search

aoterodelaroza / critic2

Analysis of quantum chemical interactions in molecules and solids.
Other
100 stars 35 forks source link

typnuc 3, and then the integration is done around the nuclear still #22

Closed Ceasea closed 5 years ago

Ceasea commented 5 years ago

hi @aoterodelaroza , I try to make charge integration of non nuclear basins of ELF.

in my poor opinion, the ELF value at nuclears should be CCP, so i try "TYPNUC 3" keyword after loading the ELFCAR.

"AUTO" command provides the nuclear as CCPs ( ncp is 1 and 2 in the output below).

however, when i use "YT" to do the integration, the charge density of nuclear basins are still integrated. (Keyword noatoms is also considered, but no changes)

is this possible?

here is my input. ........................................................................................................

crystal POSCAR load ELFCAR typnuc 3 load CHGCAR reference 1 integrable 2 root test auto cpeps 1 cpreport short yt nnm wcube

here is part of the output of "AUTO" and "YT". ........................................................................................................ Morse sum: 0 ncp pg type CPname position (cryst. coords.) 1 D4h (3,3 ) cage 0.00000000 0.00000000 0.00000000
2 D4h (3,3 ) cage 0.50000000 0.00000000 0.50000000
3 C4v (3,-3) nucleus 1.00000000 0.50000000 0.22172302
4 D2h (3,-3) nucleus 0.25000000 0.25000000 0.50000000
5 Cs (3,-1) bond 0.19167809 0.30832191 0.35675904
6 D4h (3,-1) bond 1.00000000 0.50000000 0.00000000
7 D2h (3,-1) bond 0.75000000 0.25000000 0.00000000
8 C2v (3,1 ) ring 0.00000000 0.21883710 0.50000000
9 C2v (3,1 ) ring 0.15552983 0.34447017 1.00000000
10 D4h (3,1 ) ring 0.00000000 0.00000000 0.50000000
........................................................................................................

List of attractors integrated Id cp ncp Name Z mult Position (cryst.) 1 1 1 Ti 22 -- 0.0000000 0.0000000 0.0000000 2 2 1 Ti 22 -- 0.5000000 0.5000000 0.0000000 3 3 2 Al 13 -- 0.5000000 0.0000000 0.5000000 4 4 2 Al 13 -- 0.0000000 0.5000000 0.5000000

aoterodelaroza commented 5 years ago

Hi @Ceasea: sorry for the wait. The ELF value at nuclei should indeed be a maximum but if you are using VASP you must be using pseudopotentials/PAW datasets and so it is likely that the atomic shell structure is gone altogether. I recommend plotting a line from any of the nuclei towards the interstitial to see if this is the case.

AUTO is likely not going to help very much in this case because CP localization doesn't work very well with grids, because of #7

Your best bet to find all the maxima in the system is probably the YT keyword, which not only gives you the integral of the charge density in the basins but also the positions of the maxima.

If you want me to take a look more in detail, I will also need the ELFCAR, POSCAR, and CHGCAR.

Ceasea commented 5 years ago

the shell structure is gone as you said. The problem is caused by the pseudopotentials method i used. Thus, we are turning to all-electron calculations.

thanks. your reply helps me a lot.