Effects of Composition Type and Activator on Fly Ash-Based Alkali Activated Materials

The compressive strengths of fly ash-based alkali-activated materials (AAM), produced using various activators of only sodium hydroxide, were measured. Fly ash-based AAM specimens, produced by mixing different kinds of fly ash and ground granulated blast-furnace slag (GGBFs) with an activator containing only sodium hydroxide, were cured at ambient temperature, and then placed in air for different numbers of days. The short- and long-term compressive strengths and shrinkage of fly ash-based AAM were measured and compared to one another. The effects of type of fly ash, alkali-equivalent content, GGBFs replace percentage, and ages on the compressive strengths and shrinkage of fly ash-based AAM were investigated. Even when different fly ash was used as the raw material for AAM, a similar compressive strength can be achieved by alkali-equivalent content, GGBFs replaces percentage. However, the performance of shrinkage due to different types of fly ash differed significantly.


Introduction
The latest UN figures suggest that despite current government commitments to reduce greenhouse gas emissions, atmospheric concentrations continue to rise, keeping the earth on a trajectory to levels of warming that will precipitate further environmental, social, and economic disruption and suffering on unprecedented scales [1]. Global GHG emissions increased by 1.5-fold since 1990. With several greenhouse gases in the atmosphere, including methane, nitrous oxide, and ozone-depleting substances, and greatly reducing CO 2 is the most straightforward method to solve the continuous rise of air temperature [2]. As a major emitter of CO 2 in the United States, the top three industries of CO 2 emission are transportation, electricity, and manufacturing and building industry, which contribute 35%, 31%, and 16% of CO 2 emissions, respectively, meaning the three industries account for more than 80% of total CO 2 emissions [3].
Concrete is an extensively used construction material, and Portland cement is the foremost binder material, as well as the main source of concrete hydraulicity. However, a lot of CO 2 is generated during the production process of any type of cement, which adversely affects the environment. Therefore, research is needed to find a substitute cement and reduce carbon emissions [4][5][6].
Geopolymer technology is a very effective industrial innovation, which uses the material made of aluminosilicate mineral and alkaline activity bath reaction, wherein most of the silicate mineral can be replaced by industrial waste. It was named "geopolymer" by Joseph Davidovits in 1979 [7]; the amorphous semicrystalline tri-dimensional alumino-silicates can rapidly form natural alumino-silicates solid materials under normal temperature through alkali reactivity. In addition, the geopolymer has an excellent heat resistance and fire resistance, and its carbon emission is lower than that of the conventional Portland cement [8]. Hence, it is regarded as a potential substitute building material that can be extensively In this study, we used three types of fly ash and their chemical properties are presented in Table 1. The X-ray diffraction (XRD) patterns of the powder shown in Figure 1, which were obtained at a scanning rate of 2θ/min and over a scanning range of 10 • -80 • , revealed different amorphous characteristics. The amorphous percentage of the four raw materials were 46.2, 56.7, 67.8, and 94.9%, respectively. The size distribution of the fly ash, as characterized using an INSITEC laser diffraction particle size analyzer, is shown in Figure 2. The mean geometric sizes of the three fly ash kinds were estimated to be 21.01 µm, 27.1 µm, and 26.5 µm, with standard deviations of 3.897, 3.568, and 5.758 respectively.

Activator
The fly ash and ground granulated blast-furnace slag (GGBFs) were alkali-activated by mixing with an activator in the production of AAM. The activator used here was a mixture of water and sodium hydroxide (NaOH) (reagent grade, 97% purity: Showa Chemical Industry Co., LTD, Tokyo, Japan). The chemical reaction, microstructure, and properties of an AAM can be affected dramatically by the amount of water in an activator. Therefore, the water/solid ratio was fixed at 0.3 for all fly ash-based AAM specimens. Only alkali-equivalent content parameters were employed in this study. The water/solid ratio (W/S) is the weight ratio of water to the sum of solid (includes powders and NaOH). The alkali-equivalent content, denoted by AE%, is defined as the weight fraction of Na 2 O to powders.

Sample Preparation, Mixing, and Curing
The activator plays an important role in determining the microstructure and properties of fly ash-based AAM specimens. To evaluate the effects of NaOH on the compressive strengths of fly ash-based AAM, activators with various alkali-equivalent content of AE% = 3%, 4%, 5%, 6%, 7%, and 8% were used in the production of fly ash-based AAM specimens, and replaced by 10%, 20% and 30% slag with a water/solid ratio of 0.3. The specimen proportions example is given in Table 2. For each activator, the required amounts of NaOH, and water were weighed, mixed, and then placed in a container until room temperature was reached. Next, powders were added to the container and stirred vigorously for 2 min in a 5L Hobart mixer. After complete mixing, the AAM paste was poured into 3 × 3 × 3 cm steel molds, with a total of 18 cubes cast of each mix for the compressive strength tests and further compacted on a vibrating table (CONTROLS, frequency of 60 Hz) to get rid of any air bubbles. The steel molds were covered with plastic wrap to prevent the evaporation of moisture and then cured at an ambient temperature. The specimen size used for the shrinkage test was 25 × 25 × 285 mm, mixed, molded, and cured as same as above, with a total of three samples for each proportion. One day later, the specimens were demolded.

Pozzolanic Strength Activity Index
This test was carried out by reference to ASTM C311 [17], in which the 7-and 28-day compressive strengths of mortar cubes with a 20% mass replacement of cement by fly ash were compared to those of control without fly ash, at constant flow conditions. This was used to investigate the activity of different types of raw materials.

Workability
According to the mixing conditions in Section 2.3, the flow test was performed for the mixed specimen referring to ASTM C230 [18]. The paste was poured into the top split conical ring, and the flow table was bounced 25 times within 15 s after the conical ring was removed to measure the flowability of the mixture. The influence of the flowability of different slag substitution amounts was observed in different alkali-equivalent content conditions on workability.

Setting Time Test (Vicat Needle)
Firstly, in line with Section 2.3, AAM pasteis mixed by pouring in the conical ring (a height of 40 mm, an inside diameter at the bottom of 70 mm, and an inside diameter at the top of 60 mm). Periodic penetration tests are performed on this paste by allowing a 1 mm Vicat needle to settle into this paste. The Vicat initial time of setting is the time elapsed between the initial contact of cement and water and the time when the penetration is measured or calculated to be 25 mm. The Vicat's final time of setting is the time elapsed between initial contact of cement and water and the time when the needle does not leave a complete circular impression in the paste surface. Since the setting time test is quite sensitive, the sample needed to be placed at a temperature of 23 • C and relative humidity of not less than 95%.

Compressive Strength Test
To determine the influence of the mixtures of different mix proportions on compressive strength, the specimens of all mix proportions were mixed according to Section 2.3 and made into a 3 × 3 × 3 cm specimen, which was hardened and cured under normal temperature. The compressive strength tests were carried out with reference to ASTM C109 [19]. The specimen was placed into the compression tester according to the curing ages of days 3, 7, 14, 28, 56, and 91 for compressive strength tests, and the strength was measured and recorded. The average of three specimens was used for each test.

Drying Shrinkage Test
The alkali-activated cementing material always has problems in volume stability, especially the alkali-activated slag, its dry shrinkage is quite large. Besides using fly ash as a base to produce AAM, this study also adopted a small amount of slag to replace fly ash. Therefore, it was necessary to measure the long-term volume stability. The drying shrinkage mold used in this study was a 25 × 25 × 285 mm steel die, and the measurement ages included days 3, 7, 14, 28, 56, and 91. The drying shrinkage test result showed that the volume stability of materials is very important for the usability of materials.

Microscopic Test
The raw materials were mixed according to Section 2.3. The reaction of fly ash in the specimen was considered more complete and the structure was more intact after 91 days of curing; hence, the specimen was cured at room temperature until 91 days as a microscopic test sample. The specimens made of three different materials were extracted from the mix proportion with the maximum compressive strength for microscopic test analysis. Firstly, the samples were pulverized and then vacuum dried until they reached a constant weight. Then, a portion of the sample was analyzed by XRD to determine the mineral components of the products. The XRD measurement was done with a D4 (Bruker) using a Co-Tube and equipped with a LynxEye detector. The settings were fixed divergence slits (0.5 • ), 0.04 rad Soller slits, and a step size of 0.02. The other part of the sample was placed on the Fourier transform infrared (FTIR) spectrometer for the recording of their infrared spectrum. The FTIR spectrum was recorded using a BRUKER, TENSOR II FT-IR Spectrometer over the wavelength range of 400 cm −1 to 4000 cm −1 . The resolution of the measurement was 4 cm −1 . After extracting and crushing some of the specimens, the fine particle samples were dried in a vacuum environment until they reached a constant weight. They were then Polymers 2022, 14, 63 6 of 15 placed in the Scanning Electron Microscope (SEM) to observe the extent of reaction of fly ash and the pore structure of products.

Pozzolanic Strength Activity Index
The test was performed referring to ASTM C311 [17], and the result is shown in Table 3. The activity indexes of the three kinds of fly ash exceeded 95% on Day 7, proving good activity, and the activity indexes were excellent at 117%, 125%, and 127% on Day 28. The activity index of GGBFS on Day 7 was 96%, and on Day 28 was 117%, meeting the Grade 120 furnace slag of ASTM C989 [20]. Data in Table 3 indicate that fly ash C has the highest activity index of all materials at 28 days and all three kinds of fly ash activity index are greater than GGBFs.

Flowability
The flow of mix proportions of AE% = 3, 5, and 8% of the three kinds of fly ash was tested as shown in Figure 3. As the setting time of fly ash C was much longer than that of fly ash A and B, it is less likely to cure and has a much higher flow rate than the other two kinds of fly ash. The flowability of various pastes increased with the alkali-equivalent content, because in the case of the same total mixing amount, the mixture with higher alkali-equivalent content has a lower total content of aggregate, and the flow of paste increased slightly with the slag substitution amount. However, when AE = 5 and 8%, and the slag replacement rate increased to 20%, the flow decreased because the rapid setting reaction of slag [21] reduced the flowability. However, when the slag replacement was 10%, the effect of slag on workability might be slighter than the effect of fly ash, as the fineness of the three kinds of fly ash was much lower than that of slag.

Setting Time
The setting time of mix proportions of AE% = 3, 5, and 8% of the three kinds of fly ash was tested. Wang et al. [21] reported a rapid setting problem of the slag. Consequently, our research modified the overlong setting time of pure fly ash.
As shown in Figure 4, the setting time of the AAM made of fly ash C was apparently longer than that of the other two kinds of fly ash, which results from different properties of the raw materials. Therefore, according to the XRF of the raw materials in Table 1, the CaO content in the fly ash C was 15.95%. This might not were caused by the insufficient CaO content to form AAM or other C-S-H colloids with silicates. Mortureux et al. [22] mentioned that the alkali-activated colloid in Ca form is likely to form in the case of high NaOH concentration; however, the colloid types formed by geopolymerization at a low alkali liquor concentration in this study were SiO 4 and AlO 4 -tetrahedral structures. After QXRD by Rietveld quantitative analysis, the composition of the feed material is known, as shown in Section 2.1, the amorphous content of fly ash C was the highest at 67.8%, and that of fly ash B and fly ash A were 56.7% and 46.2%, respectively. This might be because the amorphous content was too high. In the case of low alkali liquor concentration and slag replacement rate, the originally unlikely polymerization was difficult to happen, and the paste could not be hardened. When AE% = 5 and 8%, as shown in Figure 4c-f, when the slag replacement rate was 20%, the mixture setting time could be shortened greatly. When the slag replacement rate was 30%, its effect on shortening the setting time was slighter. content, because in the case of the same total mixing amount, the mixture with higher alkali-equivalent content has a lower total content of aggregate, and the flow of paste increased slightly with the slag substitution amount. However, when AE = 5 and 8%, and the slag replacement rate increased to 20%, the flow decreased because the rapid setting reaction of slag [21] reduced the flowability. However, when the slag replacement was 10%, the effect of slag on workability might be slighter than the effect of fly ash, as the fineness of the three kinds of fly ash was much lower than that of slag.

Setting Time
The setting time of mix proportions of AE% = 3, 5, and 8% of the three kinds of fly ash was tested. Wang et al. [21] reported a rapid setting problem of the slag. Consequently, our research modified the overlong setting time of pure fly ash.
As shown in Figure 4, the setting time of the AAM made of fly ash C was apparently longer than that of the other two kinds of fly ash, which results from different properties of the raw materials. Therefore, according to the XRF of the raw materials in Table 1, the CaO content in the fly ash C was 15.95%. This might not were caused by the insufficient CaO content to form AAM or other C-S-H colloids with silicates. Mortureux et al. [22] mentioned that the alkali-activated colloid in Ca form is likely to form in the case of high NaOH concentration; however, the colloid types formed by geopolymerization at a low alkali liquor concentration in this study were SiO4 and AlO4-tetrahedral structures. After QXRD by Rietveld quantitative analysis, the composition of the feed material is known, as shown in Section 2.1, the amorphous content of fly ash C was the highest at 67.8%, and that of fly ash B and fly ash A were 56.7% and 46.2%, respectively. This might be because the amorphous content was too high. In the case of low alkali liquor concentration and slag replacement rate, the originally unlikely polymerization was difficult to happen, and the paste could not be hardened. When AE% = 5 and 8%, as shown in Figure 4c-f, when the slag replacement rate was 20%, the mixture setting time could be shortened greatly. When the slag replacement rate was 30%, its effect on shortening the setting time was slighter. According to the increasing alkali equivalent concentration, when AE% = 8%, as shown in Figure 4e,f, the overall setting time was shorter than that of AE% = 3%, corroborating the result of Gebreziabiher et al. [23], which indicated that the setting time can be shortened by using high concentration alkaline activator. When the alkali-equivalent content was increased to AE% = 8%, the paste of this study had the fastest polymerization, and the reaction rate increased with the slag replacement ratio. When AE% = 8% and the slag replacement rate was 30%, the initial setting time of the pastes made of fly ash A, fly ash B, and fly ash C was 50, 65, and 50 min, respectively; thus, there were few differences in the setting time of the three kinds of fly ash from different plants.

Compressive Strength
The compressive strength of all mixtures was tested. When AE% = 3%, as shown in Figure 5a, the compressive strengths of fly ash A were 16 rate was 30% on Day 91, while the strength of replacement rate of 10% or 20% decreased slightly on Day 91. This is because the optimum amount of alkali required for fly ash differs from that required for slag, and there was no excess free alkali at 30% slag replacement to cause a late strength decline. The fly ash C had better compressive strength than the other two kinds when the slag replacement rates were 20% and 30%, with compressive strengths were 30.24 and 28.88 MPa, standard deviations of 2.34 and 1.27 on Day 91. However, when the slag replacement rate was 10%, the strength was much lower than that of the same mix proportion of the other two kinds of fly ash, with a compressive strength of 6.79 MPa and a standard deviation of 0.88. Based on the setting time of the mix proportions shown in Figure 4, it was suspected to be because the paste does not have adequate alkali content, resulting in inadequate reaction and polymerization and that the incompletely reacted alkali-activated liquid in the paste could not be retained. As a result, the strength did not yet improve after a longer age.

Drying Shrinkage
This study employed AE% = 3, 5, and 8% and the alkali-equivalent content in Figure 6 for the drying shrinkage test. When AE% = 5%, as shown in Figure 6b, and the slag replacement rate was 10%, the fly ash B expanded acutely and shrank gradually after the age of 14 days. In the mix proportions of fly ashes A and C, the shrinkage amplitude increased gradually with age; the shrinkage amplitude was larger than that of AE 3%. When AE% = 8% as shown in Figure 6c, and the slag replacement rate of the fly ash B was 10%, the fly ash B exhibited more severe expansion than AE 5%, and the expansivity was 0.5813% at the age of 91 days.
The mix proportion of fly ash C had a severe shrinkage. At the age of 91 days, for AE% = 3% as shown in Figure 6a, when the slag substitution amounts were 10, 20, and 30%, the amounts of change in length were −1.718%, −2.315%, and −2.497%, respectively.
The shrinkage of mix proportion of fly ash A was next to fly ash C, when AE% = 3% and the slag substitution amounts were 10, 20, and 30%, the amounts of change in length are −1.1893%, −1.768%, and −2.309%, respectively.  According to the increasing alkali equivalent concentration, when AE% = 8%, as shown in Figure 4e,f, the overall setting time was shorter than that of AE% = 3%, corroborating the result of Gebreziabiher et al. [23], which indicated that the setting time can be shortened by using high concentration alkaline activator. When the alkali-equivalent content was increased to AE% = 8%, the paste of this study had the fastest polymerization, portion of slag substitution amount of 10% decreased slightly during Day 56 to Day 91, with a compressive strength of 14.58 MPa at 56 days, a standard deviation of 0.82, and 11.05 MPa at 91 days, a standard deviation of 1.51. The other mix proportions did not have an obvious uptrend or downtrend. When AE = 5%, as shown in Figure 5c, the strength of various mix proportions was smooth after Day 56. Contrary to the smooth development in the strength when AE% = 4% on Day 28, the strength increased in the later stage with the increase in alkali-equivalent content. When the alkali equivalent concentration was AE% = 6%, as shown in Figure 5d, the strength of various mix proportions increased with age. The strength of various mix proportions of fly ash A still increased significantly on Day 91, the compressive strengths of 24.53, 33.12, and 37.22 MPa with standard deviations of 1.35, 1.48 and 1.49 at the slag replacement rates of 10%, 20%, and 30% on Day 91. When AE% = 7 and 8%, as shown in Figure 5e,f, the optimum alkali-equivalent content of mix proportion of fly ash A was for the drying shrinkage test. When AE% = 5%, as shown in Figure 6b, and the slag replacement rate was 10%, the fly ash B expanded acutely and shrank gradually after the age of 14 days. In the mix proportions of fly ashes A and C, the shrinkage amplitude increased gradually with age; the shrinkage amplitude was larger than that of AE 3%. When AE% = 8% as shown in Figure 6c, and the slag replacement rate of the fly ash B was 10%, the fly ash B exhibited more severe expansion than AE 5%, and the expansivity was 0.5813% at the age of 91 days. The mix proportion of fly ash C had a severe shrinkage. At the age of 91 days, for AE% = 3% as shown in Figure 6a, when the slag substitution amounts were 10, 20, and 30%, the amounts of change in length were −1.718%, −2.315%, and −2.497%, respectively.
The shrinkage of mix proportion of fly ash A was next to fly ash C, when AE% = 3% and the slag substitution amounts were 10, 20, and 30%, the amounts of change in length are −1.1893%, −1.768%, and −2.309%, respectively.
According to the test result, the higher the GGBFs replacement rate was, the more severe the shrinkage, which is due to the very large degree of autogenous shrinkage of the stone itself [24]. The higher the alkali-equivalent content of fly ash C and fly ash A According to the test result, the higher the GGBFs replacement rate was, the more severe the shrinkage, which is due to the very large degree of autogenous shrinkage of the stone itself [24]. The higher the alkali-equivalent content of fly ash C and fly ash A was, the larger the amount of deformation from shrinkage was. However, the expansion of fly ash B increased with alkali-equivalent content. According to the compound composition of various raw materials in Table 1, the highest CaO content was in fly ash C at 16.88%, so its expansion may be the volume expansion resulting from the reaction of a high content of F-CaO and water.

X-ray Diffraction Analysis
The maximum compressive strength of the specimens in this study occurred in the fly ash A specimen when AE% = 7% and slag replacement rate was 20%; hence, the specimens of three kinds of fly ash and the same mix proportion were extracted for XRD analysis. Figure 7 shows the diffractograms of three results. When 2θ = 25 • -26 • , there was an obvious quartz mineral diffraction peak, forming the mineral components of the main crystal with Mullite. When 2θ = 28 • , the presence of calcium carbonate was detected, meaning that Ca in the raw material had not fully entered into polymerization to generate strong minerals. ysis. Figure 7 shows the diffractograms of three results. When 2θ = 25°-26 obvious quartz mineral diffraction peak, forming the mineral componen crystal with Mullite. When 2θ = 28°, the presence of calcium carbonate meaning that Ca in the raw material had not fully entered into polymerizat strong minerals.

Scanning Electron Microscopy Observation
The mixture of three different materials in the same mix proportion wa ordinary temperature for 91 days before SEM tests. As shown in Figure 8 sample exhibited fewer residual fly ash particles than fly ash B and fly ash ture was denser. This result is the same as the result of compressive streng The fly ash A exhibited the highest compressive strength. As fly ash C cont unreacted fly ash, the alkali-activated liquid was suspected to increase con more spherical vitreous surface of fly ash could be damaged to dissolve th silica constituent, and a higher strength specimen could possibly be polym quently, different kinds of fly ash possessed different optimum alkali-equiv and the reactive fly ash contents in the same mix proportion were differen

Scanning Electron Microscopy Observation
The mixture of three different materials in the same mix proportion was cured under ordinary temperature for 91 days before SEM tests. As shown in Figure 8, the fly ash A sample exhibited fewer residual fly ash particles than fly ash B and fly ash C, and its texture was denser. This result is the same as the result of compressive strength in Figure 5. The fly ash A exhibited the highest compressive strength. As fly ash C contained the most unreacted fly ash, the alkali-activated liquid was suspected to increase continuously, the more spherical vitreous surface of fly ash could be damaged to dissolve the internal rich silica constituent, and a higher strength specimen could possibly be polymerized. Subsequently, different kinds of fly ash possessed different optimum alkali-equivalent contents, and the reactive fly ash contents in the same mix proportion were different. Figure 9 shows the infrared spectrogram of the three specimens. An apparent peak was observed at 1020 cm −1 , which represents the Si-O and Si-O-Si stretching vibration absorption peak [25]. There is a micro peak at 875 cm −1 , wherein the absorption band and 1400 cm −1 are recognized as the vibration of the O-C-O bond of carbonate [26]. As the raw material was free of carbonate, certain carbonization might have occurred during the polymerization process. This result matched with the carbonate product in the XRD analysis shown in Figure 7, it is universally regarded as the cause for deterioration of the geopolymerization product [27].  Figure 9 shows the infrared spectrogram of the three specimens. An apparent peak was observed at 1020 cm −1 , which represents the Si-O and Si-O-Si stretching vibration absorption peak [25]. There is a micro peak at 875 cm −1 , wherein the absorption band and 1400 cm −1 are recognized as the vibration of the O-C-O bond of carbonate [26]. As the raw material was free of carbonate, certain carbonization might have occurred during the polymerization process. This result matched with the carbonate product in the XRD analysis shown in Figure 7, it is universally regarded as the cause for deterioration of the geopolymerization product [27].

Conclusions
(1) The ultimate strength of specimens occurred in the specimen of fly ash A in this study. The activity index on Day 28 of fly ash A was only 117%, and the amorphous form was the lowest among the three raw materials, indicating the activity index and amorphous

Conclusions
(1) The ultimate strength of specimens occurred in the specimen of fly ash A in this study. The activity index on Day 28 of fly ash A was only 117%, and the amorphous form was the lowest among the three raw materials, indicating the activity index and amorphous content are not in absolute relation to AAM strength. (2) The optimum alkali-equivalent content in fly ash A was AE% = 7%, with the best compressive strength 38.58 MPa at 20% slag substitution amount. The mix proportion with the second-best compressive strength occurred in fly ash C, at AE% = 8%. The compressive strength was 36.63 MPa when the slag substitution amount was 20%. The fly ash B exhibited the lowest strength. The compressive strength was 33.77 MPa when AE% = 7% and the slag substitution amount was 30%. (3) The fly ash C had the longest setting time and was the most amorphous. Without adequate alkali-equivalent content and slag substitution amount, the polymerization was harder to happen. (4) According to the drying shrinkage test, the AAM of fly ash C had very large shrinkage because it was unlikely to set. The AAM of fly ash B exhibited too high f-CaO content, as it reacts with water and expands. Therefore, the fly ash material greatly influences the basic properties.