Comsuite User Guide

Interface

Developed by Hubertus Van Dam
Comsuite will be accessed on the level of a unified software interface in form of a Python script that will also run on HPC systems. Its main purpose will be to provide an easy to use and robust mechanism to set up and control calculations for non-expert users.  Learn more

Ab initio electronic structure platform

Developed by Andrey Kutepov
The electronic structure lies at the heart of Comsuite, as it is a basic building block for the theoretical calculation of material properties. It can be understood as an eigenvalue problem in which the determination of the electron self­-energy matrix Σ is the main challenge. The ab initio electronic structure platform offers several first­ principle and many body diagrammatic approaches (LDA, HF, scGW, LQSGW) with different levels of approximations to Σ. These approaches allow one to study the electronic structure, i.e., the ground and excited state electronic properties of weakly to moderately correlated materials. One can study the properties of strongly correlated materials by combining these approaches with DMFT or GRISB. In all cases, the material must have a known, crystalline (periodic) structure.

Our ab initio electronic structure platform is implemented in a fully relativistic way (based on Dirac equations), which is unique for both scGW and LQSGW. It is based on the stand­alone software package FlapwMBPT.
Learn more

Currently, we provide ab initio LQSGW+DMFT (PDF) and charge self-consistent LDA+DMFT (PDF) within Comsuite.

ComDMFT for ab initio LQSGW+DMFT

Developed by Sangkook Choi, Patrick Sémon, Byungkyun Kang, Andrey Kutepov, and Gabriel Kotliar
ComDMFT is a massively parallel computational package to study the electronic structure of correlated-electron systems (CES). Our approach is a parameter- free method based on ab initio linearized quasiparticle self-consistent GW (LQSGW) and dynamical mean field theory (DMFT). The non-local part of the electronic self-energy is treated within ab initio LQSGW and the local, strong correlation is treated within DMFT. 

We provide a detailed example showing how ab initio LQSGW+DMFT (PDF) can be used to compute the electronic structure of MnO.

ComDMFT for charge self-consistent LDA+DMFT

Developed by Sangkook Choi, Patrick Sémon, Byungkyun Kang, Andrey Kutepov, and Gabriel Kotliar
ComDMFT is a massively parallel computational package to study the electronic structure of correlated-electron systems (CES). In addition to ab initio LQSGW+DMFT, charge self-consistent LDA+DMFT methodology is also implemented, enabling multiple methods in one platform for the electronic structure of CES. We provide a detailed example showing how charge self-consistent LDA+DMFT (PDF) can be used to compute the electronic structure of MnO.

ComCTQMC

Developed by Patrick Sémon
In implementing DMFT, one must always solve an effective quantum impurity problem. In both our GW+DMFT and LDA+DMFT packages, we use a continuous time quantum Monte Carlo (CTQMC) impurity solver to tackle this problem.

ComRISB

Developed by Yongxin Yao, Sangkook Choi, Byungkyun Kang, Andrey Kutepov, and Gabriel Kotliar
Gutzwiller rotationally invariant slave­ boson method (Com­RISB) solves a generic multi­band Hubbard model (including local correlated orbitals and nonlocal orbitals). Combined with our ab initio electronic structure platform, Com­RISB can describe realistic materials with different degrees of electron correlations. Com­RISB yields ground state properties with comparable accuracy to ComDMFT, but it is over two­ orders of magnitude faster. It can handle all of the possible local symmetries without introducing further approximations. Com­RISB consists of programs, executables, and scripts written in Fortran90, C (C++), and Python2.7. The Com­RISB code is based on the stand­alone software package CyGutz. We provide a detailed example showing how ComRISB (PDF) can be used to compute the electronic structure of of MnO.

Post-processing tools for theoretical spectroscopies

Developed by Ran Adler and Yilin Wang
Post­processing tools for theoretical spectroscopies compute physical observables, namely transport properties such as the optical conductivity or the Seebeck coefficient. The (renormalized) electron density of states (spectral function) that is directly obtained from the converged many­-body Green's function (or self­-energy) of previous methods serves as input, together with light­ electron matrix elements corresponding to the experimental setup.

This unique spectroscopy toolbox will allow scientist to perform direct comparisons between theoretical Comsuite predictions and experimental results. In the long term, we believe that Comsuite will facilitate material design projects.

We plan to release the post-processing modules in the next code release.