Published in Meteoritics, 29, p. 504.
ALH 84001, a ferroan martian orthopyroxenite, originally consisted of three petrographically defined components: a cumulus assemblage of orthopyroxene + chromite, a trapped melt assemblage of orthopyroxene(?) + chromite + maskelynite + apatite + augite ± pyrite, and a metasomatic assemblage of carbonate ± pyrite . We present here the results of INAA study of five bulk samples of ALH 84001, combined with SIMS data on the orthopyroxene , in order to attempt to set limits on the geochemical characteristics of the latter two components, and therefore on the petrogenesis of ALH 84001.
The INAA data support the petrographic observations, suggesting that there are at least three components in ALH 84001. The most incompatible trace lithophile elements show wide variations in abundances. For example, La varies from 0.27× to 2.7×, Sm from 0.54× to 1.8×, Eu from 0.41× to 1.3×, and Hf from 1.2× to 9.9× CI chondrites. These abundances generally increase with increasing Na, but do not form two-component mixing lines on element-Na plots. The sample with the highest Na content has a LREE-enriched pattern, with La/Yb ratio of 1.3× CI, requiring that one of the non-cumulus components be LREE enriched, as previously suggested . For the following discussion, we will assume that each of the three geochemically required components can be equated with one of the petrographically observed components.
Both trapped melt and metasomatic components in ALH 84001 have higher Na than orthopyroxene based on compositions of maskelynite, apatite, and carbonate . For the metasomatic component, we will assume its Na content is that of carbonate, while for a trapped melt component, we will use a typical Na content inferred for martian meteorite parent melts, ~1 wt% Na2O . Under these assumptions, we can set limits on the LREE/HREE ratios of the components, and use this information to compare the petrogenesis of ALH 84001 with other martian meteorites. For the assumed Na contents, one of the components has higher La content than the other. If the trapped melt contains the higher La content, then, depending on the Na content, this trapped melt has a CI-normalized La/Yb ratio >2.6. The metasomatic component in this case would be HREE enriched with a La/Yb ratio of <0.3. The alternative scenario, in which the metasomatic component contains higher La content, requires that the trapped melt be LREE-depleted (La/Yb < 0.3). The metasomatic component would then have La/Yb >1.8. The SIMS data on orthopyroxene are incompatible with this latter model, as calculated equilibrium melts are not as LREE depleted as calculated here , and could be LREE enriched, depending on the partition coefficients used. Hence, based on our preliminary calculations, it appears more likely that the trapped melt in ALH 84001 was LREE enriched. Mass balance arguments would then suggest that the parent melt of ALH 84001 was also LREE enriched, similar to those of Nakhla and Chassigny . We will attempt to refine these calculations to better understand the parent melt of ALH 84001, and its relationship to the other martian meteorites.
The above calculations assume that the bulk samples are representative of different portions of ALH 84001, i.e., that we have not artificially generated a range in trace-element contents by under- or oversampling a mineral phase such as apatite. We will also evaluate the possible heterogeneous distribution of mineral phases in the bulk samples as the cause of compositional heterogeneity in our samples.
References:  Mittlefehldt (1994) Meteoritics, 29, 214.  Papike et al. (1994) LPS XXV, 1043.  Longhi and Pan (1989) Proc. LPSC 19th, 451.  Longhi (1991) Proc. LPS, Vol. 21, 695.