EPMA analyses of olivine and metal grains in Apollo 17 dunite thin section 72415,53. Thin section analyses by optical microscope and electron microprobe were performed at the Birkbeck, University of London. BSE imaging and EDS and WDS compositional data collection were performed on the Jeol JXA8100 Superprobe with an Oxford Instruments Aztec energy-dispersive system (EDS) at Birkbeck.

WDS point analyses were performed over several sessions. The first batch, including points along profiles, were performed with a 50 nA, 15 kv focused beam. Counting times and standards were as follows: Mg, Si, Fe (10 s, St Johns Olivine); Ca (30 s, andradite); Al (180 s, corundum); Mn, Cr, Co (30 s, pure metals); Ni (60 s, pure metal). Backgrounds were measured for half the peak count time. Other than for the olivine standard, calibrations were performed at 20 nA to reduce high count rates. A XPP Φ(ρz) matrix correction was applied. Analyses were filtered to remove low-quality points (e.g. at grain edges at profile margins) and to remove olivine data points that were contaminated with the micro-scale inclusions that these olivines are known to contain, usually identified by anomalous elevated Al, Ca and/or Cr contents. Typically, three or more measurements were made in different core locations on most grains. 

The second batch of measurements were performed similarly but at 80 nA with the following differences in calibration: Ti (60 s, ilmenite); Co, Cr (60 s); Ni (90 s). The change was to improve counting statistics for minor elements. Secondary standards San Carlos Olivine, SCO; St John’s Olivine, SJO; and MongOL Sh11-2 were measured at regular intervals, allowing us to check the measurement accuracy and apply a minor drift correction to the SiO2 measurements as the narrow peak shifted slightly off the measurement position due to temperature changes during this session (other elements were unaffected). Most points were measured in the same location as those of the first session, allowing datasets to be merged. The analytical sessions are labelled in the data. 

Apart from CoO, our measurements of the three olivine secondary standards showed good agreement with published data. Corrections were applied to measured CoO concentrations on the basis of the Fe-kβ overlap and additional consistent offset from published standard concentrations. Because of the corrected nature of the final CoO concentrations, we interpret them with caution and consider them to be approximately correct but not high-precision estimates.

Metals were analysed with the following conditions: 25 nA, 15 kV focused beam current, calibrated with pure metal standards except for Al, Si and P for which we used SiO2, Al2O3 and Durango apatite standards. Measurement times were: Fe, Ni: 20 s; Al, Si, P, Cr, Co: 40 s, with each background measured for half the peak time. A Φ(ρz) matrix correction was applied. A small correction was applied to remove the effect of the Fe-kβ peak from the measured Co-kα peak. Whilst Al is not routinely measured in Fe-Ni metals, we included it because an Al peak was ubiquitous in preliminary EDS spectra. High Al concentrations measured in metal by WDS are a surprising result and should be interpreted with caution.