[Pw_forum] How to shift the fermi energy? Use tot_charge ?

[gbliu]

Dear all,

       I want to calculate the dielectric function of  (3,3)-SWCNT.
       However, I need to shift the fermi energy down 1eV first.
       How can I do this?  Can I do it through setting the parameter 
tot_charge?
       If so, what's its value?  tot_charge=+1, right ?

       Another question, does caculating dielectric function need finer 
k-mesh?
       If so, can I do a scf run with, e.g. 1x1x24 k-mesh, and then a 
nscf run wiht 1x1x40  k-mesh, and then epsilon.x ?
       Is it necessary to do the scf run with a finer k-mesh?  Is it 
enough to use a finer k-mesh in a nscf run?

       B.T.W.:  I have calculated the epsilon for (3,3)-SWCNT without 
shifting fermi energy, using 41 k-points. But the curve of epsilon have 
many oscillations, especially in the range of interest 0~6eV.  Should I 
need more k-points? How many is enough?

See the attachment for input files of pw.x.

Best wishes!

Liu Gui-Bin

------------------input for epsilon.x is as follow-------------
&inputpp
    outdir='./work'
    prefix='cnt33'
    calculation='eps'
/
&energy_grid
    smeartype='gauss'
    intersmear=0.1d0
    intrasmear=0d0
    wmax=6d0
    wmin=0d0
    nw=600
    shift=0d0
/

-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: cnt33.nscf4eps.in
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20110601/4c1f3960/attachment.ksh>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: cnt33.scf.in
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20110601/4c1f3960/attachment-0001.ksh>