2021-04-02 WBO/Impropers meeting notes

Owned by Pavan Behara
Last updated: Feb 15, 2022 by Diego Nolasco (Deactivated)
2 min read

Participants

@Christopher Bayly

@David Mobley

@Jessica Maat (Deactivated)

@Trevor Gokey

@Simon Boothroyd

@Pavan Behara

 

Discussion

JM: Here are the plots with the three same-handed impropers (out of six) for a dataset of improper scans.

DLM/CB: This looks good and informative. Let’s pick a graph where we have a different behavior at the minima.

JM: Here you go

CB: When we look at the angle 40, we have two impropers that have non-zero forces (overlapping orange and green) and one with zero (in blue). Although the potential would be different and misleading to merge all three of them into one, I propose to plot the average of the three. Because when we do the refits then we will have a bunch of graphs and it would be difficult clearly analyze.

I would also propose a 2D heatmap with different heatlines for different FFs and the dimesions being improper and valence angles.

DLM: Yeah, they all seem to be good ideas, I am still on the fence about averaging, I would prefer to go with the current way of picking one improper definition, others have any thoughts?

SB: I think we should look at the data first and then make a judgement whether we will be seeing vastly different profiles among the impropers

CB: I can quote few cases where we have a trivalent nitrogen with identical pyramidalization and yet the valence angles being different, a conjugate nitrogen amine…

JM: I am not quite sure how the 2D plots would look but will for the data as SB says?

DLM: Okay, sounds good. What’s the status of other things regarding fitting?

JM: Once we have the dataset then I can begin to look at introducing parameters and think about fitting.

SB: Yeah, the aniline dataset PR is up on qca-submission but there are some issues server side blocking the compute, it may take a while to get that done, may be week or two.

DLM: Okay, I thought we were looking at datasets at hand, that’s fine.

 

CB: Regarding the dataset design, a preliminary geo-opt QM calculation with different substituents along the spectrum of electron withdrawing and electron donating groups would give us an idea about the pyramidalization of Nitrogen and helps us targeting the right substituents that cover the planar, intermediate and pyramidal cases perfectly on a grid.

SB: Okay, will add some more molecules to TG’s optimization dataset.