Bulk nickel optimization cannot converge under specific multiplicity

133 views
Skip to first unread message

GENG YUAN

unread,
May 3, 2023, 1:58:56 PM5/3/23
to cp2k
Dear all,

I am trying to optimize a bulk nickel with 96 atoms. As Ni is supposed to be ferromagnetic, I performed the optimization with a multiplicity of 96*2+1=193 but the energy couldn't converge within 500 SCF steps. However, I found that it could converge if I only change the multiplicity to its default value (1). 

My inputs and outputs for both cases are attached, the only difference between these two cases is the multiplicity in the input file (one is 1 while the other is 193). I am wondering why the optimization couldn't converge with the specific multiplicity and I would appreciate any ideas on how to figure out this issue.

Many thanks,
Geng
Multiplicity1.inp
Multiplicity1.out
Ni96.xyz
Multiplicity193.out
Multiplicity193.inp

Krack Matthias

unread,
May 4, 2023, 3:11:14 AM5/4/23
to cp...@googlegroups.com

Hi

 

Did you try already a larger smearing temperature than 500 K?

 

Matthias

--
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+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/8c533b46-b7bd-469b-8e31-4b6558f7dd13n%40googlegroups.com.

GENG YUAN

unread,
May 4, 2023, 11:36:24 AM5/4/23
to cp2k
Hi Matthias, 

Many thanks for your advice, yes, I tried higher temperature at 1000 and 1500 K for both cases. 

Still, the optimization with default value of multiplicity (1) could converge at either 1000K or 1500K, but it couldn't converge with multiplicity of 193 under these higher temperatures.

Should I try much higher temperatures or please let me know if you have any other thought.

Sincerely,
Geng

Krack Matthias

unread,
May 4, 2023, 11:50:11 AM5/4/23
to cp...@googlegroups.com

Hi Geng

 

Did you try to add k points already, because you have bulk metal system?

 

HTH

GENG YUAN

unread,
May 10, 2023, 10:48:29 AM5/10/23
to cp2k
Hi Matthias, 

I added k points (I tried two sets of k points: 2x2x1 & 3x3x2) to both cases (Multiplicity =1 &193) for my bulk Ni. I also played with different smearing temperatures (500 & 1000 & 1500 K) for each simulation with specified k points.

My simulations with Multiplicity = 1 always converged no matter what combination of settings, however, none of my simulations converged with Multiplicity = 193. Therefore, my current understanding is that the k points may not be the issue. What else do you think I can give a shot?

My inputs and outputs with k points & higher smearing temperature under multiplicity 1 & 193 are attached. 

Many thanks,
Sincerely,
Geng

M1_KP_Temp.inp
M193_KP_Temp.inp
M193_KP_Temp.out
M1_KP_Temp.out

Krack Matthias

unread,
May 10, 2023, 1:05:19 PM5/10/23
to cp...@googlegroups.com

Hi Geng

 

Maybe, something is wrong with your atomic coordinates which you did not reveal.

GENG YUAN

unread,
May 12, 2023, 11:04:51 AM5/12/23
to cp2k
Hi Matthias,

Many thanks for the suggestion, I will carefully check my atomic coordinate. 

Please allow me to ask you a question following this point about coordinates: If that's the case, why would all my simulations converge when the multiplicity is 1?

Sincerely,
Geng

Reply all
Reply to author
Forward
0 new messages