No abstract

Chalcopyrite (CuFeS2) is the main source of copper in the world. The development of hydrometallurgical processes to extract copper from chalcopyrite is challenging due to the low leaching kinetics. The main difficulty is in the fact that the kinetics of the leaching process decreases very rapidly, marginally stopping the reaction. A passivation process of the surface has been proposed for explaining the low reaction kinetics. However, the leaching mechanism and the reactants which are involved in the passivation process are still a matter of debate. Therefore, understanding the chalcopyrite surface reactivity and the intricate reaction occurring in the solid/solution interface is of fundamental importance. In the present study, DFT calculations within the plane wave framework were performed to understand the reconstruction of (001), (100), (111), (112), (101), and (110) chalcopyrite surfaces. Metal and sulfur terminated surfaces have been investigated. The structural and electronic properties of the reconstructed surfaces have been discussed in detail. Three different mechanisms of the chalcopyrite surface reconstructions emerged from this study. It is clear that the chalcopyrite surface undergoes important reconstruction in which the sulfide, S2–, ions migrate to the surface which tend to oxidize, forming disulfides, S2 2–, and, concomitantly, reducing the superficial Fe3+ to the Fe2+. It is also observed that the metal atom moves downward to the surface, forming metallic-like bidimensional alloys underneath the surface.

The interaction of water molecule with the reconstructed (001) chalcopyrite surfaces has been investigated by means of density functional calculations. All of the calculations were performed using periodic boundary conditions with SIESTA code. The structural parameters were compared with those obtained through PWscf code in order to evaluate the pseudopotentials and numerical basis set developed for this work. Two different surfaces were studied, namely, sulfur-terminated, (001)-S, and metal-terminated, (001)-M. The (001)-S surface reconstructs, forming disulfide dimers with a bond length of 2.23 Å. The (001)-M surface reconstructs, reordering the metal atoms in order to form planes of metal atoms and interlaced sulfur atoms. Different adsorption sites for the water molecule were investigated. The dissociative mechanism of the water molecule has also been analyzed in detail. For the (001)-S surface, the water adsorption on the iron atom is the preferred mechanism. The dissociative mechanism leads to structures which are, at least, 13 kcal mol À1 higher in energy than the water adsorbed on iron atom. For the (001)-M surface, no minima in the potential energy surface were found, and the water molecule prefers to form a hydrogen bond with the sulfur atoms. The dissociative mechanism for the water adsorption on (001)-M surface is thermodynamically unfavorable. The metal-alloy-like structure underneath of the sulfur atoms and the unfavorable water adsorption indicate that the surface presents some hydrophobic character. The influence of the water molecule in the reconstruction of the (001) chalcopyrite surface and in its reactivity is discussed.

No abstract

A new dynamical discrete/continuum solvation model was tested for NH(4)(+) and OH(-) ions in water solvent. The method is similar to continuum solvation models in a sense that the linear response approximation is used. However, different from pure continuum models, explicit solvent molecules are included in the inner shell, which allows adequate treatment of specific solute-solvent interactions present in the first solvation shell, the main drawback of continuum models. Molecular dynamics calculations coupled with SCC-DFTB method are used to generate the configurations of the solute in a box with 64 water molecules, while the interaction energies are calculated at the DFT level. We have tested the convergence of the method using a variable number of explicit water molecules and it was found that even a small number of waters (as low as 14) are able to produce converged values. Our results also point out that the Born model, often used for long-range correction, is not reliable and our method should be applied for more accurate calculations.