I am a beginning with computation, and this is my first time working with an organic molecule.I have attached my input and output files. I was able to get the organic portion of the molecule converged, but I need a sulfonated version, and added SO3 where I need to (bottom of the ATOMIC POSITIONS name-card). For some reason, when I add the sulfite, the molecule fails to converge after 100 iterations. For a molecule this large, it finishes in a several minutes which is abnormally fast. I original had the mixing_beta and diagonalization set to the quantum espresso defaults, but changing them did not allow it to converge. I also tried increasing the electron_maxstep to 200, but that did not help either. Any suggestions I could attempt?
P.S. I wanted a negative charge on one of my oxygen atoms, and after adding tot_charge=-1, it gave me an error saying smearing was required and that my electron count was incorrect. I do not know why that is the case, but I disregarded that in the mean time. If anyone knows why that is occurring, I'd appreciate some input on that as well.
In the input file celldm(1) = 20 , since celldm takes bohr as the unit, this means that the cell dimension is approximately 0.529*20 Angstrom == 10 Angstrom (approx). This is quite large and can be the main reason why your relaxation calculation is not converging. Moreover specifying the &CELL name card is not required when youre doing "relax" calculations. It's used when youre doing "vc-relax" calculations.
So the take away is to try and reduce the cell parameter value. It would nice if you could visualize the crystal in something like Xcrysden to check whether this is the case.
You should use charge mixing mode "TF" or local TF for better convergence. For negative charge, you must set occupation to smearing and type of smearing to gaussian. You must also test pseudopotential and cutoff energy. It sometimes fails due to bad pp.
Enamul Haque
I foolishly mixed up the the atomic positions, which prevented it from converging. After viewing it in Xcrysden I realized the mistake. And thank you for letting me know about the negative charge. I did end up getting it to converge.
- Badges
- Science topic