Quantum ESPRESSO input generator and structure visualizer

This tool takes as input a crystal structure given in a variety of formats (native PWscf, XYZ, XCrysDen, CIF, VASP, Castep, and PDB), and prepares an input file for the PWscf code of Quantum ESPRESSO, using reliable standard parameters that can be used to perform a self-consistent calculations for the chosen structure. It's not meant to be a replacement for your own understanding, but it can be used as a sanity check - in particular we provide pseudopotentials that have been pre-screened using the SSSP protocol, with reliable cutoffs for the charge density and the wavefunctions.

In addition, there is an online visualizer that can be used to study the structure that you have just chosen - it could be even your own PWscf input file, to double-check a structure you generated yourself.

Note that we do not keep track or store any of the files you upload to the service, and thus obviously we do not share or disclose them to anyone - ourselves or third parties. More details can be found in the Terms of Use.

A detailed discussion of all input parameters for PWscf can be found here, while a more in-depth documentation is here.

Here, we want just a minimal setup, where you

1. Input the crystal structure.
2. Choose a pre-defined protocol for your use case that determines the pseudopotential, k-point sampling, smearing, and convergence thresholds. The parameters have been benchmarked by Gabriel de Miranda Nascimento et al. (in preparation). See below for a table of all the choices.
3. Choose the exchange-correlation functional.
4. Decide if you want to treat your system as a non-magnetic metal (i.e. using smearing/fractional occupations), as a non-magnetic insulator (i.e. using fixed occupations equal to 2), or as a magnetic system (in this case, it's safer to use smearing/fractional occupations, unless you know what you are doing). When in doubt re metallicity, just choose it and see if all states above the Fermi energy are empty. When in doubt re magnetism, choose it and see if the absolute magnetization is zero (note that magnetic calculations are twice more expensive, more difficult to converge, and there can be different self-consistent states with different magnetizations - non-magnetic state, ferro, anti-ferro, ferri...). For fractional occupations, we are using here cold (i.e. Marzari-Vanderbilt-De Vita-Payne) smearing, but note that if your system is actually an insulator or a molecule a monotonic occupation function (Gaussian) would be safer in a few cases (cold smearing can put the Fermi energy of an insulator just below the valence, and Methfessel-Paxton can put it both just below the valence, or just above the conduction).
5. In the advanced settings, you can override any of inputs determined by the protocol (see below).

Protocols

k-points and smearing

Convergence thresholds

Some extra points of consideration:

• Regarding k-point sampling: we change the smearing (degauss) together with k-points grids since a coarser sampling benefits from more smearing. For insulators, we recommend to always use the "medium" as opposed to the "fine" sampling, as they are less problematic to converge with respect to the k-point sampling than metals. For metals with d of f electrons (lanthanide/actinide elements), the "fine" sampling is the most recommended choice, as these states may introduce a sharp density of states near the Fermi level that affects the smearing/k-points convergence behavior. Keep in mind: there is no point in sampling the Brillouin Zone in a direction in which there shouldn't be any dispersion (as is the case for, say, a 0-, 1- or 2-dimensional system where 3, 2, or 1 directions shouldn't be sampled); in these cases you should update the input file by hand a posteriori. You can also decide to use a shifted Monkhorst-Pack mesh (just change 0 0 0 -> 1 1 1 in the last 3 numbers of the k-point entry). If you are a PWscf Jedi, for large cells you could use the Baldereschi point (1/4 1/4 1/4) and nosym=.true. to get much better sampling than Gamma-only (at twice the cost).
• Note that for a 'scf' calculation etot_conv_thr and forc_conv_thr are not relevant (but would be a good choice for a 'relax' calculation). Both energy thresholds ('etot_conv_thr' and 'conv_thr') are per atom quantities, so that they get scaled by the number of atoms in the cell. Regarding the conv_thr, if you have 1) a lot of electrons/atom, 2) planning to do a phonon calculation afterwards, 3) calculating a "single-atom" property in a large cell (e.g. a formation energy) you might want to manually use a tighter threshold - say 2.0 · 10^-11 times the number of atoms.
• This tool is only a preliminary step to running a full PWscf calculation (where e.g. you might want to relax the atomic positions or the cell geometry, or perform more complex operations). Note also that even with the parameters provided the calculation is not guaranteed to converge (decrease the mixing_beta parameter, as a first remedy) or be suitable for your purpose, and it's certainly not optimized for speed.
• More complex but standardized workflows are being developed, exploiting the workflow engine of AiiDA. The Quantum Mobile virtual machine that you can run on any computer (Windows, Mac, Linux, etc...) provides an Ubuntu environment which comes with Quantum ESPRESSO, AiiDA, and all the other MaX codes preinstalled and ready to run. Additionally, the AiiDAlab QE app can be used to run standard Quantum ESPRESSO workflows.

The pseudopotentials provided here are generated according to the SSSP protocol (https://materialscloud.org/sssp). Please make an effort to acknowledge all the original authors (click here for the acknowledgements list) and of being compliant with the corresponding licenses. Also, to ensure reproducibility of your calculations always list or cite all pseudopotentials used.

We acknowledge financial support by the MARVEL NCCR, the H2020 MaX Centre of Excellence and the swissuniversities P-5 project "Materials Cloud" for the implementation of this tool.