In search of the ground state crystal structure of Ta$_2$O$_5$ from ab-initio and Monte Carlo simulations

TL;DR

This study combines ab-initio and Monte Carlo simulations to identify the gamma phase of Ta₂O₅ as the most accurate structural model, matching experimental energy loss spectra and clarifying its polymorphic nature.

Contribution

The paper demonstrates that the gamma phase of Ta₂O₅ best explains experimental energy loss data, using advanced simulations including local field effects and spin-orbit coupling.

Findings

The gamma phase explains experimental reflection energy loss signals.

Including local field effects and spin-orbit coupling is essential.

Monte Carlo simulations match experimental spectra well.

Abstract

Tantalum oxides (Ta$_2$O$_5$) are characterised by attractive physical and chemical properties, such as high dielectric constants and anti-reflection behaviour. Recently, Ta$_2$O$_5$ nanoparticles have also been proposed as possible enhancers of the relative biological effectiveness in hadrontherapy for cancer treatment. In principle, their electronic properties can be accurately investigated from first-principles simulations. However, the existence of several stable polymorphs of these oxides represents a major difficulty in order to calculate and disentangle their respective spectral features. To assess this problem, we use linear-response time-dependent density functional to investigate the energy loss function $\mbox{Im} (-1/\bar{\epsilon})$, which is a unique fingerprint of the material, in the optical limit for various polymorphs. We show that the experimental reflection energy loss signals can be rationalized and interpreted by assuming that the $\gamma$-phase of Ta$_2$O$_5$ represents the underlying structural model. We notice that both the inclusion of local field effects and spin-orbit coupling are crucial to compute the energy loss functions of this material. Finally, to further validate the $\gamma$-Ta$_2$O$_5$ polymorph as a model for experimental tantalum oxide, we compute the reflection energy loss spectra using a Monte Carlo approach, finding an excellent agreement with the experimental data.