Multipole splats for optimized and inverted effective potentials

Abstract:Approximate single-particle descriptions of correlated matter remain the principal source of actionable predictions in quantum science. In modern density functional theory (DFT), hybrid effective Hamiltonians are constructed to reproduce equilibrium densities and forces. However, their nonlocal and orbital-dependent potentials leave systematic errors in response properties and real-time dynamics. Local optimized effective potentials (OEP) or inverted Kohn-Sham (IKS) potentials would remove this limitation, but the numerical fragility in finite orbital bases has long prevented their wide adoption. We introduce multipole splats, a class of trial potentials that carry the correct asymptotic decay by construction. By showing that OEP and IKS map to variational and supervised variants of Hamiltonian learning, we recast both problems as stable nonlinear optimization formulated directly in standard orbital basis sets and applicable to any hybrid functional approximation. We show improvements to three canonical DFT failure modes. Firstly, by comparing approximate and near-exact inverted effective correlation potentials, we isolate the spatial signatures of self-interaction, delocalization, and static correlation errors. Secondly, we recover the Rydberg series without empirical asymptotic corrections, demonstrating access to precise spectral properties. Finally, we compute exact-exchange potentials for larger $\pi$-conjugated systems, establishing scaling capacity. Multipole splats extract new first-principles insights from established approximations and provide capacity for robust dataset generation for downstream processing and learning.

Chemical Physics (physics.chem-ph); Computational Physics (physics.comp-ph); Quantum Physics (quant-ph)

arXiv:2609.09280 [physics.chem-ph]

(or arXiv:2609.09280v1 [physics.chem-ph] for this version)

https://doi.org/10.48550/arXiv.2609.09280 Focus to learn more arXiv-issued DOI via DataCite (pending registration)

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.