Hi Jiwon,
Thank you for being interested in extending the metal ion parameters to the electrolyte interface. Please see my response below:
1.
Either JC or Li/Merz 12-6 parameter can be a good candidate to start your simulation. I might be biased to Li/Merz parameter since it does offer a 12-6-4 option which allows the tuning of r^4 term for versatile systems. For example, there have been works to use RESP charge fitting to develop Zn2+-pigment forcefield to reproduce this photochemical experimental work
https://pubmed.ncbi.nlm.nih.gov/27597406/, but neither JC nor Li/Merz was able to reproduce the values. Still, these parameters can build a good starting point for further fine-tuning, though.
2.
I don’t have clear evidence that DFT-RESP works better than HF-RESP, but the most recent MCPB.py uses DFT by default to parameterize metal center charges, so I believe the developer has considered the pros and cons rigorously before making this decision.
Hope my two cents help! I agree that AMBER may need electrolyte parameters one day (although it focuses on protein mostly).
Best regards,
Zhen.
From: 김지원 via AMBER <amber.ambermd.org>
Date: Monday, September 7, 2026 at 3:51 AM
To: amber.ambermd.org <amber.ambermd.org>
Subject: [AMBER] Choice of Li+ nonbonded parameters and RESP charge models for GAFF2 simulations of ether-based electrolytes
Dear Amber community,
I am planning to perform classical MD simulations of lithium battery
electrolytes using LAMMPS or GROMACS. My aim is to investigate both the
bulk solvation structure and the electric double-layer structure at
polarized electrode/electrolyte interfaces for ether-based electrolytes
containing different lithium salts.
I have been constructing a GAFF2-based force field in which partial charges
for the solvent molecules and anions are obtained by RESP fitting to
HF/6-31G* electrostatic potentials, while the bonded and Lennard–Jones
parameters are assigned from GAFF2.
I would greatly appreciate advice on the following two issues.
1.Selection of Li+ nonbonded parameters for GAFF2/RESP electrolytes
Amber provides several Li+ Lennard–Jones parameter sets, including the
Joung–Cheatham and Li/Merz parameters. However, 내가 아는 게 맞다면 these
parameters were generally optimized for particular aqueous models.
Would it be appropriate to use one of these water-model-specific Li+
parameter sets directly in an ether-based organic electrolyte described
using GAFF2 and RESP charges? If not, what criteria are generally used to
select and validate Li+ parameters for nonaqueous electrolytes?
2. Use of DFT-derived RESP charges with GAFF/GAFF2
My understanding is that the conventional GAFF protocol uses RESP charges
fitted to an HF/6-31G* electrostatic potential, with the remaining
parameters taken from GAFF or GAFF2. However, I have found a number of
recent electrolyte studies in which the electrostatic potential was
calculated using a DFT method, such as B3LYP, followed by RESP fitting.
Is there evidence that DFT-derived RESP charges systematically provide
better condensed-phase properties for nonaqueous electrolytes, or is the
choice mainly system-dependent? In addition, is it considered internally
consistent to combine DFT-derived RESP charges with the bonded and
Lennard–Jones parameters of GAFF2, or could this compromise the balance
among electrostatic, van der Waals, and intramolecular interactions
established during force-field development?
I would be very grateful for any recommendations, relevant references, or
examples of validated workflows for ether-based lithium electrolytes. Even
brief guidance on appropriate validation targets would be extremely helpful.
Thank you for your time and assistance.
Best regards,
Jiwon Kim
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
https://urldefense.com/v3/__http://lists.ambermd.org/mailman/listinfo/amber__;!!HXCxUKc!wWT1CF547Yl-zytD3Qn9SP7j_X3iFFNt1YcOX2sAZSwbTgT1d8UB0vCtvtOyuQD3H-LTziAcrcElu5cn2Mfm2kk$
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Sep 09 2026 - 12:00:02 PDT
