Effect of Heat Curing Method on the Mechanical Strength of Alkali-Activated Slag Mortar after High-Temperature Exposure

The aim of this work was to study the mechanical strength and microstructure changes of alkali-activated slag mortar (AAS mortar) after being heat treated in the temperature range of 200–1000 °C. The AAS mortar was cured in the ambient condition (20 ± 5 °C, 60 ± 5% RH) (Relative humidity: RH) and high temperature condition (80 °C) for 27 days with three different heating regimes: curing in a dry oven, curing in sealed plastic bags, and in a steam environment. The activator for the AAS synthesis was a mixture of sodium silicate solution (water glass) and sodium hydroxide (NaOH) with a SiO2/Na2O weight ratio of 1, and a dosage of 4% Na2O by slag weight. Thermogravimetric analysis (TGA) and scanning electron microscopy (SEM) incorporated with energy-dispersive X-ray spectroscopy (EDX) were used to assess the mortar microstructure change. The results revealed that the curing method significantly affected the mechanical strength of AAS at temperatures lower than 800 °C. The heat treatment at late age of 28 days was more beneficial for compressive strength enhancement in specimens without using heat curing methods.


Introduction
Alkali-activated ground granulated blast furnace slag has recently attracted strong attention in the literature as a potential alternative to Portland cement [1][2][3]. When used as the main binder, slag is commonly activated by an alkaline agent, such as sodium hydroxide, sodium carbonates, sodium silicates, or other alkalis [4]. Several studies have concluded that the main product of this activation process is generally a low crystalline calcium silicate hydrate (C-S-H) with a low CaO/SiO 2 ratio [5,6].
The properties of alkali-activated slag (AAS) were concluded to be highly dependent on the type and content of the alkaline activator, and on the curing condition [3,4,[7][8][9][10][11][12][13][14]. Investigation of the hydration process of AAS revealed that it was influenced by the sodium dosage and silica modulus M s (a SiO 2 /Na 2 O weight ratio) of the alkaline solution [15]. In addition, AAS cured at an elevated temperature can achieve a high strength at an early age when compared with the ambient cured material [16][17][18]. Elevated-temperature curing greatly accelerates the slag activation process, resulting in the fast development of mechanical strength. In addition, a curing condition with high humidity was reported to be more beneficial than a dry condition since moisture retention is required for hydration [18]. Furthermore, an autoclave curing condition was reported to be appropriate for AAS mortar using an alkaline solution with low Na 2 O concentrations and low weight ratios SiO 2 /Na 2 O SiO 2 /Na 2 O weight ratio of 3.22. The silica sand with a nominal maximum size of 4 mm and a fineness modulus of 2.45 was used in the mortar mixture as fine aggregate.

Mixture Proportion
The mortar mixture had a fine aggregate to slag weight ratio of 2.75 and a water to solid weight ratio of 0.45. The solid weight was the total weight of slag and solid component in the activator. The Na 2 O concentration of the activator was 4% by slag weight.

Experimental Methods
Mortar mixtures were mixed in a 5 L planetary blending machine in accordance with the specifications of ASTM C305 [31]. The mortar specimens were cast in different shaped molds and kept for 24 h in the ambient condition (20 ± 5 • C and 60 ± 5% RH). The 50 mm cube specimens, 40/40/160 mm prismatic specimens, and 8-shaped specimens were used for the compressive strength test, flexural strength test and tensile strength test, respectively. The mortar specimens were then unmolded and cured in the following four different curing regimes for 27 days prior to heat treatment: (a) A specimens were stored in the ambient condition (20 ± 5 • C, 60 ± 5% RH) (A method); (b) H specimens were cured in a dry oven at 80 • C (H method); (c) B specimens were kept in two sealed plastic bags and then cured in the oven at 80 • C (B method); and (d) S specimens were cured in a steam environment at 80 • C (S method) by using a humidity curing cabinet (Heungjin, model HJ-1450, Gimpo-si, Korea).
After curing, the mortar compressive and flexural strength were tested on the hydraulic press testing machine (Shimadzu, 250 kN, Kyoto, Japan) in accordance with ASTM C109 for compressive strength [32], and ASTM C348 for flexural strength [33]. The tension tester (Heungjin, 5 kN, Gimpo-si, Korea) was used to measure the mortar tensile strength as specified in ASTM C190 [34]. For heat treatment test, the specimens were exposed to high temperatures of 200, 400, 600, 800, and 1000 • C (T c ) by using a Muffle furnace (Lindberg/Blue, 1300 • C Box Furnace, Waltham, MA, USA) with a heating rate of approximately 6.67 • C/min. The specimens were stored in the furnace at the target temperature (T c ) for 2 h and then allowed to cool naturally to the laboratory temperature. The temperature cycle in the furnace is depicted in Figure 1.

of 17
The activator with a silica modulus (Ms) of 1 was made by dissolving sodium hydroxide pellets (NaOH) in sodium silicate solution (water glass) that contained 65.4% H2O by mass and had a SiO2/Na2O weight ratio of 3.22. The silica sand with a nominal maximum size of 4 mm and a fineness modulus of 2.45 was used in the mortar mixture as fine aggregate.

Mixture Proportion
The mortar mixture had a fine aggregate to slag weight ratio of 2.75 and a water to solid weight ratio of 0.45. The solid weight was the total weight of slag and solid component in the activator. The Na2O concentration of the activator was 4% by slag weight.

Experimental Methods
Mortar mixtures were mixed in a 5 L planetary blending machine in accordance with the specifications of ASTM C305 [31]. The mortar specimens were cast in different shaped molds and kept for 24 h in the ambient condition (20 ± 5 °C and 60 ± 5% RH). The 50 mm cube specimens, 40/40/160 mm prismatic specimens, and 8-shaped specimens were used for the compressive strength test, flexural strength test and tensile strength test, respectively. The mortar specimens were then unmolded and cured in the following four different curing regimes for 27 days prior to heat treatment: (a) A specimens were stored in the ambient condition (20 ± 5 °C, 60 ± 5% RH) (A method); (b) H specimens were cured in a dry oven at 80 °C (H method); (c) B specimens were kept in two sealed plastic bags and then cured in the oven at 80 °C (B method); and (d) S specimens were cured in a steam environment at 80 °C (S method) by using a humidity curing cabinet (Heungjin, model HJ-1450, Gimpo-si, Korea).
After curing, the mortar compressive and flexural strength were tested on the hydraulic press testing machine (Shimadzu, 250 kN, Kyoto, Japan) in accordance with ASTM C109 for compressive strength [32], and ASTM C348 for flexural strength [33]. The tension tester (Heungjin, 5 kN, Gimposi, Korea) was used to measure the mortar tensile strength as specified in ASTM C190 [34]. For heat treatment test, the specimens were exposed to high temperatures of 200, 400, 600, 800, and 1000 °C (Tc) by using a Muffle furnace (Lindberg/Blue, 1300 °C Box Furnace, Waltham, MA, America) with a heating rate of approximately 6.67 °C/min. The specimens were stored in the furnace at the target temperature (Tc) for 2 h and then allowed to cool naturally to the laboratory temperature. The temperature cycle in the furnace is depicted in Figure 1. After cooling to ambient temperature, the mortar specimens were tested to assess the remaining mechanical strength. With reference to some previous publications [23,30,35], selected fragments After cooling to ambient temperature, the mortar specimens were tested to assess the remaining mechanical strength. With reference to some previous publications [23,30,35], selected fragments from the crushed sample after the compressive strength test were soaked in acetone for 72 h for hydration stopping, and then clarified from acetone and finally dried in a vacuum desiccator for 24 h. Some fragments were chosen to use for the SEM (S-4100, Hitachi, Tokyo, Japan) equipped with EDX analysis. Part of the dried samples was crushed into powder for TGA analysis (SDT Q600, New Castle, DE, USA).

Reference Mechanical Strengths
The mechanical strength parameters of the AAS mortar specimens without elevated-temperature treatment at 28 days are given in Table 2. As seen in Table 2, the mechanical strength parameters were greatly influenced by curing conditions. Compared with the A method, the S method greatly increased all mortar mechanical strength parameters, while the flexural strength was reduced under the H and B methods. In addition, the material brittleness was higher when the flexural strength to compressive strength ratio (FS/CS) reduced and the angle of internal friction (Φ) increased [36][37][38]. Consequently, Table 2 shows that the H specimens were the most brittle. Heat curing treatment at early age greatly enhanced the compressive strength of AAS mortar. For instance, the H, B, and S specimens exhibited strength gains of 44.3%, 44.7%, and 65.9%, respectively, in comparison with the A specimens. However, despite being cured at the same temperature (80 • C), the S specimens achieved approximately 15% greater strength than the H or B specimens. This was due to the quicker escape of water from the sample's structure when cured in the dry oven and plastic bags. In addition, retained moisture in the steam condition was required for hydration [18]. Table 2 shows that the H and B methods degraded the flexural strength of AAS mortar. For instance, the strength was reduced by 33.8% and 17.6% for the H and B specimens, respectively. Ya-min et al. [39] reported that flexural strength was more easily influenced by the presence of micro-cracks than the compressive strength. The accelerated reaction due to heat treatment at early age caused the large chemical shrinkage and formed micro-cracks in AAS [39]. The specimen structure deteriorated by drying shrinkage due to the water evaporation at 80 • C for 27 days. This can explain why the B and H samples obtained lower flexural strength than the A samples. On the contrary, the S method resulted in a greater strength gain of approximately 64.7% compared with the A specimen. The S method was more effective in increasing flexural strength in comparison with the H or B specimens, possibly due to the limited micro-crack formation in the mortar structure of the S specimens.
The 8-shaped specimens were tested to measure the AAS mortar tensile strength at 28 days. The unexposed B and S specimens possessed similar tensile strengths of approximately twice that of the A and H specimens. Atiş et al. [36] reported that AAS mortar exhibited lower tensile strength when its brittleness increased. However, the H specimens obtained the same tensile strength as the A specimens, despite their higher brittleness.

Residual Compressive Strength
The thermal performance of AAS mortars cured in different methods was evaluated by determining the mechanical strength change after elevated-temperature treatment. The mechanical strength parameters of exposed specimens are shown in Table 3. In addition, the residual compressive strength of specimens after heat treatment at temperatures up to 1000 • C is given in Figure 2a, while Figure 2b illustrates the percent change in compressive strength throughout the range of 200-1000 • C.  [36]. (CS, FS, and TS are the compressive, flexural, and tensile strength, respectively).

Residual Compressive Strength
The thermal performance of AAS mortars cured in different methods was evaluated by determining the mechanical strength change after elevated-temperature treatment. The mechanical strength parameters of exposed specimens are shown in Table 3. In addition, the residual compressive strength of specimens after heat treatment at temperatures up to 1000 °C is given in Figure 2a, while Figure 2b illustrates the percent change in compressive strength throughout the range of 200-1000 °C.   the temperature to 400 • C reduced this strength increase, but the retained strength remained higher than the reference strengths of 30.7% and 18.5% for specimens A and B, respectively.
The compressive strength was also enhanced in the S specimens after exposure to 200 and 400 • C. The S specimens exhibited a strength increase of 18.2% and 22.4% at 200 and 400 • C, respectively. Consequently, the S specimens attained the highest strength at 400 • C instead of 200 • C for the A and B specimens.
In contrast to the specimens described above, the H specimens exhibited strength loss at all exposure temperatures. The strength reduction of this mortar was determined to be 7.8%, 23.2%, 42.3%, and 82.3% at exposure to 200, 400, 600, and 800 • C, respectively. The effect of heat treatment on the strength development of AAS has been discussed in many publications [16,18]. In high-temperature curing, the strength of AAS is significantly enhanced at early age. However, the duration of specimens in the laboratory condition prior to heat curing was pointed out to greatly affect the strength gain resulted from heat treatment [16,30]. The compressive strength enhancement of the A specimens at 200 and 400 • C was lower than the strength gain that they could achieve when curing after demolding. On the other hand, the B and S specimens gained higher strength from heat curing after demolding than that attained at 200 and 400 • C. However, the strength gain resulted from the heat treatment at 200 and 400 • C was the highest in the A specimens. In contrast, as shown in Figure 2a, heat treatment at late age did not cause any strength gain in the H specimens. The aforementioned strength enhancement (at 200 and 400 • C) was likely caused by the further hydration acceleration due to heat effect at late age. The results above reveal that the hydration enhancement at late age could be related to the availability of retained water in the mortar structure. The further hydration due to heat treatment at late age was eased with the presence of water in the mortar structure. The considerable loss of water during 27 days at 80 • C in a dry oven was one of the main causes for the lack of strength gain of the H specimens when exposed to 200 and 400 • C.
Raising the temperature beyond 400 • C led to a huge loss of compressive strength for all AAS mortars with the residual strengths converged at 800 • C ( Figure 2). The slope of the strength reduction line between 400 and 800 • C was the highest in the S specimens and the lowest in the H specimens. Despite being subjected to elevated temperatures, the S specimens achieved the highest remaining strength and the H specimens the lowest from 200 to 600 • C. Figure 2 shows that all four specimens exhibited the same remarkable trend of compressive strength change in the temperature range of 800-1000 • C. Increasing the exposure temperature to 1000 • C greatly decreased the strength loss at 800 • C for all specimens. All mortar samples exhibited a slight strength change in this temperature range. In Figure 2a, the compressive strength of the A and H specimens was lower than that of the B and S specimens at all exposure temperatures.
The visual observations of hardened mortars before and after exposure to elevated temperatures is exhibited in Figure 3. In the temperature range of 200-800 • C, the white color of the unexposed specimens changed to a darker color under a higher temperature, and then became a lighter color at 1000 • C. Figure 3 shows no sign of spalling in any of the specimens during high-temperature exposure. However, small cracks appeared on the surface of all specimens at 800 and 1000 • C.

of 17
The A and B specimens had the same trend of strength change after high-temperature treatment. At 200 °C, the A and B specimens achieved strength increases of 36.9% and 23.6%, respectively. Raising the temperature to 400 °C reduced this strength increase, but the retained strength remained higher than the reference strengths of 30.7% and 18.5% for specimens A and B, respectively.
The compressive strength was also enhanced in the S specimens after exposure to 200 and 400 °C. The S specimens exhibited a strength increase of 18.2% and 22.4% at 200 and 400 °C, respectively. Consequently, the S specimens attained the highest strength at 400 °C instead of 200 °C for the A and B specimens.
In contrast to the specimens described above, the H specimens exhibited strength loss at all exposure temperatures. The strength reduction of this mortar was determined to be 7.8%, 23.2%, 42.3%, and 82.3% at exposure to 200, 400, 600, and 800 °C, respectively. The effect of heat treatment on the strength development of AAS has been discussed in many publications [16,18]. In hightemperature curing, the strength of AAS is significantly enhanced at early age. However, the duration of specimens in the laboratory condition prior to heat curing was pointed out to greatly affect the strength gain resulted from heat treatment [16,30]. The compressive strength enhancement of the A specimens at 200 and 400 °C was lower than the strength gain that they could achieve when curing after demolding. On the other hand, the B and S specimens gained higher strength from heat curing after demolding than that attained at 200 and 400 °C. However, the strength gain resulted from the heat treatment at 200 and 400 °C was the highest in the A specimens. In contrast, as shown in Figure  2a, heat treatment at late age did not cause any strength gain in the H specimens. The aforementioned strength enhancement (at 200 and 400 °C) was likely caused by the further hydration acceleration due to heat effect at late age. The results above reveal that the hydration enhancement at late age could be related to the availability of retained water in the mortar structure. The further hydration due to heat treatment at late age was eased with the presence of water in the mortar structure. The considerable loss of water during 27 days at 80 °C in a dry oven was one of the main causes for the lack of strength gain of the H specimens when exposed to 200 and 400 °C.
Raising the temperature beyond 400 °C led to a huge loss of compressive strength for all AAS mortars with the residual strengths converged at 800 °C ( Figure 2). The slope of the strength reduction line between 400 and 800 °C was the highest in the S specimens and the lowest in the H specimens. Despite being subjected to elevated temperatures, the S specimens achieved the highest remaining strength and the H specimens the lowest from 200 to 600 °C. Figure 2 shows that all four specimens exhibited the same remarkable trend of compressive strength change in the temperature range of 800-1000 °C. Increasing the exposure temperature to 1000 °C greatly decreased the strength loss at 800 °C for all specimens. All mortar samples exhibited a slight strength change in this temperature range. In Figure 2a, the compressive strength of the A and H specimens was lower than that of the B and S specimens at all exposure temperatures.
The visual observations of hardened mortars before and after exposure to elevated temperatures is exhibited in Figure 3. In the temperature range of 200-800 °C, the white color of the unexposed specimens changed to a darker color under a higher temperature, and then became a lighter color at 1000 °C. Figure 3 shows no sign of spalling in any of the specimens during high-temperature exposure. However, small cracks appeared on the surface of all specimens at 800 and 1000 °C.  Figure 4 shows that the strength change trend of sample H was similar to that of sample B and both exhibited a lower strength loss throughout exposure temperatures to 1000 • C than mortar samples A and S. For instance, exposure to 200 • C reduced the strength by approximately 57.4% and 48.2% for specimens A and S, respectively, compared with 26.7% and 23.2% for specimens H and B, respectively. At 200 • C, the A and S specimens exhibited a drastic strength reduction, and the H and B specimens showed a lower strength drop. The further micro-crack formation from shrinkage may have caused the flexural strength loss of the AAS mortars at elevated temperatures. When subjected to a high temperature of 200 • C, the mortar experienced drying shrinkage due to water evaporation, and chemical shrinkage resulted from further reactions [40]. The formation of cracks due to this mechanism did not occur in the A specimens. However, when exposed to 200 • C, the further rapid reaction and drying shrinkage resulted in significant micro-crack formation and drastically reduced the flexural strength.
In the high-temperature condition, AAS experienced large chemical shrinkage due to rapid reaction occurrence, as mentioned by Ya-min et al. [39]. Furthermore, for S specimens, the crack formation could be restrained due to curing in the steam condition. Consequently, the drastic flexural strength reduction in these specimens could be attributed to rapid crack formation in the structure resulting from shrinkage at 200 • C. Increasing the exposure temperature to both 400 and 600 • C alleviated the strength loss for all samples. However, the residual flexural strength of all samples decreased greatly again in the temperature range of 600-800 • C. The residual flexural strength was slightly different at 800 and 1000 • C for all mortar specimens, which is similar to the compressive strength results. Throughout the temperature range from 200 to 1000 • C, the H and B specimens revealed a similar trend in residual strength variation and had a lower strength reduction than the A and S specimens. On the other hand, the A and H specimens exhibited lower flexural strength than the B and S specimens throughout the range of 200-600 • C.   Figure 4 illustrates the retained flexural strength and percent change in the strength of AAS mortars in the high-temperature range of 200-1000 °C. Exposure to the high temperature of 1000 °C reduced the flexural strength for all mortars and the strengths converged at 800 °C. Figure 4 shows that the strength change trend of sample H was similar to that of sample B and both exhibited a lower strength loss throughout exposure temperatures to 1000 °C than mortar samples A and S. For instance, exposure to 200 °C reduced the strength by approximately 57.4% and 48.2% for specimens A and S, respectively, compared with 26.7% and 23.2% for specimens H and B, respectively. At 200 °C, the A and S specimens exhibited a drastic strength reduction, and the H and B specimens showed a lower strength drop. The further micro-crack formation from shrinkage may have caused the flexural strength loss of the AAS mortars at elevated temperatures. When subjected to a high temperature of 200 °C, the mortar experienced drying shrinkage due to water evaporation, and chemical shrinkage resulted from further reactions [40]. The formation of cracks due to this mechanism did not occur in the A specimens. However, when exposed to 200 °C, the further rapid reaction and drying shrinkage resulted in significant micro-crack formation and drastically reduced the flexural strength. In the high-temperature condition, AAS experienced large chemical shrinkage due to rapid reaction occurrence, as mentioned by Ya-min et al. [39]. Furthermore, for S specimens, the crack formation could be restrained due to curing in the steam condition. Consequently, the drastic flexural strength reduction in these specimens could be attributed to rapid crack formation in the structure resulting from shrinkage at 200 °C. Increasing the exposure temperature to both 400 and 600 °C alleviated the strength loss for all samples. However, the residual flexural strength of all samples decreased greatly again in the temperature range of 600-800 °C. The residual flexural strength was slightly different at 800 and 1000 °C for all mortar specimens, which is similar to the compressive strength results. Throughout the temperature range from 200 to 1000 °C, the H and B specimens revealed a similar trend in residual strength variation and had a lower strength reduction than the A and S specimens. On the other hand, the A and H specimens exhibited lower flexural strength than the B and S specimens throughout the range of 200-600 °C.      Figure 5 shows that the slope of the residual strength reduction line of samples B and S was greater than that of samples A and H when the temperature was increased to 200 • C, i.e., their strength loss was greater. Previous research pointed out that tensile strength was strongly influenced by the paste matrix-aggregate interfacial transition zone strength [19,41]. Throughout elevated-temperature exposure, the thermal incompatibility between the paste and aggregate degraded the binding ability between the paste and aggregate, while the evolution of pores in the mortar microstructure at high temperatures also decreased the tensile strength [42]. Similar to the compressive strength case, the H specimens exhibited remaining tensile strength lower than that of the other three specimens in 200-600 • C range ( Figure 5). More brittle mortar could experience higher shrinkage [36], resulting in higher thermal shrinkage at elevated temperatures [28], and thus, leading to a degradation in mechanical strengths. This result supports the hypothesis of Atiş et al. [36] about the correlation between the shrinkage, brittleness, and tensile strength of AAS mortar. Figure 5 shows that the variation trend of tensile strength between 800 and 1000 • C was similar to that of compressive and flexural strength. Consequently, the influence of curing conditions on mechanical strength change was negligible in the temperature range of 800-1000 • C. On the other hand, the B specimens attained the highest tensile, compressive, and flexural strength at 1000 • C. Similar to the compressive and flexural strength results, the A and H specimens exhibited a low tensile strength in the range of 200-600 • C. 8 of 17 and 400 °C, and exhibited less degradation at higher temperatures. Figure 5 shows that the slope of the residual strength reduction line of samples B and S was greater than that of samples A and H when the temperature was increased to 200 °C, i.e., their strength loss was greater. Previous research pointed out that tensile strength was strongly influenced by the paste matrix-aggregate interfacial transition zone strength [19,41]. Throughout elevated-temperature exposure, the thermal incompatibility between the paste and aggregate degraded the binding ability between the paste and aggregate, while the evolution of pores in the mortar microstructure at high temperatures also decreased the tensile strength [42]. Similar to the compressive strength case, the H specimens exhibited remaining tensile strength lower than that of the other three specimens in 200-600 °C range ( Figure 5). More brittle mortar could experience higher shrinkage [36], resulting in higher thermal shrinkage at elevated temperatures [28], and thus, leading to a degradation in mechanical strengths. This result supports the hypothesis of Atiş et al. [36] about the correlation between the shrinkage, brittleness, and tensile strength of AAS mortar. Figure 5 shows that the variation trend of tensile strength between 800 and 1000 °C was similar to that of compressive and flexural strength.  Figure 6 presents the TGA curve of the AAS mortar after curing in four regimes. The H sample had the lowest mass reduction before approximately 560 °C. The mass loss before approximately 150 °C is mainly caused by the evaporation of water [30]. The lower mass reduction of H specimens revealed a lower retained water content in its mortar structure, which is unbeneficial for further hydration due to heat treatment at 28 days. In addition, the slope of the mass reduction line in the TGA curve from approximately 450 to 560 °C was steeper in samples A and H than that in samples B and S, indicating their rapid mass reduction in this temperature range. 3.5. Thermogravimetric Analysis (TGA) Figure 6 presents the TGA curve of the AAS mortar after curing in four regimes. The H sample had the lowest mass reduction before approximately 560 • C. The mass loss before approximately 150 • C is mainly caused by the evaporation of water [30]. The lower mass reduction of H specimens revealed a lower retained water content in its mortar structure, which is unbeneficial for further hydration due to heat treatment at 28 days. In addition, the slope of the mass reduction line in the TGA curve from approximately 450 to 560 • C was steeper in samples A and H than that in samples B and S, indicating their rapid mass reduction in this temperature range.  Figure 7 presents SEM images of the AAS mortar fracture surface with the four different curing methods after casting for 28 days. Figure 7 shows that the A specimens possessed a porous structure with visible unhydrated slag grains. The atomic composition of the reaction product was studied by using EDX. Figure 8a presents the EDX image of the spot marked 1 in Figure 7a where silicon (Si) and aluminum (Al) element intensity was dominant with negligible traces of calcium (Ca). Figure 7 shows that the specimen microstructure became denser following heat curing, which was consistent with the compressive strength enhancement. However, micro-cracks appeared on the surface of the B specimens due to the quick water evaporation, resulting in their low flexural strengths. In addition, both the porous structure and the grain detachment were observed in the H specimens, which may have caused its low flexural and tensile strength. On the contrary, the S specimens exhibited a massive and well-packed microstructure due to the limited water evaporation. The EDX image (Figure 8b-d) shows that the Ca element intensity at the mortar surface (at the point marked 1 in Figure 7b-d) increased remarkably to a dominant level following heat curing. These EDX spectra are also indicative of the existence of C-A-S-H, which was the major hydration product of AAS [4,12]. C-A-S-H in AAS was pointed out to be more amorphous than C-S-H in hydrated Portland cement [43,44]. This clearly demonstrated that the AAS process could be strongly accelerated by heattreatment curing despite the low Na2O concentration (4% by slag weight) of the activator.  Figure 7 presents SEM images of the AAS mortar fracture surface with the four different curing methods after casting for 28 days. Figure 7 shows that the A specimens possessed a porous structure with visible unhydrated slag grains. The atomic composition of the reaction product was studied by using EDX. Figure 8a presents the EDX image of the spot marked 1 in Figure 7a where silicon (Si) and aluminum (Al) element intensity was dominant with negligible traces of calcium (Ca). Figure 7 shows that the specimen microstructure became denser following heat curing, which was consistent with the compressive strength enhancement. However, micro-cracks appeared on the surface of the B specimens due to the quick water evaporation, resulting in their low flexural strengths. In addition, both the porous structure and the grain detachment were observed in the H specimens, which may have caused its low flexural and tensile strength. On the contrary, the S specimens exhibited a massive and well-packed microstructure due to the limited water evaporation. The EDX image (Figure 8b-d) shows that the Ca element intensity at the mortar surface (at the point marked 1 in Figure 7b-d) increased remarkably to a dominant level following heat curing. These EDX spectra are also indicative of the existence of C-A-S-H, which was the major hydration product of AAS [4,12]. C-A-S-H in AAS was pointed out to be more amorphous than C-S-H in hydrated Portland cement [43,44]. This clearly demonstrated that the AAS process could be strongly accelerated by heat-treatment curing despite the low Na 2 O concentration (4% by slag weight) of the activator.  Figure 7 presents SEM images of the AAS mortar fracture surface with the four different curing methods after casting for 28 days. Figure 7 shows that the A specimens possessed a porous structure with visible unhydrated slag grains. The atomic composition of the reaction product was studied by using EDX. Figure 8a presents the EDX image of the spot marked 1 in Figure 7a where silicon (Si) and aluminum (Al) element intensity was dominant with negligible traces of calcium (Ca). Figure 7 shows that the specimen microstructure became denser following heat curing, which was consistent with the compressive strength enhancement. However, micro-cracks appeared on the surface of the B specimens due to the quick water evaporation, resulting in their low flexural strengths. In addition, both the porous structure and the grain detachment were observed in the H specimens, which may have caused its low flexural and tensile strength. On the contrary, the S specimens exhibited a massive and well-packed microstructure due to the limited water evaporation. The EDX image (Figure 8b-d) shows that the Ca element intensity at the mortar surface (at the point marked 1 in Figure 7b-d) increased remarkably to a dominant level following heat curing. These EDX spectra are also indicative of the existence of C-A-S-H, which was the major hydration product of AAS [4,12]. C-A-S-H in AAS was pointed out to be more amorphous than C-S-H in hydrated Portland cement [43,44]. This clearly demonstrated that the AAS process could be strongly accelerated by heattreatment curing despite the low Na2O concentration (4% by slag weight) of the activator.     Figure 9 shows that the porous structure of the A specimens was significantly converted to a denser structure at 200 °C, which may have resulted from further reactions due to the heat treatment effect at 200 °C. In addition, the intensity of calcium (Ca) counts on the A specimen surface (at the spot marked 2 in Figure 9a) increased significantly after exposure at 200 °C (EDX result in Figure 10), revealing the further formation of C-S-H. This is in agreement with the great compressive strength increase and low tensile strength reduction of the A specimens at 200 °C. For the H specimens, large grains and large micro-cracks were detached in the network structure, resulting in the compressive strength drop at 200 °C ( Figure 11). The H specimens were clearly the most brittle. Therefore, this mortar exhibited a more damaged structure due to the higher shrinkage at elevated temperatures. Contrary to the A specimens, the surfaces of the S and B specimens presented several fairly large cracks and a slightly less dense area at 200 °C (Figures 12, 13). This structural transformation contrasts with the compressive strength enhancement of both specimens when exposed at 200 °C.    Figure 9 shows that the porous structure of the A specimens was significantly converted to a denser structure at 200 °C, which may have resulted from further reactions due to the heat treatment effect at 200 °C. In addition, the intensity of calcium (Ca) counts on the A specimen surface (at the spot marked 2 in Figure 9a) increased significantly after exposure at 200 °C (EDX result in Figure 10), revealing the further formation of C-S-H. This is in agreement with the great compressive strength increase and low tensile strength reduction of the A specimens at 200 °C. For the H specimens, large grains and large micro-cracks were detached in the network structure, resulting in the compressive strength drop at 200 °C ( Figure 11). The H specimens were clearly the most brittle. Therefore, this mortar exhibited a more damaged structure due to the higher shrinkage at elevated temperatures. Contrary to the A specimens, the surfaces of the S and B specimens presented several fairly large cracks and a slightly less dense area at 200 °C (Figures 12, 13). This structural transformation contrasts with the compressive strength enhancement of both specimens when exposed at 200 °C.  Figure 9 presents the significant transformation of the A specimen microstructure from 200 to 1000 • C. Figure 9 shows that the porous structure of the A specimens was significantly converted to a denser structure at 200 • C, which may have resulted from further reactions due to the heat treatment effect at 200 • C. In addition, the intensity of calcium (Ca) counts on the A specimen surface (at the spot marked 2 in Figure 9a) increased significantly after exposure at 200 • C (EDX result in Figure 10), revealing the further formation of C-S-H. This is in agreement with the great compressive strength increase and low tensile strength reduction of the A specimens at 200 • C. For the H specimens, large grains and large micro-cracks were detached in the network structure, resulting in the compressive strength drop at 200 • C ( Figure 11). The H specimens were clearly the most brittle. Therefore, this mortar exhibited a more damaged structure due to the higher shrinkage at elevated temperatures. Contrary to the A specimens, the surfaces of the S and B specimens presented several fairly large cracks and a slightly less dense area at 200 • C (Figures 12 and 13). This structural transformation contrasts with the compressive strength enhancement of both specimens when exposed at 200 • C.        Figures 9 and 11-13 show that the mortar structure was significantly degraded after heat treatment in the temperature range of 400-800 • C, causing high strength loss. Crack and pore formation was more intensive in the A and H specimens, resulting from the rapid mass reduction in the temperature range of 450-560 • C (TGA curve). In addition, the network structure of the H sample exhibited significant grain detachment. These results are consistent with the low residual mechanical strengths of the H and A specimens.  As shown in Figures 9 and 11-13, all mortar samples exhibited a remarkable transformation in network structure in the temperature range of 800-1000 • C. Raising the temperature to 1000 • C healed the cracks and discrete areas in the mortar structure due to sintering. This change in the mortar structure of all mortars considerably reduced the mechanical strength deterioration at 800 • C regardless of the curing method. Figure 12 shows that the B specimens possessed the most condensed microstructure after exposure at 1000 • C. This could explain for the highest mechanical strength of the B specimens when cured at 1000 • C.  Figures 9 and 11-13 show that the mortar structure was significantly degraded after heat treatment in the temperature range of 400-800 °C, causing high strength loss. Crack and pore formation was more intensive in the A and H specimens, resulting from the rapid mass reduction in the temperature range of 450-560 °C (TGA curve). In addition, the network structure of the H sample exhibited significant grain detachment. These results are consistent with the low residual mechanical strengths of the H and A specimens.

Microstructural Analysis
As shown in Figures 9 and 11-13, all mortar samples exhibited a remarkable transformation in network structure in the temperature range of 800-1000 °C. Raising the temperature to 1000 °C healed the cracks and discrete areas in the mortar structure due to sintering. This change in the mortar structure of all mortars considerably reduced the mechanical strength deterioration at 800 °C regardless of the curing method. Figure 12 shows that the B specimens possessed the most condensed

Conclusions
From the experimental results and discussion above, the following conclusions can be drawn: The heat curing method significantly affected the mechanical strength of AAS mortar at 28 days. The S method induced the greatest strength gain of the four mortar specimens, whereas the H specimens exhibited flexural strength degradation and increased mortar brittleness.
The A, B, and S specimens exhibited an increased compressive strength at 200 and 400 • C. These 200 and 400 • C heat treatments at 28 days were the most beneficial for compressive strength enhancement in the A specimens. The A specimens achieved a strength gain of 36.9% at 200 • C, compared with 23.7% and 18.2% for the B and S specimens, respectively. By contrast, there was no strength enhancement in the H specimens, which were the most brittle.
AAS mortar specimens with all curing methods exhibited degraded flexural and tensile strength at all exposure temperatures (200-1000 • C). In the range of 200-600 • C, the A and H specimens possessed low residual mechanical strengths due to their damaged structure.
Due to the significant transformation in structure, all AAS mortars exhibited slightly changed mechanical strengths in the temperature range from 800 to 1000 • C regardless of curing methods. The curing method had a negligible effect on the strength change of the mortar in the range of 800-1000 • C. In addition, the B specimens attained the highest residual mechanical strength at 1000 • C.
The SEM and EDX analyses revealed that the porous microstructure of the A specimens was converted to a compact structure with a greatly increased calcium (Ca) amount on the fragment surface after exposure at 200 • C.
Finally, the B and S methods were beneficial for AAS mortar with respect to residual mechanical strength in the range of 200-600 • C. Moreover, the A method was the most effective in enhancing the compressive strength of the mortar at 200 and 400 • C.