Re: [Wien] LDA vs GGA

2020-04-10 Thread Xavier Rocquefelte
I usually apply the following recipe Ueff(LDA) = Ueff(GGA) + 2 in eV. You could test many Ueff values and check how the magnetic moment evolves. Here are papers which could help. In the first one they did LDA+U+SOC with U = 0, 1 and 2 eV. https://www.nature.com/articles/srep30309

Re: [Wien] LDA vs GGA

2020-04-10 Thread Amit Chauhan ph17d008
Yes, I have done LDA/GGA + U + SOC. And both are predicting the ground state properly. But my query is related to the magnetic moment. Why it is magnetic in GGA but non-magnetic in LDA for U values I have suggested. शुक्र, 10 अप्रैल 2020, 15:37 को Xavier Rocquefelte <

Re: [Wien] LDA vs GGA

2020-04-10 Thread Xavier Rocquefelte
One more comment, LDA+U (+SOC) will be better than everything in a DFT level, because of iridium. Le 10/04/2020 à 12:02, Xavier Rocquefelte a écrit : I really recommend to look carefully at the litterature of this system. Using mbj or hybrid will work but for what? In addition, the states

Re: [Wien] LDA vs GGA

2020-04-10 Thread Amit Chauhan ph17d008
No, because my query is not related to band gap corrections. शुक्र, 10 अप्रैल 2020, 15:30 को Wasim Raja Mondal ने लिखा: > You can use mbj and hybrid functional. It is straight forward to include > spin-orbit in such calculations. > > Of course, DMFT will be the best solution. For that, you can

Re: [Wien] LDA vs GGA

2020-04-10 Thread Xavier Rocquefelte
I really recommend to look carefully at the litterature of this system. Using mbj or hybrid will work but for what? In addition, the states at the Fermi to be properly described need something you will not have in your functional. Cheers Xavier Le 10/04/2020 à 12:00, Wasim Raja Mondal a

Re: [Wien] LDA vs GGA

2020-04-10 Thread Wasim Raja Mondal
You can use mbj and hybrid functional. It is straight forward to include spin-orbit in such calculations. Of course, DMFT will be the best solution. For that, you can use packages like TRIQS, Kerstn's full LDA+DMFT code is available in his website. https://triqs.github.io/triqs/latest/

Re: [Wien] LDA vs GGA

2020-04-10 Thread Xavier Rocquefelte
For such a system DFT is not sufficient. You must use DMFT. In addition, you should include spin-orbit coupling. It is thus a very difficult situation because DMFT+SOC is not trivial at all. Cheers Xavier Le 10/04/2020 à 11:39, Amit Chauhan ph17d008 a écrit : Dear All: I am working on a

[Wien] LDA vs GGA

2020-04-10 Thread Amit Chauhan ph17d008
Dear All: I am working on a strongly spin-orbital coupled d^5 system, SrIrO3 (orthorhombic). It's a Dirac semimetal with weakly correlated (U=0.3-0.4 eV) non-magnetic ground state. With GGA+U, for U=1eV, the system is showing a magnetic metal phase but with LDA it is showing complely non-magnetic