Beyond LSDA and Thomas-Fermi: A Hartree-Fock Analysis of Spherical Nanoparticles in the Jellium Approximation
Michael Píro, Jaroslav Hamrle · 2025
Read Beyond LSDA and Thomas-Fermi: A Hartree-Fock Analysis of Spherical Nanoparticles in the Jellium Approximation by Michael Píro, Jaroslav Hamrle (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-2507-05002.txt.
Introduction
Michael Píro and Jaroslav Hamrle study the ground state of tiny round grains in a simple model of matter. Their paper is from two thousand twenty-five, though its text head bears a shorter title. The model, called jellium, holds electrons in a smooth field of plus charge. They use a Hartree-Fock method for grains with up to a few hundred electrons.
They check how well local density shortcuts match the fuller method for exchange and motion energy. Large gaps show up, both deep inside the grain and at its skin. They give refined forms for the local energy, tuned by grain size and charge density. They also set these next to gradient-based forms in wide use.
Catalog record
| CREATOR | Michael Píro, Jaroslav Hamrle |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 43 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-05002
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Beyond LSDA and Thomas-Fermi: A Hartree-Fock Analysis of Spherical Nanoparticles in the Jellium Approximation Michael Píro, Jaroslav Hamrle Abstract A comprehensive Hartree-Fock analysis of the ground-state electronic structure of spherical gold nanoparticles modeled within the jellium approximation is carried out for all systems containing up to 132 delocalized electrons. Emphasis is placed on resolving the energy-level ordering, properly describing the electron density tail and associated charge spill-out, and assessing the accuracy of local exchange and kinetic energy potentials. The calculations are performed on a high-resolution real-space grid to ensure numerical precision. Significant discrepancies are observed between the exchange energy given by the Hartree-Fock approximation and the local spin density approximation (LSDA) in both the inner and outer regions of the nanoparticle. To address these differences, a refined expression of the one-electron exchange energy density as an explicit function of the charge density is proposed. Similarly, the failure of the Thomas-Fermi kinetic energy model near the surface of the nanoparticle is resolved by introducing an improved expression for the one-electron kinetic energy density. I Introduction Nanoparticles of different sizes and shapes are known to exhibit fundamentally different optical, magnetic, and catalytic properties compared to their bulk material counterparts Kelly et al. 2003 ; Sosa et al. 2003 ; Hernando et al. 2005 ; Kawawaki and Negishi 2020 . These properties arise from the confinement of electrons at the nanoscale and are essential for a wide range of modern technological applications, including drug design and delivery, and bioimaging in medicine Afzal et al. 2022 ; Harish et al. 2022 , photocatalysis Bowker et al. 2022 , or surface-plasmon-based sensing Khurana and Jaggi 2021 . The rapid development of nanofabrication techniques has increased the demand for theoretical models that can accurately predict the behavior of electronic states in nanoparticles with high precision and minimal computational cost. A complete quantum-mechanical description of multi-electron nanoparticles quickly becomes intractable as system size increases. To address this, simplified models have been developed that still capture the essential physics. One widely used approach is the spherical jellium model, where the discrete ion background is approximated as a uniform positive charge distribution Lang and Kohn 1970 ; Brack 1993 . This model has been particularly successful in describing the delocalized conduction electrons of alkali-metal clusters. The work focused mainly on the use of density functional theory (DFT) methods to calculate energy spectra Ekardt 1984a ; Ekardt 1984b ; Redfern et al. 1985 ; Knight et al. 1987 , although several special cases of closed shells in sodium clusters were calculated using the Hartree-Fock approximation Guet and Johnson 1992 ; Hansen and Nishioka 1993 ; Madjet et al. 1995 . Beyond standard DFT, alternative approaches such as quantum hydrodynamic models (QHD) Ciracì and Della Sala 2016 ; Zhou et al. 2023 and quantum-corrected classical models (QCM) Esteban et al. 2012 ; Zapata et al. 2015 have gained traction, particularly in the field of plasmonics. These methods treat the electron fluid collectively, while incorporating quantum effects such as the Bohm potential or pressure terms derived from the kinetic energy density. They enable simulations of larger systems and dynamical phenomena, but their precision is heavily based on the quality of the input energy functionals. A common limitation of both the DFT and QHD models lies in the exchange and kinetic energy functionals used. In particular, the local spin density approximation (LSDA) can lose its accuracy significantly on surfaces of spatially confined systems Krotscheck and Kohn 1986 . More advanced generalized gradient approximations (GGAs) such as the Perdew-Burke-Ernzerhof functional (PBE) offer limited improvements when the electron density gradient is small relative to the density itself. Similarly, the Thomas-Fermi kinetic energy functional severely underestimates surface effects, failing to describe the electron spill-out or electron density decay rate outside the nanoparticle. In this work, we address these limitations by explicitly computing the ground-state electronic structure of spherical gold nanoparticles in the finite jellium model using the Hartree-Fock approximation for systems with up to 132 delocalized electrons. Electrons are added one by one to correctly determine the energy-level ordering. Upon analyzing the energy spectra and the shapes of the electron densities, we calculate the one-electron exchange and kinetic energy densities and find appropriate fits to express them as functions of the electron density. For simplicity, we adopt Hartree atomic units throughout the text. II Model description In the finite spherical jellium model, the nanoparticle is treated as a solid ball of radius R composed of N atoms. Each atom contributes a positively charged ionic core (consisting of the nucleus and tightly bound inner electrons) and \nu delocalized electrons. The positive charge is assumed to be uniformly distributed throughout the interior of the ball, resulting in the ionic charge density: \displaystyle n_{\text{I}}(\textbf{r})=\begin{cases}n_{\text{I}}&|\textbf{r}|\leq R\\ 0&|\textbf{r}|>R,\end{cases} (1) where n_{\text{I}} is the background density constant. The total positive charge, given by Q=\nu N , can then be expressed as Q=\frac{4}{3}\pi R^{3}n_{\text{I}} . Taking into account the original crystalline arrangement of the atoms, the radius R is related to the number of atoms by R=r_{s}N^{1/3} , where r_{s} is the Wigner-Seitz radius. Substituting this into the expression for Q , the ionic density constant becomes: \displaystyle n_{\text{I}}=\frac{3}{4\pi}\frac{\nu}{{r_{s}}^{3}}. (2) As shown in the Supplementary Material, Section S1, this positive background creates the electrostatic nanoparticle potential V(r) : \displaystyle V(r)=\begin{cases}\frac{\nu}{r_{s}^{3}}\frac{r^{2}-3R^{2}}{2}&r\leq R\\[2.84526pt] -\frac{\nu}{r_{s}^{3}}\frac{R^{3}}{r}&r>R.\end{cases} (3) Its shape is depicted in Fig. 1 . The parameters r_{s} and \nu depend on the specific material. In this study, we focus on gold atoms, where only the 6s electron is treated as delocalized. This yields r_{s}=3.01 Bohr and \nu=1 . The system of a single electron in the nanoparticle potential is spherically symmetric. The electron wave function is therefore separated into its angular and radial components: \displaystyle\psi_{nlm}(r,\theta,\phi)=\frac{1}{r}u_{nl}(r)Y_{l}^{m}(\theta,\phi), (4) where n , l , and m denote the principal, azimuthal, and magnetic quantum numbers, respectively. Since the angular wave functions Y_{l}^{m}(\theta,\phi) are determined analytically and are independent of the shape of the radial potential V(r) , we further examine only the radial components. To address the multi-electron problem, we add the electron-electron interaction to the Hamiltonian. This is given by the potential V_{\text{e-e}} , which accounts for all pairwise interactions between electrons: \displaystyle V_{\text{e-e}}(\textbf{r}_{1},\dots,\textbf{r}_{N})=\sum_{\begin{subarray}{l}i,j=1\\ \ i<j\end{subarray}}^{N}V_{2}(\textbf{r}_{i},\textbf{r}_{j})=\sum_{\begin{subarray}{l}i,j=1\\ \ i<j\end{subarray}}^{N}\frac{1}{|\textbf{r}_{i}-\textbf{r}_{j}|}. (5) Next, the spatial one-electron wave function is extended to include spin. We consider the electrons to be only fully in the spin-up or spin-down state. The total many-electron wave function is given by the Slater determinant. The interaction potential then decomposes into two distinct contributions: the Hartree potential \hat{V}_{\text{H}} and the exchange potential \hat{V}_{\text{x}} , defined as: \displaystyle\hat{V}_{\text{H}}\psi_{i}(\textbf{r})=\sum_{\begin{subarray}{l}j=1\\ j\neq i\end{subarray}}^{N}\int_{\mathbb{R}^{3}}\psi_{j}^{*}(\textbf{r}^{\prime})V_{2}(\textbf{r},\textbf{r}^{\prime})\psi_{j}(\textbf{r}^{\prime})d\textbf{r}^{\prime}\psi_{i}(\textbf{r}), (6) \displaystyle\hat{V}_{\text{x}}\,\psi_{i}(\textbf{r})=-\sum_{\begin{subarray}{l}j=1\\ j\neq i\end{subarray}}^{N}\delta_{s_{i}s_{j}}\int_{\mathbb{R}^{3}}\psi_{j}^{*}(\textbf{r}^{\prime})V_{2}(\textbf{r},\textbf{r}^{\prime})\psi_{i}(\textbf{r}^{\prime})d\textbf{r}^{\prime}\,\psi_{j}(\textbf{r}), (7) where s_{i} denotes the spin of the i -th electron. It is important to note that the inclusion of self-interaction terms (contributions with i=j ) in both \hat{V}_{\text{H}} and \hat{V}_{\text{x}} results in their mutual cancellation, leaving the total electron–electron interaction potential unchanged. This small change is useful for the simplification of further calculations. The one-electron wave functions are again decomposed into their radial and angular parts, as in Eq. ( 4 ), now with the addition of the spin quantum number s to the radial functions. To preserve spherical symmetry of the solution, electrons in partially filled shells are averaged over all quantum numbers m