|
大家好,我目前在做毕设需要使用VASP做一些结构优化和静态计算。其中有在U3Si2中加入一个He原子的结构优化,输出的CONTCAR中U3Si2结构发生了太大变化——原本在同一个面的U与Si原子分离成了两个面,请教导师后被告知是胞太小了,我改用2*2*3的超胞以后,情况好了点,但是在一些结构中仍然出现了两种原子所在平面分离的情况,导师也不知道是怎么回事了,我觉得应该不是胞体大小的问题了,因为看了一些论文后感觉2*2*3的超胞已经足够大了(大概14A*14A*12A),但是我本科不是学的材料物理或者化学,对此也不是很懂,所以来请教大家
Startparameter for this run
PREC = Accurate
ISTART = 0 ! 0-new 1-restart 2-samcut
ICHARG = 2 ! 0 from iniwave; 1 from CHGCAR; 2 from atom; 11 for band calculation
ENCUT = 350 ! 252.50_from OUTCAR
Electronic relaxation
NELM = 80 ! Max of electronic SC steps
NELMIN = 2 ! Min of electronic SC steps
EDIFF = 1.E-06 ! stopping-criterion of Etot for electronic loop
LREAL = Auto ! projection done in real space, before:F
ALGO = Normal ! IALGO=38 in NELMDL then changed to 48, change to normal(IALGO=38) if unconverged, changed to Very_Fast 10-09,before: Normal
GGA = PE
NBANDS = 1260 ! before =80, changed to 1200 2018-10-06
Ionic relaxation
EDIFFG = -0.03 ! stopping-criterion of force for ionic loop (negative)
NSW = 500 ! ionic steps
IBRION = 2 ! 1-quasi-Newton(if has good guess of initial positions) 2-CG 0-MD
ISIF = 2 ! 3 for CG(calculate stress tensor and cell change); 0 for MD(no stress tensor and no cell change)
POTIM = 0.25 ! for CG it's relative with the length of trial step; for MD it's time step in fs
DOS related values
# EMIN = 0
# EMAX = 10
# NEDOS = 800
# ISMEAR = 0 ! -5 for semiconductors or insulators, 0 for large cell or few k-points, 1 for metal and set SIGMA
# SIGMA = 0.05
Write flags
LWAVE = F
LCHARG = F
LORBIT = 12
NPAR = 4
附图是POSCAR以及产出的CONTCAR的侧面视图,可以发现很明显的变形,附图的是在超胞中弄一个Si原子缺陷,然后把He原子填补到那个Si缺陷的位置的结构优化先谢谢大家了
|
-
CONTCAR.png
(107.82 KB, 下载次数 Times of downloads: 36)
CONTCAR
-
POSCAR.png
(99.01 KB, 下载次数 Times of downloads: 29)
|