Bhavnesh Jangid

Iron–sulfur clusters play central roles in biological photosynthesis, respiration, and catalysis. The [Fe4S4(SCH3)4]2– cluster possesses a dense manifold of low-lying electronic states arising from strong magnetic interactions among the Fe centers, making its electronic structure challenging to describe theoretically. Previous multireference studies have generally optimized excited states individually while neglecting the interactions among low-lying states. In this work, we investigate the low-lying electronic states of [Fe4S4(SCH3)4]2– using localized active space state interaction singles (LASSIS) combined with pair-density functional theory (PDFT). By treating all states within a common model-space Hamiltonian, LASSIS generates a single set of orthogonal low-lying states. The resulting spin-state energies organize into distinct manifolds. The same overall pattern is obtained for active spaces ranging from 22 electrons in 20 orbitals to 54 electrons in 36 orbitals. To enable these calculations, we also implement GPU acceleration for the most computationally demanding components of the LASSIS workflow.

Lanthanide molecular magnets are promising candidates for quantum information technologies, but their discovery is limited by the computational cost of accurately predicting their electronic and magnetic properties. Multireference methods provide the required accuracy, yet their application at scale has been hindered by the need to manually select the model space (the states considered in the spin–orbit treatment). We present an automated multireference workflow for single-lanthanide complexes, based on complete active space wave functions followed by state-interaction spin–orbit (SISO) coupling. Its central feature is an iterative protocol for constructing the SISO model space, which enables reliable and reproducible calculations across diverse coordination environments. Benchmark calculations on lanthanide hexachlorides show that including spin–orbit-free states up to approximately 5 eV above the ground state reproduces experimental ligand-field splittings with mean absolute errors of 100–200 cm–1 across the lanthanide series. We further examine how the model-space size affects magnetic anisotropy and g-tensors in representative dysprosium- and erbium-based single-ion magnets and in the [Ln(DOTA)(H2O)]− family. Finally, we apply the protocol to 299 lanthanide complexes to generate a consistent data set of ligand-field splittings, g-tensors, magnetic axes, and approximate barriers to magnetization reversal. This work establishes a practical framework for high-throughput multireference calculations and data generation for lanthanide molecular magnets.

While multiconfigurational approaches have long been relegated to expert practitioners working on a case-by-case basis, recent developments have increasingly made these methods more routine and applicable to broader sets of systems. This review outlines the state-of-the-art in multiconfigurational approaches, with an emphasis on moving from delicate hand-selected pathways through configuration space toward more robust and efficient approaches to treating a host of challenging chemical systems accurately. First, we overview recent work in automated active-space selection, which has enabled increasingly large-scale applications of multireference methods to modeling vertical excitations and reactivity. Second, we highlight the increasingly efficient methods for recovering correlation energy beyond the active space, as headlined by extensions of pair-density functional theory and its role in accurate and efficient treatment of excited-state dynamics and its utilization to train machine-learned potentials. Finally, we highlight recent efforts to treat extended systems that until recently have lied beyond the traditional limits of active-space methods, giving center stage to product-form wave functions of the localized active space family of methods that allow for the computation of multiconfigurational band structures. These recent advancements point to a broader use of multireference approaches for high-impact chemical and materials science applications.

One of the primary challenges in quantum chemistry is the accurate modeling of strong electron correlation. While multireference methods effectively capture such correlation, their steep scaling with system size prohibits their application to large molecules and extended materials. Quantum embedding offers a promising solution by partitioning complex systems into manageable subsystems. In this Review, we highlight recent advances in multireference density matrix embedding and localized active space self-consistent field approaches for complex molecules and extended materials. We discuss both classical implementations and the emerging potential of these methods on quantum computers. By extending classical embedding concepts to the quantum landscape, these algorithms have the potential to expand the reach of multireference methods in quantum chemistry and materials.

The number of materials that "bridge the gap" between single molecules and extended solids, such as metal-organic frameworks and organic semiconductors, has been increasing. Consequently, there is a growing need for modeling approaches that effectively integrate the real-space molecular perspective employed by computational chemists and the reciprocal-space dispersive perspective employed by computational physicists. Here, we propose the localized active space (LAS) approach as a promising method to successfully bridge this gap. The LAS approach extends the active space concept from multiconfigurational methods such as complete active space self-consistent field theory to multiple molecular fragments via a product-form wave function ansatz. Here, we apply this method to solid state phenomena by treating each unit cell as a fragment with different sets of local quantum numbers (e.g., charge and excitation number). State interaction between these LAS states (LASSI) thus provides a comprehensive basis for the study of charge and energy transfer, meeting and surpassing the capabilities of single-reference fragmentation approaches such as constrained density functional theory (cDFT). Most centrally, we show how combining this LASSI approach with multiconfigurational pair-density functional theory (MC-PDFT) provides an elegant and efficient method to compute band structures that capture multiconfigurational character. We apply the LASSI band structure approach to the computation of band gaps in stretched hydrogen chain, polyacetylene, and bulk nickel oxide (NiO), finding good or excellent quantitative agreement with reference values in all cases. Additionally, we use the LAS basis in one-dimensional model systems to demonstrate its ability to treat difficult solid-state phenomena such as exciton transfer and excitation at p-n junctions.

We present the inclusion of spin-orbit coupling (SOC) effects in linearized pair-density functional theory (L-PDFT), which is a multi-state extension of multiconfiguration pair-density functional theory (MC-PDFT). Both 1-electron and 2-electron SOC integrals are computed using Breit-Pauli and Douglas-Kroll-Hess Hamiltonians in the atomic mean-field approximation. SO-L-PDFT removes the unphysical the J-symmetry breaking observed in MC-PDFT. The accuracy SO-L-PDFT is validated by calculations of zero-field splittings, fine-structure excitation energies, and low-energy excited-state spectra for a diverse group of atoms and molecules spanning the whole range of the periodic table, including atoms of groups 3, 11, 13-17, the Ce3+ and U5+ ions, group 16 monohydrides, group 17 monoxides, lanthanide hexachlorides, actinyl ions and tricarbonatoactinyl complexes. We also compare the results to new spin-orbit-inclusive calculations by single-state and multi-state multireference perturbation theory

While multiconfigurational approaches have long been relegated to expert practitioners working on a case-by-case basis, recent developments have increasingly made these methods more routine and applicable to broader sets of systems. This perspective outlines the state-of-the-art in multiconfigurational approaches, with an emphasis on moving from delicate hand-selected pathways through configuration space towards more robust and efficient approaches to treating a host of challenging chemical systems accurately. First, we overview recent work in automated active-space selection, which has enabled increasingly large-scale applications of multireference methods to modeling vertical excitations and reactivity. Second, we highlight the increasingly efficient methods for recovering correlation energy beyond the active space, as headlined by extensions of pair-density functional theory and its role in accurate and efficient treatment of excited-state dynamics and its utilization to train machine-learned potentials. Finally, we highlight recent efforts to treat extended systems that until recently have lied beyond the traditional limits of active-space methods, giving center stage to product-form wave functions of the localized active space family of methods that allow for the computation of multiconfigurational band structures. These recent advancements point to a broader use of multireference approaches for high-impact chemical and materials science applications.

Efficient and accurate computation of molecular Auger electron spectra for larger systems is limited by the rapid increase in the number of doubly ionized final states as the system size grows. In this work, we benchmark the application of multiconfiguration pair-density functional theory with a restricted active space (RAS) reference wave function for computing the carbon K-edge decay spectra of 20 organic molecules. Decay rates are computed within the one-center approximation. We evaluate the performance of different basis sets and on-top functionals and find that multiconfiguration pair-density functional theory achieves accuracy comparable to RAS followed by second-order perturbation theory, but at significantly lower computational cost.