Pseudopotentials for Orbital-Free DFT: Capturing Nonlocality and Correcting Functional Approximants

Valeria Rios-Vargas, Ezekiel Oyeniyi, Xuecheng Shao, Wala Fathelrahman Ibrahim Elsayed, Sunday Joseph Ogenyi, Alex Okello, Michele Pavanello · 2025

Read Pseudopotentials for Orbital-Free DFT: Capturing Nonlocality and Correcting Functional Approximants by Valeria Rios-Vargas, Ezekiel Oyeniyi, Xuecheng Shao, Wala Fathelrahman Ibrahim Elsayed, Sunday Joseph Ogenyi, Alex Okello, Michele Pavanello (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-19892.txt.

Introduction

Valeria Rios-Vargas and six co-authors take on a weak spot in orbital-free density functional theory. This method models electrons without orbitals, which is fast, but reliable pseudopotentials, the stand-in potentials for inner electrons, are hard to build, above all for transition metals.

The authors set out a theory for judging pseudization choices and propose building local pseudopotentials that match trusted Kohn-Sham ones through an optimized effective potential method. They produce four sets of local potentials and test their accuracy and transferability across the transition metals. The new sets improve clearly on older ones and give a firmer base for later method work.

Catalog record

CREATORValeria Rios-Vargas, Ezekiel Oyeniyi, Xuecheng Shao, Wala Fathelrahman Ibrahim Elsayed, Sunday Joseph Ogenyi, Alex Okello, Michele Pavanello
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE49 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-19892

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Pseudopotentials for Orbital-Free DFT: Capturing Nonlocality and Correcting Functional Approximants Valeria Rios-Vargas, Ezekiel Oyeniyi, Xuecheng Shao, Wala Fathelrahman Ibrahim Elsayed, Sunday Joseph Ogenyi, Alex Okello, Michele Pavanello Abstract Developing reliable pseudopotentials for orbital-free density functional theory (OF-DFT), especially for transition metals, remains a significant challenge. In this study, we provide a theoretical framework for analyzing pseudization strategies for OF-DFT calculations. From the analysis arises a proposed pseudization method which involves constructing local pseudopotentials by targeting existing Kohn–Sham DFT pseudopotentials through an optimized effective potential procedure. We produce four distinct sets of local pseudopotentials and evaluate their accuracy and transferability on the transition metal elements. Our results indicate a substantial improvement over previously available pseudopotentials. Although current OF-DFT functionals still only reach a qualitative accuracy for transition metals, our newly developed pseudopotentials provide a rigorous framework for further methodological advancements. I Introduction and Background Among all approaches to modeling materials at the atomic level, Density Functional Theory as formulated by Kohn and Sham 24 (KS-DFT) stands out for its impact. Its most widely used methods have so far been cited over 200,000 times 35 . KS-DFT’s implementations offer a good compromise between accuracy and computational cost. However, independently from the choice of the underlying approximations, it becomes too computationally expensive when large, realistically-sized systems (e.g., quantum dots, proteins and complex materials interfaces) are targeted. The reason for this is the unavoidable costs of building and diagonalizing the KS Hamiltonian matrix. In particular, diagonalizations carry a cost of \mathcal{O}(N^{3}) number of operations, where N is measure of system size. There are many approaches that tackle KS-DFT’s steep computational scaling 27 ; 7 ; 32 ; 49 . Among them, a promising alternative is orbital-free DFT (OF-DFT) 28 ; 45 . OF-DFT is based on the same theoretical footing as KS-DFT 24 . In particular, it borrows KS-DFT’s expression for the total electronic energy given as a functional of the electronic density: E[n]=T_{s}[n]+E_{Hxc}[n]+\int v_{ext}({\mathbf{r}})n({\mathbf{r}})d{\mathbf{r}} (1) where T_{s} , E_{Hxc} , and v_{ext}({\mathbf{r}}) are the noninteracting kinetic energy, the Hartree exchange-correlation (xc) energy, and the interaction with the external potential (typically the Coulomb electron-nuclear attraction potential), respectively. The noninteracting kinetic energy is given in terms of KS orbitals (single particle functions, hence the “s” subscript) as follows T_{s}[n]\equiv T_{s}[\{\phi_{i}[n]\}]=-\frac{1}{2}\sum_{i}n_{i}\langle\phi_{i}|\nabla^{2}|\phi_{i}\rangle, (2) where n_{i} are occupation numbers. The electron density is simply n({\mathbf{r}})=\sum_{i}n_{i}|\phi_{i}({\mathbf{r}})|^{2} (3) The Hartree functional is defined as the electronic Coulomb self-repulsion, E_{H}[n]=\frac{1}{2}\int\frac{n({\mathbf{r}})n({\mathbf{r}^{\prime}})}{|{\mathbf{r}}-{\mathbf{r}^{\prime}}|}d{\mathbf{r}}d{\mathbf{r}^{\prime}} and the xc functional, E_{xc}[n] is the energy term that is added to T_{s}[n] , E_{H}[n] and \int v_{ext}({\mathbf{r}})n({\mathbf{r}})d{\mathbf{r}} to recover the total electronic energy. Following from Eq. ( 3 ), we can use the KS orbitals as variational functions to minimize the total energy, leading to the canonical KS equations -\frac{1}{2}\nabla^{2}\phi_{i}({\mathbf{r}})+v_{s}({\mathbf{r}})\phi_{i}({\mathbf{r}})=\varepsilon_{i}\phi_{i}({\mathbf{r}}) (4) where v_{s}({\mathbf{r}})=v_{Hxc}[n]({\mathbf{r}})+v_{ext}({\mathbf{r}}) defining the Hxc potential as a functional derivative, v_{Hxc}[n]({\mathbf{r}})=\frac{\delta E_{Hxc}}{\delta n({\mathbf{r}})} . In OF-DFT, the noninteracting kinetic energy, T_{s} , is evaluated approximately by a pure density functional, which we call henceforth the kinetic energy density functional or KEDF. The accuracy of OF-DFT depends on the KEDF T_{s}[\rho] , with magnitude equal to the total energy, for highly inhomogeneous electron densities the devolopment of KEDFs is still a challenging task, however there are good functionals sunch as LMGP 30 ; 29 , Huang-Carter (HC) 19 , and revHC 41 which provides one of the most accurate descriptions of metallic systems currently available. Using KEDFs allows the energy minimization of the functional in Eq. ( 1 ) to be based solely on the electron density, completely bypassing the KS orbitals, and generally leading to algorithms of linear scaling complexity. 28 . One thorn on the side of OF-DFT has persisted. Namely the handling of the electron-ion interaction. While this seems a non-issue (the v_{ext}({\mathbf{r}}) is conceptually trivial), when approaching condensed phases the pseudopotential (PP) approximation is usually invoked 43 ; 44 ; 26 ; 2 ; 4 ; 31 ; 22 . Natural basis functions used to expand the KS orbitals for materials are plane waves, \chi_{\mathbf{k}}({\mathbf{r}})=\frac{1}{\sqrt{\Omega}}e^{i\mathbf{k}\cdot{\mathbf{r}}} , where \mathbf{k} denotes a wave vector in the reciprocal lattice, and \Omega is the real-space volume of the primitive cell such that the plane waves are orthonormal over \Omega . To maintain a reasonable number of plane waves, a cutoff is imposed |\mathbf{k}|<k_{max} . Thus, PPs achieve two objectives: (1) at the expense of introducing nonlocality in the pseudopotential, they remove the need to explicitly consider core electrons as these are, in most applications, chemically inert; and (2) they eliminate the singularity of the Coulomb electron-nuclear potential at the nuclear positions 11 ; 17 . Unfortunately, accurate and transferable nonlocal PPs (NLPPs) commonly used in KS-DFT, are not applicable to OF-DFT because they require KS orbitals to couple to the nonlocal part of the PP. In OF-DFT, as we have seen, orbitals are not immediately available. Although we refer to recent reviews for additional details 28 ; 45 , in the past decades several attempts have produced effective local PPs (LPPs, hereafter) that have opened OF-DFT’s applicability to main group metal elements, groups III-V elements third row and below, and a selected number of transition metal elements (such as Ag) 31 ; 18 ; 51 ; 8 ; 26 . The guiding principles to build LPPs have been to employ them in a KS-DFT calculation (i.e., with exact T_{s} ) imposing it to reproduce the electron density, lattice constant and even the band structure of solids 31 ; 18 ; 51 . For those elements for which this is possible, it is said LPPs can be employed in OF-DFT calculations. It is certainly surprising that band structures can be reproduced in pseudopotential calculations having only a local part. However, in many cases it was shown to be possible 42 . It turns out that LPPs can be successfully constructed for the main-group elements mentioned above and also for a large array of transition metals (provided that the semicore states are excluded from the pseudization 4 ). The most accurate LPPs available to date for transition metal elements are the recently developed highly accurate LPPs by the Huang group 4 . The valence density of transition metals consists of both localized d electrons and delocalized s – p conduction electrons, a regime in which most existing KEDFs are constructed from the response of homogeneous or nearly homogeneous electron gas struggle to provide accurate kinetic energies. In addition, reliable LPPs must account for semicore s -like states that lie close in energy to the valence d bands and influence bonding and screening. Since OF-DFT approximates T_{s}[\rho] strictly as a functional of the electron density, it lacks explicit angular-momentum dependence and therefore cannot distinguish d -state character from s or p contributions. These combined factors explain why transition metals are particularly demanding within OF-DFT and motivate the construction of pseudopotentials specifically designed to mitigate these limitations. In this work, semicore states are not explicitly included in the optimization procedure to preserve numerical stability across the periodic table; however the proposed methodology corrects the KEDF and the PP at the same time by reproducing correct scattering properties and remain transferable in bulk and cluster environments. A flaw, however, is that the LPPs have so far been developed based on comparison between KS-DFT calculations

Read the complete text