José Sánchez-Marín

Self‐consistent intermediate Hamiltonians: A coupled cluster type formulation of the singles and doubles configuration interaction matrix dressing

This paper presents a new self‐consistent dressing of a singles and doubles configuration interaction matrix which insures size‐consistency, separability into closed‐shell subsystems if localized molecular orbitals (MOs) are used, and which includes all fourth order corrections. This method yields, among several schemes, a reformulation of the coupled cluster method, including fully the cluster operators of single and double excitations, and partially those of the triples (Bartlett’s algorithm named CCSDT‐1a). Further improvement can be easily included by adding exclusion principle violating corrections. Since it leads to a matrix diagonalization, the method behaves correctly in case of near degeneracies between the reference determinant and some doubles. Due to its flexibility this formulation offers the possibility of consistent combination with less expensive treatments for the study of very large systems.

Theoretical study of the low‐lying states of trans‐1,3‐butadiene

We present extensive ab initio calculations on the low‐lying electronic states of trans‐1,3‐butadiene within the multireference configuration interaction (MRCI) framework by selecting the configurations with a perturbative criterion. The X 1Ag ground state and 1 3Bu, 1 3Ag, 2 1Ag, and 1 1Bu valence excited states have been calculated at a fixed geometry. The results obtained are in good agreement with previous experimental and calculated values, and could help to understand polyene spectroscopy, photochemistry, and photophysics. The advantages of a MRCI method where the most important contributions to the total MRCI wave function, perturbatively selected, are treated variationally, and the remaining terms are evaluated by means of a perturbational approach, are also discussed. Furthermore, a criterion in order to build a correlation‐consistent configuration interaction space is stated and, therefore, a reliable approximation to achieve accurate energy differences is obtained. Several monoelectronic molecu...

Universal model for the calculation of all organic solvent–water partition coefficients

Abstract We present the basis for building a universal organic solvation model to calculate solubility in any organic solvent and in water, as well as the organic solvent–water partition coefficient ( P ). Log P values are of the same order of magnitude as reference calculations but for a few cases which are discussed. Normalized log P contributions are sensitive to the rest of the atoms. When comparing porphin with phthalocyanine, the latter results in an amphipathic molecule. For C 70 , the contribution of a–e carbons to log P correlates with the distances from the nearest pentagon. The method has been also applied to benzobisthiazole oligomers and phenyl alcohols.

Applicability of size-consistent self-consistent configuration interaction to excited states

Abstract The size-consistent self-consistent configuration interaction (SC) 2 CI method is applied to the calculation of excitation energies. This method diagonalizes a dressed SDCI matrix built for the ground state but it is shown that the extra eigenvalues are not phony solutions but represent the excited roots of the system. Numerical tests are encouraging. The method is further improved by using the (SC) 2 MRCI scheme where the desired excited determinant is added to the ground state determinant as a generator of the CI space. Excellent results are obtained with this method for single and double replacement excitations in CH 2 .

Interacting induced dipoles polarization model for molecular polarizabilities. Reference molecules, amino acids and model peptides

Abstract We outline a method for the calculation of molecular dipole ( μ ) and quadrupole ( Θ = ) moments and dipole–dipole polarizabilities ( α = ) which we have successfully applied to a series of reference molecules, amino acids and model peptides. The results for μ are in line with CPHF reference calculations. In particular, the calculated positive value of CO is in agreement with both experimental and CI calculations. The computation of ( α = ) has been performed by the interacting induced dipoles polarization model that calculates tensor effective anisotropic point polarizabilities (method of Applequist et al.). The POLAR program cannot be used as a black box. Some tests should be performed when a new molecule is calculated. The POLAR program was designed for large molecules. Although in some large molecules the POLAR program has been successfully applied to predict trends, the test with small molecules shows that we have to touch up the approximations along the formulation. The results for ( α = ) of reference molecules are shown better for the POLAR program than for the PAPID algorithm. In the former case, the POLAR-AP model is revealed superior to the POLAR-IP method. The results for the amino acids and model peptides show that, as a rule, the PAPID program produces the better results, while the POLAR program produces acceptable results and converges in all the cases. In general, POLAR-IP underestimates the molecular polarizabilities and anisotropies while POLAR-AP overestimates these properties.

Size-consistent single-reference methods for electronic correlation: a unified formulation through intermediate hamiltonian theory

AbstractUsing the intermediate hamiltonian theory as a unique conceptual frame and the technique of CI matrix dressing, a wide series of single-reference methods for the treatment of the ground state correlation are reviewed, compared, and sometimes improved. These methods range from independent excitation approximation (the very next step beyond MP2) to coupled cluster, going through the so-called electron pair approximations and the (SC)2CI formalism. A hierarchy of these methods can be established according to two criteria:1.The physical effects incorporated in the model space, the choice of which is flexible.2.The quality of the evaluation of the coefficients of the external space determinants. This evaluation, which remains based on a single reference expansion of the wave function, may simply ensure the size consistency or incorporate the linked contributions from the outer space. These formulations in terms of diagonalizations of dressed CI matrices avoid convergence problems, but their main advantage is their flexibility, since they apply to multi-reference SDCI spaces as well as to SDCI spaces. The use of a common frame allows one to propose consistent combinations of methods of various costs for the treatment of various parts of the correlation energy.

Elementary presentation of self‐consistent intermediate Hamiltonians and proposal of two totally dressed singles and doubles configuration interaction methods

Intermediate Hamiltonians are effective Hamiltonians which are defined on an N‐dimensional model space but which only provide n<N exact eigenvalues and the projections of the corresponding eigenvectors onto the model space. For a single root research, the intermediate Hamiltonian may be obtained from the restriction of the Hamiltonian to the model space by an appropriate, uniquely defined dressing of the diagonal energies or of the first column. Approximate self‐consistent dressings may be proposed. The simplest perturbative form gives the same result as the original 2nd order intermediate Hamiltonian or the ‘‘shifted Bk’’ technique but it is of easier implementation. Self‐consistent inclusion of higher order exclusion principle violating corrections greatly improves the results, especially for nearly degenerate problems, as shown on several illustrative applications. Possible generalizations to enlarged or reduced model spaces are discussed.

On the formation of methane and hydrogen during cracking of alkanes

Abstract A qualitative explanation of the low dehydrogenation/cracking ratio observed in reactions with long-chain n-alkanes catalyzed by superacids is attempted via calculation of the MINDO/3 reaction profiles for the complete dehydrogenation ⇆ intermolecular interconversion ⇆ cracking (D.I.C.) set of unimolecular processes that a protonated n-alkane can undergo. It is shown that the three simplest systems can suitably model all possible D.I.C. surfaces that can be found in a long-chain linear alkane. The importance of the primary protonation site in the distribution of products is discussed. Some experimental results from cracking of n-heptane over a set of acid zeolites are used, together with those of this theoretical study, to show that a recent controversy about the presence of methane and hydrogen as primary or secondary products in saturated alkane cracking catalyzed by zeolites can be settled.

Pair potential calculation of molecular associations: a vectorized version

Abstract The program AMYRVF is a vectorized and largely modified version of a previous program called AMYR for calculating molecular associations by means of Fragas pair-wise atom-atom potential. Three new minimization procedures have been implemented as well as other improvements such as the inclusion of new pair-wise dispersion energy terms with damping functions, and the calculation of topological indices. Benchmark tests have been carried out on an IBM 3090 150E VF; the timings for the new vector algorithms and for the standard scalar computations, as well as the dependence of the overall performance gain on the size of interacting systems are reported for the VS Fortran 2.4 compiler enhanced by the ESSL vectorized library. The new program vectorizes 97% of the computation.

Are most of the stationary points in a molecular association minima? Application of Fraga's potential to benzene-benzene

The importance of characterizing the stationary points of the intermolecular potential by means of Hessian eigenvalues is illustrated for the calculation of the benzene–benzene interaction using an atom‐to‐atom pair potential proposed by Fraga (FAAP). Two models, the standard one‐center‐per atom and another using three‐centers‐per atom due to Hunter and Sanders, are used to evaluate the electrostatic contributions and the results are compared. It is found in both cases that although using low‐gradient thresholds allows optimization procedures to avoid many stationary points that are not true minima computing time considerations makes the usual procedure of using high‐gradient thresholds [say, 10−2 kj/(mol Å)] as the most efficient. Moreover, this later procedure can be recommended because the actual minima can be characterized by means of Hessian eigenvalues even if these high‐gradient thresholds are used, and further decreasing of the convergence criterion does not imply significant modifications in the geometric parameters of the minima. The possible advantages of using the three‐centers‐per‐atom model for the calculation of molecular associations between delocalized systems are also discussed on the basis of the agreement of the benzene–benzene results with experimental and theoretical data taken from the literature.