Large Number of nuclei and 2 electrons

General forum for EasySpin: questions, how to's, etc.
Post Reply
niconeuman
User
Posts: 11
Joined: Mon Oct 22, 2018 1:16 am

Large Number of nuclei and 2 electrons

Post by niconeuman »

Greetings,
I'm trying to simulate an isotropic signal of a spin coupled system which contains two weakly coupled S = 1/2 spins and several nuclei. I have 8 N atoms and two I = 1/2 nuclei. All of them are coupled but the system is highly symmetric so all the nuclei are equivalent. garlic doesn't seem to allow more than one unpaired electron. And pepper doesn't allow me to define several equivalent nuclei.
I only need to fit sys.ee, but each spectrum takes quite a while to simulate, even though I am simulating only one crystal orientation (to emulate the isotropic solution spectra).
My question is if there is some way of speeding this up. Would it be faster if I calculate the spectrum in frequency domain, and then just convert it to magnetic field domain? g is around 2.05.
I appreciate any help.
Thank you very much!
Nicolas
katarkon
Local Expert
Posts: 191
Joined: Mon Jan 12, 2015 4:01 am

Re: Large Number of nuclei and 2 electrons

Post by katarkon »

It should be better to attach your script for simulation. Your system looks extremely large for treating with matrix diagonalization. Also, it completely unclear why Sys.ee is used for isotropic system. Sys.J should be enough. Is it able to treat some nuclei perturbationally via hybrid method? It should speed-up the calculations substationally, but it is appropriate when J>>A only.
niconeuman
User
Posts: 11
Joined: Mon Oct 22, 2018 1:16 am

Re: Large Number of nuclei and 2 electrons

Post by niconeuman »

Hi, thanks for the answer. The system is indeed too large, and I'm using the hybrid option; full diagonalization would be impossible. I don't think that J >> A, I would say is is perhaps in the order of magnitude, maybe it is larger. Anyway my simulations don't seem to be too sensitive to J. Then perhaps I should test the differences between full and hybrid in a dummy model with less N atoms. So is there no way of accelerating the simulation other than using hybrid? I didn't know that Sys.J was there, but what would be the difference between defining Sys.J = value or Sys.ee = value?
Thank you very much!
Best regards,
Nicolas
katarkon
Local Expert
Posts: 191
Joined: Mon Jan 12, 2015 4:01 am

Re: Large Number of nuclei and 2 electrons

Post by katarkon »

I think, the best way is writing a custom simulation function. It looks not extremely hard if Your system may be treated by first-order perturbation limit.
The calculation of the line positions and intensities may be found elsewhere, like Sankarapandi al al., doi.org/10.1006/jmra.1993.1147
Fast algorithm for the case with equivalent nuclei may be found in Stefan Stoll's exchange function viewtopic.php?f=8&t=309.
EasySpin has middle-level functions makespec and convspec for calculating of the spectrum.
Good luck.
Post Reply