Structural relaxation does not converge for CO/Rh(111) although SCF converges

[samomoyed]

Details

Problem

I am performing a structural relaxation calculation for a CO/Rh(111) system using ABACUS.

The SCF calculation converges normally, but the structural relaxation does not converge. I have already increased relax_nmax to 400 steps, but the largest force remains significantly larger than the force threshold of 0.02 eV/Angstrom.

I noticed that the C and O atoms have relatively large forces in the z direction.

For example, near the end of the relaxation:

C1 Fz = -0.151966 eV/Angstrom
O1 Fz = 0.091577 eV/Angstrom

The k-point mesh is 8 × 6 × 1.

I would like to know whether this behavior is more likely caused by:

1. insufficient k-point sampling,
2. an unreasonable initial adsorption structure, or
3. some other setting in the relaxation calculation.

The SCF calculations converge normally, but the relaxation does not.

I have attached the INPUT, STRU, and running_relax.log files for reference.

running_relax.log
STRU.txt
INPUT.txt

Have you read FAQ on the online manual http://abacus.deepmodeling.com/en/latest/community/faq.html

• Yes, I have read the FAQ part on online manual.

Task list for Issue attackers (only for developers)

• Understand the problem or question described by the user.
• Check if the issue is a known problem or has been addressed in the documentation.
• Test the issue or problem on a similar system or environment, if possible.
• Identify the root cause or provide clarification on the user's question.
• Provide a step-by-step guide, including any necessary resources, to resolve the issue or answer the question.
• If the issue is related to documentation, update the documentation to prevent future confusion (optional).
• If the issue is related to code, consider implementing a fix or improvement (optional).
• Review and incorporate any relevant feedback from users or developers.
• Ensure the user's issue is resolved or their question is answered and close the ticket.

[samomoyed]

Following the suggestions, I made the following changes:

1. I included the relevant input and output files for reference, including INPUT, STRU, KPT, pseudopotential/orbital files, and running_relax.log.

2. I changed the relaxation method from cg to `bfgs

issue.zip

[QuantumMisaka]

@samomoyed Thanks for your issue!

1. it is recommended to use bfgs (in LTS: bfgs_trad) for fixed-cell relaxation especially surface ads
2. it is recommended to change the vacuum direction to X(A) or Y(B) but not C(Z) for better efficiency