Matlantis has introduced Matlantis CSP, a groundbreaking addition to its universal atomistic simulator, aimed at rapidly identifying previously unknown stable crystal structures. This capability addresses the challenges of decarbonisation and next-generation energy by reducing the dependency on repeated synthesis experiments with low success rates. Matlantis CSP integrates computational screening early in the research process, enabling teams to eliminate implausible options and concentrate on the most promising candidates.
Honda R&D has embraced Matlantis CSP to enhance exploration efficiency in materials development, particularly in multi-component systems and metastable structures, which have traditionally been challenging to assess due to computational demands.
Crystal structure prediction has been limited by persistent barriers such as time-consuming DFT-based evaluations, biased search processes, and the need for complex setups. Matlantis CSP overcomes these by leveraging its machine-learning interatomic potential, PFP, alongside proprietary algorithms and parallel processing optimized for large-scale CSP. This innovation allows for high-throughput structure evaluation, significantly improving search efficiency and enabling comprehensive exploration of composition spaces.
Mitsumoto Kawai, Chief Engineer at Honda R&D, expressed high expectations for CSP’s potential to enhance exploration efficiency in materials development. CSP’s ability to feasibly search crystal structures, including complex systems, promises to increase the likelihood of achieving next-generation materials while reducing development timelines.
Matlantis CSP has already demonstrated success across various systems, discovering over 10 new stable crystals. In the Ga–Au–Ca system alone, it identified 13 new crystals, significantly updating the phase diagram. Matlantis CEO Daisuke Okanohara highlighted CSP’s practicality at research scale, allowing for earlier identification of promising candidates and reducing time spent on low-probability experiments.
