Comparing hydrothermal sintering and cold sintering process: mechanisms, microstructure, kinetics and chemistry
Summary¶
ZnO is densified by hydrothermal sintering (HS, closed) vs cold sintering (CSP, open) at 155 ± 5 °C and 320 MPa with dwell times 0–80 min. Both routes reach very high relative density with similar ZnO structure, but differ in stabilized side phases, grain size evolution, and trapped solvent/organics. Zinc acetate “bridges” appear in HS samples. ReaxFF simulations compare HS-like vs CSP-like solvent environments and connect acetate clustering to observations (e.g., Zn–Zn pair correlations in clusters, pathways toward acetate crystals consistent with XRD/SEM).
Methods¶
- Sintering: HS vs CSP as in the abstract—155 ± 5 °C, 320 MPa, dwell 0, 20, 40, 80 min; characterization includes XRD/Rietveld, electron microscopy, Raman, TG-MS, etc. (per section 2).
- ReaxFF setup: Nonpolar ZnO (10̄10) wurtzite slab (14 layers, 3528 atoms); solvent with 100 Zn²⁺, 50 H₂O and two acetates per Zn²⁺ (total 16500 atoms in solvent region) for interface equilibration at 300 K; CSP-like boxes add vacuum normal to the surface while keeping liquid density; HS-like boxes heated toward 500 K with 0.002 K/fs ramp.
- Potential & integrator: ZnO/acetic-acid/water ReaxFF from cited prior work with reoptimization for sub/supercritical acetic-acid–water mixtures; MD in ADF; 0.25 fs timestep; velocity Verlet; Berendsen thermostat (0.1 ps damping).
- Extended clustering run: Surface removed; 20100-atom solvent (100 Zn²⁺, 200 acetate, 6200 H₂O) simulated 1 ns to probe hydrothermal clustering.
MD protocol (same ReaxFF runs, complementing the bench sintering above). All trajectories are 3D cells in Amsterdam Density Functional (ADF) using ReaxFF-based reactive molecular dynamics with PBC for the wurtzite (10̄10) ZnO slab and the bulk-solvent boxes. Integration uses 0.25 fs timesteps, velocity Verlet, and a Berendsen thermostat (0.1 ps damping) under NVT-like control. CSP-like and HS-like solvent boxes differ by vacuum along the surface normal and, for HS-like cells, ramped heating (0.002 K fs⁻¹ toward 500 K in the paper). Barostat / NPT: N/A — these interfacial and clustering MD stages are not the uniaxial NPT sintering in the autoclave (155 ± 5 °C, 320 MPa; dwell 0, 20, 40, 80 min). Electric field: N/A — not used. Shear / shock: N/A — not used. Replica or enhanced sampling: N/A — not used. 2 — Force-field training: N/A — a published ZnO–acetic-acid–water ReaxFF is reoptimized/used for sub/supercritical acetic–water conditions as cited, not a new full parameterization in this work. 3 — Static QM: N/A — primary quantum benchmarks are via cited prior work and comparison to experiment, not a new DFT surface study here.
Findings¶
Relative densities above 98% are reported at 0 min dwell for both hydrothermal sintering (HS, closed) and cold sintering (CSP, open), but later density trends are not monotonic in the same way for the two routes—so a single “best dwell” is route-dependent. XRD, microscopy, Raman, and TG-MS separate oxygen-related defect populations, grain structure, and zinc acetate signatures; HS samples show bridging / zinc acetate features that CSP does not reproduce under the same T/P path. The ReaxFF MD then connects HS-like solvent and temperature histories to acetate-mediated Zn clustering and pair-correlation behavior consistent with larger zinc acetate crystals in HS-processed pellets—a mechanistic story for microstructure divergence even when final relative density is high in both routes at 155 ± 5 °C and 320 MPa with dwells 0, 20, 40, 80 min. The authors’ comparison stresses reactor path (closed hydrothermal vs open CSP), humidity, and trapped organics, not just peak density. As they frame it, short interfacial and 1 ns clustering simulations do not stand in for full autoclave kinetics, grain growth, or volatile escape at the reactor scale.
Limitations¶
Simulation sizes/times capture precursor clustering chemistry, not full pellet-scale sintering kinetics; ReaxFF for ZnO–acetate–water is specialized to this interface chemistry.
Relevance to group¶
Penn State MRI/MSE/ME collaboration (Randall, van Duin, Ndayishimiye, Sengul, Bang, Tsuji) with international coauthors.