The development of rare PGE analysis in natural insulating materials by AMS has continued at the IsoTrace Laboratory. The passivation of the insulating rocks to give steady ion currents from a sputter ion source has been accomplished by increasing the amount of the local neutral caesium vapour relative to the Cs⁺ ion current. Stable operating conditions were attained so that, for example, the platinum isotopes could be detected in Rhum anorthosite, in addition to the more easily-detectable gold. Besides the low counting rate the presence of interference from the molecules of the more-abundant elements was a problem with the present system. For example, the ⁶⁵Cu₃^- anion is always present with ¹⁹⁵Pt and this generates three ions that can be in coincidence. A higher resolution ionisation chamber solved this problem as the ions are all well below the Bragg peak. This resulted in, for example, the easy resolution of the ⁶⁵Cu⁺⁺ triple-add-up peak from the weak ¹⁹⁵Pt⁺⁶ peak. The other ions interfering with the detection of the PGE will be discussed. These included ¹⁶⁵Ho⁺⁵, a triply charged molecule and surprisingly ¹³³Cs⁺⁴ from the CsO_n^- ions.

The measurement of the relative abundance of the PGE, for example Pt and Ir, with respect to a standard presents problems in understanding due the unknown details of the sputtering process and the negative ion parentage. Consequently, the structure of the negative ions, the sputtering reaction mechanism and the method of passivation imply that only ion implantation of the rock with known quantities of an isotope of the element of interest is likely to be successful for determining the absolute amounts of PGE. The creation of such an internal standard is therefore necessary.

The general problem of correctly combining AMS with sputter milli- and microprobes for very low-level PGE measurements will also be discussed. It essentially requires the non-dispersive combination of an electric and magnetic analyser with enough resolution of both to separate the interference by the difference between the mass defects. The resolution of each should be at least 2500 for the Pt⁺⁶ isotopes but lower values can be exploited by using lower charge states if a flexible charge state strategy can be adopted.