Hello Leila, I see that a spin-polarized calculation with your input file gives a system with two spin-up electrons. In this case, yes, the H atoms want to move apart to become two independent atoms. So, in the spin-polarized calculation, you need to fix the initial spin to zero. FixSpin T TotalSpin 0.0 Then I get the H2 binding energy of -4.8 eV. (It will be a little smaller when corrected by BSSE).
Spin polarization is essential for the correct description of the H atom (not essential for the molecule). Also, as Haibo said, only the gamma point is needed for an isolated molecules, more k-points just take more time. Regards, Natalia On 2/28/11, Haibo Guo <[email protected]> wrote: > Only gamma point is needed for isolated molecules or atoms. Spin > polarization can be included; you need a initial antiferromagnetic > configuration for H2. But without spin, the results should be close > because H2 molecules don't have net spin polarization. > > H2 should behave well with most functionals. Here are some of my > results of the binding energy (eV per bond): > > PBE (siesta): 4.76 > PBE (vasp): 4.54 > HSE06 (vasp): 4.53 > > Hope this helps. > > Haibo > > > > 2011/2/28 Süle Péter <[email protected]>: >> >> Dear Leila, >> >> for me using siesta-2.0.2 with your fdf file (GGA/PBE, 5X5) >> >> 0.75 -31.5205 eV/H2 >> 10. -25.4867 >> inf -24.3496 >> >> Ec= ~ -3.5854 eV/atom (e.g. -7.17 eV bond energy) >> >> which is not bad. I mean one can not necessariliy expect >> with GGA a better value. Hybdrid functionals should give better >> results. The issue of better VASP values should be discussed >> in a more detail by experts since indeed quite oftenly happens. >> >> Best regards, Péter Sule >> >> >> ---------------------------------------------------- >> Dr. Peter Sule >> senior research associate >> Institue for Technical Physics and Materials Science >> (MTA-MFA), Konkoly-Thege M. 29-33, Budapest, Hungary >> tel.: (36) 1 392 2222/3605 >> fax.: (36) 1 392 22 73 >> www.mfa.kfki.hu/~sule / [email protected] >> ----------------------------------------------------- >> >> >> On Mon, 28 Feb 2011, leila wrote: >> >>> >>> Hi, everyone, >>> >>> >>> >>> I want to do some calculations on the H effect in Fe, before which I have >>> to >>> promise the accuracy of H potential. I first make a simple tests of the >>> H2 >>> properties, including the H-H distance in H2, and also the binding energy >>> (Defined as the energy difference of the H2 molecular and two isolated H >>> atom). >>> >>> I find the H-H distance is around 0.75 Ang, which is near to the >>> experimental value of 0.741 Ang, for both LDA and GGA pseudopotential. >>> However, the binding energy is always not correct, calculated to be >>> around >>> -7.2 eV (The experimental and VASP results are all around -4.55 eV). My >>> input file for H are in the attachments. >>> >>> I don’t know where the issues are. Could anybody help me? >>> >>> >>> >>> In addition, I find that the spin polarization contributes a lot in the >>> H2 >>> calculations. When I add the spin polarization effect in the VASP and >>> SIESTA >>> calculations, the H-H distance becomes to 2.89 Ang, and the binding >>> energy >>> is approaching to 0 eV. Can anyone explain this to me. Because Fe is >>> magnetic materials, I have to consider spin polarization effect in my >>> calculation. >>> >>> >>> >>> Thank you in advance! >>> >>> Yours sincerely, >>> >>> Leila >>> >>> >>> >>> >> >
