An Atomistic Carbide-Derived Carbon Model Generated Using ReaxFF-Based Quenched Molecular Dynamics
Abstract
ReaxFF quenched MD in LAMMPS generates nanoporous carbide-derived carbon models whose structure (g®, pores, rings) and adsorption properties are compared to experiment, including optional post-quench compression.
Summary¶
The authors generate atomistic models of carbide-derived carbon (CDC) using ReaxFF reactive molecular dynamics in LAMMPS (version 27-Jul-2013). Initial configurations place 20,000 carbon atoms at random positions in a periodic cell of 7.488 nm side length (~0.95 g/cm³ initial density). Quenched MD runs from 3500 K to 3000 K with quench durations 5–500 ps (rates 1–100 K/ps). The ReaxFF parameterization targets carbon materials from the literature. Thermostat: Nosé–Hoover with 10 fs damping; ensemble NVT; timestep 0.5 fs (larger than typical 0.25 fs because only carbon is present). Optional NPT compression after quenching uses an NPT integrator at 3000 K and 20,000 atm with 10 fs temperature and 100 fs pressure damping (timestep 0.5 fs) to shift pore statistics toward experimental targets. Structural diagnostics include pair distribution functions, pore-size distributions, adsorptive properties vs experiment, ring statistics (including non-hexagonal rings), and comparison of compressed vs uncompressed models.
Methods¶
Force-field training / fitting. The study uses a published ReaxFF parameterization for carbon aimed at nanoporous / carbide-derived carbon chemistry; the article does not report a new QM refit in this work.
MD application (atomistic dynamics). Simulations use LAMMPS (27-Jul-2013) with ReaxFF. Initial states place 20,000 carbon atoms at random positions in a cubic cell of side 7.488 nm (~0.95 g/cm³ initial density), i.e. three-dimensional periodic boundary conditions in the disordered bulk model. Quenched MD cools 3500 K → 3000 K with quench durations 5–500 ps (1–100 K/ps). Production quenches use NVT integration with a Nosé–Hoover thermostat (10 fs damping) and timestep Δt = 0.5 fs (the authors note this is larger than typical 0.25 fs settings because only carbon is present). An optional post-quench compression stage uses NPT at 3000 K and 20,000 atm with 10 fs temperature damping and 100 fs pressure damping (same 0.5 fs timestep). N/A — static electric fields are not applied. N/A — no umbrella sampling, metadynamics, or other enhanced sampling; the protocol is direct quench (plus optional compression). Structural and adsorption analysis uses g®, pore-size distributions, ring statistics, and comparison to experimental adsorption data as reported in the article.
Static QM / DFT. N/A — DFT is not the engine used to generate the CDC models summarized here.
Review / non-simulation framing. This is the primary C (MDPI) research article (DOI 10.3390/c3040032). A separate corpus slug may hold a proof PDF sibling: [[2017matthew-w-thompson-we-report-a-atomistic-carbide-derived]].
Findings¶
Outcomes and mechanisms. Quenched ReaxFF MD yields disordered, nanoporous carbon models whose pair correlations, pore statistics, and adsorption-related metrics can align with experimental CDC data, especially when the optional high-pressure NPT compression step after quenching is included. Ring statistics retain substantial non-hexagonal character relative to simple graphitic motifs, as the authors emphasize for CDC-like disorder.
Comparisons. The manuscript compares simulated structure and adsorption behavior to experimental references for CDC-type carbons (tables and figures in the PDF are authoritative for numerics).
Sensitivity and design levers. Quench rate (1–100 K/ps over the stated temperature window) and the presence or absence of the post-quench compression stage shift pore-size distributions and related observables; follow the article for quantitative trends.
Limitations and outlook (as authored). High-temperature quenches and extreme compression are computational expedients to access rearrangement kinetics and target porosity; they are not literal replicas of experimental synthesis paths—see ## Limitations below for KB context.
Corpus honesty. Claims here follow the indexed PDF at pdf_path and the article text; refresh after any corpus PDF swap.
Limitations¶
High-temperature quench and compression steps are motivated by accessible kinetics in MD and are not direct replicas of experimental synthesis paths; timestep and temperature ranges should be interpreted in that context.
Relevance to group¶
Adri C. T. van Duin is a co-author; the work demonstrates ReaxFF-based CDC model building and validation against experiment.
Citations and evidence anchors¶
- DOI:
10.3390/c3040032