Abstract
Spin-orbit coupling (SOC) drives interesting and non-trivial phenomena in solid state physics, ranging from topological to magnetic to transport properties. Thorough study of such phenomena often require effective models where SOC term is explicitly included. However, estimation of SOC strength for such models mostly depend on the spectroscopy experiments which can only provide a rough estimate. In this work, we provide a simple yet effective computational approach to estimate the on-site SOC strength using a combination of the $ab$ $initio$ and tight-binding calculations. We demonstrate the wider applicability and high sensitivity of our method considering materials with varying SOC strengths and the number of SOC active ions. The estimated SOC strengths agree well with the proposed values in literature lending support to our methodology. This simplistic approach can readily be applied to a wide range of materials.
Introduction
In the field of condensed matter physics, Spin-Orbit Coupling (SOC) is a fundamental relativistic interaction where an electron's spin couples to its orbital motion. This interaction is the primary driver behind exotic states of matter such as topological insulators, Weyl semi-metals, and high-temperature superconductors. To accurately study these materials, physicists rely on "effective models"—simplified mathematical frameworks that focus on the most critical electronic interactions. These models require a precise value for the SOC strength parameter, denoted as $\lambda$.
Historically, determining $\lambda$ has relied on experimental spectroscopy. However, these measurements are often "noisy"; structural distortions or competing electronic interactions can mask the intrinsic SOC strength, leading to inaccurate models. While standard Density Functional Theory (DFT) can incorporate SOC, it often acts as a "black box," making it difficult to extract the specific $\lambda$ value needed for intuitive physical models. This work introduces a new computational framework to address this gap.
Methodology: The TBSOC Framework
The researchers propose a three-step workflow dubbed TBSOC to precisely extract SOC parameters:
- Baseline DFT and Wannierization: A non-spin-polarized ab initio calculation (DFT) is performed to establish the basic electronic structure. From this, a Tight-Binding (TB) model is constructed using Wannier functions, creating a simplified map of electron hopping.
- Relativistic DFT: A second ab initio calculation is performed, this time fully including SOC effects self-consistently. This represents the "true" electronic behavior of the material.
- Optimization: A mathematical SOC term ($H_{soc}$) containing the unknown parameter $\lambda$ is added to the TB model. Using the Nelder-Mead optimization algorithm, $\lambda$ is tweaked until the TB model’s energy bands perfectly match the relativistic DFT bands.
Results and Discussion
The TBSOC method was validated across several materials representing diverse electronic environments, from 3d to 5d transition metal compounds.
$\alpha$-RuCl$_3$ and Transition Metal Oxides
For $\alpha$-RuCl$_3$, a 4d transition metal compound famous for its potential "Kitaev quantum spin liquid" state, the method estimated $\lambda \approx 0.120$ eV. This value shows excellent agreement with the best available literature. As shown in Figure 1, the optimized tight-binding model (red dashed lines) provides a near-perfect fit to the relativistic DFT band structure (blue solid lines).
The method also demonstrated high sensitivity across scales, successfully capturing $\lambda$ values ranging from the relatively weak 65 meV in Cobalt (3d) to the massive 380 meV in Iridium (5d).
Multi-Ion Systems: TaAs
A significant advantage of TBSOC is its ability to disentangle SOC contributions from different atomic species in the same crystal. In the Weyl semi-metal TaAs, the method successfully extracted separate parameters for Arsenic ($\lambda_p = 0.18$ eV) and Tantalum ($\lambda_d = 0.21$ eV).
As illustrated in Figure 3, using these extracted values allowed for the accurate reconstruction of the characteristic "spoon-shaped" surface states, a hallmark of Weyl semi-metals. This confirms the model's predictive power for complex topological materials.
Key Insights and Conclusion
A major finding of this study is the method's coordinate independence. The choice of coordinate system (local vs. global axes) does not change the estimated $\lambda$, making the method highly robust for complex crystal structures. Furthermore, the authors suggest that computational estimates may be more reliable than experimental ones in certain cases, as experiments often underestimate $\lambda$ due to structural distortions that the computational model can isolate.
In conclusion, the TBSOC method provides a simplistic yet powerful approach to building accurate effective models for SOC-active systems. It avoids the high computational overhead of more complex many-body techniques while maintaining high sensitivity, making it readily applicable to a wide range of materials in modern condensed matter research.
Summary updated