Modelling of AlF3 catalytic surfaces
Aluminium fluoride (AlF3) is an industrially important material. It is used as a strong Lewis acid catalyst in chlorine-fluorine exchange reactions, and in the production of hydrofluorocarbons (HFCs), which is a less harmful material compared to CFC (Cholorofluorocarbon). AlF3 is also used as a coating material for mirrors in excimer lasers. Knowledge about the stability of AlF3 surfaces and their local structures is important to understand catalytic reactions although little is known about the atomic scale structure of AlF3 surfaces.
We employ surface thermodynamic calculations based on hybrid-exchange DFT to predict compositions and structures of AlF3 surfaces in contact with a gaseous environment of HF and H2O which mimics actual experimental situations. Our calculation confirms the experimental observation that a clean (001) α-AlF3 surface is not Lewis acidic. We predict that for a stable and stoichiometric (012) surface where the Al ions are five-coordinated, α-AlF3 surface is Lewis acidic. We also predict that the Lewis acidic (012) α-AlF3 surface, whose acidity is comparable to the Lewis acidity of SbF5 and of β AlF3 can only be realised at very high temperatures and extremely dry condition. The phase stability due to hydroxylation of the surfaces is shown below.
Calculations are performed using the CRYSTAL code and the B3LYP hybrid exchange functional. We have considered 54 distinct surface terminations to get the above phase diagram. Each slab containing 15 layers of atoms (or equivalent thicker slabs for non-stoichiometic systems) has been used throughout the current study. These large scale simulations have been performed using HPCx system using 128 or 256 processors. Computer time has been obtained from EPSRC sponsored Material Chemistry Consortium. This work is supported by European frame work 6 project, FUNFLUOS.