Michiel Sprik

Acidity of the Aqueous Rutile TiO2(110) Surface from Density Functional Theory Based Molecular Dynamics.

The thermodynamics of protonation and deprotonation of the rutile TiO2(110) water interface is studied using a combination of density functional theory based molecular dynamics (DFTMD) and free energy perturbation methods. Acidity constants are computed from the free energy for chaperone assisted insertion/removal of protons in fully atomistic periodic model systems treating the solid and solvent at the same level of theory. The pKa values we find for the two active surface hydroxyl groups on TiO2(110), the bridge OH (Ti2OH(+)), and terminal H2O adsorbed on a 5-fold Ti site (TiOH2) are -1 and 9, leading to a point of zero proton charge of 4, well within the computational error margin (2 pKa units) from the experimental value (4.5-5.5). The computed intrinsic surface acidities have also been used to estimate the dissociation free energy of adsorbed water giving 0.6 eV, suggesting that water dissociation is unlikely on a perfect aqueous TiO2(110) surface. For further analysis, we compare to the predictions of the MUltiSIte Complexation (MUSIC) and Solvation, Bond strength, and Electrostatic (SBE) models. The conclusion regarding the MUSIC model is that, while there is good agreement for the acidity of an adsorbed water molecule, the proton affinity of the bridging oxygen obtained in the DFTMD calculation is significantly lower (more than 5 pKa units) than the MUSIC model value. Structural analysis shows that there are significant differences in hydrogen bonding, in particular to a bridging oxygen which is assumed to be stronger in the MUSIC model compared to what we find using DFTMD. Using DFTMD coordination numbers as input for the MUSIC model, however, led to a pKa prediction which is inconsistent with the estimates obtained from the DFTMD free energy calculation.

Vibrational Sum Frequency Generation Spectroscopy of the Water Liquid–Vapor Interface from Density Functional Theory-Based Molecular Dynamics Simulations

The vibrational sum frequency generation (VSFG) spectrum of the water liquid-vapor (LV) interface is calculated using density functional theory-based molecular dynamics simulations. The real and imaginary parts of the spectrum are in good agreement with the experimental data, and we provide an assignment of the SFG bands according to the dipole orientation of the interfacial water molecules. We use an instantaneous definition of the surface, which is more adapted to the study of interfacial phenomena than the Gibbs dividing surface. By calculating the vibrational (infrared, Raman) properties for interfaces of varying thickness, we show that the bulk spectra signatures appear after a thin layer of 2-3 Å only. We therefore use this value as a criterion for calculating the VSFG spectrum.

Oxide/water interfaces: how the surface chemistry modifies interfacial water properties

The organization of water at the interface with silica and alumina oxides is analysed using density functional theory-based molecular dynamics simulation (DFT-MD). The interfacial hydrogen bonding is investigated in detail and related to the chemistry of the oxide surfaces by computing the surface charge density and acidity. We find that water molecules hydrogen-bonded to the surface have different orientations depending on the strength of the hydrogen bonds and use this observation to explain the features in the surface vibrational spectra measured by sum frequency generation spectroscopy. In particular, ice-like and liquid-like features in these spectra are interpreted as the result of hydrogen bonds of different strengths between surface silanols/aluminols and water.

The oxidation of tyrosine and tryptophan studied by a molecular dynamics normal hydrogen electrode

The thermochemical constants for the oxidation of tyrosine and tryptophan through proton coupled electron transfer in aqueous solution have been computed applying a recently developed density functional theory (DFT) based molecular dynamics method for reversible elimination of protons and electrons. This method enables us to estimate the solvation free energy of a proton (H(+)) in a periodic model system from the free energy for the deprotonation of an aqueous hydronium ion (H(3)O(+)). Using the computed solvation free energy of H(+) as reference, the deprotonation and oxidation free energies of an aqueous species can be converted to pK(a) and normal hydrogen electrode (NHE) potentials. This conversion requires certain thermochemical corrections which were first presented in a similar study of the oxidation of hydrobenzoquinone [J. Cheng, M. Sulpizi, and M. Sprik, J. Chem. Phys. 131, 154504 (2009)]. Taking a different view of the thermodynamic status of the hydronium ion, these thermochemical corrections are revised in the present work. The key difference with the previous scheme is that the hydronium is now treated as an intermediate in the transfer of the proton from solution to the gas-phase. The accuracy of the method is assessed by a detailed comparison of the computed pK(a), NHE potentials and dehydrogenation free energies to experiment. As a further application of the technique, we have analyzed the role of the solvent in the oxidation of tyrosine by the tryptophan radical. The free energy change computed for this hydrogen atom transfer reaction is very similar to the gas-phase value, in agreement with experiment. The molecular dynamics results however, show that the minimal solvent effect on the reaction free energy is accompanied by a significant reorganization of the solvent.

Density-functional molecular-dynamics study of the redox reactions of two anionic, aqueous transition-metal complexes

The thermochemistry of the RuO(4)(2-)+MnO(4)(-)-->RuO(4)(-)+MnO(4)(2-) redox reaction in aqueous solution is studied by separate density-functional-based ab initio molecular-dynamics simulations of the component half reactions RuO(4)(2-)-->RuO(4)(-)+e(-) and MnO(4)(2-)-->MnO(4)(-)+e(-). We compare the results of a recently developed grand-canonical method for the computation of oxidation free energies to the predictions by the energy-gap relations of the Marcus theory that can be assumed to apply to these reactions. The calculated redox potentials are in good agreement. The subtraction of the half-reaction free energies gives an estimate of the free energy of the full reaction. The result obtained from the grand-canonical method is -0.4 eV, while the application of the Marcus theory gives -0.3 eV. These should be compared to the experimental value of 0.0 eV. Size effects, in response to increasing the number of water molecules in the periodic model system from 30 to 48, are found to be small ( approximately 0.1 eV). The link to the Marcus theory also has enabled us to compute reorganization free energies for oxidation. For both the MnO(4)(2-) and RuO(4)(2-) redox reactions we find the same reorganization free energy of 0.8 eV (1.0 eV in the larger system). The results for the free energies and further analysis of solvation and electronic structure confirm that these two tetrahedral oxoanions show very similar behavior in solution in spite of the central transition-metal atoms occupying a different row and column in the periodic table.

Absolute pKa Values and Solvation Structure of Amino Acids from Density Functional Based Molecular Dynamics Simulation

Absolute pKa values of the amino acid side chains of arginine, aspartate, cysteine, histidine, and tyrosine; the C- and N-terminal group of tyrosine; and the tryptophan radical cation are calculated using a revised density functional based molecular dynamics simulation technique introduced previously [ Cheng , J. ; Sulpizi , M. ; Sprik , M. J. Chem. Phys. 2009 , 131 , 154504 ]. In the revised scheme, acid deprotonation is considered as a dissociation rather than a proton transfer reaction, and a correction term for treating the proton as a hydronium ion is suggested. The acidity constants of the amino acids are obtained from the vertical energy gaps for removal or insertion of the acidic proton and the computed solvation free energy of the proton. The unsigned mean error relative to experimental results is 2.1 pKa units with a maximum error of 4.0 pKa units. The estimated mean statistical uncertainty due to the finite length of the trajectories is ±1.1 pKa units. The solvation structures of the protonated and deprotonated amino acids are analyzed in terms of radial distribution functions, which can serve as reference data for future force field developments.

Aqueous Redox Chemistry and the Electronic Band Structure of Liquid Water

The electronic states of aqueous species can mix with the extended states of the solvent if they are close in energy to the band edges of water. Using density functional theory-based molecular dynamics simulation, we show that this is the case for OH(-) and Cl(-). The effect is, however, badly exaggerated by the generalized gradient approximation leading to systematic underestimation of redox potentials and spurious nonlinearity in the solvent reorganization. Drawing a parallel to charged defects in wide gap solid oxides, we conclude that misalignment of the valence band of water is the main source of error turning the redox levels of OH(-) and Cl(-) in resonant impurity states. On the other hand, the accuracy of energies of levels corresponding to strongly negative redox potentials is acceptable. We therefore predict that mixing of the vertical attachment level of CO2 and the unoccupied states of water is a real effect.

Thermal versus electronic broadening in the density of states of liquid water

Abstract The one-electron density of states of liquid water computed from an ab initio molecular dynamics trajectory is analyzed in terms of interactions between effective molecular orbitals localized on single molecules. These orbitals are constructed from the occupied extended (Kohn–Sham) orbitals using the maximally localized Wannier function method. Band positions are related to average orbital energies. The width of a band is resolved into contributions from thermal fluctuations in the orbital energies and the electronic broadening due to intermolecular coupling. It is found that the thermal and electronic broadening are of comparable magnitude with electronic broadening being the leading effect.

The electron attachment energy of the aqueous hydroxyl radical predicted from the detachment energy of the aqueous hydroxide anion.

Combining photoemission and electrochemical data from the literature we argue that the difference between the vertical and adiabatic ionization energy of the aqueous hydroxide anion is 2.9 eV. We then use density functional theory based molecular dynamics to show that the solvent response to ionization is nonlinear. Adding this to the experimental data we predict a 4.1 eV difference between the energy for vertical attachment of an electron to the aqueous hydroxyl radical and the corresponding adiabatic electron affinity. This places the state accepting the electron only 2.2 eV below vacuum or 7.7 eV above the edge of the valence band of water.

Hole Localization and Thermochemistry of Oxidative Dehydrogenation of Aqueous Rutile TiO2(110)

ion of an H atom, or oxidative dehydrogenation (ODH), is an energetically demanding process. The reversible potential for ODH of a water molecule (H2O!OHC+1/2 H2) is 2.72 V versus the standard hydrogen electrode (SHE). This is more than the double of the 1.23 V needed for the full four electron oxidation of H2O. The extra energy must also be provided by the external force driving the reaction. In photoelectrochemical splitting this is the valence band hole generated by photoexcitation. Photo-anodes must, therefore, have deep lying valence bands, which is why metal oxides are attractive materials for photo-catalytic water oxidation. TiO2 is historically one of the first metal oxides shown to be capable of evolving oxygen when exposed to light and has become the model photoanode. The valence band maximum (VBM) for the rutile form is 2.95 V in the dark and 3.25 V in the light versus SHE at pH 0. While exceeding the thermodynamic minimum, the TiO2 VBM seems to leave surprisingly little margin for the energy losses that can be expected if the localization of free holes is fast. In photocatalytic water splitting, the holes do not have to supply the full energy for ODH of H2O in the liquid. It may be easier to eliminate hydrogen from an adsorbed water molecule with the product OHC remaining as an intermediate on the surface. The question is, therefore, how much energy is saved by adsorption? This energy is not directly available from experiment but has been computed for TiO2 anodes using density functional theory (DFT) methods. 4] The focus in these calculations was on electro-oxidation. Overpotentials for electrolysis can be estimated without explicit computation of the energy of holes (the same procedure has been applied to metallic oxides such as RuO2 [5, ). Indeed, owing to the shortcomings in the generalized gradient approximation (GGA) to the density functionals used in the calculations of References [3, 4] the question of localization of holes could not be addressed in principle. Self-trapping of holes in titania has been the subject of several recent computational investigations, which have utilized advanced DFT techniques. Here we make the link to oxidation catalysis using ODH of a terminal water at the aqueous rutile TiO2 (110) interface as an example. Applying a functional, including a fraction of exact exchange (HSE06), we find that localization reduces the oxidative power of a photogenerated hole by 0.6 V. However, viewed from an electrolysis perspective, this relaxation is crucial. Without it the ODH of an adsorbed water molecule would require an even higher potential than ODH in solution. Our calculations also indicate that the net activation of ODH by the TiO2 surface is almost entirely attributable to an increase of the acidity of an adsorbed water molecule. These results have been obtained by using a DFT based molecular dynamics (DFTMD) simulation of a fully atomistic periodic slab model of the interface (see Figure 1). A similar model system was used in preliminary DFTMD calculations of the acidity and band edge energies of the aqueous TiO2 electrode. ODH is well-known in physical organic chemistry for which it is the key process in proton coupled electron transfer (PCET). For the analysis of the thermochemistry of a PCET reaction it has been found useful to resolve the reaction into Figure 1. DFTMD model of a rutile TiO2 (110) water interface from which a proton and electron has been removed. Ti, O and H atoms are represented in yellow, red, and white, respectively. The highlighted surface hydroxide is the deprotonated water. The spin density of the electronic hole is visualized by green isosurfaces with the density of 5 10 . Parts a) and b) are snapshots of MD trajectories generated using PBE and HSE06, respectively. [a] Dr. J. Cheng, Dr. M. Sulpizi, Prof.Dr. M. Sprik Department of Chemistry University of Cambridge Cambridge CB2 1EW (United Kingdom) E-mail : jc590@cam.ac.uk [b] Dr. J. VandeVondele Physical Chemistry Institute University of Zurich Winterthurerstrasse 190, CH-8057 Zurich (Switzerland) [] Present address: Johannes Gutenberg University Mainz Staudingerweg 7, 55099 Mainz (Germany) [] Present address: Department of Materials ETH Zurich Wolfgang-Pauli-Strasse 27, CH-8093 Zurich (Switzerland) Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/cctc.201100498. 636 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemCatChem 2012, 4, 636 – 640 a deprotonation and oxidation step. This is visualized for our homogeneous reference reaction in the triangle diagram of Scheme 1 a. Heterogeneous ODH will be subjected to a similar analysis. The dehydrogenation of terminal H2O is written as the expression in Equation (1): TiOH2 ! TiOHCþHþþe ð1Þ and separated in a deprotonation and ionization by Equations (2a) and (2b): TiOH2 ! TiOH þHþ ð2aÞ