By splitting the Coulomb interaction into long-range and short-range components, we decompose the energy of a quantum electronic system into long-range and short-range contributions. We show that the long-range part of the energy can be efficiently calculated by traditional wave function methods, while the short-range part can be handled by a density functional. The analysis of this functional with respect to the range of the associated interaction reveals that, in the limit of a very short-range interaction, the short-range exchange-correlation energy can be expressed as a simple local functional of the on-top pair density and its first derivatives. This provides an explanation for the accuracy of the local density approximation (LDA) for the short-range functional. Moreover, this analysis leads also to new simple approximations for the short-range exchange and correlation energies improving the LDA.

An adiabatic-connection fluctuation-dissipation theorem approach based on a range separation of electron-electron interactions is proposed. It involves a rigorous combination of short-range density functional and long-range random phase approximations. This method corrects several shortcomings of the standard random phase approximation and it is particularly well suited for describing weaklybound van der Waals systems, as demonstrated on the challenging cases of the dimers Be2 and Ne2.Density functional theory (DFT) is a powerful approach for electronic-structure calculations of molecular and condensed-matter systems [1]. However, one difficulty in its Kohn-Sham (KS) formulation using local density or generalized-gradient approximations (LDA and GGA) is the description of non-local correlation effects, such as those involved in weak van der Waals complexes, bound by London dispersion forces [2]. The adiabaticconnection fluctuation-dissipation theorem (ACFDT) approach is one of the most promising ways of constructing highly non-local correlation functionals. This approach, introduced in wave function theory [3] and in DFT [4,5], consists in extracting non-local ground-state correlations from the linear charge density response function.Recently, the ACFDT approach has received renewed interest for implementing the random phase approximation (RPA) or other related approximations for atoms, molecules and solids [6][7][8][9][10][11]. The RPA correlation energy is consistent with the use of the exact, self-interactionfree exchange energy. In spite of a number of encouraging results, such as the correct description of dispersion forces at large separation [12], the proper reproduction of cohesive energies and lattice constants of solids [10,13,14] and an improved description of bond dissociation [6,7,15], several aspects of the RPA are still unsatisfactory.First, the RPA is a poor approximation to short-range correlations, leading to correlation energies that are far too negative [16]. Second, in a Gaussian localized basis, RPA calculations have a slow convergence with respect to the basis size [6]. Third, the presence of an unphysical maximum (bump) at medium distances in dissociation curves of simple diatomic molecules [6,15] indicates an inherent problem which has not yet a fully clarified origin. Fourth, although in principle the orbitals should be calculated self-consistently [17], most RPA implementations consist of a post-KS single-iteration calculation, making the choice of the input orbitals sometimes critical. Last * Electronic address: julien.toulouse@upmc.fr † Electronic address: janos.angyan@crm2.uhp-nancy.fr but not least, although the main advantage of the RPA is supposed to be the description of dispersion forces, rare gas dimer potential curves calculated from LDA or GGA orbitals are often qualitatively wrong, as shown later.The poor short-range behavior can be corrected by adding a GGA functional constructed from the difference of the exact and RPA correlation energies of the uniform electron gas [16]...

We study three wave function optimization methods based on energy minimization in a variational Monte Carlo framework: the Newton, linear, and perturbative methods. In the Newton method, the parameter variations are calculated from the energy gradient and Hessian, using a reduced variance statistical estimator for the latter. In the linear method, the parameter variations are found by diagonalizing a nonsymmetric estimator of the Hamiltonian matrix in the space spanned by the wave function and its derivatives with respect to the parameters, making use of a strong zero-variance principle. In the less computationally expensive perturbative method, the parameter variations are calculated by approximately solving the generalized eigenvalue equation of the linear method by a nonorthogonal perturbation theory. These general methods are illustrated here by the optimization of wave functions consisting of a Jastrow factor multiplied by an expansion in configuration state functions (CSFs) for the C2 molecule, including both valence and core electrons in the calculation. The Newton and linear methods are very efficient for the optimization of the Jastrow, CSF, and orbital parameters. The perturbative method is a good alternative for the optimization of just the CSF and orbital parameters. Although the optimization is performed at the variational Monte Carlo level, we observe for the C2 molecule studied here, and for other systems we have studied, that as more parameters in the trial wave functions are optimized, the diffusion Monte Carlo total energy improves monotonically, implying that the nodal hypersurface also improves monotonically.

In many cases, the dynamic correlation can be calculated quite accurately and at a fairly low computational cost in Kohn-Sham density-functional theory (KS-DFT), using current standard approximate functionals. However, in general, KS-DFT does not treat static correlation effects (near degeneracy) adequately which, on the other hand, can be described in wave-function theory (WFT), for example, with a multiconfigurational self-consistent field (MCSCF) model. It is therefore of high interest to develop a hybrid model which combines the best of both WFT and DFT approaches. The merge of WFT and DFT can be achieved by splitting the two-electron interaction into long-range and short-range parts. The long-range part is then treated by WFT and the short-range part by DFT. In this work the authors consider the so-called "erf" long-range interaction erf(micror12)/r12, which is based on the standard error function, and where mu is a free parameter which controls the range of the long-/short-range decomposition. In order to formulate a general method, they propose a recipe for the definition of an optimal microopt parameter, which is independent of the approximate short-range functional and the approximate wave function, and they discuss its universality. Calculations on a test set consisting of He, Be, Ne, Mg, H2, N2, and H2O yield microopt approximately 0.4 a.u.. A similar analysis on other types of test systems such as actinide compounds is currently in progress. Using the value of 0.4 a.u. for micro, encouraging results are obtained with the hybrid MCSCF-DFT method for the dissociation energies of H2, N2, and H2O, with both short-range local-density approximation and PBE-type functionals.

No abstract

We present a simple, robust, and highly efficient method for optimizing all parameters of many-body wave functions in quantum Monte Carlo calculations, applicable to continuum systems and lattice models. Based on a strong zero-variance principle, diagonalization of the Hamiltonian matrix in the space spanned by the wave function and its derivatives determines the optimal parameters. It systematically reduces the fixed-node error, as demonstrated by the calculation of the binding energy of the small but challenging C(2) molecule to the experimental accuracy of 0.02 eV.

We pursue the development and application of the recently introduced linear optimization method for determining the optimal linear and nonlinear parameters of Jastrow-Slater wave functions in a variational Monte Carlo framework. In this approach, the optimal parameters are found iteratively by diagonalizing the Hamiltonian matrix in the space spanned by the wave function and its first-order derivatives, making use of a strong zero-variance principle. We extend the method to optimize the exponents of the basis functions, simultaneously with all the other parameters, namely, the Jastrow, configuration state function, and orbital parameters. We show that the linear optimization method can be thought of as a so-called augmented Hessian approach, which helps explain the robustness of the method and permits us to extend it to minimize a linear combination of the energy and the energy variance. We apply the linear optimization method to obtain the complete ground-state potential energy curve of the C(2) molecule up to the dissociation limit and discuss size consistency and broken spin-symmetry issues in quantum Monte Carlo calculations. We perform calculations for the first-row atoms and homonuclear diatomic molecules with fully optimized Jastrow-Slater wave functions, and we demonstrate that molecular well depths can be obtained with near chemical accuracy quite systematically at the diffusion Monte Carlo level for these systems.

We provide a rigorous derivation of a class of double-hybrid approximations, combining Hartree-Fock exchange and second-order Møller-Plesset correlation with a semilocal exchange-correlation density functional. These double-hybrid approximations contain only one empirical parameter and use a density-scaled correlation energy functional. Neglecting density scaling leads to a one-parameter version of the standard double-hybrid approximations. We assess the performance of these double-hybrid schemes on representative test sets of atomization energies and reaction barrier heights, and we compare to other hybrid approximations, including range-separated hybrids. Our best one-parameter double-hybrid approximation, called 1DH-BLYP, roughly reproduces the two parameters of the standard B2-PLYP or B2GP-PLYP double-hybrid approximations, which shows that these methods are not only empirically close to an optimum for general chemical applications but are also theoretically supported.