Hydration Mechanism of Reactive and Passive Dicalcium Silicate Polymorphs from Molecular Simulations
Evidence and attribution¶
Authority of statements
Prose below summarizes the J. Phys. Chem. C article identified by doi and pdf_path. It is not new primary claims by this wiki.
Summary¶
Dicalcium silicate (C\(_2\)S) exists as multiple polymorphs with markedly different hydraulic reactivity in cement clinker, yet the atomistic origins of “reactive” versus “passive” behavior remain debated. This Journal of Physical Chemistry C article combines DFT and classical molecular dynamics to compare hydration of β-C\(_2\)S and γ-C\(_2\)S. The study computes cleavage energies for low-index surfaces, constructs Wulff equilibrium morphologies, maps adsorption energy surfaces for water, locates transition-state structures for chemisorption pathways, and runs ~2 ns room-temperature MD to observe spontaneous surface reactions. The overarching thesis is that reactive site density—not only intrinsic per-site reactivity—controls net hydration kinetics for the polymorphs compared.
Methods¶
MD (ReaxFF): LAMMPS Reax/c (build 1 Feb 2014, Computational Details in papers/ReaxFF_others/Wang_Manzano_JPCC_2015_CaSiO2.pdf) simulates β- and γ-C₂S slabs with explicit water in 3D PBC supercells; the article uses external potential / dipole correction language for charged slabs. Production segments are NVT at 300 K, 0.2 fs velocity-Verlet, Nosé–Hoover thermostat (20 fs damping), ~2 ns trajectories on selected surfaces to capture spontaneous surface chemistry. No barostat, controlled pressure, electric field, or enhanced sampling is used for those NVT runs.
Force-field training: N/A — published ReaxFF parametrizations for C–S–H / silicate chemistry are cited and applied.
Static QM / DFT: DMol³ GGA-PBE, DNP basis, 5.5 Å orbital cutoff; Monkhorst–Pack 1×2×1 (β slab) and 1×1×1 (γ slab); external potential in the vacuum region when needed to suppress spurious slab dipoles. Dispersion: N/A — the Computational Details block quoted in papers/ReaxFF_others/Wang_Manzano_JPCC_2015_CaSiO2.pdf does not specify an empirical dispersion add-on for DMol³ (the article later contrasts accuracy with plane-wave literature that included dispersive corrections). Computed quantities include relaxed β/γ bulk (Table 1), surface energies, Wulff morphologies, water adsorption maps, and NEB-style pathways for water activation.
Findings¶
Water dissociation proceeds via multi-step sequences (rotation → dissociation → diffusion) with different barriers on β- vs γ-C₂S facets. γ-C₂S can be locally more favorable for some elementary steps than β-C₂S, but γ exposes far fewer reactive sites on its Wulff shape, limiting net hydration and reconciling slower macroscopic kinetics with locally attractive chemistry (abstract and discussion). Table 1 and related figures compare bulk and surface metrics to experiment and prior work. Polymorph, facet, and reactive-site density are the main explanatory knobs. Ideal single-crystal surfaces omit paste pore fluid, alkali, and grain-boundary effects in real cements. External JPCC reference—not a van Duin parametrization paper.
Limitations¶
Ideal surfaces omit grain boundaries, impurities, and alkali chemistry present in real cement pastes.
Reader notes (MAS / retrieval)¶
Use this page when the question mentions β–γ belite differences, Wulff morphology, or water dissociation sequences on C\(_2\)S surfaces; pair with cement microstructure pages for paste-scale context.
The ~2 ns classical MD complements static barriers in the article.
Relevance to group¶
Cementitious silicate simulation adjacent to group expertise in oxide–water interfaces and reactive modeling culture.
Citations and evidence anchors¶
- DOI:
10.1021/acs.jpcc.5b05257—papers/ReaxFF_others/Wang_Manzano_JPCC_2015_CaSiO2.pdf.