H2O vibration: Difference between revisions
No edit summary |
No edit summary |
||
Line 41: | Line 41: | ||
== Calculation == | == Calculation == | ||
*It is strongly recommended to set the energy cutoff manually in the INCAR file, | |||
as it gives you more control over the calculations and forces you to think about | |||
this aspect. | |||
*For the ionic optimisation the DIIS algorithm is used. | |||
This algorithm builds an approximation of the Hessian matrix and usually | |||
converges faster than the conjugate gradient algorithm. | |||
It is, however, recommended to specify the number of independent | |||
degrees-of-freedom manually. | |||
How many zero frequency modes should be observed and why? | How many zero frequency modes should be observed and why? |
Revision as of 14:17, 8 May 2017
Task
Calculation of the vibrational frequencies of a molecule.
Input
POSCAR
H2O _2 1.0000000 8.0000000 0.0000000 0.0000000 0.0000000 8.0000000 0.0000000 0.0000000 0.0000000 8.0000000 1 2 cart 0.0000000 0.0000000 0.0000000 0.5960812 -0.7677068 0.0000000 0.5960812 0.7677068 0.0000000
INCAR
SYSTEM = H2O vibration PREC = A # IBRION = 1 ; NSW = 10 ; NFREE = 2 ; EDIFFG = -1E-4 ENMAX = 400 ISMEAR = 0 # Gaussian smearing IBRION = 6 # finite differences with symmetry NFREE = 2 # central differences (default) POTIM = 0.015 # default as well EDIFF = 1E-8 NSW = 1 # ionic steps > 0
KPOINTS
Gamma-point only 0 Monkhorst Pack 1 1 1 0 0 0
Calculation
- It is strongly recommended to set the energy cutoff manually in the INCAR file,
as it gives you more control over the calculations and forces you to think about this aspect.
- For the ionic optimisation the DIIS algorithm is used.
This algorithm builds an approximation of the Hessian matrix and usually converges faster than the conjugate gradient algorithm. It is, however, recommended to specify the number of independent degrees-of-freedom manually.
How many zero frequency modes should be observed and why? Try to use the linear response code (IBRION=8 and EDIFF=1E-8) to obtain reference results. For finite differences, are the results sensitive to the step width POTIM. In this specific case, the drift in the forces is too large to obtain the zero frequency modes "exactly", and it is simplest to increase the cutoff ENCUT to 800 eV. The important and physically meaningful frequencies are, however, insensitive to the choice of the cutoff.
Download
To the list of examples or to the main page