Discrepancies in Trajectory Analysis

[kkyra99]

Hello SHARC Developers,
I am currently working through Section 2.8 ("Analyzing a Single Trajectory") of the SHARC4.0 tutorialand have encountered some inconsistencies in my results. Specifically: 1.The generated plots (Figures 2.2-2.6) differ significantly from those shown in the manual. 2.When analyzing H−C=C−H dihedrals variations using $SHARC/geo.py, I observed apparent anomalies in the results. Could you kindly advise on potential causes for these discrepancies and suggest troubleshooting steps?

Thank you ,
kyra

[maisebastian]

Dear kyra,
as I wrote recently in another issue (#12), getting different results as in the tutorial is normal, unavoidable, and not a problem:

That you obtain different results than in the example plots in the Tutorial can have various reasons, the most likely are that you use a different Python and/or Fortran version that we did when making the tutorial. This leads to different random number sequences in Wigner sampling, excited state selection, surface hopping, ..., which leads to different results. The trajectories will not match individually, but the swarm should still be qualitatively consistent.

The example trajectory plots that you show appear mostly in order. The surface hop at about 19fs from S1 to S2 crosses a large energy gap, which is something that should happen only very rarely (you can check the corresponding probabilities and random number). If you experience something like that often, consider switching to ekincorrect parallel_nac.

The dihedral angle plot is also completely normal. The dihedral angles that SHARC computes are bounded by -180 to 180 degrees, which is a very standard definition that one also finds in other software. I suggest that you plot the same curve shifted by +360 degree together with the origonal one. If you need to have smooth dihedral angles (e.g., for investigating a molecular rotor), you need to write a small script that shifts the curve by +/-360 degrees whenever it makes a step.

Best,
Sebastian

[kkyra99]

Dear Sebastian,
Thank you for your detailed explanation and helpful suggestions. I appreciate you taking the time to clarify these technical points.

Best,
kyra