Carbon sequestration mechanisms and performance evolution of alkali-activated slag during the mixing process

2.1.1. Precursor material

GGBFS is a by-product of ironmaking in blast furnace processes. It is rapidly quenched with water to form an amorphous glassy structure, which reveals latent hydraulic reactivity. Its main chemical composition consists of CaO, SiO₂, Al₂O₃, and MgO, in which CaO and SiO₂ serve as the dominant phases that provide essential calcium and silicon sources for subsequent alkali-activation reactions. In this study, the GGBFS has a specific gravity of 2.90 g/cm³ and a specific surface area of 564 m²/kg. The activity index corresponds to Grade 120, in accordance with ASTM C989/C989M-25 “Standard Specification for Slag Cement for Use in Concrete and Mortars” [20]. The material was supplied by Jinyu Industrial Co, Ltd. Table 1 shows the chemical composition and physical properties of the material. The relatively high CaO content of the GGBFS is particularly relevant to the proposed CO₂-coupled alkali activation mechanism. Under alkaline conditions, calcium-containing phases in the slag can dissolve and release Ca²⁺ into the reaction environment. The availability of Ca²⁺ provides a favorable chemical pathway for the reaction with dissolved CO₂ and subsequent formation of CaCO₃. Therefore, the Ca-rich composition of the slag can be considered advantageous for in situ CO₂ mineralization. However, CaO content alone does not determine the carbonation capacity of the material, because Ca²⁺ availability also depends on the dissolution kinetics of the slag, the alkalinity of the activating solution, pore structure, and CO₂ transport. Therefore, the high CaO content should be regarded as a favorable chemical characteristic for carbonation rather than as an independent indicator of carbonation performance.

2.1.2. Alkali activators

Sodium hydroxide (NaOH) with a purity of 95% was used as the primary alkali activator to provide a highly alkaline environment, promoting the dissolution and depolymerization of the slag precursor. Sodium silicate (Na₂SiO₃) solution was employed as both a silica source and a supplementary activator, enabling adjustment of the SiO₂/Na₂O ratio in the activator solution. This ratio plays a critical role in controlling the formation and structural evolution of hydration and geopolymeric reaction products. The sodium silicate solution contained 32.2% SiO₂ and 13.18% Na₂O, with a Baume degree (°Bé) of 50.

2.1.3. CO₂

Carbon dioxide gas with a purity of approximately 99.9% was supplied from a high-pressure gas cylinder. The gas was injected into the reaction system during carbonation experiments to evaluate the carbon uptake behavior and sequestration capacity of the AAMs.

2.1.4. Fine aggregates

Natural fine aggregates were used in this study, with SiO₂ as the dominant constituent. The specific gravity and water absorption were 2.65 and approximately 1.2%, respectively. The particle size distribution (Figure 2) followed the standard requirements for fine aggregates, with a maximum particle size of 4.75 mm and a fineness modulus of 2.47. The particle-size distribution of the fine aggregate is an important factor in determining particle packing and the resulting pore structure of the mortar. A well-graded aggregate can improve packing efficiency by reducing the volume of interparticle voids. Consequently, it decreases the amount of paste required to fill the aggregate skeleton. This can influence water demand, porosity, water absorption, and ultimately compressive strength. Therefore, the aggregate grading was considered not only for compliance with the relevant specification but also to maintain a consistent particle-size distribution. This was necessary to minimize variations in packing and isolate the effects of alkali activation and CO₂ incorporation.

Figure 2

2.3.1. Gravimetric evaluation of CO₂ sequestration capacity

The CO₂ sequestration capacity of the specimens was evaluated using a gravimetric method based on mass variation before and after carbonation. Immediately after mixing under vacuum conditions, the initial mass of each specimen (W₀) was measured as the baseline value and recorded. The specimens were subsequently exposed to a CO₂ atmosphere to promote carbonation reactions, during which CO₂ reacted with reactive components within the matrix to form carbonate products. After the carbonation process was completed, the specimens were removed, and their post-carbonation mass (W₁) was measured under the same conditions. The estimated CO₂ uptake was calculated from the relative mass increase between the pre- and post-carbonation states according to equation (1).

2.3.2. TGA test method

TGA was employed to evaluate the mass change behavior of the materials during heating. This enables characterization of the thermal decomposition behavior of reaction products and quantification of carbonate content to estimate the CO₂ sequestration capacity of AAMs. TGA effectively distinguishes mass loss behaviors across various temperature ranges, including the dehydration of hydration products (such as C-(A)-S-H gel and bound water) and the thermal decomposition of carbonate phases (mainly CaCO₃). To remove the background influence of non-carbonate components, an uncarbonated specimen was used as a blank reference. The net CO₂ uptake was estimated from the difference in LOI between carbonated and uncarbonated specimens, as expressed in equation (2).

2.3.3. Microstructural analysis

This study employed SEM coupled with Energy dispersive X-ray spectroscopy (EDS) to investigate the microstructural evolution and elemental composition of AAMs under different alkali contents and carbonation conditions. The AAS mixtures were first mixed under vacuum conditions to minimize interference from atmospheric gases and moisture. CO₂ gas was then introduced directly into the mixing chamber during the designated carbonation period. This allowed CO₂ to interact with the freshly activated slag paste during mixing. After reaching the designated curing age, the specimens were placed in an oven at 110°C ± 5°C for 24 h to stop further reaction before sampling. SEM/EDS specimens with approximate dimensions of 10 × 10 × 5 mm were then prepared from the dried material for microstructural observation and elemental analysis. TGA specimens were similarly dried and subsequently ground into powder before thermal analysis.

SEM was used to observe surface morphology, pore structure, formation of reaction products, and the distribution of unreacted particles. In addition, SEM was used to evaluate the extent of alkali activation and the densification behavior of the microstructure. Observations at multiple magnifications were performed to compare microstructural variations before and after CO₂ exposure, including the development of C-(A)-S-H gel, precipitation of CaCO₃ crystals, and pore-filling phenomena. EDS analysis was used to provide qualitative and semi-quantitative elemental characterization of localized regions, focusing on elements including Ca, Si, Al, Na, and C. The elemental composition data were used to identify changes in alkali-activation products and carbonate formation, thereby providing additional evidence for interpreting CO₂ sequestration mechanisms and carbonation effectiveness.

2.3.4. Compressive strength and water absorption tests

Compressive strength and water absorption tests were used following the standard ASTM C109/C109M-21 “Standard test method used to measure the compressive strength of hydraulic cement mortars” [21] and ASTM C642-21 “Standard Test Method for Density, Absorption, and Voids in Hardened Concrete” [22], respectively. Cubic specimens with dimensions of 50 × 50 × 50 mm were prepared, and three replicate specimens were tested for each mixture proportion at each designated curing age. Aggregate content was selected as the main experimental variable, and the mass ratios of binder to aggregate were designed as 1:1, 1:2, and 1:3. Different aggregate contents were employed to generate varying pore structures, thereby enabling evaluation of the influence of carbonation on the mechanical performance and pore compaction behavior of AAS materials.

3.1.1. Evolution behavior of CO₂ during the mixing process of AAS paste

Figure 3 presents the mass variation curves of AAS with different alkali contents during the mixing process in a CO₂ environment, which reflect the CO₂ fixation behavior of the paste. The overall results indicate that CO₂ uptake increased continuously with mixing time. This suggests progressive reactions between dissolved CO₂ and reactive species in the system, including Ca²⁺, silicate, and aluminate ions, leading to the gradual formation of carbonate-bearing products. In terms of time-dependent behavior, all mixtures showed a continuous increase in CO₂ fixation with increasing mixing time, demonstrating the kinetic nature of the carbonation process.

Figure 3

The results showed that as the alkali content increased, both the magnitude of mass change and the reaction rate increased simultaneously. This behavior is mainly related to the higher OH⁻ concentration in highly alkaline environments [23], which enhances slag dissolution and accelerates the release of Ca²⁺ and dissolved silicate species [24,25]. Consequently, the system more readily reaches a state of local supersaturation, promoting the formation of carbonate phases such as CaCO₃ [26]. Meanwhile, accelerated formation of C-(A)-S-H gel provided additional reactive interfaces, further facilitating continuous CO₂ uptake at the solid–liquid interface [27,28].

3.1.2. Early-stage reaction of AAS paste under CO₂ environment (30 s)

Figure 4 presents the CO₂ uptake behavior of AAS with different alkali contents and CO₂-exposed AAS paste (AASS) during the early stage (30 s). The AAS samples mixed under a sealed environment showed negligible mass variations, ranging from approximately 0.003 to 0.05%. Since no CO₂ exposure occurs throughout the test, these minor fluctuations were attributed to instrumental noise and transient physical effects, such as fluid redistribution, particle rearrangement, and short-term variations in sensor response, rather than to chemically induced mass changes.

Figure 4

In contrast, AASS specimens showed significant differences in early-stage CO₂ uptake depending on alkali content. Specimens with low alkali contents (2–3%) revealed CO₂ uptake values of 0.22–0.29%, indicating limited early reactivity. This was mainly related to insufficient slag dissolution and restricted ion release under low-alkaline conditions, which delayed the formation of C-(A)-S-H gel and the onset of carbonate reactions.

When the alkali content was increased to above 4%, CO₂ uptake increased significantly to about 0.42–0.46% and gradually stabilized, indicating that the system entered a more effective reaction regime. At this stage, the higher alkalinity enhanced the slag dissolution and accelerated the release of Ca²⁺ and dissolved silicate species, promoting local supersaturation. This facilitated the formation of C-(A)-S-H gel and carbonate phases, thereby shifting the CO₂ fixation behavior from a kinetically limited regime to a more stable reaction regime.

3.1.3. Maximum carbon sequestration capacity of AAS under CO₂ mixing conditions with different alkali contents

The results showed clear differences among specimens with varying alkali contents in the maximum CO₂ uptake behavior under continuous CO₂ exposure. Specimens with low alkali contents (2–3%) showed relatively limited maximum CO₂ uptake capacities of 0.28 and 0.49%, respectively, indicating poor long-term carbon sequestration performance. Increasing the alkali content to 4% markedly increased CO₂ uptake capacity over exposure time. Among all mixtures, the AASS7 specimen (7% alkali content) revealed the best performance. The sample reached a total CO₂ uptake of approximately 1.9% after approximately 300 s of continuous carbonation (Figure 5). The results suggested that a highly alkaline environment effectively enhanced slag dissolution kinetics and increased the supply of reactive ionic species. This enabled the system to sustain both carbonate precipitation and gel formation under prolonged CO₂ exposure conditions, thereby improving overall carbon sequestration capacity.

Figure 5

3.2.1. Quantitative evaluation of thermal analysis

It is important to distinguish between the 30 s and maximum CO₂-exposure conditions used in this study. The 30 s condition represents the practical CO₂ injection period incorporated into the 90 s mixing procedure and was therefore used to evaluate early-stage CO₂ fixation under a realistic mixing condition. In contrast, the maximum CO₂ uptake was determined under continuous CO₂ exposure for approximately 300 s to investigate the cumulative carbonation capacity of the material beyond the practical mixing period. Therefore, the higher CO₂ uptake observed under the maximum-exposure condition is expected because the system was allowed substantially more time for slag dissolution, CO₂ dissolution, Ca²⁺ release, and carbonate precipitation. The two conditions were used to evaluate different aspects of the carbonation process and should not be interpreted as equivalent exposure conditions. Figures 6 and 7 illustrate the TGA and derivative thermogravimetric (DTG) curves of AAS with different alkali contents under various CO₂ exposure conditions. The TGA results indicated continuous mass loss during the heating process in all specimens. However, the overall mass reduction rate did not show any significant changes related to either alkali content or CO₂ exposure conditions. Further interpretation of the DTG curves revealed a clearer identification of characteristic thermal events in different temperature ranges. Distinct peaks were observed at approximately 100°C, 600°C, 750°C, and 800°C, which are related to different dehydration and decomposition processes in the alkali-activated matrix and its carbonation products.

Figure 6

Figure 7

In the low-temperature range (approximately 60–160°C), the observed mass loss was primarily associated with the evaporation of physically bound water and the dehydration of reaction products [9,16]. This mass loss is mainly attributed to the release of structural water from C-(A)-S-H gel and other hydration/activation products, resulting in the initial dehydration peak. The DTG curves mainly reflected continued decomposition of hydration products within the intermediate temperature range (approximately 300–600°C). In AAS systems, this region is generally associated with progressive dehydration of C-(A)-S-H gel and other metastable phases, producing a broader and more gradual mass loss profile [23]. In addition, if calcium hydroxide-containing phases are present, dehydroxylation reactions may occur within approximately 425–480°C; however, the contribution of portlandite in AAS systems is typically limited compared with OPC systems [16,23].

At higher temperatures (approximately 600–800°C), noticeable DTG peaks were linked with the thermal decomposition of carbonate phases formed during carbonation. In this temperature range, calcium carbonate (CaCO₃) undergoes decarbonation, releasing CO₂ (CaCO₃ → CaO + CO₂) [29] and generating characteristic mass loss peaks that are commonly used to quantify carbonate content in alkali-activated systems. The presence and intensity of these peaks provide important evidence for carbonate formation and can be used to evaluate the effectiveness of CO₂ sequestration under different alkali contents and carbonation conditions. Furthermore, variations in peak position and intensity may reflect differences in carbonate type, crystallinity, and degree of carbonation within the matrix. The mass loss within the 600°C–800°C range was extracted and converted into CO₂ content (%) using the corresponding calculation method (Figure 6).

Figure 8 shows the mass loss related to carbonate decomposition at a temperature of 600°C–800°C. The results demonstrated that AAS paste has CO₂ sequestration capability and that CO₂ uptake increased with increasing alkali concentration and exposure duration. The AAS2 and AAS4 showed nearly negligible CO₂ content (approximately 0%). In contrast, AASS2 (30 s), AASS4 (30 s), AASS4 (max), and AASS7 (max) showed values of 0.21, 0.48, 1.06, and 1.96%, respectively.

Figure 8

3.2.2. Ca²⁺-dominated carbonate formation mechanism and alkali content effect

The thermal decomposition behavior observed in the TGA/DTG results revealed that the carbon sequestration behavior of AAS is mainly controlled by a Ca²⁺-mediated carbonate formation mechanism, in which the C-(A)-S-H gel structure plays an important role in regulating ion transport and controlling reaction interfaces. GGBFS mainly consists of Ca, Si, Al, and Mg, which are mostly present in oxide forms such as CaO, SiO₂, Al₂O₃, and MgO. Under alkaline activation conditions, Na⁺ and OH⁻ increase dissolution and depolymerization of the amorphous slag phase by breaking Si–O–Si and Si–O–Al bonds, resulting in the release of reactive ionic species, including Ca²⁺, Mg²⁺, silicate, and aluminate species. These dissolved species subsequently participate in polymerization and precipitation reactions, leading to the formation of a C-(A)-S-H-dominated gel network while simultaneously increasing the Ca²⁺ concentration within the pore solution, which provides an essential reactive source for subsequent carbonation reactions [9,30–32]. The results revealed that under CO₂ exposure, carbonation was controlled by ionic reactions occurring in the pore solution. Therefore, Ca²⁺ serves as the dominant reactive species. Dissolved Ca²⁺ reacted with carbonate species generated from dissolved CO₂ to form stable CaCO₃ precipitates, which represent the major pathway for carbon fixation [33]. Although Mg²⁺ might also participate in carbonation reactions through the formation of Mg-containing carbonate phases (MgCO₃), its contribution to overall carbon sequestration is generally limited because of slower reaction kinetics and the lower extent of carbonate formation. Therefore, specimens with higher alkali contents show enhanced CO₂ uptake capacity due to accelerated slag dissolution and greater availability of dissolved Ca²⁺ for carbonation reactions. In contrast, Si and Al primarily contribute to the formation of the C–(A)–S–H framework rather than directly participating in carbonate formation. However, they indirectly influence carbonation behavior by regulating pore structure, ion transport pathways, and reaction kinetics.

Furthermore, correlation analysis between the mass-balance method and CO₂ contents derived from TGA demonstrated an approximately linear relationship with a coefficient of determination (R ²) of 0.99, indicating excellent agreement between the two quantification approaches (Figure 9). These results confirm that the mass-based method provided a reliable and reproducible approach for evaluating CO₂ uptake in AAS systems. In addition, this strong agreement supports that carbonate formation by Ca²⁺-driven mineralization governs the overall carbonation behavior of the system.

Figure 9

3.3.1. Compressive strength

Figure 10 presents the compressive strength results of carbonated AAS systems (alkali dosage of 4%) with different slag-to-aggregate ratios (1:1, 1:2, and 1:3). The results indicated the highest compressive strength in specimens with a slag-to-aggregate ratio of 1:2. However, the 1:1 and 1:3 mixtures showed comparatively lower mechanical performance. The lower strength observed in the 1:1 mixture is due to the relatively high binder content, where strength development is mainly controlled by the gel matrix and might be more sensitive to shrinkage, pore formation, or matrix heterogeneity.

Figure 10

In contrast, the 1:3 mixture contained a higher aggregate fraction, which reduced the continuity of the binding matrix and limited effective stress transfer within the composite system, while decreasing the overall mechanical performance. The 1:2 ratio appeared to provide a more balanced relationship between binder matrix formation and aggregate framework support, resulting in improved load transfer efficiency and enhanced strength development.

Under similar mixture proportions, after short-term CO₂ exposure (30 s), all specimens showed a slight reduction in compressive strength at 1 day of curing, about 1–9%. This behavior might be due to the rapid carbonation reaction consuming Ca²⁺ in the pore solution at early ages. This reaction can temporarily interfere with initial hydration and geopolymerization processes. However, with increasing curing age, compressive strength gradually recovered and showed a clear increasing trend. At 28 days of curing, all carbonated groups indicated an overall increase in compressive strength of approximately 15–25% compared with the non-carbonated specimens. This suggested that short-term CO₂ exposure could not hinder long-term strength development and might enhance compressive strength. The improvement is mainly attributed to the formation of calcium carbonate precipitates, which act as pore-filling products, refining the microstructure and increasing matrix densification. The results show enhanced compressive strength development under CO₂ curing [34,35]. This suggests a beneficial role of carbonation reactions in improving the performance of alkali-activated systems through microstructural densification. Furthermore, comparable strength of the 1:3 mixture with the 1:2 mixture at 28 days suggests that continued hydration and carbonation reactions partially compensate for the lower initial gel content and reduced matrix continuity in mixtures with higher aggregate fractions.

3.3.2. Water absorption

Figure 11 shows the water absorption results of AAMs under different mix proportions and CO₂ exposure conditions. The water absorption values of all specimens range from approximately 9.8 to 11%, with relatively minor variations among groups. This indicated that both aggregate ratio and short-term CO₂ exposure have a limited influence on the overall water absorption behavior. For the non-carbonated specimens, the 1:3 mix showed the highest water absorption (approximately 11%), which can be attributed to the increased porosity and reduced continuity of the binding matrix associated with higher aggregate content. In contrast, the 1:2 mix indicated relatively lower water absorption, suggesting a more compact pore structure and improved matrix packing. For the CO₂exposed specimens (1:1 C, 1:2 C, and 1:3 C), the water absorption values were slightly lower than their corresponding non-carbonated groups. This trend suggested that carbonation might promote the formation of CaCO₃ precipitates, which partially fill capillary pores and contribute to microstructural refinement, thereby reducing water uptake capacity. Kim et al. [36] reported that reductions in porosity and pore size due to carbonate formation led to slightly lower water absorption [36]. Although their study was conducted on Portland cement paste rather than AAS, the observed relationship between carbonation and pore-structure modification provides supporting evidence for the pore-filling interpretation proposed in the present study. Hwalla et al. [37] stated that water absorption decreased with increasing carbonation-curing duration and pressure. Their results indicated that carbonation conditions can influence the transport-related properties of AAS materials [37]. In addition, the small error bars across all groups indicated acceptable experimental repeatability and stability. Overall, short-term CO₂ exposure leads to a slight reduction in water absorption, suggesting a limited but measurable densification effect induced by carbonation in AAMs.

Figure 11

The correlation between water absorption and compressive strength was further analyzed. The results indicated a clear linear relationship between the two parameters, with a coefficient of determination (R ²) of 0.81, which showed a moderate-to-strong correlation between water absorption and compressive strength (Figure 12). The overall trend showed that higher compressive strength resulted in lower water absorption.

Figure 12

The findings suggested that carbonation-induced CaCO₃ formation reduced water transport pathways and enhanced mechanical performance simultaneously through pore filling and microstructural densification. This reflects a coupled relationship between moisture transport behavior and strength development in AAM systems.

3.3.3. Microstructural analysis

Figure 13 shows the SEM observations of AAS mortars under different carbonation conditions. Image (a) shows the uncarbonated specimen at 500× magnification, where localized microcracks and relatively loose interfacial regions can be observed, particularly around aggregate particles. In contrast, image (b) illustrates the carbonated specimen, in which granular precipitates are clearly distributed within interparticle voids and pore regions. At a higher magnification of 3,000× (image c), numerous white precipitates are observed along the particle interfaces and within the pore spaces, indicating pronounced precipitation and pore-filling characteristics. EDS analysis revealed that these regions are mainly enriched in Ca, C, and O, with dominant Ca and O signals, suggesting that the precipitates are likely associated with carbonate-rich phases, predominantly CaCO₃. The detected Si signal might originate from the surrounding slag matrix or adjacent aggregate particles. Wang et al. [38] demonstrated that carbonation can substantially modify the multiscale pore structure of AAS, while the extent of the modification depends on the characteristics of the slag and the carbonation conditions [38]. The comparison between Figure 13(a, b) provides qualitative evidence of carbonation-induced modification of the interfacial transition zone (ITZ). In the uncarbonated specimen, the ITZ appears relatively open, with localized microcracks around the aggregate particles. After carbonation, the observed precipitates occupy interparticle voids and regions adjacent to the aggregate surfaces, indicating local filling of the interfacial pores and partial closure of microcracks. This indicates that carbonation modifies and densifies the ITZ through carbonate precipitation. This microstructural change is also consistent with the reduction in water absorption and improvement in compressive strength observed after carbonation.

Figure 13