Purely single-bonded spiral nitrogen chains stabilized by trivalent lanthanum ions.

Inspired by the single-bonded nitrogen chains stabilized by tetravalent cerium, pentavalent tantalum, and hexavalent tungsten atoms, we explored the possibility of single-bonded nitrogen polymorphs stabilized by trivalent lanthanum ions. To achieve this, we utilized the crystal structure search method on the phase diagram of binary La-N compounds. We identified three novel thermodynamically stable phases, the C2/c LaN3, P-1 LaN4, and P-1 LaN8. Among them, the C2/c phase with infinite helical poly-N6 chains becomes thermodynamically stable above 50 GPa. Each nitrogen atom in the poly-N6 chain acquires one extra electron, and the spiral chain is purely single-bonded. The C2/c phase has an indirect band gap of ∼1.6 eV at 60 GPa. Notably, the band gap exhibits non-monotonic behavior, decreases first and then increases with increasing pressure. This abnormal behavior is attributed to the significant bonding of two La-N bonds at around 35 GPa. Phonon spectrum calculations and AIMD simulations have confirmed that the C2/c phase can be quenched to ambient conditions with slight distortion, and it exhibits excellent detonation properties. Additionally, we also discovered armchair-like nitrogen chains in LaN4 and the armchair and zigzag-like mixed nitrogen chains in LaN8. These results provide valuable insights into the electronic and bonding properties of nitrides under high pressure and may have important implications for the design and development of novel functional materials.

MAGUS: machine learning and graph theory assisted universal structure searcher.

Crystal structure predictions based on first-principles calculations have gained great success in materials science and solid state physics. However, the remaining challenges still limit their applications in systems with a large number of atoms, especially the complexity of conformational space and the cost of local optimizations for big systems. Here, we introduce a crystal structure prediction method, MAGUS, based on the evolutionary algorithm, which addresses the above challenges with machine learning and graph theory. Techniques used in the program are summarized in detail and benchmark tests are provided. With intensive tests, we demonstrate that on-the-fly machine-learning potentials can be used to significantly reduce the number of expensive first-principles calculations, and the crystal decomposition based on graph theory can efficiently decrease the required configurations in order to find the target structures. We also summarized the representative applications of this method on several research topics, including unexpected compounds in the interior of planets and their exotic states at high pressure and high temperature (superionic, plastic, partially diffusive state, etc.); new functional materials (superhard, high-energy-density, superconducting, photoelectric materials), etc. These successful applications demonstrated that MAGUS code can help to accelerate the discovery of interesting materials and phenomena, as well as the significant value of crystal structure predictions in general.

Prediction of surface reconstructions using MAGUS.

In this paper, we present a new module to predict the potential surface reconstruction configurations of given surface structures in the framework of our machine learning and graph theory assisted universal structure searcher. In addition to random structures generated with specific lattice symmetry, we made full use of bulk materials to obtain a better distribution of population energy, namely, randomly appending atoms to a surface cleaved from bulk structures or moving/removing some of the atoms on the surface, which is inspired by natural surface reconstruction processes. In addition, we borrowed ideas from cluster predictions to spread structures better between different compositions, considering that surface models of different atom numbers usually have some building blocks in common. To validate this newly developed module, we tested it with studies on the surface reconstructions of Si (100), Si (111), and 4H-SiC(11̄02)-c(2×2), respectively. We successfully gave the known ground states, as well as a new SiC surface model, in an extremely Si-rich environment.