Chen Zhixing

Rotation procedure in intrinsic reaction coordinate calculations

A procedure for calculating the distance between two points on an intrinsic reaction coordinate (IRC) obtained from two separate runs is proposed. The procedure mainly involves the rotation of the geometrical configuration of one point in order to obtain a configuration not rotated relative to the other point.

Quantum chemical study on acidity of silica-alumina catalysts

Abstract Acidities of silica-alumina systems are studied by the CNDO/2-FA method. A hypothetical atom is suggested for saturating the covalence of the outermost oxygen atoms of the models. It turns out that the hydrogen charge Q H can serve as a rough measure of Bronsted acidity of silica-alumina systems, whereas the LUMO energy as a measure of Lewis acidity can only be used for comparing the generally accepted Lewis sites of the same kind of atom in the same valence state and with the same coordinate number. Using the Q H data, the strong Bronsted site is assigned to the structure AlOHSi, and the moderate one is assigned to the SiOH in connection with a four-coordinate aluminum. The Q H data do not support the view that Si-OH in connection with a three-coordinate aluminum or AlOH is the strong Bronsted site.

An empirical scheme for evaluating in situ bond energy from MNDO data

The in situ bond energy is evaluated from the resonance integral contribution to energy with a correction term of monatomic energy obtained from MNDO calculation. The sum of the in situ bond energies in a molecule is expected to be equal to its atomization energy. Root-mean-square error of heat of formation calculated from in situ bond energy for seventy nine molecules containing hydrogen, carbon, nitrogen, oxygen, and fluorine amounts to 5.5 kcal/mol. Correlations of in situ bond energy with contributional bond energy, bond dissociation energy, isolated stretching frequency, and bond length are performed and discussed.

On methods for converging self-consistent-field iterations

Abstract A method using linear combinations of successive eigenvectors based on energy minimization is presented for converging self-consistent-field iterations. It is applied to a number of divergent or poorly convergent examples in semi-empirical CNDO, INDO, and ab initio STO-3G calculations. The CNDO/2 results are compared with those obtained by the level-shifting and damping techniques.

A further study on the CNDO/2 structures of water dimer

The structure of water dimer has been further studied by the CNDO/2 method. Some results have been found to be different from those in Thiels paper.

Effect of conjugation beyond multiple bonds: Part III. The nature of carbon—carbon bond length

Abstract The effect of the change in the form of a carbon atom (tetrahedral, trigonal or linear) on the carbon—carbon bond length is studied by a statistical method. The interference of conjugative effects on bond length is eliminated by regression with the π-resonance integral contribution to energy. The charge or inductive effect is found to be insignificant in the regression. The shortening of the CC bond length in the change from a tetrahedral to a trigonal carbon atom is found to be 0.007 A, while the change from a trigonal form to a linear form shortens the CC bond by 0.0395 A. It is thus concluded that the change in form of the carbon atoms is not a major factor in shortening the middle CC bond of butadiene, as compared with that of ethane.

On some problems in the theorems of alternant hydrocarbons

A number of theorems of alternant hydrocarbons presented by Dewar are proved to be wrong. A statement of the related theorem instead of Dewars is presented.

Partitioning of MNDO energy and description of bonding

Abstract The MNDO energy is partitioned into monatomic and diatomic contributions. A term E R , the resonance integral contribution, is separated out from the diatomic contribution E AB . A nonclassical term, E NC , as the sum of E R and the exchange integral contribution E K , as also separated out. Fairly good correlations of E AB , E R , and E NC with bond length show that they can serve as a measure of bonding. E R gives the best correlation, and E AB the worst, showing that E R is the best measure of covalent bonding, and E NC the second. E AB seems to be an overall measure of bond strength, but this is not true. The reason is that the monatomic contributions to energy also take a share in bond strength. Bond energies are also correlated with these energy contributions. The correlations are not very good, because bond energy is defined by some arbitrary postulates and does not serve as a rational measure of bond strength.

Effect of conjugation beyond multiple bonds: Part II. On the relative stability of various conformations of carbonium ions☆

Abstract MNDO geometry optimization indicates that the coplanar forms of the allylcarbinyl cation are more stable than the corresponding perpendicular forms. An explanation is given from the viewpoint of hyperconjugation. It can then be concluded that for cations of the type XCH 2 CH 2 + the adjacent double bond in X will strengthen the hyperconjugation in the coplanar conformation, thus lowering the energy to a greater extent.

A comparison of various MO methods for predicting regioselectivity in Diels-Alder reactions

Various semiempirical MO methods for predicting regioselectivity in Diels-Alder reactions are compared in order to examine their reliability. They are the frontier molecular orbital method using CNDO/2, HMO and MNDO molecular orbitals and the HMO approach on pericyclic reactions presented by Tang.