Hello Md. Torikul Islam

What DFT code did you use? And could you show your input file? I am not sure if I will be able to help you with this, I use Quantum ESPRESSO and I haven't used it for heterojunctions, but many conversion problems are related to bad input construction.

Hi Md. Torikul Islam

The output info is somewhat hard to follow, but apparently the forces are still large, indicating the system hasn’t found its ground state yet. Consider sharing which DFT code you’re using and your inputs, as suggested by Ricardo Tadeu Maia

Best,

Gabriel Vinicius

Dear Md. Torikul Islam

After reviewing the output data you shared, it appears that the atomic forces are already within a reasonably acceptable range (below 0.05 eV/Å). However, if you’re targeting a tighter force convergence threshold, I recommend considering the following adjustments:

- Mixing Beta:

Lowering the mixing_beta value (e.g., < 0.7) can enhance convergence stability.

- Diagonalization Scheme:

Switching to c-g (conjugate gradient) diagonalization is more robust than David, although it tends to be slower.

- K-point Grid:

Using a denser k-point mesh can improve accuracy, especially for small unit cells.

- Pseudopotentials:

Make sure you are using pseudopotentials from the SSSP library (Efficiency or Precision version). After selecting appropriate PPs, verify and apply the recommended ecutwfc and ecutrho values found in the pseudopotential files.

Once you’ve applied these refinements, I suggest rerunning the relaxation and allowing it to continue until it naturally reaches the ground-state configuration.

Let me know if you’d like help modifying the input or checking the convergence trend.

Best regards,

Idriss

@Recardo Tadeu maia

I have used quantum espesso DFT code.Here is input relax file

@Idriss Mohdeb

Dear @Idriss Mohdeb

For your more information, input relax file is attached here

Dear Md. Torikul Islam

I took a look at your input file, and below are my comments and suggestions that I believe may help you reach your desired force threshold:

- Dipole Correction: Since you’re working with an asymmetric slab—specifically a heterojunction between 2D materials—I strongly recommend applying a dipole correction to eliminate any artificial electric field across the vacuum.

- Plane-Wave Cutoff: If you’re using PAW pseudopotentials, consider increasing the ecutwfc value to somewhere between 60–90 Ry. This often improves force accuracy and stability.

- Mixing Mode: Based on my experience, using the mixing_mode = 'local-TF' can lead to more stable self-consistent convergence, especially for complex systems.

- Convergence Threshold: Finally, keep in mind that a force convergence threshold below 0.05 eV/Å is already considered decent and acceptable for most publications.

Hope these points help!

Best regards,

Idriss