| Slide 2 o Motivation: Add capability to simulate infrared and Raman spectra to the Phono(3)py code: Phonopy-Spectroscopy o Features: • (Complex) IR dielectric function 𝜀(𝜔) → (complex) refractive index & 𝑛 𝜔 = 𝑛(𝜔) + 𝑖𝑘(𝜔) → reflectivity 𝑅(𝜔), absorption 𝛼(𝜔), loss function 𝐿(𝜔) • Ionic and static dielectric constants 𝜺!"#!$ /𝜺% + POP frequency 𝜔&" • Raman polarizability tensors 𝜶'( , activities 𝐼'( and differential cross sections ⁄ 𝑑𝜎'( 𝑑Ω, with a partial description of resonance effects • Include calculated phonon linewidths from e.g. Phono3py • Simulate measurements on single crystals and powders, or powders with preferred orientation, for arbitrary instrument geometries • Python API designed for easy interoperability with other codes and for implementing more advanced experiments (e.g. Jupyter notebooks)
2024 | Slide 11 Lasers with photon energy 𝐸 > 𝐸6 can couple to electronic states → resonance effects Common to use the far-from-resonance approximation (FFR): 𝜺 𝐸 = 𝜺 𝐸 = 0 ≡ 𝜺. Unlikely to be reasonable for SnS/SnSe because direct 𝐸6 ≈ 1 eV (1240 nm) is lower than all common laser wavelengths Can partially capture resonance effects by using 𝜺 𝐸 - but need accurate model for dielectric function TD-DFT with dielectric-dependent hybrid functional in principle makes it possible to calculate 𝜺 𝐸 similar accuracy to “gold standard” 𝐺𝑊 + BSE
2024 | Slide 17 Python API mostly complete and available on develop branch of Phonopy-Spectroscopy GitHub repo: https://github.com/skelton- group/Phonopy-Spectroscopy Will shortly add minimal working examples for reproducing SnS/SnSe calculations using a Jupyter notebook Front-end CLI for setting up calculations and post processing in progress
| Slide 18 Implementation (inc. maths + sanity checks): o Jonathan Skelton (UoM) o Anuradha Pallipurath (UoL) o Chidimma Umeh (UoM) Initial testing (i.e. bug squashing): o Joseph Flitcroft (UoM) o Guanping Li (UoM) o David Collins (UoL) … plus other collaborators too numerous to mention