Absorption correction procedures for quantitative analysis of fluid inclusions using synchrotron radiation X-ray fluorescence
Résumé
The main limiting factor associated with individual fluid inclusion analysis using synchrotron radiation X-ray fluorescence (SRXRF) is the quantification of absorption of the incident and fluorescent radiation by the fluid and the host material. Theoretically, the measured (Kα/Kβ)z intensity ratio from a given element Z in solution is directly proportional to the thickness of material traversed and therefore could be used as a reliable term for the absorption correction. In order to constrain experimentally the relationship between (Kα/Kβ)z, host material thickness and fluid inclusion size and salinity, a three-step protocol has been developed using the LURE (Orsay, France) photon microprobe installed on line D15. These are: (1) interception of the X-ray fluorescent beam emerging from pure-metal targets (Mn, Fe, Ni, Cu, Zn) using thin plates of aluminium and quartz of known thicknesses; (2) feasibility tests on silica-glass capillaries containing known concentrations (∼5000 ppm) of various metals (Mn, Ni, Zn) and salts (0 and 15 wt.% NaCl); (3) confirmation tests on synthetic fluid inclusions containing known metal contents (∼1000 ppm Ni and Zn) hosted in halite crystals. The results of experiment 1 indicate that the evolution of the (Kα/Kβ)z ratio as a function of host material thickness is in perfect agreement with theoretical predictions and is independent of the local analytical environment (reproducibility of the measurements). For experiments 2 and 3, the (Kα/Kβ)z ratio of a reference element in solution (Zn), was used to estimate the thickness of host material traversed. Establishing that the X-ray peak intensity of Zn corresponds to a given concentration allowed the concentration of other trace metals in solution to be determined. In both experiments, the precision is high with standard deviations (1σ) from 1 to 8% of the mean, but the accuracy can be relatively poor with mean values ranging from 3–11% (capillary tests) to as high as 17% (inclusion tests) of the known concentrations. Tight ranges of calculated metal concentrations at values higher than the known concentrations strongly suggests that the Zn concentration used to calibrate the Zn X-ray peaks may be different from that effectively measured using SRXRF. A knowledge of Mn, Ni and Zn chloride and silica speciation during SRXRF measurements is required to evaluate better this potential source of error. Despite this, results are commonly better than 20% (experiment 2) and 32% (experiment 3) relative to the known concentrations. This implies that elemental concentrations in individual fluid inclusions can be quantitatively determined using the SXRF technique without precise knowledge of the inclusion depth and geometry.