Our Alexandria database has a series of different datasets.

These files were generated with pymatgen. They can be loaded by using the following code:

import json, bz2
from pymatgen.entries.computed_entries import ComputedStructureEntry

with"dataset.json.bz2") as fh:
data = json.loads('utf-8'))

entries = [ComputedStructureEntry.from_dict(i) for i in data["entries"]]

print("Found " + str(len(entries)) + " entries")
print("\nEntry:\n", entries[0])
print("\nStructure:\n", entries[0].structure)

Other datasets

Materials Cloud
Here one can find all our datasets in materials cloud.

Transfer learning on large datasets for the accurate prediction of material properties [submitted]
You can download all models from the paper (WARNING: 72Gb).

Exchange-correlation functionals for band gaps of solids: Benchmark, reparametrization and machine learning [NPJ Comput. Mater. 6, 96 (2020)]
Machine learning models for the prediction of more accurate band gaps can be found here.

High-throughput study of oxynitride, oxyfluoride and nitrofluoride perovskites [J. Mater. Chem. A (2021)]
You can download an array of ComputedStructureEntry for all structures or directly the cif files.

Validation of Pseudopotential Calculations for the Electronic Band Gap of Solids [J. Chem. Theor. Comput. (2020)]
Here you can download the HLE16 and SLOC pseudopotentials.

Finding new crystal compounds using chemical similarity [NPJ Comput. Mater. 7, 12 (2021)]
You can download an array of ComputedStructureEntry for all structures (0, 1, 2) or directly the cif files.

Aenet patches [Phys. Chem. Chem. Phys. (2019)]
Here you can find patches for the aenet package that allow for the training of the forces and stresses. You can download a diff for versions 1.0.0 and 2.0.3, the full source code of versions 1.0.0 and 2.0.3, and an example.

Convex hull (2018.5.26)
In our internal database we have more than 3000 compounds that are on the convex hull and that are not included (yet) in the Materials Project database. If you want to study these materials, or just to include them in your own convex hull, you can download the cif files with the crytallographic structures, or a pickle with the ComputedStructureEntrys (compatible with Materials Project) that you can easily add to your scripts. You can also download our version of the script.

Prediction of stable tI10 and tP10 ternary compounds [J. Chem. Phys. (2018)]
You can download an array of ComputedStructureEntry for all tI10 and tP10 studied in this work.

DFTB+ parameter sets for group IV elements [J. Chem. Theory Comput. (2018)]
You can download both several versions of the parameters, as well as the corresponding training and test sets.

High-throuput study of Heuslers [unpublished]
Here you can find an array of ComputedStructureEntry for all all cubic Heusler compounds (space group 255) with composition ABC2. You can also download the relaxed crystal structures (cif format) for all systems that are close to the convex hull of stability.

High-throuput study of perovskites [Chem. Mat. (2017)]
Here you can find a pickle or a json with the DFT calculations of around 228000 (relaxed) cubic perovskites (space group 221) with composition ABC3, where A, B, C are all element up to Bi excluding rare gases and lanthanides. You can load directly the pickle with the BorgQueen module. These quantities were obtained with pymatgen and the materials project database.

A modified Pettifor scale [New J. Phys. (2016)]

  • This is the actual modified Pettifor scale: He, Ne, Ar, At, Rn, Fr, Es, Fm, Md, No, Lr, Kr, Xe, Cs, Rb, K, Na, Li, Ra, Ba, Sr, Ca, Yb, Eu, Y, Sc, Lu, Tm, Er, Ho, Dy, Tb, Gd, Sm, Pm, Nd, Pr, Ce, La, Cf, Bk, Cm, Am, Pu, Np, U, Pa, Th, Ac, Zr, Hf, Ti, Nb, Ta, V, Mo, W, Cr, Tc, Re, Mn, Fe, Os, Ru, Co, Ir, Rh, Ni, Pt, Pd, Au, Ag, Cu, Mg, Hg, Cd, Zn, Be, Tl, In, Al, Ga, Pb, Sn, Ge, Si, B, Bi, Sb, As, P, Po, Te, Se, S, C, I, Br, Cl, N, O, F, H

  • substitution.dat: NxN matrix, row/column indices correspond to elements (increasing Z). First row/column (corresponding to Z=0) filled with zeroes. Matrix elements are the number of materials with element A substituted by element B.

  • Substitution tables per element:


TD-ELF [Phys. Rev. A (Rap. Comm.) 71, 10501 (2005)]
The time-dependent electron localization function in action.

Carbon nanotubes under stress [Nano Lett. 4, 811-815 (2004)]
Ever wandered what happens if you pull a nanotube?

A primer in DFT
Extra information for this book chapter, published by Springer.