explicit A-matrix for saffron ENDOR simulations

[yukdo]

Dear all, I've encountered peculiar artefacts when simulating Mims ENDOR angular dependence using saffron with an explicit dipole-dipole hyperfine matrix. The matrix is based on eqs. 14 and 15 in https://pubs.acs.org/doi/pdf/10.1021/ja00311a012 for the case of an axial g-tensor (also https://pubs.acs.org/doi/10.1021/ja00311a013).

When the hf interaction is given using 3 canonical values and the AFrame, the 1H ENDOR spectra are symmetrical around the Euler angle β =90°. The angular dependence (see the attached figure) looks smooth, as expected. But when the explicit 3x3 hf matrix is given, the ENDOR spectra are no longer symmetric for angles close to 90°and 0°/180°. You can see the gaps in the angular dependence on the right side.

To better visualize the asymmetry, I also plotted the RMSD between ENDOR traces simulated for the angles β and (180°−β): 0° and 180°, 1° and 179°, 2° and 178°, etc. When the canonical values + AFrame are used, the RMSD is zero for each angle (between 0 and 90°). But when the explicit matrix is used, RMSD is no longer zero on either side of the β range. I.e. spectra are no longer symmetric with respect to β = 90° in the ranges of 80°–90°/90°–100° and 0°–10°/180°–170°. But for all other beta values the RMSD is still zero. The "jumps" in the RMSD dependence are very sharp, and their exact positions depend on the distance between the electron and nuclear spins.

If anyone has an idea about the origin of these effects, I would really appreciate it.

[ninowili]

If I see it correctly, when you do it explicitly, you include the g-anisotropy. You don't do that for the "implicit" calculations. is that correct?

[yukdo]

Thanks for the answer. Yes, you are correct. The g-anisotropy is taken into account explicitly in the 3x3 A-matrix. However, it is not the source of the "artifacts". Even if I use an isotropic g-factor (g1=g2=g3=g_iso), the issue is still present (see the new figure attached).

In the other route (3 canonical hf values + A-frame - from my first post), the g-tensor as found in Sys.g was still axial.
I only used g_iso to construct the canonical hyperfine values [-T -T +2T].

[Stefan Stoll]

It could be that the automatic grid symmetry determination is failing in this case.

Try setting Opt.GridSymmetry = 'D2h' or Opt.GridSymmetry = 'Ci' and look if this eliminates the difference.

[Stefan Stoll]

Looking more into this, the reason for EasySpin's failure to determine the correct symmetry is that the magnitude of the tensor is so small compared to the microwave frequency. We'll fix this.