A tool for atomic scale simulation of corrosion: applications to Mg and Ti alloys

点击翻译按钮获取中文摘要

In 2008 the annual financial cost to the UK arising from corrosion damage to metals was $70.6 billion. In addition to the financial cost, the threat of corrosion limits the range of materials that can be safely or reliably deployed. The principal reason why magnesium alloys are not as ubiquitous as the denser aluminium is their susceptibility to aqueous corrosion even in reasonably dry air. Titanium alloys are more corrosion resistant, but in the aerospace sector suffer stress corrosion cracking following degreasing in chlorine-containing agents, or even after handling by salty fingers! When titanium is used for medical implants, corrosion is a principal cause of failure; for example, localised wear of an implant exposes a small area of metal establishing a large anodic current density and localised metal wastage. Both industry and academia agree there is an urgent need to arrive at an understanding of corrosion and passivation at the atomic scale. Modern experimental methods include electrochemical scanning tunnelling microscopy in which the tip makes an atomic resolution image of the surface as it is corroding under an applied overpotential. It is time for theory and simulation to catch up! First principles quantum mechanics has been used very successfully to make accurate and detailed calculations of both measurable and not measurable quantities central to the theory and practice of corrosion: for example, the potential of zero charge and the Galvani potential. But these are equilibrium quantities. We need to know the transmission coefficient, to solve the Butler-Volmer equation and to use atomistic simulation as it is intended: as a "microscope" to view the dynamical world at the scale of electrons and atoms. Then we can delve into the structure of the non equilibrium double layer. We here propose a novel scheme for the simulation of corrosion using molecular dynamics and kinetic Monte Carlo methods, in which we can follow the dissolution of metal ions and their transport through the double layer and into the electrolyte, and the formation and transport through a passive film, at both constant overpotential and constant current. Our most ambitious vision is to deal with localised attack: pitting and crevice corrosion. Our claim is that we can achieve this by marrying two recently demonstrated theories. These are the polarisable-ion tight binding theory (PITB), and the Hairy Probes formalism for electron open boundaries. The tight binding hamiltonian is an empirical surrogate for that of the first principles density functional theory, and the PITB is able to describe quantum electrons, ionic, covalent and metallic bonding, bond making and breaking, charge transfer and polarisable ions. The method was demonstrated recently for a wide spectrum of condensed matter, including metals, metal oxides and water. The Hairy Probes refer to a way to inject electrons into a "device region" by the maintenance of controlled electrochemical potentials at the left and right hand parts of a simulation box of atoms. The two investigators have developed these theories. Our approach will be further to develop the required computer codes and to extend the tight binding method to describe magnesium, in addition to titanium for which a tight binding hamiltonian already exists. We will demonstrate uniform corrosion and after validating against known electronic structure calculations, we will leap into the unknown. We will test current thinking about the asymmetry of the electro-capilliarity curves; make simulations of uniform corrosion of bare metal and through oxide layers and compute Evans diagrams. In parallel we will address two case studies in corrosion that have been proposed to us by our industrial project partners. These are to look at the negative difference effect in magnesium and to investigate failure of the passive layer in aerospace titanium alloys when exposed to chloride environments.