A fundamental question in the study of magmatic-hydrothermal ore deposits is whether the mineralization potential of intrusions was already predetermined by the metal content of the exsolving fluids. The present study aims at addressing this question by reviewing the large number of microanalytical data (mostly laser-ablation ICP-MS data) obtained on fluid inclusions from this type of ore deposits over the last 20 years. Published data sets were screened for analyses of high-temperature fluid inclusions that are representative of premineralization fluids. A set of criteria was developed to distinguish such fluids from later, lower temperature fluids. In order to compensate differences in absolute metal concentrations caused by fluid immiscibility, all element concentrations were normalized to Na. A numerical model was developed to explore at which stage different metals are most efficiently extracted from a cooling pluton. The results suggest that the timing of most efficient metal extraction varies from metal to metal and strongly depends on pressure, the fluid/melt partition coefficient and the bulk mineral-melt partition coefficient. As a consequence, fluid compositions were chosen over the entire range of Cs/Na ratios recorded from a given pluton, as this ratio gives an indication of the fractionation degree of the silicate melts from which a fluid exsolved. In order to avoid bias toward occurrences from which a large amount of data are available, maximum four intermediate-density (ID)-type fluid inclusion assemblages plus four brines assemblages were chosen from each occurrence.
Using the above-mentioned criteria, 169 fluid compositions from 12 Cu (Mo, Au) mineralized intrusions, 10 Sn/W mineralized intrusions, two Mo mineralized intrusions, and one U-Th-REE mineralized intrusion were finally chosen and plotted in graphs of X/Na versus Cs/Na. The results reveal that Sn- and Cu-mineralizing fluids contained more Sn and Cu, respectively, than the fluids analyzed from barren and Mo or U-Th REE mineralized intrusions. Positive correlations between fluid metal content and mineralization potential may exist also for W and REEs, whereas for Mo no such trend is evident. Therefore, at least for certain metals, the metal content of high-temperature fluid inclusions can be used as an indicator of the type and extent of mineralization. However, elevated metal concentrations are present also in some fluids from barren intrusions, which implies that the mineralization potential additionally depends on other factors such as the size of the intrusion and the development of structures that promote focused fluid flow.