Machine-Intelligence-Driven High-Throughput Prediction of 2D Charge Density Wave Phases

Charge density wave (CDW) materials are an important subclass of two-dimensional materials exhibiting significant resistivity switching with the application of external energy. However, the scarcity of such materials impedes their practical applications in nanoelectronics. Here we combine a first-principles-based structure-searching technique and unsupervised machine learning to develop a fully automated high-throughput computational framework, which identifies CDW phases from a unit cell with inherited Kohn anomaly. The proposed methodology not only rediscovers the known CDW phases but also predicts a host of easily exfoliable CDW materials (30 materials and 114 phases) along with associated electronic structures. Among many promising candidates, we pay special attention to ZrTiSe₄ and conduct a comprehensive analysis to gain insight into the Fermi surface nesting, which causes significant semiconducting gap opening in its CDW phase. Our findings could provide useful guidelines for experimentalists.