Modeling the Polymerization Process for Geopolymer Synthesis through Reactive Molecular Dynamics Simulations
Summary¶
This Journal of Physical Chemistry C study simulates geopolymer gelation—the solution-mediated formation of aluminosilicate networks used in cementitious geopolymer binders—using reactive molecular dynamics with the Feuston–Garofalini aluminosilicate reactive potential rather than ReaxFF. The authors first use density functional theory to prerelax reactive silicate and aluminate monomer motifs that seed the MD, then heat and condense systems at Si:Al ratios of 2 and 3 over a 650–1800 K temperature window to capture oligomerization, aggregation, and condensation stages. The scientific goal is to connect atomistic network topology statistics—Si–O–Al connectivity and Si(nAl) distributions—to radial distribution functions and bulk density trends that can be compared with experimental geopolymer literature benchmarks cited in the paper.
Methods¶
The workflow couples DFT-relaxed monomer inputs to large-scale reactive molecular dynamics using the Feuston–Garofalini reactive aluminosilicate potential (bond-making/breaking formalism distinct from ReaxFF). Simulations sweep temperature (650–1800 K in the abstract) and compare Si:Al ratios of 2 and 3 to interrogate oligomerization, aggregation, and condensation kinetics plus gel microstructure. Analysis emphasizes Si–O–Al connectivity and Si(nAl) distributions alongside RDFs and bulk density vs experiment-focused literature benchmarks cited in the paper (papers/Others/Zhang_Tao_JPCC_2018_Geopolymer_Garofalini.pdf).
MD application (Garofalini RMD)¶
Engine / code: Reactive MD with the Feuston–Garofalini potential (N/A — LAMMPS vs in-house integrator name not on indexed excerpt—confirm in Methods). System & PBC: periodic aluminosilicate melt/gel supercells built from DFT-prerelaxed monomers (atom totals in article). Ensemble: NVT/NPT staging for the heating and condensation protocol is defined in Methods (exact labels per stage in PDF). Thermostat / barostat / timestep: N/A — not transcribed here beyond the 650–1800 K temperature window quoted in the abstract. Duration / stages: equilibration and production segments with lengths in ps/ns accompany the three-stage oligomerization → aggregation → condensation narrative in Methods. Pressure: N/A — not highlighted in the abstract-level summary used here. Electric field: N/A — not used. Enhanced sampling: N/A — not indicated for the geopolymerization protocol summarized in the abstract.
Findings¶
The authors report that computationally synthesized geopolymer-like gels can reproduce key Si(nAl) distribution signatures and RDF features relative to selected experimental references. Higher temperature accelerates condensation kinetics and yields lower final bulk density in the modeled gels, while lower Si/Al produces denser networks under the reported simulation trends. The paper is an important contrast case in this corpus because it demonstrates reactive MD for geopolymer chemistry using a specialized silicate potential rather than the van Duin-group ReaxFF ecosystem.
Corpus honesty. This work is not a ReaxFF parameterization; cite reactive-md-generic tooling and the Garofalini potential sections in the PDF for reproducibility details.
Limitations¶
Community tooling and transferability for Feuston–Garofalini-class potentials are narrower than for ReaxFF; alkali-ion chemistry, long-time curing, and full water activity in real geopolymers may require additional model extensions and validation beyond the published protocol.
Reproducibility notes¶
Geopolymer simulations are sensitive to initial hydroxyl speciation, alkali content (if added in extended models), and the heating schedule used to drive condensation. Readers should reproduce the authors’ temperature ramps and composition endpoints exactly before comparing RDF or Si(nAl) statistics across software ports. When cross-walking to ReaxFF cementitious literature, avoid mixing bond-order definitions: the aluminosilicate reactive form differs materially from ReaxFF’s charge equilibration and bond-order formalism. The DFT prerelaxation stage is load-bearing: small changes in silanol/aluminol protonation states in the optimized monomers can shift subsequent condensation cascades, so archive the exact quantum settings used to build MD initial states.
Relevance to group¶
Important contrast case: reactive MD for geopolymers using Garofalini-family potentials rather than ReaxFF.