Skip to content

Simulating the Geological Fate of Terrestrial Organic Matter: Lignin vs Cellulose

Source

Prose here summarizes the Energy Fuels article at pdf_path. For full protocols, parameters, and figures, use the PDF and SI.

Summary

Shale gas recovery models need nanoscale descriptions of kerogen matrices that adsorb and trap hydrocarbons in tortuous nanopores. Atmani et al. apply replica-exchange molecular dynamics to compare geological maturation of lignin versus cellulose—two abundant terrestrial biopolymers that contribute differently to sedimentary organic matter—into kerogen-like condensed phases plus methane. The abstract reports that lignin yields roughly twice as much kerogen and five times more methane than cellulose under the simulated protocol, consistent with pyrolysis experiments. Even when average composition and bonding statistics appear similar, ex-lignin kerogen is about an order of magnitude more compliant and microporous than ex-cellulose kerogen, implying different effective transport properties in pore-network models of shale where compliance controls fracture response and microporosity controls accessible surface area.

Methods

1 — MD application (LAMMPS + ReaxFF + REMD). Interactions use ReaxFF (Reax2013 C–C parameters together with C–O, C–H, O–O, O–H, H–H sets cited in the paper) in the large-scale atomic/molecular massively parallel simulator (LAMMPS) package.

  • System size & composition. I-β cellulose crystal, ~4200 atoms, O/C = 0.83, H/C = 1.67. Lignin pack: 18 oligomers from a softwood lignin model (Crestini et al., as cited), O/C = 0.34, H/C = 1.11, in orthorhombic cells. Both use 3D periodic boundary conditions.

  • Boundaries / periodicity. 3D PBC throughout.

  • Pre-REMD equilibration. NPT at 423 K and 25 MPa until the cell volume converges (about 100 ps for the disordered lignin case; relaxation details differ slightly by precursor).

  • Ensemble, timestep, integrator. NVT and NPT legs as described in Energy Fuels: Nose–Hoover thermostat; Nose–Hoover–Andersen barostat for NPT legs. Temperature and pressure coupling time constants 0.05 and 0.5 ps, velocity–Verlet integration, 0.1 fs timestep.

  • REMD (enhanced sampling). NVT sampling on an exponential temperature ladder from 423 K to 3500 K with 96 replicas; replica exchange attempts every 10 fs with ~20% acceptance. NVT is used for REMD to avoid unphysical first-order transitions at fixed volume; because volatiles increase internal stress, the authors interrupt runs and re-equilibrate the 423 K replica (and reinitialize all replicas) after NPT relaxation at 25 MPa (e.g. 100 ps). Cellulose: when pressure on the 423 K replica exceeds 200 MPa, a relaxation stage is applied (three such stages in their protocol). Lignin: scheduled NPT relaxations every 200 ps. Cumulative REMD until a near-equilibrium state at 423 K takes about 750 ps (cellulose) and 1300 ps (lignin) for the initial confined-pyrolysis leg; a second REMD after fluid removal runs 1800 ps (see paper for the fluid-expulsion protocol).

  • Barostat & pressure. NPT at 25 MPa during the relaxation/reinitialization windows described above. N/A — no external static electric field across the cell.

  • Shear, shock, AIMD cutoffs, QEq. N/A — not the focus; standard ReaxFF/QEq usage per the manuscript.

2 — Force-field training. N/A — the study applies published Reax2013 and related ReaxFF parameter files referenced in the article, rather than reporting a new fit in this work.

3 — Static QM. N/A — reactivity and thermochemistry come from ReaxFF/REMD, not a standalone DFT results section (QM references in the text support comparisons and field context, not a separate static-QM “application” block for maturation).

Analysis (brief). Molecular clusters are binned (gas C₀–C₄, light tar, heavy tar, kerogen/char) using bond cutoffs in the Chemical Analysis section; kerogen/char structure metrics include H/C, O/C, coordination, RDFs, and ring / pore descriptors as reported.

Findings

Yields (abstract). Lignin produces about twice as much kerogen and ~5× more methane than cellulose, consistent with pyrolysis experiments referenced in the paper.

Structure–property. Despite similar average composition and bonding statistics, ex-lignin kerogen is about an order of magnitude more compliant and more microporous than ex-cellulose kerogen—morphology (pores, compliance) diverges where chemical averages do not.

Modeling use. The authors position the results as nanoscale building blocks for bottom-up shale models that couple organic texture to recovery when combined with continuum fracture/flow solvers.

Corpus honesty. Consult papers/Atmani_Energy_Fuels_2020.pdf and SI for full replica statistics and any chemical kinetic checks; this page stays at abstract-level yield/modulus claims.

Limitations

Simplified biopolymer models and idealized thermal histories omit mineral catalysis, confinement pressure, clay surfaces, and basin-specific kerogen variability that alter absolute yields in reservoir rocks.

Relevance to group

van Duin as co-author on organic matter diagenesis and kerogen properties, relevant to geochemistry and energy applications of MD.

Citations and evidence anchors

papers/Atmani_Energy_Fuels_2020.pdf — abstract (yield ratios, modulus/porosity contrast). https://doi.org/10.1021/acs.energyfuels.9b03681 — full Methods and SI tables should be consulted for replica-exchange parameters and statistical convergence checks.