Oxidation is often associated with corrosion, but under the right conditions it can lead to oxide layers which can be applied e.g. as protective layers against corrosion, as insulating layers in microelectronic devices and as catalytic devices[1, 2]. For late transition metals and noble metals (e.g. Pd and Ag), it is now understood that the oxidation proceeds through ultra-thin oxide layers, which are thermodynamically stable at intermediate oxygen potentials and can exhibit astonishing complexity[3–6]. It is still unclear whether the same holds for transition metals and to this end the Rh (111) surface was studied extensively experimentally [7]. To supplement these experimental studies, we performed density functional calculations with the Vienna Ab Initio Simulation Package (VASP) [8] using plane waves and the PAW method [9] as well as generalized gradient approximations [10]. The experimental studies show a Moire like pattern in STM, indicative of a (8×8) oxide layer on a (9×9) supercell of the Rh(111) substrate. This phase is formed at intermediate oxygen pressures, whereas thick corundum-like Rh2O3 is only formed at significantly higher pressures and temperatures. The theoretical calculations indicate that the Moire phase can be rationalised by an ultrathin O-Rh-O trilayer surface oxide. Contrary to the experimental observation, the bulk Rh2O3 oxide is however thermodynamically more stable than ultra-thin layers. This discrepency can be understood by the surface phase diagram shown in Fig. 1(a). It indicates that a single oxygen trilayer on Rh(111) is in fact only metastable, as the trilayer forms only under conditions, where bulk Rh2O3 is already stable (to the right of the thick grey line in Fig. 1(a)). Hence the trilayer is only kinetically stabilised, contrary to the situation on Pd(111) where a Pd5O4 ad-layer is thermodynamically stable for intermediate oxygen potentials [4]. An important hint to the reason of the kinetic stability of the O-Rh-O trilayer is given by the results for the three and four layer thick oxides. Four (three) layer oxides have a lower stability than the single trilayer for oxygen potentials μO <−0.99 eV (μO <−0.78 eV) corresponding to 10 mbar (1 bar) at 800 K. The formation of the bulk oxide must however proceed through thicker oxide layers that present a kinetic barrier for the formation of the bulk oxide at too low chemical potentials. The most favorable structures for 2, 3 and 4 oxygen layers are depicted in Fig. 1(b). In contrast to oxygen atoms on the clean metal, the oxygen atoms at the oxide/metal interface are located preferentially on top of the surface Rh atoms (shifted slightly towards the bridge site). Remarkably, the trilayer termination remains favorable even for thicker oxides. For three oxygen layers (3L) the topmost surface layer contains three Rh atoms, and a fourth single Rh atom is located in the second oxide layer. For four oxygen layers (4L), the trilayer is found at both sides of the oxide, and a single Rh atom interlinks the two O-Rh-O layers. Poster
We present large-scale simulations of photovoltaics materials in the Vienna Ab initio Simulation Package [1] (VASP) that are only possible by pushing the boundaries of practically solvable eigenproblems. Thus, we enable analyses for semiconductors that were inaccessible within state-of-the-art predictive methods before. To achieve this, we a) implemented a distributed eigensolver in cuSOLVERMp that fully runs on the GPU, b) adapted VASP's Bethe-Salpeter Equation (BSE) algorithm for GPUs and to employ cuSOLVERMp, and c) dramatically improved the BSE workload distribution to yield near perfect load-balancing at scale. With a size of two million, we solve one of the largest dense, complex eigenproblems reported so far in under 5 hours sustaining 7.8 PFLOPS using 34.5 GJ on 4096 GPUs of NVIDIA's Selene supercomputer. Our results facilitate new breakthroughs in material science. These improvements in compute and energy efficiency apply to other domains relying on solving large eigenproblems, as well.
Electronic correlation energies from the random-phase approximation converge slowly with respect to the plane wave basis set size. We study the conditions under which a short-range local density functional can be used to account for the basis set incompleteness error. Furthermore, we propose a one-shot extrapolation scheme based on the Lindhard response function of the homogeneous electron gas. The different basis set correction methods are used to calculate equilibrium lattice constants for prototypical solids of different bonding types.
Density-functional calculations for the (0001)-Zn surface of wurtzite ZnO are reported. Different stabilization mechanisms, such as metallization of the surface layer, adsorption of OH groups or O adatoms, the formation of Zn vacancies, and large scale triangular reconstructions are considered. The calculations indicate that isolated Zn vacancies or O adatoms are unfavorable compared to triangular reconstructions. In the absence of hydrogen, these triangular features are stable under any realistic temperature and pressure. When hydrogen is present, the reconstruction is lifted, and hydroxyl groups stabilize the ideal otherwise unreconstructed surface. The transition between the unreconstructed hydroxyl covered surface and the triangular shaped features occurs abruptly; OH groups lift the reconstruction, but their adsorption is energetically unfavorable on the triangularly reconstructed surface.
Over the past few years considerable progress has been achieved in ab initio calculations of the structural, electronic, and dynamic properties of liquid and amorphous systems. The fundamental idea was that the equations of motion for the ionic and electronic degrees of freedom may be integrated simultaneously if the electrons are described by a pseudo-Newtonian dynamics. In metallic systems, the original Car-Parrinello approach fails because a transfer of energy from ions to electrons leads to an increase of electronic kinetic energy so that the electrons drift away from the Born-Oppenheimer surface. An alternative is to calculate the Kohn-Sham ground state of the electrons and the exact Hellmann-Feynman forces at each molecular dynamics step. This calculation is now possible using efficient conjugate-gradient techniques for energy minimization and using a subspace alignment for the prediction of the wavefunctions in new ionic configurations.
First-principles calculations of the band offsets between Cu(2)ZnSnS(4) (CZTS) and XS (X = Cd, Zn) are performed. While the interface dipole contribution for the band offsets is calculated using the Perdew-Burke-Ernzerhof functional, the Heyd-Scuseria-Ernzerhof hybrid functional is employed to introduce the quasiparticle corrections to the band offsets. The calculated conduction band offset between CZTS and CdS is 0.2 eV, validating CdS for the buffer layer of the CZTS solar cell. The small conduction band offset stems from the band gap narrowing of CdS under the interface strain caused by the lattice misfit with CZTS. A large valence band offset over 0.9 eV between CZTS and ZnS indicates that precipitated ZnS is regarded as an inactive insulator phase in CZTS absorbers.
No abstract is provided for this article.
No abstract is provided for this article.
The chemical nature and aggregate state of superheavy copernicium (Cn) have been subject of speculation for many years. While strong relativistic effects render Cn chemically inert, which led Pitzer to suggest a noble‐gas‐like behavior in 1975, Eichler and co‐workers in 2008 reported substantial interactions with a gold surface in atom‐at‐a‐time experiments, suggesting a metallic character and a solid aggregate state. Herein, we explore the physicochemical properties of Cn by means of first‐principles free‐energy calculations, which confirm Pitzer's original hypothesis: With predicted melting and boiling points of 283±11 K and 340±10 K, Cn is indeed a volatile liquid and exhibits a density very similar to that of mercury. However, in stark contrast to mercury and the lighter Group 12 metals, we find bulk Cn to be bound by dispersion and to exhibit a large band gap of 6.4 eV, which is consistent with a noble‐gas‐like character. This non‐group‐conforming behavior is eventually traced back to strong scalar‐relativistic effects, and in the non‐relativistic limit, Cn appears as a common Group 12 metal.
For semiconductor modeling, a major shortcoming of density functional theory is that the predicted band gaps are usually significantly too small. It is generally argued that this shortcoming is related to the fact that density functional theory is a ground state theory, and as a result, one is not allowed to associate the one-electron energies with the energies of quasi-particles. Although this fundamental objection is certainly correct, the modeling of the positioning of donor and acceptor levels in semiconductors faces serious limitations with present density functionals. Several solutions to this problem have been suggested. A particular attractive and fairly simple one is the inclusion of a small fraction of the non-local exchange in the Hamiltonian (hybrid functionals). This approach leads to sensible band gaps for most semiconductors, but fails for ionic solids. A more reliable approach is via many-electron Green's function techniques, which have made tremendous advances in recent years. Here GW calculations in various flavors are presented for small gap and large gap systems, comprising typical semiconductors (Si, SiC, GaAs, GaN, ZnO, ZnS, CdS and AlP), small gap semiconductors (PbS, PbSe, PbTe), insulators (C, BN, MgO, LiF) and noble gas solids (Ar, Ne). The general finding is that single-shot G(0)W(0) calculations based on wavefunctions obtained from conventional density functional theory yield too small band gaps, whereas G(0)W(0) calculations following hybrid functional calculations tend to overestimate the band gaps by roughly the same amount. This is at first sight astonishing, since the hybrid functionals yield very good band gaps themselves. The contradiction is resolved by showing that the inclusion of the attractive electron-hole interactions (excitonic effects) is required to obtain good static and dynamic dielectric functions using hybrid functionals. The corrections are usually incorporated in GW calculations using 'vertex corrections', and in fact inclusion of these vertex corrections rectifies the predicted band gaps. Finally, in order to remove the dependence on the initial wavefunctions, self-consistent GW calculations are presented, again including an approximate treatment of vertex corrections. The results are in excellent agreement with experiment, with a few per cent deviation for all materials considered. We conclude that predictive band gap engineering is now possible with the theoretical description approaching experimental accuracy.
Linear optical properties can be calculated by solving the time-dependent density functional theory equations. Linearization of the equation of motion around the ground state orbitals results in the so-called Casida equation, which is formally very similar to the Bethe-Salpeter equation. Alternatively one can determine the spectral functions by applying an infinitely short electric field in time and then following the evolution of the electron orbitals and the evolution of the dipole moments. The long wavelength response function is then given by the Fourier transformation of the evolution of the dipole moments in time. In this work, we compare the results and performance of these two approaches for the projector augmented wave method. To allow for large time steps and still rely on a simple difference scheme to solve the differential equation, we correct for the errors in the frequency domain, using a simple analytic equation. In general, we find that both approaches yield virtually indistinguishable results. For standard density functionals, the time evolution approach is, with respect to the computational performance, clearly superior compared to the solution of the Casida equation. However, for functionals including nonlocal exchange, the direct solution of the Casida equation is usually much more efficient, even though it scales less beneficial with the system size. We relate this to the large computational prefactors in evaluating the nonlocal exchange, which renders the time evolution algorithm fairly inefficient.
First-principles molecular dynamics simulations based on density-functional theory and the projector augmented wave (PAW) technique have been used to study the structural and dynamical properties of liquid iron under Earth's core conditions. As evidence for the accuracy of the techniques, we present PAW results for a range of solid-state properties of low- and high-pressure iron, and compare them with experimental values and the results of other first-principles calculations. In the liquid-state simulations, we address particular effort to the study of finite-size effects, Brillouin-zone sampling and other sources of technical error. Results for the radial distribution function, the diffusion coefficient and the shear viscosity are presented for a wide range of thermodynamic states relevant to the Earth's core. Throughout this range, liquid iron is a close-packed simple liquid with a diffusion coefficient and viscosity similar to those of typical simple liquids under ambient conditions.
