3.1. Electrical Resistivity
Figure 3 shows the 24 h measured electrical resistivity of AAS pastes that were mixed with deionized water (D40, D45, and D50;
Figure 3a) and seawater (S40, S45, and S50;
Figure 3b).
The results represent the difference in electrical resistivity and are compared with different
a/b ratios: 0.4, 0.45, and 0.5. The trend of electrical resistivity as a function of time follows the results of cement-based materials in a previous study [
10]. It was reported that the measured electrical resistivity provides three parameters to characterize the microstructural evolution during hydration. These are identified as the initial resistivity, rising time to indicate the onset of an increase in electrical resistivity, and increasing ratio of electrical resistivity after the rising time. Current flow and measured electrical resistivity are determined using the water network through GGBFS particles. Therefore, the value of the initial electrical resistivity is established using the spacing of GGBFS particles with percolated water, which can reflect the initial microstructure of suspension as the
a/b ratio. Here, the results of the initial resistivity of GGBFS-based alkali-activated pastes with seawater or deionized water is determined at an average time of 30 min, and its value is approximately 0.19 Ωm and 0.20 Ωm, respectively. Hence, the state of the initial microstructures of all the AAS pastes is similar and they exhibit sufficient electrical conductivity. For hours after mixing, microstructural change is not sufficiently large to affect the electrical resistivity. This is ascertained by the constant initial value of electrical resistivity observed for a few hours of the inactive period. However, the electrical resistivity reaches a critical point after 3 h and gradually increases. Here, a critical point is the rising time as an indicator of the setting time. This is because the coagulated GGBFS particles lead to a closing of water network and solid percolation. Further, the increasing ratio of electrical resistivity after the rising time describes the hardening phase of pastes owing to a solid network evolution caused by alkali activation.
In this study, the rising time is determined by the measured electrical resistivity that is 5 times higher than the initial value, and the increasing ratio of resistivity (Ωm/h) is the average slope of the curve after the rising time of 24 h. The calculated parameters are reported in
Table 4.
Comparing all the AAS pastes, it was observed that the rising time of the pastes with seawater or deionized water was delayed with higher a/b ratios (0.5) than with the lower a/b ratios (0.4 and 0.45). While the change in rising time was not high, the increased activator content tended to induce a delay in setting time. The pattern of delayed setting time was similar to the increasing resistivity after the rising time. Different types of mixing water and their mixing ratios influence the degree of alkali activation. This phenomenon demonstrates that a lower activator content leads to an increase of reaction products with capillary pores depercolation and cutting of conductive wires in a sample. Here, the state of sample was no longer the suspension. This supports the notion that the higher activator content improves the dispersion of GGBFS particles but does not influence the initial resistivity, and different activator content can induce a change in rising time and its increasing ratio. The changed ratio of both parameters between a/b 0.45 and 0.5 was larger than that between a/b 0.4 and 0.45. In particular, the trends of changed electrical resistivity between S40 and S45 was almost similar, it was hard to find the effect of a/b ratio on setting and hardening of GGBFS. This is because the optimized a/b ratio for hydration was between 0.4 and 0.45 for early age activation of GGBFS within 24 h, regardless of the mixing water type, and more water content remained as free water in the mixture. This influences the electrical resistivity and its increasing ratio, and leads to the difference in setting time and hardening process.
The results of three groups (
a/b ratio of 0.4, 0.45, and 0.5) in
Figure 4 demonstrate the effects of the types of mixing water.
Figure 3 shows that the lower
a/b leads to a higher electrical resistivity and faster hardening. It was observed that the rising time was similar in both pastes with the same
a/b ratio, and this trend was represented especially in the increasing ratio of electrical resistivity until approximately 12 h after rising time. Hence, the type of mixing water does not significantly affect the setting time by the alkali activation of GGBFS, but the increasing ratio of electrical resistivity and values of electrical resistivity differed as a function of time until 24 h with different mixing water under the same
a/b ratio. This supports the notion that an activation degree after setting time is dependent on the mixing water type, and seawater can better promote microstructural evolution of GGBFS than deionized water. Additionally, the difference in the degree of hardening was remarkable with higher
a/b ratio. Here, the difference of electrical resistivity at 24 h under 0.4, 0.45, and 0.5 of
a/b was 7.8%, 17.8, and 25.1%, respectively. The mixing water type rather than the dissolution and solid volume fraction of GGBFS particles can control the hardening of the AAS pastes.
3.3. XRD Analysis
Figure 6 shows the XRD patterns of deionized water-mixed AAS pastes at 24 h after the electrical resistivity test. The phase changes observed in
Figure 6 are listed in
Table 5. For seawater-mixed AAS paste samples, XRD results and phase changes are presented in
Figure 7 and
Table 6, respectively. Studies [
33,
34,
35,
36] have reported that the main reaction products in alkali-activated slag are C–S–H(I), C–A–S–H(I), hydrogarnet, C
4AH
13, and hydrotalcite. In AAS paste samples with deionized water (D40, D45, and D50), various reaction products are identified, such as C–S–H(I), C–S–H, C–A–S–H, Ca(OH)
2, K
2SO
4, hydrocalumite (3CaO∙Al
2O
3∙CaCO
3∙11H
2O), and hydrotalcite, which are also detected in seawater-mixed samples (S40, S45, and S50). In addition, akermanite, gypsum, anhydrite, and calcite contained were identified the raw GGBFS.
When the slag mainly consists of amorphous phase, it does not produce calcium hydroxide (Ca(OH)
2) as a reaction product [
37]. It is reported that Ca(OH)
2 can be formed in CaO or Ca(OH)
2-activated slag. XRD patterns of the original GGBFS (
Figure 1) showed relatively strong gypsum and anhydrite peaks in comparison with existing literature [
23,
37,
38]. Ca ions from gypsum and anhydrite can be consumed, producing a Ca-containing phase [
39]. Considering that the reflection intensities of gypsum in AAS pastes were significantly decreased and the anhydrite was decreased or absent, it is believed that the formation of Ca(OH)
2 in this study may be produced from the gypsum and anhydrite. It was expected that K
2SO
4 would be produced by the activator (KOH) and SO
42− contained in seawater. However, considering that regardless of whether or not seawater was used, K
2SO
4 was detected in all samples. The sulfate from the gypsum and anhydrite contained in the raw GGBFS may have reacted with K
+ in KOH to form K
2SO
4. This demonstrates that the reaction products of AAS can depend on certain crystalline phases caused by the GGBFS and type of activator. For each AAS sample with seawater or deionized water, the reflection intensities of K
2SO
4 increases as the
a/b ratio increases, which is due to the increase in the amount of KOH solution per amount of binder.
D45 sample exhibited less C–S–H(I) and C–A–S–H than D40 (
Table 5). It suggests that they would have affected the value of electrical resistivity between the two samples. D50 exhibited the newly formed phases (C
4AH
13 and unidentified peak). However, undissolved anhydrite remained in D50, but was absent in D40 and D45 (
Figure 6). This suggests that the degree of D50 hydration was lower than that in the D45 and D50 samples. This may allow water percolation, and thus, cause the low electrical resistivity.
When the seawater was used in AAS, Cl-bearing hydrocalumite (3CaO∙Al
2O
3∙CaCl
2∙10H
2O), AlOCl (aluminium oxide chloride), aluminum chloride hydrate, and gismondine (CaAl
2Si
2O
8∙4H
2O) were formed, unlike in the deionized water-mixed AAS. Here, it can be observed that Cl-hydrocalumite, AlOCl, and aluminum chloride hydrate are reaction products pertaining to the chloride ions in seawater. Hydrocalumite belongs to a group of layered double hydroxides (LDHs), which exhibits an anion-exchange capacity [
38,
40]. The hydrocalumite formed in the seawater-mixed AAS paste is a Cl-exchanged phase with the strongest peak of 11.362° (2θ). It is reported that Cl in seawater can be present in the form of OCl
– under the alkaline environment [
41,
42]. It is expected that the presence of AlOCl in the seawater-mixed AAS may be due to the reaction between Al from GGBFS and OCl
– in the alkaline solution with seawater. Zeolites are generally observed in alkali-activation of fly ash [
43]. Gismondine, which is a zeolitic aluminosilicate, was observed in alkali-activation of GGBFS and GGBFS/metakaoline blends based on sodium silicate solution and sodium hydroxide [
44]. The observation of gismondine in S50 may indicate that the seawater-mixed AAS may contain a zeolitic phase.
S45 sample showed relatively less reaction products (the reduction of C–S–H(I), C–S–H, Ca(OH)
2, hydrocalumite, hydrotalcite etc.) than S40 (
Table 6). However, the value of electrical resistivity of S45 was higher than that of S40, while the compressive strength was lower. This may imply that a specific reaction product is responsible for the capillary pores depercolation. However, it does not lead to the strength development. Here, the reaction products would be C
4AH
13 and the unidentified crystalline phase (
Figure 7), as they were present in S45 but absent in S40. Although gismondine was formed only in S50, its XRD patterns showed relatively low peak intensities for the commonly identified reaction products in seawater-mixed samples. This would result in a low electrical resistivity in S50.
For each a/b ratio (0.40, 0.45, and 0.50), seawater-mixed AAS paste samples hardened faster than the deionized water-mixed AAS samples. This could be attributed to the crystalline phases formed using seawater. S40 exhibited Cl-bearing phases, such as Cl-bearing hydrocalumite, AlOCl, and aluminum chloride hydrate, unlike D40. S45 produced C4AH13, unidentified phase, and Cl-bearing phases when compared with D45. For S50, the Cl-bearing phases and gismondine were formed, but they were not observed in D50. It is believed that the reaction products in seawater-mixed AAS samples was helpful in preventing the water percolation, and thus, the electrical resistivity increased compared with the deionized water-mixed sample. As the increase of electrical resistivity indicates the decrease of porosity in samples, it is expected that the seawater-mixed AAS pastes would exhibit better durability than the deionized water-mixed pastes.