Page 1 of 1
Utils to plot JASTROW term
Posted: Sun Aug 27, 2017 2:29 pm
by Vladimir_Konjkov
Hello, CASINO developers.
Is there any tools/utilities to plot JASTROW term value vs e-e, e-n distance.
In the documentation I find parameter JASTROW PLOT (Block ) which allows the user to plot the u(rij ), χ(ri ), f (ri, rj , rij ), p(rij ), ucyl (rij ) and q(ri ) terms in the Jastrow factor.
But does it suitable for CASL format?
Best, Vladimir.
Re: Utils to plot JASTROW term
Posted: Sun Aug 27, 2017 4:26 pm
by Pablo_Lopez_Rios
Hi Vladimir,
You can use the "plot" input block for this, which is activated by setting RUNTYPE to "plot". E.g.,
Code: Select all
%block plot
gjastrow_terms
electron 1 spin 1
2D
grid 100 100
A -1 -1 0
B 1 -1 0
C -1 1 0
fix electron 1 spin 2 @ 0.5 0.5 0
%endblock plot
will plot the value of all Jastrow factor terms as a function of the position of the first up-spin electron on a 100x100 grid spanning the parallelogram defined by A, B, and C (the fourth corner is B+C-A), with the first down-spin electron fixed at (0.5, 0.5, 0) and the remaining electrons at random positions.
Best,
Pablo
Re: Utils to plot JASTROW term
Posted: Wed Aug 30, 2017 5:46 am
by Vladimir_Konjkov
Hello Pablo
I want to plot particular JASTROW term vs e-e, e-n distances.
For example:
term rank [2,0] only depends on e-e distance so I plot this term to N2 molecule.
JASTROW was stored in parameters.9.casl file an graphic is in attachments emin_8_8_44.png
I also built graphic with term rank [2,0] is of r/(r^b+a) power type and have noticed that JASTROW term vs e-e distances differs significantly from term rank [2,0] is of natural power but VMC energy isn't.
Also very strange why natural power term changed its sign when e-e distance ~ 1.5 au.
I've tried to automate this procedure and write some script but if one already exists.
Best Vladimir.
Re: Utils to plot JASTROW term
Posted: Wed Aug 30, 2017 6:39 am
by Pablo_Lopez_Rios
Hi Vladimir,
I suspect your plots are missing the value of the c_1 parameter, which is omitted from the parameters.casl file since it is determined by the cusp condition. This is likely the main reason for the apparent difference between the two parametrizations in your plots. The following block will produce a correct plot of the 1-2 channel of the natural-power [ 2, 0 ] term:
Code: Select all
%block plot
gjastrow_terms
electron 1 spin 1
1D
grid 100
A 0 100 100
B 5 100 100
fix electron 1 spin 2 @ 0 100 100
%endblock plot
If you use this block for the rational parametrization you will get the correct shape, but the value will be offset by the r->\infty value of the contributions from all other electrons. For [ 1, 1 ] terms you can do something similar, fixing all other electrons at positions away from the nucleus to eliminate their contribution.
Best,
Pablo