ti count rate of approximately 1.5 X 106 cts/sec and at this count rate, a I ns bias in the dead time determination would produce a ca. 1.5% offset in the isotopic ratios. The dead time of the ion- counting system was determined by Ti isotopic measurements on terrestrial samples before and between analyses of the meteoritic hibonites. The four measured isotopic ratios were used to determine the instrumental isotopic mass-fractionation and counting-system dead time. The error in the dead time determined by this procedure was approximately 0. I ns which corresponds to an error of ca. 0.15960 in the normalised Ti isotopic ratios.

The titanium isotopic mass-fractionation, A&Ti, for the measured (46TipTih,, is expressed in delta notation relative to the standard 46Ti/‘% ratio of 0.108548 (NIEDERER et al., 198 l),

A’% = -1000/2[(~Ti~sTi),~0.108548 - l] @/amu)

where positive AtiTi is defined as enrichment in the heavy isotopes. The intrinsic Ti-isotopic mass-fractionation for hibonite, FTi, is the measured A’% corrected for matrix-dependent instrumental mass- fractionation as determined on the terrestrial Madagascar hibonite, and is approximzited by the relationship

FTi = (AtiTl)mm - (Aqi)Md Hib

where (A&Ti),, is the measured fractionation for the meteoritic hibonite and (A’6TihtiHib is the mean A46Ti for the Madagascar hibonite during each data collection period.

The measured “Ti/?i, ‘%/4*Ti, and qi/‘*Ti ratios were corrected for isotopic mass-fractionation by a power law of the form

[(46Ti/48Ti),,,J0. 108548]” = [(~i/48Ti),,,&Ti/48Ti)WJ

for i = 47,49, and 50, and m = (48 - 1~12. Delta values were obtained from

6’Ti = IOOO[‘R,,~R,., - 1] (960)

where ‘& is the normal&d %/‘*Ti ratio corrected for mass fractionation and ‘R,, is the terrestrial value for that normalised ratio. The standard ratios were obtained from a suite of terrestrial samples which Were measured on the Faraday cup to avoid any problems associated with counting-system dead time (Table 1). Is&ark inter- ference conzc$ons were less than 1% except for the %?a correction in perovskite which was 2.6%. Each analysis represents the mean of ten ratios with 10 s count times on each peak. The grand mean vaties. obtained with a ‘6Ti/‘?i of 0.108548. were “Ti/‘% = 0.099309 + 0.~~12 (2a), 49ki/‘?Ti = 0.074404’+ 0.000015~ and ‘%/‘*Ti = 0.072389 + 0.000022, which differ by -0.06 + 0.12%0, -0.79 + 0.21960, and -0.40 f 0.30%0, respectively, from the terrestrial valuesof NIEDERER eta/. (1981).

The three Ti-metal samples have a range in A’% beyond that expected from the relatively constant matrix-dependent mass-frac+ tionation. This is probably due to real variations in the intrinsic fractionation of the samples because they come from a getter pump which operates by subliming a Ti-metal charge. TI 1 comes from a relatively fresh section of a Ti charge, T12 is a sample of Ti that was condensed onto the vacuum-pump housing, and T13 comes from a sample of an almost exhausted Ti-metal charge. The relative fractionations are consistent with TI2 being an isotopically-light condensate derived from TI 1, and T13 an isotopically-heavy residue.

The Ti isotopic data set for the Murchison hibonites was obtained in three data collection periods. The first two periods were separated by two weeks and the obtained data were presented by IRELAND et al. ( 1985). Isotopic measurements of terrestrial standards were consistent with a constant, but slightly different, dead time for the counting system during each period. The data on interspersed standards of the third collection period indicakd a systematic increase of the dead time by approximately 0.05 ns/hour. The cause of this change in multiplier characteristics is unclear, but may be a result of age. The dead time for the third period was determined by fitting a linear

Table 1. Ti isotopic analyses of terrestrial samples.

Samplet A‘%?$ S”Tii2t~ S4?ii (%&mu) (%o) cow

TIl -11.3 -1.llf1.26 -0.36kO.75 -11.9 -0.6Sf1.24 0.7OkO.94 -12.0 -0.56k1.22 -0.67Y2.35

T12 -15.8 -0.31M.80 0.69k1.36 -16.0 -1.76f3.68 0.8M1.58

T13 -8.5 -0.29zk1.48 0.18zk2.16 ILM -20.7 -0.26kO.64 0.66k2.08

-19.9 0.62kO.90 0.3010.88 -20.7 1.04f2.24 2.96kS.64 -19.8 O.lO+zO.76 0.33fl.40

RUT -21.7 0.3CkO.46 -0.49f1.32 -22.2 -0.03kO.37 -0.llM.39 -22.1 -0.19zkO.24 0.23M.94 -21.9 O.llM.35 -0.16kO.26 -21.8 0.19kO.23 -0.13kO.19 -21.7 0.25kO.27 0.01M.26 -21.5 -0.08kO.U 0.08fl.78

0.16kO.66 1.29k2.20 1.48k3.56 1.58ti.66 2.07ti.34 1.2Sk2.14 l.39k2.09 0.53f1.52 3.53f7.04

-0.46zt1.84 -0.99k1.72 -0.lW.52 O.OSM.92

-0.46ztO.52 -0.39kO.56 -0.02ztO.61 0.16?1.10

-21.0 0.3010.43 0.01?0.57 -0.36k1.11 -21.3 0.06s.66 0.01fo.68 -0.ow.94 -21.1 -0.24M.46 -0.llkO.84 -0.59kO.69 -21.3 -0.09kO.48 -0.2Sfl.24 -0.351t2.25 -20.8 0.33rkO.70 -1.03M.62 -1.31f1.14 -20.9 O.lW.21 -0.65kO.65 -0.79zt1.36 -21.4 0.27kO.49 0.76k1.48 0.69i1.34 -21.1 -0.21+0.45 -0.58kO.89 - 1.34k2.38 -21.2 0.3Ok1.08 -0.08kO.90 -0.4Sf1.43 -21.0 0.59+0.82 -0.41kO.75 -0.9Sf2.34 -20.9 O.lSM.90 O.llk1.28 0.04k2.17 -20.8 -0.253.38 -0.73M.97 -0.98f 1.74

PVS -27.2 0.79rt1.32 0.04k1.63 -0.74k2.86 -27.0 0.21f1.48 0.6Sf1.57 1.5Sf2.18 -26.8 0.69H.97 1.66k1.71 0.6Sf2.67 -27.8 -0.01+1.27 0.74ti.71 -0.17M.54 -25.5 0.58f1.18 1.29zt3.81 1.2Ok5.01 -25.5 0.33f1.64 1.222.12 2.74k2.34 -25.9 0.97ti.82 2.77k5.98 3.34f7.10 -25.8 0.86k2.70 1.23f3.86 1.524.06 -25.5 0.54f1.14 0.01+2.36 0.48+2.18

tTIl,2,3, = Ti metal; ILM = ilmenite; RUT = rutile; PVS = perovskite.

$A44&ri = -1000/2[(~~%),J0.1@8548 - l] (%~amu)

positive fractionation I heavy enriched

ratios corrected by power mass fractionation law:

[(‘?i/@Ti),~0.108S48]m = [(‘TiP&Ti)~(‘Ti/48Ti)Mnl

where m = (48-i)/2

, 8Ti = 1CC10[(‘Tii8Ti)_/(‘Ti/%), - l] (%xI)

Terrestrial ratios: 47Tii% = 0.099309k12,

4%/48Ti = 0.074404f15, ‘@W4*Ti = 0.07238%22

least-squares regression to the data from the interspersed standards as a function of time.

Measurements with the ion counter of the krrestrial standards (Table 5) show good agreement with the terrestrial values determined on the Faraday cup. The mean 6”Ti is -0.11 f 0.1 I %O (2u), the mean b’% is 0.04 + 0.2 1960, and the mean Syi is 0.22 + 0.27960. The reduced chi-square for these ratios are 0.90, 0.6 1, and 0.80, respectively, indicating that the variations are consistent with the cited errors.

3. MORPHOLOGY AND CHEMISTRY

The twenty-four hibonite-bearing refi-actory grains analysed in this study have been morphologically and mineralogically divided into three categories. The PLACs are composed predominantly of single, colourless PLAty Crystal fragments of hibonite. The BAGS, for Blue AGregates, are. small unusual

2830 7‘. K. Ireland

grains composed predominantly ofblue hibonite. The SHIBs consist of Spinel-HIBonite-perovskite grains which show a wide variety of textures.

Eleven of the fourteen PLACs are single crystal fragments of hibonite. They range in size from 70 to 150 Km in the longest dimension and appear to be broken fragments from once larger inclusions. They are colourless to light-blue in colour. No inclusions were observed in these hibonite grains by optical microscopy, however the SEM revealed thin films of perovskite within the hibonite of a number of the PLACs (Fig. IA). Remnants of Fe-silicate rims were occasionally preserved on some of these fragments.

Three other grains are classified with the PLACs. Grain 7- 644 is a single crystal fragment of hibonite (100 Km long) but the outer edges of the crystal have been replaced and pseudomorphed by spine1 and perovskite. The spine1 and perovskite appear to be secondary phases and so this grain was not classified with the SHIBs. Grains 7-4 12 (Fig. 1 B) and 7-658 are hibonite crystal aggregates with Fe-silicate rims. Grain 7-4 12 also has Fe-silicate, clinopyroxene, and Fe sulfide infilling voids between hibonite laths, however, the fine grain size of the material precluded analysis of individual phases. Fe-silicate and clinopyroxene rims from Murchison refractory grains have been described previously by MACDOUGALL (198 1). The Fe-silicate is poorly characterised but was described by MACPHERSON et al. (1984) as a phyllosilicate based on low electron microprobe totals, its poor crystallinity, and its similarity in composition to the matrix. The Fe silicate and clinopyroxene rims are best preserved on the SHIBs which are described below.

Three Blue pleochroic hibonite AGgregates (BAGS) comprise the second hibonite group. They are less than 80 pm in diameter and appear to be composite grains of hibonite plates and fragments. Grain lo-43 (Fig. 1C) is composed predominantly of crystal plates of hibonite, while 10-3 1 and lo- 6 1 are composed of smaller 2 to 20 pm hibonite crystal fragments. These grains have distinctive scalloped grain margins and a characteristic adamantine lustre. Perovskite grains less than 2 pm in diameter are common, especially in grain lo- 3 1. There is no spine1 associated with the BAGS which dis- tinguishes them from the SHIBs.

The SHIBs comprise mineralogically complex grains composed of hibonite, spinel, and perovskite. They range in size from 50 to 300 km and display a range of morphologies from compact spherules (Fig. lE), to rather porous crystal aggregates(Fig. 1G). Grains 7-19, 7-143 (Fig. lD), 7-170 (Fig. lE), and 7-789 are spherules and compact aggregates with cores of hibonite laths intergrown with spinel. These grains are surrounded by rims of spine1 (+ perovskite), Fe silicate, and finally clinopyroxene (Die5 to D&o). The Fe-rich silicate is also observed cutting across the spine1 layer and filling the interstices between hibonite laths; void space between the laths is also common. Grain 7-664 is a porous aggregate of hibonite, spinel, and perovskite (Fig. 1F). Grain 7-953 is a rather fluflj~ aggregate of hibonite blades and perovskite grains (Fig. 1G). Spine1 is most common around the outer margins of the grain where it pseudomorphs hibonite. The remnants of an Fe-silicate rim are also present.

The remaining hibonite grain, 7-753, is unusual in that it has a distinctively rounded hibonite core (Fig. IH). A few

perovskite inclusions are present and it is rimmed by Fe silicate and clinopyroxene. This grain has no clear affiliation to the three groups described above although it does show some similarities to the two PLACs with rims. This grain has been left unclassified and is designated NC in Figs. 2, 3, and 4.

Chemical compositions of hibonite within the refractory grains are presented in Table 2. A reconnaissance analysis of SHIB 7-19 indicated a Ti02 concentration of -5.5% which is within the range shown by the other SHIBs. The hibonite compositions are stoichiometric with MgO and TiOz concentrations ranging from near detection limits up to 3.7% and 7.6%, respectively. The hibonite compositions presented in Table 2 appear to be dominated by a coupled substitution of

Ti“+ + Mg” = 2A13+

as indicated by the similar proportions of Ti4+ and Mg2+ cations and the stoichiometric cation totals (ALLEN et al.,

1978). However, substantial amounts of Ti’+ may be present as indicated by direct measurements of Ti3+ by electron spin spectroscopy (BECKETT et al.. 1988). These authors found that 23% of the Ti in the Murchison hibonite inclusion SH- 7 is present as Ti3+.

A histogram of the maximum Ti02 concentrations in the hibonite by morphological type is presented in Fig. 2. The PLACs show a restricted range in MgO and TiOZ concentrations from near detection limits to 1.2% and 2.6%, respectively. The three BAGS have 2.6 to 2.9% MgO and 5.1 to 6.5% Ti02. The six SHIBs are character&d by relatively high concentrations of MgG and TiOz, with MgO between 2.2 and 3.7%, and Ti02 between 3.6 and 7.6%. The minor elements are not always distributed uniformly through the hibonite of the SHIBs. The TiOz content of hibcmite in the core of 7-170 ranges from 0.5 to 5.0%, with a concomitant variation in MgO from near detection limits up to 2.5%. The unclassified grain 7-753 has hibonite with MgO of 0.9% and Ti02 of 1.9%. It would appear that while the chemical composition cannot be used to classify a given hibonite into one of the morphological groups, there does appear to be a correlation between the hibonite groups and the minor element chemistry.

Refractory grains with close affinities to the PLACs and SHIBs have previously been described in the literature. MAC- DOUGALL and PHINNEY (1979) located a single 120 pm PLAC-type hibonite crystal in Murchison matrix. MAC- PHERSON et al. (1980) recovered a single crystal from a dis- aggregated sample of Murchison which became the first of the so-called DJ hibonites (MACPHERSON et al., 1983). HUTCHEON et al. (1980) described a large inclusion (MUCH- 1) which was composed of bladed hibonite crystals. The similarity in physical appearance and chemical and isotopic compositions of DJ hibonite crystals and MUCH-l suggested that MUCH-type inclusions were a likely source of DJ hibonite fragments. The SHIBs are anakrgous to spinel-hibonite inclusions described by MACDOUGALL (1979, 1981) and MACPHERSON et al. (1983, 1984).

The classification used in this study is not intended to be definitive for all hibonite-bearing grains that have been described from Murchison; the groups are based solely on the


FIG. 1. Representative back-scattered-electron images of Murchison hibonites. (A) IO-70 is a typical PLAty Crystal fragment, or PI&C. Sputter holes are apparent on the surface of the grain and perovskite arms as the bright fitm within the hibonite; (B) PLAC7-4 12 is a ~~ni~~~~i aggregate with well-preserved Fe-silicate rims and Fe silicate, Fe sulfide, and clinopyroxene filling interstices; (C) lo-43 is a Blue AGgregate, or BAG, composed of hibonite crystal plates. The BAGS have a distinctive adamantine lustre and scalloped margins; (D) 7-143 is a Spinel-HIBonite (SHIB) compact aggregate of hibonite surrounded by spine1 and Fe-silicate. Fe-silicate also fills interstices between hibonite crystals although void spaces are common; (E) 7- 170 is a SHIB spherule with a hibonite core and is rimmed sequentially by spine1 (+ perovskite), Fe silicate, and clinopyroxene. Void spaces and Fe silicate between hibonite laths are again common; (F) 7-664 is a SHIB crystal aggregate composed of hibonite blades and anhedral perovskite grains within spinel. Dark areas are voids. (G) 7-953 is a SHIB crystal aggregate composed of fine hibonite blades with common perovskite inclusions. Spine1 pseudomorphs after hibonite are apparent around the rim; (H) 7-753 was not categorised. It has a rounded hibonite core with perovskite inclusions and is surrounded by Fe-silicate and chnopyroxene. [We bar in all photographs = 10 rm; Legend: hi = hibonite, sp = spine& pv = perovskite, fe = Fe-silicate, cp = clinopyroxene, fs = Fe sulfide.]

2832 T. R. Ireland

co.5 1 2 3 4 5 6 7 z-7.5

%TiO,

FIG. 2. Histogram of TiOl concentrations of Murchison hibonites according to group.

24 grains analysed here. The SHIBs comprise a variety of texturally diverse grains, however, owing to the limited number of SHIBs presented here, further subdivision is unwar- ranted at this stage. The number of groups required could increase with the analysis of further hibonite grains, but the three groups proposed here provide a useful initial subdivision.

4. ISOTOPIC COMPOSITIONS

Mg isotopic compositions

Mg isotopic compositions, 21A1/24Mg ratios, and apparent initial (26A1/27Al)o ratios for 20 of the hibonites am presented in Table 3. The Mg isotopic systematics of 7-97 1 were previously described by IRELAND and COMPSTON (1987), and the compositions of 1 O-3 1 and lo-6 1 were reported by IRE- LAND er al. ( 1986a). There was insufficient sample remaining for analysis of 10-52.

20

. ‘o-5 mi , x,o~5 .,.,,,,,.,..........

.. ,_,,........... ..-- _:~:::~~::::~~~::::~:::::..,--n_....- ,,_......

4 4 1.

0 Ii0 zoo

“‘A 1 /24Mg

FIG. 3. Al-MO iscchron diagram for hibonites analyszd in this study. The PLACs generally show high Al/Mg and low excesses of 26Mg. The SHIBs have relatively low Al/Mg and the correlation of excess 26Mg with 27A1/24Mg is indicative of in situ decay of 26Al with an initial 26A1/2’AI of ca. 5 X 10e5. One unclassified hibonite is distinctive in having negative 626Mg. [Error bars are 2a.l

N(’ /

FIG. 4. The Ti isotopic compositions of the three morphological groupings all show a range in 6’@Ti. The PLACs show the greatest spread, from -47 to + 16969, while the SHIBs have compositions that are close to normal except for 7- 170.

The range in 626Mg for all of the PLACs is from - 1.38 (k1.90) to +16.05 (+1.66)460 (Table 3). Ten of the PLACs show only small excesses of26Mg (<5%0) despite having high 27A1/24Mg ratios; the (26A1/27Al)o for these hibonites is less than 1 X lo-’ (Fig. 3). PLACs 7-412 and 7644 are distinctive in that they have the largest excesses of 26Mg (+ 16.1 St 1.7Y~ and + 11.1 k 2.0960) and have (26A1/27Al)o ratios of 5.5 X lo-’ and 7.7 X 10e5, respectively. Hibonite 7-97 1 has large variable excesses of 26Mg up to 400%, but the maximum (26A1/27Al)o required is less than 1 X 10V5 (IRELAND and COMPSTON, 1987). The Mg in grain 7-97 1 was also reported to be signif- icantly mass-fractionated, with the heavy isotopes enriched in the core of this grain by ca. +90960/amu relative to terrestrial Mg. However, the range in FM, for the other PLACs in Table 3 is from -3.2 to +2.6!&/amu, and so there is no evidence for large mass fractionation effects (> 10460/amu) in these grains.

Mg isotopic compositions of the three BAGS presented in IRELAND et al. (1986a) have positive mass-fractionation and variable residual 626Mg (Table 4). A replicate analysis of grain lo-43 on the Cameca IMS3f yielded FM8 of + 11.6Y&/amu and a 626Mg of -0.36 + 1.30%0 (Table 3). This is consistent with the SHRIMP analysis given the analytical uncertainty of the FM, determination.

Mg isotopic analyses of the SHIBs show 626Mg values that correlate with 27A1/24Mg (Table 3, Fig. 3). The excess 26Mg for individual grains is consistent with (26Al/27Al)o of ca. 5 X 10-s and the projected intercept of the correlation is close to the normal Mg isotopic composition. The 27Alf24Mg ratios are in general lower than expected for hibonite; the minimum 27A1/24Mg should be ca. 19 for stoichiometric hibonite with 9% Ti02 and 4.5% MgO. The lower 27A1/24Mg ratios are due to the presence of fine lamellae of spine1 (27A1/z4Mg = 2.5) within the hibonite that could not be excluded from the ion microprobe analyses. The correlation for the SHIBs in Fii. 3 is therefore not strictly an isochron, but a mixing line between hibonite and spinel compositions. The Fhle values for the SHIBs range from -0.4 to +5.49&o/amu which is close to the limits of detection in this work.

The unclassified hibonite 7-753 has a 626Mg of -4.74 + 2.28% (2~) (Table 3), which is more than 4a below the terrestrial 26Mgf24Mg. There is no detectable intrinsic Mg isotopic mass-fractionation in 7-753.

Hihonite from Murchison 2833

Table 2. Major element electron miCroprobe analysts of Murchisun hibenitcs.

Type Al203 CaO MgO Fe0 SiO2 T102Toral Al Cn Mg Fe SI TI TWII

weight % cationsl[l9 Oxygen]

7-29 PLAC 87.30 8.33 1.09 nd 1.06 1.34 99.12 Il.566 I.003 0.183 0119 0.113 12.984

7-48 PLAC 89.72 8.54 0.98 nd nd 1.78 101.02 11.679 I.010 0.162 0.148 12.999

7.51 PLAC 88.44 8.25 I.00 nd nd 1.73 99.42 Il.673 0.990 0.166 0.146 12.975

7-143 SHIB 79.41 8.64 3.65 nd nd 7.31 99.01 10.707 1.059 0.623 0.628 13.017

7.170 SHIB 84.12 8.61 2.76 0.21 nd 4.99 100.69 11.087 I.031 0.460 0.019 0.420 13.017

7412 PLAC 89.45 8.56 1.20 0.25 nd 1.63 101.09 II.651 I.014 0.198 0.023 0.135 13.021

7-W PLAC 89.20 8.48 0.95 nd nd 1.24 99.87 11.736 1.015 0.158 0.104 13.013

7-658 PLAC 88.76 8.44 1.06 0.15 nd 1.59 1OO.M) 11.688 I.010 0.177 0.014 0.133 13.022

7.664 SHIB 79.73 8.80 3.48 nd nd 7.57 99.58 10.694 1.073 0.590 0648 13.005

7-753 NC 88.49 8.&l 0.85 nd nd 1.86 99.80 Il.676 I.031 0.141 0.156 13.004

7-789 SHIB 83.08 8.40 2.22 nd nd 3.56 97.26 11.247 1.033 0.380 0.308 12.968

7.953 SHIB 81.62 8.15 3.34 0.23 nd 6.53 99.87 10.878 0.987 0.563 0022 0.555 13.06

7.971 PLAC 91.48 8.38 0.14 nd nd nd 10000 11.985 0.999 0.022 13.007

7-980 PLAC 89.56 8.46 0.79 nd nd I I6 99.97 II.761 I.010 0.131 0.097 12.999

7-981 PLAC 89.66 8.46 0.73 nd nd I.19 100.04 11.778 I.011 0.122 0.100 I3011

IO-20 PLAC 88.58 8.91 0.58 nd nd I.78 99.85 11.688 I.069 0.097 0.150 13.ow

IO-31 BAG 83.28 8.46 2.63 nd nd 5.14 99.51 II 105 1.026 0.444 0.437 13.012

IO-43 BAG 82.09 8.51 2.86 0.04 nd 6.44 99.90 10.927 1.030 0.483 0.004 0.547 12.991

IO-52 PLAC 87.90 8.55 0.96 nd nd 1.96 99.37 11.652 1.030 0.162 0.165 13.00!!

lo-54 PLAC 87.67 8.68 0.90 nd nd 2.62 99.87 11.578 1.042 0.151 0.220 12.991

IO-55 PLAC 88.29 8.60 0.51 nd nd 1.76 99.16 Il.719 1.037 0.086 0.149 12.991

lo-61 BAG 82.60 8.66 2.89 nd nd 6.34 100.49 IO.935 1.042 0.485 0.536 12.998

10-70 PLAC 88.72 8.67 0.66 nd nd 1.83 99X8 11.695 1039 0.110 0.154 12.‘!98

nd = not detected

Titanium isotopic compositions

The original Ti isotopic data from the 8 hibonites reported in IRELAND et al. ( 1985), and data from a further 15 hilxmites from mount #7 have been normalised with the 46Ti-48Ti pair to correct for isotopic mass-fractionation (Table 5). Hibonite 7-97 1 was not analysed because of large isobaric interferences from %Za, “V, and “Cr. Isotopic anomalies are common in

Table 3. AI-MI isotopic analyses of Murchison hibonites

SHIB -8.80 f 1.62 PLAC -10.88 f 1.30 PLAC -I 1.55 f 0.70 PLAC -14.01 f0.88 SHIB -12.23 f 0.50 SHIB -10.77 f 0.94 PLAC -15.08f 2.50 PLAC -9.29 f 1.02 PLAC -11.17f0.58

7.19 7-29 7-48 7.51 7.143 7.170 7-412 7-644 7-658 7-664 7-753 7.789 7.789A 7-953 7-980 7-98 I IO-20 IO-43 IO-54 IO-55 IO-70

SHIB 18.71 + 0.44 NCt -11.64f I.00 SHIB -8.30 f 1.34 SHIB -6.44 + 0.84 SHIB -10.07 f0.66 PLAC -9.91 f 0.62 PLAC -9.24 + 0.74 PLAC -13.365 1.20 BAG -0.30 + 2.04 PLAC -11.45f 1.08 PLAC -9.67 f 1.60 PLAC -15.25f 1.00

3.0 6.04 f 1.50 15. I7 f 0.28 5.55 f 1.48 1.0 -1.38 f 1.90 148.65 f 4.42 =a.05 0.3 2.56 f 1.08 107.72 f 1.40 0.33 f 0.16

-2.2 4.OOf 1.38 126.61 f 9.38 0.44 f 0.16 -0.4 2.73 f 0.92 5.27 f 0.26 7.22 f 2.80 1.1 l.76f 1.28 3.64 f 0.76 6.74 f 5.10

-3.2 16.05+ 1.66 40.10 f 3.40 5.58 + 0.58 2.5 lI.Wfl.96 19.97 f 1.30 7.72 + 1.38 0.7 4.58 + 1.94 65.73 f 3.76 0.97 t 0.42 3.1 3.47f 1.10 5.53 f0.16 8.74 f 2.78 0.2 -4.74 + 2.28 30.89 ?r 2.04 3.5 12.72 + 2.16 37.32 f 0.66 4.75 f 0.82 5.4 14.04 t 1.48 44.14 t 0.98 4.43 + 0.48 1.8 5.20 zt 1.24 13.88 * 0.76 5.22 f 1.28 I.9 l.87? 1.48 151.50f4.60 0.17+0.14 2.6 0.81 + 1.82 138.90 f 0.54 0.08f0.18

-1.5 2.46? 1.64 42.85 + 2.10 0.80 f 0.54 11.6 -0.36 k 1.30 12.99 f 5.40 <I.01 0.4 0.63 + 1.52 92.04 f 1.06 0.10 .+ 0.24 2.2 -0.35 + 2.10 125.34 f 2.72 <O.l9

-3.4 3.09+ 1.82 154.17f4.30 0.28 C 0.16

Melilite -13.6f 1.4 o.cNl * 0.95 3.89 f 0.06 -11.3+ 1.3 0.47 * 0.99 4.26 f 0. I3

-9.8 + I.2 0.92 f 1.05 4.34 f 0.06 -15.0f I.6 0.38 k 0.37

Olivinc n.m.$

-13.9f 1.6 0.01 f 1.26 <O.Ol Hibonite -8.9f 1.2 -0.70f 1.10 61.05 f 1.86 Spine] -13.4f 1.8 0.50 f 0.53 n.m.$

-8.9f 1.2 -0.27 f 0.81 2.65 + 0.03

Mean -11.9f0.9 EO.0 0.25 f 0.26

@Mvlg = I@WaMp/24Mg)M.12663 - I] (%&mu)

F ME = AaMg - (@M&,.nT~n~.tia

@*Mg = lCW(26Mp/idM~)KI.13932 - I] -2A.“Mg (%a)

(=A@Al), = [626Mg/1000 x 0.139321 /(*‘Ap4Mg)

All crmrs an k 20

tNC = not classified $n.m. = not measured.

Table 5 and only one hibonite out of the 24 analysed has all three normalised Ti ratios within 95% error limits of the terrestrial Ti isotopic composition. The hibonite data set shows that the two heaviest isotopes 49Ti and “?i are the most anomalous with variations from -6.5 to +6.5% and -49.9 to +16.8%0, respectively, while 647Ti shows smaller effects from -2.2 to +3.3%. Intrinsic mass-fractionation, FTi, in the meteoritic hibonites covers a narrow range from - 1.9 to +2.4%/amu. It is unclear if this is a real variation in the intrinsic mass fractionation, since this range is small compared with the total isotopic mass-fractionation (instrumental + intrinsic) of cu. 18L/amu.

Hibonite 7-170 is unusual in that despite having a large negative 6”Ti, it has a positive 649Ti. A replicate analysis of this hibonite was carried out on the Cameca IMS-3f at Wash- ington University following the procedures described by FAHEY et al. (1987a). The Ti isotopic analysis yielded a4’Ti = -2.94 + 1.51% (2a), 649Ti = +1.86 f 1.55960, and 65@Ti = -54.70 & 1.86%. This analysis verifies the absence of a significant deficit in 49Ti despite the large negative 6%. The 65?i values differ beyond the cited errors and it is unclear whether this is due to isotopic heterogeneity or if it is an analytical artifact of isotopic measurements on two different mass spectrometers.

IRELAND et al. (1985) carried out replicate analyses of grains 10-20, 10-55, and lo-70 in order to investigate the isotopic homogeneity of the hibonites and found that the replicate analyses agreed within analytical uncertainty. How-

Table4Mg Lsotopicanalyses ofBAGs+.

Sample

IO-31 10.4 i 0.4 9.14 * 0.77 IO-43 17.0 * 0.3 -1.79 f 0.82 IO-61 16.9 + 0.2 -2.50 * 0.39

t from Ireland et a!. (1986a).

2834 T. R. Ireland

Table 5. Ti isolopic anafyses of Murchison hibonites.

Sample A4%20 F.r,? 647TikZo (“iw/amu) (a~wkmu) (O/w)

Ei4’Ti.k2a 8SoTti2a %a$ “V$ %r$ t& (~~~) Pm) (%I) (%J) (%I) (ns)

I MadHib-17.69k0.14 0.18ti.42 -0.12~0.59 -0.31kO.74 2.61 0.02 0.61 22.0 IO-5JA -18.93X1.35 -1.2 1.W.65 6.47f1.25 16.78+1.41 7.73 0.06 0.24 22.0 IO-54 [email protected] -1.2 3.29zkO.31 2.23*0.36 3.82zt0.58 5.51 0.09 0.19 220 IO-70A -16.41k0.13 +I.3 -0.7eO.33 -3.86*0.4? -27.3750.75 7.53 0.11 0.26 22.0 IO-20A -18.31+0.23 -0.6 -l.l7+0.43 -2.94k0.68 -46.94k0.91 8.45 0.39 0.33 22.0

lo-708 -16.15+0.13 lo-7OC -16.55i0.15 IO-52 -17.211tO.12 Mad Hib-18.57M.12 Kaersut -19.01X0.20 I o-43 -17.63kO.20 IO-61 -16.99zkO.16 10-31 -18.49k0.25 Kaersut -19.61~0.12 Kaersut -19.32+0.12 IO-20B -17.01M.35 10”55B -18.47kO.19

MadHib~l8.7~.~2 7-51 -17.4840.25 7-48 -17.1410.22 7.29 -17.48kO.20 7-19 -20.37M.28 7-143 -18.71+0.16 MadHib-18.4&0.12 7-412 -17.14kOo.16 7-664 -19.2tXO.21 7-658 -18.56ti.15 7-644 -18.3X0.16

+2.4 +2.0 +1.4

+0.9 +I.6 +O.l

+1.6 10.1

+I.0 +1.4 +l.O -1.9 -0.2

+I.4 -0.7 -0.1 +o.t

-1.231t0.33 ~1.32rtO.34 0.99M.35

-0.02ti.32 -0.02iO.46 I .48ti.4 1 0.94fo.40 - 1.2010.48 -0.44fl.42 0.16kO.52

-1.OOkO49 0.18tO.52

-0.1*0.50 -1.17~0.41 -1.35kO.36 0.86kO.53

-0.29kO.68 -0.48M.45 -0.07fl.33 -1.w.43 - 1.40M.47 -2.2O&k52 -0.14?0.55

-6.48+0.60 -5.41ti.60 -0.5lM.53 -o.l3*0.54 -0.22N.88 - 1 AOkO.57 -3.463t0.61 - 1 .OtiO.76 -0.09~0.57 0.4OkO.62

-3.38rto.90 4.33kO.80

-30.5310.75 9.03 0.12 0.19 21.5 -29.76fl.75 8.37 0.11 0.16 21.5 -1.2850.77 7.04 0.32 l.il 21.5 0.071tO.78 2.70 0.02 0.63 21.5

-0.6Ukl.22 2.76 0.18 0.06 21.5 -20.87kO.79 2.81 3.73 1.19 21.5 -23.55k0.84 4.47 3.74 2.21 21.5

0.59+1.14 2.17 1.99 7.15 21.5 0.36kO.66 2.78 0.19 0.01 21.5 0.61M.86 2.75 0.18 0.06 21.5

-46.80+1.05 12.77 0.61 2.17 21.5 15.23kf.38 7.53 0.10 0.28 21.5

-0.3310.82 O.ZW.85 2.07 0.02 0.54 22.1 -3.2W.85 -25.14?1.02 6.81 0.10 0.30 22.1 -3.26H.97 -25.29kl.51 6.73 0.13 0.31 22.2 -0.Olfo.94 -10.9;%1.12 8.74 0.15 0.51 22.2 1.04f1.01 2.86k1.05 2.41 0.69 2.00 22.3 O.llkO.82 3.15kOo.76 2.44 1.66 4.52 22.3 0.08~00.57 0.6MO.78 2.05 0.02 0.42 22.5

-1.89kO.70 1.6pNl.88 7.51 0.98 9.00 22.9 -2.87s.85 -1.4W1.11 1.75 2.33 5.45 23.0 -5.49ko.64 1.9tXO.79 6.31 0.48 2.64 23.1 0.32kO.86 0.95zk1.01 6.48 0.31 1.46 23.1

7-753 -17.82f0.21 +0.7 -1.OOkO.50 1.76tO.85 11.28+1.22 7.86 0.31 3.93 23.2 7-789 -18.18?00.12 +0.3 -1.13+0.34 -2.11+0.56 -1.431t0.87 3.42 2.90 1.69 23.2 7-981 -17.09kO.10 +1.4 -0.24W.43 -3.70k0.68 -16.Slf0.71 8.64 0.17 0.92 23.3 MadHib-18.21+0.12 -0.38kO.38 0.39kO.60 0.18kO.87 1.90 0.03 0.50 23.3 MadHib-18.7WO.11 -0.15M.48 0.17+0.69 0.48kO.86 2.04 0.02 0.39 23.7 7-980 -17.01kOo.20 +1.5 -0.68MS5 -4.04~1.11 -16.Slk1.24 9.48 0.15 3.92 23.8 7-953 -17.9m.13 +0.6 -1.3lM.54 -2.03+0X7 -0.74k1.18 2.26 0.69 2.58 23.8 7- 170 -19.251tO.13 -0.7 -1.54M.37 2.34M.60 -49.9010.76 4.10 2.75 5.19 23.9

f FT, = A@Ti,,, - 4% ~~~~ = intrinsic Ti isotopic fractionation of Murchison hibonite.

$ ‘@kZa, %‘, “Cr are isobaric corrections to respective Ti isobars. P h = counting system deadtime. Terrestrial Standark Mad Hib = Madagascar hibonite; Kaenut = kaersutite. Other symbols deftned in Table 1.

ever, after the 46Ti-48Ti normalisation, the analyses of grain lo-70 show a spread of over 70 in both b49Ti and S’@Ti and the variations are correlated. The constancy of 647Ti attests to the suitability of the dead time correction since variation in the dead time would produce correlated variations of similar magnitudes in all normahsed ratios. The isotopic heterogeneity was disguised in the 46Ti-48Ti normaked data because the variation in 49Ti was assigned to mass fractionation. But even in this hibonite, the heterogeneity is smalI compared with the magnitude of the anomalies and is close to the limits of the precision of the data.

The different morphoIogic.al types of the Murchison hibonites show only a minimal correlation with Ti isotopic systematics (Fig 4). The PLACs are rather evenly distributed over the range in #‘?‘i from -47 to + 166. Two of the BAGS have similar compositions character&d by depletions in 5”Ti of approximately -20460, whereas the Ti isotopic composition of lo-31 is close to terrestrial. The SHIBs generally show small effeeta with aqi within 15k of terrestrial, but 7- 170 is highly anomalous with a b’@Ii of -50960. The unclassified grain 7-753 has a S5@Ti of + 11%.

5. IMPLICATIONS FOR THE EARLY SOLAR SY!5TEM

Ti isotopic anomalies

Previous studies have ~~ that meteoritic h&on&es commonly contain isotopicahy anomaIous Ti with variations in s@Ti/c8Ti which range from -7% to +10% relative to terrestrial Ti (HUTCHEON et al., 1983; FAHEY et al., 1985a, 1987a; IRELAND et al., 1985; HINTON et al., 1987). The anomalous compositions have been interpreted a3 signatures of distinct nucIeosyn~~c components that contxibutexI to the solar Ti inventory. The nucIeosynthetie environments which might produce the various Ti isotopic components have been discuss&d in detail by NIEMEYER and LUGMA~R ( 1984) and NIEDERER et al. (1985). FAHEY et al. (1987a) found that a plane could not be sati&ctorily &ted through the Ti isotopic data from h&mites and other meteoritic am- pies, and concluded that at least four distkt Ti-isotopic components were required. Since four components are required and there are only three normahsed ratios, the measured compositions cannot be uniquely deconvoIved into


specific isotopic components (NIEMEYER and LUGMAIR,

1984). The high variability of qi relative to the Other iSOtOPeS

has focused attention on processes which might have a high yield of Vi. However, only two hibonites in this study, lo- 55 and 7-753, have ar@Ti greater than +5%0, and none has #@Ii greater than +20% In fact hibonites with large positive anomalies are rather rare with only MY-H3/MY-H4 having b’@Ti greater than +209& (FAHEY et al., 1987a). The ‘@Ti excesses measured From lo-55,7-753, and MY-H3/MY-H4 are accompanied by ‘Ti excesses. This correlation in the abundances of the neutron-rich Ti nuclides can be produced in nucleosynthetic processes which involve conditions ap proaching nuclear statistical equilibrium. The multiple-zone- mixing model of HARTMANN et al. (1985) has been partic- ularly su& in its predictions of enrichments in the heavy Ti isotopes and other n-rich nuclides of the Fe-group elements such as ‘*Ca, “Cr, and wNi. Enrichments in all of these nu- elides have now been measured in refractory inclusions from Allende (NIEDERER and PAPANASTASSIOU, 1984; JUNGCK ef al., 1984; NIEMEYER and LUGMAIR, 1984; BIRCK and AL LEGRE, 1984; NIEDERER et af., 1985; PAPANASTASSIOU, 1986; BIRCK et al., 1987).

Hibonites with large negative b)ori (less than -209bo) are more common. In Table 5, this includes seven hibonites which, except for 7-170, also have deficits at ‘vi. Nucleo- synthetic processes which produce compositions depleted in qi and 49Ti, i.e. enriched in 4&ri relative to ‘% and %, include quasiequilibrium silicon burning (B~DANSKY ef al., 1968). explosive silicon burning ( WOOSLEY ef al., 1973) and explosive carbon burning on seed nuclei (HOWARD et al.. 1972). WOOSLEY and WEAVER (1982) modelled nucleosynthesis, including quasi-equilibrium and explosive burning stages, in a 25-solar-mass Type-II supernova with trace elements in solar proportions. If Ti produced in such a supernova is mixed with isotopically normal Ti then compositions which are very similar to the qi and 49Ti deficient compositions measured from hibonites can be produced (FAHEY, 1988). Therefore, nucleosynthesis in supernovae can produce Ti isotopic compositions that are depleted, or enriched, in the heavy isotopes of Ti depending on the physicochemical conditions.

The hibonite compositions do not show a simple correlation between b5% and 6’% (Fig. 5). The PLACs with b5”Ti less than -20%~ have 65@Ti/649Ti around 8: 1, whereas PLAC lo-55 has a relatively high 649Ti of +5%0 for the Svi of + 16%, and PLAC 7-658 has b’% of -5% and a 2% excess of vi. The most unusual composition is that of SHIB 7- 170 which lies some 271~ off the best-fit plane of FAHEY et

al. (1987a). The Ti isotopic composition of 7-l 70 is unusual in that despite the large negative 6-i of -50960, the a”% is positive. In all other cases in Table 5 where #‘?i is less than -20960, 649Ti is negative. It is unclear whether the anomalous composition of 7-170 is the signature of a distinct nucleosynthetic process which overproduces “?i and underproduces Vi, or if it is a mixture of a number of components.

For all hibonites with 6% less than -20460, 6”Ti is aIs0 negative and may be a feature of that nucleosynthetic component character&d by the depletions in the heavy Ti iso_

tom. Anomalies in 6”Ti for the other hibonites are largely

-lOI -60 40 -20

6 5oii (?W

20

FIG. 5. The Ti isotopic compositions of the hibonites appear to he dominated by the mixing of two components which are enriched and depleted in the heavy isotopes of Ti relative to terrestrial Ti. Hibonite 7-170 is unusual in that despite the large negative 6%. the s*‘% is positive. However, the hibonite data are not colinear, nor can they be satisfactorily fitted to a plane, and so a minimum of four components is required to model the variations. [Error bars are 2a.l

decoupled from a”%. For the hibonites with I @@Ii I s 5960, 6”Ti ranges from -2.2960 in 7-658 to +3.3% in 10-54. The range in d“?i for these hibonites is from -5.5% in 7-658 to +2.2%0 in 10-54. There is also the hint of a correlation between the anomalies in a”Ti and 64% which suggests that a component with sympathetic variations in “Ti and ‘vi could be present. Ti enriched in “Ti and ‘vi is produced in explosive carbon burning on seed nuclei HOWARD ef al. (1972). However, the positive identification of a distinct component such as this is difficult owing to the small variations in i?Ti and 6”Ti, and the possible presence of anomalies in these isotopes associated with neutron-rich or neutron- deficient material.

“Ai in the early solar system

There is substantial evidence for live 26A1 in the early solar system with an initial 2aAl/27Al of about 5 X 10m5 recorded in a large number of refractory inclusions. While 26Mg excesses in single mineral phases can be ascribed to primary Mg isotopic anomalies or fossil 26Mg from prior 26Al decay (e.g. CLAYTON, 1986), the correlation of excess 26Mg with *‘Al/*‘Mg in a number of coexisting phases from a single inclusion is best explained by the in situ decay of 26Al (LEE

et al., 1977). The (26A1/27Al)o for individual PLACs ranges from 7.7

X 1 0m5 in 7644 to leas than 5 X lo-’ in 7-29, with the majority of grains having (26A1/27Al)o less than 1 X lo-‘. The excess 26Mg from different grains does not correlate with 27A1/24Mg and therefore the apparent initial (26Al/27Al)o is heterogeneous within the PLAC population.

The low (26A1/27Al)o in hibonites such as the PLACs could be interpreted as re8ecting formation after the Type B Allende inclusions (e.g. HINTON and BISCHOFF, 1984). However, FAHEY et al. (1987a) argued that the large Ti isotopic anomalies preserved in these grains were not compatible with late formation and they preferred a heterogeneous distribution of 26A1 in the solar nebula. IRELAND and COMP~TON (1987) found that distillation may have been an important process

282h ‘I’. R. Ireland

in a small subset of hibonites which have very high “‘Al/ *‘Mg. Distillation would remove Mg, including the radiogenic ‘“Mg, from the hibonite which woufd reduce the (‘hAI/‘7Al)C). However, it is unclear if distillation has affected the Mg isotopic systematics of the PLACs. In particular, isotopically normal Mg would have to be returned to the hibonites following distillation in the correct proportions to restore stoi- chiometry.

The Murchi~n hibonite*~a~ng inclusion BB-5 (BAR- MATTHEWS d a!., 1982) probabty formed by the ame process as the PLAC hibonites. It has a large “‘Ti deficit of -70% (HISTOX et al., 1987), it has a Group III rare-earth-element pattern (EKAMBARAM et al.. 1984) which is typical of the PLACs (IRELAND, 1987b), and it has Iow excess 26Mg despite high 27Al/2dMg (BAR-MATTHEWS Q at., 1982). BB-5 also contains corundum with a markedfy higher 27Al/24Mg than the hibonite, however the Mg isotopic compositions of both phases are consistent with a uniform @‘Mg of around +7%. The inferred (26A1/27Al)o of the hibonite is 8 X IO-“, but the maximum value for in siru 2nAl decay could only be 1.5 X 1 O-’ given the high 2’Al/2QMg of the corundum. The excess 26Mg in this inclusion cannot be easily ascribed to in sirzc ‘“Al decay. BAR-MATTHEWS d al. (1982) preferred spatial heterogeneity of the Mg isotopic composition in the early solar nebula to account for the Mg isotopic systematics of BB-5. The large Ti and Ca isotopic anomalies of BB-5 and the other PLACs suggest that the PLACs are some of the most primitive refractory inclusions yet found. The lack of effects expected from the in sdu decay of 26AI suggests that they formed in an area of the early solar nebula which had Mg isotopic heterogeneities and was largely devoid of ‘“AI.

The BAG hibonites contain mass fractionated Mg and variable 6”Mg (IRELAND et al., t 986a). All three BAGS have positive intrinsic mass fractionation which is consistent with these grains being residues from a distillation event. The ‘“Mg excess in 10-3 1 could be attributed to 2hMg addition from in situ ‘hAl decay with (2hA1/27Alk, of approximately 4 X 10 ‘. However, two of the hibonites have slightly negative b2*Mg which indicates the presence of nucleosynthetic anomalies in the Mg isotopic com~sit~on. Therefore, there can be no unique solution as to the possible Mg isotopic components in these grains.

Extreme Mg mass-fractionation has been reported previously by MACIXWGALL and PHINNEY (1979) and H~JTC-HEON

et al. (1983) from the Murchison hibonite-bearing inclusion MH-8. The Mg isotopic mass-fractionation in hibonite was observed to vary from approximately +25 to +345%/amu during analysis, and spine1 within the hibonite also contained mass fractionated Mg. However, this inclusion does not appear to be related to the BAGS. MH-8 is a large (200 pm) inclusion composed of hibonite laths enclosing a small amount of spinel, and has lower MgO and TiOz concentrations than BAGS. The heterogeneity in the Mg isotopic mass- fractionation of MH-8 is a unique characteristic of this inclusion.

The Mg isotopic systematics of the SHIBs are consistent with the in situ decay of **Al with (2hA1/‘7Al)o of ca. 5 X 1 O-‘. which is typica of (Z6Al/‘7AIk in spinel-bearing refractory inclusions (HUTCHEOK ef a!., 1986).

Wibonite 7-753 has a clearly resolved deficit in ‘hMg. The

Presence of “‘Mg depletions has been used to suggest that the solar System Mg isotopic composition was originally depleted in 26Mg, and the ‘<‘Mg abundance was subsequently aug- mented by the decay of “‘AI (ESAT r~ ai., 1980). CI.AYTC)N ( 1986) has calculated that common interstellar Mg has a “‘Mg

deficit of 1.9%. and the preservation of this signature might be expected in at least some grains, However, it is unclear at this stage whether ‘“Mg depletions in the hibonites are a signature of an initial deptetion of the entire solar system. or if the anomalies are due to Mg isotopic heterogeneities in the hibonite precursors.

The isotopic and chemical characteristics of hibonites can provide important constraints on the processing of material in the solar nebula. Hibonite is a refractory phase. and as such high temperatures are required to form the present morphologies. The preservation of isotopic anomalies in these grains indicates they have escaped the degree of homogeni- sation that character&s most soIar materials. Isotopic anomalies not associated with radioactive decay have some- times been used as evidence of presolar materials. While it is almost certain that the Ti isotopic anomalies are the signatures of different nucleosynthetic sites, there is no evidence that the hibonite grains themselves are interstellar. FA~IEY et al. (1987a) have previously argued for a solar system origin for isotopically anomalous hibonites.

Two main types of model have been proposed for the nature of the thermal event which produced meteoritic hibonites: condensation of grains from a cooling gas. or melting/ distillation of dust aggregates.

GROSSMAN (1972) used equilibrium thermodynamics to predict the sequence of phases that would condense from a cooling gas of solar composition. The sequence bea= a re- markable resemblance to the phase assemblages of refractory inclusions from the Allende carbonaceous chondrite. The rather porous textures associated with some hibonite-bearing grains have also been used as compelling evidence for condensation (MACPHERSON c’t al., 1983). The physical param- eters used in the condensation model were derived from a solar system evolution model proposed by CAMERON i 19623. In this model interstellar dust and gas collapstxl in a refatively short period of time to produce a protosun. The mid-plane pressure of the solar nebula at 3 AU was - 10‘ ’ atm. and temperatures in excess of 1800 K were attained. Under these conditions the material falling into the solar nebula would have been vaporised.

However, LARSON { t 969) showed that material was gradually added to the evolving solar nebula and the time scale for disk accretion was much longer than the collapse time envisaged by CAMERON ( 1962). Recent models for the solar nebula predict midplane pressures of only lo”“? to 1O-5 atm. at 3 AU and temperatures no greater than a few hundred K {for a review see Wooa and MORFILL., 1988). Therefore, rc- fractory grains such as the hibonites, which preserve the signatures of high temperature events, were not melted by the nebula gas but probably obtained the heat required for their formation from local transient events (MORFNJ., 1983).

KORSKKI and FEGLEI’ ( 1984) used thermodynamic calculations to predict a theoretical phase diagram which in-

Hibonitc from Murchison 2831

eluded the Ca-aluminates for the first time. A stability field for calcium dialuminate was found bctwcea the condensation temperatures for hibonite and melilite. Both calcium dialuminate and melilite are generally absent from Murchison refractory grains and KORNACKI and PEGLEY (1984) favoured an origin by melting of dust aggregates. However, GEIGER et al. (1988) have used new experimental data to show that calcium dialuminate may not be stable under the conditions expected for the solar nebula, and MACPHERSON and GROSSMAN ( 1984) have suggested that the absence of melilite from the meteoritic assemblages is due to metastable condensation of spine1 prior to melilite. Solid solution effects and partial pressures of other components may also have an affect on the stability fields of various phases (MACPI~ERSON and GROSSMAN, 1984). Therefore, the exact predictions for the phase assemblage are not clear and the presence or absence of a phase cannot be used as evidence either for or against the condensation model at this time.

LUGMAIR, 1984). Each grain could have a unique isotopic composition because the precursor dust aggregates carried different proportions of dust from various nucleosynthetic sites. A bulk average of the dust gives the solar isotopic abundances, but both positive and negative anomalies can be Pro- duced by varying the proportions of dust from different sources.

The best candidate for a grain formed directly from gas to solid condensation in this study is 7-953 (Fig. IG), which is very similar to the grain SH-6 studied by MACPHERSON et al. (1984). Grain 7-953 is a fluffy aggregate of hibonite blades with perovskite inclusions and spine1 pseudomorphs alter hibonite. However, while the spine1 pseudomorphs are consistent with the reaction of hibonite with an Mg-rich gas, it is not clear that the hibonite-perovskite assemblage formed directly as a condensate. In particular, it would be expected that perovskite would condense at a lower temperature onto the hibonite, yet the perovskite often occurs as anhedral inclusions within the hibonite (Fig. 1G).

The PLACs, BAGs, and SHIBs are all probably derived from melts of refractory dust aggregates. The three generations of hibonite, each with distinct chemical and isotopic char- a~eristics, may represent discrete cycles in a transient heat source. The rather clear-cut distinction for the three hibonite groups analysed here indicates that once formed, the grains have been largely isolated from later events. The presence of mass fractionated Mg in the BAGS indicates that these grains may have been distilled, but the extent of the heating was not sufficient to cause significant Ti isotopic mass-fractionation and the composition of the BAG precursors must have been refractory as well.

6. CONCLUSIONS

The absence of isotopic anomalies in meteoritic samples was used as evidence that the solar system was originally hot and homogeneous (BEGEMANN, 1980). The preservation of large isotopic anomalies in individual hibonite grains presents a problem to a condensation origin. The gas associated with each grain would have had to remain isolated in the face of turbulence and admixture of more material and the gas must then renucleate into relatively large i~topi~y-horn~n~~ grains. The preservation of diverse Ti isotopic compositions within hibonite groups which have similar chemical and Mg isotopic characteristics does not appear to be consistent with condensation of these grains from a gas.

Twenty-four hibonite gmins from the Murchison carbonaceous chondrite have been divided into three distinctive groups. The first group is composed of fourteen Colourless PLAty Crystal fragments (PLACs) that have less than 2.7% TiOr . Three Blue AGgregates (BAG@ composed of hibonite crystal plates and fragments have 5.1-6.546 TiOr . Spinel-HI- Bonite grains (SHIBs), generally with abundant perovskite inclusions, show a range in morphologies from compact spherules to porous crystal aggregates and are commonly rimmed by spinel, an Fe-silicate, and clinopyroxene. Ti02 concentrations in hibonite from SHIBs range from 3.5% to 7.696, but the TiOr concentration in hibonite from one grain was heterogeneous. The hibonite compositions in all hibonite types are stoichiometric with a predominantly coupled substitution of Mg* and Ti’+ for 2A13+. One grain with a rounded hibonite core does not clearly fit into the three designated hibonite types.

An alternative model postulates that the present hibonites formed from melts (e.g. KORNACKI and FEGLEY, 1984, and references therein). The precursors to the hibonites were probably either aggregates of intemtellar dust grains or possibly grains which had experienced some degree of processing within the solar system. Distillation of chondritic dust aggregates could also form a refractory phase assemblage by removing the volatile components to leave a refractory residue. NOTSU er al, (1978), HASHIMOTO ef al. (1979), and IRELAND and &AT ( 1986) have all produced refractory phase aggregates by distillation. However, distillation From near chondritic wmblages should involve significant isotopic

mass fractionation. The general absence of fractionation effects, except for Mg in the BAGS, suggests that the dust assemblages were initially refractory in composition.

Magnesium isotopic analyses also reflect the morphological types. The PLACs have high *‘AV2’Mg but usually show only small excesses in 26Mg corresponding to (26A1/27A1)o of less than I X 10W5. However, two PLACs have (26~/2’Alb of 5.5 X lop5 and 7.7 X lo-’ which suggests that, if present, 26A1 was heterogeneously distributed in the formation region of the PLACs. Alternatively, the excess 26Mg might reflect Mg isotopic heterogeneities. The BAGs have positive mass fractionation and variable 826Mg values. Contributions to the Mg isotopic composition of the BAGS cannot be uniquely assigned and they may preserve components from nucleosynthetic anomalies, distillation, or in situ 26Ai decay. The SHIBs preserve 26Mg excesses which are correlated with 27~1/ “Mg. The Mg isotop’ 1~ systematics of these grains are consistent with in situ 26A1 decay with an initial 2bA1f27Al of co. 5 X 10-j. The unclassified hibonite has a clearly resolved negative @‘Mg which indicates that primary Mg isotopic anomalies are occasionally preserved in hibonites.

Melting ofdust aggregates is also compatible with the preservation of isotopic anomalies by the cosmic chemical mem-

The Ti isotopic compositions have only a minimal cor-

ory model (CLAYTON, 1977, 1978, 1981; NIEMEYER and relation with morphological type. The PLACs appear to be the most anomalous with eight out of thirteen having 1 65@Ii 1

2838 T. R. Ireland

greater than 5% and @“Ti covering a range from -47 to + 16%. Two of the BAGS have b”‘Ti of -20%. while the other BAG has only small Ti isotopic anomalies. Five of the six SHlBs have 1 bs’% 1 within -tS?& of terrestrial, but SHIB 7-l 70 has a 6”Ti of -50% indicating that extreme compositions are preserved in the SHIBs as well. For hibonites with large anomalies in b’“Ti, the 6““Ti is generally of the same sign. This is consistent with models of nucleosynthesis within supernovae. However, the composition of the SHIB 7-170 is

highly unusual in that despite 6% of -5O?&, it has a positive a4’Ti. It is unclear whether this composition represents a distinct nucleosynthetic component, or ifit is a mixture ofcom-

ponents. The three types of hibonite probably represent distinct for-

mation episodes within the solar system and each may be associated with individual transient heating events. The cor- relations in morphology, mineralogy, hibonite chemistry, and Mg isotopic systematics suggests that these features are pri- marily a signature of the formation environment. Ti isotopic compositions are not directly correlated with the hibonite types because each grain reflects variations in the proportions of interstellar dust from different nucleosynthetic sites. The preservation of diverse Ti isotopic compositions within each group is difficult to reconcile with condensation of these grains from a gas. The hibonite-bearing grains probably formed from the melting of refractory dust aggregates composed predom-

inantly of Al. Ca, Ti, and Mg oxides.

Acknowledgements-This paper is the result of a Ph.D. study at the Australian National University under the supervision of Prof. W. Compston whose guidance in matters isotopic is greatly appreciated. Instrumental support on SHRIMP by 1. S. Williams, J. J. Foster, and N. Scramm was indispensible. I would also like to thank Albert Fahey for introducing me to the Cameca IMS-3f. This paper has benefitted greatly from discussions with A. J. Fahey. E. K. Zinner, and R. M. Walker. Reviews by I. Hutcheon. G. Lugmair, and an anonymous reviewer resulted in significant improvements to the manuscript. This work was partially supported through NASA grant #NAG9-55.

Eduorial handlrnX- H. Y. McSween. Jr.

REFERENCES

ALLEK J. M., GROSSMAN L., DAVIS A. M. and HUTCHEON 1. D. (I 978) Mineralogy, textures and mode of formation of a hibonite- bearing Allende inclusion. Pram. Lunar Planer. Sci. Con/: 9rh. l209- 1233.

BAR-MATTHEWS, M., HUTCHEON 1. D., MACPHERSON G. J. and GRWMAN L. (1982) A corundum-rich inclusion in the Murchison carbonaceous chondrite. Geochim. Cosmochim. Acla 46, 3 i-4 I.

BECKETT J. R., LIVE D., RN-Dow T., GROSMAN L. and STOLPER E. ( 1988) Ti” in meteoritic and synthetic hibonite. Geochim. Cas- mochim. Acta 52, 1479-1495.

BEGEMANN F. (1980) Isotopic anomalies in meteorites. Repf. Prog. Phvs. 43. 1309-1356.

BIRCH J. L. and ALL~GRE C. J. (1984) Chromium isotopic anomalies in Allende teftactoty inclusions. Geophys. Res. Letr. 11.943-946.

BIRCK J. L.. PROMBO C. A. and LUGMAIR G. W. (1987) Ni and Cr isotopes in Allende inclusions. Lunar Planer. Sci. XVliI. 79-80.

BODANSKY D.. CLAYTON D. D. and FOWLER W. A. (1968) Nucleo- synthesis during silicon burning. Phys. Rev. Left 20, 161-164.

CAMERON A. G. W. (1962) The formation of the sun and planets. Icarus 1, 13-69.

CA~ANZARO E. J., ML‘RPHY T. J., GARNER E. L. and SHIELDS W. R. (1966) Absolute isotopic abundance ratios and atomic weights of magnesium. J. Res. Var. Eur. Sfand. 709,453-458.

CLAYTON D. D. (1977) Solar system isotopic anomalies: Supernova neighbour or presolar carriers? Icarus 32.255-269.

Ct A\ 1‘0% D. D. (1978) Precondensed matter: Key to the early solar system. iMoon and Plunels 19, 109- 137.

CI.AY~OZ D. D. (198 I ) Some key issues in isotopic anomalies: As- trophysical history and aggregation. Proc. Lunar fluner. .%i I.?rh. 1781-1802.

CLAY I ON D. D. (I 986) Interstellar fossil “‘Mg and its possible relationship to excess meteoritic 26Mg. Ap. J. 310.490-498.

ESAT T. M.. PAPANAST~SSIOU D. A. and WASSERBURC, G. J. (1980) The initial state of 26AI and 26Mg12JMg in the early solar system. Lunar Planet. Sri. XI. 262-263.

EKAMBARAM V., KAWAB~ 1.. TAUAKA T., DAVIS A. M. and GROSS- MAN L. ( 1984) Chemical compositions of refractory inclusions in the Murchison C2 chondrite. Geochim. Cosmochim. Acra 48,2089- 2105.

FAH~Y A. J. (1988) Ion microprobe measurements of Mg. Ca, TI. and Fe isotopic ratios and trace element abundances in hibonite- beanng inclusions from primitive meteorites. Ph.D. thesis, Wash- ington University.

FAHEY A. J.. GOS~AMI J. N., MCKEEGAN, K. D. and Z~NNER E. (1985a) Evidence for extreme “‘Ti enrichments in primitive meteorites. Ap. J. 296, Ll7-LZO.

FAHE~ A. J.. GOSWAMI J. N.. MCKEEGAX K. D. and ZINNER E. (1985b) Ion probe and track studies of CM hibonites: ‘*Al. 2uPu and REE. Lunar Planer. .Sct. Xb’L 227-228.

FAHE~’ A. J.. GOSWAMI J. N., MCKJXGAN K. D. and ZINNER E. (1987a) ‘“Al, ‘“Pu. ‘(‘Ti. REE and trace element abundances in hibonite grains from CM and CV meteorites. Geachim. Casmo- chim. Acta 51, 329-350.

FAHT-y A. J., GOSWAMI J. N., MCKEEGAN K. D. and ZIK~‘;ER E. ( 1987b) I60 excesses in Murchison and Murray hibonites: A case against a late supernova injection origin of isotopic anomalies in 0, Mg, Ca. and Ti. .4p. J. (Lnrs.) 323, L91-L95.

FLEX-H G. D., CAPEI.LEI‘; J. and SVEC H. J. (1966) The abundance of the vanadium isotopes from sources of geochemical interest. Advan .Masr Spec 111, 57 I-574.

GEIGER C. A., KLEPPA 0. J.. MYSE~’ B. 0.. LAT~?MER J. M. and GROSSMAN L. (1988) Enthalpies of formation of CaAI.0, and Ca.AI,20,9 (hibonite) by high temperature, alkali borate solution calorimetry. Geochim. Casmachim. .4na 52, I729- 1736.

GROSSMAX L. (1972) Condensation in the primitive solar nebula. Geachim. Cosmochim. Acra 36, 597-619.

HARTMASN D.. WOOSLEY S. E. and EL EID M. F. (1985) Nucleo- synthesis in neutron-rich supernova ejecta. Asfraphys. J. 297. 837- 845.

HASHIMOTO A., KUMAZAWA M. and ONUMA N. (1979) Evaporation metamorphism of primitive dust material and in the early solar nebula. Earth Plnnn. Sri. Left. 43, 13-2 I.

HEYDECX;ER H. R., FOSTER J. J. and COMBTON W. (1979) Evidence of a new isotopic anomaly from titanium isotopic ratios in meteoric materials. h’arure 278, 704-707.

HtN.ros R. W. and Btsctiort A. (1984) Ion microprobe magnesium isotope analysis of plagioclase and hibonite from ordinary chondrites. Nu!ure 308. I69- 172.

HINTON R. W., DAVIS A. M. and SCATENA-WACHEL D. E. (1985) lon microprobe determination of REE and other trace elements in meteoritic hibonite. Lunar Plane!. Sci. XVI. 352-353.

HINTON R. W.. DAVIS A. M. and SCATENA-WACHEL D. E. (1987) Large negative “‘Ti anomalies in refractory inclusions from the Murchison carbonaceous chondrite-evidence for incomplete mixing of neutron-rich supernova ejecta into the solar system. Ap. J. 313,420-428.

HOWARD W. M., ARNET~ W. D., CLAYTON D. D. and WRIGLEY S. E. ( 1972) Nucieosynthesis of rare nuclei from seed nuclei in explosive carbon burning. Ap. J. 175, 201-206.

Ht:T(‘tiEoN 1. D.. BAR-MATTHEWS M.. TANAKA T., MACPHERS~N G. J.. GROSSMAN L. and KAWABE 1. (1980) A Mg isotope study of hibonite-bearing Murchison inclusions. Me~eornics IS, 306- 307.

HUTCHEON I. D.. STEELE I. D.. WACHEL D. E. S., MACDOUGA~ L J. D. and PHIKXEY D. (1983) Extreme Mg fractionation and ev-


idence of Ti isotopic anomalies in Murchison mfmctory inclusions. Lunar Planet. Sci. XIV, 339-340.

HUTCHEON I. D., ARMSTRONG J. T. and WASSERBURG G. J. (1986) Mg isotopic studies of CA1 in C3V chondrites. Lunar Planet. Sci. XVII, 312-373.

IRELAND T. R. (1987a) Correlated morphological, chemical, and isotopic systematics from Murchison (C2M) hibonites. Lunar Planet. Sri. XVIII, 45 1-452.

IRELAND T. R. (1987b) Correlated morphological, chemical, and isotopic systematics from Murchison (C2M) hibonites. Meteoritics 22,416-417.

IRELAND T. R. and COMPSTON W. (1987) Large heterogeneous 26Mg excesses in a hibonite from the Murchison meteorite. Nature 327, 689-692.

IRELAND T. R. and BAT T. M. (I 986) Laboratory synthesis ofcalcium dialuminate and hibonite. Lunar Planet. Sci. XVII, 378-379.

IRELAND T. R., COMP~TON W. and HEYDEGGER H. R. (1985) Ti- tanium isotopic anomalies in hibonites from the Murchison carbonaceous chondrite. Geochim. Cosmochim. Acta 49,1989- 1993.

IRELAND T. R., COMPSTON W. and ESAT T. M. (1986a) Magnesium isotopic compositions of olivine, spine1 and hibonite from the Murchison carbonaceous chondrite. Geochim. Cosmochim. Acta 50, 1413-1421.

IRELAND T. R., ESAT T. M. and COMP~TON W. (1986b) Magnesium isotopic systematics of Murchison refractory inclusions. Lunar Planet. Sci. XVII, 380-38 I.

IRELAND T. R., FAHEY A. J. and ZINNER E. K. (I 988) Trace element abundances in hibonites from the Murchison carbonaceous chondrite: Constraints on high-temperatum processes in the solar nebula. Geochim. Cosmochim. Acta 52.2841-2854 (this issue).

JUNCCK M. H. A., SHIMAMURA T. and LUGMAIR G. W. (1984) Ca isotope variations in Allende. Geochim. Cosmochim. Acta 48, 265 l-2658.

KORNACKI A. S. and FEGLEY B. JR. (1984) Origin of spinel-rich chondrules and inclusions in carbonaceous and ordinary chondrites. J. Geophys. Res. Suppl. 89, B588-B596.

LARSON R. B. ( 1969) Numerical calculations of the dynamics of a collapsing prom-star. Mon. Not. Roy. Astr. Sot. 145, 271-295.

LEE T., PAPANASTASSIOU D. A. and WASSERBURG G. J. (1977) 2bAl in the early solar system: Fossil or fuel? Astrophys. J. (Letts.) 211. L107-LI IO.

MACD~UGALL J. D. (1979) Refractoryclement-rich inclusions in CM meteorites. Earth Planet. Sci. Lett. 42, l-6.

MACDOUGALL J. D. (1981) Refractory spherules in the Murchison meteorite: Are they chondrules? Geophys. Res. Len. 8, 966-969.

MACDOUGALL J. D. and PHINNEY D. (1979) Magnesium isotopic variations in hibonite from the Murchison meteorite: An ion microprobe study. Geophys. Res. Lett. 6, 2 15-2 18.

MACPHERSON G. J. and GROSSMAN L. (1984) “Flufy” Type A Ca-

Al-rich inclusions in the Allende meteorite. Geazhim. Cosmochim. Acta 4&29-46.

MACPHERSON G. J., BAR-MATTHEWS M., TANAKA T., OLSEN E. and GROSSMAN L. (1980) Refractory inclusions in Murchison: Recovery and mineralogical description. Lunar Planet. Sri. XI. 660-662.

MACPHERSON G. J., BAR-MATTHEWS, M., TANAKA T., OLSEN E. and GROSSMAN L. (1983) Refractory inclusions in the Murchison meteorite. Geochim. Cosmochim. Acta 47, 823-840.

MACPHERXIN G. J., GROSSMAN L., HASHIMOTO A., BAR-MAT- THEWS M. and TANAKA T. (I 984) Petrographic studies of refractory inclusions from the Murchison meteorite. J. Geophys. Res. 89, C299-C3 12.

MCLEOD L. C., RUDOWSKI R. and ULLMAN W. J. (1985) Pure gyp- sum separates from “dirty” evaporites. J. Sediment. Petrol. 55, 596-597.

MORFILL G. E. (1983) Some cosmochemical consequences of a tur- bulent protoplanetaty cloud. Icarus 53.4 l-54.

NIEDERER F. R. and PAPANASTASSIOU D. A. (1984) Ca isotopes in refractors inclusions. Geochim. Cosmochim. Acta 48.1279- 1293.

NIEDERER F. R., PAPANASTASSIOU D. A. and WASSE~BURG G. J. (198 I) The isotopic composition of titanium in the Allende and Leoville meteorites. Geochim. Cosmochim. Acta 45, 10 I7- IO3 I.

NIEDERER F. R., PAPANASTASSIOU D. A. and WA~SERBURG G. J. ( 1985) Absolute isotopic abundances of Ti in meteorites. Geochim. Cosmochim. Acta 49,835-85 I.

NIEMEYER S. and LUGMAIR G. W. (1984) Titanium isotopic anomalies in meteorites. Geochim. Cosmochim. Acta 48, 140 I - 14 16.

NOTSU K., ONUMA J. and NISHIDA N. (1978) High temperature heating of the Allende meteorite. Geochim. Cosmochim. Acta 42, 903-907.

PAPANASTASSIOU D. A. (1986) Cr isotopic anomalies in the Allende meteorite. Ap. J. (Lens.) 308, L27-L30.

SHIELDS W. R. (1966) NBS Technical Note 277. U.S. Govt. Printing Office. Washington, DC.

WOOD J. A. and MORFILL G. E. (1988) A review of solar nebula models. In Meteorites and the Early Solar System (eds. J. F. KER- RICGE and M. S. MATTHEWS). Univ. of Arizona Press (in press).

WOOSLEY S. E. and WEAVER T. A. (1982) Nucleosynthesis in two 25 solar-mass stars of different population. In Essays in N&ear Astrophysics (eds. C. A. BARNES, D. D. CLAYTON and D. N. SCHRAMM), pp. 377-399. Cambridge Univ. Press.

WOOSLEY S. E., ARNETT W. D. and CLAYTON D. D. (1973) The explosive burning of oxygen and silicon. Ap. J. Suppl. 26, 231- 312.

ZINNER E. K., FAHEY A. J.. GOSWAMI J. N., IRELAND T. R. and MCKEEGAN K. D. (1986) Large ‘%a anomolies are associated with Ti anomalies in Murchison and Murray hibonites. Ap. J. (Lerfs.) 311, L103-L107.

correlated morphological, chemical, and isotopic...

Documents