hi all siesta users i want to to do convergence study for HgTe but i found below error Error in Cholesky factorisation in cdiag ERROR STOP from Node: 0 please help me where my structure input data is wrong.. i am attaching my fdf file with mail. thanks in advance to all.
On Tue, Feb 16, 2016 at 2:27 PM, Riccardo Rurali <[email protected]> wrote: > Dear Barnali, > > On 16/02/16 08:25, Barnali Bhattacharya wrote: > >> Respected SIESTA users, >> >> I am facing a problem regarding HOMO-LUMO gap calculation of a >> graphene flake using SIESTA 3.2. I have calculated HOMO-LUMO gap (1.43 >> eV) from *.EIG file as the HOMO is the energy level right below EF, and >> the LUMO is right above EF. But when I am plotting projected density of >> state calculated from *.PDOS file using fmpdos.f, the distance betwee >> the first state at VB and CB is comparatively less (1.213) than >> HOMO-LUMO gap. >> Why these two values differ? Is it expected? I am totally >> conused. >> I will be grateful if anyone help me? I am waiting for any response. >> > > I see two possible sources for this discrepancy. > > 1) The first one is that in the DOS (and then in the PDOS as well) you > integrate over all the k-points you have in your calculation. If you have > more k-points you should check the HOMO-LUMO gap in the EIG file for each > of them. > > Now, if what you have is 'flake', i.e. a zero-dimensional system, you > should have performed your calculation at the Gamma point. Therefore, the > only other possibility that comes to mind is: > > 2) When you calculate DOS (and consequently PDOS) you normally add a > broadening, meaning that you replace each eigenvalue with a lorentzian with > a finite width. It might be that the shrinking you observe simply comes > from the tails of the HOMO and LUMO lorentzians. Repeat the calculation > with a different broadening and see if things change. > > Good luck. > > Riccardo > > > >
HgTe.fdf
Description: Binary data
