In by admin

Name Mr. Xiang Zhong
Organization or Institution Department of Chemistry, University of Miami
Presentation Type Poster
Topic Computational Chemistry

Development of OPLS parameters for pyridinium-based ionic liquids


Xiang Zhong, Brian Doherty, and Orlando Acevedo*

Author Institution(s)

Department of Chemistry, University of Miami, Coral Gables, FL 33146


Development of force fields for imidazolium-based ionic liquid simulations have gained considerable attention with multiple parameter sets reported in the literature; however, accurate potentials for pyridinium-based molten salts are lacking. With an increasing amount of experimental data of physical properties for N-alkylpyridinium [RPyr] (R = E for ethyl, B for butyl, etc.) based solvents now readily available, the accuracy of existing parameters, including our own OPLS-2009IL force field, can be dramatically improved. In this work, new OPLS parameters are reported for the ion pair combinations of N-butylpyridinium [BPyr] with the anions tetrafluoroborate [BF4], chloride [Cl], bromide [Br], and bis(trifluoromethylsulfonul)amide [NTf2]. The computed solvent density, heats of vaporization, viscosities, and other relevant solvent data are compared to experiment. Different parameterization techniques were tested from traditional QM-based charges to more novel machine learning techniques featuring genetic algorithms. Large improvements in the prediction of bulk properties will be reported. For example, overall mean absolute errors (MAEs) of 2.5% and 3.0% for the heat of vaporization and density, respectively, were found using our new parameter set compared to 17.0% and 3.4% for our earlier OPLS-2009IL force field. Correct local interactions, including cation-anion interactions and pi-pi stacking between the pyridinium rings, were also emphasized.