Questions on Ps2 example

General discussion of the Cambridge quantum Monte Carlo code CASINO; how to install and setup; how to use it; what it does; applications.
Post Reply
Dmitry_Zvezhinsky
Posts: 8
Joined: Sat May 03, 2014 12:38 pm

Questions on Ps2 example

Post by Dmitry_Zvezhinsky »

Dear Casino users (and especially Neil)!
Could you please give me any hints on some questions concerning the example distributed within input/exmol directory?
1. Firstly, have you published these (or similar) results produced by Casino, especially concerning typical benchmarks for positron systems: Ps2, Ps-, PsH etc. and including any related technical details? What references could you advice on this?
2. You have used the H term in the Jastrow factor which is not documented in casino manual (at least I have not found anything). I have only realized that this is some three-body term which somehow takes into account the information from %block custom_striplet_dep.
I wonder how does this term look like and how does the input block above enters its calculation?
3. I have done the DMC simulation based on your example. Here is the plot of binding energy vs. time step. (I have tried the calculations without the mysterious H term in Jastrow and without backflow transformation). It seems that the H term is quite useful for calculations, while the backflow correction is not. Am I wrong? What do you think about this?
Attachments
Ps2Bind.png
Ps2Bind.png (31.07 KiB) Viewed 7690 times
Neil Drummond
Posts: 113
Joined: Fri May 31, 2013 10:42 am
Location: Lancaster
Contact:

Re: Questions on Ps2 example

Post by Neil Drummond »

Dear Dmitry,

To my knowledge no results on Ps2, etc., have been published using exmol. It was implemented by Pablo and you might want to ask him about this...

The H term is a polynomial three-body Jastrow term, a bit like the f term (which is a three-body term for two electrons and an ion). The custom_striplet block lets you define the types of particle triples that have the same or different three-body terms. In a Ps2 molecule there is only one kind of triple: triples are either two electrons and a positron or two positrons and an electron, and these should have the same three-body terms. If you were looking at a biexciton with different electron and hole masses, there would be two types of triple: two heavy particles and light particle or two light particles and a heavy particle. However you would also need to define a "null" spin-triple type (for the non-exist electron-electron-electron triples, etc.). Something like

%block custom_striplet_dep
no_striplet_deps 1
striplet_dep -1 3
1=1=1,2=2=2,3=3=3,4=4=4,1=1=2,1=1=3,1=1=4,2=2=1,2=2=3,2=2=4,3=3=1,3=3=2,3=3=4,4=4=1,4=4=2,4=4=3 # Unused triples
1=2-3,1=2-4 # Two electrons and a hole
3=4-1,3=4-2 # Two holes and an electron
%endblock custom_striplet_dep

You can then use this by setting the spin-dependence to -1 in the H term.

The systems you are looking at do not have nodal surfaces. The node-altering abilities of backflow are not helpful, although backflow can provide some description of three-body correlations.

Best wishes,

Neil.
Post Reply