Delayed setting time for alkali-activated slag composites using activator containing SiO₂ and Na₂O

Alkaline activators are made up of water, sodium silicate, and sodium hydroxide. The content of SiO₂ and Na₂O in a given alkaline activator mix design depends on the volume of alkaline activator. To determine the SiO₂: Na₂O ratio, the sodium silicate solution was first weighed to determine the weight of Na₂O and H₂O. These values were then used to calculate the amount of NaOH and Na₂O required to equal the Na₂O content in the alkaline activator mix. Finally, the weight of water (distillate water) consumed in the preparation of alkaline activator is equal to the weight of total water minus the weight of water in the sodium silicate solution. In cases where phosphoric acid is added, we deduct the weight of phosphoric acid from the weight of water consumed in the preparation of the alkaline activator.

The process of mixing water, sodium silicate, sodium hydroxide, and phosphoric acid must be implemented in a systematic manner. This process begins with the mixing of sodium silicate with water, to which is added sodium hydroxide under stirring until it dissolves entirely (including sediment). A reaction between sodium silicate and sodium hydroxide increases the temperature of the resulting solution. In the event that phosphoric acid is added to the solution, it is important to maintain stirring until all of the sediment (i.e., sodium ions and phosphoric acid) disappears. When the alkaline activator reaches room temperature, it is ready for use in the production of AAS.

Solid sodium hydroxide (in the form of pellets), liquid sodium silicate, and pure water were used in the preparation of alkaline activators with various SiO₂ and Na₂O ratios. The slag in the current study was ground granulated blast-furnace slag, the composition of which is listed in Table 1. The fineness of slag was 4,000 cm²/g, and the grain size distribution of the slag is summarized in Table 2. The sodium hydroxide used in the experiments was of reagent grade, the chemical composition of which is shown in Table 3. The sodium silicate was extrapure reagent grade, the chemical composition of which is provided in Table 4.

In all AASC designs in this study, the liquid/binder ratio (i.e., the ratio of alkaline activator to slag by weight) was fixed at 0.5. The mix designs used in this study are listed in Table 5. In each series of experiments, the pH of the powder samples ground from the concrete specimens was measured in accordance with the specifications of the American Society for Testing and Materials (ASTM)-E70. Initial and final setting times were established using the Vicat test in accordance with ASTM C191 specifications. Scanning electron microscopy (SEM) analysis was used to characterize representative samples measuring 1 mm × 1 mm × 1 mm, in accordance with ASTM C1723 specifications. The specimen was mixed and placed in a steel mold and left at room temperature for 24 h before unmolding. The AASC specimens were then cured in saturated lime water for 3 days and subjected to SEM. The specimens for SEM were cleaned and placed in an oven (105°C, 24 h), dried, and then removed to room temperature (approx. 5 h); next, they were placed in a vapor-plating apparatus for gold plating. The specimens were then vacuum-dried (120 s) to complete the preparation process.

NaOH, Na₂SiO₃, and H₂O were combined in various proportions to form the various activators. Note that using aqueous NaOH alone would likely result in excessively rapid setting. Hence, sodium silicate was added to the aqueous NaOH solution to retard this effect. Note that the addition of sodium silicate to highly alkaline NaOH lowered the pH. The silicate-based activator is a mixture of sodium hydroxide and sodium silicate; therefore, the production of SiO₂ can be traced back to Na₂SiO₃ (aq.), whereas the production of Na₂O can be attributed mainly to NaOH (s.) with a small amount derived from Na₂SiO₃ (aq.). As shown in Figure 1, increasing the concentration of SiO₂ with a fixed concentration of Na₂O reduced the pH value of the activator. Increasing the concentration of Na₂O with a fixed concentration of SiO₂ increased the pH of the activator. This can be attributed to the fact that a higher concentration of SiO₂ resulted in the production of larger quantities of SiO32− {\rm{SiO}}_3^{2 - } capable of reacting with H⁺ ions from the activator. Conversely, a higher concentration of Na₂O made available a larger number of Na⁺ ions to react with OH⁻ ions from the activator.

Fig. 1

The composition of alkaline activators can be characterized in terms of alkali modulus and total alkali content. Alkali modulus refers to the weight per unit volume of SiO₂ divided by the weight per unit volume of Na₂O (SiO₂/Na₂O). Total alkali content refers to the total weight per unit volume of SiO₂ plus the total weight per unit volume of Na₂O (SiO₂+Na₂O). As shown in Figure 2, increasing the dosage of SiO₂+Na₂O produced a corresponding increase in the pH value when the alkali modulus (SiO₂/Na₂O) was fixed at 1.0. Increasing the dosage of SiO₂+Na₂O increased the availability of both SiO₂ and Na₂O; however, the quantity of OH⁻ (from Na₂O) exceeded the quantity of H⁺ (from SiO₂), which caused the increase in pH. As shown in Figure 3, when the dosage of alkaline activator was fixed at 160 g/l or 200 g/l, the higher pH values were due to a lower relative proportion of SiO₂ or higher proportion of Na₂O. In other words, with the total quantity of SiO₂ and Na₂O fixed, the higher pH values can be attributed to a lower alkali modulus.

Fig. 2

Fig. 3

As shown in Figure 4, increasing the alkali modulus led to a corresponding decrease in the pH value of the activator, i.e., pH values varied only with a change in the alkali modulus. Figure 5 illustrates the relationship between pH and SiO₂+Na₂O. The analysis of variations in pH under a fixed SiO₂+Na₂O dosage indicates that the SiO₂+Na₂O dosage had no effect on pH. The alkali modulus (SiO₂/Na₂O) can be used to indicate an increase in SiO₂ content, resulting in a lower pH. This means that adding sodium silicate to a silicate-based activator (to retard the setting process) increases the alkali modulus (SiO₂/Na₂O) to above that of aqueous NaOH solution, while reducing the pH to below that of aqueous NaOH. Zhang's [28] study indicated that excessive SiO₂ content could hamper the alkali activation reaction and the excessive SiO₂ content decreases the overall pH, leading to a slower rate of reaction. The conclusion of the cited study is consistent with that of our study.

Fig. 4

Fig. 5

The effects of SiO₂ and Na₂O on the setting time were examined by varying their respective concentrations. As shown in Figure 6, increasing the concentration of SiO₂ under a fixed Na₂O concentration reduced the setting time. It was due to the rapid production of the initial calcium silicate hydrate (C–S–H) gel as a result of the bonding of the calcium ions present in the slag to the silicate ions in the pore solution, which can harden the AASC paste. It is also consistent with the conclusions of the studies by Chang [29] study and Tong et al's [30]. At SiO₂ concentrations of 100 g/l or 120 g/l, increasing the concentration of Na₂O decreased the initial and final setting times. However; the effect was insignificant at other SiO₂ concentrations.

Fig. 6

Previous studies have demonstrated a correlation between the setting time of AAS pastes [29] and the concentration of SiO₂ and Na₂O in the alkaline activator [30, 31]. It was then sought to determine whether the observed effect was the result of decreases in pH. Therefore, experiments were performed comparing the influence of alkali modulus (SiO₂/Na₂O) and total alkali content (SiO₂+Na₂O) on the setting time. Under a constant alkali modulus, increasing the SiO₂+Na₂O dosage decreased the setting time (see Figure 7), while increasing the content of both sodium silicate and NaOH. It was observed that higher reactant concentrations produced higher reaction rates. Increasing the content of aqueous silica led to a higher SiO32− {\rm{SiO}}_3^{2 - } concentration, with a corresponding increase in the reaction rate. Increasing the NaOH concentration was also shown to increase the reaction rate, due to an increase in the [OH⁻] concentration, which facilitated the dissolution of Ca²⁺ from the slag.

Fig. 7

As shown in Figure 8, increasing the concentration of SiO₂ decreased the setting time, regardless of the SiO₂+Na₂O dosage. In other words, a higher alkali modulus (indicating a larger quantity of Na₂O dissolved from the sodium silicate) could be used as an index indicating a shorter setting time. The results confirmed that NaOH facilitated the dissolution of Ca²⁺ from slag particles; however, the content of aqueous silica also played a key role in the early hydration of AAS. From the results, it can be seen that the alkali content of SiO₂ had a significantly higher effect on the setting time than Na₂O. For a fixed total alkali content (SiO₂+Na₂O), the higher the SiO₂ content, the shorter was the setting time, which is in line with previous studies [29].

Fig. 8

Regression analysis was used to compare the effects of alkali modulus (SiO₂/Na₂O) and total alkali content (SiO₂+Na₂O) on the setting time. As shown in Figure 9, there was a relationship between alkali modulus and setting time; however, a clear link was observed between alkali modulus and pH. From this, we can conclude that pH did not affect the initial or final setting time.

Fig. 9

As shown in Figure 10, it was confirmed that SiO₂+Na₂O dosage (i.e., the total concentrations of SiO₂ and Na₂O) had a pronounced effect on setting time, in a dose-dependent manner. This suggests that retardation of setting time cannot necessarily be attributed to a reduction in pH. The tests revealed that at SiO₂ = 60 g/l, the setting time retardation effects became evident under a high alkali modulus and became far more pronounced under a low alkali modulus. The effect of total alkali content on the setting time presented a clearer correlation.

Fig. 10

SEM was used to examine the effects of SiO₂+Na₂O dosage on the microstructure of AASC specimens. As shown in Figure 11, two specimens of the same alkali modulus presented a fine pore structure. Although the difference in pH values between the two specimens was not significant, the difference in SiO₂+Na₂O dosage resulted in different microstructures. The differences in microstructure were mainly influenced by the setting time. SEM images obtained at 3 days revealed a large number of microfractures on the surface of the samples prepared using a paste containing 120 g/l of SiO₂ and 120 g/l of Na₂O. This can be attributed to the production of AAS amorphous reaction products under the effects of rapid hydration and condensation. These results support those of previous studies by Collins and Sanjayan [32], Yang et al. [33], and Opiso et al. [34]. By comparison, specimens prepared using paste containing 80 g/l of SiO₂ and 80 g/l of Na₂O revealed totally and partially hydrated slag grains (due to a slow hydration reaction), as well as fewer microcracks. The main reaction products were fibrous structures corresponding to C–S–H gel, dense gel phase, partially reacted cenosphere, and a fully reacted matrix. A small quantity of calcium aluminosilicate hydrate gel (C–A–S–H gel) was also observed.

Fig. 11

The partially hydrated particles would continue to increase the strength of the paste as it was cured over time. Under the activator composition of SiO₂/Na₂O = 1, the pastes set rapidly, but no visible cracks were observed on the microstructure surface. The faster setting time resulted in more partially hydrated slag grains on the surface of the specimens. The primary benefit of slowing the setting process through the addition of silicate-based activator is enhanced workability and the production of a complex microstructure with fewer cracks and more stable strength development. The results are consistent with previous studies by Atiş et al. [35] and Mastali et al. [36]. AAS has a more compact and dense crystal structure than the structure of C–S–H, formed during the hydration of ordinary Portland cement-based composites proposed by Alharbi et al. [37] and Paradiso et al. [38].