The valence orbitals of NH3 by electron momentum spectroscopy: Quantitative comparisons using Hartree-Fock limit and correlated wavefunctions

The complete valence shell binding energy spectra and valence orbital electron momentum distributions for NH₃ have been measured by high-momentum-resolution electron momentum spectroscopy (EMS). The results are quantitatively compared with theoretical calculations using SCF wavefunctions ranging from DZ quality to a newly developed 126-GTO wavefunction essentially at the Hartree-Fock limit. The 3a₁ and to a lesser extent the 2a₁ valence orbital are not adequately described even at the Hartree-Fock limit with basis set saturation including diffuse functions. The differences between theory and experiment are largely resolved by ion-neutral overlap calculations using CI wavefunctions to incorporate the effects of electron correlation. The 126-G (CI) wavefunctions provide accurate calculation of a wide range of electronic properties of NH₃ and also give good quantitative prediction of the three valence orbital momentum distributions as well as a reasonable prediction of the many-body pole strength distribution observed in the (2a₁)⁻¹ inner valence binding energy spectrum. The present EMS results are compared with recent investigations of wavefunction tails by exterior electron distribution calculations and Penning ionization electron spectroscopy measurements reported by Ohno et al.     

On the SCF calculation of excited states: Singlet states in the two-electron problem

The problem of determining SCF wave functions for excited electronic states is examined for singlet states of two-electron systems using a Lowdin natural orbital transformation of the full CI wave function. This analysis facilitates the comparison of various SCF methods with one another. The distribution of the full CI states among the natural orbital MCSCF states is obtained for the S states of helium using a modest Gaussian basis set. For SCF methods that are not equivalent to the full CI wave functions, it is shown that the Hartree-Fock plus all single excitation wave functions are equivalent to that of Hartree-Fock plus one single excitation. It is further shown that these wave functions are equivalent to the perfect pair or TCSCF wave functions in which the CI expansion coefficients are restricted to have opposite signs. The case of the natural orbital MCSCF wave function for two orbitals is examined in greater detail. It is shown that the first excited state must always be found on the lower natural orbital MCSCF CI root, thus precluding the use of the Hylleras-Undeim-MacDonald (HUM) theorem in locating this state. It is finally demonstrated that the solution obtained by applying the HUM theorem (minimizing the upper MCSCF CI root with respect to orbital mixing parameters) is an artifact of the MCSCF method and does not correspond to any of the full CI states.     

The spin-polarized extended Brueckner orbitals

Conventional natural and Brueckner orbitals (BOs) are rather frequently used for improving active orbital spaces in various configuration interaction (CI) approaches. However, the natural and Brueckner single-determinant models per se fail to give an adequate picture of highly correlated and quasidegenerate states such as open-shell singlet and dissociative states. We suggest the use of the spin-polarized extended BOs formally defining them in the same manner as in Lo?wdin's spin-extended Hartree-Fock method. Such BO orbitals turn out to be quite flexible and particularly useful for analyzing highly correlated electronic states. It is shown that the extended BOs always exist, unlike the usual unrestricted BOs. We discuss difficulties related to violation of size-consistency for spin projected determinant models. The working algorithm is proposed for computing BOs within the full CI and related complete active space methodology. The extended BOs are analyzed in terms of the special density-like matrices associated with spin-up and spin-down BO orbitals. From these density matrices, the corresponding spin-polarization diagrams are produced for effectively unpaired (essentially correlated) electrons. We illustrate the approach by calculations on cyclic hydrogen clusters (H(4), H(6), and H(8)), certain carbene diradicals and monoradicals, and low-lying excited states. The computations show that the BO spin-projected determinant provides a strong overlap with the multi-configurational state even for quasidegenerate states and bond breaking processes.     

Approximate MCSCF optimization for multiconfigurational spin‐tensor electron propagator method (MCSTEP): The vertical ionization potentials of CO,HCN, HNC,H2CO,and O3

The multiconfigurational spin tensor electron propagator method (MCSTEP) was developed as an implementation of electron propagator/single particle Green's function methods. MCSTEP was specifically designed for open shell and highly correlated (nondynamically correlated) initial states. The initial state used in MCSTEP is typically a small complete active space (CAS) with multiconfigurational self‐consistent field (MCSCF) state. In some cases, because of our use of a small CAS in MCSTEP, the Lagrangian eigenvalues of the MCSCF reference state are in an undesired order (u). The desired order (d) can usually be obtained by excluding one or more orbital rotations in MCSCF optimization between the doubly occupied and partially occupied orbitals. We systematically examine several cases where the undesired order occurs for the low‐lying vertical MCSTEP ionization potentials (IPs) of the molecules CO, HCN, HNC, H₂CO, and O₃ with our recently established CAS choices for MCSCF/MCSTEP. By excluding one or more orbital rotations between the partially and doubly occupied orbitals, an approximate MCSCF reference state with the same CAS choice is obtained for use in standard MCSTEP calculations that, in general, gives more reliable vertical MCSTEP IPs. © 2007 Wiley Periodicals, Inc. J Quantum Chem, 2008     

Implementation of a general multireference configuration interaction procedure with analytic gradients in a semiempirical context using the graphical unitary group approach

The graphical unitary group approach has been applied in an efficient implementation of a general multireference configuration interaction (MRCI) method for use with small active molecular orbital spaces in a semiempirical framework. Gradients can be computed analytically for molecular orbitals from a closed-shell or a half-electron open-shell Hartree-Fock calculation. CPU times for single point energy and gradient calculations are reported. The code allows MRCI geometry optimizations of large molecules, as illustrated for the singlet ground state and the four lowest triplet states of fullerene C(76).     

Generalized-molecular-orbital theory: Simple multiconfiguration self-consistent-field method

Even after completing a multiconfiguration self-consistent-field (MCSCF ) calculation, one must often include additional configuration interaction (CI ) to obtain quantitative or semiquantitative results. There is some question of whether the prior MCSCF calculation is worthwhile, if additional CI is needed later. We have developed a new MCSCF computational method, which, because of our assumptions about the nature of the configurations, yields one Fock-like operator for all the “filled” orbitals (high occupation numbers) and a second Fock-like operator for all the “virtual” orbitals (low occupation numbers). Since there are only two matrices to build, our method is considerably faster than other MCSCF approaches. Because of these similarities to standard molecular-orbital (MO ) calculations, we have termed our approach generalized-molecular-orbital (GMO ) theory. However, the “virtual” orbitals, unlike those of standard MO theory, are optimized to correlate the “filled” ones and can he used in a subsequent CI calculation. Results are presented for the correlation energy of H₂O, the spectroscopic constants of N₂, the singlet–triplet energy separations in CH₂, and the nature of the chromium–chromium quadruple bond. Although these results are at a very low level of CI , the GMO approach appears to correct for the gross deficiencies of the single-determinant SCF procedure.     

Hartree-Fock orbitals for complex-scaled configuration interaction calculation of highly excited Feshbach resonances

We examine a complex-scaled configuration interaction [(CS)CI] for highly excited Feshbach resonances, where we study the 2s(2) resonance of helium as a test case. Sizable full-CI calculations are reduced by using a correctly defined minimum active space. We compare the convergence of the minimum active space for conventional Hartree-Fock (HF) orbitals obtained as solutions to Hermitian HF equations, to the convergence of minimum active space for complex orbitals obtained as solutions to complex-scaled HF equations. Ground-state optimized orbitals are compared to a simple modification of the HF method using the excited-state mean-field potential.     

A classical versus quantum mechanics study of the \hbox{OH}\,+\,\hbox{CO} \rightarrow\,\hbox{H}\,+\,\hbox{CO}_2 (J?=?0) reaction

When appropriately used, the multiconfigurational self-consistent field (MCSCF) approximation is useful in discerning correct electronic structure results. However, with the increasing size of chemical systems of interest, MCSCF rapidly becomes unfeasible due to the requirement of larger active spaces, which lead to computationally unmanageable numbers of configurations. This situation is especially true for complete active space self-consistent field (CASSCF). In particular, reducing this computational expense by using restricted active spaces in solving for gradients and nonadiabatic couplings (NACs) during dynamics runs would save computer time. However, the validity of such restricted spaces is not well-known even for recovering the majority of the nondynamical correlation and inevitably varies between chemical systems across a range of nuclear geometries. As such, we use the recently implemented coupled perturbed–occupation restricted multiple active space (CP-ORMAS) equations (West et al., unpublished) to verify the accuracy of this approximation for gradients and NACs vectors around two specific conical intersection geometries for the silaethylene and butadiene systems. Overall, no excitations between appropriate subspaces show qualitatively reasonable results while single excitations significantly improve ORMAS results relative to the CASSCF level in these particular systems. However, single excitation schemes do not always lead to the correct orbital subspaces, and as a result, seemingly smooth potential energy surfaces (PES) do not always result in smooth analytical gradients and NACs. In addition, while some of the single excitation ORMAS and CASSCF schemes have improper orbitals rotate into the active space, the schemes without excitations (even with more subspaces) do not exhibit this behavior.     

Pathway analysis of super-exchange electronic couplings in electron transfer reactions using a multi-configuration self-consistent field method

We present a novel pathway analysis of super-exchange electronic couplings in electron transfer reactions using localized molecular orbitals from multi-configuration self-consistent field (MCSCF) calculations. In our analysis, the electronic coupling and the tunneling pathways can be calculated in terms of the configuration interaction (CI) Hamiltonian matrix obtained from the localized MCSCF wave function. Making use of the occupation restricted multiple active spaces (ORMAS) method can effectively produce the donor, acceptor, and intermediate configuration state functions (CSFs) and CIs among these CSFs. In order to express the electronic coupling as a sum of individual tunneling pathways contributions, we employed two perturbative methods: L?wdin projection-iteration method and higher-order super-exchange method. We applied them to anion couplings of butane-1,4-diyl and pentane-1,5-diyl. The results were (1) the electronic couplings calculated from the two perturbative methods were in reasonable agreement with those from a non-perturbative method (one-half value of the energy difference between the ground and first excited states), (2) the main tunneling pathways consisted of a small number of lower-order super-exchange pathways where bonding, anti-bonding, or extra-valence-shell orbitals were used once or twice, and (3) the interference among a huge number of higher-order super-exchange pathways significantly contributed to the overall electronic coupling, whereas each of them contributed only fractionally. Our method can adequately take into account both effects of non-dynamical electron correlation and orbital relaxation. Comparing with the analyses based on the Koopmans' theorem (ignoring both effects) and the ORMAS-CIs from frozen localized reference orbitals (ignoring the effect of orbital relaxation), we discuss these effects.     

Extended Koopmans' theorem ionization potentials for beryllium atom shake-up transitions

The ionization potentials were calculated for Be using the extended Koopmans' theorem (EKT ) using several full configuration interaction (CI ) and multiconfigurational-self-consistent-field (MCSCF ) wave functions as reference wave functions. The wave functions used account for 89.7–96.7% of the correlation energy. Comparisons are made with experimental values and with δCI values calculated as the difference in energy obtained from CI wave functions for Be and Be⁺. The best EKT IP differed from the δCI value by 0.0003 eV for the lowest IP and by 0.0006 eV for ionization into the lowest ²P state of Be⁺. A calculation of ionization into the second ²P state of Be⁺ requires diffuse orbitals that are unimportant in the wave function for the ground state of Be. This results in small natural orbital occupation numbers for natural orbitals needed in the EKT calculation. © 1994 John Wiley & Sons, Inc.     

A full-configuration interaction "nuclear orbital" method to study doped 3HeN clusters (N< or =4)

An efficient full configuration interaction (FCI) treatment, based on the Jacobi-Davidson algorithm, is developed in order to study small doped (3)He(N) clusters. The state of each He atom in a given cluster is described by a set of wave-functions which by extention of the quantum-chemistry notation are caller here "nuclear orbitals". The FCI treatment is applied to the calculation of binding energies and helium natural orbitals of (3)He(N)...Br(2)(X) complexes. In agreement with our previous calculations using a Hartree-Fock approach [Phys. Rev. Lett. 93, 053401 (2004)], in which the He-He interaction is modified at small distances to account for short-range correlation effects, the lowest-energy states of each multiplet are found to be very close in energy. The natural orbital analysis, in turn, indicates the adequacy of the "nuclear orbital" approach in these systems.     

Equivalent orbitals for multiconfigurational spin‐tensor electron propagator method (MCSTEP): The vertical ionization potentials of B,NO, CF,and OF

The multiconfigurational spin tensor electron propagator method (MCSTEP) was developed as an implementation of electron propagator/single particle Green's function methods for ionization potentials (IPs) and electron affinities (EAs). MCSTEP was specifically designed for open shell and highly correlated (nondynamically correlated) initial states. For computational efficiency the initial state used in MCSTEP is typically a small complete active space (CAS) multiconfigurational self‐consistent field (MCSCF) state. If in a molecule there are some degenerate orbitals which are not fully or half occupied, usual MCSCF calculations will make these orbitals inequivalent, i.e., the occupied ones will be different from the nonoccupied ones, so that the degeneracy is broken. In this article, we use a state averaged MCSCF method to get equivalent orbitals for the initial state and import the integrals into the subsequent MCSTEP calculations. This gives, in general, more reliable MCSTEP vertical IPs. © 2008 Wiley Periodicals, Inc., 2008     

Strong electron correlation in the decomposition reaction of dioxetanone with implications for firefly bioluminescence

Dioxetanone, a key component of the bioluminescence of firefly luciferin, is itself a chemiluminescent molecule due to two conical intersections on its decomposition reaction surface. While recent calculations of firefly luciferin have employed four electrons in four active orbitals [(4,4)] for the dioxetanone moiety, a study of dioxetanone [F. Liu et al., J. Am. Chem. Soc. 131, 6181 (2009)] indicates that a much larger active space is required. Using a variational calculation of the two-electron reduced-density-matrix (2-RDM) [D. A. Mazziotti, Acc. Chem. Res. 39, 207 (2006)], we present the ground-state potential energy surface as a function of active spaces from (4,4) to (20,17) to determine the number of molecular orbitals required for a correct treatment of the strong electron correlation near the conical intersections. Because the 2-RDM method replaces exponentially scaling diagonalizations with polynomially scaling semidefinite optimizations, we readily computed large (18,15) and (20,17) active spaces that are inaccessible to traditional wave function methods. Convergence of the electron correlation with active-space size was measured with complementary RDM-based metrics, the von Neumann entropy of the one-electron RDM as well as the Frobenius and infinity norms of the cumulant 2-RDM. Results show that the electron correlation is not correctly described until the (14,12) active space with small variations present through the (20,17) space. Specifically, for active spaces smaller than (14,12), we demonstrate that at the first conical intersection, the electron in the σ(?) orbital of the oxygen-oxygen bond is substantially undercorrelated with the electron of the σ orbital and overcorrelated with the electron of the carbonyl oxygen's p orbital. Based on these results, we estimate that in contrast to previous treatments, an accurate calculation of the strong electron correlation in firefly luciferin requires an active space of 28 electrons in 25 orbitals, beyond the capacity of traditional multireference wave function methods.     

Multiconfiguration self-consistent-field theory based upon the fragment molecular orbital method

The fragment molecular orbital (FMO) method was combined with the multiconfiguration self-consistent-field (MCSCF) theory. One- and two-layer approaches were developed, the former involving all dimer MCSCF calculations and the latter limiting MCSCF calculations to a small part of the system. The accuracy of the two methods was tested using the six electrons in six orbitals complete active space type of MCSCF and singlet spin state for phenol+(H(2)O)(n), n=16,32,64 (6-31G( *) and 6-311G( *) basis sets); alpha helices and beta strands of phenylalanine-(alanine)(n), n=4,8,16 (6-31G( *)). Both double-zeta and triple-zeta quality basis sets with polarization were found to have very similar accuracy. The error in the correlation energy was at most 0.000 88 a.u., the error in the gradient of the correlation energy was at most 6.x10(-5) a.u./bohr and the error in the correlation correction to the dipole moment was at most 0.018 D. In addition, vertical singlet-triplet electron excitation energies were computed for phenol+(H(2)O)(n), (n=16,32,64), 6-31G( *), and the errors were found to be at most 0.02 eV. Approximately linear scaling was observed for the FMO-based MCSCF methods. As an example, an FMO-based MCSCF calculation with 1262 basis functions took 98 min on one 3.0 GHz Pentium4 node with 1 Gbyte RAM.     

Valence-bond calculations with polarized atomic orbitals

A new method for computing polarized atomic orbitals (PAOs) is described: this method leads to very easy calculations. The space of the resulting PAOs is close to that of MC SCF MOs. Using these PAOs in the frame of a VB calculation has led to the same level of accuracy as the comparable MC SCF calculation for the dissociation energies and the lowest electronic transition energies of H₂, H₃ and N₂.     

A complete active space SCF method (CASSCF) using a density matrix formulated super-CI approach

A density matrix formulation of the super-CI MCSCF method is presented. The MC expansion is assumed to be complete in an active subset of the orbital space, and the corresponding CI secular problem is solved by a direct scheme using the unitary group approach. With a density matrix formulation the orbital optimization step becomes independent of the size of the CI expansion. It is possible to formulate the super-CI in terms of density matrices defined only in the small active subspace; the doubly occupied orbitals (the inactive subspace) do not enter. Further, in the unitary group formalism it is straightforward and simple to obtain the necessary density matrices from the symbolic formula list. It then becomes possible to treat very long MC expansions, the largest so far comprising 726 configurations. The method is demonstrated in a calculation of the potential curves for the three lowest states (¹Σ⁺_g, ³Σ⁺_u and ³Π_g) of the N₂ molecule, using a medium-sized gaussian basis set. Seven active orbitals were used yielding the following results: D_e: 8.76 (9.90), 2.43 (3.68) and 3.39 (4.90) eV; r_e: 1.108 (1.098), 1.309 (1.287) and 1.230 (1.213) Å; ω_e: 2333 (2359), 1385 (1461) and 1680 (1733) cm^?1, for the three states (experimental values within parentheses). The results of these calculations indicate that it is important to consider not only the dissociation limit but also the united atom limit in partitioning the occupied orbital space into an active and an inactive part.     

Local orbitals by minimizing powers of the orbital variance

It is demonstrated that a set of local orthonormal Hartree-Fock (HF) molecular orbitals can be obtained for both the occupied and virtual orbital spaces by minimizing powers of the orbital variance using the trust-region algorithm. For a power exponent equal to one, the Boys localization function is obtained. For increasing power exponents, the penalty for delocalized orbitals is increased and smaller maximum orbital spreads are encountered. Calculations on superbenzene, C(60), and a fragment of the titin protein show that for a power exponent equal to one, delocalized outlier orbitals may be encountered. These disappear when the exponent is larger than one. For a small penalty, the occupied orbitals are more local than the virtual ones. When the penalty is increased, the locality of the occupied and virtual orbitals becomes similar. In fact, when increasing the cardinal number for Dunning's correlation consistent basis sets, it is seen that for larger penalties, the virtual orbitals become more local than the occupied ones. We also show that the local virtual HF orbitals are significantly more local than the redundant projected atomic orbitals, which often have been used to span the virtual orbital space in local correlated wave function calculations. Our local molecular orbitals thus appear to be a good candidate for local correlation methods.     

Natural transition orbitals for complex two-component excited state calculations

While the natural transition orbital (NTO) method has allowed electronic excitations from time-dependent Hartree-Fock and density functional theory to be viewed in a traditional orbital picture, the extension to multicomponent molecular orbitals such as those used in relativistic two-component methods or generalized Hartree-Fock (GHF) or generalized Kohn-Sham (GKS) is less straightforward due to mixing of spin-components and the inherent inclusion of spin-flip transitions in time-dependent GHF/GKS. An extension of single-component NTOs to the two-component framework is presented, in addition to a brief discussion of the practical aspects of visualizing two-component complex orbitals. Unlike the single-component analog, the method explicitly describes the spin and frequently obtains solutions with several significant orbital pairs. The method is presented using calculations on a mercury atom and a CrO₂Cl₂ complex.     

Symmetry and equivalence restrictions in electronic structure calculations

A simple method for obtaining MCSCF orbitals and CI natural orbitals adapted to degenerate point groups, with full symmetry and equivalence restrictions, is described. Among several advantages accruing from this method are the ability to perform atomic SCF calculations on states for which the SCF energy expression cannot be written in terms of Coulomb and exchange integrals over real orbitals, and the generation of symmetry-adapted atomic natural orbitals for use in a recently proposed method for basis set contraction.