Mn environment in doped SrTiO3 revealed by first-principles calculation of hyperfine splittings

Abstract

In order to take advantage of the strong correlation between geometric changes and electronic properties, better understanding of the structure and properties of perovskites doped with paramagnetic transition metal ions is required. Computational characterization of these dopants and calculated hyperfine values provide a guide that can ultimately aid in the interpretation of experimentally obtained electron paramagnetic resonance (EPR) spectra. In this study, we perform ab initio calculation of the hyperfine splitting parameter for Sr-substituted Mn2+ in SrTiO3 in various geometries in order to assign experimentally reported EPR peaks to exact dopant structures. Additionally, we calculate the hyperfine parameters for Mn4+ and Mn2+∕4+ with adjacent oxygen vacancy, which remain to be assigned peaks in experimental EPR spectra. Calculation of Ti-substituted Mn is largely hindered by the highly correlated electronic structure; however, it is shown that reasonable hyperfine values for Ti-site defects can be obtained by tuning functional parameters. Overall, this study demonstrates that calculating the hyperfine splitting for a transition metal dopant in a complex oxide is feasible and can provide a fingerprint for different geometries in equivalently defected systems.

A-Site Doping to Alter Oxygen Vacancy Diffusion in SrTiO3

“…Therefore, a careful analysis of any DFT solution is required. In fact, even hybrid methods like HSE06 are unable to correctly predict the experimental spin state of Mn doped on the B-site, 31 limiting the current study to A-site dopants. In all calculations, Bader analysis 33 revealed positive charges of +1.7, +1.4, and +1.4 for Cr 3+ , Fe 2+ , and Mn 2+ dopants, respectively (Table S2).…”

Lessons Learned from Catalysis to Qubits: General Strategies to Build Accessible and Accurate First-Principles Models of Point Defects

“…Often dopant geometries are calculated in isolation as idealized structures, as including real crystal features like other defects and interfaces is nontrivial. Thus, standard practice for dopant modeling is to begin with a pure crystal representation of the host material and insert the dopant atom at either a substitutional or an interstitial site. − , In principle, this idealized structure can then be optimized to the lowest energy configuration.…”

“…Indeed, recent exemplary insights into experimental data by comparison to DFT are far too numerous to capture in a single publication. However, scanning tunneling microscopy and spectroscopy, − vibrational properties, electron paramagnetic and nuclear magnetic resonance, Mössbauer spectroscopy, ,, and transport and optical properties are amenable to comparison with experiment. Computational studies should begin reporting these values to evaluate the quality of a computational model and to provide direct comparison to experiment.…”

“…23 From a more practical standpoint, constructing the allelectron charge density from PPs using the projector augmented wave (PAW) method has been shown to be successful in calculating the hyperfine parameters of some transition-metal-doped systems. 24 A comparison between the different approaches was attempted by Ma et al 25 in 2022. Calculated hyperfine parameters for shallow dopant systems using the PP-PAW method in VASP 26,27 and AE and PP-AE in CP2K 28 were compared.…”

“…Thus, standard practice for dopant modeling is to begin with a pure crystal representation of the host material and insert the dopant atom at either a substitutional or an interstitial site. [8][9][10]24 In principle, this idealized structure can then be optimized to the lowest energy configuration.…”