Skip to content

Development of a ReaxFF Reactive Force Field for Tetrabutylphosphonium Glycinate/CO2 Mixtures

Evidence and attribution

Authority of statements

Prose sections below (Summary, Methods, Findings, etc.) are curated summaries of the publication identified by doi, title, and pdf_path in the front matter above. They are not new primary claims by this wiki.

Summary

This Journal of Physical Chemistry B article develops a ReaxFF parametrization for the ionic liquid tetrabutylphosphonium glycinate interacting with carbon dioxide, denoted [P(C\(_4\))\(_4\)][Gly], motivated by tunable CO\(_2\) capture in ionic liquids spanning physisorption through chemisorption. The glycinate anion contains both carboxylate and amine-like functionality, so the authors expect simultaneous physical, complexation, and reactive interactions with CO\(_2\). The training set combines periodic density functional theory pathways for CO\(_2\) reacting with the anion in the condensed phase, condensed-phase molecular dynamics configurations, gas-phase ion–CO\(_2\) interactions, and gas-phase cluster models for intra-ion interactions. Bo Zhang, Adri C. T. van Duin, and J. Karl Johnson validate the optimized force field against QM references and selected experimental trends. The introduction further motivates ionic-liquid CO\(_2\) capture by comparing aqueous amine scrubbing costs and solvent degradation to the tunability of ionic liquid–CO\(_2\) interaction strengths, and it explains why simple nonreactive force fields are insufficient when chemisorption and charge-transfer complexes appear, motivating a unified reactive treatment such as ReaxFF for simultaneous physisorption and reaction pathways. Experimental context for phosphonium amino-acid ionic liquids includes supported-IL formats that mitigate viscosity limitations during uptake, with spectroscopic evidence for carbamate-like products discussed in the cited primary literature.

Methods

Training-set construction for the reactive parametrization follows the abstract: periodic DFT pathways for CO\(_2\) reacting with the glycinate anion in the condensed phase, condensed-phase MD snapshots for structural diversity, and gas-phase cluster models capturing ion–CO\(_2\) and intra-ion interactions. After fitting, production molecular dynamics is carried out in LAMMPS. The methods section reports NpT equilibration and density measurements using a 0.5 fs timestep with Nosé–Hoover thermostat (50 fs damping) and Nosé–Hoover barostat (500 fs damping) at 1 atm, supplemented by NVT segments and short NVE segments at 0.25 fs timestep for energy conservation checks. These protocols are used to compare ReaxFF-based structural sampling with DFT-based molecular dynamics distributions for key intermolecular coordinates, emphasizing distribution-level agreement for the C(CO\(_2\))–N(anion) distance and CO\(_2\) bending angle rather than single average bond lengths.

MD application (distribution + density benchmarks). Production MD uses LAMMPS with NpT equilibration/density measurements at 0.5 fs using Nosé–Hoover thermostat (50 fs damping) and Nosé–Hoover barostat (500 fs damping) at 1 atm, supplemented by NVT segments and short NVE energy-conservation checks at 0.25 fs (methods summary on this page). System sizes, full staging tables, trajectory lengths, and electrostatic/cutoff conventions are N/A here—see papers/Zhang_Johnson_IL_CO2_capture_JPC_2014.pdf.

Force-field training (IL + CO\(_2\) ReaxFF). Training-set composition matches the abstract (condensed-phase DFT pathways, MD snapshots, gas-phase clusters) as summarized in ## Summary; optimization software, weights, and full QM settings are N/A on this page (PDF/SI).

Static QM. Periodic DFT pathways and DFT-based MD anchor training/validation (abstract + methods); detailed DFT protocols are N/A for transcription here.

Findings

The optimized ReaxFF parametrization is reported to capture essential features of both physical and chemical interactions between CO\(_2\) and [P(C\(_4\))\(_4\)][Gly] relative to the QM training data and selected comparisons to van der Waals-corrected DFT or classical descriptions where applicable. Probability distributions for the distance between the carbon atom of CO\(_2\) and the nitrogen atom of the anion, and for the CO\(_2\) bending angle, agree in broad terms between ReaxFF MD and DFT-based MD at the thermodynamic states considered, supporting the claim that the parametrization captures chemisorption motifs and thermal fluctuations sampled in the condensed-phase reference simulations. The authors further predict that mixture density increases with CO\(_2\) concentration up to at least 50 mol % CO\(_2\), attributing part of the densification to the comparatively small effective volume occupied by chemisorbed CO\(_2\) species. They note that this density trend may be tested experimentally.

Limitations

Transferability to other ionic-liquid chemistries requires retraining; long-time diffusion and viscosity are not the sole focus of the parametrization exercise emphasized in the abstract.

Relevance to group

van Duin co-authorship on ReaxFF for CO\(_2\) capture in phosphonium glycinate melts—bridges reactive IL simulation to separation applications.

Citations and evidence anchors

Reader notes (navigation)

  • Ionic liquids + CO\(_2\): separation / capture thread; reaxff-family (van Duin co-author).