Nano-TiO₂ accelerates cement early but quietly locks its crystals later

Nano titanium dioxide has long been cast in a single supporting role in cement chemistry: an inert template that speeds up hydration by giving calcium silicate hydrate extra surfaces to nucleate on. A new study published in Case Studies in Construction Materials argues that this role is only half the story. Using a battery of techniques spanning isothermal calorimetry, X-ray diffraction, electron microscopy, low-field nuclear magnetic resonance and solid-state ²⁹Si NMR, a research team led by Li-yuan Gao shows that crystalline nano-TiO₂ exerts two opposing effects at different stages of the same hydration process. In the first hours it accelerates crystallization; later, it quietly inhibits the maturation of the very crystals it helped to form.

The distinction matters because the strength and durability of Portland cement rest almost entirely on calcium silicate hydrate, or C-S-H, the binding phase that interlocks as clinker dissolves and hydrates precipitate. Nanomaterials influence this process through two classical mechanisms. The nucleation effect describes nanoparticles acting as additional sites where C-S-H embryos can form, lowering the energy barrier for precipitation and boosting early reaction rates. The pozzolanic effect, by contrast, is chemical: highly reactive nanomaterials such as nano-SiO₂ react secondarily with calcium hydroxide released during hydration, consuming that relatively weak phase and generating additional C-S-H. Traditionally, TiO₂, Fe₂O₃, ZrO₂ and carbon nanotubes have been filed under the first category as physically inert nucleation templates, while SiO₂ and γ-Al₂O₃ belong to the second as pozzolanically active reactants.

What remained unresolved was whether nano-TiO₂ might switch from promotion to inhibition as hydration progresses, and how such a switch relates to the evolving crystallinity of C-S-H. To isolate the question, the team synthesized stoichiometric C-S-H with a calcium-to-silicon ratio of approximately 1 by ball milling mixtures of calcium hydroxide and amorphous nano-SiO₂, introducing 1.5 weight percent nano-TiO₂ in the modified system. This streamlined synthesis eliminated confounding phases such as ettringite and unhydrated clinker, allowing any structural impact to be attributed cleanly to the surface chemistry of the nanoparticle. Amorphous nano-SiO₂, with its well-documented pozzolanic reactivity, served as the benchmark for comparative assessment in real Portland cement pastes prepared at a water-to-cement ratio of 0.4.

The calorimetric evidence reveals a striking temporal profile. During the initial exothermic peak, the 1.5 percent nano-SiO₂ blend produced the sharpest first heat pulse, consistent with its ultrafine particles providing abundant heterogeneous nucleation sites and its anionic silica surface adsorbing calcium ions to drive further clinker dissolution. The nano-TiO₂ blend did not reach that apex, yet remarkably, after only six minutes its curve overtook the nano-SiO₂ trace. The authors attribute this crossover to the pinning of calcium ions in Ti–O⁻ pockets on the particle surface, an interaction that sustains a subdued but prolonged C-S-H precipitation process and keeps cumulative heat release ahead of the other mixtures. Within the first 43.7 hours, the nano-TiO₂ system generated more total heat than either the plain reference or the nano-SiO₂ counterpart; beyond that point, the ranking reversed as the sustained pozzolanic reaction of nano-SiO₂, a strongly exothermic process with an enthalpy of roughly –80 kilojoules per mole, took over thermal output.

Pore structure evolution tracked by low-field NMR told the same story from a different angle. At seven days, the nano-TiO₂ paste showed the greatest volume of gel pores and the lowest total porosity, at 25.32 percent versus 28.15 percent for the control, confirming its superior early-age nucleation potency. The transition-pore signal also shifted to shorter relaxation times, reflecting pore refinement: the high nucleation density induced by nano-TiO₂ produces smaller, more densely packed C-S-H clusters, narrowing the inter-cluster gaps between them. By 28 days, however, the ranking had inverted. The nano-SiO₂ paste drove total porosity to a minimum of 24.96 percent and delivered the lowest combined capillary and macropore fraction, the pores wider than 100 nanometers that universally limit durability by governing permeability, freeze–thaw damage and carbonation. The nano-TiO₂ system, by contrast, exhibited the largest macropore fraction at later age, a consequence of the particle exceeding its dispersion ceiling and forming fractal agglomerates.

The mechanistic core of the study lies in the ball-milled C-S-H synthesis, where X-ray diffraction exposed the dual temporal effects with unusual clarity. In the nanoparticle-free system, patterns collected after one hour showed a single low-intensity peak near 29.2 degrees, corresponding to overlapping (220) and (222) reflections of tobermorite-type C-S-H whose inability to separate indicates poor crystallinity. At two and three hours, partial peak separation occurred alongside a weak (531) reflection, and by five hours the (220) and (222) peaks were fully resolved and shifted to higher angles while new peaks appeared at 18, 34, 47, 50 and 54 degrees. These correspond precisely to xonotlite, indexed as the (200), (310), (330), (420) and (511) planes. The data document a mechanochemically driven phase transformation from layered tobermorite to chain-structured xonotlite, accompanied by an increase in silicate chain polymerization from single to double chains.

In the nano-TiO₂-incorporated samples, the transformation never happened. Patterns collected at one, two, three and five hours consistently displayed the tobermorite (220) reflection near 29.4 degrees alongside the (222) peak, with relative intensities and positions essentially unchanged as reaction time extended. Nano-TiO₂, long regarded as lacking pozzolanic reactivity and valued solely as a nucleation agent or filler, was revealed as a structural stabilizer that imparts metastable phase stability to C-S-H. Electron microscopy corroborated the crystallography. The control samples exhibited the characteristic tobermorite morphology of highly curved, crumpled nanosheets, whereas nano-TiO₂ incorporation induced a fibrous bundle-like or foil-like architecture with enhanced porosity and reduced structural ordering. Transmission electron microscopy of the five-hour products showed the modified system as loosely packed, crumpled nanosheets with large interparticle voids and reduced image contrast, while the nanoparticle-free control formed dense, block-like clusters stacking tightly into uniform agglomerates hundreds of nanometers in size, with smoother edges and higher crystallinity.

Solid-state ²⁹Si NMR supplied the molecular-scale mechanism. The deconvoluted spectrum of pure C-S-H revealed three primary silicate structural units: Q¹(p) near –80 ppm, bridging Q¹(b) tetrahedra near –83 ppm, and pairing Q² tetrahedra near –86 ppm. The pronounced Q¹(b) intensity indicates a tobermorite-type structure in which silicate chains are organized in ordered dreierketten configurations with predominant bridging sites, reflecting relatively high crystallinity. The nano-TiO₂ spectrum changed substantially: a distinct Q⁰ peak corresponding to isolated tetrahedra emerged near –72 ppm, the relative Q²/Q¹ ratio decreased, and overall peak broadening signified markedly enhanced structural disorder with diverse local coordination environments. Together these observations demonstrate that strong interfacial interactions between nano-TiO₂ surfaces and silicate species impede the ordered assembly of silicate chains, inhibiting further polymerization and chain elongation and effectively locking the metastable tobermorite structure against transformation to xonotlite.

The macroscopic payoff is measurable. At 1.5 percent dosage, nano-TiO₂ raised the three-day compressive strength of the paste from 18.9 to 26.3 megapascals, a 39 percent gain, and the 28-day strength from 58.5 to 70.9 megapascals, a 21 percent gain. Nano-SiO₂ outperformed it at equivalent content, reaching 29.3 megapascals at three days and 73.6 megapascals at 28 days, reflecting its combined nucleation, pozzolanic and microfilling effects. The authors also acknowledge limitations: the dosage-dependent phase stabilization effect awaits systematic quantification, comparative studies with other inert nanoparticles are needed to generalize the interfacial stabilization mechanism, and whether the C-S-H–nano-TiO₂ interaction is governed by physical adsorption, chemisorption or epitaxial lattice matching remains to be visualized through in-situ synchrotron characterization, environmental TEM and molecular dynamics simulations.

What the study delivers now is a conceptual recalibration with practical consequences for time-matched nanomaterial deployment in cementitious systems. Nano-TiO₂ acts as a nucleation catalyst within a limited time window, increasing cumulative hydration heat release by 22 percent within 24 hours, while nano-SiO₂ provides sustained pozzolanic reactivity that dominates later stages. The finding challenges the conventional binary classification of nanomaterials as either inert nucleation templates or active reactants, indicating that a single nanoparticle can perform both promotion and inhibition at different stages of the same reaction. It also opens a tantalizing synergy: because nano-TiO₂ is photocatalytic, exploiting its phase-stabilization capacity alongside its light-driven functionality could point toward multifunctional cement-based materials that both regulate their own microstructure and actively degrade pollutants, provided the long-term stability of the locked metastable phase under carbonation, chloride ingress and sulfate attack can be validated in service.

Subject of Research: Materials Science

Article Title: Dual temporal effects of nano-TiO₂ in cement hydration: Crystallization acceleration and inhibition

Article References: Gao, L.-Y., Liu, Y., Li, Y.-X., Wang, W.-H., Liu, X.-C., Chen, M., Wang, X.-N., Li, Z.-J., & Wang, Y.-M. (2026). Dual temporal effects of nano-TiO₂ in cement hydration: Crystallization acceleration and inhibition. Case Studies in Construction Materials, 25, Article e06633. https://doi.org/10.1016/j.cscm.2026.e06633

Image Credits: AI Generated

DOI: 10.1016/j.cscm.2026.e06633

Keywords: Dual, temporal, effects, nano-TiO, cement, hydration, Crystallization, acceleration, inhibition, scientific research