arXiv subjects
Publications and source records attributed to J. R. Trail.
We present an overview of the variational and diffusion quantum Monte Carlo methods as implemented in the CASINO program. We particularly focus on developments made in the last decade, describing state-of-the-art quantum Monte Carlo algorithms and software and discussing their strengths and their weaknesses. We review a range of recent applications of CASINO.
We use coupled-cluster quantum chemical methods to calculate the energetics of molecular clusters cut out of periodic molecular hydrogen structures that model observed phases of solid hydrogen. The hydrogen structures are obtained from Kohn-Sham density functional theory (DFT) calculations at pressures of 150, 250 and 350 GPa, which are within the pressure range in which phases II, III and IV are found to be stable. The calculated deviations in the DFT energies from the coupled-cluster data are reported for different functionals, and optimized functionals are generated which provide reduced errors. We give recommendations for semi-local and hybrid density functionals that are expected to accurately describe hydrogen at high pressures.
We report a systematic analysis of the performance of a widely used set of Dirac-Fock pseudopotentials for quantum Monte Carlo (QMC) calculations. We study each atom in the periodic table from hydrogen (Z=1) to mercury (Z=80), with the exception of the 4f elements (57 <= Z <= 70). We demonstrate that ghost states are a potentially serious problem when plane-wave basis sets are used in density functional theory (DFT) orbital-generation calculations, but that this problem can be almost entirely eliminated by choosing the s channel to be local in the DFT calculation; the d channel can then be chosen to be local in subsequent QMC calculations, which generally leads to more accurate results. We investigate the achievable energy variance per electron with different levels of trial wave function and we determine appropriate plane-wave cutoff energies for DFT calculations for each pseudopotential. We demonstrate that the so-called "T-move" scheme in diffusion Monte Carlo is essential for many elements. We investigate the optimal choice of spherical integration rule for pseudopotential projectors in QMC calculations. The information reported here will prove crucial in the planning and execution of QMC projects involving beyond-first-row elements.
A scheme is developed for creating pseudopotentials for use in correlated-electron calculations. Pseudopotentials for the light elements H, Li, Be, B, C, N, O, and F, are reported, based on data from high-level quantum chemical calculations. Results obtained with these correlated electron pseudopotentials (CEPPs) are compared with data for atomic energy levels and the dissociation energies, molecular geometries and zero-point vibrational energies of small molecules obtained from coupled cluster single double triple (CCSD(T)) calculations with large basis sets. The CEPPs give better results in correlated-electron calculations than Hartree-Fock-based pseudopotentials available in the literature.
Atomic force calculations within the variational and diffusion quantum Monte Carlo (VMC and DMC) methods are described. The advantages of calculating DMC forces with the "pure" rather than the "mixed" probability distribution are discussed. An accurate and practical method for calculating forces using the pure distribution is presented and tested for the SiH molecule. The statistics of force estimators are explored and violations of the Central Limit Theorem are found in some cases.
One of the goals in the development of large scale electronic structure methods is to perform calculations explicitly for a localised region of a system, while still taking into account the rest of the system outside of this region. An example of this in surface physics would be to embed an adsorbate and a few surface atoms into an extended substrate, hence considerably reducing computational costs. Here we apply the constrained electron density method of embedding a Kohn-Sham system in a substrate system (first described by P. Cortona\cite{1} and T.A. Wesolowski\cite{2}), within a plane-wave basis and pseudopotential framework. This approach divides the charge density of the system into substrate and embedded charge densities, the sum of which is the charge density of the actual system of interest. Two test cases are considered. First we construct fcc bulk aluminium by embedding one cubic lattice of atoms within another. Second, we examine a model surface/adsorbate system of aluminium on aluminium and compare with full Kohn-Sham results.
A method for the calculation of the damping rate due to electron-hole pair excitation for atomic and molecular motion at metal surfaces is presented. The theoretical basis is provided by Time Dependent Density Functional Theory (TDDFT) in the quasi-static limit and calculations are performed within a standard plane-wave, pseudopotential framework. The artificial periodicity introduced by using a super-cell geometry is removed to derive results for the motion of an isolated atom or molecule, rather than for the coherent motion of an ordered over-layer. The algorithm is implemented in parallel, distributed across both ${\bf k}$ and ${\bf g}$ space, and in a form compatible with the CASTEP code. Test results for the damping of the motion of hydrogen atoms above the Cu(111) surface are presented.
Electron-hole pair creation by an adsorbate incident on a metal surface is described using \textit{ab initio} methods. The approach starts with standard first principles electronic structure theory, and proceeds to combine classical, quantum oscillator and time dependent density functional methods to provide a consistent description of the non-adiabatic energy transfer from adsorbate to substrate. Of particular interest is the conservation of the total energy at each level of approximation, and the importance of a spin transition as a function of the adsorbate/surface separation. Results are presented and discussed for H and D atoms incident on the Cu(111) surface.
We investigate the properties of norm-conserving pseudopotentials (effective core potentials) generated by inversion of the Hartree-Fock equations. In particular we investigate the asymptotic behaviour as $\mathbf{r} \to \infty$ and find that such pseudopotentials are non-local over all space, apart from a few special special cases such H and He. Such extreme non-locality leads to a lack of transferability and, within periodic boundary conditions, an undefined total energy. The extreme non-locality must therefore be removed, and we argue that the best way to accomplish this is a minor relaxation of the norm-conservation condition. This is implemented, and pseudopotentials for the atoms H$-$Ar are constructed and tested.
We report smooth relativistic Hartree-Fock pseudopotentials (also known as averaged relativistic effective potentials or AREPs) and spin-orbit operators for the atoms H to Ba and Lu to Hg. We remove the unphysical extremely non-local behaviour resulting from the exchange interaction in a controlled manner, and represent the resulting pseudopotentials in an analytic form suitable for use within standard quantum chemistry codes. These pseudopotentials are suitable for use within Hartree-Fock and correlated wave function methods, including diffusion quantum Monte Carlo calculations.
Quantum Monte Carlo and quantum chemistry techniques are used to investigate pseudopotential models of the lithium hydride (LiH) molecule. Interatomic potentials are calculated and tested by comparing with the experimental spectroscopic constants and well depth. Two recently-developed pseudopotentials are tested, and the effects of introducing a Li core polarization potential are investigated. The calculations are sufficiently accurate to isolate the errors from the pseudopotentials and core polarization potential. Core-valence correlation and core relaxation are found to be important in determining the interatomic potential.
A new method is presented for obtaining all-electron results from a pseudopotential calculation. This is achieved by carrying out a localised calculation in the region of an atomic nucleus using the embedding potential method of Inglesfield [J.Phys. C {\bf 14}, 3795 (1981)]. In this method the core region is \emph{reconstructed}, and none of the simplifying approximations (such as spherical symmetry of the charge density/potential or frozen core electrons) that previous solutions to this problem have required are made. The embedding method requires an accurate real space Green function, and an analysis of the errors introduced in constructing this from a set of numerical eigenstates is given. Results are presented for an all-electron reconstruction of bulk aluminium, for both the charge density and the density of states.
Highly accurate experimental structure factors of silicon are available in the literature, and these provide the ideal test for any \emph{ab initio} method for the construction of the all-electron charge density. In a recent paper [J. R. Trail and D. M. Bird, Phys. Rev. B {\bf 60}, 7863 (1999)] a method has been developed for obtaining an accurate all-electron charge density from a first principles pseudopotential calculation by reconstructing the core region of an atom of choice. Here this method is applied to bulk silicon, and structure factors are derived and compared with experimental and Full-potential Linear Augmented Plane Wave results (FLAPW). We also compare with the result of assuming the core region is spherically symmetric, and with the result of constructing a charge density from the pseudo-valence density + frozen core electrons. Neither of these approximations provide accurate charge densities. The aspherical reconstruction is found to be as accurate as FLAPW results, and reproduces the residual error between the FLAPW and experimental results.
The constrained electron density method of embedding a Kohn-Sham system in a substrate system (first described by P. Cortona, Phys. Rev. B {\bf 44}, 8454 (1991) and T.A. Wesolowski and A. Warshel, J. Phys. Chem {\bf 97}, 8050 (1993)) is applied with a plane-wave basis and both local and non-local pseudopotentials. This method divides the electron density of the system into substrate and embedded electron densities, the sum of which is the electron density of the system of interest. Coupling between the substrate and embedded systems is achieved via approximate kinetic energy functionals. Bulk aluminium is examined as a test case for which there is a strong interaction between the substrate and embedded systems. A number of approximations to the kinetic-energy functional, both semi-local and non-local, are investigated. It is found that Kohn-Sham results can be well reproduced using a non-local kinetic energy functional, with the total energy accurate to better than 0.1 eV per atom and good agreement between the electron densities.
We demonstrate that unrestricted Hartree-Fock theory applied to electrons in a uniform potential has stable Wigner crystal solutions for $r_s \geq 1.44$ in two dimensions and $r_s \geq 4.5$ in three dimensions. The correlation energies of the Wigner crystal phases are considerably smaller than those of the fluid phases at the same density.
The combination of continuum Many-Body Quantum physics and Monte Carlo methods provide a powerful and well established approach to first principles calculations for large systems. Replacing the exact solution of the problem with a statistical estimate requires a measure of the random error in the estimate for it to be useful. Such a measure of confidence is usually provided by assuming the Central Limit Theorem to hold true. In what follows it is demonstrated that, for the most popular implementation of the Variational Monte Carlo method, the Central Limit Theorem has limited validity, or is invalid and must be replaced by a Generalised Central Limit Theorem. Estimates of the total energy and the variance of the local energy are examined in detail, and shown to exhibit uncontrolled statistical errors through an explicit derivation of the distribution of the random error. Several examples are given of estimated quantities for which the Central Limit Theorem is not valid. The approach used is generally applicable to characterising the random error of estimates, and to Quantum Monte Carlo methods beyond Variational Monte Carlo.
Expectation values of physical quantities may accurately be obtained by the evaluation of integrals within Many-Body Quantum mechanics, and these multi-dimensional integrals may be estimated using Monte Carlo methods. In a previous publication it has been shown that for the simplest, most commonly applied strategy in continuum Quantum Monte Carlo, the random error in the resulting estimates is not well controlled. At best the Central Limit theorem is valid in its weakest form, and at worst it is invalid and replaced by an alternative Generalised Central Limit theorem and non-Normal random error. In both cases the random error is not controlled. Here we consider a new `residual sampling strategy' that reintroduces the Central Limit Theorem in its strongest form, and provides full control of the random error in estimates. Estimates of the total energy and the variance of the local energy within Variational Monte Carlo are considered in detail, and the approach presented may be generalised to expectation values of other operators, and to other variants of the Quantum Monte Carlo method.
A method is presented for calculating electron-hole pair excitation due to an incident atom or molecule interacting with a metal surface. Energy loss is described using an \textit{ab initio} approach that obtains a position-dependent friction coefficient for an adsorbate moving near a metal surface from a total energy pseudopotential calculation. A semi-classical forced oscillator model is constructed, using the same friction coefficient description of the energy loss, to describe excitation of the electron gas due to the incident molecule. This approach is applied to H and D atoms incident on a Cu(111) surface, and we obtain theoretical estimates of the `chemicurrents' measured by Nienhaus et al [Phys. Rev. Lett. \textbf{82}, 446 (1999)] for these atoms incident on the surface of a Schottky diode.