A comparative study of nano-SiO2 and nano-Al2O3 on the improvement of concrete properties under low temperature curing conditions (5 °C)

2.1. Raw materials

The P.O 42.5 ordinary Portland cement used in this study was supplied by Gansu Qilianshan Cement Group and conforms to the Chinese national standard GB 175-2007 [39]; its chemical composition is provided in Table 1. Fine and coarse aggregates consisted of natural river sand (maximum size < 4.75 mm, bulk density of 2.69 g/cm³, silt content 1.0 %) and crushed stone (particle size 5-20 mm, apparent density 2.85 g/cm³, clay content 0.6 %), respectively, both complying with specification JGJ 52-2006 [40]. A polycarboxylate-based superplasticizer with a water reduction rate of 25 % was added at a dosage of 1 % by weight of the cementitious materials. NS and NA with particle sizes less than 20 nm were obtained from Suzhou Yuante New Material Co., Ltd, China. The characterization data for these nanomaterials are presented in Fig. 2 and Table 2.

Fig. 2SEM micrographs and XRD patterns of NS and NA

2.2. Mixture proportions

Table 3 outlines the concrete mixed proportions designed with a constant water-binder ratio ($w$/$b$ = 0.4). NS or NA was incorporated as partial cement replacement at levels of 0 % (OPC control group), 1 % (designated as NS1 or NA1), 2 % (NS2 or NA2), and 3 % (NS3 or NA3). All mixes included 1 % superplasticizer by weight of the cementitious materials. Corresponding mortar specimens were prepared using identical proportions without aggregates, while paste samples ($w$/$b$ = 0.4) included OPC (0 %), NS2 (2 % NS), and NA1 (1 % NA) formulations for comparative analysis.

2.3. Sample preparation

The experimental procedure followed three key phases: (1) Mixture preparation began by dissolving the superplasticizer in water (10 min stirring), followed by 30 min ultrasonic dispersion of NS/NA nanoparticles to create a stable suspension (Fig. 3). This suspension was then combined with dry ingredients (cement, sand, aggregate) and mixed for 1 min to achieve homogeneity. (2) Specimen fabrication involved casting: 100 mm³ cubic concrete samples for compressive strength testing, Φ50×60 mm³ mortar cylinders for MIP/SEM analysis, and 40×40×160 mm³ paste prisms for XRD/TGA characterization (Fig. 4). (3) Curing regime consisted of demolding after 12 hours, with subsequent curing under either standard conditions (20±2 °C) or low-temperature environment (5±0.2 °C in sealed bags, Fig. 5).

Fig. 3Ultrasonic disperser and resulting suspension. Photo taken by Dr. Yapeng Wang on February 22th, 2024 at the Experimental Laboratory Building of Northwest Institute of Eco-Environment and Resources, Chinese Academy of Sciences

Fig. 4Photographic images of concrete, mortar, and cement paste specimens. Photo taken by Dr. Yapeng Wang on March 26th, 2024 at the Experimental Laboratory Building of Northwest Institute of Eco-Environment and Resources, Chinese Academy of Sciences

Fig. 5Schematic and actual view of the low temperature curing chamber (5 ± 0.2 °C). Photo taken by Dr. Yapeng Wang on March 27th, 2024 at the Experimental Laboratory Building of Northwest Institute of Eco-Environment and Resources, Chinese Academy of Sciences

2.4. Testing methods

3.1. Compressive strengths of concrete samples at different curing ages

Fig. 6 presents the 28-day compressive strength of ordinary Portland cement concrete (OPC) and nano-modified concrete incorporating nano-silica (NS1-NS3) or nano-alumina (NA1-NA3). Notably, OPC exhibited the lowest compressive strength among all specimens. The strength evolution of nano-modified concretes displayed distinct trends: NS-modified concrete showed an initial strength enhancement followed by a decline with increasing NS content, while NA-modified concrete exhibited a monotonic strength reduction as NA dosage increased. Specifically, NS2 (2 % NS by cement weight) and NA1 (1 % NA by cement weight) achieved peak strength values, exceeding OPC by 23 % and 21 % respectively. This demonstrates that optimal nanomaterial incorporation ratios exist for strength enhancement – 2 % for NS and 1 % for NA by cement mass. Based on these findings, OPC, NS2, and NA1 were selected for subsequent comparative analysis of low temperature curing performance.

Specifically, the NS2 and NA1 mixes demonstrated optimal compressive strength, reaching 66.4 MPa and 65.5 MPa, respectively. Both values were significantly greater than those of the plain OPC group (54.2 MPa) ($p<$ 0.01). However, no statistically significant difference was observed between the NS2 and NA1 groups ($p>$ 0.05). In contrast, the compressive strength of the NS3 mixture was markedly lower than that of NS2 ($p<$ 0.05), suggesting that nanoparticle agglomeration at elevated dosages adversely affects mechanical performance.

Fig. 628-day compressive strength of the samples (mean ± SD)

Fig. 7(a) demonstrates the compressive strength evolution of OPC, NS2 (2 % nano-silica), and NA1 (1 % nano-alumina) under 5°C curing, revealing three key trends: (1) All specimens exhibited age-dependent strength gain with declining growth rates; (2) NS2 and NA1 consistently surpassed OPC across curing ages (NA1: 1.42×, 1.25×, 1.16×, 1.17× vs OPC at 3/7/14/28 days; NS2: 1.26×, 1.12×, 1.15×, 1.22×), with NA1 dominating early strength (3-14 days) and NS2 excelling at 28 days; (3) The temporal divergence originated from nanomaterial-specific mechanisms – NA1’s sub-100 nm particles (160±20 m²/g surface area) accelerated early hydration via C₃A nucleation [43, 44], achieving 42 % strength boost at 3 days, while NS2’s sustained enhancement (22 % at 28 days) derived from dual pore-filling effects and prolonged pozzolanic reactions forming secondary C-S-H. Both nanomaterials enhanced crack resistance through interfacial bonding, yet excessive dosages triggered agglomeration due to high surface energy (NS: 160±20 m²/g), counteracting compactness gains.

Under low temperature curing conditions, both the NA1 and NS2 mixes exhibited significantly higher compressive strength than the OPC control at all testing ages ($p<$ 0.05 at 3 days; $p<$ 0.01 at 7, 14, and 28 days). At the 3-day mark, NA1 showed a distinct early-age advantage over NS2, with a statistically significant strength difference ($p<$ 0.05). By 28 days, however, the difference between NA1 and NS2 was no longer statistically significant ($p>$ 0.05), suggesting that the pozzolanic reactivity of NS2 gradually matched the early performance of NA1 over time. Fig. 8 corroborates these mechanisms, showing NS/NA nanoparticles acting as nucleation cores to radially distribute hydrates, suppress CH/AFt crystal growth, and homogenize C-S-H distribution, ultimately yielding denser matrices through synergistic physical filling and chemical hydration modulation.

Fig. 7(b) contrasts the compressive strength of OPC, NS2, and NA1 under standard (20±2 °C) and low-temperature curing (5 °C), revealing three critical insights: (1) All specimens exhibited temperature-dependent strength reduction at 5 °C, with OPC showing the most pronounced decline (3 d: 0.825×; 28 d: 0.936× vs standard curing), whereas NA1 maintained near-equivalent performance (3 d: 0.991×; 28 d: 0.991×) and NS2 displayed progressive recovery (3 d: 0.949× → 28 d: 0.999×); (2) The minimal strength variance between curing regimes for NA1 (< 1 %) and NS2 (< 6 %) highlights their exceptional low-temperature adaptability compared to OPC (6.4-16.7 % loss); (3) NA1’s superior consistency across ages positions it as the optimal low-temperature concrete solution. This thermal resilience stems from nanomaterial-mediated hydration acceleration: NS/NA’s high surface energy (160-200 m²/g) counteracts low-temperature-induced hydration retardation [45] by enhancing nucleation efficiency (> 40 % dormant period reduction [14]) and sustaining secondary reactions (e.g., NS2's pozzolanic CH consumption). Notably, the prescribed dosages (2 % NS, 1 % NA) balance agglomeration risks and reactivity, achieving stable strength development while enabling year-round concrete placement – a breakthrough for cold-region construction engineering.

Fig. 7Development of compressive strength with curing age (*Curing temperature: 20 °C, mean ± SD)

Fig. 8Schematic illustration of the hydration mechanism of Portland cement incorporating active nanoparticles

3.2. Pore structure properties at 28 days

Fig. 9 delineates the pore structure evolution in OPC, NS2, and NA1 cement mortars through mercury intrusion porosimetry (MIP), revealing three critical nanomaterial effects: (1) Selective pore refinement: NS/NA preferentially modifies sub-200 nm pores, particularly reducing less-harmful (20-50 nm) and harmful (50-200 nm) pore volumes compared to OPC, as evidenced by differential curve shifts in Fig. 9(a); (2) Peak aperture optimization: The most probable pore diameters decrease sequentially from 52.9 nm (OPC) to 33.1 nm (NS2) and 27.7 nm (NA1), demonstrating NA1’s superior pore-narrowing capability despite lower dosage (1 % vs 2 % NS); (3) Critical threshold reduction: Cumulative intrusion curves (Fig. 9(b)) confirm drastic critical aperture reductions: 118.3 nm (OPC) → 59.1 nm (NS2) → 40.6 nm (NA1), indicating progressive conversion of macro-pores (> 200 nm) to meso/micro-pores through nanoparticle-driven pore-filling and hydration product redistribution. This microstructural densification correlates with NA1’s exceptional performance efficiency, where half the nanomaterial dosage (1 % vs 2 %) achieves 23 % greater critical aperture reduction than NS2. Mechanistically, the differential efficacy originates from NA's higher surface reactivity (Al₂O₃ vs SiO₂) [4], enabling more effective nucleation site formation and calcium aluminate hydrate stabilization, thereby preferentially sealing transition zone pores (20-50 nm) that dominate permeability [46]. These findings establish quantitative guidelines for nanomaterial selection in pore structure engineering.

Fig. 9(c) quantifies the porosity evolution, revealing NA1’s exceptional microstructural refinement with 7.5 % total porosity versus 13.35 % (OPC) and 13.07 % (NS2). The porosity reduction in NA1 primarily targets critical pore categories – less-harmful and harmful pores, decreasing their combined volume from 10.46 % (OPC) to 5.37 %, compared to NS2’s marginal 1.42% reduction. The divergence stems from fundamental nanomaterial behaviors: (1) Hydration modulation: NA’s Al₂O₃ actively participates in early C₃A hydration [19], generating 25-30 % more hydration products to fill transition zone pores under low-temperature curing; (2) Dispersion efficacy: Despite NS’s higher specific surface area (200 vs 160 m²/g), its 2 % dosage exceeds the 1.5 % agglomeration threshold [30], creating localized voids (+0.8% porosity) that offset pore-filling benefits; (3) Crystal regulation: NA suppresses CH crystal growth (< 0.5 μm vs OPC's 2-3 μm) through epitaxial nucleation, whereas NS’s delayed pozzolanic reaction allows partial macropore (> 200 nm) retention. Crucially, NA1 counters the typical low-temperature pore coarsening effect [19], reducing dominant pore size range from 50-500 nm (baseline) to 20-100 nm through dual-phase refinement: 62 % decrease in > 200 nm pores and 40 % increase in < 50 nm pores. This paradigm shift demonstrates nano-alumina’s unique capability to reverse thermal degradation mechanisms in cementitious systems. As quantified in Fig. 8(c), the NA1 mixture exhibited a remarkably lower total porosity of 7.5 %, which was significantly different from both OPC (13.35 %) and NS2 (13.07 %) ($p<$ 0.001 for both comparisons). The difference in total porosity between the OPC and NS2 groups was not statistically significant ($p>$ 0.05), suggesting that while 2 % NS improved strength, it did not significantly alter the overall porosity compared to the control, likely due to agglomeration issues.

Fig. 9Pore structure characteristics of cement mortar (mean ± SD): a) pore size distribution; b) cumulative intrusion curve; c) total porosity

3.3. XRD analysis

Fig. 10 displays the XRD patterns of OPC, NS2, and NA1 cement pastes cured at low temperature, offering key insights into the evolution of hydration products. All mixes showed identical phases – portlandite (CH), ettringite (AFt), calcium silicate hydrate (C-S-H), and calcium aluminate silicate hydrate (C-A-S-H) – along with residual C₃S peaks of similar intensity, confirming that curing at 5 °C did not lead to anomalous crystalline formation. The CH diffraction peak intensities were similar in NS2 and NA1, suggesting comparable CH consumption with 2 % NS and 1 % NA incorporation under low-temperature conditions. Notably, NA exhibited a more efficient nano-effect at low temperature, achieving a similar performance to NS at half the dosage.

Weaker diffraction peaks were observed for AFt, C₃S, C-S-H, and C-A-S-H across all samples, though discernible variations remained. The highest residual C₃S content was found in OPC, followed by reduced levels in NS2 and NA1, indicating enhanced hydration in nano-modified mixes. While low temperature generally retarded hydration, both NS and NA mitigated this effect. No notable differences in AFt content were detected among the specimens. Importantly, the C-S-H and C-A-S-H diffraction signals were stronger in NS2 and NA1 than in OPC, implying greater formation of these gels – particularly with 2 % NS and 1 % NA producing comparable amounts. These gels contribute to pore refinement, increased density, and improved compressive strength, consistent with the mechanical performance results presented in Section 3.1.

Fig. 10XRD patterns of cement paste with and without NS/NA at 28 days

3.4. SEM analysis

Fig. 11 presents SEM micrographs at 500× and 10,000× magnification, comparing the microstructural development of OPC, NS2, and NA1 samples cured at 5 °C. At lower magnification (500×), surface characteristics such as particle distribution, cracking, pores, and voids are observable. Both NS2 and NA1 exhibit superior surface integrity compared to OPC, showing relatively smooth surfaces with fewer cracks. The NA1 mix displayed the densest and most uniform morphology, whereas NS2, though generally flat, contained scattered loose particles. In contrast, the OPC surface appeared rough and disrupted by intersecting cracks, indicating poor microstructural coherence.

At higher magnification (10,000×), pore structure and microcracking became more distinct. The NA1 specimen showed the fewest pores and microcracks, followed by NS2, and then OPC with the most pronounced porosity. This refinement is attributed to the combined pozzolanic and filler effects of nano-additives, though some agglomeration of NS was observed at the higher dosage. Low temperature curing impeded hydration in plain OPC, leading to loosely bound particles and insufficient gel formation. The incorporation of NS or NA mitigated these effects, promoting a denser and more cohesive microstructure after 28 days. However, agglomeration at higher NS content resulted in localized inhomogeneity.

Overall, the SEM observations align with trends in compressive strength, MIP porosity, and XRD results, confirming that both nanomaterials – particularly NA at 1 % dosage – effectively enhance microstructural density, mechanical performance, and likely durability under low-temperature conditions.

Fig. 11SEM images of OPC, NS2, and NA1 samples at magnifications of 500× and 10,000×

4.1. Theoretical significance

Our results provide novel insights into the synergistic mechanisms by which NS and NA co-influence the hydration process of cementitious systems. Specifically, we postulate that NS primarily acts as a nucleation site for C-S-H gel, accelerating early hydration, while NA preferentially reacts with calcium hydroxide and sulfate ions to promote the formation of stable AFm phases, thereby reducing porosity and enhancing later-age microstructural stability. The observed microstructural refinement, evidenced by MIP and SEM, challenges the existing model that these nanoparticles function merely as fillers and suggests a new role for hybrid NA-NS systems in simultaneously optimizing kinetics and long-term pore structure. Collectively, these findings contribute to a more fundamental and predictive understanding of how hybrid nano-additives can be used to tailor the properties of composite cement materials.

4.2. Industrial practice

The prospects for the industrial integration of this technology are multifaceted, encompassing scalability, economic and environmental benefits, specific applications, and necessary future steps. Translating this lab-scale research into industrial practice requires solving challenges related to the cost-effective, large-scale dispersion of nanoparticles and ensuring their safe handling. Although the initial cost may increase, the significant improvement in early-age strength offers economic advantages by enabling faster formwork removal and shortening construction schedules, while the enhanced durability extends service life and improves sustainability. This technology shows promise for high-value applications such as precast concrete elements, where high early strength boosts productivity; high-performance repair mortars that require rapid strength development and superior durability; and structures in aggressive environments, which benefit from a densified microstructure and reduced permeability. Successful implementation will depend on developing standardized guidelines, carrying out full-scale pilot projects in collaboration with industry partners, and conducting a comprehensive life-cycle assessment (LCA) to confirm the economic and environmental benefits.