Reduced density matrix functional theory (RDMFT) provides a rigorous framework for treating strong electron correlation, yet its application to periodic systems has long been hindered by two fundamental challenges: the lack of transferable approximate functionals and the poor efficiency of orbital-occupation optimization. Here we show that these two longstanding obstacles can be overcome simultaneously. We develop a periodic implementation of RDMFT based on a coupled optimization framework that enables the efficient simultaneous optimization of natural orbitals and occupation numbers under periodic boundary conditions. Using this implementation, we demonstrate that physically motivated short-range screening transforms the Power functional into a transferable functional applicable to both molecules and periodic solids. Remarkably, short-range screening is found to play a dual role: besides substantially improving energetic accuracy, it fundamentally reshapes the optimization landscape, producing robust optimization step sizes and dramatically accelerating convergence. The coupled optimization framework consistently requires substantially fewer optimization iterations than conventional decoupled optimization for periodic calculations, while the screened $ω$P22 functional outperforms semilocal and hybrid density functionals in predicting representative surface reaction barriers. These results establish a computationally efficient periodic RDMFT framework and identify short-range screening as a promising design principle for developing transferable one-body reduced-density-matrix functionals.
Reduced density matrix functional theory (RDMFT) provides a rigorous framework for treating strong electron correlation, yet its application to periodic systems has long been hindered by two fundamental challenges: the lack of transferable approximate functionals and the poor efficiency of orbital-occupation optimization. Here we show that these two longstanding obstacles can be overcome simultaneously. We develop a periodic implementation of RDMFT based on a coupled optimization framework that enables the efficient simultaneous optimization of natural orbitals and occupation numbers under periodic boundary conditions. Using this implementation, we demonstrate that physically motivated short-range screening transforms the Power functional into a transferable functional applicable to both molecules and periodic solids. Remarkably, short-range screening is found to play a dual role: besides substantially improving energetic accuracy, it fundamentally reshapes the optimization landscape, producing robust optimization step sizes and dramatically accelerating convergence. The coupled optimization framework consistently requires substantially fewer optimization iterations than conventional decoupled optimization for periodic calculations, while the screened ωP22 functional outperforms semilocal and hybrid density functionals in predicting representative surface reaction barriers. These results establish a computationally efficient periodic RDMFT framework and identify short-range screening as a promising design principle for developing transferable one-body reduced-density-matrix functionals.
Reduced density matrix functional theory (RDMFT) offers a route beyond Kohn-Sham density functional theory for strongly correlated systems, yet practical calculations for periodic solids are still out of reach. We formulate RDMFT for extended systems in a basis-independent way and present a planewave implementation using iterative minimization for periodic solids, evaluating nonlocal exchange-correlation functionals through the existing adaptive compressed exchange machinery. Natural occupations are optimized under N-representability constraints with a spectral projected gradient (SPG) method or an first-order explicit-by-implicit (EBI) map, while natural orbitals are updated by Riemannian optimization on the complex Stiefel manifolds. Benchmarks on typical systems of \ce{H2}, silicon, and sodium with the Hartree-Fock functional show that SPG reproduces converged hybrid references, whereas EBI can stall when occupations approach 0 or 1. With the power and Müller functionals, SPG yields lower energies and more stable convergence than EBI. Applications to fractionally charged \ce{LiH}, dissociating \ce{H2} and \ce{N2} molecules, and equation of state of silicon show that the algorithm presented in this implementation is reliable and robust.
From drug design to material science, computational chemistry has become a valuable tool for many fields of applications in the past decades. Quantum chemistry in particular, can provide accurate descriptions of molecules and solids by relying on principles of quantum mechanics with the Schrödinger equation at its core. Unfortunately this equation becomes prohibitively expensive to solve for systems made of more than a few electrons and one has to rely on other approaches. Among the most popular is Density Functional Theory, relatively inexpensive and exact in principle. One has, however, to introduce approximations in practice, deteriorating its accuracy for challenging, so-called strongly correlated, systems. On the contrary, other methods like CASSCF remain reliable for these systems, but although cheaper than the solution of the Schrödinger equation, remain too expensive for extended systems. There is therefore room for methods in between, that could handle large correlated systems with reasonable accuracy. One such method is the One Body Reduced Density Matrix (1-RDM) Functional Theory (RDMFT), based on the fact that an electronic system can, in principle exactly, be described by its 1-RDM only. Although its theory dates back to the 1970s, computational hindrances have prevented this method from shining so far. Indeed, computations in RDMFT rely on a constrained optimization of the eigenvalues and eigenvectors of the 1-RDM, that can be reformulated as an unconstrained optimization by parameterizing the latter but in both cases suffers from convergence problems. In this thesis, the role of this parametrization is analyzed and motivates a now second order optimization algorithm, which improves the convergence speed by two orders of magnitude w.r.t. the state of the art algorithm. This understanding of the problem also inspires an alternative approach with the 1-RDM as the optimization variable. This however requires the development of new approximations and a first investigation in that direction, based on Machine Learning is performed. Lastly, a generalization of the method to excited states is attempted, highlighting some key computational challenges.
Density-Functional Theory (DFT), although widely used and very successful in the calculation of several observables, fails to correctly describe strongly correlated materials. In the first part of this work we, therefore, introduce reduced-density-matrix-functional theory (RDMFT) which is one possible way to treat electron correlation beyond DFT. Within this theory the one-body reduced density matrix (1-RDM) is used as the basic variable. Our main interest is the calculation of the fundamental gap which proves very problematic within DFT. In order to calculate the fundamental gap we generalize RDMFT to fractional particle numbers M by describing the system as an ensemble of an N and an N+1 particle system (with N≤ M≤N+1). For each fixed particle number, M, the total energy is minimized with respect to the natural orbitals and their occupation numbers. This leads to the total energy as a function of M. The derivative of this function with respect to the particle number has a discontinuity at integer particle number which is identical to the gap. In addition, we investigate the necessary and sufficient conditions for the 1-RDM of a system with fractional particle number to be N-representable. Numerical results are presented for alkali atoms, small molecules, and periodic systems. Another problem within DFT is the description of non-relativistic many- electron systems in the presence of magnetic fields. It requires the paramagnetic current density and the spin magnetization to be used as basic variables besides the electron density. However, electron-gas-based functionals of current-spin-density-functional Theory (CSDFT) exhibit derivative discontinuities as a function of the magnetic field whenever a new Landau level is occupied, which makes them difficult to use in practice. Since the appearance of Landau levels is, intrinsically, an orbital effect it is appealing to use orbital-dependent functionals. We have developed a CSDFT version of the optimized effective potential (OEP) method which allows for the use of explicitly orbital-dependent functionals in the context of CSDFT. We present the derivation of corresponding equations and show results for a quantum dot in external magnetic fields in the second part of this thesis.
Recently, an approximate theoretical framework was introduced, called local reduced density matrix functional theory (local-RDMFT), where functionals of the one-body reduced density matrix (1-RDM) are minimized under the additional condition that the optimal orbitals satisfy a single electron Schrödinger equation with a local potential. In the present work, we focus on the character of these optimal orbitals. In particular, we compare orbitals obtained by local-RDMFT with those obtained with the full minimization (without the extra condition) by contrasting them against the exact NOs and orbitals from a density functional calculation using the local density approximation (LDA). We find that the orbitals from local-RMDFT are very close to LDA orbitals, contrary to those of the full minimization that resemble the exact NOs. Since local RDMFT preserves the good quality of the description of strong static correlation, this finding opens the way to a mixed density/density matrix scheme, where Kohn-Sham orbitals obtain fractional occupations from a minimization of the occupation numbers using 1-RDM functionals. This will allow for a description of strong correlation at a cost only minimally higher than a density functional calculation.
Recently, an approximate theoretical framework was introduced, called local reduced density matrix functional theory (local-RDMFT), where functionals of the one-body reduced density matrix (1-RDM) are minimized under the additional condition that the optimal orbitals satisfy a single electron Schrödinger equation with a local potential. In the present work, we focus on the character of these optimal orbitals. In particular, we compare orbitals obtained by local-RDMFT with those obtained with the full minimization (without the extra condition) by contrasting them against the exact NOs and orbitals from a density functional calculation using the local density approximation (LDA). We find that the orbitals from local-RMDFT are very close to LDA orbitals, contrary to those of the full minimization that resemble the exact NOs. Since local RDMFT preserves the good quality of the description of strong static correlation, this finding opens the way to a mixed density/density matrix scheme, where Kohn-Sham orbitals obtain fractional occupations from a minimization of the occupation numbers using 1-RDM functionals. This will allow for a description of strong correlation at a cost only minimally higher than a density functional calculation.
The last several decades have seen significant advances in the theoretical modeling of materials within the fields of solid-state physics and materials science, but many methods commonly applied to this problem struggle to capture strong electron correlation accurately. Recent widespread interest in quantum materials---where strong correlation plays a crucial role in the quantum effects governing their behavior---further highlights the need for theoretical methods capable of rigorously treating such correlation. Here, we present a periodic generalization of variational two-electron reduced density matrix (2-RDM) theory, a bootstrapping-type method that minimizes the ground-state energy as a functional of the 2-RDM without relying on the wave function. The 2-RDM is computed directly by semidefinite programming with N-representability conditions, ensuring accurate treatment of strongly correlated electronic systems. By exploiting translational symmetry, we significantly reduce computational scaling, enabling applications to realistic materials-scale systems. Additionally, we introduce an alternative to conventional energy band structures: natural-orbital occupation-number bands, which, being independent of mean-field assumptions, offer deeper insights into electron correlation effects. We demonstrate the effectiveness of this approach by applying the theory to hydrogen chains, molybdenum disulfide, and nickel oxide, showing that natural-orbital occupation bands correctly capture electronic character in regimes where density functional theory fails. This work represents a major step toward accurately describing the electronic structure of quantum materials using reduced density matrices rather than wave functions.