Strength development and microstructure properties of slag activated with alkaline earth metal ions: a review study

Abstract This article reviews the microstructure properties and strength development of slag activated with calcium oxide, calcium hydroxide, magnesium oxide, barium hydroxide, carbide slag, and calcined dolomite. X-ray diffraction, scanning electron microscopy, energy-dispersive X-ray spectroscopy, and thermogravimetric/differential thermal analysis were used to evaluate the microstructure characteristics of activated slag. Furthermore, compressive strength test results were studied to evaluate the development of the strength of slag activated with alkaline earth metal ions. The results showed that slag samples activated with carbide slag and barium hydroxide, respectively, produced the lowest and highest compressive strengths. Moreover, among the activators of alkaline earth metal ions studied thus far, barium hydroxide produced the densest microstructure, lowest calcium-to-silicon ratio, and most hydration products. Slag activated with carbide slag produced the lowest calcium–silicate–hydrate (C–S–H) content compared with other activators of alkaline earth metal ions.


Introduction
The use of ordinary Portland cement requires high energy consumption and is a factor contributing to global warming (Anand et al., 2006;Juenger et al., 2011); thus, most researchers have sought to minimize the use of this type of cement as much as possible.It was reported that the use of blast furnace slag to produce alkaline activated slag (AAS) offers large-scale production potential without requiring ordinary Portland cement (Provis & Deventer, 2014).According to extensive research, AAS has advantages like improved chemical attack resistance (Bakharev et al., 2002(Bakharev et al., , 2003;;Roy et al., 2000), increased mechanical strength (Bakharev et al., 1999;Collins & Sanjayan, 1999;Fern� andez-Jim� enez et al., 1999), lower heat required for hydration (Shi et al., 2006;Wang et al., 1995), and a stronger cement matrix formation (Brough & Atkinson, 2000;Shi & Xie, 1998).
The three most popular alkaline activators for AAS are sodium carbonate (Rashad, 2013), potassium hydroxide, and sodium silicate.The highest compressive strength and densest microstructure can be created by activating slag with sodium hydroxide and sodium silicate activators.However, these activators' production processes are highly energy-intensive, rendering them uneconomical (Rashad, 2013;Shi & Xie, 1998).
Issues such as the highly corrosive nature of alkaline solutions, fast setting time, soluble alkaline viscosity, and heat released from the dissolution of alkaline compounds, especially alkaline hydroxide, should be considered when they are used for slag activation (Gong & Yang, 2000;Yang et al., 2008Yang et al., , 2012)).These problems and disadvantages have prompted a search for other activators of alkaline earth metal ions, whether for use as an additive or main activator, that offer greater durability and lower cost (Kim et al., 2013).
Li and Yi (2020) investigated the strength development and microstructure properties of slag activated by carbide slag (CS), an industrial by-product of acetylene production, as a potential substitute for hydrated lime (HL) for slag activation.They prepared samples of slag activated at different CS and HL contents and then tested their compressive strengths.They also performed thermogravimetric (TG) analysis, X-ray diffraction (XRD), Fourier transform infrared spectroscopy (FTIR), scanning electron microscopy (SEM), and to examine the hydration properties of CS-activated slag.The results showed that the AAS produced the same compressive strength as that activated with hydrated lime at the same CS contents over curing times of 28 and 56 days.These results mean that CS could be used to replace HL with active ground-granulated blast-furnace slag (GGBS).The calcium hydroxide in CS further accelerated the hydration of GGBS and the hydration products; however, a very high CS content was found to reduce the GGBS content and increase the formation of crystals.Jin et al. (2015) examined reactive magnesium oxide as a potentially more viable and greener alternative.Hydration kinetics of MgO-GGBS by TG analysis, XRD, and SEM showed that reactive magnesium oxide (MgO) was more effective than hydrated lime for activation.It was suggested that reactive MgO has the potential to act as an efficient and economical activator for GGBS.Jeong et al. (2017) studied barium hydroxide as a major activator for GGBS.Except after 3 days of curing, barium hydroxide-activated slag had a significantly higher compressive strength than calcium hydroxide-activated GGBS, mainly due to the increased production of hydration products, which resulted in a significant decrease in the number of pores.Although the main product of the hydration reactions in both mixtures is hydrated CS, the calcium-to-silicon ratio differed in the two mixtures.
AAS binders are produced by sodium hydroxide (NaOH) and Na 2 O.nSiO 2 activators, which are expensive, set quickly, and have high pH-related toxicity.As a result, researchers have searched for more economical and less dangerous activators of alkaline earth metal ions for mass production.Considering the ideal characteristics of alkali-activated slag (including high-temperature resistance, high compressive strength, and excellent resistance to chloride and sulfate ions), one solution for this is the complete replacement of cement with slag.
The effects of various commercial and by-product activator types on ground-granulated blast furnace slag-based concrete were thoroughly examined previously (Cheah et al., 2021).The GGBS activated with by-product sodium silicate produced nearly identical mechanical properties to GGBS activated with industrial sodium silicate (Cheah et al., 2021).However, no article thus far has comprehensively reviewed the effects of activators of alkaline earth metal ions on the microstructure characteristics and strength development of slag activated with activators of alkaline earth metal ions.The present review assessed the effects of magnesium oxide (MgO), calcium oxide (CaO)/calcium hydroxide Ca(OH) 2 , barium hydroxide (Ba(OH) 2 ), calcined dolomite (CaMg(CO 3 ) 2 ), and carbide slag on the microstructure characteristics and strength development of the concrete.
To assess the microstructure of concrete thus prepared, the results of a variety of tests including thermogravimetry/differential thermal analysis (TG/DTA), scanning electron microscopy/energy-dispersive X-ray spectroscopy (SEM/EDS), X-ray diffraction (XRD), Fourier transform infrared spectroscopy/nuclear magnetic resonance spectroscopy (FTIR/NMR), accompanied with compressive strength test were studied.Furthermore, to promote future research, the present paper reviewed published articles on the microstructure characteristics and strength development of slag activated with activators of alkaline earth metal ions.Since research on this topic has been limited to certain circumstances, the investigation and comparisons of the present review can help researchers to access information about limitations on this topic for future use.

Information about the composition of activators and final optimum mixtures
Figure 1 provides the XRD and TGA results for calcined dolomite at both 800 � C (D800) and 1000 � C (D1000) (Gu et al., 2016).The figure demonstrates that D800 was made up primarily of undecomposed dolomite, MgO, CaO 3 , and CaCO 3 .It was very challenging to find a specific percentage of each composition due to the unknown percentage of undecomposed dolomite.Due to the lack of dolomite in D1000, it was estimated that it contained MgO, CaO 3 , and CaCO 3 with specific proportions of 35.8, 54.6, and 9.6%, respectively, based on the TG curve (Figure 1).
Figure 2a displays the TGA/DTG results of GGBS, carbide slag (CS), and hydrated lime (Li & Yi, 2020).The TGA test showed that GGBS loses almost no mass, but the DTG curves for CS and hydrated lime (HL) showed two significant peaks.Calcium-based constituents were the primary constituents of both CS and HL, and the relatively high loss on ignition (LOI) of CS and hydrated lime (HL) was caused by the decomposition of Ca(OH) 2 (380-620 � C) and CaCO 3 (620-900 � C).As shown in Figure 2, compared to HL, CS has less Ca(OH) 2 and more CaCO 3 .The carbonation of CS in the air during outside stacking is most likely what causes the formation of CaCO 3 in CS.
XRD evaluated the mineralogy of GGBS, HL, and CS, as depicted in Figure 2b (Li & Yi, 2020).Portlandite, calcite, dolomite, and MgO, were the primary components of HL and CS.The CS contained intense calcite peaks and a small amount of carbon and calcium carbide.Peaks of calcium aluminates containing silicates (gehlenite, C 2 AS), akermanite, merwinite, anhydrite, and calcite were found for GGBS.
The ideal activators of alkaline earth metal ions percentage for slag activation is displayed in Table 1.In this table, A/B indicates the amount of activator in the binder, and W/B indicates the content of the water used in the binder.This table shows that slag activation has increased as a result of the use of    10% calcium oxide, calcium hydroxide, and barium chloride.Additionally, 20, 5, and 15%, respectively, were the ideal quantities of calcined dolomite, carbide slag, and magnesium oxide.

Ca(OH) 2 /CaO-activated GGBS
To create the highest compressive strength in slag activated with calcium oxide and calcium hydroxide, researchers seek to determine their optimal percentage as activators.It has been shown that the slag samples activated with calcium oxide had higher compressive strength than the samples activated with calcium hydroxide when their optimal percentages were used (Kim et al., 2013;Neville, 1995;Shi et al., 2006;Yang et al., 2010Yang et al., , 2012aYang et al., , 2012b)).Activation of slag with 6.25% calcium oxide (of the total weight of the binder) showed a greater compressive strength over time compared to activation with the same amount of calcium hydroxide.At 56 days of curing, the compressive strength of the calcium oxide-activated samples was about 32 MPa higher than that of the calcium hydroxide-activated samples (Figure 3a) (Kim et al., 2013).It could be claimed that calcium oxide was a more effective activator than calcium hydroxide (Yang et al., 2012b).
The use of auxiliary activators for better activation of slag activated with calcium oxide and calcium hydroxide has been considered to get higher compressive strength.For example, the use of 1% sodium sulfate as an auxiliary activator along with calcium hydroxide for the activation of slag was found to increase the compressive strength by about 30% (Figure 3b).Also, the use of 2% sodium carbonate as an auxiliary activator for calcium hydroxide increased the compressive strength by 20% (Yang et al., 2012b).The reason for improving the compressive strength of slag samples activated with CaO and Ca(OH) 2 accompanied by auxiliary activators (sodium sulfate and sodium carbonate) may be attributed to creating a denser microstructure and more hydration products (i.e., C-S-H).
Research has shown that the curing method used for slag samples activated by calcium hydroxide and auxiliary activators improved the compressive strength.Samples that had been stored in water exhibited higher compressive strengths than samples stored in the air (Yang et al., 2012) because of the creation of more hydration products in samples stored in water.In fact, the water treatment significantly increased the consumption of activators in the calcium hydroxide-activated slag binders over a long curing time (Shi et al., 2006;Yang et al., 2010), which increased the amount of the hydration products.In addition, an increase in the volume of fine aggregates, which increased internal microcracking and the voids between the paste and aggregates, decreased the compressive strength of the samples (Neville, 1995).

MgO-activated GGBS
The use of magnesium oxide as a slag activator was considered by several researchers (Bernal et al., 2014;Burciaga-D� ıaz & Betancourt-Castillo, 2018;Ke et al., 2017;Park et al., 2018Park et al., , 2020;;Yoon et al., 2018).Some experts have been interested in the topic because the ideal MgO content used in all binders may change over time.It has been reported that, over a short curing time of 7 days, the highest compressive strength was recorded for slag samples activated with 10% MgO.Over a long curing time of 28 days, however, the highest compressive strength results were obtained for samples activated with 20% MgO (Figure 4) (Park et al., 2020).This depended on the degree of hydration of the AAS.As the degree of hydration increased, the formation of C-A-S-H increased (Yi et al., 2015).However, other studies have stated that, with an increase in the water-to-binder ratio, the optimal amount of MgO activator at 28 days of curing, was 10% but, with an increase in activator from 10% to 20%, the compressive strength decreased (Yi et al., 2015).This occurred because the water-to-binder ratio accelerated the rapid hydration of the AAS.Therefore, the slag content, which affected the 28-day compressive strength, was more important at a higher water-to-binder ratio.It should be noted that the optimal percentage of MgO for slag activation depends on its reactivity.For high-reactivity MgO, the optimal percentage for activation is 20%.For low-reactivity MgO, the optimal percentage for slag activation is 10%.As the percentage of MgO increases, the compressive strength decreases because of the decrease in the percentage of activated slag and the formation of more cracks (Jin et al., 2015).

Calcined dolomite-activated GGBS
The use of calcined dolomite as a slag activator is a good alternative to calcium oxide and MgO activators (Djayaprabha et al., 2017;Gu et al., 2014a,b;2016).However, the calcination temperature of dolomite strongly influenced the compressive strength (Gu et al., 2014a,b).Dolomite calcinated at 1000 � C produced a greater compressive strength after slag activation than that calcinated at 800 � C (Figure 5) (Gu et al., 2014a,b;Figure 5a).In addition, an increase in the percentage of calcined dolomite used to activate the slag was found to increase its compressive strength.This was more evident for calcined dolomite calcinated at 800 � C. The use of 15% (1000 � C) calcined dolomite to activate slag after 60 and 150 days of curing showed increases in the compressive strength of 2.9% and 7.4%, respectively, compared to the use of 10% (1000 � C) calcined dolomite (Figure 5b and c).However, similar amounts of calcined dolomite at 800 � C, increase compressive strength by 23.6% and 13.5%, respectively (Gu et al., 2014a,b).This is likely caused by the increased formation of C-A-S-H gel in calcined dolomite at 1000 � C slag, which also increased the density of the microstructure of samples.
Research has shown that by increasing the percentage of calcined dolomite up to 20% as a slag activator, the compressive strength has decreased.Meanwhile, increasing the percentage of calcined dolomite from 20% to 30%, and then to 40%, has reduced the compressive strength (Figure 5).This issue can be attributed to the weakening of the microstructure and the reduction of hydration products formed by increasing the percentage of calcined dolomite from 20%.
The compressive strength of slag activated by calcined dolomite had a lower compressive strength over a long curing time of 90 days than slag samples activated by MgO and calcium oxide.However, at short curing times of 7 days, the compressive strength of the calcined dolomite-activated samples was higher than that of MgO-activated samples and less than that of calcium oxide-activated samples.The formation of hydrotalcite-like phases in the microstructure likely strengthened the mechanical properties of MgO-activated samples over long curing times.

Barium hydroxide-activated GGBS
Barium hydroxide has a 2678 times higher solubility product constant than Ca(OH) 2 .When both are dissolved in water, Ba(OH) 2 may have significantly higher amounts of OH -than Ca(OH) 2 .This means that it is likely to have more benefits for slag activation because it has a higher pH than other alkaline activators (Lambert & Clever, 2013;Song et al., 2000;Song & Jennings, 1999;Yang et al., 2010).As a result, barium hydroxide-activated slag samples are likely to have greater compressive strengths than those activated with other activators.(Yi et al., 2015); with different MgO percentages (Park et al., 2020).
Studies have shown that, at the same percentage of barium hydroxide and calcium hydroxide as activators, the compressive strength in samples activated with barium hydroxide was about 38% higher than for samples activated by calcium hydroxide (Figure 6) (Jeong et al., 2017(Jeong et al., , 2019;;Shi & Day, 1995).In fact, the use of a barium hydroxide activator produced a denser microstructure than for slag activated with calcium hydroxide.Additionally, the amount of C-S-H and hydrotalcite in its microstructure increased and the amount of portlandite was minimal, both of which indicate that the maximum compressive strength had been achieved.

Carbide slag-activated GGBS
According to the chemical analysis of carbide slag, it consists of more than 80% calcium hydroxide.Therefore, researchers have used carbide slag as an activator in AAS.For samples activated with carbide slag after 7 days of curing, the optimal percentage of carbide slag used to obtain the highest compressive strength was 10%.This value decreased with an increase in the compressive strength of the carbide slag (Li & Yi, 2020).The expanding use of carbide slag resulted in a higher calcium hydroxide content, which was detrimental to the development of strength (Pacheco-Torgal et al., 2008).As curing increased from 7 to 28 days and beyond, the optimum percentage of carbide slag used to activate the slag was found to be 5% (Li & Yi, 2020).This could be because the positive effects of the increase in activator became more evident during the shorter curing periods.After 7 days of curing, the use of hydrated lime as a slag activator increased the compressive strength results over those for carbide slag.However, over time, the compressive strength of slag activated by hydrated lime and carbide slag were very similar (Li & Yi, 2020).
Hydrotalcite peaks also can be seen in XRD profiles of slag activated with calcium oxide and calcium hydroxide (Ben Haha et al., 2011, 2012;Escalante-Garc� ıa et al., 2003;Oh et al., 2010;Van Jaarsveld et al., 1997;Wang & Scrivener, 1995) because of the presence of sufficient magnesium in the activated slag samples.Calcium hydroxide peaks also can be seen in the XRD profiles of both calcium hydroxide-activated and calcium oxide-activated slag samples (Figure 7) (Kim et al., 2013), although, after 28 days of curing, these peaks disappeared for the calcium hydroxide-activated samples due to the high consumption of calcium hydroxide and the formation of calcium carbonate.
XRD profiles of samples activated with calcium oxide and calcium hydroxide showed that the intensity of the peaks of C-S-H in the slag samples activated with calcium oxide was higher than in the slag samples activated with calcium hydroxide (Kim et al., 2013).In addition, the amount of unactivated slag in the slag samples activated with calcium hydroxide was higher than in the slag samples activated with calcium oxide.Therefore, the use of calcium oxide as an activator, by minimizing the unactivated slag and creating more hydration products, causes more compressive strength in the samples activated with them compared to the samples activated with calcium hydroxide (Kim et al., 2013).
XRD analysis showed that the use of sodium sulfate and sodium carbonate as auxiliary activators of calcium hydroxide/calcium oxide formed C-S-H.The increases in the C-S-H peaks in the XRD profile indicate an improvement in the quality of the microstructure of the AAS.The type of auxiliary activator also affected the results of the XRD analysis.The use of sodium sulfate as an auxiliary activator of calcium hydroxide intensified the C-S-H peaks compared to sodium carbonate (Yang et al., 2012b).
The results of XRD analysis indicated that, in slag samples activated with 5%, 10%, and 20% MgO, 18%, 16%, and 13% of the MgO remained un-hydrated, respectively.This indicates that a small amount of MgO in the binder matrix did not react.All samples displayed peaks of a similar magnitude that are associated with ettringite, it appears that the incorporation of MgO does not affect the formation of the substance.In fact, the formation of ettringite did not depend on the amount of MgO, but on the amount of calcium sulfate or gypsum in the slag (Jin et al., 2014;Myers et al., 2017;Park et al., 2016b).
Hydrotalcite is a double-layered hydroxide of Mg-Al, so providing more of the MgO could help it form more effectively (Figure 8) (Yi et al., 2014).In addition, an increase in the percentage of MgO used to activate the slag caused a decrease in the intensity of the hydrated phase of gehlenite.
The XRD chart showed the high-reactivity MgO resistance phases more clearly.In actuality, highly reactive MgO reacted with the slag more quickly.As a result, more hydrotalcite formed (Jin et al., 2015).In addition, the amount of MgO that did not participate in the hydration reactions increased with the use of low-reactivity MgO.With highly reactive MgO, the Mg was consumed rapidly and formed M-S-H, which is difficult to detect by XRD analysis (Jin & Al-Tabbaa, 2013;Zhang et al., 2011).

Calcined dolomite-activated GGBS
XRD analysis of slag activated by calcined dolomite, as for the other activators, showed the presence of hydrotalcite-like and C-S-H phases (Gu, 2014;Gu et al., 2014b).The calcium carbonate peaks associated with high percentages of activator were very strong, and in some cases overlap with the C-S-H-related peaks.The XRD diagrams for slag samples activated by calcined dolomite (Figure 9) (Gu et al., 2014b) showed that the C-S-H-related peaks in the samples activated by calcined dolomite (1000 � C) were more than those for samples activated by calcined dolomite (800 � C).It can be claimed that more slag was activated by these activators.
The use of calcined dolomite produced more portlandite and C-S-H than calcium oxide and MgO activators, and less hydrotalcite over time.The slag samples activated by calcined dolomite had the weakest microstructure of any other activators of alkaline earth metal ions.By using auxiliary activators, calcined dolomite-activated specimens may have their microstructure characteristics strengthened.This should be given special attention in future research because calcined dolomite can be used to produce economically viable AAS.
The witherite (barium carbonate, BaCO 3 ) observed in the XRD diagram formed when barium was present along with sulfate and carbonate sources (Mobasher et al., 2014), which could be due to the barium hydroxide reacting with calcite.The use of barium hydroxide in the slag samples created a higher pH and the hydrotalcite phase was dependent on the dissolution of Mg ions (Khan & O'Hare, 2002;Shigeo & Teruhiko, 1973).As a result, the barium hydroxide activator caused a greater amount of hydrotalcite phase than other activators (Jeong et al., 2017).
The presence of portlandite in the XRD diagram depended on the reaction of barium hydroxide and calcite.In the XRD diagram (Figure 10), no ettringite can be observed because of the reaction between the barium ions and sulfate, which prevented its formation.This could have been the cause of the low compressive strength of the barium hydroxide-activated samples at short curing times (3 days), where ettringite played an important role in the initial strength of the cementitious systems (Jeong et al., 2017).
XRD analyses of slag samples activated with calcium oxide, calcium hydroxide, and barium hydroxide have shown that most C-S-H-related peaks are created in slag samples activated with barium hydroxide.In addition, it can be stated that a denser microstructure is created in the slag samples activated with barium hydroxide compared to the samples activated with calcium oxide and calcium hydroxide.Therefore, creating a denser microstructure and more C-S-H-related peaks in the slag samples activated with barium hydroxide causes the highest compressive strength in these samples compared to the slag samples activated with other activators.

Carbide slag-activated GGBS
Samples of slag activated by carbide slag (depending on the addition of aluminum to C-S-H) can be seen in the XRD diagrams' C-A-S-H phase.Additionally, the XRD diagram for slag activated with carbide slag (Li & Yi, 2020) shows the presence of a hydrotalcite phase, which was dependent on the formation of calcium hydroxide and carbonate ions.The kuzelite (Ca 4 Al 2 (SO 4 )(OH) 12 .6H 2 O) can be revealed in the XRD diagram of slag activated with carbide slag (Jeong et al., 2016a;Lothenbach et al., 2008) because there was a sufficient source of calcium sulfate in carbide slag.The anion exchange of mono sulfate with CO3 2-ions were what causes monocarboaluminate to form, and the sample with the highest CS content exhibits the peak of this substance more clearly (Bakolas et al., 2006;Yum et al., 2018).Over time, in slag samples activated with carbide slag, a stratlingite phase could be observed, which reduced the volume and increased the porosity, thus reducing the strength (Li & Yi, 2020).

Ca(OH) 2 /CaO-activated GGBS
TG/DTA is used to quantify hydration products.Different materials can be distinguished based on their thermal properties by measuring weight loss and its rate of occurrence directly and quickly using TG and DTA (Atis, 2004;Bakharev et al., 2001;Bakolas et al., 2006;Ngala & Page, 1997).TG/DTA indicated that calcium oxide-activated slag samples had slightly more C-S-H than the calcium hydroxide-activated samples, which improved their mechanical properties (Kim et al., 2013).For both samples, the weight loss on the TG analysis curve from 0 to 200 � C was 6-7%, which is similar to that for sodium hydroxide and metasilicate activators (Ben Haha et al., 2012).
The TGA curves showed that between 0 and 200 � C, a phase resembling hydrotalcite formed (Figure 11).It can be inferred that at 28 days, the intensity of the C-S-H peaks corresponding to the samples activated with calcium oxide has increased, in contrast to samples cured after just one day.Nevertheless, at 28 days, there was little difference between samples activated with calcium oxide and calcium hydroxide in terms of compressive strength.
In the DTA diagram (Figure 11), the samples activated with calcium oxide and calcium hydroxide showed C-S-H peaks at 200 to 250 � C and a second broad peak at 371 � C (MacKenzie et al., 1993).It can be said that the effects of these activators in the DTA diagram were similar to those of the other slag activators.
The DTA and TGA information also revealed that the samples activated with calcium hydroxide (CHSb) carbonation at 28 days.The Ca (OH) 2 decomposition temperature range, or the TGA curve's inflection points and DTA curve peaks, were not present in the CHSb sample (Figure 11) (Kim et al., 2013).
Due to carbonation, which is known to be directly proportional to compressive strength in alkali-activated GGBFS and cement pastes, samples activated with calcium hydroxide have a lower compressive strength at 28 days than samples activated with calcium oxide (CSb).It is because the porosity (and pores) of the paste directly affect how easily carbonation occurs in hydrated pastes or alkali-activated paste.

MgO-activated GGBS
The DTG diagram (Figure 12) showed that in MgO-activated slag samples, the peaks formed at 90 � C due to the evaporation of the water absorbed by the C-S-H gel (Maciejewski et al., 1994) and at 680 � C due to the decarbonization of calcite (Cho et al., 2019;Theiss et al., 2013).In MgO-activated slag samples, the intensity of the DTG peaks increased at 60 � C and was related to the removal and absorption of water between the layers.From the DTG diagrams (Figure 12), it can be inferred that an increase in the percentage of MgO increased the formation of hydration products (Trittschack et al., 2014).As the percentage of MgO increased, a peak formed in the graph between 260 and 450 � C (Figure 12).In the TGA diagrams (Figure 12), hydrotalcite and brucite can be seen to form at 310 and 400 � C, respectively (Ben Haha et al., 2011).Therefore, from the DTG diagrams, it can be stated that an increase in the percentage of MgO required to activate the slag caused an increase in the hydrotalcite and brucite contents (Park et al., 2016).
It is claimed that there are too many overlaps within each temperature range for thermal analysis to be able to quantify each hydration phase.Weight loss is thus divided into three main phases without making a distinction between them (Ben Haha et al., 2011;Gruskovnjak et al., 2011).It has been stated that an increase in the percentage of MgO as a slag activator increased the weight difference between 50 and 250 � C and 250 and 550 � C. This, in turn, increased the degree of hydration occurring in the MgOactivated slag.TG analysis and calculation of the weight differences at different temperatures indicated that, in slag activated with high-reactivity MgO, the weight differences at those temperature ranges were higher than for slag activated with low-reactivity MgO.The degree of hydration of the activated slag was higher and more hydration products formed, resulting in more optimal mechanical properties (Jin et al., 2015).A comparison of the TG analysis of slag activated with MgO, calcium oxide, and calcium hydroxide indicated that the use of MgO at an optimal percentage produced the most C-S-H peak at about 100 � C. No peak between 400 and 500 � C was detected in the MgO-activated samples, as shown in Figure 13, except the peak represents CH (Yi et al., 2014).Thus, it can be stated that the use of MgO to activate slag created better mechanical properties and a denser microstructure than calcium oxide and calcium hydroxide activators.
In the DSC analysis of these samples, a single endothermic peak formed at 795 � C and ended at 820 � C as can be observed in Figure 14.The experimental results from the work of different researchers confirm that a calcination temperature of 900 � C resulted in the formation of calcined dolomite with a stable physical and chemical composition (Djayaprabha et al., 2017).

Barium hydroxide-activated GGBS
The majority of weight losses between 0 and 200 � C were caused by stratlingite, ettringite, or C-S-H dehydration.Due to the lack of ettringite formation, the barium hydroxide-activated samples displayed a fairly wide peak in this temperature range (Figure 15).A large hump at about 350 � C can be seen in the   TG diagram of the barium hydroxide-activated samples which corresponds to the hydrotalcite phase decomposition.In addition, a peak formed between 180 and 210 � C in the TG diagram for barium hydroxide-activated samples, indicating a stratlingite phase decomposition.In addition, the weight loss at about 400 � C can be related to portlandite decomposition (Figure 15) (Jeong et al., 2017).Because their decomposition temperatures were higher than 1000 � C, barite and witherite were not visible in the TGA curves for barium hydroxide-activated slag (Carmona-Quiroga & Blanco-Varela, 2013;Mobasher et al., 2014).
The weight loss for barium hydroxide-activated slag after 28 days of curing was greater than after 3 days of curing, indicating that more hydration products were produced over time in the samples.A comparison of the TG curves for the barium hydroxide and calcium hydroxide-activated samples showed that the weight loss for barium hydroxide-activated samples was greater than for the calcium hydroxideactivated samples.This indicates that more hydration products formed in samples activated with barium hydroxide (Jeong et al., 2017).

Carbide slag-activated GGBS
The weight loss from 30 to 200 � C in the TG diagram is related to the decomposition of C-S-H and AFm (Al 2 O 3 -Fe 2 O 3 -mono); this loss helps to explain the development of strength in samples of slag activated by carbide slag (Jeong et al., 2016b;Park et al., 2016).In this temperature range, the carbide-activated slag weight losses were higher than for pure slag, indicating the role of the carbide slag activator in slag hydration (Figure 16) (Li & Yi, 2020).
The DTA diagram for carbide-activated slag (Figure 15) (Santacruz et al., 2016), shows peaks at about 140 and 170 � C which indicate the presence of AFm.The peaks for samples activated by 5% carbide slag were greater than those activated by 20% carbide slag.The peak at about 100 � C is related to the presence of C-S-H and was greater in samples activated by 5% carbide slag than by 20% carbide slag.
Lower C-S-H and more AFt (typically Ca 6 Al 2 (SO 4 ) 3 (OH) 12 .26H 2 O) percentages for the slag samples activated with percentages greater than 5% carbide slag indicate that the optimum carbide slag content for slag was 5%.Weight loss at 400-500 � C indicated the removal of water in calcium hydroxide and was lowest for slag activated by more than 5% carbide slag (Li & Yi, 2020).

Ca(OH) 2 /CaO-activated GGBS
The EDS results in Figure 17 indicate that the calcium-to-silicon ratio in the C-S-H matrix of the calcium oxide-activated slag samples was between 1.04 and 1.15 (Liu et al., 2021).SEM analysis of the calcium oxide-activated slag samples showed that the slag particles were surrounded by hydration products, resulting in a dense matrix.However, micro-cracks and particles with a low degree of hydration of approximately 5 lm in diameter were evident and resulted in low compressive strength.
SEM analyses of slag samples activated with calcium oxide and calcium hydroxide have shown that a denser microstructure is created in slag samples activated with calcium oxide.In addition, EDAX analysis has shown that the degree of geo-polymerization of slag samples activated with calcium oxide is higher than that of slag samples activated with calcium hydroxide.Therefore, denser microstructure in slag samples activated with calcium oxide creates more optimal mechanical characteristics.
A denser microstructure could be created using auxiliary activators.For example, the use of fluorgypsum will create ettringite crystals that are similar to fibers between the pores and the slag particles and will produce a more uniform matrix and improved mechanical properties (Ashok et al., 2020;Kunther et al., 2015;Park et al., 2016).Of course, more C-A-S-H gel also will be produced and could be an additional factor in strengthening the mechanical properties of calcium oxide-activated slag and auxiliary activators of fluorgypsum (an anhydrite byproduct obtained after the reaction of sulfuric acid and fluorite; H 2 SO 4 þCaF 2 ¼2HF þ CaSO 4 ).

MgO-activated ground GGBS
The calcium-to-silicon ratio can be calculated using EDS results (Bahmani et al., 2020a,b).For MgO-activated slag, the ratio will be 1 and 1.53 at different places (Figure 18) (Yi et al., 2014), but the percentage of activator did not affect this ratio.Analysis indicated that the concentration of Mg in the MgO-activated slag makes it difficult to distinguish between outer and inner products (Brough & Atkinson, 2002;Richardson et al., 2004;Richardson & Groves, 1992).As a result, the magnesium concentration was probably much higher than in slag activated by other activators.
EDS and regression analysis showed that the aluminum-to-silicon and magnesium-to-silicon ratios at different points on the diagram for activated slag were linearly related to MgO (Yi et al., 2014).This confirms the presence of C-S-H mixed with the hydrotalcite-like phase (Fernandez et al., 2005;Wang & Scrivener, 1995).
All samples of slag activated with different percentages of MgO have been shown in SEM analysis to have a dense microstructure with C-S-H gel.Portlandite and hydrotalcite-like sheets also were evident in all samples.Over time, denser microstructures formed in the MgO-activated samples.SEM analysis indicates that, with an increase in the percentage of MgO-activators, hydrotalcite-like sheets became more evident (Gu et al., 2014a,b).

Calcined dolomite-activated GGBS
SEM analysis of calcined dolomite-activated samples shows that the major hydration products of slag activated by calcined dolomite were C-S-H, calcite, and portlandite (Figure 19) (Djayaprabha et al., 2017).The use of calcined dolomite compared to pure slag showed C-S-H and more calcite in the SEM images of the samples, indicating a denser microstructure than for pure slag.An increase in the percentage of calcined dolomite from 20% weakened the microstructure of the activated slag and also weakened the mechanical properties.

Barium hydroxide-activated GGBS
SEM analysis of barium hydroxide-activated slag samples showed cracks of 0.1-5 lm in size in the microstructure because of the drying of the samples in a vacuum.The microstructure of the barium hydroxideactivated slag was brighter than that of the other activators.This possibly could be due to the increase in the number of barium atoms compared to other elements (Ca, Si, Al).In fact, the barium atoms were Figure 18.Slag activated with MgO (GM) EDS micro-analysis of: (a) calcium:silicon against magnesium:silicon of 28-day; (b) aluminium:silicon against magnesium:silicon 28-day (Yi et al., 2014).not concentrated at one point but were widely distributed throughout the matrix.EDS analysis showed that the Ca/Si ratio in barium hydroxide-activated slag samples was about 1.1 (Figure 20) (Jeong et al., 2017).
Barium hydroxide formed a denser microstructure than the other slag activators.The lowest Ca/Si ratio occurred in samples activated by barium hydroxide and the unreacted slag particles were minimal in the barium hydroxide-activated slag compared to the other activators.As a result, it was clear that most hydration products can be created using this activator.It can be said that barium hydroxide was the best activator for slag from among the available activators of alkaline earth metal ions.

Carbide slag-activated GGBS
In the slag samples activated with carbide slag, in addition to C-S-H, aluminum and magnesium are found in the EDS C-S-H results.This indicates that part of the silicon had been replaced with aluminum in the C-S-H gel (Figure 21) (Li & Yi, 2020).In samples of slag activated by carbide slag, the presence of Aft was more prominent and caused the pores to fill with AFt.This is a demonstration of the positive effects of carbide slag on strength development.With an increase in the percentage of carbide slag, a stratlingite phase can be seen in the SEM images (Li & Yi, 2020).This also explains the decrease in compressive strength of the samples with an increase in the percentage of carbide slag.

29 Si and 27 Al MAS-NMR spectroscopy
Raw GGBFS revealed an asymmetric and broad spectrum of the fourfold coordinated aluminum site (Al (IV)), according to the 29 Si MAS-NMR spectrometry range in Figure 22a (Jeong et al., 2017).Even after activation, all specimens still included considerable amounts of Al (IV) spectra.In the slag samples activated with calcium hydroxide, no additional peak of Al (IV) was formed, but in the samples activated with barium hydroxide, the sharp peak of Al (IV) was formed, most likely as a result of stratlingite and Al substitution in C-S-H.As a result, the samples activated with 10% Ba(OH) 2 significantly changed the 27 Al MAS-NMR spectrometry range, which should be connected to the highest degree of Al dissolution from the GGBFS.
Small fivefold [Al (V)] peaks and strong sixfold [Al (VI)] peaks appearance were both present in samples activated with barium hydroxide and calcium hydroxide.Ettringite was formed for the samples activated with Ca(OH) 2 near 13 ppm peaks.Monocarboaluminate and hydrotalcite for samples activated with barium hydroxide were formed in 8-12 ppm peaks.
The 27 Al MAS NMR spectra of the slag activated with MgO are displayed in Figure 22b (Park et al., 2020).Depending on the tetrahedral sites, increasing the amount of magnesium oxide as a slag activator drastically reduced the spectrum's resonance intensity.This may be because more reaction rate activation causes less aluminum to be present in the unreacted slag.

FTIR test
Figure 23a shows the FTIR bands of 28-day slag activated with MgO and Ca(OH) 2 (Yi et al., 2014).Regardless of the activator type and content used all spectra exhibit bands that are strongly similar, indicating that hydration products were very similar in nature.The main hydration product present is Si-O, which bends at about 970 cm À 1 , and Si-O-Si, which bends at 640 cm À 1 , which was typical of CSHlike compounds.The presence of CO3 2-is indicated by recognizable bands at 1400, 1500, and 860 cm À 1 , which can be explained by the existence of a hydrotalcite-like phase.
The FTIR results are shown in Figure 23b for the samples activated with carbide slag (Li & Yi, 2020).The pure GGBS sample does not contain any obvious O-H bands, but the samples activated with carbide slag accomplish and show an increase in this band as the carbide slag content increases.All of the specimens contained CO3 2-peaks, indicating the formation of calcite and hydrotalcite.The typical hydration product in all samples, i.e.C-S-H, was present in the Si-O band at 972 cm À 1 , and this peak gets stronger as carbide slag content rises.

Comparison of mechanical characteristics and microstructure of samples containing different activators
The compressive strength outcomes of samples activated with various activators and the resulting hydration products are compared in Table 2.The highest and lowest compressive strengths were produced in the slag samples activated with barium hydroxide and carbide slag, respectively, according to this table.A denser cement matrix and more hydration products of slag samples activated with barium hydroxide were to blame for this outcome.
It can be concluded that calcium silicate hydration products were produced in the slag samples activated with each of the activators of alkaline earth metal ions by comparing the results, which are from the primary hydration products of slag activated with alkaline earth metal ions.Except for the activated slag with calcined dolomite at 800 � C, the calcium hydroxide (CH) phase was also produced in all the samples.The CH total consumption of activated slag with calcined dolomite at 800 � C during the hydration process is the cause of this problem.Str€ atlingite was also only produced in slag samples that had been activated with calcium oxide, barium hydroxide, and carbide slag.This outcome demonstrates how much aluminum oxide is present in these samples.
Barite and witherite were formed only in the slag samples activated with barium hydroxide.When calcium coexists with salt and carbonate sources, barite and witherite, two Ba-bearing products, frequently form.In particular, the reaction between calcium sulfate and barium hydroxide can produce barite.The formed witherite may be calcite, which was initially present in the raw GGBFS, reacting with barium hydroxide.Only samples activated with carbide slag, which can reduce volume, increase porosity, and produce samples with the lowest compressive strength were able to experience AFM phases.

Recommendations for future research
Considering that the effects of activators of alkaline earth metal ions as activators of slag and types of aggregates used in AAS have not been yet considered at the same time, this issue can be considered in future research.In addition, the effects of fibers on the characteristics of ductility and energy absorption capacity of slag activated with activators of alkaline earth metal ions can be of interest to researchers.
Researchers may be interested in the effects that activators of alkaline earth metal ions had on the durability of samples placed in environments that are acidic, sulfate, and chloride destructive.Moreover, future research may take into account the effects of using silica fume, metakaolin, and other pozzolans to enhance the microstructure and mechanical properties of activated slag samples with activators of alkaline earth metal ions.Finally, the characteristics of slag activated with activators of alkaline earth metal ions exposed to high temperatures can be of interest to researchers.

Conclusions
Thus far, no review article has examined the microstructure and strength development of slag activated with activators of alkaline earth metal ions.The current study examined the effects of calcium oxide/calcium hydroxide, magnesium oxide, barium hydroxide, calcined dolomite, and carbide slag.For research purposes, the results reported for SEM/EDS, TG/DTA, XRD, and compressive strength tests were studied.The following are the general findings of the review article.
1.The use of calcium oxide as a slag activator caused the greatest number of hydration products to be created compared to calcium hydroxide based on the XRD analysis and SEM images.The compressive strength of the calcium oxide-activated slag samples was about 32 MPa greater than that of the calcium hydroxide-activated samples.The use of auxiliary activators increased the compressive strength of samples activated with calcium hydroxide.The use of 1% sodium sulfate and 2% sodium carbonate increased the compressive strength by 30% and 20%, respectively.2. For the slag activated by MgO, the reactivity was very effective.For high-reactivity MgO, the optimal activating percentage was 20%, and for low-reactivity MgO, the optimal activation percentage was 10%.MgO-activated samples produced the densest microstructure, lowest calcium-to-silicon ratio, and the greatest number of hydration products from the calcium oxide and calcium hydroxide activators, and TG analysis showed that no calcium hydroxide was found in the microstructure of slag activated by MgO. 3. The use of barium hydroxide as a slag activator caused a compressive strength of more than 38% compared to the calcium hydroxide activator.The calcium-to-silicon ratio in the barium hydroxideactivated slag samples was about 1.1, which was lower than for the calcium hydroxide-activated slag.The higher compressive strength of the slag samples activated by barium hydroxide could have resulted from the denser microstructure.4. The calcination temperature of calcined dolomite affected the microstructure of the activated slag such that the slag was activated by (1000 � C) calcined dolomite compared to the slag activated by (800 � C) calcined dolomite had a larger C-A-S-H gel and hydrotalcite phases.5.At 7 days of curing, the highest compressive strength results were obtained for samples activated by 10% carbide slag.At 28 days of curing, however, 5% carbide slag produced the highest compressive strength.The use of a higher percentage of carbide slag as an activator produced the lowest calcium-silicate-hydrate (C-S-H) content and caused the stratlingite phase to become more visible in the SEM images.

Table 1 .
The final optimum mixtures.

Table 2 .
Compressive strength and hydration: products of samples containing different activators.