The Effect of Sodium and Magnesium Sulfate on Physico-Mechanical and Microstructural Properties of Kaolin and Ceramic Powder-Based Geopolymer Mortar

: Recent trends in reducing the ecological footprint of the construction industry have increased the attention surrounding the use of alternative binding systems. Among the most promising are geopolymer binders, which were found to have the capability to substantially reduce the environmental impact of Portland cement use. However, even the use of this alternative binding system is known to be heavily dependent on the use of industrial byproducts, such as precursors and an alkaline source, produced through an energy intensive process. To address this and provide a greener route for this binding system, this study adopts the use of natural kaolin and raw ceramic powder as the main precursors. The activation process is performed by using solid potassium hydroxide in conjunction with sodium and magnesium sulfate, which are naturally available, to produce geopolymers. To assess the resulting geopolymer samples, 28 mixes are produced and a series of physico-mechanical and microstructural analyses is conducted. The results show that the use of ceramic powder can improve the physico-mechanical properties by reducing porosity. This, however, requires a relatively higher alkalinity for activation and strength development. These ﬁndings are further conﬁrmed with the XRD and FTIR results. Nonetheless, the use of ceramic powder with sodium and magnesium sulfate is found to result in a more coherent and homogenous microstructure, compared to the geopolymers produced with potassium hydroxide and kaolin. The ﬁndings of this study point to the suitability of using sodium and magnesium sulfate for the cleaner production of kaolin and ceramic powder-based geopolymers.


Introduction
The increasing socioeconomic effect of climate change and natural disasters has caused a growing body of international and intergovernmental entities to advocate carbon neutrality and aim for a net-zero carbon dioxide production. Yet, the incorporation of carbon neutrality into construction requires a fundamental shift from the commonly accepted use of polluting practices, such as the heavy dependence on ordinary Portland cement (OPC) to alternative cementing materials and systems, such as geopolymer technology [1]. According to previous studies that conducted a life cycle assessment (cradle-to-grave), for instance, geopolymer binders have been reported to be able to reduce the overall CO 2 production of concrete by approximately 40% [2,3] to 75% [4] when compared to Portland cement concrete. Although this reduction rate was considerable [5][6][7], after approximately a century of development, geopolymer binders are not commonly being used for major infrastructural projects. In part, this could be due to the need for the utilization of polluting Sustainability 2022, 14, 13496 3 of 24 In that respect, in this study, two novel additives of sodium and magnesium sulfate that are naturally occurring minerals with formulas Na 2 SO 4 and MgSO 4 , respectively, are used as additives to enhance the activation process. Although both chemical compounds are highly soluble in water, the pH of Na 2 SO 4 is reported to be approximately 8.5 [49], while MgSO 4 has a lower pH of 6-7 [50]. Using lower pH activators can have major applications, such as the conservation of reactive metals and resins that do not react with encasing materials if the pH is not very high [51]. To date, however, very few studies have utilized naturally occurring sodium and magnesium sulfate as additives in geopolymer concrete. Rashad et al. [51], for instance, utilized sodium sulfate as a naturally occurring material to activate blast furnace slag. In their analysis, it was reported that the fineness of the precursor particles was more effective for the geopolymerization rate than increasing the sodium sulfate content. Rattanasak et al. [52] used sodium sulfate, among other additives, for its lower pH and reactivity to control the setting time of geopolymer concrete. It was reported that the presence of sodium sulfate retarded the leaching of silica and alumina from high-calcium coal fly ash (class C) and increased the setting time significantly. The reported results, however, showed an enhanced compressive strength later on. Ma et al. [48] reported that the addition of sodium sulfate could reduce the heat of geopolymerization due to its lower pH, and, thus, reduce the drying shrinkage of geopolymer specimens. Nonetheless, the addition of sodium sulfate has been reported to increase the overall content of microcracks. Qing-feng et al. [53] studied the microstructural properties of a coal fly ash-based geopolymer activated using sodium silicate and sodium sulfate. In their analysis, it was reported that only a higher content of sodium sulfate could cause a higher peak at the symmetrical stretching vibration of Si-O, Si-O-Si and Al-O-Si, which translated into promoting the formation of N-A-S-H gels. Jun and Oh [54] utilized sodium sulfate as an additive to coal fly ash (F)-based geopolymer concrete, and reported an enhanced strength development due to a change in pore size distribution. It was also reported that the addition of sodium sulfate resulted in higher strength values for lower Si/Al ratios of reaction products.
Although the aforementioned studies provided a significant contribution to alternate routes for the cleaner production of geopolymer concrete, the used precursors and activators in most of the mentioned studies were common coal fly ash and blast furnace slag, activated using a liquid sodium silicate [55] activator. To address this and practice an alternative route for the cleaner production of geopolymer binders, this study adopts the use of naturally occurring kaolin and raw ceramic powder, activated using potassium hydroxide (KOH), while sodium and magnesium sulfate are also added to evaluate their effectiveness on enhancing the activation process.
KOH is a strong base that is produced through the electrolysis of potassium chloride [56,57]. KOH is a major activator used in alkali-activated materials, and can be found in pellet, flake and powder forms [58]. The result of this study is found to be significant, and points to the potential application of sodium and magnesium sulfate to partially substitute commonly used activators for the production of greener geopolymer concretes. The following sections further elaborate on the experimental program of this study.

Precursors
In this study, kaolin and ceramic powder with a SiO 2 content of~70% and 54% and a specific gravity of 2.53 and 2.72, respectively, were used as the main precursors. The specific surface areas of the ceramic and kaolin powders were obtained as 2720 cm 2 /g and 2530 cm 2 /g, respectively. Table 1 presents the results of an X-ray fluorescence (XRF) test conducted on the kaolin and ceramic powder based on mass percentage. In addition, a particle size analysis was also conducted on the kaolin and ceramic powder, the results of which are presented in Figure 1a. Figure 1b presents the X-ray diffraction (XRD) of the raw minerals. For this figure, the XRD analysis of raw materials was conducted using Bruker Phaser instruments with CuKα radiation (λ = 1.542 A • ) at 40 kV. As shown in this figure, the major crystalline phases were quartz (SiO 2 ) (PDF#01-083-2466), kaolinite (Al 2 Si 2 O 5 (OH) 4 ) (PDF#01-072-2300) and anorthite (CaAl 2 Si 2 O 8 ) (PDF#01-078-0432) in kaolin, while quartz, kaolinite and calcite (CaCO 3 ) (PDF#01-072-1937) were observed in the ceramic powder, which was similar to that mentioned in Refs. [59][60][61][62]. conducted on the kaolin and ceramic powder based on mass percentage. In addition, a particle size analysis was also conducted on the kaolin and ceramic powder, the results of which are presented in Figure 1a. Figure 1b presents the X-ray diffraction (XRD) of the raw minerals. For this figure, the XRD analysis of raw materials was conducted using Bruker Phaser instruments with CuKα radiation (λ = 1.542 A°) at 40 kV. As shown in this figure, the major crystalline phases were quartz (SiO2) (PDF#01-083-2466), kaolinite (Al2Si2O5(OH)4) (PDF#01-072-2300) and anorthite (CaAl2Si2O8) (PDF#01-078-0432) in kaolin, while quartz, kaolinite and calcite (CaCO3) (PDF#01-072-1937) were observed in the ceramic powder, which was similar to that mentioned in Refs. [59][60][61][62].

Aggregate
A basalt aggregate with a specific gravity of 2.74 g/cm 3 and a maximum size of 4 mm was used in this study. Figure 2 presents the sieve analysis (based on ASTM C136) conducted on the fine aggregate.   3 and a maximum size of 4 mm was used in this study. Figure 2 presents the sieve analysis (based on ASTM C136) conducted on the fine aggregate.

Aggregate
A basalt aggregate with a specific gravity of 2.74 g/cm 3 and a maximum size of 4 mm was used in this study. Figure 2 presents the sieve analysis (based on ASTM C136) conducted on the fine aggregate.

Activator
To increase the alkalinity of the mixes, solid potassium hydroxide (KOH) with a molecular weight of 56.11 g/mol with a density of 2.04 g/cm 3 was used. The ratio of KOH to

Activator
To increase the alkalinity of the mixes, solid potassium hydroxide (KOH) with a molecular weight of 56.11 g/mol with a density of 2.04 g/cm 3 was used. The ratio of KOH to binder was kept at 20% throughout all the mixes. Table 2 provides further information on the properties of the KOH used in this study.

Mix Proportions
In this study, a total of 28 mixes was prepared with a constant liquid-to-binder ratio of 0.5. In general, the liquid-to-binder ratio is a common term used in geopolymer binders, and it refers to the content of the activator to precursor. In the mixes, the kaolin content was replaced with ceramic powder at 0, 10, 20 and 30 vol.%. In the same way, sodium (Na 2 SO 4 ) and magnesium sulfate (MgSO 4 ) were also added at 0, 3, 6 and 9 vol.%, substituting the kaolin content. The reason for the mentioned ratios was the authors' preliminary tests on the fresh properties of the mixes that led to choosing such specific ratios. Table 3 further shows the mix proportions used in this study. In this table, the mixes were labelled as follows: letters CP, NS and MS refer to ceramic powder, sodium and magnesium sulfate, respectively. The numbers following each letter indicate the content of each material used in the mix (% for sodium and magnesium sulfate and kg/m 3 for the ceramic powder). For instance, CP30-NS9 represents a mix containing 30% ceramic powder and 9 vol% sodium sulfate.

Specimen Preparation and Test Methods
In this study, geopolymer mortar samples were prepared by placing the freshly mixed materials in proper molds and then into an oven with a constant temperature of 115 • C for 24 h. Finally, the samples were removed from the oven and kept in ambient temperature until tested. Figure 3 presents the detailed mixing procedure used in this study.

Specimen Preparation and Test Methods
In this study, geopolymer mortar samples were prepared by placing the freshly mixed materials in proper molds and then into an oven with a constant temperature of 115 °C for 24 h. Finally, the samples were removed from the oven and kept in ambient temperature until tested. Figure 3 presents the detailed mixing procedure used in this study.  To analyze the mechanical properties of the mixes, after 28 days of curing, compressive and flexural strength tests were conducted by using 40 × 40 × 160 mm samples based on EN 1015-11 [63]. In that respect, three specimens were created for each flexural strength test. The compressive test was applied on 6 half samples obtained from the flexural test. In addition, to assess the physical properties of the produced geopolymer samples, dry bulk density, porosity and water absorption tests were conducted on specimens in accord-  To analyze the mechanical properties of the mixes, after 28 days of curing, compressive and flexural strength tests were conducted by using 40 × 40 × 160 mm samples based on EN 1015-11 [63]. In that respect, three specimens were created for each flexural strength test. The compressive test was applied on 6 half samples obtained from the flexural test. In addition, to assess the physical properties of the produced geopolymer samples, dry bulk density, porosity and water absorption tests were conducted on specimens in accordance to ASTM C642-13 [64]. The dry bulk density, porosity and water absorption of the produced mortars were calculated with Equations (1)-(3), respectively, where the mass of the ovendried sample was represented with A. The surface-dry sample mass after immersion and immersion-boiling is represented by B and C, respectively. D represents the mass of the sample in water. Additionally, the term ρ refers to the density of water.
In conjunction, ultrasonic pulse velocity (UPV) as a nondestructive testing (NDT) method was performed to control the level of uniformity of the samples. X-ray powder diffraction (XRD) was employed to characterize the crystalline phases of the raw materials and produced geopolymers. To evaluate the produced geopolymer chemical bonds and stability, a spectroscopic investigation was carried out using a Fourier transform infrared (FTIR) spectrometer (Shimadzu IRSpirit model) over a spectrum range of 4000-400 cm −1 . The microstructures and surface morphologies of the specimens were obtained with scanning electron microscopy (SEM, TESCAN MAIA3 XMUv). The elemental distribution of hydration products was analyzed with energy-dispersive X-ray spectroscopy (EDS). Figure 4a,b represents the 28-day compressive strength of the geopolymer mortars according to the Na 2 SO 4 or MgSO 4 contents (0%, 3%, 6% and 9%), with the replacement of 10%, 20% and 30% of kaolin with ceramic powder. At 0% Na 2 SO 4 and MgSO 4 content, the increase in ceramic powder content from 0% to 10 and 20 wt.% increased the compressive strength from 6.5 to 7.9 and 11.5 MPa, respectively. This finding could be explained by the fact that ceramic powder has a higher specific surface area than kaolin, creating a denser matrix composition. Moreover, Xu et al. [65] reported in a study on using natural aluminosilicate materials to synthesize geopolymers that the rate of Al separation from lowreactivity kaolin is mostly insufficient to create a gel formation. In this regard, it requires extra time for interactions among raw materials due to the weak reactivity of kaolinite. However, using low-reactivity kaolinite solely does not create a good synergy between the aluminosilicate sources, leading to the formation of a poor structure [65]. In parallel with this finding, the increase in the content of ceramic powder supports gel formation, as it increases the amount of kaolinite and other aluminosilicates simultaneously, effectively improving the compressive strength of the material. The results demonstrated that the substitution of 20 wt.% ceramic powder decreased the compressive strength by 5%. This reduction might be explained due to the use of ceramic powder above 20 wt%, lowering the polymerization reaction and not contributing to the strength values. The increased ceramic powder content reduced the amount of SiO 2 in the total binder, which caused a decrease in the compressive strength. tively improving the compressive strength of the material. The results demonstrated tha the substitution of 20 wt.% ceramic powder decreased the compressive strength by 5% This reduction might be explained due to the use of ceramic powder above 20 wt%, low ering the polymerization reaction and not contributing to the strength values. The in creased ceramic powder content reduced the amount of SiO2 in the total binder, whic caused a decrease in the compressive strength.  In the Na2SO4 introduced groups, the compressive strength rose gradually when creasing the replacement levels of kaolin with ceramic powder. Based on the obtained t results, it was clearly seen that the addition of Na2SO4 increased the compressive streng of the specimens regardless of the ceramic powder content, and the highest compress strength was obtained as 15.7 MPa in the sample with 9% Na2SO4 and 30% ceramic po der. These results may be similar to the results obtained in sodium-silicate-activated ash-based geopolymer containing various proportions of sodium sulfate in Ref.

Compressive and Flexural Strength
[5 Qing-feng et al. reported that the acceptable introduction of Na2SO4 significantly creased the strength of the geopolymers, as it highly promoted geopolymerization [5 With the usage of higher amounts of Na2SO4, the dissolution rate of aluminosilicates creased considerably, and the formation of N-A-S-H gel was accelerated [53]. Con quently, the occupation of Na2SO4 in geopolymer systems is advantageous for developi compressive strength values, which is parallel with the results of the current study. Equation 1, the reaction products of Na2SO4 and KOH were given. As a result of the re tion, the presence of additional NaOH was witnessed in the medium. It is known th K2SO4 can also be used as an activator [66]. K2SO4 and the additional NaOH released d ing the reaction was also effective in increasing the compressive strength.
Although the increase in the amount of MgSO4 (from 0% to 9%) for the mortars w a 0% and 10% content of ceramic powder remarkably increased the compressive streng In the Na 2 SO 4 introduced groups, the compressive strength rose gradually when increasing the replacement levels of kaolin with ceramic powder. Based on the obtained test results, it was clearly seen that the addition of Na 2 SO 4 increased the compressive strength of the specimens regardless of the ceramic powder content, and the highest compressive strength was obtained as 15.7 MPa in the sample with 9% Na 2 SO 4 and 30% ceramic powder. These results may be similar to the results obtained in sodium-silicate-activated fly ash-based geopolymer containing various proportions of sodium sulfate in Ref. [53]. Qing-feng et al. reported that the acceptable introduction of Na 2 SO 4 significantly increased the strength of the geopolymers, as it highly promoted geopolymerization [53]. With the usage of higher amounts of Na 2 SO 4 , the dissolution rate of aluminosilicates increased considerably, and the formation of N-A-S-H gel was accelerated [53]. Consequently, the occupation of Na 2 SO 4 in geopolymer systems is advantageous for developing compressive strength values, which is parallel with the results of the current study. In Equation 1, the reaction products of Na 2 SO 4 and KOH were given. As a result of the reaction, the presence of additional NaOH was witnessed in the medium. It is known that K 2 SO 4 can also be used as an activator [66]. K 2 SO 4 and the additional NaOH released during the reaction was also effective in increasing the compressive strength.
2 KOH + Na 2 SO 4 → K 2 SO 4 + 2 NaOH (4) Although the increase in the amount of MgSO 4 (from 0% to 9%) for the mortars with a 0% and 10% content of ceramic powder remarkably increased the compressive strength values by 88% and 48%, it gradually reduced the strength values of those with a 20% and 30% content of ceramic powder by approximately 25% and 42%, respectively. In samples with MgSO 4 , while the mortar with 0% ceramic powder and 9% MgSO 4 (CP0MS9) had the highest compressive strength value of 11.91 MPa, the lowest value was determined as 6.3 MPa in the mortar containing 30% ceramic powder with 9% MgSO 4 (CP30MS9). A study conducted by Kang et al. [67] using various contents (2, 4, 6, 8 and 10%) of MgSO 4 as the activator in slag-based geopolymer paste showed an explicit reduction of almost 50% in compressive strength with the increase in the amount of MgSO 4 from 6% to 10%. This finding was attributed to the lower activation effect of MgSO 4 , leading to the formation of a porous structure, significantly decreasing the compressive strength. Moreover, it was reported that lower pH environments (pH < 11.5) lowered the compressive strength values because of the inadequate activation of slags [52,67]. In parallel with the results of this study, the addition of higher amounts of MgSO 4 deteriorated the compressive strength, as it remarkably reduced the pH of the environment. In general, the compressive strength test results revealed that the addition of Na 2 SO 4 significantly improved the compressive strength, while the addition of MgSO 4 did not show a similar trend. In Equation (2), the reaction products of MgSO 4 and KOH were given. It could be seen that the additional Mg(OH) 2 and K 2 SO 4 were formed as a result of the reaction. Mg(OH) 2 is the aqueous form of MgO. It was stated that the MgO additive in geopolymer concrete can increase the compressive strength [68]. The type and amount of the activator are effective in increasing and decreasing the compressive strength [69,70]. Therefore, the compressive strength of the samples containing MgSO 4 did not increase significantly.
2 KOH + MgSO 4 → K 2 SO 4 + Mg(OH) 2 (5) The flexural strength of the geopolymer mortars was investigated at 28 days, and the effect of various contents of Na 2 SO 4 , MgSO 4 and ceramic powder is presented in Figure 5a,b. Similar to the overall trend witnessed in compressive strength, the addition of ceramic powder to the mixtures without salts remarkably increased the geopolymer mortars' flexural strengths. The introduction of ceramic powder from a 0 to 30% replacement level almost doubled the flexural strength values (1.23 to 2.40 MPa). The increase in the flexural strength might have been attributed to the ceramic powder's higher specific surface area, saturating the geopolymer matrix's pores and improving the interface zone that strengthened the bonding to maintain the flexural loading. Furthermore, the incorporation of Na 2 SO 4 into the geopolymer mixtures remarkably developed the flexural strength, owing to the contribution of Na 2 SO 4 to the geopolymerization reaction, regardless of the utilization of ceramic powder [53]. The increase in flexural strength was found to be 56 and 117% with an increasing Na 2 SO 4 content from 0 to 3 and 9%, respectively. However, the effect of the MgSO 4 increase on the flexural strength did not show a clear trend, as seen in the Na 2 SO 4 addition. The addition of MgSO 4 (0 to 9%) on the mixes incorporating 0% and 10% ceramic powder increases the flexural strength values by almost 95% and 70%, respectively. However, the MgSO 4 addition to the specimens, including ceramic powder of above 10 wt.%, adversely affected the flexural strength values. Similar to the compressive strength test results, this reduction was due to the fact that the addition of MgSO 4 decelerated the geopolymerization by lowering the ambient pH and preventing the formation of gels. 0% and 10% ceramic powder increases the flexural strength values by almost 95% 70%, respectively. However, the MgSO4 addition to the specimens, including ce powder of above 10 wt.%, adversely affected the flexural strength values. Similar t compressive strength test results, this reduction was due to the fact that the additi MgSO4 decelerated the geopolymerization by lowering the ambient pH and preve the formation of gels.  Figure 6a,b plot the dry unit weight of the geopolymer samples, ranging be 2.02 g/cm 3 and 2.13 g/cm 3 . In the specimens with 0% Na2SO4 and MgSO4, the incre the ceramic powder content from 0% to 10%, 20% and 30 wt.% enhanced the unit w from 2.03 to 2.05, 2.07 and 2.12 g/cm 3 , respectively. The ceramic powder 's surface a reaction was higher than kaolin's (2720 cm 2 /g and 2530 cm 2 /g, respectively), leading formation of a denser matrix and with a higher unit weight. In addition, the unit w of samples containing Na2SO4 varied from 2.05 to 2.13 g/cm 3 , while it ranged betwee and 2.07 g/cm 3 for specimens incorporating MgSO4. The highest unit weight was obt for the mixture incorporating the combined use of 30% ceramic powder and 9% N (CP30NS9) at 2.13 g/cm 3 . The geopolymer specimens with 0% MgSO4 revealed a gro trend in the unit weight value, with a rise in ceramic powder or Na2SO4 content. How this variation was less prominent in the series containing MgSO4. The lowest o weights was obtained for the mixture containing 30% ceramic powder and 9% M (CP30MS9) at 2.02 g/cm 3 .  Figure 6a,b plot the dry unit weight of the geopolymer samples, ranging between 2.02 g/cm 3 and 2.13 g/cm 3 . In the specimens with 0% Na 2 SO 4 and MgSO 4 , the increase in the ceramic powder content from 0% to 10%, 20% and 30 wt.% enhanced the unit weight from 2.03 to 2.05, 2.07 and 2.12 g/cm 3 , respectively. The ceramic powder 's surface area of reaction was higher than kaolin's (2720 cm 2 /g and 2530 cm 2 /g, respectively), leading to the formation of a denser matrix and with a higher unit weight. In addition, the unit weight of samples containing Na 2 SO 4 varied from 2.05 to 2.13 g/cm 3 , while it ranged between 2.02 and 2.07 g/cm 3 for specimens incorporating MgSO 4 . The highest unit weight was obtained for the mixture incorporating the combined use of 30% ceramic powder and 9% Na 2 SO 4 (CP30NS9) at 2.13 g/cm 3 . The geopolymer specimens with 0% MgSO 4 revealed a growing trend in the unit weight value, with a rise in ceramic powder or Na 2 SO 4 content. However, this variation was less prominent in the series containing MgSO 4 . The lowest of unit weights was obtained for the mixture containing 30% ceramic powder and 9% MgSO 4 (CP30MS9) at 2.02 g/cm 3 . and 2.07 g/cm 3 for specimens incorporating MgSO4. The highest unit weight was obt for the mixture incorporating the combined use of 30% ceramic powder and 9% N (CP30NS9) at 2.13 g/cm 3 . The geopolymer specimens with 0% MgSO4 revealed a gro trend in the unit weight value, with a rise in ceramic powder or Na2SO4 content. How this variation was less prominent in the series containing MgSO4. The lowest o weights was obtained for the mixture containing 30% ceramic powder and 9% M (CP30MS9) at 2.02 g/cm 3 .

Dry Unit Weight
Sustainability 2021, 13, x FOR PEER REVIEW 12 Figure 6. Dry bulk density (unit weight) of mixtures for mixes produced (a) with Na2SO4 a with MgSO4.

Porosity and Water Absorption
Test results revealed that the porosity of mortars without salts decreased wi increased ceramic powder content, which could be attributed to the higher specific su area of ceramic powder than kaolin (see Figure 7a,b). To illustrate, the increase i amount of ceramic powder from 0% to 30% reduced the porosity by 38%. In cases w Na2SO4 was introduced, the highest porosity of 12% was obtained in the series wi Na2SO4 and 0% ceramic powder (CP0NS3), while the lowest porosity was observed mortar containing 30% ceramic powder and 9% Na2SO4 (CP30NS9). The results c showed that the addition of Na2SO4 played a critical role in reducing the porous stru of the mortars. The highest porosity was found in the sample with 30% ceramic po and 9% MgSO4 (CP30MS9), while the lowest porosity was found in the specimen wi MgSO4 and 10% ceramic powder (CP10MS6) as 13.3% and 9.8%, respectively. Acco to the results obtained from previous studies, it was reported that aluminosilicate hy gels show less formation in environments where the pH is lower than 9.5 [71,72]. Fo reason, it is inevitable to obtain a high porosity and low compressive strength in pH environments. Therefore, the utilization of supplementary materials with a low (<9.5), such as MgSO4, causes the formation of a porous matrix, remarkably reducin mechanical performance [67].

Porosity and Water Absorption
Test results revealed that the porosity of mortars without salts decreased with an increased ceramic powder content, which could be attributed to the higher specific surface area of ceramic powder than kaolin (see Figure 7a,b). To illustrate, the increase in the amount of ceramic powder from 0% to 30% reduced the porosity by 38%. In cases where Na 2 SO 4 was introduced, the highest porosity of 12% was obtained in the series with 3% Na 2 SO 4 and 0% ceramic powder (CP0NS3), while the lowest porosity was observed in the mortar containing 30% ceramic powder and 9% Na 2 SO 4 (CP30NS9). The results clearly showed that the addition of Na 2 SO 4 played a critical role in reducing the porous structure of the mortars. The highest porosity was found in the sample with 30% ceramic powder and 9% MgSO 4 (CP30MS9), while the lowest porosity was found in the specimen with 6% MgSO 4 and 10% ceramic powder (CP10MS6) as 13.3% and 9.8%, respectively. According to the results obtained from previous studies, it was reported that aluminosilicate hydrate gels show less formation in environments where the pH is lower than 9.5 [71,72]. For this reason, it is inevitable to obtain a high porosity and low compressive strength in lower pH environments. Therefore, the utilization of supplementary materials with a lower pH (<9.5), such as MgSO 4, causes the formation of a porous matrix, remarkably reducing the mechanical performance [67]. The water absorption test results are displayed in Figure 8a,b. The pore-filling of the ceramic powder was more pronounced in specimens without Na2SO4 and M in which the water absorption reduced from 6.34% in CP0 to 3.86% in CP30. In add the mortars containing a higher kaolin content were found to demand a higher amou water because of the characteristics of the material, being an uncalcined clay and ca of absorbing considerable amounts of water [73]. While the water absorption of the tars incorporating Na2SO4 ranged between 3.16% and 6.34%, the values fluctuate tween 4.78% and 6.62% for those with MgSO4. The specimen with 9% Na2SO4 and 3 ceramic powder (CP30NS9) revealed the lowest level of water absorption at 3.2%, w clearly exhibited the effect of Na2SO4 on the water absorption characteristics of the polymers. It might have also been attributed to the reaction between Na2SO4 and th isting alkali activators, reducing the void ratio.
On the other hand, the addition of MgSO4 (0 to 9%) reduced the water absorpti mortars with 0 and 10% of ceramic powder, while it did not positively affect the w absorption of geopolymers containing 20% and 30% of ceramic powder. To demons the increase in the amount of MgSO4 (0 to 9%) decreased the water absorption of CP The water absorption test results are displayed in Figure 8a,b. The pore-filling effect of the ceramic powder was more pronounced in specimens without Na 2 SO 4 and MgSO 4 , in which the water absorption reduced from 6.34% in CP0 to 3.86% in CP30. In addition, the mortars containing a higher kaolin content were found to demand a higher amount of water because of the characteristics of the material, being an uncalcined clay and capable of absorbing considerable amounts of water [73]. While the water absorption of the mortars incorporating Na 2 SO 4 ranged between 3.16% and 6.34%, the values fluctuated between 4.78% and 6.62% for those with MgSO 4 . The specimen with 9% Na 2 SO 4 and 30% of ceramic powder (CP30NS9) revealed the lowest level of water absorption at 3.2%, which clearly exhibited the effect of Na 2 SO 4 on the water absorption characteristics of the geopolymers. It might have also been attributed to the reaction between Na 2 SO 4 and the existing alkali activators, reducing the void ratio.
CP10 by 22% and 12%; however, it increased the water absorption of CP20 and CP 10% and 70%, respectively. As mentioned earlier, MgSO4 created a porous matrix d the lower pH [67], increasing water absorption in series containing MgSO4 at higher r Consequently, Na2SO4 played a more influential role in water absorption characte of geopolymers than MgSO4.  Figure 9 plots the relations between the porosity, water absorption and dry weight of the geopolymer mortars. Rather linear relations between the porosity and absorption and between the porosity and dry unit weight were determined, with 0.9 0.85 coefficients of correlation, respectively. On the other hand, the addition of MgSO 4 (0 to 9%) reduced the water absorption of mortars with 0 and 10% of ceramic powder, while it did not positively affect the water absorption of geopolymers containing 20% and 30% of ceramic powder. To demonstrate, the increase in the amount of MgSO 4 (0 to 9%) decreased the water absorption of CP0 and CP10 by 22% and 12%; however, it increased the water absorption of CP20 and CP30 by 10% and 70%, respectively. As mentioned earlier, MgSO 4 created a porous matrix due to the lower pH [67], increasing water absorption in series containing MgSO 4 at higher ratios. Consequently, Na 2 SO 4 played a more influential role in water absorption characteristics of geopolymers than MgSO 4 . Figure 9 plots the relations between the porosity, water absorption and dry unit weight of the geopolymer mortars. Rather linear relations between the porosity and water absorption and between the porosity and dry unit weight were determined, with 0.99 and 0.85 coefficients of correlation, respectively. Sustainability 2021, 13, x FOR PEER REVIEW 15 of 27 Figure 9. Correlation of porosity vs. water absorption and dry unit weight.

Ultrasound Pulse Velocity
The ultrasonic pulse velocity (UPV) test was employed to evaluate the internal conditions of the concrete nondestructively [74]. The mean velocity of the ultrasonic pulse in the geopolymer mortars measured after 28 days of curing is plotted in Figure 10a,b. Based on the test results, in mixes without Na2SO4 and MgSO4, the increased level of ceramic powder was found to increase the UPV results of the mortars. This was believed to have been caused by higher compaction due to the pore-filling effect of the ceramic powder. In detail, the UPV value of CP0 was measured as 1697 m/s, while that of CP30 was measured as 2102 m/s. This showed a ~19% increase in the UPV value when 30% of ceramic powder was used to substitute the kaolin. Similarly, since the addition of Na2SO4 supported the geopolymerization reaction, its contribution to the compressive strength was also supported by the UPV results. To illustrate, an increase in the Na2SO4 content from 0 to 9% enhanced the UPV value of the specimen without ceramic powder from 1696 to 1992 m/s. The improved UPV findings could be attributed to the formation of a denser structure in the geopolymer mortar matrix [75].

Ultrasound Pulse Velocity
The ultrasonic pulse velocity (UPV) test was employed to evaluate the internal conditions of the concrete nondestructively [74]. The mean velocity of the ultrasonic pulse in the geopolymer mortars measured after 28 days of curing is plotted in Figure 10a,b. Based on the test results, in mixes without Na 2 SO 4 and MgSO 4 , the increased level of ceramic powder was found to increase the UPV results of the mortars. This was believed to have been caused by higher compaction due to the pore-filling effect of the ceramic powder. In detail, the UPV value of CP0 was measured as 1697 m/s, while that of CP30 was measured as 2102 m/s. This showed a~19% increase in the UPV value when 30% of ceramic powder was used to substitute the kaolin. Similarly, since the addition of Na 2 SO 4 supported the geopolymerization reaction, its contribution to the compressive strength was also supported by the UPV results. To illustrate, an increase in the Na 2 SO 4 content from 0 to 9% enhanced the UPV value of the specimen without ceramic powder from 1696 to 1992 m/s. The improved UPV findings could be attributed to the formation of a denser structure in the geopolymer mortar matrix [75].
Although the addition of MgSO 4 (0 to 9%) in the specimens with 0% and 10% ceramic powder slightly increased the UPV values by 14% and 9%, it reduced the UPV values in the samples with 20% and 30% ceramic powder by almost 15% and 30%, respectively. This reduction could be explained by the fact that a higher ceramic powder content did not contribute to geopolymerization, and the excessive utilization of MgSO 4 reduced the mix's pH, resulting in a lower reaction. The trend was similar to the compressive strength test results reported in Section 3.1.
A linear regression approach was implemented to establish the correlation of UPV and the unit weight of geopolymer mortars, and the results were plotted in Figure 11. As can be seen, a linear relationship was witnessed among the unit weight and UPV values with reasonable confidence for the correlation factor (R 2 ) of 0.89. The influence of supplementing ceramic powder and/or Na 2 SO 4 in geopolymers was evident on the velocity of the ultrasonic pulse and its respective relationship with the unit weight values. This showed an overall improvement, which may be owed to the formation of a denser matrix when ceramic powder or Na 2 SO 4 were added. Based on the figure, the higher reactivity of the ceramic powder and Na 2 SO 4 accelerated geopolymerization and led to a more consolidated and robust composition. Although the addition of MgSO4 (0 to 9%) in the specimens with 0% and 10% ceramic powder slightly increased the UPV values by 14% and 9%, it reduced the UPV values in the samples with 20% and 30% ceramic powder by almost 15% and 30%, respectively. This reduction could be explained by the fact that a higher ceramic powder content did not contribute to geopolymerization, and the excessive utilization of MgSO4 reduced the mix's pH, resulting in a lower reaction. The trend was similar to the compressive strength test results reported in Section 3.1.
A linear regression approach was implemented to establish the correlation of UPV and the unit weight of geopolymer mortars, and the results were plotted in Figure 11. As can be seen, a linear relationship was witnessed among the unit weight and UPV values with reasonable confidence for the correlation factor (R 2 ) of 0.89. The influence of supplementing ceramic powder and/or Na2SO4 in geopolymers was evident on the velocity of the ultrasonic pulse and its respective relationship with the unit weight values. This showed an overall improvement, which may be owed to the formation of a denser matrix when ceramic powder or Na2SO4 were added. Based on the figure, the higher reactivity of the ceramic powder and Na2SO4 accelerated geopolymerization and led to a more consolidated and robust composition. Figure 11. Correlation of UPV and unit weight.

Microstructural Analysis
3.5.1. X-ray Diffraction (XRD) Figure 12 presents the XRD patterns of the selected mixtures characterized at 28 days. The XRD patterns of CP0 and CP30 revealed the presence of quartz and kaolinite, earlier recognized as chief minerals in raw kaolin and ceramic powder. The XRD spectrum of  3.5. Microstructural Analysis 3.5.1. X-ray Diffraction (XRD) Figure 12 presents the XRD patterns of the selected mixtures characterized at 28 days. The XRD patterns of CP0 and CP30 revealed the presence of quartz and kaolinite, earlier recognized as chief minerals in raw kaolin and ceramic powder. The XRD spectrum of CP0 showed the intense phase of quartz, with reasonably intense peaks of kaolinite and muscovite, roughly similar to those witnessed in raw kaolin (see Figure 1b). However, the patterns of CP30 exhibited a decrease in the intensity of quartz, especially at approximately 26.5 • (2θ). This finding could be explained by the increase in the amount of ceramic powder; significant parts of the crystalline structures were spent during the geopolymerization, producing an amorphic or semicrystalline structure and developing a compressive strength of composition [76,77]. Furthermore, various amounts of anorthite (CaAl 2 Si 2 O 8 ), which is a calcium-based mineral in the feldspar group derived only from a ceramic powder precursor, phases resided in CP30. The XRD spectra of the geopolymers incorporating Na2SO4 (CP30NS9) and MgSO4 (CP30MS9) were also presented in Figure 12. Based on CP30NS9, the presence of lowconcentration peaks of kaolinite, moderate-intensity peaks of anorthite with the intense phase of quartz could be seen. The intense peak of anorthite phases exhibited a slight decrease in CP30NS9 than CP30. This indicated that the addition of Na2SO4 could dissolve part of the crystalline phase and convert it into a semicrystalline geopolymer gel. In line with the compressive strength results (see Figure 4), the XRD patterns of CP30NS9 presented the lowest intensity of quartz, followed by CP30 and CP0, respectively. For CP9NS9, Na2SO4 did not lead to the formation of a different crystalline phase in the geopolymers. An intense peak of sodium sulfate was not obtained on the diffractogram, showing that Na2SO4 reacted almost entirely in the geopolymer system. The study conducted by Qing-feng determined that there was no sodium sulfate peak in the geopolymer with a low salt content [53]. On the other hand, the introduction of MgSO4 into the geopolymers resulted in boosted peaks of quartz compared to other mixes, indicating the lower utilization of these crystalline forms throughout the reaction processes of CP30MS9, causing an important reduction in the compressive strength of the geopolymers. With the addition of MgSO4 to the geopolymers, the anorthite peaks at approximately 23.5°, 27.6° and 30.7° 2θ were replaced with diopside (CaMgSi2O6). N(C)-A-S-H gel (geopolymer gel) was observed at approximately 37° and 41° 2θ in the XRD patterns of all produced samples with different intensities. The increase in this gel resulted in an increase in strength values [78][79][80][81]. The XRD spectra of the geopolymers incorporating Na 2 SO 4 (CP30NS9) and MgSO 4 (CP30MS9) were also presented in Figure 12. Based on CP30NS9, the presence of lowconcentration peaks of kaolinite, moderate-intensity peaks of anorthite with the intense phase of quartz could be seen. The intense peak of anorthite phases exhibited a slight decrease in CP30NS9 than CP30. This indicated that the addition of Na 2 SO 4 could dissolve part of the crystalline phase and convert it into a semicrystalline geopolymer gel. In line with the compressive strength results (see Figure 4), the XRD patterns of CP30NS9 presented the lowest intensity of quartz, followed by CP30 and CP0, respectively. For CP9NS9, Na 2 SO 4 did not lead to the formation of a different crystalline phase in the geopolymers. An intense peak of sodium sulfate was not obtained on the diffractogram, showing that Na 2 SO 4 reacted almost entirely in the geopolymer system. The study conducted by Qingfeng determined that there was no sodium sulfate peak in the geopolymer with a low salt content [53]. On the other hand, the introduction of MgSO 4 into the geopolymers resulted in boosted peaks of quartz compared to other mixes, indicating the lower utilization of these crystalline forms throughout the reaction processes of CP30MS9, causing an important reduction in the compressive strength of the geopolymers. With the addition of MgSO 4 to the geopolymers, the anorthite peaks at approximately 23.5 • , 27.6 • and 30.7 • 2θ were replaced with diopside (CaMgSi 2 O 6 ). N(C)-A-S-H gel (geopolymer gel) was observed at approximately 37 • and 41 • 2θ in the XRD patterns of all produced samples with different intensities. The increase in this gel resulted in an increase in strength values [78][79][80][81].

FTIR
The result of the FTIR test is presented in Figure 13. It can be seen in Figure 13a that, for most of the mixes, two broad absorbance bands at approximately 450-1000 and 3600-3700 cm −1 were observable. According to Ref. [82], the asymmetric stretching from 950 to 1200 was the result of the Si-O-Si and Al-O-Si vibrational bands. Ref. [30] noted that the bands at approximately 3600 corresponded to the O-H stretching that reflected the presence of structural water. It can be seen from Figure 13a that the inclusion of 30% of ceramic powder slightly reduced the transmittance of the two bands at 3688 and 1001. Based on Ref. [30], this small shift to a lower frequency was associated with the penetration of Al 4+ (tetrahedral aluminum atoms) into the reacting binder system. bands at approximately 3600 corresponded to the O-H stretching that reflected the presence of structural water. It can be seen from Figure 13a that the inclusion of 30% of ceramic powder slightly reduced the transmittance of the two bands at 3688 and 1001. Based on Ref. [30], this small shift to a lower frequency was associated with the penetration of Al 4+ (tetrahedral aluminum atoms) into the reacting binder system. Further, in Figure 13b, it can be seen that the inclusion of 9% magnesium sulfate (CP0MS3 versus CP0MS9) slightly reduced the bands at ~1000 cm -1 . In Figure 13c, it can be seen that variations in mixes of 9% magnesium versus sodium sulfate did not change the FTIR bands considerably. These results agreed with those reported by Bakharev in [83], who exposed a coal fly ash (F)-based geopolymer to a sodium and magnesium sulfate solution and noted insignificant changes to the FTIR transmittance bands. Further, Figure  13d presents the effect of the magnesium and sodium sulfate content on mixes with 0% of ceramic powder. Based on this figure, the transmittance of mixes with 9% of magnesium sulfate appeared to experience lower bands compared to their companion mixes with 3% of magnesium sulfate. The lower intensity changes of specimens that were exposed to sodium sulfate versus those of magnesium sulfate were confirmed by Ref. [84].

SEM and EDS
Figure 14a-d represent SEM micrographs with magnifications of 20 µm and 5 µm of fractured surfaces of specimens CP0, CP30, CP30NS9 and CP30MS9, following compression tests after 28 days of curing, respectively. The SEM micrograph of specimen CP0 demonstrated a highly heterogeneous microstructure with a significant portion of unreacted kaolin granules, revealing the lower dissolution of the kaolin precursor powder in this mix. On the other hand, a relatively more cohesive and denser geopolymer matrix could be seen in CP30 compared to CP0. This indicated that the increase in the concentration of ceramic powder from 0% to 30% caused a higher degree of dissolution of aluminosilicates and a powerful formation of geopolymer gels in CP30. By comparing Figure  14a,b, it can be seen that the sample incorporating 30% of ceramic powder had a lower porosity and a more consolidated microstructure than CP0. This illustrated the reason for the CP30 sample having a higher compressive strength than CP0. Additionally, the number of cracks and their respective widths were observed to be larger in CP0 as compared to CP30. Further, in Figure 13b, it can be seen that the inclusion of 9% magnesium sulfate (CP0MS3 versus CP0MS9) slightly reduced the bands at~1000 cm -1 . In Figure 13c, it can be seen that variations in mixes of 9% magnesium versus sodium sulfate did not change the FTIR bands considerably. These results agreed with those reported by Bakharev in [83], who exposed a coal fly ash (F)-based geopolymer to a sodium and magnesium sulfate solution and noted insignificant changes to the FTIR transmittance bands. Further, Figure 13d presents the effect of the magnesium and sodium sulfate content on mixes with 0% of ceramic powder. Based on this figure, the transmittance of mixes with 9% of magnesium sulfate appeared to experience lower bands compared to their companion mixes with 3% of magnesium sulfate. The lower intensity changes of specimens that were exposed to sodium sulfate versus those of magnesium sulfate were confirmed by Ref. [84]. Figure 14a-d represent SEM micrographs with magnifications of 20 µm and 5 µm of fractured surfaces of specimens CP0, CP30, CP30NS9 and CP30MS9, following compression tests after 28 days of curing, respectively. The SEM micrograph of specimen CP0 demonstrated a highly heterogeneous microstructure with a significant portion of unreacted kaolin granules, revealing the lower dissolution of the kaolin precursor powder in this mix. On the other hand, a relatively more cohesive and denser geopolymer matrix could be seen in CP30 compared to CP0. This indicated that the increase in the concentration of ceramic powder from 0% to 30% caused a higher degree of dissolution of aluminosilicates and a powerful formation of geopolymer gels in CP30. By comparing Figure 14a,b, it can be seen that the sample incorporating 30% of ceramic powder had a lower porosity and a more consolidated microstructure than CP0. This illustrated the reason for the CP30 sample having a higher compressive strength than CP0. Additionally, the number of cracks and their respective widths were observed to be larger in CP0 as compared to CP30.

SEM and EDS
In the case of the geopolymer specimens incorporating 9% Na 2 SO 4 and 30% ceramic powder, as outlined in (Figure 14c), a consolidated microstructure was acquired with rare partially reacted particles. The images exhibited that the introduction of Na 2 SO 4 enhanced the uniformity degree of particles, allowing for their uniform dispersal within the matrix. Indeed, this reinforcement in the microstructure could be associated with an acceleration in the geopolymerization with the introduction of Na 2 SO 4 and the formation of more aluminosilicate gels. The results were in line with the findings of the study conducted by Qing-feng et al. [53], that the salt content might benefit the dissolution of aluminosilicate raw precursors and enhance the degree of geopolymerization [53]. On the other hand, it could be observed that CP30MS9 (Figure 14d) displayed a higher porosity and lower densification. This was because the lower pH of MgSO 4 decreased the activation effect, leading to a high porosity and low compressive strength [67]. By comparing Figure 14c,d, it could be seen that the geopolymer with Na 2 SO 4 had a more compact morphology than the mortar with MgSO 4 . Moreover, the inclusion of 9% of sodium sulfate produced certain crystals that bridged the gaps in the microstructures of the produced geopolymers. This observation was consistent with the compressive strength test results (Figure 5), and could explain why CP0NS9 exhibited the maximum compressive strength recorded for all tested mortars.  Table 4 represents the EDS analysis of the areas in Figure 14 marked from 1 to 8. The EDS table gives the dominant peaks for Al and Si and various minor peaks for Fe, Na, Mg, S, K and Ca due to precursors, the activator, aggregate and some supplements. To illustrate, in the first area of specimen CP0, the formation was thought to be a basalt aggregate due to the reasonable amount of Ca and Si peaks. However, the second area showed the presence of an acceptable amount of K, which was probably detected from the geopolymer gel. Additionally, the spectra of three and four also represented geopolymer gel formation, which was confirmed by the presence of Al, Si and K as the major elements. With the addition of Na 2 SO 4 , some minor peaks for Na were detected in the spectrum, namely, five and seven, in specimen CP30NS9. The presence of Ca and Si represented the basalt aggregate in spectrum six. The improvement in Na 2 SO 4 to the formation of geopolymer gel was also confirmed with the XRD analysis, which showed the reduced peaks for crystalline phases compared to the specimen without sodium sulfate. The EDS analysis indicated that spectrum eight represented the geopolymer gel, because of the higher contents of the elements Al, Si and K. It is noteworthy to mention that it was challenging to evaluate the geopolymerization rate for the specimens because of the small scan area.

Conclusions
In this research study, the effect of sodium and magnesium sulfate on the mechanical and microstructural properties of kaolin and ceramic powder-based geopolymer concrete was evaluated. To that end, a total of 28 mixes was prepared with the kaolin content being replaced with ceramic powder at 0, 10, 20 and 30 vol.%. In the same way, sodium (Na 2 SO 4 ) and magnesium sulfate (MgSO 4 ) were also added at 0, 3, 6 and 9 vol.% by substituting the kaolin content. Based on the results of this study, the following conclusions could be drawn:

•
Replacing kaolin with ceramic powder increased the compressive and flexural strengths of the geopolymer mortars. This increase was attributed to the ceramic powder's higher specific surface area compared to kaolin, which created a denser matrix formation when replaced with kaolin. At a given ceramic powder content, the geopolymers containing MgSO 4 exhibited lower strength values than those incorporating Na 2 SO 4 . This was due to the lower pH value of MgSO 4 than Na 2 SO 4 , causing a porous matrix and leading to lower strengths.

•
The porosity and water absorption of the geopolymers decreased when ceramic powder was replaced with kaolin. In addition, mortars containing MgSO 4 exhibited higher porosity values than those with Na 2 SO 4 , showing sodium sulfate's contribution to geopolymerization, creating a denser microstructure. Additionally, a strong correlation is found to exist between the water absorption and porosity values in all mixes. • UPV values were found to exhibit a very conforming trend with the unit weight, and showed a notable improvement with the increase in the ceramic powder content. Consistent with the compressive strength results, the increase in Na 2 SO 4 was found to increase the UPV values, as it contributed to the formation of gel and reduced the void ratio. However, MgSO 4 showed a decreasing trend in UPV values, indicating of a more porous structure. • Through the XRD analyses, intense peaks for quartz, kaolinite, muscovite, anorthite and diopside were identified. A lowest quartz peak was observed in the geopolymer mortar incorporating 30% of ceramic powder with 9% Na 2 SO 4 , which showed that a combination of Na 2 SO 4 and ceramic powder could dissolve considerable parts of the crystalline phases and convert them into a semi-crystalline geopolymer gel, thereby contributing to the mechanical and microstructural properties of the geopolymers.

•
The result of the FTIR in this study showed that the inclusion of ceramic powder slightly reduced the transmittance values, likely due to the penetration of Al 4+ (tetrahedral aluminum atoms) into the reacting binder system. In the same way, the inclusion of sodium versus magnesium sulfate did not considerably alter the stretching vibrational bands, which could be due to the similar effect of the two compounds on the microstructural development of the geopolymer concrete. • Based on the SEM images, the geopolymers without ceramic powder showed a heterogeneous microstructure with a considerable number of unreacted particles; however, a more coherent geopolymer matrix could be seen with the addition of the ceramic powder. Further to this, it is believed that sodium sulfate could bridge the crack gaps and accelerate the geopolymerization. This could lead to the formation of more aluminosilicate gel.
In the end, the results of this study were found to be significant and pointed to the suitability of utilizing ceramic powder and kaolin as precursors, activated, in part, using sodium and magnesium sulfate for the cleaner production of geopolymer concrete. Nonetheless, future studies in this area could provide further information on the thermodurability properties of geopolymer concretes produced with sodium and magnesium sulfate.