Reactive molecular dynamics modeling of collision-induced dissociation of 1-ethyl-3-methylimidazolium tetrafluoroborate ionic liquid ions
Summary¶
The paper models collision-induced dissociation (CID) of gas-phase ions of the ionic liquid EMIM–BF\(_4\) ([EMIM]\(^+\), [BF\(_4\)]\(^-\), and 2:1 and 1:2 clusters) colliding with N\(_2\) using ReaxFF in LAMMPS, spanning 10–100 eV laboratory-frame collision energies. B3LYP/6-31G GAMESS calculations supply relaxed precursor structures and supplementary thermochemistry on selected pathways; tandem mass spectrometry (MS/MS) on an electrospray platform provides experimental product distributions for comparison.
Methods¶
- Quantum chemistry: Geometries for precursor ions/clusters use DFT B3LYP/6-31G in GAMESS. Cluster conformers are explored with CHARMM MD in LAMMPS (annealing cycles) before returning minima to DFT. Selected pathways are evaluated for ΔH\(_{0\,\mathrm{K}}\) and ΔG\(_{298\,\mathrm{K}}\) to compare dissociation favorability.
- Reactive MD: LAMMPS ReaxFF parameters follow Medina-Ramos et al. with EMIM–BF\(_4\) refinements from Tahsin & Petro. Each collision starts from 300 K NVT thermalization with Nosé–Hoover thermostats as implemented for canonical sampling, then switches to NVE while assigning an approach velocity for the target lab-frame collision energy (10–100 eV in 10 eV steps). Angular sampling uses a grid of y-angles (0–300°, six values) and z-angles (0–180°, seven values). Impact parameters span 0–4 Å relative to N\(_2\). Nonbonded N\(_2\)–ion interactions use a 4 Å cutoff. Each trajectory uses a 0.1 fs timestep and 2 ps total integration time.
- Experiments: SCIEX ZenoTOF 7600 MS2 on ~100 μM EMIM–BF\(_4\) in methanol infusion, N\(_2\) collision gas, 0–100 eV collision energy in 10 eV steps (positive and negative modes).
MD application (head-on collisions): Isolated ion + N\(_2\) pairs in a gas-phase molecular dynamics cell; N/A — full 3D PBC periodic supercell for a condensed solid (non-periodic / minimal box in LAMMPS-style collision setups, per article); after 300 K NVT with Nose-Hoover thermostat, NVE microcanonical production for 2 ps at 0.1 fs timestep; N/A — barostat; N/A — hydrostatic pressure coupling; N/A — electric field; N/A — metadynamics; N/A — umbrella bias.
Findings¶
- MD predicts [EMIM]\(^+\) onset near ~20 eV (lab frame) and [BF\(_4\)]\(^-\) remains intact below ~40 eV, with [BF\(_4\)]\(^-\) fragmenting to [F]\(^-\) + BF\(_3\) and [BF\(_2\)]\(^+\) at higher energies (≥60 eV features discussed), including reaction pathways the authors highlight.
- [EMIM]\(^+\) fragmentation is dominated by 3-methylimidazolium at lower collision energies; cyanide becomes prominent at higher eV; 2:1 and 1:2 clusters fragment at 10 eV, showing sensitivity to cluster concentration-like stoichiometry. Outcomes are compared to electrospray MS2 experiments with agreement described in the main text, although internal energy distributions in experiment limit exact matching (limitation noted in the article).
- Simulated spectra align versus the ZenoTOF measurements; caveat—short (2 ps) gas-phase trajectories miss slow rearrangements. Version-of-record PDF governs all quantities; this page is a curated extract-aligned summary.
Limitations¶
- Short (2 ps) gas-phase trajectories and a specific ReaxFF fit may miss slow rearrangements or electronic effects important at some energies.
- Electrospray experiments probe soft ionized ensembles; high-energy CID modeling omits solvent and internal energy distributions of the experimental ion population beyond the stated protocol.
Relevance to group¶
Demonstrates benchmarked ReaxFF CID workflows for ionic-liquid ions in LAMMPS, relevant to mass-spectrometry and low-pressure plasma fragmentation contexts.