SimplySQS: An automated and reproducible workflow for special quasirandom structure generation with ATAT
Journal of Computational Science, 2026
Please see the full published article: https://doi.org/10.1016/j.jocs.2026.102846
Highlights:
Automated SQS workflow with ATAT
Introduces SimplySQS, a reproducible workflow for generating special quasirandom structures using ATAT mcsqs.Interactive browser-based interface
Guides users through structure import, composition setup, supercell definition, and SQS search parameters.Automatic input and script generation
Creates ATAT input files and an all-in-one monitor.sh script for reproducible SQS searches.Reduced manual errors
Standardizes SQS setup, execution, and output analysis to improve reliability and accessibility.Randomness-score assessment
Uses Warren–Cowley (WC) parameters to judge whether random structures are sufficient or SQS optimization is needed.Output analysis and format conversion
Visualizes mcsqs convergence and converts optimized SQSs into POSCAR, CIF, LMP, and XYZ formats.Demonstration on Pb1-xSrxTiO₃
Generates 320-atom SQSs across the full composition range and relaxes them using MACE MATPES-r²SCAN.Cubic-to-tetragonal transition
Captures the experimental phase transition near x ≈ 0.5.Mixing free-energy prediction
Shows that configurational entropy stabilizes the mixed Pb₁₋ₓSrₓTiO₃ phase at 300 K.