[CP2K-user] how to choose the size of the cell correctly?

[Travis]

Hi,

Switch to non-periodic boundaries (PERIODIC NONE in both &CELL and &POISSON 
sections, use the WAVELET solver in &POISSON, and add 

&CENTER_COORDINATES
&END CENTER_COORDINATES

to &TOPOLOGY section. The molecule will be default centered at L/2 where L 
is the box length in a given direction. Add a 6 Angstrom buffer in your 
&CELL section between the molecule and the box edge on each side: top, 
bottom, left, right, front, back. If you need to increase the buffer 
distance, you'll see a warning printed at each SCF iteration that says the 
electron density isn't close enough to zero at the box edge. Just increase 
your buffer in 1 Angstrom increments until the error no longer shows up. 
You don't want to immediately set the box size to be very large as the 
vacuum is filled with plane waves and so adds considerable computational 
expense. You can have different size a, b, and c lattice vectors for such a 
calculation.

-T

On Saturday, January 30, 2021 at 5:44:03 PM UTC-5 mal... at gmail.com wrote:

> Hello everyone!
>  I'm trying to calculate homo lumo gap of single walled carbon nanotubes, 
> for the &SUBSYS &CELL I'm choosing the ABC as following: add 2 angstroms to 
> the longest distance in each dimension. Am I doing everything right?
> I attached my input file and the structure of the CNT (9,0) which i'm 
> studying now.
> best regards,
> Malin 
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20210131/d4f2db01/attachment.htm>