QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: Twisty on July 5, 2010, 07:40

Title: About Si Nanowire
Post by: Twisty on July 5, 2010, 07:40

Dear sirs,
to make a Si nanowire, one can start with silicon from database then he can take it to bulk builder to make a bulk si. then we can take it to nanowire of custom builder. If he chooses the bandstructure from nanoscript writer, vnl will show the bandstructure of silicon nanowire.

But if he chooes only si from database to make nanowire in the custom builder without making it bulk, does the bandstructure show the correct result?

it can be also applied in GaAs or InP nanowire.
please let me know which one is correct one. And another thing, why only bulk silicon does not show the correct bandgap? it shows 2.3eV where i have change all parameter but it does not show correct result. Again it does not show band splitting in the valence band. I think i lag some theoretical aspects or the software has some technical trick? please let me know about it.

thanks.

Title: Re: About Si Nanowire
Post by: Nordland on July 5, 2010, 11:37

I am not sure what you want to know so I will try a little all-around:

First of all are you talking about Extended Huckel or DFT ?

The bandstructure and hence the band gap of a bulk silicon can be quite different from a nano-wire due to the surface of the nano-wire. It is only in the "infinite" limit that the two bandstructures becomes identical.

If it is DFT, it is well established truth that the DFT underestimates the band gap quite a lot,
but in the just released version there is support for LDA+U ( & GGA+U) which can correct for this.

I hope this was a little help.

Title: Re: About Si Nanowire
Post by: Twisty on July 5, 2010, 13:48

Dear sir,
thanks for your reply. Actually i was simulating a Si nanowire by DFT+LDA. First i want to see the bandstructure by only bulk silicon. i tried by Extended Huckel, found 2.3eV. Then i tried DFT+LDA, i found less than 1eV. But according to your advice, i have 1.5eV by DFT+LDA+U. it seems rather correct than others.

I think it will also work for nanowire also. thanks again.

Title: Re: About Si Nanowire
Post by: Nordland on July 5, 2010, 15:24

What parameter have you used for U ?

Title: Re: About Si Nanowire
Post by: Twisty on July 5, 2010, 17:53

Dear sir,
i m sorry that i mistakenly took the bandgap between gamma-k band where i should have take gamma-x band which is less than 1eV.
Specifically i dont have much idea of the following parameter:
 electron temp:300
mesh cut: 75
iteration control:
maximum steps:100
algorithm:pulayMixer
damping factor: 0.1
only ka=15 kb=15 kc=15
i mean all are default. and the curve i have here it is attached.