the problem of calcuate dipolemoment of layered system
Posted: Tue Dec 21, 2010 8:04 am
Dear VASP head administrator:
I am calculating the dipole moment of a layered composite system. It seems that there are some uncertainties when VASP deals with the dipole moment calculation.
Our system is n layers of graphene on Ni(111) surface. Before this, we tried a much simpler case -- the HF molecule and successfully got a reasonable value (0.35 el*Angstrom) of dipole moment.
Howerver，when it goes to our complicated system, the value gets wired. Here is what I did step by step:
(1) Self-consistent calculation.
========== INCAR ============
SYSTEM = Ni-C surface
ENCUT = 380
ISTART = 0
ISPIN = 2
ISMEAR = -5
ICHARG = 2
EDIFF = 1E-7
PREC = High
NELM=200
============================
and poscar,
=================== POSCAR:======================
4Ni-graphene
2.48901587
0.866025404 -0.500000 0.00000000
0.866025404 0.500000 0.00000000
0.000000000 0.000000 12.05295650
4 10
Cartesian
0.00000000 0.000 0.0000000000 #Ni
0.577350269 0.000 0.8164953970 #Ni
1.15470054 0.000 1.6256505400 #Ni
1.73205081 0.000 2.4336003900 #Ni
1.73205081 0.000 3.2861461800 #1C
2.30940108 0.000 3.2861461800 #1C
2.30940108 0.000 4.6320596600 #2C
2.88675135 0.000 4.6320596600 #2C
2.88675135 0.000 5.9779731300 #3C
3.46410162 0.000 5.9779731300 #3C
1.73205081 0.000 7.3238866100 #4C
2.30940108 0.000 7.3238866100 #4C
2.30940108 0.000 8.6698000800 #5C
2.88675135 0.000 8.6698000800 #5C
==================================================
(2) Non-self consistent calculation.
======= INCAR ==========
SYSTEM = Ni-C surface
ENCUT = 380
ISTART = 1
ISPIN = 2
ISMEAR = -5
ICHARG = 11
IDIPOL = 3
DIPOL = 0 0 zEDIFF = 1E-7
PREC = High
NELM=200
=========================
According to the VASP manual, for surface only the component normal to the surface is meaningful, only z direction is considered. Using the tags: ‘DIPOL = center of cell (in direct coordinates)’, we defined different center of the cell (0 0 z). However, the value of dipole moment varies dramatically with the option of z value, as shown in the table below,
z----- dipole (el*Ang)
0.070 4.576176
0.080 12.183869
0.088 29.701620
0.091 38.923637
0.094 50.129609
0.097 63.314839
0.100 78.310293
0.110 -110.230370
0.200 31.971017
0.250 -45.747171
0.270 -1.679887
0.700 134.440373
0.800 -15.121036
0.900 -43.481955
1.000 -97.462555
1.100 78.308795
Our question is:
(a) Why does the dipole change so abruptly with z value?
(b) How to choose an appropriate position of z? The VASP suggests the center of mass along z direction, then how to compare dipole moment from system to system if the center of mass changes with number of layers?
(c) The KPOINTS was chosen close to the limit of VASP, namely 37*37*1, is there any space to improve the precision of this kind of calculation?
Any suggestion is appreciated. Thanks a lot!
Sincerely,
HHGUO
I am calculating the dipole moment of a layered composite system. It seems that there are some uncertainties when VASP deals with the dipole moment calculation.
Our system is n layers of graphene on Ni(111) surface. Before this, we tried a much simpler case -- the HF molecule and successfully got a reasonable value (0.35 el*Angstrom) of dipole moment.
Howerver，when it goes to our complicated system, the value gets wired. Here is what I did step by step:
(1) Self-consistent calculation.
========== INCAR ============
SYSTEM = Ni-C surface
ENCUT = 380
ISTART = 0
ISPIN = 2
ISMEAR = -5
ICHARG = 2
EDIFF = 1E-7
PREC = High
NELM=200
============================
and poscar,
=================== POSCAR:======================
4Ni-graphene
2.48901587
0.866025404 -0.500000 0.00000000
0.866025404 0.500000 0.00000000
0.000000000 0.000000 12.05295650
4 10
Cartesian
0.00000000 0.000 0.0000000000 #Ni
0.577350269 0.000 0.8164953970 #Ni
1.15470054 0.000 1.6256505400 #Ni
1.73205081 0.000 2.4336003900 #Ni
1.73205081 0.000 3.2861461800 #1C
2.30940108 0.000 3.2861461800 #1C
2.30940108 0.000 4.6320596600 #2C
2.88675135 0.000 4.6320596600 #2C
2.88675135 0.000 5.9779731300 #3C
3.46410162 0.000 5.9779731300 #3C
1.73205081 0.000 7.3238866100 #4C
2.30940108 0.000 7.3238866100 #4C
2.30940108 0.000 8.6698000800 #5C
2.88675135 0.000 8.6698000800 #5C
==================================================
(2) Non-self consistent calculation.
======= INCAR ==========
SYSTEM = Ni-C surface
ENCUT = 380
ISTART = 1
ISPIN = 2
ISMEAR = -5
ICHARG = 11
IDIPOL = 3
DIPOL = 0 0 zEDIFF = 1E-7
PREC = High
NELM=200
=========================
According to the VASP manual, for surface only the component normal to the surface is meaningful, only z direction is considered. Using the tags: ‘DIPOL = center of cell (in direct coordinates)’, we defined different center of the cell (0 0 z). However, the value of dipole moment varies dramatically with the option of z value, as shown in the table below,
z----- dipole (el*Ang)
0.070 4.576176
0.080 12.183869
0.088 29.701620
0.091 38.923637
0.094 50.129609
0.097 63.314839
0.100 78.310293
0.110 -110.230370
0.200 31.971017
0.250 -45.747171
0.270 -1.679887
0.700 134.440373
0.800 -15.121036
0.900 -43.481955
1.000 -97.462555
1.100 78.308795
Our question is:
(a) Why does the dipole change so abruptly with z value?
(b) How to choose an appropriate position of z? The VASP suggests the center of mass along z direction, then how to compare dipole moment from system to system if the center of mass changes with number of layers?
(c) The KPOINTS was chosen close to the limit of VASP, namely 37*37*1, is there any space to improve the precision of this kind of calculation?
Any suggestion is appreciated. Thanks a lot!
Sincerely,
HHGUO