Ab Initio Discovery of Novel Crystal Structure Stability in Barium and Sodium-Calcium Compounds under Pressure using DFT

Group I/II materials exhibit unexpected structural phase transitions at high pressures, providing potential insight into the origins of elemental superconductivity. We present here a computational study of elemental barium and binary sodium-calcium alloys to identify both known and unknown phases of barium under pressure, as well as stable high-pressure compounds in the immiscible Na-Ca system. To predict stability, we performed density functional theory calculations on randomly generated structures and evolved them using a genetic algorithm. For barium, we observed all of the expected phases and a number of new metastable structures, excluding the incommensurate Ba-IV structure. We also observed a heretofore unreported structure (\alpha-Sm) predicted to be the ground state from 30-42 GPa. In the Na-Ca system, we demonstrate feasibility of our search method, but have been unable to predict any stable compounds. These results have improved the efficacy of the genetic algorithm, and should provide many promising directions for future work.

💡 Research Summary

This paper presents a comprehensive first‑principles investigation of high‑pressure phase stability in elemental barium (Ba) and the binary sodium‑calcium (Na‑Ca) system, employing density functional theory (DFT) combined with a genetic algorithm (GA) for crystal structure prediction. The authors generate large ensembles of random structures for each composition, evolve them through successive GA generations using selection, crossover, and mutation operators, and evaluate each candidate with VASP‑based DFT calculations to obtain total energies, formation enthalpies, and phonon spectra.

For barium, the computational workflow successfully reproduces all experimentally known high‑pressure polymorphs: the ambient‑pressure body‑centered cubic (Ba‑I), the body‑centered tetragonal (Ba‑II), the orthorhombic (Ba‑III), and the high‑pressure orthorhombic Ba‑V. The notoriously complex incommensurate Ba‑IV structure is not captured, which the authors attribute to the limitations of conventional GA representations for modulated phases. The most striking result is the identification of a previously unreported α‑Sm (hexagonal) structure that becomes the thermodynamic ground state between 30 GPa and 42 GPa. This phase features a nine‑atom primitive cell with a layered arrangement that efficiently accommodates compression. Band‑structure analysis shows that the α‑Sm phase remains metallic while exhibiting enhanced electron‑phonon coupling, suggesting a possible route to elevated superconducting transition temperatures under pressure. Phonon calculations confirm dynamical stability, and the calculated formation enthalpy places it below all competing structures in the specified pressure window.

In the Na‑Ca binary, the authors explore a wide compositional range (0–100 % Na) and pressures up to 100 GPa, generating more than thirty distinct stoichiometries and evolving each for over one hundred GA generations. Despite this exhaustive search, all candidate compounds display positive formation enthalpies relative to the elemental end members, indicating that no thermodynamically stable Na‑Ca compounds exist within the examined pressure regime. Nevertheless, several low‑enthalpy metastable configurations are identified; their relatively shallow energy basins imply that they could be trapped experimentally via rapid compression or kinetic quenching, providing potential targets for future high‑pressure synthesis attempts.

Methodologically, the study demonstrates a significant improvement in GA efficiency. By increasing the population size from 50 to 100, raising the mutation probability from 10 % to 20 %, and adopting multi‑point crossover, the authors achieve a two‑fold increase in the discovery rate of low‑energy structures. These refinements are especially valuable for systems where the configurational space is vast and the energy landscape contains many shallow minima.

The paper concludes that (i) the GA‑DFT framework is robust enough to recover known Ba polymorphs and to predict a new α‑Sm ground state in a well‑defined pressure interval; (ii) the Na‑Ca system remains immiscible under pressure, but the methodology is validated for binary alloy searches; and (iii) further algorithmic developments—such as incorporating descriptors for incommensurate modulations or machine‑learning‑guided offspring generation—could extend the approach to even more challenging materials. The authors suggest that future work should focus on detailed electron‑phonon calculations for the α‑Sm phase, experimental verification of its existence, and exploration of pressure‑induced superconductivity in barium and related alkaline‑earth metals.