GGA (Generalized Gradient Approximation) are a group of functional, not a particular one. PBE0 is a hybrid functional (different class than GGA) which is based on PBE. Hybrid functionals include an additional term representing the HF exchange, not present in GGA functionals. 

There is no way to answer your question easily, because everything depends on the system, model etc. A simpliest thing you can do is to run calculations with both functionals, e.g. PBE and PBE0 and compare the results. You may also take into account meta-GGA functionals like M06L. 

Good luck!

Dear Elisa, yes! there are important differences between hybrid exchange-correlation functionals (PBE0) and GGAs. PBE0 contains an exact-exchange contribution and GGAs do not. Thus, the effective potential imposed by these functionals is different. Usually Hartree-Fock method overestimates the GAP and GGAs underestimate this quantity. Therefore, the mix exact-exchange plus GGAs yields reasonable band gaps. I am assuming you are going to perform solid-state calculations. If this is the case then I suggest the U+GGA approximation, in that way you will avoid to use exact-exchange contributions, which is good for many solid-state codes. 

I have done the calculation, however I found different results between PBE0 and GGA-PBE. And also I tried LDA. PBE0 calculation is more similar to LDA result trend. I wonder which result should I believe 

I would not recommend to take LDA into account. This is rather sum of errors, that makes the results of LDA and hybrid functionals compareable. The difference between PBE and PBE0 has its source in HF exchange term. You may consider PBE with Hubbard correction (DFT+U), as mentioned above. 

BTW - what is exactly the difference in energy between PBE and PBE0?

i agree with daniels answer, check the difference between both the functionals . only then it can be commented furthur on.

thank u

LDA, PBE, HSE, PBE0, mBJ - which approach is the best? Depends on the system under consideration. In case of germanium the better choice is GW on GGA.

O thanks.. 

I just found that PBE calculation fails to give correct band order in the Ge bulk, thus it might be the reason why I got different value when I calculate dipole matrix elements on GeNW within PBE and PBE0. 

what is sufficent number of Kpoint when we use hybride funtional PBE0 for calculation band strucure

PBE0 is a hybrid functional. Compared with HSE06, PBE0 need more dense K-grids to calculate the electronic structure of materials.

Two are different. Please read any DFT review paper.

Hi all,

i read the responses to the question. Now got a bit apprehension. I use VASP mostly. So, when i use geometry optimise, or HSE06, Wannier90 whatever, i use the POTCAR (i.e. pseudo potential files where LEXCH = PE or something is mentioned) i.e its of PBE functional. I do not use any explict tag in INCAR to override the POTCAR. Means all my calculations are based on PBE hybrid functional method? and no GGA?

As far as I know, Hybrid functionals like PBE0 might open the bandgap of semiconductors, compared to that of PBE-GGA.

Please check this paper: Article Accurate Band Gaps for Semiconductors from Density Functional Theory