[CP2K-user] Inconsistent cell dimensions from cell optimization with TZV2P-MOLOPT-GTH basis set

[Matthew Okenyi]

Hello,

I have recently carried out a cell optimization of a (3,2,1) supercell of 
croconic acid (see croc_uc.xyz) with CP2K v6.1. I have attached the input 
file as crca_20181229_112044_000.inp and linked the output files (in 2 
parts) crca_20181229_112044_000.out 
<https://drive.google.com/file/d/1x7jmZEkIpeROmqoBn-2B8t6N6XLoCgxv/view?usp=sharing> 
and crca_20181229_112044_000_restarted.out 
<https://drive.google.com/file/d/1x7jmZEkIpeROmqoBn-2B8t6N6XLoCgxv/view?usp=sharing>. 
I am having problems with the values that CP2K returned for the relaxed 
cell dimensions, however. These were [25.474, 10.328, 10.866]. I thought 
that this implied unit cell dimensions of [8.491, 5.164, 10.866]; however, 
the displacement between corresponding atoms in unit cells that are 
alongside each other in the x-direction within the supercell is [8.721, 
0.0, 0.0]. Similarly, the displacement between corresponding atoms in unit 
cells that are alongside each other in the y-direction is [0.0, 5.274, 0.0].

These do not appear to be consistent with the relaxed cell dimensions 
produced by the CP2K cell optimization. To investigate further, I generated 
two (6,2,1) supercells made by joining two (3,2,1) supercells in the 
x-direction. (6,2,1) supercell A has the second (3,2,1) supercell displaced 
by [25.474, 0.0, 0.0] from the first (i.e. the a-dimension given by CP2K's 
cell optimization), while (6,2,1) supercell B has the second (3,2,1) 
supercell displaced by [26.163, 0.0, 0.0] = 3*[8.721, 0.0, 0.0] from the 
first. For supercell A, at the boundary between the (3,2,1) supercells, the 
O-H bonds are vertical, instead of diagonal to be consistent with the 
crystal structure. For supercell B, at O-H bonds at the boundary are 
diagonal and consistent. I've attached the xyz files and VESTA screenshots 
of the 2 supercells.

Could somebody, please, explain why the supercell dimensions given in the 
cell optimization output are not consistent with the true cell dimensions? 
Or am I misinterpreting these values? I have done cell optimizations with 
both the TZV2P-MOLOPT-GTH and DZVP-GTH-PBE basis sets and have only 
experienced this problem with TZV2P-MOLOPT-GTH.

Best wishes,
Matthew

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20181231/9b3344c9/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: supercell_A.png
Type: image/png
Size: 91271 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20181231/9b3344c9/attachment.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: supercell_A.xyz
Type: chemical/x-xyz
Size: 28848 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20181231/9b3344c9/attachment.xyz>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: supercell_B.xyz
Type: chemical/x-xyz
Size: 28848 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20181231/9b3344c9/attachment-0001.xyz>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: crca_20181229_112044_000.inp
Type: chemical/x-gamess-input
Size: 13952 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20181231/9b3344c9/attachment.inp>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: supercell_B.png
Type: image/png
Size: 99863 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20181231/9b3344c9/attachment-0001.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: croc_uc.xyz
Type: chemical/x-xyz
Size: 1969 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20181231/9b3344c9/attachment-0002.xyz>