Answers are grounded in this paper and cite their sources.
An algorithm for the calculation of energy gradients has been implemented into the linear combination of atomic orbitals-Hartree–Fock–Slater (LCAO-HFS) program due to Baerends et al. [Chem. Phys. 2, 41 (1973)], and tested in connection with geometry optimizations of transition-metal complexes, organic molecules and compounds of main group elements. The algorithm is based on a self-consistent nonlocal density-functional method (NL-SCF) combining the nonlocal exchange correction by Becke [Phys. Rev. A 38, 3098 (1988)] with the nonlocal correlation correction by Perdew [Phys. Rev. B 33, 8822 (1986)]. It is shown that nonlocal corrections to the local-density approximation (LDA) improves the calculated metal-ligand bond distances considerably. Full optimizations of MnO−4, Cr(CO)6, Mo(CO)6, Fe(CO)5, Ni(CO)4, HMn(CO)5, HCo(CO)4, Fe(C5H5)2, and C5H5NiNO revealed that the LDA underestimates metal-ligand bond distances by an average of 0.05 Å, whereas nonlocal corrections, on the average, reduce the error to 0.01 Å. Nonlocal corrections are not as important for organic molecules and compounds of main group elements where LDA already affords bond distances in good agreement with experiment. Bond angles are, in general, not influenced much by nonlocal corrections. The nonlocal corrections increase the computational time involved by a factor of 4 on a scalar machine and a factor of 2 on a vector machine.
No takes yet. Share an insight, caveat, or question.
Fan et al. (1991) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: