[Pw_forum] about energy gap

[sun jason]

Dear all,

I send it again for more discussions.

Much appreciation to Mr. Stefano de Gironcoli for your kind answers!
but I still have some doubts.

1) we mentioned that LDA underestimate band gaps, the experimental gaps for 

c-BN and diamond are 6.4 eV and 5.6 eV, and the LDA results are about 4.5 
and 4.1, respectively.
so why the LDA gap of B-C-N system is not in the range of 4.1~4.5 eV?

2)N doping may couse gap reduction, but should I looks my B-C-N system as N 

doping in boron carbide? doping may couse impurity (donor or acceptor), but 

my systems are perfect single crystals.I have not consider the amorphous 
systems
they are in the case of doping?

3)as above one, are there "disorders" in perfect single crystals( in 
computational 
model)?

4)yes, the top of the valence band or the bottom of conduction band is a 
little localized, but very small, within 0.5 eV. the gap is still small 
compared with the parents materials even we add the lineup value.


Thank you very much!

====================================================
Message: 2
Date: Thu, 07 Jul 2005 11:27:01 +0200
From: Stefano de Gironcoli <degironc at sissa.it>
To: pw_forum at pwscf.org
Subject: Re: [Pw_forum] about energy gap
Reply-To: pw_forum at pwscf.org

1) LDA strongly underestimate band gaps
how large are your gaps for the parents materials compared to literature
values (Exp and Theory)

2) N doping is known to produce large reduction (giant negative bowing)
in GaAs.
A few % of N actually reduces the gap.

3) disorder tends to close the gaps

4) are your valence-band top and conduction -band bottom states
localized on different atoms species.
is your cell a kind of nano-heterojunction ? in that case you might have
some band-lineup effects that
could reduce the gap

Dig better in the literature and in your calculations.
Compare with known results in similar systems.

best regards,
Stefano de Gironcoli


sun jason wrote:

> Dear all,
>
> I'm a freshman in computational materials, I'm puzzled by a simple
> question about energy gap, when I calculated the C2-BN system which
> was construtured by mixing cubic BN and diamond.
> Well, the computed LDA band gap of c-BN and diamond is around 4.5 and
> 4.1 eV, respectively. Therefore, by mixing two compounds with such a
> wide band gap (roughly of the order of 4
> eV) one might expect to get a BC2N compound with an Eg between 4.5 and
> 4.1eV. Hence, why is our calculated Eg (about 1.0 to 2.0 eV) is much
> lower than expected for a C2-BN system?
> the unit cell I calculated is crystallined in diamond-like structure
> and the atoms are sp3 hybridization.
>
> Thank you very much!