QE nscf band structures

These calculations give the band structure of each MC3D structure along its high-symmetry k-point path, and the Fermi level, metallicity and band gap read off it.

Published as r3byg-xp284 — 19,405 structures, with the eigenvalues themselves, not only the derived numbers.

What was calculated

Setting	Value
Code	Quantum ESPRESSO pw.x
Task	nscf band structure, from the charge density of a self-consistent parent
Functional	PBEsol
Pseudopotentials	SSSP
Spin	None
Cell	Primitive, as SeeKpath standardises it
k-point path	SeeKpath, through PwBandsWorkChain

Same settings and the same structures as the SCF k-point sweeps, so a material's two records describe the same system computed the same way.

Reproduce it

The workflow is the standard PwBandsWorkChain, which standardises the cell, builds the SeeKpath path, runs the self-consistent parent, and then the nscf calculation on its charge density. You need AiiDA with a configured pw.x code and an installed pseudopotential family.

uv sync --extra aiida

from aiida import load_profile, orm
from aiida.engine import submit
from aiida.plugins import WorkflowFactory
from ase.io import read

load_profile()

PwBandsWorkChain = WorkflowFactory("quantumespresso.pw.bands")
pseudo_family = "SSSP/1.3/PBEsol/efficiency"

builder = PwBandsWorkChain.get_builder_from_protocol(
    code=orm.load_code("pw-7.5@your-computer"),
    structure=orm.StructureData(ase=read("100115.cif")),
    protocol="balanced",
    overrides={
        "bands_kpoints_distance": 0.1,
        "scf": {"pseudo_family": pseudo_family},
        "bands": {"pseudo_family": pseudo_family},
    },
)
node = submit(builder)
print(node.pk)

bands_kpoints_distance is what controls how densely the path is sampled. The built-in protocols use 0.1 (fast), 0.025 (balanced) and 0.015 (stringent) Å⁻¹, and the override above pins it rather than inheriting it, which matters — see the note below.

When the workchain finishes, node.outputs.band_structure carries the eigenvalues and the labelled k-points, and node.outputs.band_parameters the Fermi energy.

Pin the path density, and compare paths before eigenvalues

The published record was not produced with a single bands_kpoints_distance, so its density varies from structure to structure. Take it from the published k-point coordinates rather than assuming a protocol default.

Two paths that differ in sampling density are not comparable row by row even when both are correct. Compare kx, ky, kz and the high-symmetry label first; only then compare eigenvalues.

Explore the published data

The record is self-contained — it carries the cell each calculation ran on, so you never have to re-derive a structure from MC3D to work with it.

example.py in the record loads the summary, the structures and the bands straight out of the tarballs and plots one band structure:

python example.py                 # the summary, then a plotted band structure
python example.py 100115          # one structure by source_db_id

CIF_files/<source_db_id>.cif is the primitive cell the bands were computed on, not a conventional cell that then needs standardising. bands/<source_db_id>.csv carries kx, ky, kz and label alongside the eigenvalues, so a regenerated path can be checked point for point against the published one.

metallicity and band_gap_ev are both derived from the published eigenvalues and Fermi energy, so neither has to be taken on trust. The threshold used for metallicity is a plain zero — see the record page for why no thresholded label is published, and pick your own cut on band_gap_ev if your work needs one. The record page also explains how the k-point coordinates were reconstructed and on what evidence each one was accepted.