Variational solution of the single-particle Dirac equation in the field of two nuclei using relativistically adapted Slater basis functions

A variational method for solving the time-independent single-particle Dirac equation in the Coulomb field of two nuclei is described. A minimax variational principle and basis functions that have the proper analytic behavior, i.e. behave like r ^γ,γ non-integer, in the neighborhood of a nucleus, are used. A momentum space integration scheme for computing the necessary two-center integrals is described. Results are given for a standard test problem on two nuclei with Z=90 with an internuclear separation of 2.0/Z. The results confirm those of a previous calculation [F.A. Parpia and A.K. Mohanty, Chem Phys Lett 238: 209 (1995)]. Received: 13 May 1998 / Accepted: 22 June 1998 / Published online: 28 August 1998     

Generalization of coupled-cluster response theory to multireference expansion spaces: application of the coupled-cluster singles and doubles effective Hamiltonian

Employing separate cluster ansatz in time-independent and time-dependent wave-operators, coupled-cluster (CC) response theory is generalized to multireference (MR) expansion spaces. For state energies, this corresponds to the MR secular problem with an arbitrary similarity-transformed effective Hamiltonian, H˜=Ω⁻¹ HΩ. The effective Hamiltonian can be generated via size-extensive CC methods. Thus the states in MR linear response theory (MRLRT) maintain the usual CC core-extensive properties. We have used the Gelfand unitary group basis of the spin-adapted configurations to construct the matrix of H˜ in the MR excitation space. As a preliminary application, the CC singles and doubles effective Hamiltonian is applied to excitation and photoionization energies of the CH⁺ and N₂ molecules, and is compared with experimental results and results from other numerical procedures including conventional CC linear response theory (CC-LRT), MR and full configuration interaction (MRCI and FCI) methods. The numerical results indicate that MRLRT reproduces valence and external excited states quantitatively, combining the best features of CC-LRT and MRCI. Received: 2 July 1998 / Accepted: 28 August 1998 / Published online: 11 November 1998     

Argon-matrix-isolation Raman spectra and density functional study of 1,3-butadiene conformers

s-trans, s-cis and gauche conformers of 1,3-butadiene have been studied using density functional theory and the coupled-cluster method using double substitutions (CCD). Matrix isolation Raman and IR data for the minor conformer were obtained and are used in combination with the theoretical results to resolve earlier ambiguities in vibrational assignments. Based on high-quality Hessians, new harmonic stretching force constants are reported for the carbon backbone of s-trans-1,3-butadiene. For the minor conformer the best unscaled root mean square error of the calculated frequencies for the s-cis and gauche geometries are 17.5 cm⁻¹ and 7.4 cm⁻¹, respectively, primarily due to a better agreement of the gauche results for the vibrations at 983 cm⁻¹, 596 cm⁻¹ and 470 cm⁻¹ which depend strongly on the torsional angle. Although this points towards the gauche form rather than the s-cis form, the calculated transition dipole moment directions at the CCD/6-311G(d,p) level confirm the earlier conclusion that the minor conformer has C _{2 v} symmetry in the matrix. It is concluded that either the better agreement between the frequencies calculated for the gauche form and the observed values is coincidental, or that the molecule is indeed nonplanar in the matrix and tunnels very rapidly between the two mirror-image forms (or its lowest vibrational level lies above the barrier). Received: 1 July 1998 / Accepted: 26 October 1998 / Published online: 15 February 1999     

Temperature dependency of the intrinsic carrier density of hydrogenated amorphous silicon in MOS structures

The admittance versus frequency of a hydrogenated amorphous silicon metal oxide semiconductor capacitor is measured at a fixed bias in inversion and for temperatures in the range of 20–50 °C. The data are fitted to theoretical capacitance and conductance curves where the time constant of inversion is the result of the fit. In turn, the time constant can be converted to the (minority) carrier lifetime so that a lifetime value for each measurement temperature is available. The conversion from the time constant to the minority carrier lifetime requires the knowledge of the temperature-dependent intrinsic carrier density or rather its activation energy. The criterion for the correct choice is a temperature-independent carrier lifetime. Three published room temperature values of the intrinsic carrier density have been tested. The carrier lifetime activation energy is E _a = 0.70 ± 0.03 eV. Received: 17 June 1998 / Accepted: 23 October 1998     

Multireference perturbation CI III. Fast evaluation of the one-particle density matrix

Within the frame of multireference perturbation configuration interaction we have developed a fast algorithm, based on diagrammatic techniques, for the calculation of the first-order correction to the one-particle density matrix. As an example of an application we have chosen the evaluation of the dipole moment of the CO molecule, where utilization of the first-order density is shown to corroborate the variational calculation. Received: 4 August 1998 / Accepted: 21 September 1998 / Published online: 16 November 1998     

Global geometry optimization of small silicon clusters at the level of density functional theory

By an application to small silicon clusters Si _N (with N = 4,5,7,10) it is shown that truly global geometry optimization on an ab initio or density functional theory level can be achieved, at a computational cost of approximately 1–5 traditional local optimization runs (depending on cluster size). This extends global optimization from the limited area of empirical potentials into the realm of ab initio quantum chemistry. Received: 24 February 1998 / Accepted: 6 March 1998 / Published online: 17 June 1998     

Correlated and idempotent Dirac first-order density matrices with identical diagonal Fermion density: a route to extract a one-body potential energy in TDDFT

After a brief introduction to the use of the idempotent Dirac first-order density matrix (DM), its time-dependent generalization is considered. Special attention is focused on the equation of motion for the time-dependent DM, which is characterized by the one-body potential V(r, t) of time-dependent density functional theory. It is then shown how the force –∇ V(r, t) can be extracted explicitly from this equation of motion. Following a linear-response treatment in which a weak potential V(r, t) is switched on to an initially uniform electron gas, the non-linear example of the two-electron spin-compensated Moshinsky atom is a further focal point. We demonstrate explicitly how the correlated DM for this model can be constructed from the idempotent Dirac DM, in this time-dependent example.     

Multiconfigurational expansions of density operators: equations of motion and their properties

 Multiconfigurational expansions of density operators which may be used in numerical treatments of the dynamics of closed and open quantum systems are introduced. The expansions of the density operators may be viewed as analogues of those used in the multiconfiguration time-dependent Hartree (MCTDH) method, which is a well-established and highly efficient method for propagating wavepackets in several dimensions. There is no unique multiconfigurational representation of a density operator and two sensible types of MCTDH-like expansions are studied. Equations of motion for these multiconfigurational expansions are presented by adopting the Dirac–Frenkel/McLachlan variational principle (or variants of thereof). Various properties of these sets of equations of motion are derived for closed and open system dynamics. The numerical and technical aspects of this approach have been recently discussed by us [(1999) J Chem Phys 111: 8759]. Here we discuss the formal aspects of the approach in a more general context. Received: 26 January 2000 / Accepted: 8 February 2000 / Published online: 12 May 2000     

Convergence of the density functional one-centre expansion for the molecular continuum: N2 and (CH3)3N

A large-scale one-centre expansion with a radial B-spline basis set is implemented for bound and continuum states. A Kohn-Sham hamiltonian is employed with Hartree and exchange-correlation potentials calculated from the SCF electron density taken from a previous LCAO calculation. An inverse iteration method is used to obtain the continuum wavefunction, from which the cross section and asymmetry parameter are calculated. The convergence with respect to angular momentum and cut-off radius is analysed for N₂. The relevance of multipolar contributions even at large distances is shown and suggestions for further improvements are given. In order to show that the present method is suitable to treat systems of moderate size, the (CH₃)₃N molecule has also been calculated and the results are compared with experiment. Received: 19 May 1998 / Accepted: 20 August 1998 / Published online: 7 December 1998     

An implementation of the conductor-like screening model of solvation within the Amsterdam density functional package

The conductor-like screening model (COSMO) of solvation has been implemented in the Amsterdam density functional program with maximum flexibility in mind. Four cavity definitions have been incorporated. Several iterative schemes have been tested for solving the COSMO equations. The biconjugate gradient method proves to be both robust and memory-conserving. The interaction between the surface charges and the electron density may be calculated by integrating over either the fitted or exact density, or by calculating the molecular potential. A disk-smearing algorithm is applied in the former case to avoid singularities. Several self-consistent field/COSMO coupling schemes were examined in an attempt to reduce computational effort. A gradient-preserving algorithm for removing outlying charge has been implemented. Preliminary optimized radii are given. Applications to the benzene oxide-oxepin valence tautomerization and to glycine conformation are presented. Received: 13 November 1998 / Accepted: 16 December 1998 / Published online: 16 March 1999     

Efficient boundary basis functions for time-independent quantum scattering calculations

Novel parametrically energy-dependent boundary functions, F, combined with a finite L ² basis set, permit accurate and efficient calculation of scattering wavefunctions for many energies. Both accuracy and efficiency are achieved simultaneously because all the necessary integrals are energy-independent and also certain functionals, (H−E)|F>, in the Schr?dinger equation are allowed to satisfy desirable boundary conditions. In addition, slight modification of the Schr?dinger equation in the boundary region is shown to be useful for improving the numerical accuracy when a cutoff-radius-truncated basis set is used. The advantages of the present approaches are demonstrated for the one-dimensional Eckart barrier problem. Received: 17 June 1998 / Accepted: 13 July 1998 / Published online: 18 September 1998     

Density functional study of ethylene oxidation on an Ag(111) surface

The mechanism of ethylene epoxidation on Ag surfaces has been investigated using the density functional method and Ag _n clusters (n = 3 to 10) modeling the Ag(111) surface. The adsorption energy of O₂ to the Ag clusters was strongly dependent on the HOMO level of the cluster, and the clusters with higher HOMO levels afforded larger O₂ adsorption energies. The energetics was investigated for both the molecular and atomic oxygen epoxidation mechanisms. For the atomic oxygen mechanism, epoxidation was found to proceed without an activation energy, whereas a small amount of activation energy (about 5 kcal/mol) was calculated for the molecular oxygen mechanism. Received: 2 July 1998 / Accepted: 9 September 1998 / Published online: 8 February 1999     

Cyanovinyl radical: an illustration of the poor performance of unrestricted perturbation theory and density functional theory procedures in calculating radical stabilization energies

Stabilization energies for the 1-cyanovinyl radical (CH₂=CCN) have been calculated using a variety of conventional ab initio (M?ller–Plesset, quadratic configuration interaction and coupled-cluster) and density functional theory (B-LYP, B3-LYP) procedures, as well as with a range of compound methods. Compared with a high-level benchmark value (that predicts a stabilization energy of 17.1 kJ mol⁻¹), UMP2 and UMP4 give the wrong sign and magnitude of the stabilization energy (both methods predicting desta- bilization instead of stabilization), while B-LYP and B3-LYP overestimate the degree of stabilization. The RMP2, RMP4, QCISD(T) and CCSD(T) techniques, and several, but not all, variants of G2 and CBS theories give radical stabilization energies in good agreement with the benchmark value. Received: 15 June 1998 / Accepted: 19 August 1998 /  Published online: 15 February 1999     

Theoretical studies of the potential energy surface and wavepacket dynamics of the Li3 system

The three-dimensional (3D) potential energy surface of the ground state of Li₃ was determined by the multireference configuration interaction method. The vibrational motions and pseudorotation were investigated by a 3D time-dependent wavepacket formalism. The analytical expression of the 3D surface is given and the results of vibrational analyses at several critical points are presented. The low-lying excited states of Li₃ were examined for the C _{2 v} structure and the vertical and adiabatic excitation energies were calculated. The ground and singlet excited states of Li₂ were calculated and their spectroscopic constants compare well with the experimental values. A 3D wavepacket calculation was performed for simulations of the stimulated emission pumping spectrum in which the A state was taken as an intermediate. The recurrences of the autocorrelation functions were characterized by classical trajectory calculations. The autocorrelation functions obtained by wavepacket propagation are reproduced well by the accumulation of the classical trajectories in the short-time region. Received: 2 July 1998 / Accepted: 3 September 1998 / Published online: 8 February 1999     

Equation of motion of the correlated first-order density matrix for the ground-state of the Hookean atom with two electrons

The ground-state wave function Ψ for a given force constant k = 1/4 a.u. of the two-electron Hookean atom is known in exact analytical form. Here the corresponding first-order density matrix γ(r, r′) is studied, particular attention being focussed on its equation of motion. The exact form which results from the known Ψ is displayed, and given a physical interpretation. Harmonic confined model two-electron atoms with arbitrary interaction u(r ₁₂) are also briefly referred to in the present context.     

Towards an order-N DFT method

One of the most important steps in a Kohn-Sham (KS) type density functional theory calculation is the construction of the matrix of the KS operator (the “Fock” matrix). It is desirable to develop an algorithm for this step that scales linearly with system size. We discuss attempts to achieve linear scaling for the calculation of the matrix elements of the exchange-correlation and Coulomb potentials within a particular implementation (the Amsterdam density functional, ADF, code) of the KS method. In the ADF scheme the matrix elements are completely determined by 3D numerical integration, the value of the potentials in each grid point being determined with the help of an auxiliary function representation of the electronic density. Nearly linear scaling for building the total Fock matrix is demonstrated for systems of intermediate size (in the order of 1000 atoms). For larger systems further development is desirable for the treatment of the Coulomb potential. Received: 30 March 1998 / Accepted: 6 July 1998 / Published online: 15 September 1998     

Poisson-Boltzmann calculations versus molecular dynamics simulations for calculating the electrostatic potential of a solvated peptide

We performed a very long molecular dynamics simulation of a peptide in explicit water molecules and ions and averaged the electrostatic potential caused by peptide, water and ions at eight points in the vicinity of the peptide. These electrostatic potential values were directly compared to the potential calculated by solving the non-linear Poisson-Boltzmann equation for the system, which describes the solvent using continuum electrostatics. We analyze the contribution of dielectric constant, conformational flexibility and solvation effects on the electrostatic potential at these eight points. Received: 24 April 1998 / Accepted: 4 August 1998 / Published online: 23 November 1998     

Theoretical study of the valence π → π* excited states of polyacenes: anthracene and naphthacene

The valence π → π ^* excited states of anthracene and naphthacene are studied with multireference perturbation theory with complete active space self-consistent field reference functions. The predicted spectra provide a consistent assignment of all one- and two-photon spectra and T-T spectra of low-lying valence π → π ^* excited states of anthracene and naphthacene. The present theory predicts the valence π → π ^* excitation energies with an accuracy of 0.15 eV for anthracene and of 0.25 eV or better for naphthacene. The excited states of anthracene and naphthacene are compared with those of benzene and naphthalene studied previously. The present calculations predict that, going from anthracene to naphthacene, there is a symmetry reversal of the two lowest singlet state transitions, but not for the triplet, just as indicated by the experimental data. Some general trends of polyacene excited states are discussed based on the calculated results for benzene to naphthacene. Conclusive results obtained for anthracene and naphthacene can be used as a model for understanding the excited states of larger polyacenes. Received: 22 April 1998 / Accepted: 6 July 1998 / Published online: 28 September 1998     

Grid-free DFT implementation of local and gradient-corrected XC functionals

Following an approach to density functional theory calculations based on the matrix representation of operators, we implemented a scheme as an alternative to traditional grid-based methods. These techniques allow integrals over exchange-correlation operators to be evaluated through matrix manipulations. Both local and gradient-corrected functionals can be treated in a similar way. After deriving all the required expressions, selected examples with various functionals are given. Received: 7 March 1998 / Accepted: 21 May 1998 / Published on line: 6 August 1998