Data mining to discover hundreds of new 2D and 1D materials

< back to research overview

We created a database of hundreds of 2d layered materials and 1d molecular wires.
Click here to download the database! (Updated October 2019)

Layered materials have vast potential in applications, both in the bulk form and in a single layer. Bulk layered materials have been widely used for various technologies, for example in energy storage; most commercial Li-ion batteries today use layered materials such as graphite or layered metal oxides to store lithium. In a single layer form, they provide advantages such as maximum mechanical flexibility and optical transparency. Applications in nanoscale devices are numerous, and atomically thin transistors and resonators have been demonstrated. Even though there is a lot of research interest in layered two-dimensional materials, only a few of them are being studied. Graphene, BN, transition metal dichalcogenides and some others like phosphorene dominate the research on 2D materials.

Examples of 2d layered (left) and 1d molecular wire (right) structures that our data mining algorithm discovered.

In this work, we use data-mining to screen almost 60,000 materials from the Materias Project database to identify discover 1173 2D layered materials, 487 1D materials, 98 lattice-commensurate heterostructures and 325 materials with piezoelectric monolayers. To do this, we developed a novel algorithm that can determine the dimensionalisy of the strongly bonded subunits in the crystal. We find bonds between atoms by comparing interatomic distances between all pairs of atoms in a unit cell with the sum of their bond radii. If the distances are smaller than the sum of bond radii + tolerance of 0.45 angstroms (to account for bond length variability), we consider the pair bonded. We find all connected clusters of atoms in the unit cell, and do the same for a 2x2x2 supercell of the unit cell. Now, we count the number of atoms in the connected clusters in the original unit cell and the 2x2x2 supercell. If the connected cluster is two-dimensional, the number of atoms in a cluster would have increased to 4 times the number in the original unit cell, because the size of a two-dimensional object scales as length^2.

Using this algorithm, we identified 1173 2D weakly bonded solids and 487 1D weakly bonded solids with unique chemical formulas, consisting of 23 distinct chemical families of 2D weakly bonded solids and eight chemical families of 1D weakly bonded solids. While layered materials have received considerable attention, much less was known about bulk inorganic crystals that consist of weakly bonded one-dimensional chains, or molecular wires. Less than 20 such materials were found in literature, and this work is the first to uncover a variety of 1D weakly bonded solids. The majority of weakly bonded solids identified in this work come from experimentally reported crystals composed of stacks of 2D layers or 1D chains of bonded atoms.

We also identify 98 weakly bonded lattice-commensurate heterostructures, i.e. materials with weakly bonded adjacent units with dissimilar properties. There has been much activity on assembling layers of different materials to make van der Waals heterostructures with new material properties3,4. However, the fabrication of such heterostructures present a significant experimental challenge. The materials identified in our work have the substantial manufacturing advantage of being amenable to growth using more conventional and scalable methods.

We also present data on the band gaps of the 2D and 1D materials, as well as the point groups. Crystal symmetries of bulk and monolayer 2D materials can be different for many layered materials, and the physical properties affected by them are also different. For example, TMDs in the 2H phase with inversion symmetry in the bulk form are not piezoelectric, but single layers of TMDs are piezoelectric. We developed a method to calculate monolayer point group symmetries from AA-stacking of the monolayers. Combining this with the band gap data, we identify 2D piezoelectric material candidates from point group symmetries.


Cheon, G., Duerloo, K.-A. N., Sendek, A. D., Porter, C., Chen, Y., Reed, E. J., Data Mining for New Two- and One-dimensional Weakly Bonded Solids and Lattice-commensurate Heterostructures. Nano Letters, doi:10.1021/acs.nanolett.6b05229 (2017).