2DMatPedia, an open computational database of two-dimensional materials from top-down and bottom-up approaches
Scope
Open dataset of thousands of monolayer structures derived from Materials Project bulk parents plus elemental-substitution expansions, with consistent high-throughput density-functional properties for screening and machine-learning use.
Summary¶
Two-dimensional materials discovery increasingly relies on systematic computational libraries. 2DMatPedia compiles more than six thousand monolayer structures obtained by exfoliating layered bulks identified in the Materials Project using a topology-based screening rule, then augmenting that set with bottom-up chemical substitutions within periodic-table families. Structural, electronic, and energetic descriptors are computed in a uniform high-throughput density-functional framework and released publicly for downstream analytics. The Scientific Data data descriptor positions the resource as bridging top-down exfoliation hypotheses—starting from known layered parents—and bottom-up elemental substitutions that explore hypothetical 2D stoichiometries not yet isolated experimentally.
Methods¶
Structure discovery (top-down). All bulk entries in the Materials Project are screened for layered structures with a topology-based algorithm; candidate bulks are theoretically exfoliated to monolayers (Fig. 1 workflow in the paper).
Structure discovery (bottom-up). Additional 2D stoichiometries are generated by elemental substitution of atoms in known 2D structures with other elements from the same group in the periodic table, starting from unary and binary compounds produced in the top-down stage.
DFT / database (high-throughput). Structural, electronic, and energetic properties of the resulting >6,000 monolayers are computed in a uniform high-throughput density-functional framework aligned with Materials Project conventions: PBE GGA functional, plane-wave basis with PAW pseudopotentials, Monkhorst–Pack k-point meshes scaled to structure size, and Hubbard / spin–orbit treatments where noted per entry class. Dispersion: GGA semilocal errors for weak interactions are mitigated per the MP-compatible protocol referenced in the paper (vdW-consistent recipes as documented on 2dmatpedia.org and the Sci. Data descriptor). Computed properties include formation energies, band gaps, exfoliation energies, and related electronic-structure DOS / band summaries used for screening.
Access. Records, provenance, and validation are released at http://www.2dmatpedia.org/ with downloadable flat files for screening, mining, and ML training.
MD / reactive dynamics (not applicable). N/A — the database is built from high-throughput static (or structurally relaxed) DFT; no production MD, AIMD, or ReaxFF trajectories are reported in the data descriptor focus summarized here.
Findings¶
The combined top-down and bottom-up construction yields broad coverage of hypothetical 2D chemistries beyond hand-curated sets, with homogeneous computational metadata across entries. The authors position the release as a reproducible starting point for property screening, data mining, and machine-learning interatomic potentials, and document technical validation benchmarks for database consumers. 2DMatPedia targets users who need thousands of relaxed monolayers with consistent DFT-derived descriptors (e.g., band gaps, formation energies, exfoliation energies) rather than isolated prototype flakes.
Corpus honesty. Fully precise functional, k-mesh, and basis recipes are defined in the paper and on the portal; this page does not copy every high-throughput convergence default line-by-line.
Limitations¶
Density-functional approximations and high-throughput convergence criteria limit quantitative accuracy for every niche chemistry; users must perform higher-level theory or experiment for critical candidates.
Hubbard U corrections, vdW functionals, and spin–orbit effects are not uniformly applied across every 2DMatPedia record; screening priorities should dictate whether a standard PBE entry is sufficient or requires post-processing GW/hybrid spot checks.
Relevance to group¶
Provides external DFT reference data adjacent to group workflows on two-dimensional energy materials, though not a ReaxFF study.
Citations and evidence anchors¶
- https://doi.org/10.1038/s41597-019-0097-3