Problem of large rotation angle structures of twisted bP

[Xinyu Zhang]

Dear all,

Thanks for your helpful software. I'm facing one problem in generating large rotation angle twisted bP structures or other orthorhombic structures.

When generating such structures, the generated superlattices always contain some addtional atoms that overlap with other atoms.

For instance, open the example/TwistedbP/Ang_5-6 and edit the get_ang.inp, change the theta_range as 30.0 35.0 0.1 to find a structure whose rotation angle is between 30° and 35°. Then running the calculation will find a solution and running the twister.py will lead to a structure with 557 atoms rather than expected 552 atoms. Some output information is as below:

Generating a 52 x 52 x 1 supercell
Generating a 52 x 52 x 1 supercell
Total number of atoms in superlattice: 557
Atom positions written to file: superlattice.dat
Number of atoms in layer 1: 279
Number of atoms in layer 2: 278
Areas, uc l1: 15.343020, uc l2: 15.343020, sl: 1058.668380
Expected number of atoms in layer 1 from area ratios: 276
Expected number of atoms in layer 2 from area ratios: 276
Time taken:         0.72 s

Transfer the superlattice to POSCAR format and open it in VESTA, we can observe the stucture have 5 addtional atoms that overlap other atoms: P522 P520 P279 P277 P262. The overlap will lead to subsequent problems in molecular dynamics. The POSCAR file is attached.

I want to know why the twister.py would lead to such structrues and how to get the correct twisted structures.

I would greatly appreciate your help.

Best regards,

Xinyu Zhang

SIGS, THU