To share experience including discussions about scientific questions.
Moderators: Global Moderator, Moderator
-
suojiang_zhang1
- Jr. Member

- Posts: 64
- Joined: Tue Nov 19, 2019 4:15 am
#1
Post
by suojiang_zhang1 » Mon Feb 24, 2025 1:46 pm
Hi,
I can train a MLFF to investigate the structure and its evolution in hundrads of ps time.
I wonder such strategy can be used to train a reaction, for example the CO2 reduction on Cu surface?
Or like QM/MM to provide the solvation environment to hold on the reactants/
I will appreciate your any suggetions and instruction.
Yours
-
ferenc_karsai
- Global Moderator

- Posts: 530
- Joined: Mon Nov 04, 2019 12:44 pm
#2
Post
by ferenc_karsai » Mon Feb 24, 2025 2:08 pm
I've moved the post to "users by users".
-
ferenc_karsai
- Global Moderator

- Posts: 530
- Joined: Mon Nov 04, 2019 12:44 pm
#3
Post
by ferenc_karsai » Thu Feb 27, 2025 9:49 am
You will need some kind of advanced molecular dynamics method coupled to the machine learning force fields to tackle chemical reactions. In the last VASP workshop in November 2024 these topics were discussed. The talks from the workshop are not available online at the moment but will probably made available online in the future.
We also still have no publication that fully simulates chemical reactions but we are working on it. The closest thing to that was within this publication.