Shandong Kang Crystal New Material Co., Ltd. provided ground granulated blast slag (GGBS) that met S105 grade standards. The chemical composition, density and BET area of GGBS are shown in Table 1 and the particle size distribution of GGBS is presented in Figure 1. Both the aggregates for AASM and the raw materials for WwFS use quartz sand with a particle size of 380-830 μm and a chemical composition of 99% SiO₂, which is milky white or colourless and translucent. The waterglass, model SP38, has a modulus of about 3.3 and was bought from Jiashan Yourui Refractory Co., LTD. Analytically pure KOH and NaOH were supplied from Tianjin Hengxing Chemical Reagent Manufacturing, Co., Ltd.

The production of WwFS is usually higher in industrial production because the use of waterglass in the foundry process is more common. Waterglass is an inorganic cementitious material commonly used as a binder to bind sand grains during the casting process to form the casting model. WwFS are usually subjected to a high temperature thermal field during the casting process, and the shells are heated differently due to the location of the casting, creating a temperature gradient. Using a typical WwFS treated at 100°C guarantees consistency in the treatment temperature of the waterglass layer. According to Fang (22) et al. that real WwFS usually contain more impurities, such as metallic elements and organic matter, which may affect the hydration process of AAS pastes. To ensure that the experiments were controlled, reproducible, and to avoid the influence of impurities such as metallic elements and organics, this study chose not to use actual waste samples, but to use laboratory-prepared WwFS samples and apply them to the AAS system.

In a Harbor mixer, quartz sand, waterglass, and KOH were mixed together for three minutes at a mass ratio of 100: 5: 2, as per Ref (21). Then, pure CO₂ was injected into the mixture for 10 mins at 5 L/min. After hardening, the mixture was cracked and dried at 100°C for 24 h to prepare WwFS, which was synthesized as shown in Figure 2. Also, WwFS with a particle size of 380-830 μm was selected to keep the same particle size as that of the quartz sand.

As shown in Figure 3b, compared with the sharp-edged and smooth-surfaced quartz sand (Figure 3a), the surface of WwFS is encapsulated with a thicker layer of dry waterglass, which is mainly composed of Na₂CO₃, sodium silicate gel, and a small amount of KOH (31). The width of the waterglass layer is about 1-5 μm. WwFS is composed of quartz sand and the waterglass layer on the surface, so its chemical composition is mainly SiO₂, Na₂SiO₃, Na₂CO₃ and a small amount of KOH. Figure 4 displays the particle size distribution of quartz sand and WwFS analyzed with the Malvern Mastersizer 2000 particle size analyzer. Obviously, the volume of WwFS is higher in the range of 760-1450 μm, while the volume of small (380-760 μm) and large (1450-2000 μm) particle sizes decreases. This may be explained by the mixing, hardening and crushing processes that increase the grain size of small particles and the damage to large sand sizes.

The SiO₃ ^2- released from the waterglass coating of WwFS has an important effect on the hydration of AASM. To evaluate the dissolution characteristics of WwFS, the absorbance of Si at 648 nm (32) was tested according to the method of Wang et al (21). Silicate standard solutions of 0, 2, 4, 8, 12, 16 and 20 mg/L were made up using Na₂SiO₃·9H₂O and deionized water. Using a 752UV-VIS spectrophotometer, the absorbance of the standard solution was measured. The final standard curve is shown in Equation (1) and presented in Figure 5.

Where, C is the concentration of Si and A is the absorbance of the solution.

Testing solutions were collected at 1, 3, 6, and 24 hours by adding 500 g of WwFS₂₀ or WwSF₇₀ to 500 mL of deionized water, respectively. The absorbance and silicate concentration of each solution are shown in Table 2. This suggested that the dissolution rate of WwFS₇₀ is obviously higher than that of WwSF₂₀, which may influence the hydration and mechanical properties of AASM.

Six mixtures were designed to explore the influence of the pre-wetting methods of WwFS on AASM performance, 0.55 was the ratio of activator to binder. As shown in Table 3, dried WwFS was used to replace all of the quartz sand in WwFS1. WwFS₂₀ and WwFS₇₀ represent WwFS immersed in water at 20°C and 70°C, respectively. The last number in the designation denotes the time that WwFS was immersed in water. Also, WwFS₇₀1, WwFS₇₀3 and WwFS₇₀6 can be considered as the experimental groups, and the other three groups as the control groups.

Fresh AAS mortars were prepared by the following process. Initially, NaOH was weighed and added into the water and mixed to exotherm it to room temperature, then WwFS was added and soaked for different times at two temperatures. Dry materials were mixed for three minutes using a Harbor mixer, and then wet materials (water for soaking WwFS and wet WwFS) were added and mixed for 3 mins. Two groups, WwFS0 and WwFS1, directly mixed slag, NaOH solution and sand for three minutes.

The fluidity of AAS mortars is shown in Figure 7. It is clear that the use of dried WwFS to replace quartz sand reduces the flow of the mortar by 12.7% compared to WwFS0. WwFS is fine and porous and has a larger specific area than quartz sand (12, 38), which increases the water required for particle lubrication and, thus, reduces the fluidity of the mortar. When WwFS₂₀24 was added, the fluidity of the AAS mortar increased by 10.4%, in comparison to WwFS1. Since the water absorbed by WwFS can participate in the lubrication between particles (39), the fluidity of the mortar increases. Furthermore, the fluidities of all WwFS₇₀-added AAS mortars were lower than that of WwFS1, and the longer the pre-wetting time, the lower the fluidity. Specifically, with increasing the pre-wetting time from 1 h to 6 h, the fluidity decreased by 2.1% to 29.2% compared to WwFS1. The decrease in the fluidity of WwFS₇₀-incorporated mortars may indicate that the presence of WwFS accelerated the early hydration of AASM. It could be attributed to the properties of WwFS, as WwFS₇₀ is not only rich in hydrophilic silica (40), but also WwFS₇₀ is coated with high silica modulus waterglass and KOH, which may release OH^- during the process of hot-water immersion to increase the pH conditions of pore solutions (41). The increased alkalinity in AAS system may accelerate the dissolution process of GGBS and promote the polycondensation of flocculated products (42). Consequently, higher alkaline conditions result in a faster rate of generation of hydration products, which decreases the fluidity of AASM (43).

The compressive strength of AAS mortars is depicted in Figure 8. When WwFS was used to replace all quartz sand, the compressive strength of AASM at all ages increased significantly by 22.1% to 30.8% compared to WwFS0. This may be attributed to the dried waterglass layer that can release a small amount of SiO₃ ^2- into AAS paste, providing an additional source of Si for the formation of hydration products (44). The compressive strength of WwFS₂₀24 was reduced at all ages compared to WwFS1. This suggests that the water absorbed by wet WwFS increased the local w/b ratio of AASM (45), although the water released from pre-wetted WwFS may have contributed to the hydration of the slag, but high w/b dominated the negative effect on compressive strength. When incorporating WwFS₇₀ into AASM, the compressive strength of AAS mortars was higher than that of WwFS1 and WwFS0 at all ages. As the pre-wetting time increased from 1 h to 6 h, the compressive strength of WwFS₇₀ increased by 4.5~10.7% at 3d, 3.7~14.2% at 7d, and 1.3~10.0% at 28d, in comparison with WwFS1. This finding is associated with the dissolution of the waterglass coating with high solubility and the release of ions. Pre-wetting WwFS in hot water dissolved and released more SiO₃ ^2- from the wetted waterglass layer into the mixing water, not just confined around the WwFS, but in a broader and more homogeneous way activating the binding potential of the paste matrix. Moreover, WwFS₇₀ could release SiO₃ ^2- and some KOH from the WwFS coating on the quartz particles into the pore solution, working as an auxiliary activator and forming a synergistic activating effect with NaOH on GGBS. Based on previous research by Wang et al. (21), the dried waterglass coating of WwFS could be a primary limiting factor in AASM compressive strength development. However, the waterglass layer of WwFS dissolved in hot water became thinner, weakening the defects of the sandwich structure, while the residual waterglass layer still played a role in improving the bonding between aggregate and paste (46, 47). Hence, more hydration products could be generated in the interfacial transition zone (ITZ), resulting in a denser ITZ between WwFS and the surrounding paste, thus increasing the compressive strength of AASM.

The flexural strength of AAS mortars is presented in Figure 9. It is clear that the development trend of flexural strength is similar to that of compressive strength of AAS mortars with different pre-wetted WwFS treatment methods. The flexural strength of WwFS1 is higher than that of WwFS0 by approximately 16.2~20.0% at all ages. Although the flexural strength of WwFS₂₀24 is lower than that of WwFS1 due to its high w/b, it is still higher than the flexural strength of WwFS0. In addition, the flexural strength of mortar at all ages increases with the increase of soaking time when the pre-wetting temperature of treated WwFS is kept at 70°C. At 28d, the flexural strength of WwFS₇₀ is increased by 3.1~18.3% compared to WwFS0, where WwFS₇₀6 possesses the highest flexural strength. The increase in flexural strength is attributed to the waterglass coating on WwFS₇₀, which is highly soluble as shown in Table 2. WwFS can release SiO₃ ^2- into the pore solution around the paste, providing additional Si species for the generation of hydration products (48). As a result, more stress-bearing hydration products can be produced and the microstructure of AAS can be densified, resulting in higher flexural strength.

Figure 10a presents the pore structure of AAS mortars. The pore size distributions are very similar for all groups, with a distinct single peak around 10 nm and a flat profile after 100 nm. After adding WwFS1, the peak increased obviously compared to WwFS0. This may be associated with the inherent sandwich-like structure of WwFS, where the interface between the paste and aggregate changes from dual to ternary, leading to additional pores in the microstructure system (21). The mortar with the addition of WwFS₂₀24 was similar to that of WwFS1, which suggests that room temperature pre-wetting for 24 h did not improve the pore structure. The free water and waterglass released by WwFS₂₀24 could probably produce a denser ITZ, but the effect is rather weakened by the locally high w/b. By contrast, with the increase of hot water pre-wetting time, the peak of WwFS₇₀ exhibited a tendency to gradually decrease and move to the left, indicating that WwFS₇₀ refined the pore structure of AASM. Generally, pores smaller than 10 nm mainly consist of gel pores and pores between gel units, the lowest peak of WwFS₇₀6 means that soaking WwFS in hot water results in tighter connection between generated gel (49-52). Moreover, it is also possible that waterglass has entered the paste, generating extra hydration products that fill the pores and cracks, further densifying the ITZ.

Figure 10b further shows the cumulative volume of pores in the three zones. It is evident that replacing quartz sand with WwFS increased the total pore volume of the mortar by 3.3%, indicating that the sandwich structure introduced more pores into AASM. Incorporating pre-wetted WwFS not only reduced the total porosity and macro pores volume, but also optimized the gel pores and capillary pores volume. However, the improvement in the pore structure of mortars by pre-wetting WwFS at 20°C was limited, with a slight reduction in the total porosity of mortars containing WwFS₂₀24 compared to WwFS1. With the addition of WwFS₇₀, the volumes of all three pore size zones decreased, especially the total porosity of WwFS₇₀6 decreased by 30.4% compared to that of WwFS1. Therefore, since the refined pore structure could be produced in AASM using WwFS₇₀, the hypothesis about the auxiliary activation of slag by WwFS₇₀ is further substantiated.

The typical SEM images of 28d AASM are shown in Figure 11. From Figure 11(a), cracks with a width as high as 20 μm were found between the quartz sand and the surrounding paste, which indicates that the interface is a typical physical interface with poor bonding. After substituting quartz sand with WwFS, the cracks between WwFS1 and AAS paste were clearly narrowed. Since the layer of dried waterglass on the surface of WwFS1 may dissolve into the surrounding paste in small amounts, it promoted the hydration of slag, leading to an enhanced bonding effect between WwFS1 and AAS paste. The paste around WwFS₂₀24 is loose and porous, as shown in Figure 11(c), attributed to a significant local increase in w/b near WwFS. This results in the deterioration of the microstructure (53), consistent with the decrease in compressive strength of WwFS₂₀24 in Figure 8. In contrast, AASM with the addition of WwFS₇₀1 exhibited a large number of hydration products in the microstructure, such as C-(A)-S-H gels. Dissolved SiO₃ ^2- and KOH from the hot water-soaked waterglass layer could be released into the mixing water as a source of coactivator in addition to NaOH. Moreover, the pre-wetted waterglass coating could function as reactive aggregates, with the dissolution and permeation of SiO₃ ^2- and some KOH facilitating the hydration of the surrounding slag particles, producing additional hydration products. The enhanced dense ITZ structural layer was observed in WwFS₇₀6 with tight interfaces and no cracks generated, as shown in Figure 11(e). In conclusion, the improved bonding between WwFS and surrounding paste by hot water pre-wetting of WwFS can be confirmed, consequently influencing the compressive strength and pore structure of AAS mortar.

Figure 12 shows the contour mapping of the elastic modulus of AAS mixtures determined by nano-indentation method, respectively. The average elastic modulus of AASM along the Y axis in Figure 12 was calculated to investigate the variations in ITZ properties after incorporating pre-wetted WwFS, as shown in Figure 13. There is no weak zone between the quartz sand and paste, and the elastic modulus is approximately 30~50 GPa. Since the dried waterglass layer may cause the discontinuity of the interface structure, a weak transition zone was found in WwFS1. The ITZ thickness of AAS mortar with the addition of WwFS is about 60 μm. For WwFS₂₀24, ITZ has the lowest elastic modulus. However, the ITZ with the addition of WwFS₇₀ was enhanced with an increase in elastic modulus, and in particular, the elastic modulus of ITZ with WwFS₇₀6 tended to be closer to that of WwFS0. This implies that hot water immersion treatment of WwFS did promote slag hydration, generating more hydration products in the interface. Meanwhile, these results further proved that WwFS₇₀6 can achieve better bonding between AASM matrix and aggregate, and also contribute to the strength of mortars.

Figure 14 presents the contour mapping of elastic modulus of pastes around AAS aggregates. The red areas represent unhydrated GGBS or aggregates with an elastic modulus more than 60 GPa (54), while the areas less than 60 GPa are considered to be hydration products. In Figs. 14d and 14e, the red (>60 GPa) areas are significantly decreased and the cyan (20~30 GPa) and green (30~40 GPa) areas are increased, which suggests that WwFS₇₀6-added mortars had a higher degree of hydration with less unhydrated GGBS.

Figure 15 illustrates the deconvolution results of elastic modulus and volume fraction of pastes around AAS aggregates. According to previous studies, the elastic modulus was divided into four stages in order from low to high, i.e., porous phase (PP), low-density C-A-S-H (LD), high-density C-A-S-H (HD) and unreacted GGBS (UG). The average elastic modulus of C-A-S-H for each mixture is presented in the inset table in Figure 15a. After replacing quartz sand with WwFS, the average elastic modulus of both LD and HD was higher than that of WwFS0. Compared to WwFS1, the C-A-S-H average elastic modulus is increased for all of WwFS₇₀, although WwFS₂₀24 is slightly lower. The average elastic modulus of C-A-S-H increased from 32.66 GPa for WwFS1 to 37.88 GPa for WwFS₇₀6, indicating that the addition of WwFS₇₀6 densified the microstructure of the hydration products. In Figure 15b, the hydration products with the largest volume fraction in WwFS₇₀1 and WwFS₇₀6 are both HD. WwFS₇₀6 shows a decrease in PP volume fraction from 26% to 19% and an increase in HD volume fraction from 26% to 37% compared to WwFS1, suggesting that WwFS₇₀6 promotes the formation of more HD in AAS. The above findings further demonstrate the synergistic effect on activating the slag between NaOH and sodium silicate dissolved into the mixing water, which contributes to improving the compressive strength of AAS mortars, consistent with the results of MIP.