A possible mechanism of copper corrosion in anoxic water

Recent experiments show that solid Cu reacts with anoxic water. The reaction is observed by measuring the hydrogen release. This release is continuous and stable over a period of months. We have since theoretically found that water adsorbs dissociatively at a copper surface. But this adsorption is not enough to explain the amount of hydrogen released in the experiment. This observation calls for the explanation of the removal of the reaction product from the surface to provide a clean Cu surface where the water dissociation takes place. In this paper we investigate, by first-principles calculations, two possible mechanisms for this removal: first the possibility of Cu–O–H nanoparticulate formation, and second the diffusion of the dissociation products into Cu. We show that while the formation of nanoparticulates is energetically unfavorable, the diffusion of OH along grain boundaries can be substantial. The OH being placed in a grain boundary of the Cu sample quickly dissociates and O and H atoms diffuse independently of each other. Such a diffusion is markedly larger than the diffusion in bulk Cu. Thus, grain boundary diffusion is a viable mechanism for providing a clean Cu surface for the dissociation of water at the Cu surface. An order-of-magnitude estimate of the amount of hydrogen released in this case agrees with experiment. But this mechanism is not enough to explain the result of the experiment. We propose the formation of nanocrystals of copper oxide as a second step. A decisive experiment is proposed.