I've been using SIESTA (with VIBRA) lately to calculate the vibrational frequencies of brucite (Mg(OH)2) at the gamma point. When I compare the results with some I have using CRYSTAL06, not only are some of the frequencies very different, but there is no degeneracy in the SIESTA calculations (see below)
MODE SIESTA CRYSTAL06 1 -0.0842 -6.8266 2 0.0296 -1.2079 3 0.0517 -1.2079 4 291.9689 309.1325 5 293.0351 309.1325 6 334.8122 326.3867 7 338.7810 326.3867 8 415.6654 405.3074 9 475.6206 405.3074 10 643.3266 483.2689 11 644.7176 500.6947 12 917.5655 994.5470 13 924.3611 994.5470 14 3395.2587 3514.1542 15 3469.0307 3611.5673 While some are very near degenerate, others have > 10% splitting, which seems very large to me despite the fact that SIESTA does not use symmetry. I have used atomic displacements of .04, .004, and .001 with only very little change in results. I have attached my .fdf's for both SIESTA and VIBRA. Are these discrepancies due to my misuse or is this a numerical inevitability in SIESTA? Does anyone have any advice? Thanks, Ben
bru_vib.fdf
Description: Binary data
bru_vib-siesta.fdf
Description: Binary data