Preliminary Interpretation of the Induction Period in Hydration of Sodium Hydroxide/Silicate Activated Slag

materials

Article Preliminary Interpretation of the Induction Period in Hydration of Sodium Hydroxide/Silicate Activated Slag

Yibing Zuo 1,* and Guang Ye 2 1 School of Civil and Hydraulic Engineering, Huazhong University of Science and Technology, Wuhan 430074, China 2 Faculty of Civil Engineering and Geosciences, Delft University of Technology, Stevinweg 1, 2628 CN Delft, The Netherlands; [email protected] * Correspondence: [email protected]; Tel.: +86-18771708516

Received: 13 September 2020; Accepted: 23 October 2020; Published: 27 October 2020

Abstract: Many calorimetric studies have been carried out to investigate the reaction process of alkali-activated slag paste. However, the origin of the induction period and action mechanism of soluble Si in the dissolution of slag are still not clear. Moreover, the mechanisms behind diﬀerent reaction periods are not well described. In this study, the reaction kinetics of alkali-activated slag paste was monitored by isothermal calorimetry and the eﬀect of soluble Si was investigated through a dissolution test. The results showed that occurrence of the induction period in hydration of alkali-activated slag paste depended on the presence of soluble Si in alkaline activator and the soluble Si slowed down the dissolution of slag. A dissolution theory-based mechanism was introduced and applied to the dissolution of slag, showing good interpretation of the action mechanism of soluble Si. With this dissolution theory-based mechanism, origin of the induction period in hydration of alkali-activated slag was explicitly interpreted.

Keywords: induction period; alkali-activated slag; soluble Si; dissolution; isothermal calorimetry

1. Introduction The technology of alkali activation of ground granulated blast furnace slag (GGBFS) has been increasingly used for producing a clinker-free cementitious material, since it is able to eﬃciently improve the sustainability of concrete production [1–4]. Awoyera and Adesina reviewed the application of alkali-activated slag in the view point of its sustainability and summarized that alkali-activated slag cement contributes to achieve a greener concrete, taking oﬀ the strain on exploration of natural resources for ordinary Portland cement (OPC) production and preventing not only environmental degradation due to the exploration but also contamination due to the disposal of slag [5]. The alkali-activated slag cement has the ability to address the energy and environmental concerns that are associated with OPC production. It is reported that 80% or greater reduction of CO2 emission can be achieved by alkali-activated slag cement through proper use of alkaline activator when compared to OPC [6]. Yang et al. assessed the CO2 reduction of alkali-activated slag concrete and found 55–75% reduction of CO2 emission relative to OPC concrete [7]. The CO2 reduction is mainly because of the absence of a high-temperature calcination step in the production of alkali-activated cement. Compared with OPC, alkali-activated slag cement exhibits similar or even better mechanical properties and durability performance [8–10]. All of these excellent properties are closely related to the reaction process of alkali-activated slag, because the reaction process governs the reaction products formation, microstructure development and thus microstructure-related-physical-properties, including mechanical properties and durability performance [11–13].

Materials 2020, 13, 4796; doi:10.3390/ma13214796 www.mdpi.com/journal/materials Materials 2020, 13, 4796 2 of 19

Isothermal calorimetry, as a well-accepted technique for studying the reaction process, has been successfully and prevalently applied in cementitious systems. Many isothermal calorimetric studies have been conducted to investigate the hydration process of OPC, such as the influence of moisture [14] and temperature [15], incorporation of cement replacement materials [16] as well as chemical admixtures [17], etc. As a latent hydraulic cementitious material, GGBFS is widely used to replace OPC in concrete production. In addition to the calorimetric responses that reflect the hydration of OPC, the blended systems of OPC and slag also show a calorimetric peak that reflects the reaction of slag [18,19]. The similar calorimetric responses, reflecting reaction of slag, are expected in alkali-activated slag cement. Compared with those of OPC based systems, the calorimetric responses of alkali-activated slag cement could be very different, particularly depending on the nature and concentration of alkaline activators. By studying the reaction kinetics of powder sodium silicate activated slag pastes and liquid sodium silicate activated slag pastes, the researchers found that both the powder sodium silicate activated slag pastes and the NaOH activated slag pastes showed a single calorimetric peak [20]. The liquid sodium silicate activated slag pastes and OPC pastes, however, both demonstrated multiple calorimetric peaks and an obvious induction period. In this study the induction period is defined as the duration between two calorimetric peaks on heat evolution curve, in which the rate of heat evolution is very low. The different calorimetric responses between powder and liquid sodium silicate activated slag pastes is resulted from a larger pH in the powder sodium silicate activator than that in the liquid sodium silicate activator (henceforth if not specified liquid sodium silicate is written as sodium silicate for clarity). In another study, the in-situ isothermal calorimetry was used to monitor the reaction kinetics in sodium hydroxide and sodium silicate activated slag pastes [21]. The in-situ data demonstrated only one major calorimetric peak with no induction period for sodium hydroxide activated slag pastes and multiple calorimetric peaks with an obvious induction period for sodium silicate activated pastes. Shi and Day [22] used a conduction calorimeter to monitor the reaction kinetics of alkali-activated slag pastes with different kinds of alkaline activators, such as NaOH, Na CO , Na SiO 5H O, Na PO 2 3 2 3· 2 3 4 and NaF. The calorimetric responses were dependent on the type of alkaline activator, as described by three proposed models by the researchers. For the reaction kinetics of alkali-activated blends of slag with other aluminosilicate precursors, Gao et al. studied the sodium silicate activated slag-fly ash paste [23] and Bernal et al. investigated the sodium silicate activated slag-metakaolin paste [24]. It can be found that in the viewpoint of the trend of heat evolution rate curve, both the sodium silicate activated slag-fly ash pastes and the sodium silicate activated slag-metakaolin pastes showed similar calorimetric signals to those of sodium silicate activated slag paste, as long as slag was the major component in the blends. According to the aforementioned calorimetric studies, among others in the literature like [25–28], an obvious different calorimetric characteristic can be found between the NaOH activated slag paste and the sodium silicate activated slag paste. As shown in Figure1, the heat evolution rate curve of NaOH activated slag paste demonstrates no induction period, while that of sodium silicate activated slag paste show an obvious induction period. This difference can be attributed to the presence of soluble Si in sodium silicate activator. However, the action mechanism that how soluble Si results in an induction period is still not clear. Furthermore, the mechanisms behind different reaction periods are not well described and understood. Therefore, this study aims to discuss and explicitly interpret the origin of the induction period and describe the mechanisms behind different reaction periods of alkali-activated slag paste. First, isothermal calorimetric experiments were carried out to monitor the calorimetric responses of sodium hydroxide activated slag pastes and sodium silicate activated slag pastes. Then, the effect of soluble Si on the dissolution of slag in alkaline solution was investigated through a dissolution test. In the dissolution test a recipe was carefully designed via thermodynamic analysis to avoid formation of any reaction products during the dissolution of slag. Afterwards a dissolution theory-based mechanism was introduced and applied to the dissolution of slag. Finally, the reaction process and origin of the Materials 2020, 13, 4796 3 of 19 induction period were described and interpreted, respectively. The results presented in this study will contribute to understanding the reaction kinetics of alkali-activated slag paste, particularly the effect of alkalineMaterials 2020 activator, 13, x FOR composition PEER REVIEW on the dissolution of slag. 3 of 20

Figure 1. Schematic representation of the heat evolution rate curves of NaOH activated slag paste and sodiumFigure 1. silicate Schematic activated representation slag paste. of The the shadowed heat evolution part refersrate curves to the of induction NaOH ac period.tivated slag paste and sodium silicate activated slag paste. The shadowed part refers to the induction period. 2. Materials and Methods

2.1. MaterialsTherefore, and this Mixtures study aims to discuss and explicitly interpret the origin of the induction period and describe the mechanisms behind different reaction periods of alkali-activated slag paste. First, isothermalIn this calorimetric study, GGBFS, experiments from ORCEM were carried The Netherlands, out to monitor was the used calorimetric to prepare responses alkali-activated of sodium slaghydroxide pastes. activated Table 1slag presents pastes and the sodium chemical silicate composition activated ofslag slag, pastes. as Then, determined the effect by of X-raysoluble fluorescenceSi on the dissolution spectrometry of slag (XRF) in alkaline through solution Epsilon 3XLEwas investigated spectrometer through (PANalytical, a dissolution Heracles test. Almelo, In the 2 Thedissolution Netherlands) test a test. recipe The was surface carefully area was designed determined via thermodynamic as 2.38 m /g by Geminianalysis VII to 2390 avoid (micromeritics, formation of Brussel,any reaction Belgium). products Figure 2during depicts the the dissolution particle size of distribution slag. Afterwards and X-ray a didiffssolutionraction pattern theory-based of slag, measuredmechanism by was EyeTech introduced laser-di andffraction applied (Ankersmid, to the dissolut Nijverdal,ion of slag. The Netherlands)Finally, the reaction and PW process 1830X-ray and dioriginffractometer of the induction (Philips, Ameterdam, period were Thedescribed Netherlands), and interpreted, respectively. respectively. It is clear thatThe noresults crystalline presented phase in wasthis identifiedstudy will from contribute the X-ray to pattern, understanding indicating the a totally reaction amorphous kinetics natureof alka ofli-activated slag. This resultslag paste, is in lineparticularly with the factthe effect that usually of alkaline GGBFS activator contains composition>95% vitreous on the phase dissolution [29]. of slag.

Table 1. Chemical composition of GGBFS. 2. Materials and Methods

Oxide SiO2 CaO Al2O3 MgO Fe2O3 SO3 K2O TiO2 LOI * 2.1. Materials and Mixtures Weight (%) 32.91 40.96 11.85 9.23 0.46 1.61 0.33 1.00 1.15 Materials 2020, 13, x FOR PEER REVIEW 4 of 20 In this study, GGBFS, from ORCEM* LOI refersthe Netherland to loss on ignition.s, was used to prepare alkali-activated slag pastes. Table 1 presents the chemical composition of slag, as determined by X-ray fluorescence spectrometry (XRF) through Epsilon 3XLE spectrometer (PANalytical, Heracles Almelo, the Netherlands) test. The surface area was determined as 2.38 m2/g by Gemini VII 2390 (micromeritics, Brussel, Belgium). Figure 2 depicts the particle size distribution and X-ray diffraction pattern of slag, measured by EyeTech laser-diffraction (Ankersmid, Nijverdal, the Netherlands) and PW 1830X-ray diffractometer (Philips, Ameterdam, the Netherlands), respectively. It is clear that no crystalline phase was identified from the X-ray pattern, indicating a totally amorphous nature of slag. This result is in line with the fact that usually GGBFS contains >95% vitreous phase [29].

Table 1. Chemical composition of GGBFS.

Oxide SiO2 CaO Al2O3 MgO Fe2O3 SO3 K2O TiO2 LOI * Weight (%) 32.91 (a)40.96 11.85 9.23 0.46 1.61 (b)0.33 1.00 1.15 * LOI refers to loss on ignition. FigureFigure 2.2. ParticleParticle size size distribution distribution (a) ( aand) and X-ray X-ray diffraction diffraction pattern pattern (b) of (b )slag, of slag,determined determined by laser- by laser-didiffractionffraction and diffractometer, and diffractometer, respectively. respectively.

SodiumSodium hydroxidehydroxide (analytical (analytical grade, grade,>98%, >98%, from fr Sigma-Aldrich,om Sigma-Aldrich, Darmstadt, Darmstadt, German) German) was mixed was with distilled water to prepare sodium hydroxide activator. Water glass (8.25 wt % Na O, 27.5 wt % mixed with distilled water to prepare sodium hydroxide activator. Water glass (8.25 wt2 % Na2O, 27.5 wt % SiO2 and 64.25 wt % H2O, from BRENNTAG, Apolda, Belgium), combined with sodium hydroxide and distilled water, was used to prepare sodium silicate activator with the desired contents of Na2O and SiO2. According to the presence of soluble Si, the alkaline activators are divided into two groups. One group is sodium hydroxide activator without soluble Si and the other group is sodium silicate activator with soluble Si. Table 2 presents the mix compositions of the alkali-activated slag pastes. According to the contents of Na2O and SiO2, the mixtures were denoted as N4S0, N4S5.4,

N6S0, N6S5.4, N8S0 and N8S5.4. The numbers following N and S indicate mass percentages of Na2O and SiO2 in the alkaline activator with respect to slag, respectively. For instance, sample N4S5.4 represents the alkali-activated slag paste with Na2O/slag = 4% and SiO2/slag = 5.4%. For all samples the water/slag ratio was fixed at 0.4.

Table 2. Mix compositions of alkali-activated slag pastes.

Mixture Slag (g) Na2O (g) SiO2 (g) Ms † Water (g) N4S0 100 4 0 0 40 N6S0 100 6 0 0 40 N8S0 100 8 0 0 40 N4S5.4 100 4 5.4 1.395 40 N6S5.4 100 6 5.4 0.93 40 N8S5.4 100 8 5.4 0.6975 40

† Ms is the molar modulus of alkaline activator, calculated as: Ms = SiO2 (mol)/ Na2O (mol).

2.2. Isothermal Calorimtry Test By following the procedures in ASTM C1679 [30], the heat release flow of alkali-activated slag pastes was measured by an isothermal calorimeter. Alkaline activator solutions were formulated and cooled down to the temperature of 20 °C prior to mixing with slag. The alkali-activated slag pastes were prepared by mixing slag and alkaline activator solution externally and about 10 g of fresh paste was immediately put into the isothermal calorimeter (TAM-Air-314). The measurement temperature was set at 20 °C. For each mixture, two samples from the same mixing batch were tested at the same time. The duration of measurement was one week.

2.3. Dissolution Test The dissolution test was carried out to study the effect of soluble Si on the dissolution of slag in alkaline solution. In the test, 0.1 g of slag was dissolved in 200 mL of alkaline solution. The temperature was fixed at 20 °C and a magnetic stirring of 250 rpm was applied to prevent particle

Materials 2020, 13, 4796 4 of 19

SiO2 and 64.25 wt % H2O, from BRENNTAG, Apolda, Belgium), combined with sodium hydroxide and distilled water, was used to prepare sodium silicate activator with the desired contents of Na2O and SiO2. According to the presence of soluble Si, the alkaline activators are divided into two groups. One group is sodium hydroxide activator without soluble Si and the other group is sodium silicate activator with soluble Si. Table2 presents the mix compositions of the alkali-activated slag pastes. According to the contents of Na2O and SiO2, the mixtures were denoted as N4S0, N4S5.4, N6S0, N6S5.4, N8S0 and N8S5.4. The numbers following N and S indicate mass percentages of Na2O and SiO2 in the alkaline activator with respect to slag, respectively. For instance, sample N4S5.4 represents the alkali-activated slag paste with Na2O/slag = 4% and SiO2/slag = 5.4%. For all samples the water/slag ratio was ﬁxed at 0.4.

Table 2. Mix compositions of alkali-activated slag pastes.

Mixture Slag (g) Na2O (g) SiO2 (g) Ms † Water (g) N4S0 100 4 0 0 40 N6S0 100 6 0 0 40 N8S0 100 8 0 0 40 N4S5.4 100 4 5.4 1.395 40 N6S5.4 100 6 5.4 0.93 40 N8S5.4 100 8 5.4 0.6975 40

† Ms is the molar modulus of alkaline activator, calculated as: Ms = SiO2 (mol)/Na2O (mol).

2.2. Isothermal Calorimtry Test By following the procedures in ASTM C1679 [30], the heat release ﬂow of alkali-activated slag pastes was measured by an isothermal calorimeter. Alkaline activator solutions were formulated and cooled down to the temperature of 20 ◦C prior to mixing with slag. The alkali-activated slag pastes were prepared by mixing slag and alkaline activator solution externally and about 10 g of fresh paste was immediately put into the isothermal calorimeter (TAM-Air-314). The measurement temperature was set at 20 ◦C. For each mixture, two samples from the same mixing batch were tested at the same time. The duration of measurement was one week.

2.3. Dissolution Test The dissolution test was carried out to study the effect of soluble Si on the dissolution of slag in alkaline solution. In the test, 0.1 g of slag was dissolved in 200 mL of alkaline solution. The temperature was fixed at 20 ◦C and a magnetic stirring of 250 rpm was applied to prevent particle aggregation. Three types of alkaline solution were used. The first type is 0.1 mol/L sodium hydroxide solution without soluble Si. The second and third types are 0.1 mol/L sodium hydroxide solutions with soluble Si by the concentrations of 1 mmol/L and 5 mmol/L, respectively. During the dissolution process, a small volume of solution was sampled with a syringe at different times. In order to remove the undissolved slag, filter paper was used to filter the sampled solution. Prior to element analysis in solution, the filtered solutions were diluted using nitric acid (2.0 vol.%). Afterwards, the diluted solutions were analyzed using a PerkinElmer Optima 5300DV ICP-OES spectrometer, based on which the concentrations of Si, Al and Ca in solution were obtained.

2.4. Scanning Electron Microscopy (SEM) For SEM sample preparation, the samples were crushed into small pieces (1–2 cm3), then impregnated using a low viscosity epoxy resin and afterwards polished down to 0.25 µm (for more information on preparing SEM samples, please refer to [31]). Then, the microstructure of the polished samples was examined by ESEM (Philips XL30, Eindhoven, The Netherlands). The obtained SEM images with a magniﬁcation of 500 were used to determine the degree of reaction of slag through × image analysis. More details on the image analysis method are provided in [32]. Moreover, the SEM Materials 2020, 13, 4796 5 of 19 images with a magniﬁcation of 1000 were also recorded for more clearly observing the microstructure × morphology of alkali-activated slag pastes.

2.5. Thermodynamic Analysis of Solution Thermodynamic analysis of solution was conducted to prove that no reaction products precipitated during the dissolution test in Section 2.3. In thermodynamic analysis, the saturation index (SI) was employed as a criteria to judge if reaction products can precipitate or not [33]:

SI = log(IAP/KS0) (1) where IAP and KS0 are the ion activity product and equilibrium solubility product of a solid phase, respectively. If SI > 0 it means this solid phase is likely to precipitate. If SI < 0, it indicates that this solid phase is not likely to precipitate. If SI = 0, it implies equilibrium between solution and solid. It should be noted that SI may result in confusions if someone wants to use it to analyze the precipitation probabilities of phases that dissociate into a diﬀerent number of ions (N)[33]. Therefore, eﬀective saturation index (ESI) was introduced:

ESI = SI/N (2)

In order to calculate the activities of ions, the Gibbs energy minimization software GEM-Selektor v.3 (http://gems.web.psi.ch/)[34,35] and the thermodynamic database in [36,37] for alkali-activated slag systems were used. The measured elemental concentrations were used as the input. Given the activities of ions, the ion activity products can be calculated according to the dissociation reactions of reaction products. With the ion activity products and equilibrium solubility products, the eﬀective saturation index can be determined via Equation (2).

The Considered Reaction Products in the Thermodynamic Analysis In alkali-activated slag cement, the reaction products can be categorized into primary reaction products and secondary reaction products. The primary reaction product is a calcium-sodium aluminosilicate hydrate (CNASH) [38]. Myers et al. used eight CNASH solid solution (CNASH_ss) end-members to model this primary reaction product [36]. These eight CNASH_ss end-members and their dissociation reactions and equilibrium solubility products are given in Table3.

Table 3. Chemical reactions and equilibrium solubility products at 25 ◦C and 1 bar for eight CNASH_ss end-members and secondary reaction products in alkali-activated slag.

End-Member Chemical Reactions Log Ks0 CNASH gel ideal solid solution eight end-members, ‘CNASH_ss’ model [36] (CaO)1.25 (Al2O3)0.125 (SiO2) (H2O)1.625 5CA · 2+ 2 · · 10.75 1.25Ca + SiO − + 0.25AlO− + 0.25OH− + 1.5H O − ⇔ 3 2 2 (CaO) (Al2O3)0.15625 (SiO2)1.1875 (H2O)1.65625 +0.6875OH− INFCA ·2+ · 2 · 8.90 Ca + 1.1875SiO − + 0.3125AlO− + 2H O − ⇔ 3 2 2 (CaO)1.25 (Na2O)0.25 (Al2O3)0.125 (SiO2) (H2O)1.25 5CNA · 2+ 2· · · + 10.40 1.25Ca + SiO − + 0.25AlO− + 0.5Na + 0.75OH− + H O − ⇔ 3 2 2 (CaO) (Na2O)0.34375 (Al2O3)0.15625 (SiO2)1.1875 (H2O)1.3 INFCNA ·2+ · 2 · · + 10.00 Ca + 1.1875SiO − + 0.3125AlO− + 0.6875Na + 1.3125H O − ⇔ 3 2 2 (CaO) (Na2O)0.3125 (SiO2)1.5 (H2O)1.1875 + 0.375OH− INFCN ·2+ 2· · + 10.70 Ca + 1.5SiO3− + 0.625Na + 1.375H2O − ⇔ 2+ 2 T2C* (CaO)1.5 (SiO2) (H2O)2.5 1.5Ca + SiO3− + OH− + 2H2O 11.60 · · ⇔ 2+ 2 − T5C* (CaO)1.25 (SiO2)1.25 (H2O)2. 1.25Ca + 1.25SiO3− + 2.5H2O 10.50 · · ⇔ 2+ 2 − TobH* (CaO) (SiO ) (H O) +OH− Ca + 1.5SiO − + 3H O 7.90 · 2 1.5· 2 2.5 ⇔ 3 2 − Crystalline reaction products in alkali-activated slag [33,39] 2+ 2 C2ASH8 (CaO)2 (Al2O3) (SiO2) (H2O)8 2Ca + 2AlO2− + SiO3− + 8H2O 19.10 · · · ⇔2+ − C3AH6 (CaO)3 (Al2O3) (H2O)6 3Ca + 2AlO2− + 4OH− + 4H2O 20.85 · ·2+ ⇔ − Ca(OH)2 Ca(OH)2 Ca + 2OH− 5.20 ⇔ 2+ − C4AH13 (CaO) (Al O ) (H O) 4Ca + 2AlO− + 6OH− + 10H O 25.41 4· 2 3 · 2 13 ⇔ 2 2 − Note: 5CA and INFCA are two C-A-S-H end-members; 5CNA and INFCNA are two C-N-A-S-H end-members; INFCN a C-N-S-H end-member; T2C*, T5C* and TobH* are three C-S-H end-members [36]. Materials 2020, 13, 4796 6 of 19

The secondary reaction products are crystalline phases that are identiﬁed in alkali-activated slag cement, such as hydrotalcite [40], tetracalcium aluminate hydrate (C4AH13)[40], katoite (C3AH6)[37], stratlingite (C2ASH8)[41] and portlandite (CH) [42]. The dissociation reactions and equilibrium solubility products of these crystalline phases except hydrotalcite are also given in Table3.

3. Results

3.1. Reaction Kinetics of Alkali-Activated Slag Paste The measured heat evolution rates for sodium hydroxide and sodium silicate activated slag pastes are plotted in Figure3a,b, respectively. For NaOH activated slag pastes (Figure3a), the heat release rate curves demonstrate two calorimetric peaks, between which there is no noticeable induction period. These observations are in line with [12,20–22,43]. The first calorimetric peak, i.e., P1, occurred in the first ten minutes. It results from the dissolution/wetting of slag in sodium hydroxide solution [20,43]. The second calorimetric peak, i.e., P2, appeared around 1~3 h after mixing. According to the literature [22,43], P2 results from the formation of a large amount of reaction products. An increase of Na2O content in the sample led to a higher P2. Based on P1 and P2, three reaction periods can be identified during the reaction process of sodium hydroxide activated slag paste. As shown in Figure3a, these three reaction periods are the initial dissolution period (I), acceleration/deceleration period (II) andMaterials steady 2020,period 13, x FOR (III), PEER respectively. REVIEW 7 of 20

(a) (b)

Figure 3. Heat evolution rates fo forr sodium hydroxide ( a)) and and sodium sodium silicate silicate ( (b)) activated activated slag slag pastes. pastes.

In the graph, tt1,, t t2 andand t t33 standstand for for the the transition transition time. time.

3.2. DissolutionFor sodium Results silicate activated slag paste (Figure3b), an obvious induction period appeared between the two calorimetric peaks. This is obviously diﬀerent from the calorimetric response for sodium Table 4 lists the measured concentrations of Si, Al and Ca for the dissolution of slag in 0.1 mol/L hydroxide activated slag pastes. An increase of Na O content in the sample led to both a higher sodium hydroxide solution with and without soluble 2Si. The thermodynamic analysis of solution will P and earlier occurrence of P . According to P and P , the reaction process of sodium silicate be2 carried out and the effect of 2soluble Si on the dissolution1 2 of slag will be discussed in the next two activated slag paste had four reaction periods, i.e., the initial dissolution period (I), induction period sections, respectively. (II), acceleration/deceleration period (III) and steady period (IV), respectively.

3.2. Dissolution ResultsTable 4. Measured elemental concentrations of Si, Al and Ca (mmol/L). [Si]Table 4 lists the Element measured 10 concentrations min 25 min of Si, Al 40 and min Ca for the 60 min dissolution 90 ofmin slag in 0.1120 mol min/L sodium hydroxide solution(Si) with 0.0786 and without0.0961 soluble Si.0.1400 The thermodynamic 0.1929 analysis 0.2464 of solution0.2857 will be carried out and the eﬀect of soluble Si on the dissolution of slag will be discussed in the next two 0 mmol/L (Al) 0.0415 0.0459 0.0681 0.0900 0.1111 0.1270 sections, respectively. (Ca) 0.1300 0.1650 0.1775 0.1975 0.2500 0.2700 [Si] element 10 min 30 min 60 min 90 min 120 min (Si) 1.4500 1.4821 1.5000 1.5429 1.5679 1 mmol/L (Al) 0.0333 0.0296 0.0519 0.0556 0.0704 (Ca) 0.1400 0.1175 0.1350 0.1475 0.1625 (Si) 5.1071 5.3214 5.0714 4.2143 4.7857 5 mmol/L (Al) 0.0444 0.0407 0.0407 0.0593 0.0481 (Ca) 0.0575 0.1050 0.0850 0.0725 0.0925

3.2.1. Dissolution Results The equilibrium solubility products of the eight CNASH_ss end-members and secondary reaction products in Table 3 were used to calculate the ESI as described in Section 2.5. Since the concentration of Mg in solution was below the detection limit and thus not measured, the ESI with respect to hydrotalcite was not calculated. It is noted that the ionic strength of solution was below 1.0 mol/L (see Figure 4), which is within the valid range (~1–2 mol/L) of the activity correction using the extended Debye–Huckel equation [44].

Materials 2020, 13, 4796 7 of 19

Table 4. Measured elemental concentrations of Si, Al and Ca (mmol/L).

[Si] Element 10 min 25 min 40 min 60 min 90 min 120 min (Si) 0.0786 0.0961 0.1400 0.1929 0.2464 0.2857 0 mmol/L (Al) 0.0415 0.0459 0.0681 0.0900 0.1111 0.1270 (Ca) 0.1300 0.1650 0.1775 0.1975 0.2500 0.2700 [Si] element 10 min 30 min 60 min 90 min 120 min (Si) 1.4500 1.4821 1.5000 1.5429 1.5679 1 mmol/L (Al) 0.0333 0.0296 0.0519 0.0556 0.0704 (Ca) 0.1400 0.1175 0.1350 0.1475 0.1625 (Si) 5.1071 5.3214 5.0714 4.2143 4.7857 5 mmol/L (Al) 0.0444 0.0407 0.0407 0.0593 0.0481 (Ca) 0.0575 0.1050 0.0850 0.0725 0.0925

3.2.1. Dissolution Results The equilibrium solubility products of the eight CNASH_ss end-members and secondary reaction products in Table3 were used to calculate the ESI as described in Section 2.5. Since the concentration of Mg in solution was below the detection limit and thus not measured, the ESI with respect to hydrotalcite was not calculated. It is noted that the ionic strength of solution was below 1.0 mol/L (see Figure4), which is within the valid range (~1–2 mol /L) of the activity correction using the extended Debye–HuckelMaterials 2020, 13, x equation FOR PEER [ 44REVIEW]. 8 of 20

Figure 4. Ionic strength of solution. Figure 4. Ionic strength of solution. Figures5 and6 present the calculated e ffective saturation indexes (ESI). Most of the ESI were smallerFigures than 5 0. and This 6 meanspresent thatthe thecalculated solutions effective were undersaturatedsaturation indexes with (ESI respect). Most to of all the considered ESI were reactionsmaller productsthan 0. This in alkali-activated means that the slag solutions cement. were Therefore, undersaturated these reaction with products respect are to unstableall considered in the solutionsreaction products and thus in are alkali-activated not expected to slag precipitate. cement. InTherefore, other words, these the reaction elements products like Si, are Al andunstable Ca, etc. in werethe solutions in the solution and thus after are dissolution not expected from to precipitat slag. e. In other words, the elements like Si, Al and Ca, etc. wereThe calculatedin the solution pH of after solutions dissolution is presented from slag. in Figure7. Although the addition of Si in sodium hydroxide solution decreased the pH, the reduction was smaller than 0.04. This tiny decrease of pH is not supposed to significantly affect the dissolution rate of slag. In other words, the dissolution of slag can only be influenced by the presence of soluble Si. As discussed previously that no reaction products precipitated and the elements like Si, Al and Ca, z. were all in the solution after dissolution, then variations of the concentrations of Ca and Al are able to reflect the dissolution degree of slag. Larger concentrations of Ca and Al means a larger dissolution degree of slag. Based on the variations of the concentrations of Ca and Al, the effect of soluble Si on the dissolution of slag can be deduced. This aspect will be discussed in details in the next subsection.

Figure 5. Calculated effective saturation indexes (ESI) for primary reaction products for the dissolution of slag in 0.1 mol/L sodium hydroxide solution with and without soluble Si.

Materials 2020, 13, x FOR PEER REVIEW 8 of 20

Figure 4. Ionic strength of solution.

Figures 5 and 6 present the calculated effective saturation indexes (ESI). Most of the ESI were smaller than 0. This means that the solutions were undersaturated with respect to all considered reaction products in alkali-activated slag cement. Therefore, these reaction products are unstable in Materials 2020, 13, 4796 8 of 19 the solutions and thus are not expected to precipitate. In other words, the elements like Si, Al and Ca, etc. were in the solution after dissolution from slag.

FigureFigure 5. 5.Calculated Calculated eﬀ ectiveeffective saturation saturation indexes indexes (ESI) (E forSI) primary for primary reaction reaction products products for the dissolutionfor the Materials 2020, 13, x FOR PEER REVIEW 9 of 20 ofdissolution slag in 0.1 of mol slag/L sodiumin 0.1 mol/L hydroxide sodium solution hydroxide with solution and without with an solubled without Si. soluble Si.

FigureFigure 6. 6.Calculated Calculated eﬀ ectiveeffective saturation saturation indexes indexes (ESI) (ESI) for secondary for secondary reaction reaction products products for the dissolutionfor the ofdissolution slag in 0.1 of mol slag/L sodiumin 0.1 mol/L hydroxide sodium solution hydroxide with solution and without with an solubled without Si. soluble Si.

The calculated pH of solutions is presented in Figure 7. Although the addition of Si in sodium hydroxide solution decreased the pH, the reduction was smaller than 0.04. This tiny decrease of pH is not supposed to significantly affect the dissolution rate of slag. In other words, the dissolution of slag can only be influenced by the presence of soluble Si. As discussed previously that no reaction products precipitated and the elements like Si, Al and Ca, z. were all in the solution after dissolution, then variations of the concentrations of Ca and Al are able to reflect the dissolution degree of slag. Larger concentrations of Ca and Al means a larger dissolution degree of slag. Based on the variations of the concentrations of Ca and Al, the effect of soluble Si on the dissolution of slag can be deduced. This aspect will be discussed in details in the next subsection.

Figure 7. Calculated pH.

Materials 2020, 13, x FOR PEER REVIEW 9 of 20

Figure 6. Calculated effective saturation indexes (ESI) for secondary reaction products for the dissolution of slag in 0.1 mol/L sodium hydroxide solution with and without soluble Si.

The calculated pH of solutions is presented in Figure 7. Although the addition of Si in sodium hydroxide solution decreased the pH, the reduction was smaller than 0.04. This tiny decrease of pH is not supposed to significantly affect the dissolution rate of slag. In other words, the dissolution of slag can only be influenced by the presence of soluble Si. As discussed previously that no reaction products precipitated and the elements like Si, Al and Ca, z. were all in the solution after dissolution, then variations of the concentrations of Ca and Al are able to reflect the dissolution degree of slag. Larger concentrations of Ca and Al means a larger dissolution degree of slag. Based on the variations Materialsof the concentrations2020, 13, 4796 of Ca and Al, the effect of soluble Si on the dissolution of slag can be deduced.9 of 19 This aspect will be discussed in details in the next subsection.

Materials 2020, 13, x FOR PEER REVIEW 10 of 20 FigureFigure 7.7. Calculated pH. 3.2.3.2.2.2. Effect Eﬀect of Soluble Si on the Dissolution of Slag

Figure8 8 displays displays the the measured measured concentrations concentrations of Caof andCa and Al for Al the for dissolution the dissolution of slag of in slag 0.1 molin 0.1/L mol/Lsodium sodium hydroxide hydroxide solution solution with and with without and without soluble so Si.luble Itcan Si. It be can seen be that seen the that concentration the concentration of Ca ofwas Ca reduced was reduced when morewhen solublemore soluble Si was Si added was added into the into sodium the sodium hydroxide hydroxide solution. solution. The added The solubleadded solubleSi in the Si sodium in the sodium hydroxide hydroxide solution solution was also was found also found to result to inresult a decrease in a decrease of the of concentration the concentration of Al. ofSince Al. allSince the all dissolved the dissolved elements elements were in were the solution,in the solution, reductions reductions of the concentrationsof the concentrations of Ca and of Ca Al andreﬂect Al thereflect retarded the retarded dissolution dissolution of slag dueof slag to the due soluble to the Si.soluble In other Si. In words, other thewords, soluble the Sisoluble in alkaline Si in alkalinesolution solution slowed downslowed the down dissolution the dissolution of slag. of slag.

(a) (b)

Figure 8. MeasuredMeasured concentrations concentrations of Ca ( aa)) and and Al Al ( (bb)) for for the the dissolution dissolution of of slag slag in in 0.1 0.1 mol/L mol/L sodium hydroxide solution with and without soluble Si.

The retarding eeffectffect ofof solublesoluble SiSi on on dissolution dissolution of of slag slag then then slowed slowed down down the the reaction reaction kinetics kinetics of ofsodium sodium silicate silicate activated activated slag, slag, resulting resulting in an in induction an induction period period during during the reaction the reaction process process as shown as shownin Figure in3 b.Figure The induction3b. The induction period has period been alsohas widelybeen also detected widely through detected isothermal through calorimetry isothermal calorimetrymeasurements measurements by many studies by many in the literaturestudies in [20 th–e22 literature,25–28]. Besides [20–22,25–28] the reaction. Besides kinetics, the thereaction effect kinetics,of soluble the Si effect by slowingof soluble down Si by theslowing dissolution down the of dissolution slag also a offfects slag thealso microstructure affects the microstructure formation. formation.Many studies Many reported studies a homogenousreported a homogenous microstructure microstructure in sodium silicate in sodium activated silicate slag activated as opposed slag to as a opposedheterogeneous to a heterogeneous microstructure microstructure in sodium hydroxide in sodium activated hydroxide slag [32 activated,45,46]. One slag reason [32,45,46]. is that One the reasonslowing is downthat the dissolution slowing down of slag dissolution and thus retardedof slag and reaction thus retarded kinetics leavesreaction ample kinetics time leaves for reaction ample timeproducts for reaction to form products evenly in to sodium form evenly silicate in activated sodium slag.silicate Another activated reason slag. is Another that the reason soluble is Si that can the act solubleas nucleation Si can sites.act as nucleation sites.

4. Discussion

4.1. Dissolution of Slag In aluminosilicate materials including slag, Si and Al build up the framework while alkali and alkali-earth metals like Ca and Mg modify the framework [47,48]. So, the alkali and alkali-earth metals are also called modifying elements. The framework refers to the glass network in aluminosilicate materials. In the framework Si and Al are tetrahedrally coordinated. As schematically shown in Figure 9, the dissolution of slag can be described via the following four steps [49–51].

Materials 2020, 13, 4796 10 of 19

4. Discussion

4.1. Dissolution of Slag In aluminosilicate materials including slag, Si and Al build up the framework while alkali and alkali-earth metals like Ca and Mg modify the framework [47,48]. So, the alkali and alkali-earth metals are also called modifying elements. The framework refers to the glass network in aluminosilicate materials. In the framework Si and Al are tetrahedrally coordinated. As schematically shown in MaterialsFigure9 2020, the, 13 dissolution, x FOR PEER of REVIEW slag can be described via the following four steps [49–51]. 11 of 20

(a) (b)

(d) (c)

Figure 9. SchematicSchematic illustration illustration of of the the dissolution of slag (after [[51]).51]). For For clarity, clarity, additional bonds between Si Si and and O O as as well well as as be betweentween Al Al and and O Oare are not not shown. shown. First, First, the the modifying modifying elements elements are are initiallyinitially releasedreleased through through the metalthe metal/proton/proton exchange exchange reactions, reactions, as shown as in (showna); then, in hydrolysis (a); then, of hydrolysisthe bonds between of the Albonds and Obetween starts, as Al shown and inO (starts,b); afterwards, as shown the in bonds (b); betweenafterwards, Si and the O startbonds to betweenbreak, as shownSi and inO ( cstart); finally, to break, Al and as Si shown are released, in (c); as finally, a result Al of whichand Si the are framework released, isas gradually a result ofdissolved, which the as shown framework in (d). is gradually dissolved, as shown in (d). Due to the smaller bonding energy of Al-O than Si-O, Al dissolves more easily than Si in the Due to the smaller bonding energy of Al-O than Si-O, Al dissolves more easily than Si in the dissolution of slag [47]. The initially dissolved Al changes the adjoined Si coordination condition dissolution of slag [47]. The initially dissolved Al changes the adjoined Si coordination condition from from fully coordinated to partially coordinated (Figure9c). Compared with the fully coordinated Si, fully coordinated to partially coordinated (Figure 9c). Compared with the fully coordinated Si, the the partially coordinated Si dissolves faster. So the dissolution of framework can be divided into the partially coordinated Si dissolves faster. So the dissolution of framework can be divided into the following two steps: initial release of a small amount of Al (Figure9b) and then followed by the release following two steps: initial release of a small amount of Al (Figure 9b) and then followed by the of Si that coordinates to the initially dissolved Al through O (Figure9c). release of Si that coordinates to the initially dissolved Al through O (Figure 9c). Due to the preferential dissolution of modifying elements, a leached surface layer is formed Due to the preferential dissolution of modifying elements, a leached surface layer is formed around the particle [47,48], as shown in Figure 10. This leached surface layer is mostly composed around the particle [47,48], as shown in Figure 10. This leached surface layer is mostly composed of of leftover Al and Si and it affects the dissolution of slag since it influences how fast the modifying leftover Al and Si and it affects the dissolution of slag since it influences how fast the modifying elements diffuse through it. A larger thickness of the leached surface layer obviously led to a smaller diffusion rate of the modifying element.

Materials 2020, 13, 4796 11 of 19

elements diﬀuse through it. A larger thickness of the leached surface layer obviously led to a smaller MaterialsMaterials 2020 2020, ,13 13, ,x x FOR FOR PEER PEER REVIEW REVIEW 1212 of of 20 20 diﬀusion rate of the modifying element.

Figure 10. Schematic representation of the formation of the leached surface layer. Figure 10. SchematicSchematic representation representation of the form formationation of the leached surface layer.

4.2.4.2. A A A Dissolution Dissolution Dissolution Theory Theory Based Based Mechanis Mechanis Mechanismmm Applied Applied to to the the Dissolution Dissolution of of Slag Slag BasedBased on onon the thethe theory theory of of LasagaLasaga Lasaga andand and Luttge Luttge Luttge [ 52[52] [52]] and and and the the the “vacancy “vacancy “vacancy island” island” island” theory theory theory of Doveof of Dove Dove et al. et et [53al. al.], [53],Juilland[53], Juilland Juilland et al. et putet al. al. forward put put forward forward a dissolution a a dissolution dissolution theory theory basedtheory mechanismbased based mechanism mechanism to explain to to theexplainexplain onset the the of theonset onset induction of of the the inductionperiodinduction in period hydrationperiod in in hydration hydration of crystalline of of crystalline crystalline alite [54]. alite Inalite this [54]. [54]. mechanism In In this this mechanism mechanism the dissolution the the dissolution dissolution is divided is is into divided divided three intointoforms, three three i.e., forms, forms, the formation i.e., i.e., the the formation formation of vacancy of of islands,vacancy vacancy etch islands, islands, pit formationetch etch pit pit formation formation at dislocations at at dislocations dislocations and step and retreatand step step at retreatpre-existingretreat atat pre-existingpre-existing roughness. roughnessroughness Crystallographic. . CrystallographicCrystallographic defects (acting defectsdefects as dislocations) (acting(acting asas dislocations) anddislocations) solution and saturationand solutionsolution are saturationsaturationtwo main are factorsare two two consideredmain main factors factors in considered thisconsidered mechanism. in in this this Although mechanism. mechanism. there Although Although is no crystallographic there there is is no no crystallographic crystallographic defects in slag defectsdefectsbecause in in of slag slag its amorphousbecause because of of nature,its its amorphous amorphous preferential nature, nature, dissolution pr preferentialeferential of modifying dissolution dissolution elements of of modifying modifying and partially elements elements dissolved and and partiallypartiallyframework dissolved dissolved (see Figure framework framework9) would (see (see create Fi Figuregure defects 9) 9) would would in the create dissolutioncreate defects defects of slag.inin the the These dissolution dissolution created of of defects slag. slag. These couldThese createdcreatedact as dislocations defects defects could could on act whichact as as dislocations dislocations formation of on on etch which which pit canformation formation take place. of of etch etch Furthermore, pit pit can can take take the place. place. solution Furthermore, Furthermore, saturation thetheis alsosolution solution reported saturation saturation to significantly is is also also reported reported affect the to to dissolutionsignificantly significantly of affect slagaffect [ 55the the]. dissolution Therefore,dissolution itof of is slag slag conceivable [55]. [55]. Therefore, Therefore, to use the it it isisdissolution conceivable conceivable theory-based to to use use the the dissolution dissolution mechanism theory-based theory-based to interpret themechanism mechanism dissolution to to interpret ofinterpret slag. the the dissolution dissolution of of slag. slag. AmongAmong the the three three dissolution dissolutiondissolution forms, forms,forms, the thethe first firstfirst two two ar are aree fast fast dissolution dissolution processesprocesses processes whilewhile while thethe the thirdthird third is is is a critnn aaslow slow slow dissolutiondissolution dissolution process.process. process. TheTh Thee firstfirst first twotwo two dissolutiondissolution dissolution formsforms forms haveha haveve activationactivation activation energies, energies, i.e., i.e., ∆Δ ΔGGcritcrit nand and crit ΔΔ∆GGcrit, ,respectively. ,respectively. respectively. Prior Prior Prior to to to the the the onset onset onset of of of dissolution dissolution dissolution through through through the the the first first first two two two forms, forms, forms, the the the activation activation activation energy energy energy barriersbarriers must must be be removed. removed. Figure Figure Figure 11 11 schematically schematicallyschematically demonstrates demonstrates these these two two activation activation energy energy barriers barriers asas a a function function of ofof the thethe undersaturation undersaturation ofof of solution.solution. solution. AA A reductionreduction reduction of of of the the the undersaturation undersaturation undersaturation degree degree degree leads leads leads to to anto ananincrease increase increase of of theof the the activation activation activation energy energy energy barrier. barrier. barrier. The The The undersaturation un undersaturationdersaturation of of theof the the solution solution solution provides provides provides the the the energy energy energy to totoremove remove remove the the the activation activation activation energy energy energy barriers barriers barriers for for for dissolving dissol dissolvingving through through through the the formationthe formation formation of vacancy of of vacancy vacancy islands islands islands and and etchand etchpitetch atpit pit dislocations. at at dislocations. dislocations.

Figure 11. Schematic illustration of the two activation energy barriers ∆G n (vacancy islands) and critcritn FigureFigure 11. 11. Schematic Schematic illustration illustration of of the the two two activation activation energy energy barriers barriers Δ ΔGGcrit n(vacancy (vacancy islands) islands) and and ∆G (etch pit) as a function of the undersaturation degree. (after [54]). critcrit ΔΔGGcrit (etch (etch pit) pit) as as a a function function of of the the un undersaturationdersaturation degree. degree. (after (after [54]). [54]).

AsAs shown shown in in Figure Figure 11, 11, there there are are three three regimes, regimes, i.e., i.e., re regimesgimes I, I, II II and and III, III, and and their their corresponding corresponding raterate controllingcontrolling mechanismsmechanisms areare stepstep retreat,retreat, etetchch pitpit formationformation andand vacancyvacancy islandsislands formation,formation, respectively.respectively. In In regime regime III III the the undersaturation undersaturation degre degreee is is very very large large (for (for example example at at the the beginning beginning of of dissolutiondissolution of of slag), slag), providing providing sufficient sufficient energy energy to to remove remove the the activation activation energy energy barrier barrier for for formation formation

Materials 2020, 13, 4796 12 of 19

As shown in Figure 11, there are three regimes, i.e., regimes I, II and III, and their corresponding rate controlling mechanisms are step retreat, etch pit formation and vacancy islands formation, respectively. MaterialsIn regime 2020 III, 13 the, x FOR undersaturation PEER REVIEW degree is very large (for example at the beginning of dissolution13 of of20 slag), providing suﬃcient energy to remove the activation energy barrier for formation of vacancy ofislands vacancy on theislands smooth on the surface smooth of slag.surface As of the slag. undersaturation As the undersat movesuration from moves regime from III regime to regime III to II, regimethe energy II, the provided energy provided by undersaturation by undersaturation decreases de andcreases is not and able is not to remove able to remove the activation the activation energy energybarrier ofbarrier vacancy of vacancy islands islands formation formation on the smoothon the smooth surface surface of slag. of On slag. the On other the hand,other etchhand, pits etch is pitsable is to able form to at form the dislocations at the dislocations or defects or produceddefects produced in regime in III. regime When III. the When undersaturation the undersaturation moves to movesregime I,to the regime provided I, the energy provided by undersaturation energy by undersaturation continues to decrease continues and to it isdecrease not possible and anymoreit is not possiblefor the formation anymore of for etch the pits formation at dislocations of etch orpits defects. at dislocations In regime or I thedefects. dissolution In regime rate I isthe low, dissolution since the ratedissolution is low, issince limited the dissolution to the step retreat. is limited It should to the bestep pointed retreat. out It thatshould the be step pointed retreat out process that isthe active step retreatduring theprocess whole is dissolutionactive during process. the whole In regimes dissolu II andtion III,process. however, In regimes the contribution II and III, of stephowever, retreat the to contributionthe overall dissolution of step retreat rate isto small.the overall dissolution rate is small.

4.3. Interpreting the Action Mechanism of Soluble Si in the Dissolution of SlagSlag in Alkaline Solution When slag waswas broughtbrought intointo contactcontact with with the the NaOH NaOH solution solution with with soluble soluble Si, Si, the the soluble soluble Si Si led led to toa very a very low low undersaturation undersaturation degree degree with with respect respect to Si.to Si. As As aresult, a result, the the undersaturation undersaturation was was unlikely unlikely to tosupply supply su sufficientfficient energy energy to to remove remove the the activation activation energy energy barriers barriers and and thus thus itit isis didifficultfficult for vacancy islands and etch pits to form spontaneously on th thee leached surface layer. Then the dissolution moves rapidly to regime I, in which the dissolution is limitedlimited to step retreat. Therefore, the dissolution was slow andand thusthus the the dissolution dissolution of of slag slag was was retarded. retarded Consequently,. Consequently, the concentrationsthe concentrations of Ca of and Ca Aland were Al weresmaller smaller in the in solution the solution with soluble with soluble Si than thoseSi than without those without solubleSi. soluble The retarding Si. The retarding effect of soluble effect of Si solubleon the reaction Si on the of reaction slag can of be slag further can confirmedbe further byconfirmed the degree by ofthe reaction degreeof of slag reaction as shown of slag in as Figure shown 12. inIt canFigure be seen 12. It that can the be degreeseen that of the reaction degree of of slag reaction was always of slag smaller was always for sodium smaller silicate for sodium activated silicate slag activatedthan that forslag sodium than that hydroxide for sodium activated hydroxide slag withactivated the same slag with content the of same Na2 O.content of Na2O.

Figure 12. DegreeDegree of reaction of slag, derived from SEM-image analysis, for N6S0 and N6S5.4. 4.4. Interpreting the Eﬀects of Soluble Ca and Al in Solution on the Reaction of Slag 4.4. Interpreting the Effects of Soluble Ca and Al in Solution on the Reaction of Slag Suraneni et al. used a micro-reactor approach to study the reactions of slag in alkaline solutions Suraneni et al. used a micro-reactor approach to study the reactions of slag in alkaline solutions with soluble Ca and Al [55]. Figure 13 shows the micro-reactors before immersion in solution with soluble Ca and Al [55]. Figure 13 shows the micro-reactors before immersion in solution (Figure (Figure 13a), after 2 days in 0.1 M KOH solution (Figure 13b), after 2 days in 0.1 M KOH + 20 mM 13a), after 2 days in 0.1 M KOH solution (Figure 13b), after 2 days in 0.1 M KOH + 20 mM CaCl2 CaCl2 solution (Figure 13c) and after 2 days in 0.1 M KOH + 60 NaAlO2 solution (Figure 13d). Before solution (Figure 13c) and after 2 days in 0.1 M KOH + 60 NaAlO2 solution (Figure 13d). Before immersion in solutions, the base and walls of the gaps were completely smooth. After 2 days of immersion in solutions, the base and walls of the gaps were completely smooth. After 2 days of immersion in 0.1 M KOH solution, the gap showed clear growth of reaction products (globules). immersion in 0.1 M KOH solution, the gap showed clear growth of reaction products (globules). After After 2 days of immersion in 0.1 M KOH solution with soluble Ca and Al, by contrast, the degree of 2 days of immersion in 0.1 M KOH solution with soluble Ca and Al, by contrast, the degree of dissolution of slag was significantly reduced, the walls were smooth and there was not much reaction products formed in the gap. It can be clearly seen that the soluble Ca and Al in solution significantly slowed down the reaction of slag.

Materials 2020, 13, 4796 13 of 19 dissolution of slag was signiﬁcantly reduced, the walls were smooth and there was not much reaction products formed in the gap. It can be clearly seen that the soluble Ca and Al in solution signiﬁcantly Materialsslowed 2020 down, 13, thex FOR reaction PEER REVIEW of slag. 14 of 20

(a) (b)

FigureFigure 13. Micro-reactors:Micro-reactors: (a(a)) gaps gaps before before immersion immersion in solution,in solution, (b) after (b) 2after days 2 indays 0.1 Min KOH 0.1 M solution, KOH solution,(c) after 2(c days) after in 2 0.1 days M KOHin 0.1 +M20 KOH mM + CaCl20 mM2 solution CaCl2 solution and (d) and after (d 2) days after in2 days 0.1 M in KOH 0.1 M+ KOH60 mM + 60NaAlO mM NaAlO2 solution.2 solution. (cited from(cited [ 55from]). [55]).

TheThe inhibiting inhibiting effect effect of of soluble soluble Ca Ca and and Al on Al the on reaction the reaction of slag of may slag be may also be interpreted also interpreted by the dissolutionby the dissolution theory-based theory-based mechanism. mechanism. The solu Theble soluble Ca and Ca andAl Alin insolution solution decreased decreased the the undersaturationundersaturation with with respect respect to to the the anhydrous anhydrous phases phases.. This This made made it it difficult difficult to to provide provide sufficient sufficient activationactivation energyenergy toto overcome overcome the the energy energy barriers barrier fors vacancy for vacancy island formationisland formation and etch and pit formationetch pit formationat dislocations. at dislocations. The dissolution The dissolution of slag was of mainly slag wa limiteds mainly to thelimited step to retreat the step with retreat very low with dissolution very low dissolutionrate. Consequently, rate. Consequently, the reaction the of slagreaction was slowedof slag was down. slowed It should down. be It noted should that be the noted formation that the of formationcovalent bonds of covalent between bonds adsorbed between Al speciesadsorbed and Al hydroxylated species and hydroxylated silicate surfaces, silicate which surfaces, was found which in wasthe dissolutionfound in the of dissolution tricalcium silicateof tricalcium [56], might silicate also [5 act6], might in the dissolutionalso act in the of slag.dissolution This action of slag. can This lead actionto the inhibitingcan lead to e fftheect inhibiting of soluble effect Al on of the soluble reaction Al ofon slag.the reaction of slag. 4.5. Interpreting the Reaction Process and Origin of the Induction Period of Alkali-Activated Slag 4.5. Interpreting the Reaction Process and Origin of the Induction Period of Alkali-Activated Slag The good interpretation of the action mechanisms of soluble Si, Ca and Al in the dissolution or The good interpretation of the action mechanisms of soluble Si, Ca and Al in the dissolution or reaction of slag using the three forms of dissolution demonstrates that this dissolution theory based reaction of slag using the three forms of dissolution demonstrates that this dissolution theory based mechanism is applicable for describing the reaction process of alkali-activated slag. mechanism is applicable for describing the reaction process of alkali-activated slag. In the beginning of NaOH activated slag paste, undersaturation degree was very large since there In the beginning of NaOH activated slag paste, undersaturation degree was very large since were no soluble Si, Al, Ca and Mg in the activating solution. The undersaturation was able to provide there were no soluble Si, Al, Ca and Mg in the activating solution. The undersaturation was able to sufficient energy to remove activation energy barriers and thus all the three dissolution forms took provide sufficient energy to remove activation energy barriers and thus all the three dissolution forms place. This resulted in rapid dissolution of constituents in slag, which was accompanied with rapid took place. This resulted in rapid dissolution of constituents in slag, which was accompanied with release of heat (reflected by P1, see Figure3a). The elements released into solution then existed as rapid release of heat (reflected by P1, see Figure 3a). The elements released into solution then existed as soluble elements (normally as ions). These soluble elements then reduced the undersaturation degree. As a result, the dissolution gradually moved from regime III to regime II, in which the dissolution via vacancy islands formation was not likely to take place anymore. In the meanwhile, the preferential dissolution of modifying elements and Al led to an increase of the thickness of the leached surface layer. The leached surface layer became stable when the dissolution of modifying

Materials 2020, 13, 4796 14 of 19 soluble elements (normally as ions). These soluble elements then reduced the undersaturation degree. As a result, the dissolution gradually moved from regime III to regime II, in which the dissolution via Materialsvacancy 2020 islands, 13, x FOR formation PEER REVIEW was not likely to take place anymore. In the meanwhile, the preferential15 of 20 dissolution of modifying elements and Al led to an increase of the thickness of the leached surface elementslayer. The and leached Al was surface in a dynamic layer became equilibrium stable whenwith the the dissolution dissolution of of Si. modifying When the elements concentrations and Al ofwas soluble in a dynamic elements equilibrium in solution with increased the dissolution towards of saturation Si. When theor oversaturation, concentrations ofreaction soluble products elements startedin solution to precipitate. increased towardsThe precipitation saturation of orreacti oversaturation,on products consumed reaction products soluble startedelements to in precipitate. solution, whichThe precipitation increased the of reaction undersaturation products consumeddegree and soluble thus elementsaccelerated in solution,the dissolution which increasedof slag. The the dissolutionundersaturation of slag degree then andled thusto an accelerated increase of the th dissolutione concentrations of slag. of The soluble dissolution elements of slagin solution, then led facilitatingto an increase the ofprecipitation the concentrations of reaction of products. soluble elements In this way in solution, the dissolution facilitating of slag the and precipitation formation of reaction products.products Inprogressed this way theinterdependently dissolution of slagthrough and formationthe solubl ofe reactionelements products in solution. progressed Thus, interdependentinterdependently processes through led the to soluble a rapid elements growth of in reaction solution. products, Thus, interdependent which can be processesreflected by led P to2 in a Figure 3a. rapid growth of reaction products, which can be reﬂected by P2 in Figure3a. Since these two interdependent processes facilitated each other, P2 occurred closely after P1. As Since these two interdependent processes facilitated each other, P2 occurred closely after P1. As a a result, the induction period did not appear between P1 and P2 in the sodium hydroxide activated result, the induction period did not appear between P1 and P2 in the sodium hydroxide activated slag slag paste. It should be noted that a higher alkalinity or Na2O content of sodium hydroxide activator paste. It should be noted that a higher alkalinity or Na2O content of sodium hydroxide activator would wouldaccelerate accelerate these two these interdependent two interdep processes.endent processes. This can This be conﬁrmed can be confirmed by the observations by the observations in Figure3 ina, Figure 3a, i.e., the magnitude of P2 became larger and P2 also appeared earlier when the Na2O content i.e., the magnitude of P2 became larger and P2 also appeared earlier when the Na2O content increased. increased. As schematically shown in Figure 14, it can be inferred that P2 will rise and advance earlier As schematically shown in Figure 14, it can be inferred that P2 will rise and advance earlier to merge to merge with P1 when the alkalinity continues to increase. This inference has been evidenced by the with P1 when the alkalinity continues to increase. This inference has been evidenced by the isothermal isothermalcalorimetric calorimetric results in [21 results]. in [21].

Figure 14. SchematicSchematic illustration illustration of of the the effect eﬀect of increasi increasingng the pH of NaOH activator on the reaction kinetics of NaOH activated slag paste.

InIn sodium hydroxide activated slag paste, only slag grains acted as nucleation sites. This This led led to thethe formation formation of of reaction reaction products products mainly mainly on on the the su surfacerface of of slag slag grains. grains. The The continuous continuous formation formation of reactionof reaction products products around around slag slag resulted resulted in inthe the layers layers of of reaction reaction products, products, as as shown shown in in Figure Figure 15a.15a. This is in line with the SEM observations in [[32,45,432,45,466].]. The layers of reaction products then acted as diffusiondiffusion barriers for the didiffusionffusion of OHOH−- fromfrom the the solution solution to to the the un undissolveddissolved slag. slag. Consequently, Consequently, thethe dissolution of of slag slowed down, which henc hencee retarded the chemical reactions among soluble elements. Accompanying Accompanying the the decelerated decelerated dissolution dissolution and and chemical chemical reactions, reactions, heat evolution release also slowed down after P . Then the reaction of slag gradually moved into a steady period in which also slowed down after P2. Then the reaction of slag gradually moved into a steady period in which thethe dissolution dissolution and and reaction reaction of of slag slag was was controlled controlled by by a a diffusion diffusion process of hydroxide ions through thethe layers layers of reaction products.

Materials 2020, 13, 4796 15 of 19 Materials 2020, 13, x FOR PEER REVIEW 16 of 20

(a) (b)

FigureFigure 15. Micrographs,Micrographs, derived derived by by SEM, SEM, for for N4S0 N4S0 ( (aa)) and and N4S5.4 N4S5.4 ( (bb)) at at 28 28 days. N N and and S S indicate indicate

weightweight percentages of Na 2OO and and SiO SiO22 withwith respect respect to to slag, slag, respectively. respectively.

InIn sodium sodium silicate silicate activated activated slag slag paste, paste, undersat undersaturationuration degrees degrees of of the the modifying modifying elements elements (Ca, (Ca, Mg,Mg, and and K) K) and and Al Al were were initially initially very very large large in in the the beginning. beginning. The The undersaturation undersaturation was was able able to to provide provide sufficientsufficient energy energy to to remove remove activation activation energy energy barr barriersiers and and thus thus all all the the three three dissolution forms forms took took placeplace for for the dissolution of of these elements. This This re resultedsulted in in rapid rapid release release of of the modifying elements andand Al, Al, which which was was accompanied accompanied with with rapid rapid release release of of heat heat (reflected (reflected by by P P11, ,see see Figure Figure 3b).3b). However, However, thethe soluble soluble Si Si in in sodium sodium silicate silicate activator activator led led to to a a very very low low undersaturation undersaturation degree degree with with respect respect to to Si. Si. AsAs a result, the undersaturation with with respect respect to to Si Si could could not not supply supply sufficient sufficient energy energy to to remove remove the the activationactivation energy barriersbarriers forfor thethe dissolution dissolution of of Si Si via via formation formation of of vacancy vacancy islands islands and and etch etch pits pits on theon theleached leached surface surface layer. layer. The The dissolution dissolution of Si of on Si the on leached the leached surface surface layer occurredlayer occurred in regime in regime I. In regime I. In regimeI, the leached I, the leached surface surface layer dissolved layer dissolved at a low at rate a lo sincew rate the since dissolution the dissolution was limited was limited to the stepto the retreat. step retreat.Therefore, Therefore, the modifying the modifying elements elements and Aldissolved and Al disso fasterlved than faster Siin than the Si leached in the leached surface layer.surface This layer. led Thisto a continuousled to a continuous growth of growth the leached of the surface leached layer, surface which layer, decreased which decreased the diffusion the ratediffusion of modifying rate of modifyingelements. This,elements. therefore, This, slowedtherefore, down slowed the releasedown th ofe modifying release of modifying elements. Aselements. a result, As the a reactionresult, the of reactionslag came of intoslag ancame induction into an period.induction period. InIn the the meanwhile, meanwhile, the the slowly slowly released released Ca Ca and and Al Al during during the the induction induction period period gradually gradually reacted reacted withwith soluble soluble Si Si to to produce produce reaction reaction products, products, which which progressively progressively consumed consumed the the soluble soluble elements, elements, in particularin particular of ofthe the solubl solublee Si. Si. The The consumption consumption of of soluble soluble elem elementsents increased increased the the undersaturation. undersaturation. When thethe undersaturationundersaturation with with respect respect to to Si increasedSi increased to provide to provide suffi cientsufficient energy, energy, the activation the activation energy energybarrier barrier could be could removed be removed and the and dissolution the dissolution of Si could of Si takecould place take via place formation via formation of etch of pits etch on pits the onleached the leached surface surface layer. This layer. decreased This decreased the thickness the thickness of the leached of the surfaceleached layer, surface which layer, then which increased then increasedthe diffusion the ratediffusion of modifying rate of modifying elements through elements the through leached the surface leached layer. surface As a layer. result, As the a dissolutionresult, the dissolutionof unreacted of slag unreacted was accelerated. slag was The accelerated. dissolution The of the dissolution leached surface of the layer leached and unreactedsurface layer slag thenand unreactedreleased modifying slag then elements, released Al modifying and Si into elements solution., ThisAl and increased Si into the solution. concentrations This increased of modifying the concentrationselements, Al and of Simodifying towards saturation elements, orAl oversaturation. and Si towards Then saturation reaction or productsoversaturation. grew rapidly. Then reaction The fast productsgrowth of grew reaction rapidly. products The reflected fast growth intensive of reaction reactions, products as characterized reflected by intensive P2 in Figure reactions,3b. as characterizedIn addition by to P2 slag in Figure grains 3b. the soluble Si in solution could also act as nucleation sites [57,58]. This led to formationIn addition of reactionto slag grains products the soluble not only Si aroundin soluti slagon could grains also but act also as nucleation in the solution sites space[57,58]. [53 This,54] led(as to seen formation in Figure of 15reactionb). With products the continuous not only around formation slag of grains reaction but also products, in the solution the microstructure space [53,54] of (assodium seen silicatein Figure activated 15b). With paste the became continuous denser. formation As a result, of reaction more and products, more di theffusion microstructure paths of OH of− sodiumfrom solution silicate to activated unreacted paste slag became were blocked. denser. ThenAs a theresult, dissolution more and of more slag anddiffusion reactions paths of solubleof OH- fromelements solution moved to intounreacted the steady slag periodwere blocked. controlled Then by athe di ffdissolutionusion process. of slag Therefore, and reactions the heat of evolution soluble elementsrate decreased moved gradually into the after steady P2. period controlled by a diffusion process. Therefore, the heat evolution rate decreased gradually after P2.

Materials 2020, 13, 4796 16 of 19

5. Conclusions This study used isothermal calorimetry to monitor the reaction kinetics of alkali-activated slag pastes and carried out a dissolution test to investigate the role of soluble Si in the dissolution of slag in alkaline solution. The origin of the induction period and action mechanism of soluble Si on the dissolution of slag were interpreted explicitly for the ﬁrst time in this study. The main conclusions can be drawn as follows:

1. The heat release rate curves of alkali-activated slag paste depended on the presence of soluble Si in alkaline activator. For NaOH activated slag paste, three reaction periods were identiﬁed according to P1 and P2. For sodium silicate activated slag paste, one more reaction period was found between P1 and P2, i.e., the induction period. 2. The dissolution test revealed that the soluble Si in alkaline solution slowed down the dissolution of slag. The action mechanism of soluble Si, Ca and Al in the dissolution or reaction of slag was well interpreted by the dissolution theory-based mechanism. This demonstrates that the dissolution theory-based mechanism is applicable for describing the reaction process of alkali-activated slag paste, particularly for understanding the induction period. 3. In NaOH activated slag paste, a large undersaturation degree resulted in rapid dissolution of slag, leading to no noticeable induction period. In sodium silicate activated slag paste, the undersaturation with respect to Si was very low that it could not supply suﬃcient energy to remove the activation energy barriers for the formation of vacancy islands and etch pits on the surface of slag. This retarded the dissolution of slag, as a result of which, an induction period occurred.

Author Contributions: Y.Z. designed and performed the experiments, analyzed the data and wrote the paper. G.Y. supervised the project, analyzed the data, commented and edited the paper. All authors have read and agreed to the published version of the manuscript. Funding: The first author would like to gratefully acknowledge the China Scholarship Council (the Grant Number 201406160086) and the Fundamental Research Funds for the Central Universities (HUST: 2020kfyXJJS038) for the financial support in this work. Acknowledgments: A special note of appreciation goes to Klaas van Breugel (Delft University of Technology, The Netherlands) for his insightful comments and valuable suggestions. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Wang, S.-D.; Pu, X.-C.; Scrivener, K.; Pratt, P. Alkali-activated slag cement and concrete: A review of properties and problems. Adv. Cem. Res. 1995, 7, 93–102. [CrossRef] 2. Roy, D.M. Alkali-activated cements opportunities and challenges. Cem. Concr. Res. 1999, 29, 249–254. [CrossRef] 3. Van Deventer, J.S.J.; Provis, J.L.; Duxson, P. Technical and commercial progress in the adoption of geopolymer cement. Miner. Eng. 2012, 29, 89–104. [CrossRef] 4. Pacheco-Torgal, F.; Castro-Gomes, J.; Jalali, S. Alkali-activated binders: A review: Part 1. Historical background, terminology, reaction mechanisms and hydration products. Constr. Build. Mater. 2008, 22, 1305–1314. [CrossRef] 5. Awoyera, P.; Adesina, A. A critical review on application of alkali activated slag as a sustainable composite binder. Case Stud. Constr. Mater. 2019, 11, e00268. [CrossRef] 6. Duxson, P.; Provis, J.L.; Lukey, G.C.; Van Deventer, J.S. The role of inorganic polymer technology in the development of ‘green concrete’. Cem. Concr. Res. 2007, 37, 1590–1597. [CrossRef] 7. Yang, K.-H.; Song, J.-K.; Song, K.-I. Assessment of CO2 reduction of alkali-activated concrete. J. Clean. Prod. 2013, 39, 265–272. [CrossRef] 8. Fernández-Jiménez, A.; Palomo, J.; Puertas, F. Alkali-activated slag mortars: Mechanical strength behaviour. Cem. Concr. Res. 1999, 29, 1313–1321. [CrossRef] Materials 2020, 13, 4796 17 of 19

9. Pacheco-Torgal, F.; Abdollahnejad, Z.; Camões, A.F.; Jamshidi, M.; Ding, Y. Durability of alkali-activated binders: A clear advantage over Portland cement or an unproven issue? Constr. Build. Mater. 2012, 30, 400–405. [CrossRef] 10. Bernal, S.A.; Provis, J.L. Durability of alkali-activated materials: Progress and perspectives. J. Am. Ceram. Soc. 2014, 97, 997–1008. [CrossRef] 11. Deir, E.; Gebregziabiher, B.S.; Peethamparan, S. Inﬂuence of starting material on the early age hydration kinetics, microstructure and composition of binding gel in alkali activated binder systems. Cem. Concr. Compos. 2014, 48, 108–117. [CrossRef] 12. Altan, E.; Erdo˘gan,S.T. Alkali activation of a slag at ambient and elevated temperatures. Cem. Concr. Compos. 2012, 34, 131–139. [CrossRef] 13. Haha, M.B.; Lothenbach, B.; Le Saout, G.; Winnefeld, F. Inﬂuence of slag chemistry on the hydration of

alkali-activated blast-furnace slag–Part II: Effect of Al2O3. Cem. Concr. Res. 2012, 42, 74–83. [CrossRef] 14. Gerstig, M.; Wadsö, L. A method based on isothermal calorimetry to quantify the influence of moisture on the hydration rate of young cement pastes. Cem. Concr. Res. 2010, 40, 867–874. [CrossRef] 15. Xu, Q.; Ruiz, J.M.; Hu, J.; Wang, K.; Rasmussen, R.O. Modeling hydration properties and temperature developments of early-age concrete pavement using calorimetry tests. Thermochim. Acta 2011, 512, 76–85. [CrossRef] 16. Bentz, D.P.; Sato, T.; De la Varga, I.; Weiss, W.J. Fine limestone additions to regulate setting in high volume fly ash mixtures. Cem. Concr. Compos. 2012, 34, 11–17. [CrossRef] 17. Siler, P.; Kratky, J.; De Belie, N. Isothermal and solution calorimetry to assess the effect of superplasticizers and mineral admixtures on cement hydration. J. Therm. Anal. Calorim. 2012, 107, 313–320. [CrossRef] 18. Gruyaert, E.; Robeyst, N.; De Belie, N. Study of the hydration of Portland cement blended with blast-furnace slag by calorimetry and thermogravimetry. J. Therm. Anal. Calorim. 2010, 102, 941–951. [CrossRef] 19. Han, F.; Liu, R.; Wang, D.; Yan, P. Characteristics of the hydration heat evolution of composite binder at different hydrating temperature. Thermochim. Acta 2014, 586, 52–57. [CrossRef] 20. Ravikumar, D.; Neithalath, N. Reaction kinetics in sodium silicate powder and liquid activated slag binders evaluated using isothermal calorimetry. Thermochim. Acta 2012, 546, 32–43. [CrossRef] 21. Gebregziabiher, B.S.; Thomas, R.; Peethamparan, S. Very early-age reaction kinetics and microstructural development in alkali-activated slag. Cem. Concr. Compos. 2015, 55, 91–102. [CrossRef] 22. Shi, C.J.; Day, R.L. A calorimetric study of early hydration of alkali-slag cements. Cem. Concr. Res. 1995, 25, 1333–1346. [CrossRef] 23. Gao, X.; Yu, Q.; Brouwers, H. Reaction kinetics, gel character and strength of ambient temperature cured alkali activated slag–fly ash blends. Constr. Build. Mater. 2015, 80, 105–115. [CrossRef] 24. Bernal, S.A.; Provis, J.L.; Rose, V.; De Gutierrez, R.M. Evolution of binder structure in sodium silicate-activated slag-metakaolin blends. Cem. Concr. Compos. 2011, 33, 46–54. [CrossRef] 25. Jin, Y.; Feng, W.; Zheng, D.; Dong, Z.; Cui, H.; Li, M.; Sang, G.; Tang, W. Study on the interaction mechanism between slags and alkali silicate activators: A hydration kinetics approach. Constr. Build. Mater. 2020, 250, 118900. [CrossRef] 26. Ouyang, X.; Ma, Y.; Liu, Z.; Liang, J.; Ye, G. Effect of the sodium silicate modulus and slag content on fresh and hardened properties of alkali-activated fly ash/slag. Minerals 2019, 10, 15. [CrossRef] 27. Xie, F.; Liu, Z.; Zhang, D.; Wang, J.; Huang, T.; Wang, D. Reaction kinetics and kinetics models of alkali activated phosphorus slag. Constr. Build. Mater. 2020, 237, 117728. [CrossRef] 28. Gebregziabiher, B.S.; Thomas, R.J.; Peethamparan, S. Temperature and activator effect on early-age reaction kinetics of alkali-activated slag binders. Constr. Build. Mater. 2016, 113, 783–793. [CrossRef] 29. Chen, W.; Brouwers, H. The hydration of slag, Part 1: Reaction models for alkali-activated slag. J. Mater. Sci. 2007, 42, 428–443. [CrossRef] 30. ASTM. Standard Practice for Measuring Hydration Kinetics of Hydraulic Cementitious Mixtures Using Isothermal Calorimetry; C1679-14; ASTM International: West Conshohocken, PA, USA, 2014. 31. Ye, G. Experimental Study and Numerical Simulation of the Development of the Microstructure and Permeability of Cementitious Materials. Ph.D. Thesis, Delft University of Technology, Delft, The Netherlands, 2003. 32. Zuo, Y.; Ye, G. Pore Structure Characterization of Sodium Hydroxide Activated Slag Using Mercury Intrusion Porosimetry, Nitrogen Adsorption, and Image Analysis. Materials 2018, 11, 1035. [CrossRef] Materials 2020, 13, 4796 18 of 19

33. Lothenbach, B. Thermodynamic equilibrium calculations in cementitious systems. Mater. Struct. 2010, 43, 1413–1433. [CrossRef] 34. Kulik, D.A.; Wagner, T.; Dmytrieva, S.V.; Kosakowski, G.; Hingerl, F.F.; Chudnenko, K.V.; Berner, U.R. GEM-Selektor geochemical modeling package: Revised algorithm and GEMS3K numerical kernel for coupled simulation codes. Comput. Geosci. 2013, 17, 1–24. [CrossRef] 35. Wagner, T.; Kulik, D.A.; Hingerl, F.F.; Dmytrieva, S.V. GEM-Selektor geochemical modeling package: TSolMod library and data interface for multicomponent phase models. Can. Mineral. 2012, 50, 1173–1195. [CrossRef] 36. Myers, R.J.; Bernal, S.A.; Provis, J.L. A thermodynamic model for C-(N-)A-S-H gel: CNASH_ss. Derivation and validation. Cem. Concr. Res. 2014, 66, 27–47. [CrossRef] 37. Myers, R.J.; Lothenbach, B.; Bernal, S.A.; Provis, J.L. Thermodynamic modelling of alkali-activated slag cements. Appl. Geochem. 2015, 61, 233–247. [CrossRef] 38. Provis, J.L.; Bernal, S.A. Geopolymers and related alkali-activated materials. Annu. Rev. Mater. Res. 2014, 44, 299–327. [CrossRef] 39. Lothenbach, B.; Gruskovnjak, A. Hydration of alkali-activated slag: Thermodynamic modelling. Adv. Cem. Res. 2007, 19, 81–92. [CrossRef] 40. Wang, S.-D.; Scrivener, K.L. Hydration products of alkali activated slag cement. Cem. Concr. Res. 1995, 25, 561–571. [CrossRef] 41. Richardson, I.G.; Brough, A.R.; Groves, G.W.; Dobson, C.M. The characterization of hardened alkali-activated blast-furnace slag pastes and the nature of the calcium silicate hydrate (C-S-H) phase. Cem. Concr. Res. 1994, 24, 813–829. [CrossRef] 42. Zuo, Y.; Nedeljković,M.; Ye, G. Coupled thermodynamic modelling and experimental study of sodium hydroxide activated slag. Constr. Build. Mater. 2018, 188, 262–279. [CrossRef] 43. Huanhai, Z.; Xuequan, W.; Zhongzi, X.; Mingshu, T. Kinetic study on hydration of alkali-activated slag. Cem. Concr. Res. 1993, 23, 1253–1258. [CrossRef] 44. Helgeson, H.C.; Kirkham, D.H.; Flowers, G.C. Theoretical prediction of the thermodynamic behavior of aqueous electrolytes by high pressures and temperatures; IV, Calculation of activity coefficients, osmotic coefficients, and apparent molal and standard and relative partial molal properties to 600 degrees C and 5 kb. Am. J. Sci. 1981, 281, 1249–1516. 45. Ben Haha, M.; Le Saout, G.; Winnefeld, F.; Lothenbach, B. Influence of activator type on hydration kinetics, hydrate assemblage and microstructural development of alkali activated blast-furnace slags. Cem. Concr. Res. 2011, 41, 301–310. [CrossRef] 46. Ye, H.; Radlińska,A. Quantitative analysis of phase assemblage and chemical shrinkage of alkali-activated slag. J. Adv. Concr. Technol. 2016, 14, 245–260. [CrossRef] 47. Oelkers, E.H.; Gislason, S.R. The mechanism, rates and consequences of basaltic glass dissolution: I. An experimental study of the dissolution rates of basaltic glass as a function of aqueous Al, Si and oxalic acid concentration at 25 C and pH = 3 and 11. Geochim. Cosmochim. Acta 2001, 65, 3671–3681. [CrossRef] 48. Oelkers, E.H. General kinetic description of multioxide silicate mineral and glass dissolution. Geochim. Cosmochim. Acta 2001, 65, 3703–3719. [CrossRef] 49. Gislason, S.R.; Oelkers, E.H. Mechanism, rates, and consequences of basaltic glass dissolution: II. An experimental study of the dissolution rates of basaltic glass as a function of pH and temperature. Geochim. Cosmochim. Acta 2003, 67, 3817–3832. [CrossRef] 50. Oelkers, E.H.; Schott, J.; Devidal, J.-L. The effect of aluminum, pH, and chemical affinity on the rates of aluminosilicate dissolution reactions. Geochim. Cosmochim. Acta 1994, 58, 2011–2024. [CrossRef] 51. Duxson, P.; Provis, J.L. Designing precursors for geopolymer cements. J. Am. Ceram. Soc. 2008, 91, 3864–3869. [CrossRef] 52. Lasaga, A.C.; Luttge, A. Variation of Crystal Dissolution Rate Based on a Dissolution Stepwave Model. Science 2001, 291, 2400–2404. [CrossRef] 53. Dove, P.M.; Han, N.; De Yoreo, J.J. Mechanisms of classical crystal growth theory explain quartz and silicate dissolution behavior. Proc. Natl. Acad. Sci. USA 2005, 102, 15357–15362. [CrossRef] 54. Juilland, P.; Gallucci, E.; Flatt, R.; Scrivener, K. Dissolution theory applied to the induction period in alite hydration. Cem. Concr. Res. 2010, 40, 831–844. [CrossRef] Materials 2020, 13, 4796 19 of 19

55. Suraneni, P.; Palacios, M.; Flatt, R.J. New insights into the hydration of slag in alkaline media using a micro-reactor approach. Cem. Concr. Res. 2016, 79, 209–216. [CrossRef] 56. Nicoleau, L.; Schreiner, E.; Nonat, A. Ion-specific effects influencing the dissolution of tricalcium silicate. Cem. Concr. Res. 2014, 59, 118–138. [CrossRef] 57. Lloyd, R.R.; Provis, J.L.; van Deventer, J.S.J. Microscopy and microanalysis of inorganic polymer cements. 2: The gel binder. J. Mater. Sci. 2009, 44, 620–631. [CrossRef] 58. Rees, C.A.; Provis, J.L.; Lukey, G.C.; van Deventer, J.S. The mechanism of geopolymer gel formation investigated through seeded nucleation. Colloids Surf. A Physicochem. Eng. Asp. 2008, 318, 97–105. [CrossRef]

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional aﬃliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).