[Wien] Plamon (plasma) in joint/kram

[Fred Nastos]

On November 30, 2005 04:13 pm, L. D. Marks wrote:
> My main question is whether there is a correction for the effective mass
> built in somewhere here, or not. The UG says "number of ``free`` electrons
> per unit cell assuming bare electron mass (calculated around E_f+/- 10*de
> as defined in input line 4), plasma-frequency". It's not clear if an
> estimate of the effective mass is in there. (If so, it would be nice to
> know what it is so one can do an empirical correction if needed.)

If this is indeed an approximation, then I can't see how the oscillator 
strength sum rule is satisfied.  And I was under the impression that
the optic/joint codes do a check on this at some point.

I'm guessing that in this line of joint.f (around 660 for my version):
	plasm2(ip)= plasm2fac*(EPLF*gesDENSTY(npl+1,ip))
the array gesDENSTY contains the corrections required to incorporate the 
correct computed effective mass.

My understanding is that for m* in the usual plasma frequency formula,
one needs to use an m* averaged over the Fermi surface (I have some
notes on this somewhere if you're interested).  This can easily be
computed within the integration scheme that are used in joint.
Furthermore, it is my understanding that this is what is done in the
code.  If it's not, then it can in principle be easily calculated.

I agree, the documentation (and perhaps variable names) needs some
work.  Pet-peeve: the same variables are sometimes capitalized and
other times not.

FN

> Specific locations:
>
> SRC_joint
> line 661 joint,f	plasm2(ip) is plasma frequency for
>  			a density (both spins I think). It's not
>  			transparent what this density is (few comments).
>
> Used later around line 721 to print out Drude results.
>
> SRC_kram
> kram.f	Uses the plasma frequencies (user input) with I assume is
> supposed to have been copied from a previous run of joint (mode 6) and
> adds it to the dielectric constant around line 207.
>
> On Wed, 30 Nov 2005, Fred Nastos wrote:
> > On November 30, 2005 01:47 pm, L. D. Marks wrote:
> >> From what I can see, joint will (with option 6) calculate a plasmon
> >> frequency for an effective mass of one based on the number of electrons,
> >> then kram will add this in. This strikes me as strange, probably wrong.
> >> Why does one need to do this?
> >
> > This might be related to the divergent part of the dielectric function.
> > For a metal, the Kramers-Kronig relation relating the imaginary part of
> > the dielectric function to the real part looks like:
> >
> > Im[eps(omega)] = sigma/omega - (some principal parts integral)
> >
> > But I'm not sure exactly what your are referring to when you say "kram
> > will add this in".  Can you point out roughly which parts of the code you
> > are concerned with?
> >
> > FN
> > _______________________________________________
> > Wien mailing list
> > Wien at zeus.theochem.tuwien.ac.at
> > http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>
> Note: if you have an old email address for me, please note that "nwu" has
> been changed to "northwestern".
> -----------------------------------------------
> Laurence Marks
> Department of Materials Science and Engineering
> MSE Rm 2036 Cook Hall
> 2220 N Campus Drive
> Northwestern University
> Evanston, IL 60201, USA
> Tel: (847) 491-3996 Fax: (847) 491-7820
> email: L-marks at northwestern dot edu
> http://www.numis.northwestern.edu
> -----------------------------------------------
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien