Measurement of Ti[superscript +3] / Ti[superscript +4] ratios in pyroxene in Wark-Loveing rims: Evidence for formation in a reducing solar nebula

Ti-bearing clinopyroxene, known as fassaite, is a major phase in the interiors of coarse-grained, Ca-, Al-rich refractory inclusions (CAIs). Electron microprobe (EMP) analyses of such pyroxene yield low cation sums when normalized to six oxygen anions if it is assumed that all Ti is present as Ti{sup 4+}. Instead, we can assume that there is one Ca cation [1] and two tetrahedral cations present per six oxygens, and can then calculate a Ti{sup 3+}/(Ti{sup 3+} + Ti{sup 4+}), or Ti{sup 3+}/Ti{sup tot}, ratio that gives exactly one cation in the remaining site and a total of exactly four cations per six oxygens. Additional evidence for the presence of Ti in multiple valence states includes: pleochroism in Ti-rich crystals with negligible amounts of other multivalent elements; results from measurements of optical spectra [2]; and X-ray absorption near-edge structure (XANES) analysis [3]. Calculation of accurate Ti{sup 3+}/Ti{sup tot} ratios from EMP analyses is generally limited to analyses with >4 wt% TiO{sub 2}{sup tot} (all Ti as TiO{sub 2}), because at low Ti contents the analytical uncertainties approach the magnitude of the cation deficit caused by assuming all Ti is Ti{sup 4+}. Many refractory inclusions are enclosed in sequences of mineralogically distinct layers, more » first described by [4], that must have formed after the host inclusions did. In most cases, from the CAI outward, the sequence consists of a layer of spinel {+-} perovskite; voids, melilite, or alteration products; clinopyroxene; and hedenbergite. The pyroxene layer may be immediately adjacent to spinel and is commonly zoned from Ti-rich fassaite nearest the spinel layer to Ti-poor aluminous diopside over distances of {approx}10 {micro}m [4]. A recent study of the pyroxene in the rims of one Allende and two Leoville inclusions [5] found that most points analyzed by electron probe had between 4 and 7 wt% TiO{sub 2}{sup tot} and 0.7-1.7 wt% FeO. Those authors also found < 1 Ca cation, but very close to 4 total cations, per 6 oxygen ions, and therefore little or no Ti{sup 3+}. They kindly loaned us one of their samples for analysis, and we also found low Ca contents, high FeO contents ({approx}1 wt%) and undetectable to low Ti{sup 3+} contents compared to fassaite found in the interiors of CAIs. The low-Ca analyses would seem to reflect the presence of an enstatite component in addition to the standard fassaite components [1], but despite inclusion of such a component in our calculations we have not been able to satisfactorily resolve the analyses into pyroxene endmembers. This is a hint that pyroxene analysis spots are contaminated with another phase, a serious problem for stoichiometrydependent calculation of Ti{sup 3+}/Ti{sup tot} ratios. We need to know whether or not rim pyroxene contains Ti{sup 3+} because of its importance as a recorder of nebular oxygen fugacity, but we are unsure of how to normalize EMP analyses in this case, so we cannot calculate Ti{sup 3+}/Ti{sup tot} accurately. We therefore decided to measure the Ti{sup 3+}/Ti{sup tot} of rim pyroxene directly, by XANES. We have conducted a detailed study of pyroxene in the rim on TS24, a fluffy Type A inclusion from Allende. This inclusion is large and very irregularly shaped, giving it a relatively high surface area. It has a well-developed rim sequence, from the CAI outward, of spinel, clinopyroxene zoned from Ti-, Al-rich to nearly pure diopside over {approx}20 {micro}m, and hedenbergite. The spinel-pyroxene and diopside-hedenbergite contacts are sharp. « less