Matyáš Novák | Searcharxiv

arXiv2026

ASE2SPRKKR: a unified Python framework integrating the Spin-Polarized Relativistic Korringa-Kohn-Rostoker method into the Atomic Simulation Environment

The Spin-Polarized Relativistic Korringa-Kohn-Rostoker (SPR-KKR) is an all-electron ab-initio multiple-scattering code that provides unique capabilities for treating chemical disorder, finite-temperature magnetism, relativistic effects, and spectroscopic properties of various types of solids through its fundamental formulation in terms of the single-particle Green's function rather than eigenstates. We present ASE2SPRKKR, a comprehensive Python interface that integrates SPR-KKR into the Atomic Simulation Environment (ASE), making SPR-KKR more accessible, streamlined, and uniform. Our implementation extends the ASE's Atoms object to handle fractional site occupations for coherent-potential-approximation calculations while maintaining full compatibility with ASE's extensive ecosystem of structure builders, optimizers, and analysis tools. Automated input generation with validation, comprehensive output parsing, and direct MPI support enable seamless integration into high-throughput and multi-method workflows. We demonstrate the interface through representative applications: semi-infinite surface calculations reproducing Rashba-split Au(111) surface states; one-step photoemission modeling capturing matrix-element effects; exchange-parameter extraction for atomistic spin dynamics; and X-ray absorption spectroscopy including magnetic circular dichroism. Beyond these demonstrations, ASE2SPRKKR is designed with transferability as a first-class concern. By grounding its architecture in FAIR principles of Findability, Accessibility, Interoperability, and Reusability, it establishes a replicable blueprint for bringing other specialized Green's function and first-principles codes into the collaborative, reproducible workflows that modern materials discovery requires.

cond-mat.mtrl-sci↗

arXiv2016

Isogeometric analysis in electronic structure calculations

In electronic structure calculations, various material properties can be obtained by means of computing the total energy of a system as well as derivatives of the total energy w.r.t. atomic positions. The derivatives, also known as Hellman-Feynman forces, require, because of practical computational reasons, the discretized charge density and wave functions having continuous second derivatives in the whole solution domain. We describe an application of isogeometric analysis (IGA), a spline modification of finite element method (FEM), to achieve the required continuity. The novelty of our approach is in employing the technique of Bézier extraction to add the IGA capabilities to our FEM based code for ab-initio calculations of electronic states of non-periodic systems within the density-functional framework, built upon the open source finite element package SfePy. We compare FEM and IGA in benchmark problems and several numerical results are presented.

physics.comp-ph↗

arXiv2015

Finite element method and isogeometric analysis in electronic structure calculations: convergence study

We compare convergence of isogeometric analysis (IGA), a spline modification of finite element method (FEM), with FEM in the context of our real space code for ab-initio electronic structure calculations of non-periodic systems. The convergence is studied on simple sub-problems that appear within the density functional theory approximation to the Schrödinger equation: the Poisson problem and the generalized eigenvalue problem. We also outline the complete iterative algorithm seeking a fixed point of the charge density of a system of atoms or molecules, and study IGA/FEM convergence on a benchmark problem of nitrogen atom.

physics.comp-ph↗