My Community

Hello,

My question is about the relaxation when a small molecule interactng with metal surface. For example, when I optimize Pt111 surface, it would be easier to get the force less than 0.001ev/angstrom, but when I attach a molecule, e.g.
NH3BH3, the optimization seems stuck.
The INCAR and POSCAR can be seen below.

Any opinion? Thank you in advance.

Burns

P. S.

schematic graph:

N
/
B
PtPtPtPt
PtPtPtPt
PtPtPtPt
(H atoms are omitted)

INCAR:
PREC = accurate
ISTART = 0
ICHARG = 2
GGA = 91
ISYM = 0
ALGO = Fast
NELMIN = 5
LREAL = .TRUE.
ENCUT = 450
ENAUG = 550
EDIFF = 0.1E-5
EDIFFG = -0.01
IBRION = 1
NFREE = 10
NSW = 100
NBLOCK = 1
POTIM = 0.5
VOSKOWN = 1

KPOINTS: 1x2x2

POSCAR:
Pt111
1.00000000000000
8.389436 8.389436 8.389436
5.885849 -5.885849 0.000000
0.000000 5.885849 -5.885849
27 1 1 6
Selective dynamics
Direct
0.155906 0.282652 0.162207 F F F
0.155906 0.282652 0.495541 F F F
0.155906 0.282652 0.828874 F F F
0.155906 0.615985 0.162207 F F F
0.155906 0.615985 0.495541 F F F
0.155906 0.615985 0.828874 F F F
0.155906 0.949318 0.162207 F F F
0.155906 0.949318 0.495541 F F F
0.155906 0.949318 0.828874 F F F
0.317612 0.053237 0.047838 T T T
0.317458 0.052810 0.383574 T T T
0.319590 0.055266 0.716852 T T T
0.318156 0.389019 0.048936 T T T
0.327029 0.386575 0.383503 T T T
0.317951 0.389113 0.718853 T T T
0.320248 0.721703 0.049989 T T T
0.318911 0.722932 0.383448 T T T
0.318871 0.723181 0.717919 T T T
0.000000 0.171541 0.273318 F F F
0.000000 0.171541 0.606652 F F F
0.000000 0.171541 0.939985 F F F
0.000000 0.504874 0.273318 F F F
0.000000 0.504874 0.606652 F F F
0.000000 0.504874 0.939985 F F F
0.000000 0.838207 0.273318 F F F
0.000000 0.838207 0.606652 F F F
0.000000 0.838207 0.939985 F F F
0.507322 0.298996 0.354654 T T T
0.589051 0.451423 0.418145 T T T
0.440285 0.424251 0.380742 T T T
0.505917 0.240126 0.459767 T T T
0.515181 0.191758 0.191800 T T T
0.583126 0.553657 0.554264 T T T
0.651161 0.386559 0.403546 T T T
0.591577 0.514074 0.338978 T T T

1) please start with an initial structure as good as possible (ie using experimental data like spectroscopy,...)
2) if you are not sure about the quality of your initial gues, use IBRION 2 or 3 (1 works best in the region close to the minimum).
3) please always use 1 k-point in the direction across the surface normal ! Therefore choose your Bravais matrix such that the surface normal is eg along one of the carthesian axes.