The total energy in different different K-mesh is constant in dismal. Is this problem related to input file or software problem...?
This is difficult to answer without more information about the system you're simulating, and how you're simulating it. In general, the energy is definitely *not* the same for different k-point meshes, but it's difficult to know what's happening in your case. A few notes:
- The energy *will* be the same if your system only has localised states, for example a molecule in a large simulation cell
- The convergence with respect to k-point mesh is typically oscillatory, but there can be plateaux where errors approximately cancel out. You need to test several k-point meshes to be certain. Also, at some point the k-point mesh is dense enough that the energy has converged, although it is usually very difficult to achieve meV/atom accuracy.
- If you have a very large Fermi-level broadening (smearing width) then the energy will not be very sensitive to the k-point mesh
Finally, have you double-checked that the k-point mesh is changing? If you accidentally changed the k-point mesh for a DOS calculation, for example, rather than the ground state, then the total energy would be calculated on the same mesh as before and so would be the same.
Thank you James, I am attaching here my input file. i hope it will give you enough information about my simulation work.
Usually, when the energetic difference is negligible for different k-mesh, you have reached the convergence regarding this parameter concerning the others.
We know that the k-mesh number is the ideal number in which the energy is constant and has the least value in the sense of the lowest levels of energy , and therefore the larger the number, the clearer the calculation of crystal, electronic and optical properties. Of course, we must pass through optimization.
Jayendra Kagat I'm not a Quantum Espresso expert, but your input file does not look correct to me. The K_POINTS section should have six integers, e.g.
K_POINTS{automatic}
7 7 3 0 0 0
to generate a 7x7x3 Monkhorst-Pack mesh, centred on the origin (the last 3 integers specify whether to offset the grid or not). However, your input has fractional numbers 0.5, which I don't think will work.
Hello Jayant,
I just checked your input file.
K_POINTS{automatic}
0.5 0.5 0.5 0 0 0
If we look at above block then it can be seen that you have used the fractional values. Thats the mistake that you are doing.
give the integers values like 2 2 2 0 0 0, 4 4 4 0 0 0, 6 6 6 0 0 0 and so on.
You will notice that when we are taking less no. of k-points the energy will differ for each value but if you do the same exercise then at some value may be 16 16 16 0 0 0 here you will notice that the energy is almost constant even after increasing the kpoints beyond this.
Try to look for energy convergence wrt to k-points.
I hope you got me.
Thank you all... For help me... I got my solution... And solution is the gap between "K_POINTS" AND {automatic} 😊
sometimes the spaces between blocks may leads to such problems. Good to know that the problem is resolved now.
- Badges
- Science topic
- Similar topics
- Correction