[CP2K-user] [CP2K:19830] Re: Problems in `GEO_OPT` of FeS (Mackinawite) Slab

[Ardalan Hayatifar]

Thank you!
I tried centering (different ways - let me know if you'd like the code to 
replace your Excel method) and even using a bigger cell. The system still 
implodes on itself. The convergence cycles show insane positive changes in 
energy.
No matter what I do, it is still imploding (.gif image):

[image: untitled.gif]

On Friday, January 26, 2024 at 1:26:22 PM UTC-6 Станислав Бачурин wrote:

> I think you should give it a try. There's also a rough approach - open the 
> coordinate file in an editor, copy it to a spreadsheet program like Excel, 
> find the most negative value in each column representing its coordinate, 
> and add that value to all the column values. This way, your coordinates 
> will shift in a way that allows you to accurately understand the physical 
> dimensions of the cell. I did this when I couldn't center it through 
> visualization programs.
> пятница, 26 января 2024 г. в 22:16:40 UTC+3, Ardalan Hayatifar: 
>
>> Great catch!
>> In VMD and using the same exact cell dimensions:
>> ```
>> pbc set {18.3675 18.3675 20.1312} -all
>> pbc box
>> ```
>>
>> [image: Screen Shot 2024-01-26 at 1.11.49 PM.png]
>> it is off in XY plane.
>>
>> It seems that in ASE-PyCP2K,
>> ```
>> structure = read('thickness-1-A_1.cif')
>> structure.center(axis=2)
>> ```
>> is not centering anything.
>>
>> Should I try using `&CENTER_COORDINATES` to center it?
>>
>> On Friday, January 26, 2024 at 12:59:00 PM UTC-6 Станислав Бачурин wrote:
>>
>>> Could you check if your system is aligned in the xy plane? Maybe, before 
>>> applying xy periodicity, it needs to be properly oriented.
>>>   Also, it is necessary to set the dimensions along the Z-axis with a 
>>> slight margin in the cell settings until this warning disappears. I 
>>> struggled a lot with calculations under gas-phase conditions. I had 
>>> exactly the same strange behavior - the nuclei almost merged in the organic 
>>> molecule. When I figured out the cell settings and Poisson, everything was 
>>> optimized correctly.
>>> пятница, 26 января 2024 г. в 21:49:57 UTC+3, Ardalan Hayatifar: 
>>>
>>>> Thanks for your comment. I have been trying `PERIODIC XYZ` or `PERIODIC 
>>>> XY` and that doesn't fix it either. 
>>>> Having more layers also shows the same behavior.
>>>>
>>>> On Friday, January 26, 2024 at 12:46:41 PM UTC-6 Станислав Бачурин 
>>>> wrote:
>>>>
>>>>> *** WARNING in pw_env_methods.F:738 :: The selected periodicities in 
>>>>> the *** *** sections &CELL and &POISSON do not match *** You set 
>>>>> periodicity to None, but use the periodic Poisson solver. Try wavelet or 
>>>>> switch to Periodic in your Cell's settings.
>>>>> пятница, 26 января 2024 г. в 21:38:47 UTC+3, Ardalan Hayatifar: 
>>>>>
>>>>>> I am using CP2K version 2023.2. I want to perform `GEO_OPT`on FeS 
>>>>>> layers that I have made using Pymatgen. I am using .cif file from published 
>>>>>> data.
>>>>>>
>>>>>> *Input File:*
>>>>>>
>>>>>> Note: `MAX_SCF` set to 100 results in the same behavior.
>>>>>>
>>>>>> (attached)
>>>>>>
>>>>>> *Problem:*
>>>>>>
>>>>>> This is a convergence problem. Upon visualization I can see that 
>>>>>> things get closer than they should and the system acts crazy.
>>>>>>
>>>>>> *Pictures:*
>>>>>>
>>>>>> Initial:
>>>>>>
>>>>>> [image: Screen Shot 2024-01-26 at 10.46.20 AM.png]
>>>>>>
>>>>>>
>>>>>> After some SCF cycels:
>>>>>>
>>>>>> [image: Screen Shot 2024-01-26 at 10.46.45 AM.png]
>>>>>>
>>>>>> Another representation:
>>>>>>
>>>>>> [image: Screen Shot 2024-01-26 at 10.10.58 AM.png]
>>>>>>
>>>>>> (shown atoms distance ~1A!)
>>>>>>
>>>>>> *What I’ve tried:*
>>>>>>
>>>>>>
>>>>>> - I have changed the slab size in X and Y directions to remove size 
>>>>>> effects. Tried dimensions that are at least 1X1 A and 1.8 X1.8 A dimensions 
>>>>>> - all while having 15 A vacuum in Z
>>>>>>
>>>>>> - I have monitored the input file for close contacts using 
>>>>>> `CHECK_INTERATOMIC_DISTANCES 1`. There are no pair-wise distances below 2.5 
>>>>>> A.
>>>>>>
>>>>>> - PBC effects: I have tried adding 1 A to cell dimensions `ABC` to 
>>>>>> again eliminate possible overlaps, but the behavior persists.
>>>>>>
>>>>>> - I have tried higher cutoff values (e.g. 1200).
>>>>>>
>>>>>> - I have tried writing the input so that it uses a `.xyz` file 
>>>>>> instead of PyCP2K-ASE format and that doesn’t change anything.
>>>>>>
>>>>>> - Centering the solid layer using `&CENTER_COORDINATES` also did not 
>>>>>> help.
>>>>>>
>>>>>>
>>>>>> At this point I’d like to ask for any help or suggestions. It it 
>>>>>> possible that my parameters are wrong?
>>>>>>
>>>>>>
>>>>>> Thanks in advance for your time.
>>>>>>
>>>>>>
>>>>>>
>>>>>>

-- 
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+unsubscribe at googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/50f76365-7b6e-4f7e-903c-e1282a9e961en%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20240126/8f26f9e6/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: untitled.gif
Type: image/gif
Size: 1045075 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20240126/8f26f9e6/attachment-0001.gif>