Evaluating ion implants as trace-element reference material for EPMA

To optimize gold ore processing, it is critical to evaluate the levels of “invisible” Au residing in sulphides. The EPMA remains a powerful instrument to quantitatively correlate Au concentrations with grain growth features at optimal spatial resolution. A major obstacle in evaluating the consistency and accuracy of trace Au concentrations obtained by EPMA is the lack of reliable secondary standards. Natural sulphides are poor candidates due to their heterogeneous Au distribution. Moreover, because Au has restricted solubility in most sulphides and the substitution mechanisms are poorly understood, synthesis routes are limited. In this context, ion implants were evaluated as trace-element reference material for EPMA. However, the narrow depth distribution of the implanted ions relative to typical electron ranges confers to these materials unique properties. If treated as homogeneous specimens, the X-ray intensities would yield “apparent” Au concentrations that vary with accelerating voltage. Consequently, the implants need to be treated as layered materials requiring appropriate correction procedures. We investigated cm-sized polished grains of magnetite, pyrite and galena. Implantation dose and energy were 5.0E14 Au atoms/cm2 at 3 MeV for magnetite, and 1.0E14 atoms/cm2 at 1 MeV for pyrite and galena. The Au depth profiles were calculated with the TRIM simulation and measured using a Cameca IMS 3f SIMS. Au Mα and Lα X-ray intensities were collected at accelerating voltages from 10 to 30 kV using a JEOL JXA 8900 EPMA. The k ratios were obtained relative to pure Au. To calculate the predicted “theoretical” k ratios, the continuous Au depth distribution was simulated using double- to quadruple-layered structures on a substrate. The first satisfying observation is the extent of lateral micrometer-scale homogeneity in the Au levels across the cm-sized implants. The ratio of analytical to statistical standard deviations never exceeds 1.3, with theoretical standard errors ranging from 1.5 to 9%. Secondly, the measured and predicted X-ray intensities are in very good agreement. Discrepancies arise at lower accelerating voltages, mainly related to the position of the Au concentration maximum relative to the X-ray production profile. As long as the voltage is high enough to keep the Au concentration maximum at a mass depth close to the φ(ρz) maximum observed in a hypothetical homogeneous bulk specimen, the results are insensitive to small uncertainties on the Au depth profile. At lower voltages, the predicted intensities are much more sensitive as the concentration maximum aligns with the lower part of the φ(ρz) curve where X-ray production decreases sharply. Fortunately, the depth of the concentration maximum can be controlled by varying the implantation energy to produce optimal implants for given accelerating voltage windows. Finally, the Au X-ray intensities obtained for these implants yield “apparent” bulk concentrations ranging from 35 to 320 ppm. X-ray intensities corresponding to lower Au abundances can be tested, by using lower implantation dose. In conclusion, ion implants are practical EPMA reference material, especially when conventional synthesis is difficult. Materials with specific analytical challenges can be engineered due the excellent control of the implantation parameters.