Role of Calcium Silicate Hydrate as Silicon Source in Clinkering: Mineral Evolution and Kinetics

In recent years, Portland cement production in China has exceeded 1.8 billion tonnes per year, while CO₂ emissions from limestone decarbonation and fuel combustion remain high (i.e., ≈1.1 billion tonnes), accounting for ~8% of the national total. Replacing limestone with Ca–rich, non–carbonate solid wastes (e.g., recycled concrete powder, RCP) represents a promising pathway toward low–carbon clinker production. The RCP mainly consists of quartz and hardened cement paste. The RCP with higher paste fractions is reported to have lower clinker formation temperatures. Previous studies indicate that Ca(OH)₂, unhydrated clinker particles, and certain minor elements (e.g., S) can facilitate clinkering. C–S–H is a dominant phase in hardened cement paste (>55%). Upon heating (≈500–700 ℃), C–S–H undergoes dehydration and structural depolymerization/rearrangement, potentially generating reactive Ca–Si intermediates that may affect clinkering reactions. However, most previous work treated a hardened cement paste as an integrated entity, and the specific role of C–S–H in governing clinker mineral formation, reaction kinetics, and microstructural evolution remains insufficiently understood.

To decouple the multiphase effects of RCP, this study used synthetic C–S–H to replace quartz, thereby isolating the contribution of C–S–H to clinkering reactions, phase evolution, and kinetics.

C–S–H was hydrothermally synthesized using Ca(OH)₂ and SiO₂ nanoparticles at a CaO/SiO₂ molar ratio of 1.2. Its thermal evolution was preliminarily characterized by X-ray diffraction (XRD). C–S–H lost a long-range order at ~800 ℃ and recrystallized into C₂S and wollastonite (CS) after heating to 900 ℃. At 1000 ℃, the crystallite sizes were ~15 nm for C₂S and ~50 nm for CS (the Williamson–Hall method).

Five raw meals were designed with a silica modulus (SM) of 2.8, an alumina modulus (IM) of 1.8, and a lime saturation factor (KH) of 0.91. The reference mix used analytical–grade SiO₂ together with CaCO₃, Al₂O₃·2SiO₂·2H₂O, and Fe₂O₃. In the remaining mixes, C–S–H progressively replaced SiO₂ (and the corresponding fraction of CaCO₃), with increasing the silicon substitution ratio of C–S–H to SiO₂ from 25% to 100% (on a Si basis). Pellets (Φ 30 mm × 7 mm) were pressed at 40 MPa, dried at 105 ℃, heated in a muffle furnace at 10 ℃/min to selected temperatures, and followed by air quenching.

The phase evolution of clinkers was quantified by X-ray diffraction (XRD) coupled with Rietveld refinement using the ICSD structure models and ZnO as an internal standard. To suppress the dominant endotherm from CaCO₃ decarbonation and enhance clinkering-related thermal signals at a high temperature, CaO was used instead of CaCO₃ in the tests of thermogravimetry-differential scanning calorimetry (TG–DSC), and the derived kinetic parameters were interpreted comparatively as apparent values. The non–isothermal TG–DSC measurements were performed at 5–20 K/min in N₂, and apparent activation energies and kinetic models were evaluated by the Flynn–Wall–Ozawa (FWO) isoconversional method.

Introducing C–S–H into raw meals produces a pronounced exothermic peak associated with C₂S formation at ~870 ℃, consistent with the low–temperature decomposition and recrystallization behavior of C–S–H. When quartz serves as a silicon source, C₂S formation kinetics are well described by a phase-boundary-controlled (R3) model, with an apparent activation energy of 540 kJ/mol. In contrast, when C–S–H serves as a silicon source, the best-fit model shifts to the Avrami–Erofeev (A2), and the higher activation energy (797 kJ/mol) can be attributed to a structural rearrangement coupled with C–S–H recrystallization. The Arrhenius analysis further shows that lnA increases from 54.5 to 84.0; accordingly, the instantaneous reaction rate (dα/dt) at α = 0.5 and T = 900 ℃ increases from 0.43 to 7.60 min^–1, indicating a great kinetic compensation effect likely associated with the disordered/nanostructured nature of C–S–H and the enlarged effective reaction interface.

Consistently, the XRD–detectable onset temperature for C₂S formation decreases from 1000–1100 ℃ (quartz) to 900 ℃ (C–S–H), and the C₂S content is increased by 10–24% at the same sintering temperatures (i.e., 900–1300 ℃). The incorporation of C–S–H also increases the C₃S content by 5–15% at 1350–1450 ℃. Moreover, a high defect density of C–S–H and nanocrystal formation during heating contributes to crystallite refinement, reducing the C₂S crystallite size from 87 nm to 66 nm and the C₃S size from 160 nm to 121 nm.

Mechanistically, C–S–H enhances clinker burnability via (ⅰ) structural depolymerization/rearrangement at a lower temperatures, (ⅱ) a higher defect density and early formation of nanocrystalline products that promote C₂S/C₃S nucleation, and (ⅲ) a reduced effective CaO transport requirement because C–S–H (at Ca/Si ratio of 1.2) supplies part of the Ca needed for C₂S formation, thereby shortening the effective CaO diffusion distance.

Using C–S–H as a silicon source could markedly accelerate solid-state reaction kinetics (0.43 min^–1→ 7.60 min^–1), lower the XRD-detectable onset temperature of C₂S formation (1000–1100 ℃→ 900 ℃), and promote the formation of both C₂S and C₃S. In addition, the recrystallization products of C–S–H provided abundant nucleation sites and led to refined silicate crystallites. These findings could offer a mechanistic guidance for clinkering with RCP containing high hardened cement paste (C–S–H–rich) fractions toward low–carbon clinker production.