Early-Age Strength Development and Microstructural Mechanisms of CO2-Cured Recycled Weathered Limestone Concrete

2.1. Materials

In this study, weathered limestone was collected from a section of the Gongyu Expressway located in Guoluo Tibetan Autonomous Prefecture, Qinghai Province, China. Its geographical location is shown in Figure 1.

Figure 1:

The rock subgrades of Gonghe-Yushu Expressway is shown in Figure 2. The weathered limestone collected on site was selected from the weathered rock subgrades.

Figure 2:

Weathered limestone was crushed and sieved. Coarse aggregate size was 5–15 mm, and natural river sand (0.15–4.75 mm) was used as fine aggregate; the particle-size distribution curves are shown in Figure 3. Ordinary Portland cement (P·O 42.5) and tap water were used. The mix proportions were w/c = 0.38, coarse aggregate/cement = 1.11, and sand/cement = 2.72, with 1.5% polycarboxylate superplasticizer. Cubic specimens (150 mm × 150 mm × 150 mm) were cast for subsequent curing and testing.

Figure 3:

Cylindrical specimens (50 mm in diameter and 100 mm in height) were prepared following ISRM, SL/T264-2020, and JTGE41-2005. The end surfaces were ground to a smoothness < 0.05 mm, and the maximum deviation in diameter and height was controlled within 0.3 mm.

2.2. Experimental scheme

Different curing ages and carbonization durations were designed to analyze the samples. The curing ages were 3 d, 7 d and 28 d respectively, and the carbonization time was 0 h, 2 h, 4 h and 6 h respectively.

2.3. Experimental procedure

The specimens in this study underwent CO₂ curing, subsequent standard curing, saturation treatment, and a series of laboratory tests. For clarity, the overall experimental workflow and key stages are summarized in Figure 4, while the test coverage and the number of specimens for each test item are provided in Table 1. Detailed procedures for the individual tests are presented in Sections 2.3.1–2.3.4.

Figure 4:

3.1. Mass and UPV changes

The mass and ultrasonic pulse velocity change of concrete under different curing ages and carbonation are important indicators of its microstructure evolution and macroscopic performance. These changes profoundly reflect the hydration process, pore structure change and long-term durability characteristics of the material.

Mass and UPV were monitored at curing ages of 3, 7, and 28 days under CO₂ curing durations of 0–6 h to quantify macroscopic responses to carbonation and subsequent hydration.

Figures 5 and 6 summarize the mass and UPV results of weathered limestone specimens at curing ages of 3, 7, and 28 days under CO₂ curing durations of 0, 2, 4, and 6 h. The bars represent mean values, and the error bars indicate ±SE (n=3).

Figure 5:

Figure 6:

At a given curing age, the mass increases monotonically with CO₂ curing duration. Specifically, relative to the 0 h condition, the mass changes at 2, 4, and 6 h are approximately 1.01%, 1.42%, and 1.79% at 3 d; 0.98%, 1.38%, and 1.77% at 7 d; and 0.99%, 1.38%, and 1.92% at 28 d, respectively. In contrast, at a fixed CO₂ curing duration, the mass differences among curing ages are comparatively small, indicating that, within the investigated window, CO₂ curing duration exerts a more pronounced influence on mass than curing age. In addition, the mass gain exhibits a distinct ‘fast-first-then-slow’ characteristic: most of the increase is achieved within the first 2 h, followed by smaller incremental gains at longer durations. This kinetic signature is consistent with an initially rapid reaction in accessible pore networks, followed by progressive pore refinement and the development of a carbonate-enriched layer that increases transport resistance, causing the process to become diffusion-limited and the marginal uptake to decrease with time (Padmalal et al., 2024; Ning et al., 2025).

UPV shows a complementary trend. At the same curing age, UPV increases with CO₂ curing duration, with relative increases (compared with 0 h) of approximately 0.90%, 2.10%, and 3.85% at 3 d; 1.27%, 3.62%, and 4.22% at 7 d; and 1.74%, 2.83%, and 4.11% at 28 d for CO₂ curing durations of 2, 4, and 6 h, respectively. These increases can be attributed to carbonation-induced pore refinement and localized densification via CaCO₃ precipitation, which improves wave-path continuity and enhances the apparent dynamic stiffness sensed by ultrasonic propagation. In contrast, at a fixed CO₂ curing duration, UPV increases markedly with curing age. For instance, under the 0 h CO₂-curing condition, UPV rises from 3660 m/s at a curing age of 3 days to 4310 m/s at 28 days, corresponding to an overall increase of approximately 17.8%. A comparable age-dominant increase is observed under 6 h CO₂ curing, where UPV increases from 3801 m/s at 3 days to 4487 m/s at 28 days, which is about 18.0%. These results demonstrate that curing age, reflecting the progression of hydration and associated microstructural densification, governs the UPV response, whereas short-term CO₂ curing provides an additional but smaller and consistent enhancement of roughly 4% within 6 h (Xian & Shao, 2021).

The increase in mass during early-age carbonation is primarily attributed to the rapid reaction of CO₂ with the specimen surface. Free calcium hydroxide (CH) reacts with CO₂ to form CaCO₃ and H₂O, followed by decalcification of the C-S-H hydration product, where Ca²⁺ migrates and precipitates as CaCO₃. The carbonation front advances rapidly at the early stage (0–2 h) and slows as a CaCO₃-rich surface layer, resulting in the characteristic “fast-first-then-slow” growth of mass. The UPV increases slightly due to pore sealing and local densification induced by the formation of continuous carbonate filling zones.

During standard curing, the slow increase in mass from 3 to 28 days is mainly driven by ongoing hydration. Hydration consumes free water and generates additional products, increasing the volume fraction of the gel phase and reducing capillary space. However, this hydration-induced mass gain is relatively minor compared with the mass increase caused by early-age carbonation. In contrast, UPV increases markedly with curing age because continuous hydration refines the pore structure and densifies the interfacial transition zone, enhancing the propagation of ultrasonic waves.

3.2. Stress-strain curve analysis

Figure 7 presents the stress-strain curves and corresponding uniaxial compressive strength of weathered limestone concrete specimens under different curing ages and CO₂ curing durations, in combination with Table 3. Among them, (a) (d), (b) (e), (c) (f) are the uniaxial stress-strain curves and error bar diagrams at the curing age of 3 d, 7 d and 28 d.

Figure 7:

At an early curing age of 3 days, the uncarbonated specimen exhibits the lowest overall stress-strain response, with a peak stress of 16.58 MPa. The elastic region is short and nearly non-linear, while the post-peak decline is relatively gentle, indicating a loosely hydrated microstructure. With increasing CO₂ curing time to 2 h and 4 h, the peak stress rises significantly to 19.72 MPa and 21.09 MPa, respectively, and the slope of the elastic region increases, reflecting enhanced elastic modulus and material stiffness. Extending the CO₂ curing to 6 h further increases the peak stress to 22.02 MPa, although the increment slows, demonstrating the typical marginal diminishing effect of carbonation curing. The corresponding bar chart shows that CO₂ curing from 0 h to 6 h increases strength by 32.81%, confirming that early-age carbonation plays a decisive role in rapid strength development. This effect is primarily due to the reaction of CO₂ with Ca(OH)₂ to form CaCO₃, which fills pores and densifying the ITZ, thereby enhancing compressive performance.

At 7 days of curing, a similar trend is observed, with overall strength increased compared to 3 days. The uncarbonated specimen reaches 20.02 MPa, while the specimen carbonated for 6 h attains 24.15 MPa. Although the absolute strength gain remains significant, it is less pronounced than at 3 days, reflecting the progression of hydration. The ongoing formation of C-S-H gel and partial filling of capillary pores increases the inherent density of the system, reducing the relative contribution of carbonation. The stress-strain curves show more pronounced linear elastic behavior before peak stress and slightly increased elastic modulus, while the post-peak slope remains steep, indicating enhanced brittleness and limited energy dissipation.

At 28 days, the mechanical behavior is dominated by hydration. The uncarbonated specimen achieves 36.30 MPa, and the 6 h carbonated specimen reaches 38.38 MPa, an increase of only 5.73%. This clearly demonstrates the hydration-dominant effect, where long-term hydration primarily determines the mechanical properties, and additional carbonation has limited impact. The limited effect of prolonged carbonation at this stage is attributed to two factors: (1) the formation of a dense CaCO₃ outer layer during early carbonation, which partially restricts water and ion transport, and (2) the gradual filling of capillary pores by hydration products, which reduces the available space for CO₂ reactions. Consequently, extending carbonation time beyond early stages does not significantly enhance strength.

Figure 8 presents representative fracture patterns of weathered limestone concrete cylinders tested under uniaxial compression, for which splitting-dominated and shear-dominated failures are commonly reported as typical modes in cylindrical compression. Specimen (a) fails predominantly by axial cracking, with one or more major cracks propagating parallel to the loading direction, which is characteristic of a splitting-type failure. In contrast, specimen (c) displays a distinct inclined fracture band that traverses the cylinder, indicating a shear-dominated failure mode. Specimen (b) exhibits a mixed failure pattern, where longitudinal splitting coexists with inclined cracking, suggesting a transitional response between the splitting- and shear-dominated end members. The three specimens correspond to curing ages of 3, 7, and 28 days, respectively.

Figure 8:

Overall, the strength gain from CO₂ curing is prominent at early age but becomes marginal at later age; meanwhile, longer CO₂ curing tends to increase pre-peak stiffness and steepen the post-peak drop, indicating increased brittleness likely related to pore refinement and carbonate precipitation in the ITZ (Rostami et al., 2012; Li, Z. et al., 2019).

These observations confirm that early-age carbonation not only accelerates strength development by filling pores and reinforcing the ITZ but also affects the mechanical behavior and failure patterns of the concrete.

3.3. Results of XRD analysis

Under varying curing conditions, the mineral composition changes of concrete samples were examined by XRD. The XRD patterns are presented in Figure 9, with panels (a), (b), and (c) corresponding to curing ages of 3 days, 7 days, and 28 days, respectively. Figure 10 illustrates the mineral composition of weathered limestone under different curing regimes: panels (a) and (b) show the results at a curing age of 3 days with carbonation durations of 0 h and 6 h, respectively; panels (c) and (d) correspond to a curing age of 28 days with carbonation times of 0 h and 6 h, respectively.

Figure 9:

Figure 10:

Figures 9 and 10 report the crystalline-phase evolution of weathered limestone concrete exposed to CO₂ curing for 0–6 h at curing ages of 3, 7, and 28 d. Phase identification is supported by diagnostic reflections under Cu Kα radiation: portlandite Ca (OH)₂ is tracked by its characteristic peaks near 2θ of 18.1°, 34.1°, and 47.1°. Calcite CaCO₃ is tracked by its major reflection near 29.4° together with secondary reflections such as 39.4° and 43.2°. Dolomite CaMg(CO₃)₂ shows a characteristic reflection near 31.0°. Quartz SiO₂ is evidenced by the strong reflections near 20.8° and 26.6°.

At a curing age of 3 d, the XRD patterns in Figure 9 (a) show a clear attenuation of portlandite reflections at about 18.1° and 34.1° as CO₂ curing duration increases, while carbonate-related reflections become more prominent, especially the calcite reflection near 29.4° and its accompanying peaks around 39.4° and 43.2°. The corresponding phase fractions in Figure 10 indicate that CaCO₃ increases from 55.9% at 0 h to 59.6% at 6 h, whereas CH decreases from 1.5% to 0.8%. These coupled changes provide direct evidence that Ca (OH)₂ is converted to CaCO₃ during carbonation, with portlandite being one of the first hydration products to undergo carbonation and form carbonate phases.

At a curing age of 7 d, Figure 9 (b) retains the same direction of change with increasing CO₂ curing duration, namely progressive weakening of CH reflections and strengthening of carbonate reflections. However, the contrast between adjacent CO₂ curing durations is smaller than at 3 d, as reflected by the more gradual change in peak prominence across the 0–6 h series. This observation indicates that the incremental phase conversion achievable within a short CO₂-curing window decreases with age.

At a curing age of 28 d, CH reflections in Figure 9 (c) are weak, and the remaining evolution with CO₂ curing is primarily manifested as a further increase in carbonate-related reflections together with a marginal additional reduction of CH. Figure 10 quantifies this net change between 0 h and 6 h at 28 d, where CaCO₃ increases from 58.0% to 60.5% and CH decreases from 0.8% to 0.6%. The smaller magnitude of CH depletion at 28 d compared with 3 d indicates that short-term carbonation produces a reduced additional conversion of crystalline CH at later age, while carbonate accumulation continues but at a lower marginal rate within the same CO₂-curing duration window.

In summary, the XRD results show a coupled trend of portlandite depletion and carbonate enrichment with increasing CO₂ curing duration, as reflected by the systematic weakening of CH reflections and the intensification of carbonate-related reflections (Shao et al., 2006). The magnitude of these changes is more pronounced at early age and becomes progressively less evident at later ages, indicating a reduced incremental crystalline-phase conversion within the same short CO₂-curing window as curing age increases.

3.4. NMR test results and analysis

The evolution of pore structure in recycled weathered limestone concrete is influenced by both carbonation and curing age. The NMR T₂ spectra and porosity measurements provide a quantitative basis for elucidating the underlying pore structure development mechanisms. According to NMR theory, the relationship between the T₂ relaxation time and pore size can be expressed by the following equation:

• ρ₂ - the transverse surface relaxivity,

• r_c - the pore radius,

• F_s - the geometric factor,

• T₂ - relaxation time.

Based on this relationship, the T₂ relaxation times can be classified into three categories: short relaxation (0.01–10ms), medium relaxation (10–100ms), and long relaxation (100–10,000ms). The relaxation time reflects the degree of water confinement within the pore: shorter relaxation times correspond to more strongly bound water and, consequently, smaller pore sizes.

The T₂ spectra of the samples under different curing conditions are shown in Figure 11. Where, (a), (d); (b), (e); (c) and (f) are the T₂ spectrum and porosity diagram of the sample at the curing age of 3 d, 7 d and 28 d, respectively.

Figure 11:

From the perspective of carbonation duration, increasing CO₂ curing time (0–6 h) shifts the T₂ spectra toward shorter relaxation times: the short-T₂ peak intensity increases while the long-T₂ peak area decreases, indicating that larger or connected pores are progressively refined. This is mainly driven by CaCO₃ precipitation from the carbonation of Ca(OH)₂, which deposits in capillary pores and the ITZ and reduces pore connectivity. The porosity reduction is most evident at 4–6 h, whereas further densification of the surface layer may hinder CO₂ ingress and slow the subsequent reduction rate (Li, Z. et al., 2019; Xian & Shao, 2021).

From the perspective of curing age, the porosity of specimens at 3, 7, and 28 days exhibits a continuous decrease, with the main peak of the T₂ spectrum shifting toward shorter relaxation times. This long-term trend is primarily attributed to ongoing hydration, during which unhydrated C₃S and C₂S react to form C-S-H gel and additional Ca(OH)₂. These hydration products fill capillary and some macropores, reducing overall porosity. Under CO₂ curing, part of the CH is further converted into CaCO₃, producing a synergistic effect of crystal deposition and gel filling. With increasing age, this synergy between hydration and carbonation products progressively densifies the pore structure. By 28 days, the T₂ spectrum shows that peaks in the long relaxation time region almost disappear, and porosity stabilizes, reflecting the long-term equilibrium of the hydration-carbonation coupling reaction.

In summary, CO₂ curing primarily governs the rapid transformation of pore structure at early ages through CaCO₃ deposition, while curing age controls the long-term trend of porosity reduction and matrix densification. The combined action of hydration and carbonation products explains both the shift from long to short relaxation times in T₂ spectra and the gradual decline in porosity. This microstructural optimization provides a basis for improved durability and mechanical properties, highlighting the complementary roles of CO₂ curing and age growth in enhancing concrete pore structure.

4.1. Carbonation Depth and Duration

For concrete materials, predicting durability through service-life evaluation is of great engineering significance. Regarding carbonation resistance, according to Fick’s first law of diffusion, the carbonation depth is proportional to the square root of carbonation time, expressed as:

• x - the carbonation depth,

• k - the carbonation coefficient influenced by multiple factors,

• t - the carbonation duration.

Figure 12 shows the phenolphthalein colour boundary used to determine carbonation depth (pink: uncarbonized; colourless: carbonized) for CO₂ curing durations of 0, 2, 4, and 6 h.

Figure 12:

As shown in Figure 13, the carbonation depth curve yields a carbonation coefficient k of 2.60[mm/√d], indicating favorable carbonation resistance of the concrete.

Figure 13:

4.2. Mechanism analysis

In the carbonation experiment, carbon dioxide diffuses from the surface to the inside through the concrete pore system, and dissolves in the pore liquid phase to form carbonic acid, which then undergoes a series of chemical reactions with the cement hydration products. The process follows the gas-liquid-solid three-phase reaction mechanism, in which Ca (OH)₂ is the main reactant, and CaCO₃ crystal is formed by carbonation reaction and releases water. At the same time, the silicate phase such as C-S-H gel also undergoes decalcification decomposition, resulting in the degradation of its cementitious properties.

As illustrated in Figure 14, the stick models represent the schematic progression of the main carbonation and hydration reactions of recycled weathered limestone concrete with increasing CO₂ curing time and curing age. Labels (a), (b), (c), and (d) correspond to reactions (3), (4), (5), and (6), respectively, highlighting the distinct carbonation-hydration processes at different stages.

Figure 14:

The main hydration reactions of the recycled weathered limestone concrete at different stages are summarized as follows:

By integrating the results from Figures 7, 9, and 11, the evolution of the matrix can be summarized as follows:

At 0 h carbonation, the 3-day specimen exhibits insufficient hydration, low C-S-H connectivity, and limited CH reserve. The T₂ spectrum shows a dominant peak at longer relaxation times with an extended tail, while the uniaxial stress-strain curve displays the lowest peak stress and elastic modulus. In contrast, at 7 d, the matrix is denser, the T₂ main peak shifts to shorter times, the tail is reduced, and mechanical properties increase. The 28-day specimen shows the most mature structure, with a flatter XRD background, minimal porosity, a left-shifted T₂ peak, and the highest mechanical indices.

At 2 h carbonation, all ages enter the “CH-prioritized carbonation” stage. XRD reveals increased CaCO₃ peaks and decreased CH peaks, with amplitude growing with age. T₂ spectra show systematic left-shifts and increased peak intensities, accompanied by attenuation of the long-time tail, indicating pore refinement and reduced connectivity due to CaCO₃ nucleation and filling. Correspondingly, the uniaxial curves exhibit significant increases in peak stress and initial tangent modulus, with the 28-day specimen showing the largest gains, reflecting the combined effect of sufficient CH supply and favorable pore topology for CaCO₃ deposition.

At 4 h carbonation, CH carbonation continues and reaches the stage of maximum densification. XRD shows the clearest contrast between CaCO₃ and CH, and T₂ spectra exhibit prominent short-time peaks with minimal long-time tails at 7 d and 28 d, indicating that CaCO₃ deposition has transitioned from point nucleation to sheet-like penetration, significantly reducing macropore connectivity. The uniaxial curves show improved strength and stiffness, with negligible post-peak softening, suggesting optimized load-transfer pathways within the matrix. At this stage, densification and ITZ strengthening dominate mechanical performance, while C-S-H decalcification and microcrack formation remain minimal.

At 6 h carbonation, CH carbonation persists, but age-dependent differentiation emerges. In 7-day and 28-day specimens, surface-layer C-S-H decalcification begins, CaCO₃ content further increases but shows convergence, and CH reduction is no longer linear. T₂ spectra still shift left, but with reduced rate, and some cases show shoulder peaks or minor long-tail rebound, indicating partial shortening of the effective gel connectivity network. Uniaxial curves exhibit stabilized or slightly decreased peak stress and steeper post-peak softening, reflecting toughness reduction due to surface decalcification and micro-scale volumetric changes.

In summary, early-age hydration defines the baseline availability of reactive Ca²⁺ and pore structure. early-age carbonation is initially dominated by CH carbonation, with the 2 h reaction being most pronounced. At 4 h, pore refinement and load-bearing chain optimization are achieved, while 6 h carbonation induces more pronounced C-S-H decalcification, leading to increased brittleness post-peak. Overall, early-age carbonation significantly enhances early-age strength, but its effect gradually diminishes with increasing age, demonstrating the characteristic “marginal diminishing effect” (Han et al., 2020).

5.1. Interpretation of Results

Across curing ages, mass uptake, UPV, and compressive response show a consistent cross-scale signature under different CO₂-curing durations. The measured CO₂ uptake is about 1–2% and exhibits a fast-first-then-slow trend with diminishing returns beyond 4 h, which is consistent with accelerated carbonation transitioning from reaction control to diffusion limitation due to near-surface CaCO₃ accumulation and self-sealing (Padmalal et al., 2024; Han et al., 2024; Ning et al., 2025). The same transport-limited behavior is also consistent with reported RH sensitivity and later-stage plateauing kinetics (Ashraf & Olek, 2016; Saeki et al., 2025). In parallel, UPV increases by up to 4% with CO₂ curing, indicating enhanced phase continuity and higher apparent stiffness, which aligns with porosity-reduction-driven elastic-property gains observed in carbonated cementitious materials (Zhang et al., 2020; Li et al., 2019).

Benchmarking against ordinary concrete. At 28 d, the uncarbonated weathered-limestone concrete in this study reached 36.30 MPa, which is comparable to or higher than the conventionally water-cured normal-strength Portland cement concrete reported by Wang et al. (2022), where the 28 d compressive strength of the water-cured control ranged from 26.73 to 31.58 MPa depending on cylinder size; accordingly, the present 28 d baseline exceeds that ordinary-concrete range by 4.72–9.57 MPa (approximately 14.9–35.8%) (Wang et al., 2022). In Li et al. (2019), the conventionally moisture-cured reference at 28 d is reported at approximately 69.8 MPa, but the corresponding mixture design adopts a low water-to-binder ratio (w/c = 0.35) and a high cement content (450 kg/m³), which places it in a substantially higher strength-grade regime; therefore, it should be interpreted as a high-strength benchmark rather than a direct counterpart to the present cold-region recycled weathered-limestone system (Li et al., 2019).

XRD and pore-related observations corroborate the macroscopic trends described above. The XRD patterns indicate a progressive consumption of portlandite together with an enrichment of carbonate phases, which is consistent with the carbonation reaction forming calcium carbonate (including different polymorphs) during CO₂ curing (Shao et al., 2006). The NMR results suggest pore refinement, as evidenced by a shift of the T₂ distribution toward shorter relaxation components. This observation is consistent with carbonate precipitation filling and partially bridging pores, while ongoing hydration further consumes capillary pore space, which together helps explain the simultaneous increases in UPV and early compressive strength (Li et al., 2019; Zhang et al., 2020).

5.2. Research Significance and Limitations

From an aggregate-suitability viewpoint, the recycled weathered limestone is porous and absorptive, which may disturb the local effective w/c ratio near the ITZ and reduce early stiffness/strength if unmanaged. Water-state control such as short-duration pre-saturation is an established strategy to stabilize workability with limited strength penalty (García-González et al., 2014). In addition, CO₂-based treatments for recycled aggregates have been shown to reduce absorption and improve ITZ integrity via CaCO₃ filling and interface reconstruction, providing a mechanistic basis for using short-term CO₂ curing as a compensation route for porous aggregates (Li et al., 2020; Wu et al., 2022; Lu et al., 2023; Bergmans et al., 2024).

The engineering motivation is that crushed-rock embankments in permafrost corridors are sensitive to clogging by aeolian deposits and weathering products, which degrades convective cooling and motivates local recycling of maintenance wastes into value-added materials (Chen et al., 2018; Yu et al., 2020; Kong et al., 2019). From a sustainability standpoint, CO₂ mineralization can deliver measurable but bounded net benefits that depend on process energy and CO₂ logistics (Driver et al., 2024; Zajac et al., 2022), and recycled-concrete mineralization has demonstrated regional-scale storage potential and industrially compatible operation windows (Rosa et al., 2022; Izoret et al., 2023; Suescum-Morales et al., 2023). Therefore, combining recycled weathered limestone utilization with staged CO₂ curing offers a plausible pathway for cold-region road materials, while comprehensive aggregate qualification and durability validation remain the key next steps.