Linear Scaling Approach for Optical Excitations Using Maximally Localized Wannier Functions

Konrad Merkel,Frank Ortmann

2023-09-13

Abstract:We present a theoretical method for calculating optical absorption spectra based on maximally localized Wannier functions, which is suitable for large periodic systems. For this purpose, we calculate the exciton Hamiltonian, which determines the Bethe-Salpeter equation for the macroscopic polarization function and optical absorption characteristics. The Wannier functions are specific to each material and provide a minimal and therefore computationally convenient basis. Furthermore, their strong localization greatly improves the computational performance in two ways: first, the resulting Hamiltonian becomes very sparse and, second, the electron-hole interaction terms can be evaluated efficiently in real space, where large electron-hole distances are handled by a multipole expansion. For the calculation of optical spectra we employ the sparse exciton Hamiltonian in a time-domain approach, which scales linearly with system size. We demonstrate the method for bulk silicon - one of the most frequently studied benchmark systems - and envision calculating optical properties of systems with much larger and more complex unit cells, which are presently computationally prohibitive.

Mesoscale and Nanoscale Physics

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the efficient calculation of optical absorption spectra in large - scale periodic systems. Specifically, the authors propose a method based on Maximally Localized Wannier Functions (MLWF), which can be used to calculate optical absorption properties, especially for complex unit - cell systems containing a large number of atoms. Traditional methods face huge computational challenges when dealing with these systems because optical properties are essentially affected by many - body effects, such as the Coulomb interaction between electron - hole pairs, which leads to a significant increase in the amount of calculation. By using MLWF, the authors hope to overcome these computational limitations and make it possible to calculate the optical properties of large - scale systems.

### Main contributions
1. **Linear scaling method**: The authors propose a linear scaling method that uses MLWF to represent the exciton Hamiltonian, so that the computational complexity is linearly related to the system size, rather than the traditional high - order relationship.
2. **Sparse matrix**: Due to the highly localized nature of MLWF, the generated Hamiltonian matrix is very sparse, which greatly improves the computational efficiency.
3. **Real - space calculation**: The electron - hole interaction term can be efficiently calculated in real space. In particular, for long - distance electron - hole pairs, it can be processed by multipole expansion.
4. **Time - domain method**: The authors also propose a time - domain method to calculate the macroscopic dielectric function, avoiding the eigenvalue - solving process with high computational complexity.

### Theoretical framework
- **Two - particle eigenvalue problem**: Starting from the Tamm - Dancoff approximation, the authors derive the contributions of the single - particle band structure and the screened electron - hole interaction.
- **MLWF basis**: By transforming the Bloch functions into the MLWF basis, the authors show how to represent the exciton wave function and Hamiltonian in the localized basis.
- **Overlap density and Coulomb integral**: It is discussed in detail how to calculate the overlap density and Coulomb integral in the MLWF basis. These integrals are crucial for describing the properties of excitons.

### Numerical implementation
- **Numerical calculation of Coulomb matrix elements**: The authors propose a numerical method to calculate Coulomb matrix elements. This method avoids the numerical divergence problem by introducing an auxiliary density function.
- **Calculation of LFE matrix elements**: The calculation of Local Field Effects (LFE) matrix elements is relatively simple because the potential function used does not diverge.
- **Time - domain method**: Through the time - evolution operator expanded by Chebyshev polynomials, the authors achieve an efficient calculation of the dielectric function.

### Calculation details
- **DFT calculation**: DFT calculations are carried out using the PBE exchange - correlation functional and PAW pseudopotentials to obtain electronic states.
- **Construction of Wannier functions**: MLWF is calculated using the wannier90 code, and its localization characteristics and accuracy are verified.
- **Mesh division**: Regular meshes are used in Fourier space for the calculation of two - particle integrals, and the mesh size is determined by VASP calculations.

### Results
- **Overlap density and Coulomb integral**: By analyzing the overlap density and Coulomb integral at different distances, the effectiveness of the MLWF method is demonstrated.
- **Optical absorption spectrum**: Taking the silicon crystal as an example, its optical absorption spectrum is calculated and the accuracy of the method is verified.

In summary, this paper proposes an efficient calculation method based on MLWF, which can accurately calculate the optical absorption spectrum in large - scale periodic systems, providing a new tool for materials science and optoelectronic device design.