Titanium isotopic anomalies in hibonites from the Murchison carbonaceous chondrite

Titanium isotopic anomalies in hibonites from the Murchison carbonaceous chondrite

Geachzmrca ef Cosmwhrmrca Acm Vol. 49, pp. 1989-1993 0 Pngsmon Press Ltd. 198S. Rinted in U.S.A. LETTER Titanium isotopic anomalies in hibonites fro...

602KB Sizes 0 Downloads 68 Views

Geachzmrca ef Cosmwhrmrca Acm Vol. 49, pp. 1989-1993 0 Pngsmon Press Ltd. 198S. Rinted in U.S.A.

LETTER

Titanium isotopic anomalies in hibonites from the Murchison carbonaceous chondrite T.R. IRELAND, W. COMPSTON Research School of Earth Sciences, The Austraiian National University, Canberra ACT 260 I. Australia

and H. R. HEYDEGGER* Department of Chemistry and Physics, Purdue University Calumet, Hammond, IN 46323 U.S.A. and Enrico Fermi Institute, University of Chicago, Chicago. IL 60637 U.S.A. (Received d4u.v27, 1985: uccepred in rev~sed.f~r~ Augwr 8,

1985)

Abstract-The isotopic compositions of titanium in eight grains of hibonite fCaAl,20,9) from the carbonaceous chondrite Murchison have been determined by high precision secondary ion mass spectrometry using an ion microprobe. The titanium in the hibonites varies greatly in Ti. from about -42 to +8 permil (relative to terrestrial) with smaller (up to 4 pennil). but clearly resolvable, effects in ‘6Ti and 48Ti. These results complement ion probe measurements by FAHEYet al. (1985) of a 100 permil excess of “ri in a hibonite grain from the carbonaceous chondrite Murray, and confirm the presence of widespread negative anomalies suggestedby the results of HUTCHEONet af. (1983) on hibonites from Murchi~n. The magnitude of these variations seems explicable only in terms of nucleogenic pmcesses which produced extremely variable titanium isotopic abundances in the hibonite source materials. The hibonites evidently did not participate to the same extent as most material in the mixing and homogenisation processes that accompanied the formation and later evolution of the solar system. Thus, significant source materials of the hibonites may be the supernova condensates of CLAYTON(1978) and may support the concept of “chemical memory” (CLAYTON, 1978; NIEMEYERand LUGMAIR,1984).

THE ISOTOPIC composition of titanium in refractory covered using ion microprobes on a small number of inclusions from meteorites is an important indicator hibonite grains from C2M chondrites. of nucleosynthetic processes. Titanium is on the lower HUTCHEON et al. ( 1983) first analysed four hibonite mass side of the iron abundance peak and therefore inclusions from Murchison using an ion microprobe its composition can reflect variations in the quasiset at low mass resolution. Two of the hibonites apequiIib~um conditions in stellar interiors giving rise peared to have a normal titanium isotopic com~sition to that peak, provided nuclear processing during in- but the other two were character&d by large ‘@lYdefjection into the interstellar medium is minimal icits of -8 f 2 permil (DJ-6). and -16 + 3 permil (WOOSLEYand WEAVER, 1982). Titanium is important (DJ-5). The precision of the data was not sufficient to also in having one of the most refractory oxides and resolve anomalies in the other isotopic ratios. FAWEY is concentrated in refractory inclusions by either conet af. (1985) have recently reported the titanium isoden~tion or evaporation processes (GROSSMAN,1972). topic com~sition of three hi~nit~ at high mass resA large number of titanium isotopic analyses has olution to separate all significant isobaric interferences, including 4sCa, from 48Ti. Two Murchison hibonites been carried out on individual calcium aluminiumrich inclusions (CAI) from Allende and other carbowere character&d by excesses of - +8 permil in connaceous chondrites by thermal ionisation mass spec- trast to the deficits recorded by HUTCHEON et al. (1983). A hibonite inclusion from Murray has a retrometry. The great majority of these show a 1 permil enhancement of qi relative to terrestrial ratios markable signature in having a 100 permit excess in (NIEDERER ef al., 1980, 1981, 1985; NIEMEYERand % with effects at the other two normal&d ratios LUGMAIR, 198 I, 1984; HEYDEGGER et al.. 1982). clearly resolved. In this letter we document the presence of wideAnalyses of the so called FUN (Fractionation and Unknown Nuclear) inclusions show more diversity, but spread titanium isotopic heterogeneities in Murchison even here the anomalies are limited to a range of PIi hibonites, and independently confirm the large deficits from -5to f4 pertnil (NIEDERER~~al., I981.19851. in ‘@fi indicated by the first ion micropro~ measureLargeranomalies, up to 100 permil, have been dis- ments of titanium in Murchison hibonites by HUTCHEON et al. (1983). The great virtue of ion microprobe analysis is that * N.R.C. Resident Associate, NASA Johnson Space Center, SN3, Houston, TX 77058, U.S.A.

isotopic compositions can be determined in situ on a few nanograms of sample. In this study titanium iso-

1989

I990

i K Ireland. W iompston

topic analyses have been carried out on a JO-75 pm size fraction of hibonite grains separated from the Murchison (GM) chondrite. (~‘onventional titanium isotopic analysis of such small individual grains would be extremely difficult Hibonite is an ultrarefractory mineral composed ofcalcium hexaluminate (CaAl,,O,o) with minor but significant substitution of titanium and magnesium. The Murchison hibonites were extracted by crushing 67 g of small meteorite fragments to pass a 400 pm mesh, with further sieving to collect the 4075 pm fraction. The denser fraction (p > 3.3 g/cm’) was separated in methylene iodide and the non-magnetic fraction was collected after magnetic separation in an alcohol suspension. The resulting concentrate was handpicked for blue inclusions. which were mounted in epoxy, and polished in prep aration for analysis. Of seventy blue grains separated and mounted. eight were identified by electron microprobe as dommantly hibonite and these were selected for isotopic analysis. The eight hibonite grains are morphologically divided mto three inclusions and five crystal fragments and range from 1.7 to 6.4% TiOz. Most of the remainder were spine1 with <;15 pm inclusions of hibonite and perovskite. The titanium isotopic analyses were carried out on the ion microprobe. SHRIMP. designed and constructed at the Australian National University (CLEMENTer al.. 1977). A negative oxygen primary beam of l-3 nA focused to a spot 25 pm in diameter sputters a sufficiently intense beam of Ti’ ions for precise titanium analysis at a mass resolution of 7OClO. A fraction of the secondary ion beam was monitored using a Car) 401 electrometer at a suppressed aperture before the source slit. and the counts on each isotope were divided by the charge collected simultaneously on this monitor. This procedure ap-

and H. K. Heydegger proximates to a double collecting mode ol operation ano IIsulted in increased precision. by removing the cllects ol‘ an: low frequency Instability in the primary beam The iso~opc~. of interest were measured sequentially by cyclic magnetic liclti stepping. each peak being electrostatically centered hy dellec tion plates located in front of the collector slit Flat-topped peaks. in which the top IS equal to or wider tharr one-third the base width. were maintained for each isotope h! automanr mechanical displacement of the collector assemhi\ along the beam path. This compensated for small mass-dependent zhrti% in the focal pomt of the magnetic analyser High precision titanium isotopic analyses b> ion micropior*. require an intense secondary ion beam for good counting ststistics and. stmultaneously. sufficient mass resolution to scl,arate isobaric interferences. At pi mass resolution of 7000 21 A.%{(I% valley), all molecular interferences. rnriudinp iii drides. are resolved from the titanium isotopes. (ini! 3tomii isobaric interferences from &C‘a.%V. and “Cr are not resolved these may be estimated by monuoring %t. “a’ :md “f! and then stripped from the respective titamum iroban 1 h resolution is sufficient to fully separate ‘*Ca from “‘Ti %hen the latter is centered in the collector slit: during 4 mass scan 48Ca does not enter the collector slit until “‘Ti just begins tt’ leave it. In addition. because the “Ca intensity IS ver) iov. compared with ‘8Ti in the hibomtes, there is no observable contribution (
TABLE I. Iitanium isotopic analyses of Murchison hibonite grams and terrestnal standards. -_--. ..~ (I)’ Sample 20 1 2 31 I 43 I 52 I

305

(I? 3.

-----~.-~~-

1312 rTdhTi

14)’ ,?‘TI

(5)’ f?‘OTi

(h? i’a ,.---.-_.-

329 326 324 320

-17.1 - 13.2 - 15.4 - 15.3 -14.0

0.26 -c 0.82 0.80 t I.22 / 26 c 0.94 -3.36 +_0.86 -I 75 rt 0.90

‘.4Y rt 0.50 ; 44 ‘r 1.14 I .08 + 0.62 0.03 t 0.48 0.24 2 0.36

-42.46 f 0.50 4 I .78 2 0.48 I .‘1! + 0.62 17.21 t 0.76 ---0.0I f 0.44

RR0 : I iz 6_ ,

54 I 303 55 I 302 2 330

-17.5 -- 14.6 -14.0

-3.67 + 0.54 2.34 + 0.80 I.24 1 0.98

2.45 ? 0.28 3.61 _’ 0.50 7.34 i- 0.84

I .92 f 0.50 7.89 -+ 0.46 7.77 + 1.08

61 I 325 70 I 304 2 318 3 319

-15.3 -13.5 - 13.9 -14.1

3.33 fr 0.96 -~0.46 f 0.64 - 1.41 2 0.82 -0.55 $: 0.84

1 __ ‘9 *2.63 i 3.66 ii 23 i-

ks 1 2 3 hi I I 3 ru I il I

-18.3 --17.3 -17.6 -14.9 -15.1 - 15.9 -21.6 -19.5

0.23 0.20 -0.17 -0.35 - 0.67 -0.52 0.13 --0 I3

300 .39’ _328 301 317 321

277 295

0.x4 0.36 0.46 0.80

18.28 & 0.72 2 I .48 t 0.44 71.34 i-O.&l 22.40 & 0.56

5.t> - ‘i _I ._ 4.0 16 8.i ‘X

2 0.8: + 1.10 + 0.8X 2 0.G -+ 0.94 + 0.94 I 0.26

~0.04 _t 0.56 0.0’ t 0.96 -0.36 + 0.48 0.3 I :’ 0.50 0.54 I 0.40 0.05 t 0.50 0.03 ?z 0.08

+ 0.13

-0.3’ c 0.22

0.09 f 0.50 ~-0.45 F 0.50 0.20 + 0.52 0.20 + 0.52 0.31 ? 0.46 0.71 i 0.54 0.16 t 0.16 0.24 & 0.33

2.x 2-h’ 2.’ .’ 6 I. Z.’ 2x ii..? ri.(i

\

iti i

’ Sample identification: grain no., spot number. consecutive data tile number. Terrestrial standards: ks -: kaersurue (Glen Innes. N.S.W.), hi = hibonite (Antani Mora, Madagascar), ru = rutile (beach sand. Rutile and Zircon Mines (Newcastle) Ltd.). il = ilmenite (St Urbain. Quebec). Rutile. ilmenite collected on Faraday cup. Files 300-306 deadtimc = 22 ’ ns. Files 3 17-330 deadtime = 2 I .4 ns. ’ A = - 1[(*‘Ti/?‘i).,J I .336] - I ; x lOOO/:! which ISthe fractionation per atomic mass umt relative to ai lii4” I r i .33X) (HEYDEGGERel a/.. 1979) heavy enrichment = positive. Typical analytical error 0.3 permil (2cr,). ’ Delta notation deviations from terrestrial standards of HFYDECXERL*Iuf (1982). 6’Ti = [(‘Ti!4’ri),,,/(‘ri/4”l ii,,,,,, Ij Y 1000. Errors for Murchison hibonites are 20, of 7 sets of IO ratios. ’ Ca and V + Cr are corrections in permil applied to 46Ti and “Ti respectively due to unresolved isobaric interferences from calcium. vanadium and chromium

1991

Ti isotopic anomalies high count rates and for ratios that differ significantly from unity (HAYESand SCHOELLER,1977). The titanium isotopes have a rather fo~u~tous abundance pattern in this regard, with only 48Ti being signi~cantly more abundant than the other masses. During the Murchison hibonite, terrestrial hibonite. and kaersutite analyses, the count rate was kept close to 0. I MHz for 49Ti to minim&e biassing due to error in the estimation of deadtime. However even at this count rate, a 0.1 ns error in dete~ining the deadtime would have a 0.12 permil effect on the ?i,@% ratio. Deadtime errors have no detectable effect on the other ratios owing to the similar count rates at these peaks. Deadtime was determined initially by measuring 48Ti/4vi over a wide range of beam intensities. and later. by setting the (normalised) 48Ti/49Tito the standard value. The data were corrected for deadtime. ratioed to the secondary beam monitor, ratioed to mass 49. stripped to remove the isobaric interferences, and normal&d to I .336 for 4’Ti/ ‘?i to correct for mass fractionation. Mass fractionation always accompanies the sputtering process with a systematic enhancement of the light isotopes (SHIMIZLIand HART. 1982). For titanium, the effects are large and are also matrix dependent. ranging from 0.8%/amu for Ti metal to 2.8%/amu for pcrovskite. In addition, instrumental parameters, such as the secondary ion extraction. may also contribute. but the fractionation for Ti isotopes using SHRIMP has been found to be constant to within %2 permil for a given phase in a given mount. The fmctionation is adequately described by a linear mass fractionation law allowing no~ali~tion to an arbitrary datum point, in this case to 1.336 for 4’Ti/+‘9Ti(HEYDEGCXR et al.. 1979). The actual normalisation law used will have little effect owing to the limited range of fractionation recorded in this study. The fractionation corrected ratios of the unknowns (unk) are then expressed as deviations from the standard (stdf terrestrial ratios in delta notation as 6’Ti = { [(‘Ti/4%),,,/(‘Ti/49Ti),]

- I ) X 1000

where I’may be 46,48, or 50 and the standard terrestrial values from HEYDEGGERet al. (1982). To verify the ability of SHRIMP to make titanium analyses of high precision and accuracy we cite results for a pyroxene

646Ti[4

separate from the Allende inclusion 3529. 32 which had previously been determined by thermal ionisation mass spectrometry as having a qi excess of 1.O ? 0.2 permil (HEYDECGERet al., 1982). The ion microprobe analyses of individual pyroxene fragments were carried out with 49Ti at 0.2 MHz: 4sTi was not collected during this experiment to avoid possible overloading of the ion counter. The mean ion microprobe result for three grains was found to be 1.3 -t O-3(20,) permil excess for %. and a*Ti within error of terrestrial. A more complete description of these data will be given elsewhere. The analyses reported here are weighted means of 7 sets of IO scans of the peak sequence. Each set takes approximately 25 minutes. resulting in a total collection time of -3 hours per spot analysis. The uncertainties quoted in Table 1as two standard errors of the mean are propagated from the dispersion observed amongst the seven set means for each spot. and are, m general. very close to those expected from counting statistics. The hibonite titanium analyses were carried out during two collection periods separated by a week during which SHRIMP was used for other applications. The first four hibonites were analysed in a single 24 hour period along with terrestrial kaersutite and hibonite standards. The second period was for 3 days when the remaining hibonites plus three of the previous four grains were analysed, again with interspersed kaersutite and hibonite standards. The deadtime determined for the first set of analyses was 22.2 ns and for the second set 21.4 ns. However because all the isotopes have similar abundances except 48Ti. only the 48Ti/4% ratio is sensitive to deadtime variations of this magnitude. However, during both periods, interspersed standard runs agree within error with the terrestrial norms and replicate analyses of samples 20, 55. and 70 show excellent agreement (Table I), despite the differing deadtimes used. The are the grains viduals

most striking features of the hibonite analyses diversity of the compositions measured between and the contrasting homogeneity within indias analysed in different spots (Fig. 1). As with

all other reported titanium

I I

isotopic anomalies.

n

FIG. I. isotopic data of hibonites presented in Table i plotted as S*Ti (upper) and S48Ti{lower) vs. S’@Ti; note difference in vertical and horizontal scales. Italicised numbers refer to Murchison hibonite grains listed in Table I. Boxes are two sigma uncertainties of the analyses. Replicate analyses of grains 20, 55, and 70 show excellent agreement suggesting intragrain homogeneity as opposed to the diversity of compositions between individual crystals.

the

I. K. Ireland. W. Compston and H. R. Heydegger

1992

dominant effects are in “Ti. in this case ranging from -42 to +8 permil. The range of ds% anomalies reported by HUTCHEON ef al. ( 1983). FAHEY PI~1. ( 1985 ). and HINTON EI ~1. (1985) from Murchison and Murra!, hibonites is represented in the analyses reported here. and the increased precision of the SHRIMP results and the expanded data base have allowed clear resolution of widespread anomalies in 46Ti (HEYDEGGER e/ al 1982) and 4*Ti. Despite the variety of isotopic compositions, the mass fractionation is ciose to that observed for terrestrial hibonite. Our experience with the consistency of sputtering fractionation suggests that an\; intrinsic isotopic fractionation should be less than 3 permil/amu. This observation corresponds with the low values of fractionation (< 1 permil/amu) reported from Allende inclusions (NIEDERER et al. 1985). intriguing. The deficits in 50Ti are especially HUTCHEON et al. ( 1983) described 50Ti deficits in two hibonite grains measured by ion microprobe. Recent measurements reported by HINTON et al. (1985) also using the Chicago machine included ‘% deficits up to 60 permil. Thermal ionisation analysis of an “anhydrous” separate from Murchison led NIEMEYERand LUGMAIR (1984) to report a “hint” of negative ‘vi anomalies. However. of all of the thermal ionisation results. only FUN inclusion Cl has yielded an unambiguous deficit (NIEDERER et al., 1985). Enhancements of 5% have been much more common in the analyses reported (HEYDEGGER et al.. 1979. 1982; NIEDERER et al., 1980, 198 I. 1985: NIEMEYER and LUGMAIR, 198 I, 1984). The addition

of a “Tirich component to a terrestrial-type composition was suggested from the first indication of the + I permit anomaly in the Allende inclusions (HEYDEGGER CIu/.. 1979). The observation of FAHEY ef u/. ( 1985 1recently of a 100 permil enhancement of “Ti in the hibonite grain from the Murray carbonaceous chondrite is consistent with such a model. However. the large negative anomalies are more consistent with models involving more complex mixtures of isotopically different titanium (NIEDERER el ~1.. 1981. 1985: NIEMEYER and

topes shown to exist amongst hlbomtes tram a slngir meteorite. we may infer the presence of still greater isotopic heterogeneity in their source materials. The latter may well be the micron-sized grams ofrefractor! oxides. etc. that condensed from gases qlected from one or more stellar objects predating the formatIon ui planetary bodies in the solar system (CL.&\-TOY, I978 :. The great variability in titamum isotopic zompositlon amongst these grains would reflect spatial and/or tern poral heterogeneities in the composition ofmtersteltar dust from a number of diverse nucleosynthctlc en\!. ronments in one or more stellar objects. The accrerior; of the primary objects into the hundred micron-sired objects we see today took place with a lesser degree oi homogenisation (“chemical memory”) than. apparently. characterises most solar system materials, 17~ with sufficient heating at some stage to cause the degre:. of intragrain titanium isotopic uniformit! observed. The diversity of compositions in Murchison hibon ites is in marked contrast to the titanium in 4llendc refractory inclusions. Nine out of ten Allerxie C’AIs have the + 1 permil 5@lYsignature (HE~IXXXER 1’1U: 1982; NIEMEYER and LUGMAIR. 1984) which requires a relatively homogeneous source for these objects. Thus. the reported (NIEMEYER and L%M*IR. 1383. NIEDERER VI a/., 1985) variability in titanium Isotopic composition in bulk meteorites could be due to varration in the small proportion of highly anomalous hibonite rather than large variations in the proportiori of slightly anomalous CA1 pyroxene. ,4ch-noM,/~d~~,menls-I. S. Williams and J. J. f-ester partly.

pated in the exploratory use of SHRIMP for precise titanium isotopic analyses (COMPSTON n al.. 198 I ) and together with N. Schramm and L. All&on provided essential instrumental support. This manuscript has benefitted greatly from criticai examination by 1. S. Williams. M. 7. Esat. and A. E. Ringwood and from thorough reviews by R. Walker and two anonymou+ referees. One of us (HRH) wishes to acknowledge supper! from the US-Australia Co-operative Science Program through NSF grant 83 I 1890. from the Australian National I !nivenit). and from Purdue University. Calumet.

LUGMAIR. 1981. 1984: HEYDEGGER CI ui.. 1982).

The minimum number of components necessary to account for the observed isotopic variations has increased as the precision of the measurements and the size of the data base has increased. NIEDERER et N/ ( 1980)

found that three components were required. as

did HEYDECGER ef al. (1982).

while NIEMEYER and

LUGMAIR (1984). using essentially the same data base. argued for the requirement

of four components. The hibonite measurements using the ion probe have greatly expanded the volume in “three-isotope ratio” space in which data reside, but have not yet resolved the issue because the number of components required is sensitive both to the precision of the experimental data (which is significantly degraded under renormalisation) and to the normalisation basis adopted. Even though the hibonite titanium is dominated by the terrestrial isotopic signature, from the large variations in ‘@I3abundance relative to other titanium iso-

REFERENCES CLAYTOND. D. (I 978) Precondensedmatter:keq ~1 lhe cari) solar system. Moon and f’/unrr.~ 19, 109- 137 CLEMENT S. W. J.. COMPSTON W. and NEWS IlAD ci. c IKy i Design of a large. high resolutron Ion microprobe. Prcj, In/ S.I.IV.S Con/: (Muenster). COMPSTON W., FOSTERJ. J.. WILLIAMSI. Y. iurxs J h and HEYDEGCERH. R. (198I ) Ti isotopic composition using the ion microprobe. /Inn. Rep. Rex .%h Kurrh %r I%‘/, 2 12-2 17. Australian National Universlt! FAHEYA. J.. GOSWAMIJ. N.. MCKEEGANK. D. and ZINNER E. (1985) lon probe measurements reveal large titanium isotopic effects in CM hibonites (abstr.). f.unur ~%lanc+‘Q: .YI’I, 229-230. GROSSMANL. (1972) Condensatton in the prcmmv? soi nebula. Geochim. Cosmochim. Acta 36, 597-h ! 9 HAYES J. M. and SCHOELLER D. A. f 1977) Hugh precision counting: limitations and optimal conditions. Ana/ C‘htw 49, 306-3 1I. HEYDECGER H. R.. FOSTER J. J and C’OMPS’IC)~U t iV7Q>

Ti isotopic anomalies Evidence of a new isotopic anomaly from titanium isotopic ratios in meteoric materials. Nature 278, 704-707. HEYDECGERH. R., FOSTERJ. J. and COMPSTONW. (1982) Terrestrial, meteoritic, and lunar titanium isotopic ratios reevaluated: evidence for correlated variations. Earth Plan~f. Sci. Lelt. 58, 406-4 18. HINTON R. W., DAVIS A. M. and SCATENA-WACHEL D. E. (1985) A large negative titanium-50 anomaly in a refractory inclusion from the Murchison meteorite. Mefeoritics (in press). HUTCHEON1. D., STEELE1. M., WACHEL D. E. S., MACDOUGALLJ. D. and PHINNEYD. ( 1983) Extreme Mg fractionation and evidence of titanium isotopic variations in Murchison refractory inclusions (abstr.). Lunar Planet. Sci. XIV. 339-340. NIEDERERF. R., PAPANASTASSIOU D. A. and WASSERBURG G. J. (1980) Endemic isotopic anomalies in titanium. Astrophvs. J. Lett. 240, 123-128. NIEDERERF. R., PAPANASTASSIOU D. A. and WASSERBURG

1993

G. J. (198 I) The isotopic composition of titanium in the Allende and Leoville meteorites. Geochim. Cosmochim. Ado 45, 1017-1031.

NIEDERERF. R.. PAPANASTASSIOU D. A. and WASSERBURG G. J. ( 1985). Absolute isotopic abundances ot‘ Ti in meteorites. Geochim. Cosmochim. .4cta 49, 835-85 I. NIEMEYERS. and LUGMAIRG. W. (1981) Ubiquitous isotopic anomalies in Ti from normal Allende inclusions. Eurfh Planet. Sci. Lett. 53, 21 l-225.

NIEMEYERS. and LUGMAIRG. W. (1984) Titanium isotopic abundances in meteorites. Geochim. Cosmochim. Acla 48, 1401-1416.

SH~MIZUN. and HART S. R. (1982) Applications of the ion microprobe to geochemistry and cosmochemistry. .4nn. RPY Earth Planel. Sci. 10,483-526. WOOSLEYS. E. and WEAVERT. A. (1982) Nucleosynthesis in two 25M stars of different population. In Essa.vs m Nuclear Asrrophvsics (ed. C. A. BARNESer al.), pp. 377-399. Cambridge Univ. Press.