Diffusion Monte Carlo Research Articles

Fast Near Ab Initio Potential Energy Surfaces Using Machine Learning.

A machine-learning based approach for evaluating potential energies for quantum mechanical studies of properties of the ground and excited vibrational states of small molecules is developed. This approach uses the molecular-orbital-based machine learning (MOB-ML) method to generate electronic energies with the accuracy of CCSD(T) calculations at the same cost as a Hartree-Fock calculation. To further reduce the computational cost of the potential energy evaluations without sacrificing the CCSD(T) level accuracy, GPU-accelerated Neural Network Potential Energy Surfaces (NN-PES) are trained to geometries and energies that are collected from small-scale Diffusion Monte Carlo (DMC) simulations, which are run using energies evaluated using the MOB-ML model. The combined NN+(MOB-ML) approach is used in variational calculations of the ground and low-lying vibrational excited states of water and in DMC calculations of the ground states of water, CH5+, and its deuterated analogues. For both of these molecules, comparisons are made to the results obtained using potentials that were fit to much larger sets of electronic energies than were required to train the MOB-ML models. The NN+(MOB-ML) approach is also used to obtain a potential surface for C2H5+, which is a carbocation with a nonclassical equilibrium structure for which there is currently no available potential surface. This potential is used to explore the CH stretching vibrations, focusing on those of the bridging hydrogen atom. For both CH5+ and C2H5+ the MOB-ML model is trained using geometries that were sampled from an AIMD trajectory, which was run at 350 K. By comparison, the structures sampled in the ground state calculations can have energies that are as much as ten times larger than those used to train the MOB-ML model. For water a higher temperature AIMD trajectory is needed to obtain accurate results due to the smaller thermal energy. A second MOB-ML model for C2H5+ was developed with additional higher energy structures in the training set. The two models are found to provide nearly identical descriptions of the ground state of C2H5+.

Q-AQUA: A Many-Body CCSD(T) Water Potential, Including Four-Body Interactions, Demonstrates the Quantum Nature of Water from Clusters to the Liquid Phase.

Many model potential energy surfaces (PESs) have been reported for water; however, none are strictly from "first-principles". Here we report such a potential, based on a many-body representation at the CCSD(T) level of theory up to the four-body interaction. The new PES is benchmarked for the isomers of the water hexamer for dissociation energies, harmonic frequencies, and unrestricted diffusion Monte Carlo (DMC) calculations of the zero-point energies of the Prism, Book, and Cage isomers. Dissociation energies of several isomers of the 20-mer agree well with recent benchmark energies. Exploratory DMC calculations on this cluster verify the robustness of the new PES for quantum simulations. The accuracy and speed of the new PES are demonstrated for standard condensed phase properties, i.e., the radial distribution function and the self-diffusion constant. Quantum effects are shown to be substantial for these observables and also needed to bring theory into excellent agreement with experiment.

Point defect formation energies in graphene from diffusion quantum Monte Carlo and density functional theory

Density functional theory (DFT) is widely used to study defects in monolayer graphene with a view to applications ranging from water filtration to electronics to investigation of radiation damage in graphite moderators. To assess the accuracy of DFT in such applications, we report diffusion quantum Monte Carlo (DMC) calculations of the formation energies of some common and important point defects in monolayer graphene: monovacancies, Stone-Wales defects, and silicon substitutions. We find that standard DFT methods underestimate monovacancy formation energies by around 1 eV. The disagreement between DFT and DMC is somewhat smaller for Stone-Wales defects and silicon substitutions. We examine vibrational contributions to the free energies of formation for these defects, finding that vibrational effects are non-negligible. Finally, we compare the DMC atomization energies of monolayer graphene, monolayer silicene, and bulk silicon, finding that bulk silicon is significantly more stable than monolayer silicene by 0.7522(5) eV per atom.

The binding of atomic hydrogen on graphene from density functional theory and diffusion Monte Carlo calculations.

In this work, density functional theory (DFT) and diffusion Monte Carlo (DMC) methods are used to calculate the binding energy of a H atom chemisorbed on the graphene surface. The DMC value of the binding energy is about 16% smaller in magnitude than the Perdew-Burke-Ernzerhof (PBE) result. The inclusion of exact exchange through the use of the Heyd-Scuseria-Ernzerhof functional brings the DFT value of the binding energy closer in line with the DMC result. It is also found that there are significant differences in the charge distributions determined using PBE and DMC approaches.

Diffusion quantum Monte Carlo method on diradicals using single- and multi-determinant-Jastrow trial wavefunctions and different orbitals.

In this work, the diffusion quantum Monte Carlo (DMC) method is employed to calculate the energies of singlet and triplet states for a series of organic diradicals and diatomic diradicals with π2 configuration. Single-determinant-Jastrow (SDJ) trial wavefunctions for triplet states, two-determinant-Jastrow (2DJ) trial wavefunctions for the singlet states, and multi-determinant-Jastrow (MDJ) trial wavefunctions are employed in DMC calculations using restricted open-shell B3LYP (ROB3LYP) orbitals, complete-active-space self-consistent field (CASSCF) orbitals, state-average CASSCF orbitals, or frozen-CASSCF orbitals. Our results show that DMC energies using either SDJ/2DJ or MDJ with ROB3LYP orbitals are close to or lower than those with the other orbitals for organic diradicals, while they are not very sensitive to the employed orbitals for diatomic diradicals. Furthermore, using MDJ can reduce DMC energies to some extent for most of the investigated organic diradicals and some diatomic diradicals. The importance of MDJ on DMC energies can be estimated based on the percentage of main determinants in the CASCI wavefunction. On the other hand, singlet-triplet gaps can be calculated reasonably with DMC using MDJ with a mean absolute error of less than 2 kcal/mol with all these orbitals. CASCI wavefunctions using density functional theory orbitals are preferred in constructing MDJ trial wavefunctions in practical DMC calculations since it is easier to obtain such wavefunctions than CASSCF methods.

Correlated Wave Functions for Electron-Positron Interactions in Atoms and Molecules.

The positron, as the antiparticle of the electron, can form metastable states with atoms and molecules before its annihilation with an electron. Such metastable matter–positron complexes are stabilized by a variety of mechanisms, which can have both covalent and noncovalent character. Specifically, electron–positron binding often involves strong many-body correlation effects, posing a substantial challenge for quantum-chemical methods based on atomic orbitals. Here we propose an accurate, efficient, and transferable variational ansatz based on a combination of electron–positron geminal orbitals and a Jastrow factor that explicitly includes the electron–positron correlations in the field of the nuclei, which are optimized at the level of variational Monte Carlo (VMC). We apply this approach in combination with diffusion Monte Carlo (DMC) to calculate binding energies for a positron e+ and a positronium Ps (the pseudoatomic electron–positron pair), bound to a set of atomic systems (H–, Li+, Li, Li–, Be+, Be, B–, C–, O– and F–). For PsB, PsC, PsO, and PsF, our VMC and DMC total energies are lower than that from previous calculations; hence, we redefine the state of the art for these systems. To assess our approach for molecules, we study the potential-energy surfaces (PES) of two hydrogen anions H– mediated by a positron (e+H22–), for which we calculate accurate spectroscopic properties by using a dense interpolation of the PES. We demonstrate the reliability and transferability of our correlated wave functions for electron–positron interactions with respect to state-of-the-art calculations reported in the literature.

Intrinsic Ferromagnetism of Two-Dimensional (2D) MnO2 Revisited: A Many-Body Quantum Monte Carlo and DFT+U Study

Monolayer MnO$_2$ is one of the few predicted two-dimensional (2D) ferromagnets that has been experimentally synthesized and is commercially available. The Mermin-Wagner theorem states that magnetic order in a 2D material cannot persist unless magnetic anisotropy (MA) is present and perpendicular to the plane, which permits a finite critical temperature. Previous computational studies have predicted the magnetic ordering and Curie temperature of 2D MnO$_2$ with DFT+U (Density Funtional Theory + Hubbard U correction), with the results having a strong dependence on the Hubbard U parameter. Diffusion Monte Carlo (DMC) is a correlated electronic structure method that has had demonstrated success for the electronic and magnetic properties of a variety of 2D and bulk systems since it has a weaker dependence on the starting Hubbard parameter and density functional. In this study, we used DMC and DFT+U to calculate the magnetic properties of monolayer MnO$_2$. We found that the ferromagnetic ordering is more favorable than antiferromagnetic and determined a statistical bound on the magnetic exchange parameter ($J$). In addition, we performed spin-orbit MA energy calculations using DFT+U and using our DMC and DFT+U parameters along with the analytical model of Torelli and Olsen, we estimated an upper bound of 28.8 K for the critical temperature of MnO$_2$. These QMC results intend to serve as an accurate theoretical benchmark, necessary for the realization and development of future 2D magnetic devices. These results also demonstrate the need for accurate methodologies to predict magnetic properties of correlated 2D materials.

Gradient-descent optimization of fermion nodes in the diffusion Monte Carlo technique

We present a method for optimizing the location of the fermion ground-state nodes using a combination of diffusion Monte Carlo (DMC) and projected gradient descent (PGD). A PGD iteration shifts the parameters of an arbitrary node-fixing trial function in the opposite direction of the DMC energy gradient, while maintaining the cusp condition for atomic electrons. The energy gradient is calculated from DMC walker distributions by one of three methods we derive from an exact analytical expression. We combine our energy gradient calculation methods with different gradient descent algorithms and a projection operator that maintains the cusp condition. We apply this stochastic PGD method to trial functions of Be, Li$_2$, and Ne, all consisting of a single Slater determinant with randomized parameters, and find that the nodes dramatically improve to the same DMC energy as nodes optimized by variational Monte Carlo. Our method, therefore, departs from the standard procedure of optimizing the nodes with a non-DMC scheme such as variational Monte Carlo, Density function theory, or configuration interaction based calculation, which do not directly minimize the DMC energy.

Surrogate Hessian accelerated structural optimization for stochastic electronic structure theories.

We present an efficient energy-based method for structural optimization with stochastic electronic structure theories, such as diffusion quantum Monte Carlo (DMC). This method is based on robust line-search energy minimization in reduced parameter space, exploiting approximate but accurate Hessian information from a surrogate theory, such as density functional theory. The surrogate theory is also used to characterize the potential energy surface, allowing for simple but reliable ways to maximize statistical efficiency while retaining controllable accuracy. We demonstrate the method by finding the minimum DMC energy structures of the selected flake-like aromatic molecules, such as benzene, coronene, and ovalene, represented by 2, 6, and 19 structural parameters, respectively. In each case, the energy minimum is found within two parallel line-search iterations. The method is near-optimal for a line-search technique and suitable for a broad range of applications. It is easily generalized to any electronic structure method where forces and stresses are still under active development and implementation, such as diffusion Monte Carlo, auxiliary-field Monte Carlo, and stochastic configuration interaction, as well as deterministic approaches such as the random-phase approximation. Accurate and efficient means of geometry optimization could shed light on a broad class of materials and molecules, showing high sensitivity of induced properties to structural variables.

A quantum Monte Carlo study of systems with effective core potentials and node nonlinearities

We study beryllium dihydride (BeH$_2$) and acetylene (C$_2$H$_2$) molecules using real-space diffusion Monte Carlo (DMC) method. The molecules serve as perhaps the simplest prototypes that illustrate the difficulties with biases in the fixed-node DMC calculations that might appear with the use of effective core potentials (ECPs) or other nonlocal operators. This is especially relevant for the recently introduced correlation consistent ECPs (ccECPs) for $2s2p$ elements. Corresponding ccECPs exhibit deeper potential functions due to higher fidelity to all-electron counterparts, which could lead to larger local energy fluctuations. We point out that the difficulties stem from issues that are straightforward to address by upgrades of basis sets, use of T-moves for nonlocal terms, inclusion of a few configurations into the trial function and similar. The resulting accuracy corresponds to the ccECP target (chemical accuracy) and it is in consistent agreement with independent correlated calculations. Further possibilities for upgrading the reliability of the DMC algorithm and considerations for better adapted and more robust Jastrow factors are discussed as well.

High Accuracy Transition Metal Effective Cores for the Many-Body Diffusion Monte Carlo Method.

Practical applications of the real-space diffusion Monte Carlo (DMC) method require the removal of core electrons, where currently localization approximations of semilocal potentials are generally used in the projector. Accurate calculations of complex solids and large molecules demand minimizing the impact of approximated atomic cores. Prior works have shown that the errors from such approximations can be sizable in both finite and periodic systems. In this work, we show that a class of differential pseudopotentials, known as pseudo-Hamiltonians, can be constructed for the 3d transition metal atoms, entirely removing the need for any localization scheme in the DMC projector. As a proof of principle, we demonstrate the approach for the case of Co. In order to minimize errors in the pseudo-Hamiltonian at the many-body level, we generalize the recently proposed correlation-consistent pseudopotential generation scheme to successively close semilocal representations of the differential potentials. Our generation scheme successfully produces potentials tailored specifically for real space projector quantum Monte Carlo methods with low error at the many-body level, i.e., with many-body scattering properties very close to relativistic all-electron results. In particular, we show that the agreement with respect to atomic and molecular quantities reach chemical accuracy in many cases─on par with the most accurate semilocal pseudopotentials available. Further, our pseudo-Hamiltonian generation scheme utilizes standard quantum chemistry codes designed only to work with semilocal pseudopotentials, enabling straightforward generation of pseudo-Hamiltonians for additional elements in future works.

A combined first principles study of the structural, magnetic, and phonon properties of monolayer CrI3.

The first magnetic 2D material discovered, monolayer (ML) CrI3, is particularly fascinating due to its ground state ferromagnetism. However, because ML materials are difficult to probe experimentally, much remains unresolved about ML CrI3's structural, electronic, and magnetic properties. Here, we leverage Density Functional Theory (DFT) and high-accuracy Diffusion Monte Carlo (DMC) simulations to predict lattice parameters, magnetic moments, and spin-phonon and spin-lattice coupling of ML CrI3. We exploit a recently developed surrogate Hessian DMC line search technique to determine CrI3's ML geometry with DMC accuracy, yielding lattice parameters in good agreement with recently published STM measurements-an accomplishment given the ∼10% variability in previous DFT-derived estimates depending upon the functional. Strikingly, we find that previous DFT predictions of ML CrI3's magnetic spin moments are correct on average across a unit cell but miss critical local spatial fluctuations in the spin density revealed by more accurate DMC. DMC predicts that magnetic moments in ML CrI3 are 3.62 μB per chromium and -0.145 μB per iodine, both larger than previous DFT predictions. The large disparate moments together with the large spin-orbit coupling of CrI3's I-p orbital suggest a ligand superexchange-dominated magnetic anisotropy in ML CrI3, corroborating recent observations of magnons in its 2D limit. We also find that ML CrI3 exhibits a substantial spin-phonon coupling of ∼3.32 cm-1. Our work, thus, establishes many of ML CrI3's key properties, while also continuing to demonstrate the pivotal role that DMC can assume in the study of magnetic and other 2D materials.

A pathway toward high-throughput quantum Monte Carlo simulations for alloys: A case study of two-dimensional (2D) GaSxSe1-x.

The study of alloys using computational methods has been a difficult task due to the usually unknown stoichiometry and local atomic ordering of the different structures experimentally. In order to combat this, first-principles methods have been coupled with statistical methods such as the cluster expansion formalism in order to construct the energy hull diagram, which helps to determine if an alloyed structure can exist in nature. Traditionally, density functional theory (DFT) has been used in such workflows. In this paper, we propose to use chemically accurate many-body variational Monte Carlo (VMC) and diffusion Monte Carlo (DMC) methods to construct the energy hull diagram of an alloy system due to the fact that such methods have a weakerdependence on the starting wavefunction and density functional, scale similarly to DFT with the number of electrons, and have had demonstrated success for a variety of materials. To carry out these simulations in a high-throughput manner, we propose a method called Jastrow sharing, which involves recycling the optimized Jastrow parameters between alloys with different stoichiometries. We show that this eliminates the need for extra VMC Jastrow optimization calculations and results in significant computational cost savings (on average 1/4 savings of total computational time). Since it is a novel post-transition metal chalcogenide alloy series that has been synthesized in its few-layer form, we used monolayer GaSxSe1-x as a case study for our workflow. By extensively testing our Jastrow sharing procedure for monolayer GaSxSe1-x and quantifying the cost savings, we demonstrate how a pathway toward chemically accurate high-throughput simulations of alloys can be achieved using many-body VMC and DMC methods.

Quantum states of hydrogen cations confined into spherical and nanotube-like potential wells: , (C 2h and C 2v ) and (D 3h )

In this work, the potential energy surfaces (PESs) of hydrogen-based cations and placed inside infinite potential wells, in the shape of an infinite cylinder and a sphere, are investigated. The calculations are performed using the diffusion Monte Carlo (DMC) technique. Different symmetries are studied, some of them for the first time. The PESs trends are rationalized. The Raman shift of mode of spherically confined is also calculated.

Diffusion Monte Carlo with fictitious masses finds holes in potential energy surfaces

Holes are commonly associated with high-dimensional, non-parametric representations of potential energy surfaces (PESs). These are regions of the PES with large negative values. Typically these occur at highly distorted geometries of (in reality) high energy. Finding holes in a systematic and efficient way is thus an important aspect in the development of PESs. Here we propose using diffusion Monte Carlo (DMC) with artificially reduced masses to automatically and robustly sample large regions of a fit, including those at very high energies. We demonstrate this for a recent PES of methane that was obtained using energies and gradients. The DMC software with the methane potential is available for download.

Importance of Van der Waals Interactions in Hydrogen Adsorption on a Silicon-carbide Nanotube Revisited with vdW-DFT and Quantum Monte Carlo.

Density functional theory (DFT) is a valuable tool for calculating adsorption energies toward designing materials for hydrogen storage. However, dispersion forces being absent from the local/semi-local theory, it remains unclear as to how the consideration of van der Waals (vdW) interactions affects such calculations. For the first time, we applied diffusion Monte Carlo (DMC) to evaluate the adsorption characteristics of a hydrogen molecule on a (5,5) armchair silicon-carbide nanotube (H2-SiCNT). Within the DFT framework, we benchmarked various exchange-correlation functionals, including those recently developed for treating dispersion or vdW interactions. We found that the vdW-corrected DFT methods agree well with DMC, whereas the local (semilocal) functional significantly over (under)-binds. Furthermore, we fully optimized the H2-SiCNT geometry within the DFT framework and investigated the correlation between the structure and charge density. The vdW contribution to the adsorption was found to be non-negligible at ∼1 kcal/mol per hydrogen molecule, which amounts to 9–29% of the ideal adsorption energy required for hydrogen storage applications.

Using Diffusion Monte Carlo Wave Functions to Analyze the Vibrational Spectra of H7O3+ and H9O4.

An approach for evaluating spectra from ground state probability amplitudes (GSPA) obtained from diffusion Monte Carlo (DMC) simulations is extended to improve the description of excited state energies and allow for coupling among vibrational excited states. This approach is applied to studies of the protonated water trimer and tetramer, and their deuterated analogs. These ions provide models for solvated hydronium, and analysis of these spectra provides insights into spectral signatures of proton transfer in aqueous environments. In this approach, we obtain a separable set of internal coordinates from the DMC ground state probability amplitude. A basis is then developed from products of the DMC ground state wave function and low-order polynomials in these internal coordinates. This approach provides a compact basis in which the Hamiltonian and dipole moment matrix are evaluated and used to obtain the spectrum. The resulting spectra are in good agreement with experiment and in many cases provide comparable agreement to the results obtained using much larger basis sets. In addition, the compact basis allows for interpretation of the spectral features and how they evolve with cluster size and deuteration.

Non-Covalent Forces in Naphthazarin—Cooperativity or Competition in the Light of Theoretical Approaches

Non-covalent interactions responsible for molecular features and self-assembly in Naphthazarin C polymorph were investigated on the basis of diverse theoretical approaches: Density Functional Theory (DFT), Diffusion Quantum Monte Carlo (DQMC), Symmetry-Adapted Perturbation Theory (SAPT) and Car-Parrinello Molecular Dynamics (CPMD). The proton reaction paths in the intramolecular hydrogen bridges were studied. Two potential energy minima were found indicating that the proton transfer phenomena occur in the electronic ground state. Diffusion Quantum Monte Carlo (DQMC) and other levels of theory including Coupled Cluster (CC) employment enabled an accurate inspection of Potential Energy Surface (PES) and revealed the energy barrier for the proton transfer. The structure and reactivity evolution associated with the proton transfer were investigated using Harmonic Oscillator Model of Aromaticity - HOMA index, Fukui functions and Atoms In Molecules (AIM) theory. The energy partitioning in the studied dimers was carried out based on Symmetry-Adapted Perturbation Theory (SAPT) indicating that dispersive forces are dominant in the structure stabilization. The CPMD simulations were performed at 60 K and 300 K in vacuo and in the crystalline phase. The temperature influence on the bridged protons dynamics was studied and showed that the proton transfer phenomena were not observed at 60 K, but the frequent events were noticed at 300 K in both studied phases. The spectroscopic signatures derived from the CPMD were computed using Fourier transformation of autocorrelation function of atomic velocity for the whole molecule and bridged protons. The computed gas-phase IR spectra showed two regions with OH absorption that covers frequencies from 2500 cm to 2800 cm at 60 K and from 2350 cm to 3250 cm at 300 K for both bridged protons. In comparison, the solid state computed IR spectra revealed the environmental influence on the vibrational features. For each of them absorption regions were found between 2700–3100 cm and 2400–2850 cm at 60 K and 2300–3300 cm and 2300–3200 cm at 300 K respectively. Therefore, the CPMD study results indicated that there is a cooperation of intramolecular hydrogen bonds in Naphthazarin molecule.

GPU-Accelerated Neural Network Potential Energy Surfaces for Diffusion Monte Carlo.

Diffusion Monte Carlo (DMC) provides a powerful method for understanding the vibrational landscape of molecules that are not well-described by conventional methods. The most computationally demanding step of these calculations is the evaluation of the potential energy. In this work, a general approach is developed in which a neural network potential energy surface is trained by using data generated from a small-scale DMC calculation. Once trained, the neural network can be evaluated by using highly parallelizable calls to a graphics processing unit (GPU). The power of this approach is demonstrated for DMC simulations on H2O, CH5+, and (H2O)2. The need to include permutation symmetry in the neural network potentials is explored and incorporated into the molecular descriptors of CH5+ and (H2O)2. It is shown that the zero-point energies and wave functions obtained by using the neural network potentials are nearly identical to the results obtained when using the potential energy surfaces that were used to train the neural networks at a substantial savings in the computational requirements of the simulations.

Interactions between large molecules pose a puzzle for reference quantum mechanical methods

Quantum-mechanical methods are used for understanding molecular interactions throughout the natural sciences. Quantum diffusion Monte Carlo (DMC) and coupled cluster with single, double, and perturbative triple excitations [CCSD(T)] are state-of-the-art trusted wavefunction methods that have been shown to yield accurate interaction energies for small organic molecules. These methods provide valuable reference information for widely-used semi-empirical and machine learning potentials, especially where experimental information is scarce. However, agreement for systems beyond small molecules is a crucial remaining milestone for cementing the benchmark accuracy of these methods. We show that CCSD(T) and DMC interaction energies are not consistent for a set of polarizable supramolecules. Whilst there is agreement for some of the complexes, in a few key systems disagreements of up to 8 kcal mol−1 remain. These findings thus indicate that more caution is required when aiming at reproducible non-covalent interactions between extended molecules.