maituoy
commented
3 years ago Answers to your questions:
-
You can adjust the urange in the example.py. If a positive u range is given, you will have positive u values.
-
The objective function consists of 2 parts, one is the delta band gap, the other one is delta band structure. For metal, only the delta band structure will matter. You can find the delta band value of each iteration in the 4th column of u.txt file. Here is the snapshot of u.txt file in our 2d example. And if delta band is very small for a metal, which means the band structures of HSE and DFT+U are very similar. Therefore, I don't think it's necessary to apply U values to your calculation.
-
I've never tried it before. But I think it should work since it doesn't care what you do in previous steps as long as you can generate a band structure for comparison.
Intuitively, the pressure adds strains and forces on the structures and therefore influences the electronic structures. It would be interesting to see how the U value varies as the structure changes. I've done some work related to the pressure dependence of electronic structures. You may find it from (https://doi.org/10.1016/j.synthmet.2019.04.021) if you are interested.
Good luck!
Dear Yu I have sent you e-mail via RearchGate,but I am not sure if you will receive it ,so I post it here again. I also have some other quesitons as follows. 1.I see the input ,the U is negative ,Generally set to be positive in INCAR,is it can be converted? 2.if my system is metal , how to check the appropriate U without the band gap. the output u.txt ? If the delta_band is smaller, is U more appropriate for insulator. 3.and if the scf and hse can both switch the ISIF=3 to fit the U with optimization.
I will do the pressure denpendence of the U. Could you give some help or advice. zhouchao