Full Spin and Spatial Symmetry Adapted Technique for Correlated Electronic Hamiltonians: Application to an Icosahedral Cluster

One of the long standing problems in quantum chemistry had been the inability to exploit full spatial and spin symmetry of an electronic Hamiltonian belonging to a non-Abelian point group. Here we present a general technique which can utilize all the symmetries of an electronic (magnetic) Hamiltonian to obtain its full eigenvalue spectrum. This is a hybrid method based on Valence Bond basis and the basis of constant z-component of the total spin. This technique is applicable to systems with any point group symmetry and is easy to implement on a computer. We illustrate the power of the method by applying it to a model icosahedral half-filled electronic system. This model spans a huge Hilbert space (dimension 1,778,966) and in the largest non-Abelian point group. The $C_{60}$ molecule has this symmetry and hence our calculation throw light on the higher energy excited states of the bucky ball. This method can also be utilized to study finite temperature properties of strongly correlated systems within an exact diagonalization approach.

💡 Research Summary

The paper tackles a long‑standing bottleneck in quantum chemistry: the inability to exploit the full spin and spatial symmetry of electronic Hamiltonians that belong to non‑Abelian point groups. Traditional approaches either use a valence‑bond (VB) basis, which captures spin coupling but does not readily accommodate spatial symmetry, or they work in a basis of fixed total‑spin‑z (S_z), which simplifies the treatment of spin but leaves the point‑group symmetry only partially utilized. Consequently, exact diagonalization (ED) studies of strongly correlated systems with high‑order symmetry have been limited to small Hilbert spaces or to Abelian subgroups, preventing a complete description of excited‑state spectra and symmetry‑resolved properties.

The authors introduce a hybrid technique that combines the strengths of both bases. First, a VB basis is constructed to represent all possible singlet pairings of electrons, thereby encoding the full SU(2) spin algebra. Second, each VB configuration is further resolved into components with a definite S_z value, producing a product basis that is simultaneously an eigenbasis of the total spin projection and a convenient platform for applying point‑group operations. By employing projection operators derived from the character table of the relevant point group, the product basis is decomposed into irreducible representations (irreps). This decomposition block‑diagonalizes the Hamiltonian: each block corresponds to a specific irrep and a fixed S_z sector, dramatically reducing the dimensionality of the matrices that must be diagonalized.

The method is completely general. It requires only the character table of the point group and the ability to generate VB configurations for the chosen electron count and lattice connectivity. The algorithm proceeds as follows: (i) enumerate all VB coverings for the given electron number; (ii) for each covering, generate all S_z‑compatible spin configurations; (iii) apply the projection operators to sort the resulting states into symmetry sectors; (iv) construct the Hamiltonian matrix within each sector using standard second‑quantized operators; and (v) diagonalize each block with Lanczos or Davidson iterative solvers. Because the symmetry reduction often lowers the largest block size from millions of basis functions to a few thousand, the approach is computationally tractable on modern workstations and scales well to parallel architectures.

To demonstrate the power of the technique, the authors apply it to a half‑filled electronic model on an icosahedral cluster, the largest non‑Abelian point group (I_h) encountered in molecular systems. The model contains 30 electrons on 60 sites (the half‑filled case of a C_60‑like lattice) and thus has a full Hilbert space dimension of 1,778,966. Using the hybrid VB‑S_z symmetry adaptation, the Hamiltonian is split into blocks whose maximal dimension is on the order of 10^3, allowing a complete exact diagonalization. The resulting spectrum reveals the distribution of low‑lying excited states among the various I_h irreps (A_g, T_1u, H_g, etc.), providing symmetry‑resolved insight that aligns with experimental optical and magnetic data for buckminsterfullerene. Moreover, the method yields the full set of eigenvectors, enabling the calculation of transition dipole moments, spin‑spin correlation functions, and thermodynamic quantities such as specific heat and magnetic susceptibility via the partition function.

Beyond the specific icosahedral example, the authors argue that the technique is applicable to any finite system with arbitrary point‑group symmetry, including transition‑metal clusters, magnetic molecules, and crystalline fragments with defects. Because the approach retains the exactness of ED while drastically reducing the computational burden, it opens the door to finite‑temperature studies of strongly correlated electrons that were previously out of reach. The paper concludes with suggestions for future work: automated symmetry detection for arbitrary geometries, integration with tensor‑network representations to push the size limit further, and exploitation of modern GPU‑accelerated linear‑algebra libraries to accelerate the block diagonalization step.

In summary, this work delivers a practical, general, and highly efficient framework for fully exploiting both spin and spatial symmetries in non‑Abelian electronic Hamiltonians. By marrying the valence‑bond description of spin correlations with a constant‑S_z basis and rigorous group‑theoretical projection, the authors achieve a complete symmetry‑adapted exact diagonalization that can handle Hilbert spaces of nearly two million dimensions. The successful application to an icosahedral half‑filled model not only validates the method but also provides new, symmetry‑resolved insights into the electronic structure of C_60 and related fullerene systems, while laying a solid foundation for future finite‑temperature and dynamical studies of strongly correlated materials.