High Cu concentrations in vapor-type fluid inclusions: An artifact?

Abstract

Studies of natural fluid inclusions have shown that Cu commonly occurs at higher concentrations in vapor-type inclusions than in coexisting brine inclusions, a phenomenon which has been interpreted to arise from copper partitioning into the vapor phase. In the first part of this study, we attempted to experimentally reproduce this behavior in hydrothermal experiments covering a range of P–T conditions (600–800°C, 70–130MPa), fluid compositions (NaCl±KCl±FeCl2; 0.5–6.6wt.% S), fluid acidities (quench fluid pH ⩽0.3–10), and sulfur speciation (H2S-dominated to SO2-dominated). However, as in several other studies we did not succeed in reproducing conditions under which Cu partitions into the vapor phase. In view of recent observations that quartz-hosted fluid inclusions can diffusively loose or gain Cu after entrapment, we set out to determine if the evidence from natural fluid inclusions could be compromised. For this purpose we synthesized vapor and brine inclusions from Cu–H2O–NaCl–S fluids at 800°C/130MPa and re-equilibrated them in slightly different fluids at 800°C/70MPa, measuring some inclusions by LA-ICP-MS after each step. Vapor inclusions indeed experienced a dramatic increase in Cu from 0.3±0.03 to 5.7±3.3wt.%, while brine inclusions remained largely unmodified, leading to a change in the partition coefficient DCuvap/brine from a true value (i.e., before re-equilibration) of 0.4±0.05 to an apparent value of 8.3±4.9. The requirements for substantial diffusional gain of Cu in fluid inclusions are a change in the pH of the surrounding fluid from ⩽1 to more neutral and the presence of S in the pre-existing fluid inclusions. These requirements are also fulfilled in nature: cooling magmatic-hydrothermal fluids experience a change from acidic to more neutral pH due to buffering along the feldspar-mica join, and natural vapor inclusions typically contain significant amounts of sulfur. A reversal experiment performed on natural, quartz-hosted fluid inclusions from the Erongo granite, Namibia, showed that this process can be reversed with the measured DCuvap/brine value of 11±9.3 being modified to 0.06±0.04. Thus, DCuvap/brine values >1 measured on natural fluid inclusions in quartz are likely a secondary feature caused by post-entrapment copper diffusion. Realistic DCuvap/brine values in porphyry Cu environments are between 0.11 and 0.15, and reconstructed vapor/brine mass ratios are in the order of 4–9. This suggests that the main transporting agent of Cu at the porphyry level are brines and that models based on copper into the vapor phase are incorrect.