CrystaLattE

CrystaLattE automates computation of lattice energies in molecular crystals using a many-body cluster expansion that evaluates interactions among dimers, trimers, and higher-order clusters.


Key Features:

  • Many-Body Cluster Expansion: CrystaLattE employs a many-body cluster expansion to systematically account for interactions among dimers, trimers, and higher-order N-mers in molecular crystals.
  • Parallel Computation: Independent computations on different N-mers enable inherent parallelism for scalable evaluation of cluster contributions.
  • Redundancy Avoidance: Identical N-mers are detected and redundant calculations avoided using Coulomb-matrix descriptors that exploit three-dimensional periodic symmetry.
  • Quantum Chemistry Integration: The software interfaces with the quantum chemistry package Psi4 via the Quantum Chemistry Common Driver and Databases library for electronic structure computations.
  • Dispersion-Corrected HF-3c and Three-Body Dispersion: Lattice energies are computed with the dispersion-corrected HF-3c method and supplemented by an Axilrod-Teller-Muto estimate of three-body dispersion interactions.
  • Accuracy and Convergence: In a benchmark on crystalline benzene including all symmetry-unique dimers and trimers within a 15 Å cutoff, CrystaLattE produced an error of -1.0 kJ mol^-1 relative to the estimated experimental lattice energy.

Scientific Applications:

  • Lattice Energy Prediction: Predict lattice energies of molecular crystals for studies in computational chemistry and materials science.
  • Crystal Stability and Phase Behavior: Analyze crystal stability, phase behavior, and material properties using computed lattice energies.
  • Many-Body Interaction Analysis: Investigate many-body interactions in crystalline systems by including dimers, trimers, and higher-order clusters.

Methodology:

CrystaLattE uses a many-body cluster expansion with independent computations on dimers, trimers, and higher-order clusters, detects identical N-mers via Coulomb-matrix descriptors, interfaces with Psi4 through the Quantum Chemistry Common Driver and Databases library, employs the dispersion-corrected HF-3c method with an Axilrod-Teller-Muto three-body dispersion estimate, and in a demonstrated case included all symmetry-unique dimers and trimers within a 15 Å cutoff; inputs include a crystallographic information file (CIF) and an execution-parameter input file.

Topics

Details

License:
LGPL-3.0
Programming Languages:
Python
Added:
1/9/2020
Last Updated:
12/19/2020

Operations

Publications

Borca CH, Bakr BW, Burns LA, Sherrill CD. CrystaLattE: Automated computation of lattice energies of organic crystals exploiting the many-body expansion to achieve dual-level parallelism. The Journal of Chemical Physics. 2019;151(14). doi:10.1063/1.5120520. PMID:31615262.

PMID: 31615262
Funding: - National Science Foundation: ACI-1449723 - U.S. Department of Defense: HASI