Abstract:When the spins of molecular orbitals are allowed to be aligned with different directions in space rather than being aligned collinearly, the resulting noncollinear spin orbitals add extra flexibility to variational optimization of the orbitals, and solutions obtained with collinear spin orbitals may be unstable with respect to becoming noncollinear in the expanded variational space. The goal of the present work is to explore whether and in what way the molecular orbitals of the Kohn-Sham density functional theory become noncollinear when fully optimized for multi-reference molecules, transition states, and reaction paths. (We note that a noncollinear determinant has intermediate flexibility between a collinear determinant and a linear combination of many collinear determinants with completely independent coefficients. However, the Kohn-Sham method is defined to involve the variational optimization of a single determinant, and a noncollinear determinant represents the limit of complete optimization in the Kohn-Sham scheme.) We compare the results obtained with the noncollinear Kohn-Sham (NKS) scheme to those obtained with the widely used unrestricted Kohn-Sham (UKS) scheme for two types of multi-reference systems. For the dissociation of the MnO and NiO transition metal oxides, we find UKS fails to dissociate to the ground states of neutral atoms, while NKS dissociates to the correct limit and predicts potential energy curves that vary smoothly at intermediate bond lengths. This is due to the instability of UKS solutions at large bond distances. For barrier heights of O3, BeH2, and H4, NKS is shown to stabilize the multi-reference transition states by expanding the variational space. Although the errors vary because they are closely coupled with the capability of the employed exchange-correlation functionals in treating the multi-configurational states, these findings demonstrate that results with collinear spin orbitals should be further scrutinized, and future development of exchange-correlation functionals for multi-reference systems should incorporate the flexibilities of NKS.

Study on the Construction of Satisfactory Nonorthogonal Localized Molecular Orbitals

An Efficient Method for Constructing Nonorthogonal Localized Molecular Orbitals

A method for constructing the reasonable and most contracted localized orbitals with less computational efforts

Describing Chemical Reactivity with Frontier Molecular Orbitalets

Study On Designing Effective Organic Nonlinear (Nlo) Optical Materials

Noncollinear Spin States for Density Functional Calculations of Open-Shell and Multi-Configurational Systems: Dissociation of MnO and NiO and Barrier Heights of O3, BeH2, and H4

Natural localized molecular orbitals

Linear-scaling local natural orbital CCSD(T) approach for open-shell systems: algorithm, benchmarks, and large-scale applications

Orbital Minimization with Localization

Riemannian Trust Region Method for Minimization of Fourth Central Moment for Localized Molecular Orbitals

Global Versus Local Orthogonal Collocation in Simultaneous Approach

Toward linear scaling auxiliary field quantum Monte Carlo with local natural orbitals

An examination of the nature of localized molecular orbitals and their value in understanding various phenomena that occur in organic chemistry

ff ects on frontier orbitals of conjugated molecules using an absolutely localized molecular orbital based analysis †

Outstanding Improvement in Removing the Delocalization Error by Global Natural Orbital Functional

Selected non-orthogonal configuration interaction with compressed single and double excitations

Direct determination of optimal real-space orbitals for correlated electronic structure of molecules

Exploring the Potential of Natural Orbital Functionals

Theoretical Study on Neutral Molecules with Square Planar Tetracoordinate Oxygen O(B)4 Arrangements

Electron localization function and electron localizability indicator applied to study the bonding in the peroxynitrous acid HOONO

Advancing Natural Orbital Functional Calculations Through Deep Learning-Inspired Techniques for Large-Scale Strongly Correlated Electron Systems