Skip to content

Statistical Analysis of Tri-Cresyl Phosphate Conversion on an Iron Oxide Surface Using Reactive Molecular Dynamics Simulations

Corpus PDF role

This slug tracks an author proof PDF (Khajeh_FeO_Phosphate_JPCC_2019_proof.pdf). The typeset journal PDF and aligned curation live under 2019khajeh-j-phys-chem-statistical-analysis; scientific statements below match that work.

Summary

Aviation lubricant chemistry motivates reactive modeling of tri-cresyl phosphate (TCP) on iron oxide surfaces under conditions relevant to film-forming contacts. The study develops ReaxFF parameters for Fe/P/O chemistry against DFT reference data for Fe–O–P bonding, P/PO adsorption on iron facets, and related dissociation and angle energetics, then applies the field in large-scale LAMMPS sampling. Rather than a single short trajectory, the authors launch 100 parallel replicas with one TCP molecule per cell on a passivated amorphous iron oxide substrate, exploring chemisorption regioisomers across a temperature ladder from 300 K to 700 K. The central output is site-resolved statistics for which TCP atoms form load-bearing hetero-bonds to the surface—especially Fe–C versus P=O oxygen contacts—and how those preferences shift with temperature, contextualized against experimental observations of TCP-derived films in oxygen-poor tribological environments.

Methods

Force-field training (2). ReaxFF for Fe/P/O and TCP–oxide chemisorption, fit to DFT for Fe–O–P geometry/energy, Fe–P dissociation, Fe(II)/Fe(III)–O–P bending, and P/PO on Fe facets; ReaxFF vs DFT in the manuscript; parameters in SI. Substrate: one passivated amorphous iron oxide slab.

MD application (1). Engine: LAMMPS; 0.25 fs timestep; bond order every 1.25 ps (cutoff 0.3). System: one TCP and one oxide slab; 100 replicas (50 P=O-down, 50 P=O-up). Ensemble: NVT; Nose–Hoover thermostat on non-fixed atoms. PBC and fixed substrate treatment as in 2019khajeh-j-phys-chem-statistical-analysis / PDF. Barostat / servocontrol of pressure: N/A — NVT. Temperature: 300–700 K in 100 K steps, ~1 ns/stage to plateau; last ~25 ps averaged. Electric field: N/A. Umbrella / metadynamics / replica exchange: N/A — parallel independent trajectories only.

DFT (3). Training DFT only; not a static QM application paper.

Findings

Among tracked hetero-bonds (with hydrogen-mediated contacts treated as non-load-bearing for the film-growth metrics), Fe–C bonding is the most probable contact class, supporting a picture in which aryl carbons anchor TCP-derived fragments to the ferrous interface. Overall TCP–surface bonding increases with temperature over the sampled window, but the dominant reaction sites reorganize: aryl C5/C6 positions are favored at lower temperatures, whereas 700 K statistics shift probability toward P=O oxygen atoms. The replica ensemble is essential: single short runs cannot explore the regioisomer distribution accessible to TCP on a disordered oxide, so the parallel strategy yields histogram-level insight that would be invisible from one trajectory alone.

Limitations

The oxide morphology is a single disordered realization; shear, load cycling, and multi-molecule crowding typical of full tribological contacts are not simulated. Proof-PDF pagination may differ slightly from the version-of-record PDF on 2019khajeh-j-phys-chem-statistical-analysis for locators and figure numbering.

Reader notes (navigation)