The observed reactivity of MgO with water is in apparent conflict with theoretical calculations which show that molecular dissociation does not occur on a perfect (001) surface. We have performed ab-initio total energy calculations which show that a chemisorption reaction involving a reconstruction to form a (111) hydroxyl surface is strongly preferred with ∆E = −90.2 kJ mol −1 . We conclude that protonation stabilizes the otherwise unstable (111) Magnesium oxide has long provided a prototype for the study of surface structure and chemical reactions of oxides. Naturally occurring MgO, known by its mineral name of periclase, is not a common crustal mineral, but its simple structure makes it an excellent example for the investigation of mineral surface chemistry.Reactions at mineral surfaces are responsible for much of the chemical change which occurs in the Earth's crust. Weathering reactions control the erosion of rocks and the consequent evolution of surface topography thus providing an opposing mechanism to the more dramatic process of mountain building. Aqueous reactions in sedimentary basins are responsible for the diagenetic processes which transform unconsolidated sediments into rocks. In this work we study the nature of a simple mineral surface when exposed to an aqueous environment and the chemical interaction of water with that surface. This is both a prerequisite to studying the interaction with aqueous solutions and a tractable first step towards ligand-exchange reactions in more complex silicate minerals.We have performed experiments on single-crystals of MgO prepared with high-quality (001) faces which were reacted with acidic solutions. The experiments and results are reported in detail elsewhere 1 , the main feature being the development of an altered surface layer. Elastic Recoil Detection Analysis (ERDA)2 shows protonation to a depth of 900Å with a H/Mg ratio close to 2 giving a probable chemical composition of magnesium hydroxide. Indeed brucite (the mineralogical name for Mg(OH) 2 ) is the most common alteration product of periclase in the natural environment 4 and well-crystallized intergrowths of brucite on periclase have been reported 5 . The initial stage in the reaction is hydroxylation of the surface. MgO has the cubic rocksalt structure with (001) cleavage planes. This is the most stable surface and is the only one seen experimentally 6 . The simplest possibility for a hydroxylated surface is obtained by dissociating a water molecule and placing the OH group above each magnesium ion and the H above each oxygen of the (001) surface (see Fig. 1a) as postulated by Coluccia et al.7 . Some striking hydroxylation experiments were reported by Jones et al. who studied surface roughening on (001) faces of nanocrystalline MgO in a transmission electron microscope 8 . The remarkable affinity of MgO for water is demonstrated by their in situ observation of hydration-induced surface roughening over 10 minutes under vacuum with P H 2 O < 10 −5 Pa. The presence of surface hydroxyl groups on MgO powde...

No abstract

Abstract. Ab initio total energy calculations based on the local density approximation (LDA) and the generalised gradient approximation (GGA) of density functional theory have been performed for brucite, Mg(OH)2, diaspore, A1OOH and hypothetical hydrous wadsleyite, Mg7Si4O14(OH)a. The use of a general gradient approximation (GGA) is essential to obtain a good agreement (~ 1%) of the calculated lattice parameters to diffraction data. The calculated fractional coordinates of brucite and diaspore are in good agreement (~1.5%) with experimental data. The angle of the non-linear hydrogen bond in diaspore is reproduced well, and the calculated Raman active OH stretching frequency in brucite is in very good agreement with spectroscopic data. There are no significant differences between the calculated fractional coordinates and the second derivative of the energy when GGA is used instead of standard LDA. It is concluded that the description of the static and the dynamic behavior of the OH groups in these hydroxides is very good. It is therefore inferred that the parameter free model is predictive and it has been used to evaluate a hypothetical structure of hydrous wadsleyite. The model reproduces the unusual Si-O bond length of 1.7 A, observed in [3-MgzSiO 4. It predicts an O-H distance of 0.97 ~, which is significantly shorter than the distance obtained from earlier model calculations.

A series of ab initio simulations, based on density functional theory, of the structure of the clean GaAs-(001)-(2 × 4) surface and of C 2 H 2 , C 2 H 4 , and trimethylgallium (TMGa) adsorbates are described. This surface was selected because of its importance in the growth of GaAs by molecular beam epitaxy. After summarizing briefly the theoretical basis of the computational methods used in the paper, we review critically what is known from experiment and theory about the structure of the clean surface. We argue that there is now strong evidence in favor of the "trench dimer" model for the -phase of the clean surface, while the structures of the R and γ phases are less settled. We then present ab initio simulations of the trench dimer, the three dimer, and the gallium rebonded models of the clean GaAs(001)-(2 × 4) surface and discuss their common structural and bonding features. Ab initio simulations of C 2 H 2 and C 2 H 4 adsorbates at arsenic dimers of the GaAs(001)-(2 × 4) surface are then presented. The changes in the bonding configurations of both the adsorbates and the surface arsenic dimers are explained in terms of changes in the bond orders and local hybridization states. The As dimer bond is broken in the stable chemisorbed states of the molecules. However, an intermediate state, in which the As dimer is still intact, provides a significant barrier to chemisorption in both cases. This barrier, and its absence at the Si(001) surface, stems from the two extra electrons in the As dimer compared with the Si dimer. We then go on to describe the results of 14 ab initio simulations of structures connected with the chemisorption and decomposition of TMGa on the GaAs(001)-(2 × 4) surface. TMGa is commonly used in the growth of GaAs crystals from the vapor phase. The results of these simulations are used to explain a number of experimental observations concerning the surface coverage and the decomposition of TMGa to dimethylgallium and monomethylgallium. Significant technical aspects of the calculations, notably the number of relaxed layers in the slab calculations and the necessity to use gradient-corrected adsorption energies, are stressed. The paper also contains critical comments about ab initio simulations of the GaAs-(001)-(2 × 4) clean surface and about the model based on a "linear combination of structural motifs".