Efficient Simulation\nof Inhomogeneously Correlated\nSystems Using Block Interaction Product States

The strength of the density matrix renormalization group\n(DMRG)\nin handling strongly correlated systems lies in its unbiased and simultaneous\ntreatment of identical sites that are both energetically degenerate\nand spatially similar, as typically encountered in physical models.\nHowever, this very feature becomes a drawback when DMRG is applied\nto quantum chemistry calculations for large, realistic correlated\nsystems. This is because entangled orbitals often span broad ranges\nin both energy and space, with their interactions being notably inhomogeneous.\nIn this study, we suggest addressing the strong intrafragment correlations\nand weak interfragment correlations separately, utilizing a large-scale\nmulticonfigurational calculation framework grounded in the block interaction\nproduct state formulation. The strong intrafragment correlation can\nbe encapsulated in several electronic states located on fragments,\nwhich are obtained by considering the entanglement between fragments\nand their environments. Moreover, we incorporate non-Abelian spin-SU(2)\nsymmetry in our work to target the desired states we interested with\nwell-defined particle number and spin, providing deeper insights into\nthe corresponding chemical processes. The described method has been\nexamined in various chemical systems and demonstrates high efficiency\nin addressing the inhomogeneous effects in strong correlation quantum\nchemistry.