The trust-region self-consistent field method: towards a black-box optimization in Hartree-Fock and Kohn-Sham theories

The trust-region self-consistent field (TRSCF) method is presented for optimizing the total energy E(SCF) of Hartree-Fock theory and Kohn-Sham density-functional theory. In the TRSCF method, both the Fock/Kohn-Sham matrix diagonalization step to obtain a new density matrix and the step to determine the optimal density matrix in the subspace of the density matrices of the preceding diagonalization steps have been improved. The improvements follow from the recognition that local models to E(SCF) may be introduced by carrying out a Taylor expansion of the energy about the current density matrix. At the point of expansion, the local models have the same gradient as E(SCF) but only an approximate Hessian. The local models are therefore valid only in a restricted region-the trust region-and steps can only be taken with confidence within this region. By restricting the steps of the TRSCF model to be inside the trust region, a monotonic and significant reduction of the total energy is ensured in each iteration of the TRSCF method. Examples are given where the TRSCF method converges monotonically and smoothly, but where the standard DIIS method diverges.     

Calculation of nonadiabatic couplings with restricted open-shell Kohn-Sham density-functional theory

This paper generalizes the recently proposed approaches for calculating the derivative couplings between adiabatic states in density-functional theory (DFT) based on a Slater transition-state density to transitions such as singlet-singlet excitations, where a single-determinant ansatz is insufficient. The proposed approach is based on restricted open-shell Frank et al. [J. Chem. Phys. 108, 4060 (1998)] theory used to describe a spin-adapted Slater transition state. To treat the dependence of electron-electron interactions on the nuclear positions, variational linear-response density-functional perturbation theory is generalized to reference states with an orbital-dependent Kohn-Sham Hamiltonian and nontrivial occupation patterns. The methods proposed in this paper are not limited to the calculation of derivative coupling vectors, but can also be used for the calculation of other transition matrix elements. Moreover, they can be used to calculate the linear response of open-shell systems to arbitrary external perturbations in DFT.     

Globally convergent trust-region methods for self-consistent field electronic structure calculations

As far as more complex systems are being accessible for quantum chemical calculations, the reliability of the algorithms used becomes increasingly important. Trust-region strategies comprise a large family of optimization algorithms that incorporates both robustness and applicability for a great variety of problems. The objective of this work is to provide a basic algorithm and an adequate theoretical framework for the application of globally convergent trust-region methods to electronic structure calculations. Closed shell restricted Hartree-Fock calculations are addressed as finite-dimensional nonlinear programming problems with weighted orthogonality constraints. A Levenberg-Marquardt-like modification of a trust-region algorithm for constrained optimization is developed for solving this problem. It is proved that this algorithm is globally convergent. The subproblems that ensure global convergence are easy-to-compute projections and are dependent only on the structure of the constraints, thus being extendable to other problems. Numerical experiments are presented, which confirm the theoretical predictions. The structure of the algorithm is such that accelerations can be easily associated without affecting the convergence properties.     

Non self-consistent field theory

Roothaan's SCF method [2] is reformulated so that two non-SCP methods are developed to solve the eigenequation. The results from these methods can be used as starting eigenfunctions for Roothaan's SCF method.

---

Zusammenfassung Die self-consistent field Methode von Roothaan [2] wird neu formuliert: dabei werden zwei Nicht-SCF Methoden zur Lösung der Eigenwertgleichung entwickelt. Deren Resultate können als Anfangsfunktionen bei Rechnungen mit der SCF-Methode von Roothaan benutzt werden.

---

Résumé La méthode SCF de Roothaan [2] est reformulée en développant deux méthodes non-SCF pour la solution de l'équation aux valeurs propres. Les résultats de ces deux méthodes peuvent être utilisées comme fonctions de départ pour la méthode SCF de Roothaan.

---

---

The title Theoretical Chemistry has been transferred to the Division of Physical Chemistry.     

Two-configuration,self-consistent field theory

The Hartree-Fock pseudoeigenvalues equations have been derived for configuration interaction in the case of two atomic configurations ns ² np ^N and np ^N+2. A special case of investigation is the interaction between the nearly-degenerate configurations 1s ²2s ²2p ^N and 1s ²2p ^N+2. The LCAO form of the equations has also been established.

---

Zusammenfassung Hartree-Fock Pseudoeigenwertgleichungen werden für die Konfigurationswechselwirkung im Fall der zwei Atomkonfigurationen ns² np ^N und np ^N+2 abgeleitet. Ein Spezialfall, der untersucht wird, ist die Wechselwirkung zwischen den beinahe entarteten Konfigurationen 1s ²2s ²2d ^N und 1s ²2p ^N+2. Die Gleichungen werden auch in ihrer LCAO Form angegeben.

---

Résumé On établit les équations de Hartree-Fock correspondant à une interaction de configurationentre deux configurations atomiques ns ² np ^N et np ^N+2. Le cas simple d'interaction entre les deux configurations presque dégénérées 1s ²2s ²2p ^N et 1s ²2p ^N+2 a été plus particulièrement étudié. Les équations sont également présentées sous leur forme L.C.A.O.

     

Complete multi-configuration self-consistent field theory

The two-configuration self-consistent field formalism previously presented in this Journal is extended and the CMC SCF LCAO MO (complete multi-configuration self-consistent field LCAO MO) technique is presented. The single Slater determinant for a 2n electron system is replaced by a combination of determinants built from two sets of MO's, one containingn orbitals; the second, ( —n) orbitals. All the possible double excitations from the (n) set to the ( —n) set are considered. The orbitals as well as the linear combination of determinants are simultaneously optimized making use of the self-consistent field technique.

---

Zusammenfassung Es wird die Methode des self-consistent field für eine Gesamtheit mehrerer Konfigurationen in der LCAO-Näherung entwickelt. Ein 2n-Elektronensystem wird nicht mehr durch eine einzige Slaterdeterminante, sondern durch eine Kombination von Determinanten beschrieben, die aus zwei Sätzen von Molekülfunktionen mitn bzw. ( —n) Orbitalen aufgebaut werden. Alle möglichen zweifachen Anregungen vom (n) zum ( —n) Satz werden berücksichtigt. Mit Hilfe der SCF-Teehnik werden sowohl die Orbitale als auch die Kombination der Determinanten gleichzeitig optimiert.

---

Résumé On expose la méthode du champ self-consistant pour un ensemble complet de plusieurs configurations, dans l'approximation LCAO-MO (CMC SCF LCAO MO). Le déterminant de Slater pour un système de 2n électrons est remplacé par une combinaison de déterminants construits a partir de deux ensembles d'orbitales moléculaires, l'un contenant (n) orbitales et l'autre (-n) orbitales. On considère toutes les doubles excitations possibles, de l'ensemble (n) à l'ensemble (-n). La technique du champ self-consistant permet d'optimiser simultanément les orbitales ainsi que les coefficients dans la combinaison linéaire de determinants. La méthode CMC SCF tient plainement compte de la corrélation associée à chaque paire d'électrons et fait intervenir toutes les interactions paire-paire. L'optimisation simultanée des orbitales des deux ensembles (n) et (-n) garantit une convergence rapide.

     

The Gaussian equivalent representation in thermodynamic self-consistent field theory

Pair distribution functions and thermodynamic values for systems of classical particles with the Gauss and Morse interparticle interaction potentials were calculated using the approach based on the representation of the partition function and distribution functions in the form of functional integrals. The results are compared with those obtained in molecular dynamics simulations.     

Size-extensive vibrational self-consistent field method

The vibrational self-consistent field (VSCF) method is a mean-field approach to solve the vibrational Schro?dinger equation and serves as a basis of vibrational perturbation and coupled-cluster methods. Together they account for anharmonic effects on vibrational transition frequencies and vibrationally averaged properties. This article reports the definition, programmable equations, and corresponding initial implementation of a diagrammatically size-extensive modification of VSCF, from which numerous terms with nonphysical size dependence in the original VSCF equations have been eliminated. When combined with a quartic force field (QFF), this compact and strictly size-extensive VSCF (XVSCF) method requires only quartic force constants of the ?(4)V/?Q(i)(2)?Q(j)(2) type, where V is the electronic energy and Q(i) is the ith normal coordinate. Consequently, the cost of a XVSCF calculation with a QFF increases only quadratically with the number of modes, while that of a VSCF calculation grows quartically. The effective (mean-field) potential of XVSCF felt by each mode is shown to be harmonic, making the XVSCF equations subject to a self-consistent analytical solution without matrix diagonalization or a basis-set expansion, which are necessary in VSCF. Even when the same set of force constants is used, XVSCF is nearly three orders of magnitude faster than VSCF implemented similarly. Yet, the results of XVSCF and VSCF are shown to approach each other as the molecular size is increased, implicating the inclusion of unnecessary, nonphysical terms in VSCF. The diagrams of the XVSCF energy expression and their evaluation rules are also proposed, underscoring their connected structures.     

Applications of self-consistent field theory in polymer systems

1Introduction Owing to the specificity of the long chain,polymers present complexity and versatility.These molecules in the system can be various in their topological struc-tures,such as linear,star,comb or circle structures;meanwhile they can be polymeri…     

Time-dependent density-functional theory in the projector augmented-wave method

We present the implementation of the time-dependent density-functional theory both in linear-response and in time-propagation formalisms using the projector augmented-wave method in real-space grids. The two technically very different methods are compared in the linear-response regime where we found perfect agreement in the calculated photoabsorption spectra. We discuss the strengths and weaknesses of the two methods as well as their convergence properties. We demonstrate different applications of the methods by calculating excitation energies and excited state Born-Oppenheimer potential surfaces for a set of atoms and molecules with the linear-response method and by calculating nonlinear emission spectra using the time-propagation method.     

Linear-scaling implementation of molecular response theory in self-consistent field electronic-structure theory

A linear-scaling implementation of Hartree-Fock and Kohn-Sham self-consistent field theories for the calculation of frequency-dependent molecular response properties and excitation energies is presented, based on a nonredundant exponential parametrization of the one-electron density matrix in the atomic-orbital basis, avoiding the use of canonical orbitals. The response equations are solved iteratively, by an atomic-orbital subspace method equivalent to that of molecular-orbital theory. Important features of the subspace method are the use of paired trial vectors (to preserve the algebraic structure of the response equations), a nondiagonal preconditioner (for rapid convergence), and the generation of good initial guesses (for robust solution). As a result, the performance of the iterative method is the same as in canonical molecular-orbital theory, with five to ten iterations needed for convergence. As in traditional direct Hartree-Fock and Kohn-Sham theories, the calculations are dominated by the construction of the effective Fock/Kohn-Sham matrix, once in each iteration. Linear complexity is achieved by using sparse-matrix algebra, as illustrated in calculations of excitation energies and frequency-dependent polarizabilities of polyalanine peptides containing up to 1400 atoms.     

An average-of-configuration method for using Kohn-Sham density functional theory in modeling ligand-field theory

The Amsterdam Density Functional (ADF) package has been used to constrain Kohn-Sham DFT in such a fashion that a transition from KS-DFT to ligand-field theory in the form of the parametrical d(q)() model is completely well-defined. A relationship is established between the strong-field approximation of the parametrical d(2) model for the tetrahedral complexes VCl(4)(-) and VBr(4)(-) and certain fixed-orbital ADF-computed energies. In this way values for all the parameters of the d(2)() model may be computed, thus allowing the ADF results to be expressed in terms of a KS-DFT energy matrix that can be diagonalized. This means that the KS-DFT deficiency with regard to computation of nondiagonal elements has been overcome and the KS-DFT eigenenergies have become available through the KS-DFT mimicking of the ligand-field plus repulsion model. By using mutually orthogonal strong-field energy matrices, the mimicking has been further elucidated. The computed values for the empirical parameters of VCl(4)(-) and VBr(4)(-) are in good agreement with experimental data. The spectrochemical and the nephelauxetic series have been computed by including the remaining halide complexes and the quantitatively special position of F(-)() among the halides corroborated for both series.     

Linear-scaling implementation of molecular electronic self-consistent field theory

A linear-scaling implementation of Hartree-Fock and Kohn-Sham self-consistent field (SCF) theories is presented and illustrated with applications to molecules consisting of more than 1000 atoms. The diagonalization bottleneck of traditional SCF methods is avoided by carrying out a minimization of the Roothaan-Hall (RH) energy function and solving the Newton equations using the preconditioned conjugate-gradient (PCG) method. For rapid PCG convergence, the Lowdin orthogonal atomic orbital basis is used. The resulting linear-scaling trust-region Roothaan-Hall (LS-TRRH) method works by the introduction of a level-shift parameter in the RH Newton equations. A great advantage of the LS-TRRH method is that the optimal level shift can be determined at no extra cost, ensuring fast and robust convergence of both the SCF iterations and the level-shifted Newton equations. For density averaging, the authors use the trust-region density-subspace minimization (TRDSM) method, which, unlike the traditional direct inversion in the iterative subspace (DIIS) scheme, is firmly based on the principle of energy minimization. When combined with a linear-scaling evaluation of the Fock/Kohn-Sham matrix (including a boxed fitting of the electron density), LS-TRRH and TRDSM methods constitute the linear-scaling trust-region SCF (LS-TRSCF) method. The LS-TRSCF method compares favorably with the traditional SCF/DIIS scheme, converging smoothly and reliably in cases where the latter method fails. In one case where the LS-TRSCF method converges smoothly to a minimum, the SCF/DIIS method converges to a saddle point.     

Multiconfiguration self-consistent field theory of localized orbitals: The secular problem

In MC SCF theory both the orbitals and the wavefunction expansion coefficients are optimized by minimizing the energy of the system with respect to arbitrary variations of the orbitals and the wavefunction expansion coefficients. This procedure leads to separate equations for the optimal orbitals and to the secular equation for the expansion coefficients which must be solved self-consistently. In a previous paper we discussed the properties of several different choices of localization potential which may be used in the orbital equation. In this paper we derive an alternative native secular equation by making use of the transformation properties of the localized orbitals. This secular equation is considerably simpler than the conventional secular equation and leads to a simplified scheme for the self-consistent solution of the orbital and secular equations.     

Four-component relativistic Kohn-Sham theory

A four-component relativistic implementation of Kohn-Sham theory for molecular systems is presented. The implementation is based on a nonredundant exponential parametrization of the Kohn-Sham energy, well suited to studies of molecular static and dynamic properties as well as of total electronic energies. Calculations are presented of the bond lengths and the harmonic and anharmonic vibrational frequencies of Au(2), Hg(2+)(2), HgAu(+), HgPt, and AuH. All calculations are based on the full four-component Dirac-Coulomb Hamiltonian, employing nonrelativistic local, gradient-corrected, and hybrid density functionals. The relevance of the Coulomb and Breit operators for the construction of relativistic functionals is discussed; it is argued that, at the relativistic level of density-functional theory and in the absence of a vector potential, the neglect of current functionals follows from the neglect of the Breit operator.     

Exchange methods in Kohn-Sham theory

Differences between exchange methods in exchange-only Kohn-Sham theory are highlighted by calculations of diatomic molecule total energies, uncoupled isotropic NMR shieldings, and HOMO-LUMO eigenvalue differences. Optimised effective potential (OEP) and Wu-Yang (WY) results are very similar. Localised Hartree-Fock (LHF) and Krieger-Li-Iafrate (KLI) results are close to one another, but are different to OEP and WY. Becke 1988 exchange (B88X) is different again. Shieldings reduce from OEP/WY to LHF/KLI to B88X, which is consistent with an observed reduction in HOMO-LUMO gaps. LHF, KLI, and B88X shieldings and HOMO-LUMO gaps are closer to near-exact, correlated values, than are the OEP values. These variations arise entirely due to differences in the one-electron exchange potentials, which is clearly evident in potential difference plots, relative to OEP, for the N2 molecule. Density difference plots are also presented, which exhibit a spatial correlation with the potential differences. HOMO and LUMO probability density difference plots show a contraction of the LUMO relative to OEP, which is consistent with the NMR and HOMO-LUMO findings. Plots are also presented for near-exact, correlated Kohn-Sham calculations. The features are qualitatively similar to those observed in the LHF, KLI, and B88X plots, highlighting correlated character in these approximate exchange-only calculations.     

Eigenvalue independent partitioning and direct minimization in self-consistent field theory

It is pointed out that the matrix elements of a partitioning operator, f, introduced previously for constructing effective hamiltonians, provide natural sets of unconstrained and non-redundant variables in which to formulate self-consistent field theory. The elements of f are complementary to corresponding elements of the density matrix, the relation being essentially that between covariant and contravariant vectors. Some consequences for direct minimization SCF techniques are considered.     

Car-Parrinello treatment for an approximate density-functional theory method

The authors formulate a Car-Parrinello treatment for the density-functional-based tight-binding method with and without self-consistent charge corrections. This method avoids the numerical solution of the secular equations, the principal drawback for large systems if the linear combination of atomic orbital ansatz is used. The formalism is applicable to finite systems and for supercells using periodic boundary conditions within the Gamma-point approximation. They show that the methodology allows the application of modern computational techniques such as sparse matrix storage and massive parallelization in a straightforward way. All present bottlenecks concerning computer time and consumption of memory and memory bandwidth can be removed. They illustrate the performance of the method by direct comparison with Born-Oppenheimer molecular dynamics calculations. Water molecules, benzene, the C(60) fullerene, and liquid water have been selected as benchmark systems.     

Pseudopotential method for the direct construction of localized orbitals by multiconfiguration self-consistent field theory

The orbital equations for the direct construction of localized fixed orbitals by multiconfiguration self-consistent field theory (MCSCF-FXO) are transformed without approximation into pseudopotential form by a two-step process. First the utilization of a particular family of localization is shown to separate the set of orbital equations into two sets of coupled equations, one describing “valence” orbitals and one describing “core” orbitals. In addition we obtain by appropriate choice of localization potential three different sets of MCSCF-FXO orbitals, namely: maximally screened, “one-center” and “intermediate” orbitals. In the second step the orbital equations are transformed into pseudopotential form and explicit non-local pseudopotentials yielding and core orbitals are obtained. Finally, several different physically motivated approximations to the exact pseudopotentials, and the frozen-core approximation are discussed.