Multiwfn forum

saeed_E

Member

Registered: 2019-12-21

Posts: 237

Poor correlation between Rho and IE

Dear Tian,
Recently, I studied PnB interaction in O2IPn...NCH systems (Pn= N, P, As, Sb, Bi). While, as many previous experiences, it is expected a satisfactory linear correlation exists between values of Rho at BCP and values of interaction energies, in the case under study, this correlation displays a very poor R*R value. On the other hand, a much more satisfactory correlation is obtained between values of interaction energy and Laplacian of Rho! I also found that the main problem is caused by two heaviest elements Sn and Bi. Indeed, if data associated with these two elements are removed from any regression, the resulted regression is highly improved.

Could you please let me know your highly professional explanation about such a strange and novel observation?

Sincerely,
Saeed

Last edited by saeed_E (2024-08-08 08:00:23)

sobereva

Tian Lu (Multiwfn developer)

From: Beijing

Registered: 2017-09-11

Posts: 1,773

Website

Re: Poor correlation between Rho and IE

Dear Saeed,

In principle it is not necessarily that interaction energy always has a good correlation with rho(BCP), especially when the involved elements cross many rows of periodic table (note that usually the larger the atom size, the smaller the rho at BCP, because density at valence region is sparser). After all, rho(BCP) is just a very simple descriptor of weak interaction, simultaneously considering more descriptors to train a model may significantly improve correlation.

Best,

Tian

saeed_E

Member

Registered: 2019-12-21

Posts: 237

Re: Poor correlation between Rho and IE

Dear Tian,
Many many thanks for your highly valuable, professional, and novel comments. I was so enjoyed and learned.

Sincerely,
Saeed