Dualistic distribution coefficients of elements in the system mineral-hydrothermal solution. II. Gold in magnetite

Abstract

The system magnetite-Au-hydrothermal solution was employed to continue studying the distribution coefficients of trace elements in system with real crystals. The role of surface nonautonomous phase (NP) is elucidated. The distribution coefficient of an Au structural admixture between magnetite and hydrothermal solution at the experimental conditions [450°C, 1 kbar (100 MPa), and fluid sampling by a trap] is, according to the most representative data, 1.0 ± 0.3, and Au is thus not an incompatible element in magnetite, in contrast to pyrite and arsenopyrite [1], minerals for which this coefficient is much lower than one. The NP is enriched in Au with respect to the rest of the crystal by a factor of more than 4000, and this results in an one order of magnitude increase in the bulk distribution coefficient. Similar to pyrite, the reason for the dualistic nature of the distribution coefficient is the presence of an NP, which contains ∼2000 ± 500 ppm Au. The NP occupies the approximately 330-nm surface layer of the crystal, and the chemically bound Au [Au(III), according to XPS data] admixture is evenly distributed with depth within the layer, which is the reason for the strongly determinate dependences of the concentrations of the evenly distributed Au admixture on the size and specific surface area of the crystal. The occurrence of an NP is controlled by the chemistry of the system. The partial substitution of Fe for Mn and the synthesis of a phase close to jacobsite MnFe2O4 results in the disappearance of both the NP itself and the size dependence of the Au concentration. The XPS spectra of O 1s and Fe 2p are used to analyze two models: (i) a single goethite-like (O2−/OH−∼ 1) phase of variable composition and Fe in more than one valence state and (ii) a heterogeneous structure of alternating domains of wuestite- and goethite-like NP. The reason for the “excess” admixture in the former instance can be vacancies at Fe sites, whereas that in the latter one is the interaction of the admixture with nanometer-in-size nanometer in-size strained domains on the surface of the crystal.