FINDSLAB: Software for Exfoliation and Cleavage of Crystals

FINDSLAB: Software for Exfoliation and Cleavage of Crystals Tom Barnowsky & Rico Friedrich Technische Universität Dresden & Helmholtz-Zentrum Dresden-Rossendorf, Germany This program implements the XCP method to identify 2D materials from bulk materials by estimating bonding energies using a two-body potential model [1]. Potential Models The code supports a range of two-body potential parametrizations \(V(r)\), namely: Lennard-Jones + Yukawa Morse + Yukawa Mie + Yukawa Parameters are provided as plain text files which are specified via the environment variable `FINDSLAB_POTDATA`. Note that all energies have to be multiplied by a factor two to compare to surface/bonding energies. Build Run `make` serially (without `-j n`). A binary will be created in the `bin` directory. Requirements: a recent Fortran compiler, BLAS and LAPACK. Usage FINDSLAB provides some instructions when running `findslab --help`. The code is designed to work with VASP POSCAR files, however, the reader is not fully general and expects the formatting as it is found in the AFLOW database (aflow.org) [2]. To convert general structure files (including those from other codes) to this format, use the AFLOW software to run `aflow --vasp`. The aflow code is available at github.com/aflow-org/aflow. Here we provide an example code to determine HKLSEARCH slabs from bulk Ca3N2 retrieved via the AFLOW REST API [3]: export FINDSLAB_POTDATA= export OPENBLAS_NUM_THREADS=1 curl http://aflowlib.duke.edu/AFLOWDATA/ICSD_WEB/HEX/Ca3N2_ICSD_169727/CONTCAR.relax.vasp | aflow --sconv | findslab --hklsearch Conversion to the conventional unit cell via `aflow --sconv` is optional and is only used here to relate Miller indices to the conventional cell. Release Notes v1.1: Included Morse+Yukawa potential data with code. Fixed crashes occurring at very low spacing between parallel high-index lattice planes. v1.0: Initial release. Acknowledgements The authors thank Carsten Timm, Steve Schmerler, and Moritz Leucke for fruitful discussions. Parts of this work are based on an implementation for creating Miller planes from the atomic simulation environment (ASE) [4]. Additionally, we implement the criterion of Mounet et al. [5] to identify van der Waals-bound layers in bulk structures. License This dataset is published under the Apache 4.0 license. We kindly ask works based on this software to cite this entry and/or the associated publication.