Development of the ReaxFF CBN reactive force field for the improved design of liquid CBN hydrogen storage materials
Year vs slug
The stable wiki id retains the 2015pai-… slug for link stability; bibliographic year is 2016 per the published PCCP issue pages.
Summary¶
Liquid CBN (carbon–boron–nitrogen) carriers such as 3-methyl-1,2-BN-cyclopentane are attractive because they can plug into existing liquid-fuel infrastructure, but liquid-phase hydrogenation/dehydrogenation pathways are hard to map experimentally. This PCCP article develops ReaxFF_CBN from B3LYP QM data on BN-substituted cycles and small molecules, then validates that the fitted field reproduces QM dehydrogenation pathways and energetics for training cases. ReaxFF MD in LAMMPS on ~500-monomer liquid cells (density set from experiment) compares unimolecular versus bimolecular dehydrogenation as temperature rises and contrasts pentagonal versus hexagonal liquid CBN motifs for thermal stability and dehydrogenation kinetics.
Methods¶
Force-field training¶
QM reference: DFT with the B3LYP hybrid functional (basis sets and SCF settings in the article/SI) supplies bond dissociation curves, reaction energetics, and BN-substituted cyclic geometries for C/B/N/H liquid CBN carriers. Training set / targets: reactions and RMSE tables enumerated in the main text and ESI. Optimization: ReaxFF_CBN parameters are adjusted until those QM benchmarks are satisfied within the reported tolerances. Experimental liquid mass densities constrain supercell volumes.
MD application (atomistic dynamics)¶
Engine: LAMMPS with ReaxFF_CBN after Cerius2/Dreiding pre-relaxation of liquid cells. System: cubic 3D PBC boxes of ~500 monomer units sized to experimental density for each CBN liquid composition (atom totals depend on monomer choice—see PCCP tables). Ensemble / timestep: NVT at fixed volume with Nosé–Hoover thermostat (damping 1.0 in the paper’s units) and Δt = 0.1 fs. Protocol: equilibration near 300 K, then linear heating 300 → 2000 K at 34 K/ps to accelerate dehydrogenation while respecting stated boiling constraints for the modeled liquids. Barostat / applied pressure: N/A — volume set from experimental density, not NPT pressure control. Electric field / enhanced sampling: N/A.
Findings¶
Validation: ReaxFF_CBN reproduces QM dehydrogenation pathways and energetics for the highlighted training reactions (additional tables in ESI). Mechanism vs temperature: unimolecular dehydrogenation dominates at lower T, while bimolecular channels grow in importance as temperature increases—consistent with higher collision rates in dense liquid. Motif comparison: hexagonal liquid CBN motifs are reported as preferable to pentagonal analogs when weighing combined thermal stability and dehydrogenation kinetics under the same heating ramps. Outlook (authored framing): the parametrization is positioned as a screening tool for new liquid CBN hydrogen carriers ahead of synthesis, with accelerated heating and finite system size as practical caveats on full kinetic convergence.
Limitations¶
Accelerated heating protocols and finite system sizes for liquids leave kinetic branching only partially converged; ReaxFF accuracy for B–N organics must be tracked against new QM data whenever functional groups move outside the training manifold. Regeneration cyclability claims in the introduction are engineering targets that are not fully resolved by short MD windows alone.
Relevance to group¶
ReaxFF parameterization case study on heteroatomic organic hydrogen storage.
Citations and evidence anchors¶
DOI 10.1039/C5CP05486A.