How to draw band gap of sillicon from resulst

General discussion of the Cambridge quantum Monte Carlo code CASINO; how to install and setup; how to use it; what it does; applications.

How to draw band gap of sillicon from resulst

Postby Nguyen_Oanh » Tue Sep 13, 2016 1:05 am

Hello CASINO user!
I am a new user and I am using CASINO program to check band gap of sillicon. I have a question about using program. Can you help me to answer for my question about how to plot expval file? I read CASINO manual but I don't understand the way to plot it. I used plot_expval with input file but it still error.Can you guide me to contribute input file in plot_expval?
I already calculated density and spin density, structure factor of sillicon 222G crystal. The out file have total energy, local energy and e-e vanderWald, e-i interaction. how can I draw density depend on electron energy ( band structure of sillicon crystal)
Thank you very much!
Nguyen_Oanh
 
Posts: 17
Joined: Mon Sep 12, 2016 10:32 am

Re: How to draw band gap of sillicon from resulst

Postby Ryan_Hunt » Tue Sep 13, 2016 7:33 am

Dear Nguyen_Oanh,

I think the answer to your question involves the way in which band-gaps are calculated in CASINO.The bandgap itself is not something you can access from the "plot_expval" utility. Band-gaps are obtained from total energy calculations made with different levels of excitations from the ground state (as specified by the MDET block in correlation.data, or similar in the new casl format). Info on how to do this is in the manual.

Short answer: can't get bandgap from plot_expval - need to perform separate calculations for excited states.

Hope this helps, Ryan.
Ryan_Hunt
 
Posts: 26
Joined: Thu Mar 10, 2016 10:58 am
Location: Lancaster, UK

Re: How to draw band gap of sillicon from resulst

Postby Nguyen_Oanh » Wed Sep 14, 2016 3:22 am

Thank you because of your responding. I already read Casino manual but I don't find detail about how to the way to calculate band gap. Can you issue detail informations? I want to calculate band gap of sillicon.
Thank you!
Nguyen_Oanh
 
Posts: 17
Joined: Mon Sep 12, 2016 10:32 am

Re: How to draw band gap of sillicon from resulst

Postby Ryan_Hunt » Wed Sep 14, 2016 7:08 am

Dear Nguyen_Oanh,

In some respects you are right - the manual says very little about how to calculate band gaps directly. However, the manual provides information on how one can (a) ADD/REMOVE electrons and (b) PROMOTE electrons. If you want to calculate a bandgap - firstly you should decide which one (optical or electronic - aka quasiparticle or excitonic gap). Secondly, you should decide what excited state total-energy calculations you need in order to calculate that bandgap. For example, if you want to calculate the excitonic gap of some material at the Gamma point in the Brillouin zone, then you might consider taking the total-energy difference of the FNDMC energy of a state with one electron PROMOTED from the valence band to the conduction band at Gamma (details in the manual - the MDET block) and the FNDMC energy of the ground state. For the quasiparticle gap the required "excited states" will be formed from electron ADDITIONS and REMOVALS (change neu/ned in the input file, and check the orbital occupancy in the outfile).

That's not very detailed, but it covers most of the ideas behind doing a bandgap calculation in QMC. Generally, you use energy differences between excited and ground states to extract gaps. There will, of course, be other issues to consider (general QMC ones and bandgap specific ones).

If you are in the market for some other resources on excited states in silicon, may I suggest searching for the various calculations done in the 90s by the group at Cambridge? (A scholar search of "silicon excited states monte carlo" should suffice).

Hope this is helpful, thanks, Ryan.
Ryan_Hunt
 
Posts: 26
Joined: Thu Mar 10, 2016 10:58 am
Location: Lancaster, UK

Re: How to draw band gap of sillicon from resulst

Postby Nguyen_Oanh » Wed Sep 14, 2016 8:21 am

Thank you very much. I have more a problem about using plot_expval. Can you guide me how to build input file in plot_expval utility? I see in casinohelp plot_expval but I don't understand .
The input file that I wote in plot_expval
%block plot_expval
3 # for 3-dimension
10 10 10 # 30 points along x, 20 points along y, 10 points along z
0 0 0
0 0 0
0 0 0 # max coordinate A = (10, 10, 10), min coordinate B = (0, 0, 0)
%endblock plot_expval
Can you help me checking the error in input file of plot_expval?
Yours sincerely!
Nguyen_Oanh
 
Posts: 17
Joined: Mon Sep 12, 2016 10:32 am

Re: How to draw band gap of sillicon from resulst

Postby Ryan_Hunt » Wed Sep 14, 2016 9:43 am

Dear Nguyen_Oanh,

You need to decide what region you would like to plot in and what quantity you would like to plot. Say density for example, then you would write a plot_expval block like this

%block plot_expval
3
10 10 10
v11 v12 v13
v21 v22 v23
v31 v32 v33
v41 v42 v42
%endblock plot_expval

where vAB is the Bth element of vector vA - a vector definining the "corners" of the 3D plot region. This is pretty much written verbatim from the output of
Code: Select all
casinohelp plot_expval
, but I have tested it for one of the silicon (blip wfn) examples and it does seem to produce a plot.

I don't actually use plot_expval very often, I suggest that if you still have problems with it that you start a new thread (plenty of people who might use it more often than I probably no longer get notifications about this thread! :) )

Thanks, Ryan.
Ryan_Hunt
 
Posts: 26
Joined: Thu Mar 10, 2016 10:58 am
Location: Lancaster, UK

Re: How to draw band gap of sillicon from resulst

Postby Nguyen_Oanh » Sun Sep 18, 2016 4:21 am

Thank you so much! I will read this part in manual again.
Nguyen_Oanh
 
Posts: 17
Joined: Mon Sep 12, 2016 10:32 am

Re: How to draw band gap of sillicon from resulst

Postby Nguyen_Oanh » Wed Sep 21, 2016 4:17 am

Dear ueser!
I have a wonder about expval.data. I calculated for Sillicon 222G example with plane wave function and I received these results about charge density and structrure factor. With that results, Can I draw DOS depend on energy? How to the way to draw? Can you lead me to use these results in expval.data particularly?
Yours sincerely!
Nguyen_Oanh
 
Posts: 17
Joined: Mon Sep 12, 2016 10:32 am

Re: How to draw band gap of sillicon from resulst

Postby Nguyen_Oanh » Wed Sep 21, 2016 4:18 am

Nguyen_Oanh wrote:Dear ueser!
I have a wonder about expval.data. I calculated for Sillicon 222G example with plane wave function and I received these results about charge density and structrure factor. With that results, Can I draw DOS depend on energy? How to the way to draw? Can you lead me to use these results in expval.data particularly?
Yours sincerely!
Nguyen_Oanh
 
Posts: 17
Joined: Mon Sep 12, 2016 10:32 am

Re: How to draw band gap of sillicon from resulst

Postby Ryan_Hunt » Wed Sep 21, 2016 8:56 am

Dear Nguyen_Oanh,

The density of states for a crystalline material would be calculated from a set of orbitals on a (hopefully very fine) mesh of k-point in an irreducible wedge of the Brillouin Zone. For a fixed density, these orbitals can be calculated and the DOS formed by taking some integral over the occupied states (and indeed the unoccupied ones - beyond the Fermi level, and with some "complications" that I'm neglecting). If you like, you can find more information about this by looking at the documentation for the various DOS calculating codes (OptaDOS is an example - search "OptaDOS user guide").

To attempt a QMC calculation for such a quantity (at least, in the same way as the existing codes) would likely require some serious thought and computer time. Suffice it to say, this is probably not as straightforward as adding a term to your expval block.

Sorry I can't be of more help, Ryan.
Ryan_Hunt
 
Posts: 26
Joined: Thu Mar 10, 2016 10:58 am
Location: Lancaster, UK

Next

Return to The CASINO program

Who is online

Users browsing this forum: No registered users and 1 guest

cron