RomeoDFT.jl

This software package allows one to perform a Robust Occupation Matrix Energy Optimization of a crystalline structure using Density Functional Theory.

It is most useful in materials with complex magnetic configurations or where the local orbital physics of one or more atomic species plays a crucial role in describing the ground state. In many of these cases the energy surface of the DFT functional sports many local minima, which in theory should be extensively explored in order to find the global minimum associated with the ground state.

RomeoDFT.jl facilitates a fully automated global search, exploring many different occupations of local atomic orbitals (target states) in order to reliably determine the ground state without much human interaction.

It uses a patched version of QuantumESPRESSO to run the DFT calculations for each of the target states.

Features

• "Press of the button" determination of the ground state starting from a single template input file with the structure and input parameters
• Seamless interfacing with HPC clusters using RemoteHPC.jl and DFControl.jl
• Run geometry optimization for the initial structure as well as any found self-consistent state associated with a local minimum
• Automatically and fully self-consistently determine the Hubbard U parameter for the initial structure as well as any self-consistent state associated with a local minimum
• PostProcessing steps including non-selfconsistent, bandstructure and dos/projected dos calculations
• Robust data storage using JLD2.jl with versioning
• Command line interface facilitating most of the tasks
• Built with ECS (see Overseer.jl making it very extensible

Acknowledgements

This work has been funded by H2020-OpenModel and EPFL.