The crystal structure diverged during Relaxation of both atomic positions and the shape of the unit cell using QHA

[dha2025]

Discussed in #240

^{Originally posted by dha2025 December 5, 2024}
Dear Dr. Tadano,

Recently, I installed ALAMODE to study the temperature dependence of lattice parameters in anisotropic systems using QHA and SCPH. I followed the steps provided for ZnO and BaTiO3 in the ALAMODE website. I successfully calculated the anharmonic interatomic force constants (IFCs) and performed structural optimization (relaxing atomic positions) without any errors using both methods.

However, when I enabled the “Relax both atomic positions and the shape of the unit cell” tag, I encountered the following warnings and the structure diverged:

1. Warning in compute_anharmonic_del_v0_del_umn: squared SCP frequency is negative. ik = 3
2. Warning in compute_anharmonic_v1_array: 0 acoustic modes are detected at the Gamma point.

The crystal structure diverged in both QHA and SCPH calculations, even at 10 K. I reviewed the additional input files and attempted to resolve the issue by adjusting the cluster size and increasing the number of configurations to generate IFCs.

Could you please provide insights into the possible causes of this problem and suggest potential solutions?

Thank you in advance for your assistance.