Influence of Polycarboxylate Superplasticizers with Different Functional Units on the Early Hydration of C3A-Gypsum

materials

Article Inﬂuence of Polycarboxylate Superplasticizers with Different Functional Units on the Early Hydration of C3A-Gypsum

Kuangyi Hu 1,2 and Zhenping Sun 1,2,* 1 School of Materials Science and Engineering, Tongji University, Shanghai 200092, China; [email protected] 2 Key Laboratory of Advanced Civil Engineering Materials of Ministry of Education, Tongji University, Shanghai 200092, China * Correspondence: [email protected]

Received: 3 February 2019; Accepted: 4 April 2019; Published: 6 April 2019

Abstract: Inﬂuence of polycarboxylate superplasticizers (PCs) with different functional units on the hydration heat and hydration products of C3A-gypsum was investigated. Three kinds of PCs with different monomers were discussed. It was seen that PCs mainly shortened the induction stage of C3A-gypsum hydration, and the amount of remaining gypsum was related to the duration of the induction stage. The second heat ﬂow peak of the sample with PCs was higher than that of the blank. Moreover, PC intercalation occurred during the hydration.

Keywords: polycarboxylate superplasticizers; C3A-gypsum; functional units; hydration

1. Introduction

Tricalcium aluminate (3CaO·Al2O3,C3A) is one of the essential components of Portland cement, which inﬂuences the early properties of cement pastes. Generally, calcium sulfate (CaSO4·H2O, gypsum) is added during the grinding period to adjust workability of the cement pastes before the ﬁnal set. Therefore, the early hydration of cement depends on the C3A-gypsum reaction [1]. C3A reacts with water directly to produce calcium aluminate hydrates combined with different amounts of water, and all of these calcium aluminate hydrates (C4AH13/C4AH19 and C2AH8, hydroxy-AFm) convert into hydrogarnet (C3AH6, katoite) eventually. The equation for C3A hydration is shown below (Equation (1)):

R.T. 2C3A + 27H → C4AH19(C4AH13 when R.H. < 85%) + C2AH8 → C3AH6 (1)

where

R.H.—relative humidity, R.T.—room temperature, C—CaO,

A—Al2O3, H—H2O.

C3A can also react with gypsum in water to form ettringite (C6A$3H32, AFt), which is presented as Equation (2):. C3A + 3C$H2 + 26H → C6A$3H32, (2) where:

Materials 2019, 12, 1132; doi:10.3390/ma12071132 www.mdpi.com/journal/materials Materials 2019, 12, 1132 2 of 18

$—SO3.

C3A turns AFt into AFm (C4A$H12, kuzelite) when gypsum runs out, as in Equation (3):

C3A + C6A$3H32 + 4H → 3C4A$H12, (3)

Some researchers have studied the mechanisms and hydration process of C3A-gypsum systems [1–7]. Unlike in previous studies [2–5] showing that AFt or gel-like hydroxy-AFm could provide a substantial barrier to slow down the hydration rate of C3A in the presence of gypsum, the study of Minard et al. [6] conjectured that the early deceleration of C3A hydration might be caused by adsorption of sulfate ions on the surface dissolution sites of C3A particles. Based on the conclusions above, Quennoz and Scrivener [1] obtained new results showing that the rapid reaction after the exhaustion of gypsum was influenced by the surface area of C3A particles. Nowadays, polycarboxylate superplasticizers (PCs) have been widely used in real concrete technology due to their favorable dispersibility. With their carboxylate backbone, PCs can be adsorbed on the surface of cement particles such that the steric hindrance provided by polyethylene oxide (EO) side chains can prevent the aggregation of particles during hydration. Hence, PCs influence not only the rheological behavior but also the hydration behavior of cement paste [8–12] due to its prolonged induction period of the early hydration of cement [8]. Thus, PCs may also have an impact on the C3A-gypsum early hydration. The morphology and chemistry of C3A-gypsum-PCs hydrates were also studied [13–15]. Merlini et al. [13] found that superplasticizers had an effect on the total amount of formed ettringite in the C3A-gypsum system, and available sulfates in the system could be reduced via the formation of the organo-mineral phase. This kind of phase was determined with the dissolved sulfate concentration in the pore solution [14]. Moreover, Dalas et al. [15] demonstrated that the surface area of AFt was related to the electrostatic charge and dosage of PCs. While studies have been conducted on the adsorption behavior between superplasticizers and C3A clinkers [16,17], Alonso and Puertas [16] reported that the PCs molecular structure and the sulfate concentration in solution were the important factors that influence the adsorption behavior of PCs on the cubic C3A. Myers et al. [17] showed that the diffuse layer of C3A could weaken the dissolution-inhibition effects of superplasticizers. Nonetheless, research has yet to reconcile the microscopic result and the macroscopic evidence of C3A-gypsum hydration with PCs, as well as those without. There is a lot that remains unknown about the effects of PCs, especially the effect of its different functional units on the C3A-gypsum hydration process. This paper proposes to gain further insight into the effects of PCs with different functional units on the hydration heat and hydration products of C3A-gypsum system, which could be helpful for understanding the working mechanism of PCs in the early hydration of cement.

2. Experimental

2.1. Minerals (C3A/Gypsum)

C3A was provided by Dumaite (DMT, Shanghai, China), where X-ray diffraction (XRD) and Rietveld analyses (Topas Academic, Version 5) showed that the contained free-lime (CaO) was less than 1 wt%. The calcium sulfate dihydrate from Sinopharm Chemical Reagent (Shanghai, China) was also used in this study. Particle size distribution of minerals was analyzed with the LS 230 nanolaser particle size analyzer (Beckman Coulter, Brea, CA, USA), which is listed in Table1. Materials 2019, 12, x FOR PEER REVIEW 3 of 17 Materials 2019, 12, 1132 3 of 18 2.2. PC Superplasticizers (PCs)

Three PCs (PC1, PC2, andTable PC3) 1. wereParticle synthesized size distribution via free-radical of minerals. polymerization according to our previous work [18] and were employed in this study. The chemical structure of PCs is shown in Figure 1. VC (Vitamin C, Sinopharm Chemical Reagent,Diameter Shanghai, (µm) China)-H2O2 (Aladdin, Shanghai, Minerals China) were used as initiators, and TGA (Thioglycollicd10 d50 acid, Sinopharmd90 dm Chemical Reagent, Shanghai, China) was used as the chain transfer agent to control the backbone length of PCs. Dosages of VC, C3A 3.09 6.88 15.92 8.83 H2O2, and TGA were 0.15 wt%,Gypsum 1.00 wt%, and 8.10 0.35 wt% 20.16 of the 38.00 total weight 21.79 of monomers, respectively.

TPEG2400Note: d 10(Isoprenyl, d50, and d90 representoxy poly(ethylene the particle size belowglycol) which ether the volumewith percentmolecular is 10%, 50%,weight and 90%,of 2400, respectively; Levima, Tengzhou,dm is the China) mean particle as macromonomer size. was diluted into 40 wt% water solution in a 250 mL four-neck round-bottom flask equipped with a constant stirrer and was heated in the 45 °C water bath. AA (Acrylic2.2. PC Superplasticizers acid, Sinopharm (PCs) Chemical Reagent, Shanghai, China) as a carboxylate monomer was diluted into Three40 wt% PCs solution (PC1, PC2, in beaker and PC3) A. wereAM synthesized(Acrylamide, via Aladdin, free-radical Shanghai, polymerization China) or according AMPS (2- to acrylamido-2-methylpropanesulfonicour previous work [18] and were employed acid, Aladdin, in this study. Shanghai, The chemical China) as structure a functional of PCs monomer is shown was in diluted into 40 wt% solution in beaker B. VC and TGA were mixed and diluted into a 40 wt% solution Figure1. VC (Vitamin C, Sinopharm Chemical Reagent, Shanghai, China)-H 2O2 (Aladdin, Shanghai, inChina) beaker were C. H used2O2 aswas initiators, added into and TPEG2400 TGA (Thioglycollic solution all acid, at once Sinopharm when TPEG Chemical solution Reagent, was heated Shanghai, up toChina) 45 °C. was Then used solution as the in chain beakers transfer A, B, and agent C we tore control fed dropwise the backbone with peristaltic length of pumps PCs. Dosages into TPEG of solution for 3.0 h, 3.0 h, and 3.5 h respectively. The mixture in flask was stirred at 45 °C for an extra VC, H2O2, and TGA were 0.15 wt%, 1.00 wt%, and 0.35 wt% of the total weight of monomers, hourrespectively. in order TPEG2400to react completely. (Isoprenyl Gel oxy permeation poly(ethylene chromatography glycol) ether (GPC) with anal molecularysis of the weight synthesized of 2400, PCsLevima, was Tengzhou,performed China)using Waters as macromonomer Alliance 2695 wasinstrument diluted (Waters, into 40 wt% Eschborn, water Germany). solution in Details a 250 mL of thefour-neck monomer round-bottom ratios and the ﬂask molecular equipped characteristics with a constant are shown stirrer in and Table was 2. heated in the 45 ◦C water bath. AA (Acrylic acid, Sinopharm Chemical Reagent, Shanghai, China) as a carboxylate monomer was diluted into 40 wt%Table solution 2. Monomers in beaker molar A. AM ratio (Acrylamide, and molecular Aladdin, data of PCs. Shanghai, China) or AMPS (2-acrylamido-2-methylpropanesulfonic acid, Aladdin, Shanghai, China) as a functional monomer was Sample The Molar Ratio of Monomers Molecular Characteristics diluted into 40 wt% solution in beaker B. VC and TGA were mixed and diluted into a 40 wt% solution AA AM AMPS TPEG2400 Mn (Da) Mw (Da) PDI in beaker C. H2O2 was added into TPEG2400 solution all at once when TPEG solution was heated up to 45PC1◦C. Then solution5 in beakers0 A,0 B, and C were1 fed dropwise48600 with peristaltic107400 pumps into2.21 TPEG ◦ solutionPC2 for 3.0 h,1 3.0 h, and4 3.5 h respectively.0 The1 mixture in ﬂask35740 was stirred67840 at 45 C for1.90 an extra hour in order to react completely. Gel permeation chromatography (GPC) analysis of the synthesized PC3 1 0 4 1 36030 83110 2.31 PCs was performed using Waters Alliance 2695 instrument (Waters, Eschborn, Germany). Details of Note: Mn means number average molecular mass, Mw means weight average molecular mass, and the monomer ratios and the molecular characteristics are shown in Table2. PDI means polydisperse index.

Figure 1. Molecular architecture of PCs.

Materials 2019, 12, 1132 4 of 18

Table 2. Monomers molar ratio and molecular data of PCs.

The Molar Ratio of Monomers Molecular Characteristics Sample AA AM AMPS TPEG2400 Mn (Da) Mw (Da) PDI PC1 5 0 0 1 48,600 107,400 2.21 PC2 1 4 0 1 35,740 67,840 1.90 PC3 1 0 4 1 36,030 83,110 2.31

Note: Mn means number average molecular mass, Mw means weight average molecular mass, and PDI means polydisperse index.

2.3. Preparation of C3A-Gypsum Samples

Samples of C3A in the absence and presence of various amount of calcium sulfate were prepared for study. Mix ratios were presented in Table3. The powders were dry-mixed manually for 10 min in a mortar before mixed with deionized water or PC solutions.

Table 3. Summary of the C3A-gypsum samples.

Samples C3A (g) Gypsum (g) Deionized Water/PC Solutions

C3A_0.5G 1 (3.7 mmol) 0.385 (1.9 mmol) 10 g deionized water C3A_1G 1 (3.7 mmol) 0.770 (3.7 mmol) 10 g deionized water C3A_2G 1 (3.7 mmol) 1.540 (7.4 mmol) 10 g deionized water C3A_3G 1 (3.7 mmol) 2.310 (11.1 mmol) 10 g deionized water C3A_5G 1 (3.7 mmol) 2.310 (18.5 mmol) 10 g deionized water C3A_0.06PC1 1 (3.7 mmol) 0.770 (3.7 mmol) 10 g 0.06% PC1 solution C3A_0.12PC1 1 (3.7 mmol) 0.770 (3.7 mmol) 10 g 0.12% PC1 solution C3A_0.24PC1 1 (3.7 mmol) 0.770 (3.7 mmol) 10 g 0.24% PC1 solution C3A_0.06PC2 1 (3.7 mmol) 0.770 (3.7 mmol) 10 g 0.06% PC2 solution C3A_0.12PC2 1 (3.7 mmol) 0.770 (3.7 mmol) 10 g 0.12% PC2 solution C3A_0.24PC2 1 (3.7 mmol) 0.770 (3.7 mmol) 10 g 0.24% PC2 solution C3A_0.06PC3 1 (3.7 mmol) 0.770 (3.7 mmol) 10 g 0.06% PC3 solution C3A_0.12PC3 1 (3.7 mmol) 0.770 (3.7 mmol) 10 g 0.12% PC3 solution C3A_0.24PC3 1 (3.7 mmol) 0.770 (3.7 mmol) 10 g 0.24% PC3 solution

2.4. Hydration Heat

The hydration heat of C3A-gypsum pastes was measured using the TAM Air isothermal calorimeter (TA instrument, New Castle, PA, USA) at 20 ◦C. Pastes were prepared outside the calorimeter and were introduced in the calorimeter immediately after mixing (less than 2 min).

2.5. XRD Analysis The traces of sulfur component in PC3 samples might contaminate the platinum modules during heating process in TGA analysis. Furthermore, accurate crystallographic information of some C3A-gypsum hydrates was unknown for QXRD (Quantitative X-ray Diffraction)/Rietveld analysis. Thus, it was not possible to calculate absolute amounts of phases in this study. Semi-quantitative method could be an alternative option to reveal the trend changes of phases. Samples listed in Table3 were mixed separately in sealed containers and cured at (20 ± 1) ◦C. Ethyl alcohol was added into containers immediately after 24 h of hydration to stop the reaction. Then samples were dried at 45 ◦C for 48 h in the vacuum drying oven (Shengke, Shanghai, China) before the test. XRD patterns were obtained using the Bruker AXS D8 (Karlsruhe, Germany) with a RINT2000 vertical goniometer operating at 40 kV and 250 mA. A CuKα source with a 0.3 mm slit was used. The scan was performed at the 2θ angles between 5◦ and 40◦ with an increment of 0.02◦. As for peak area (in counts) of phases, the specific diffraction peak was fitted with a Gaussian function to make it smoother first. Two end points of the smoothed curve were used as the baseline. Then, the peak area between the smoothed curve and baseline was calculated by integrating over Materials 2019, 12, 1132 5 of 18 the specific 2θ angles range. For ettringite, the preferred 2θ angles were between 8.8◦ and 9.4◦. The preferred 2θ angles of katoite were between 39.0◦ and 39.6◦. For kuzelite, the preferred 2θ angles wereMaterials between 2019, 12 9.6, x FOR◦ and PEER 10.2 REVIEW◦. Finally, for gypsum, the preferred 2θ angles were between 11.4◦ and5 of 17 12.0◦. 3. Results and Discussion 3. Results and Discussion 3.1. Early Hydration of C3A-Gypsum System

3.1. Early Hydration of C3A-Gypsum System To begin with, the hydration of C3A in the presence of various amount of gypsum was

investigated.To begin with, The thehydration hydration heat of Cevolution3A in the curves presence of ofC various3A with amountthe gypsum/C of gypsum3A molar was investigated. ratios (G/C) Thebetween hydration 0.5 and heat 5 evolutionduring 24 curvesh are plotted of C3A on with Figu there gypsum/C 2. Two exothermic3A molar peaks ratios were (G/C) observed between when 0.5 andthe 5 G/C during ratio 24 was h are less plotted than 2. on It Figurewas shown2. Two that exothermic the first one peaks was were extremely observed sharp when and the the G/C second ratio one waswas less relatively than 2. Itbroad. was shown However, that thethere ﬁrst was one only was extremelyone exothermic sharp andpeak the during second the one 24 was h when relatively more broad.gypsum However, was added. there Therefore, was only one in accordance exothermic peakwith duringprevious the research 24 h when [1,6], more this gypsum phenomenon was added. is due Therefore,to the higher in accordance gypsum dosage with previous and liquid research to solid [1 (L/S),6], this ratio. phenomenon is due to the higher gypsum dosage and liquid to solid (L/S) ratio.

450

400 70 C3A_0.5G 60 350 C3A_1G

C A_2G A) 3 A) 3 50 3 300 C3A_3G 40 C A_5G 250 3 30

200

Heat flow (mW/g C 20

150 10

Heat flow (mW/gC 0 100 0246810 Time (h) 50

0 0 5 10 15 20 25 Time (h)

FigureFigure 2. 2.Heat Heat evolution evolution curves curves of of C 3CA3A with with various various amounts amounts of of gypsum gypsum during during the the 24-h 24-h hydration. hydration.

FigureFigure3 showed3 showed the the XRD XRD results results of of the the hydrates hydrates after after the the 24-h 24-h hydration hydration of of C 3CA-gypsum.3A-gypsum. ExcessiveExcessive gypsum gypsum was was found found when when the the G/C G/C ratio ratio was was higher higher than than 2. 2. AFm AFm was was detected detected in in all all samples samples eveneven when when there there was was excessive excessive gypsum gypsum in in the the system. system. However, However, this this result result is is contradictory contradictory to to what what wewe have have known known from from Reference Reference [2 [2]] showing showing that that AFm AFm forms forms only only when when gypsum gypsum is is exhausted, exhausted, as as we we summarizedsummarized in in the the Introduction Introduction section. section. Based on the shape of the hydration heat curves, the hydration age can be divided into three stages, which are the dissolution–crystallization stage, the induction stage, and the transformation stage with the exhaustion of gypsum, respectively.

MaterialsMaterials 20192019, 12, 12, x, FOR 1132 PEER REVIEW 6 of6 of17 18

4 1 - Ettringite C3A_5G 2 - Katoite C A_3G 3 - Kuzelite 3 4 - Gypsum 4 C A_2G 4 3 C3A_1G

C3A_0.5G 4 1 2 11 1 1 1 21 1 4 3 21 32 3 1 1 2 1 2 3 14 4 33

5 10152025303540 2θ (°)

FigureFigure 3. 3.XRDXRD patterns patterns of phases of phases obtained obtained from from C3A Cwith3A withvarious various amounts amounts of gypsum of gypsum after the after 24-h the hydration.24-h hydration. 3.1.1. Dissolution–Crystallization Stage Based on the shape of the hydration heat curves, the hydration age can be divided into three stages,The which dissolution–crystallization are the dissolution–crystallization stage includes stage, the firstthe induction exothermic stage, peak. and After the C transformation3A-gypsum is in stagecontact with with the water,exhaustion hydroxy-AFm of gypsum, and respectively. AFt are both precipitated due to the different dissolution rates of C3A and gypsum. As can be seen from Figure2, the first peak width of each curve is nearly the same, 3.1.1.which Dissolution–Crystallization means the dosage of gypsum Stage had no impact on the duration of this stage. Different heights of the first peaks could be caused by the measuring error of the C A weight, the different interval between The dissolution–crystallization stage includes the first exothermic3 peak. After C3A-gypsum is in contactsample with mixing water, time hydroxy-AFm and calorimeter and measuringAFt are both time, precipitated and the variousdue to the reaction different heats dissolution of Equations rates (1) and (2). of C3A and gypsum. As can be seen from Figure 2, the first peak width of each curve is nearly the same, Becausewhich means of the the low dosage solubility of ofgypsum gypsum, had hydroxy-AFm no impact on was thefirst duration precipitated of this bystage. the dissolutionDifferent of C A. Furthermore, the formation of ettringite was the reason why consumption rate of sulfate ion heights3 of the first peaks could be caused by the measuring error of the C3A weight, the different intervalwas higher between than sample the dissolution mixing time rate and of gypsumcalorimeter [6]. measuring Therefore, time, the dissolution and the various of C3 reactionA could heats be the ofmain Equations factor (1) that and controlled (2). the reaction rate during this stage. The specific surface area of C3A could haveBecause determined of the low the hydrationsolubility of rate gypsum, of C3A-gypsum hydroxy-AFm system was [ 1first,6]. precipitated Some researchers by the believeddissolution that sulfate ions could combine with hydroxy-AFm rather than C A, which could explain the time interval of C3A. Furthermore, the formation of ettringite was the reason3 why consumption rate of sulfate ion between the dissolution of gypsum and the formation of ettringite. was higher than the dissolution rate of gypsum [6]. Therefore, the dissolution of C3A could be the main factor that controlled the reaction rate during this stage. The specific surface area of C3A could 3.1.2. Induction Stage have determined the hydration rate of C3A-gypsum system [1,6]. Some researchers believed that sulfateWhen ions could gypsum combine is not with overdosed, hydroxy-AFm unhydrated rather Cthan3A wouldC3A, which consume could AFt, explain and the as time a result, interval AFm betweenwould bethe precipitated, dissolution of and gypsum two thermal and the peaksformation would of ettringite. be found in the hydration heat flow curves as well. The interval between the two peaks is commonly known as the induction stage when the 3.1.2.hydration Induction rate isStage temporarily low and would be accelerated again in the following hours. As illustrated in Figure2, only sample C 3A_1G had a complete induction period during the 24-h hydration. When gypsum is not overdosed, unhydrated C3A would consume AFt, and as a result, AFm The formed hydrates could cover the whole surface of C A particles, thus they inhibit the would be precipitated, and two thermal peaks would be found in 3the hydration heat flow curves as hydration of C A, which was the primary explanation for the induction stage in past decades [2–5]. well. The interval3 between the two peaks is commonly known as the induction stage when the However, the morphology of ettringite was considered unlikely to block the ion exchange [5]. hydration rate is temporarily low and would be accelerated again in the following hours. As Minard et al. [6] found that the hydroxy-AFm “sheet” could be present on the surface of C3A after illustrated in Figure 2, only sample C3A_1G had a complete induction period during the 24-h a 3-min hydration of C A-gypsum. However, platelets of the same precipitation was not observed hydration. 3 after 5-min hydration of C3A-hemihydrate [7]. Therefore, the hydrates after dissolution–crystallization The formed hydrates could cover the whole surface of C3A particles, thus they inhibit the stage may not be the critical reason for a slowdown of the hydration of C3A-gypsum. Afterwards, the hydration of C3A, which was the primary explanation for the induction stage in past decades [2–5]. However, the morphology of ettringite was considered unlikely to block the ion exchange [5]. Minard

Materials 2019, 12, 1132 7 of 18

2− new kinetics was put forward and adsorption of SO4 onto etch pit sites of C3A particles was thought to be the main reason for the control of the reaction during the induction period [1,6]. Considering this result, an important characteristic at the end of the induction stage was the complete disappearance of sulfate ion [6]. It is clear that the formation of AFt continuously occurred with a slow rate during the whole stage. Nevertheless, when the G/C ratio increased from 0.5 to 5, the induction stage only existed in a range where gypsum was insufﬁcient, as shown in Figure2, which means the adsorption of sulfate ions caused a dynamic equilibrium between the dissolution of C3A and the consumption of sulfate ions in solution when gypsum was ample, AFt formed with an almost constant rate until C3A ran out, and no other visible exothermic peaks could be observed except the ﬁrst one. The exhaustion of sulfate ions in the media could induce the desorption of species in order to maintain the equilibrium mentioned above. Therefore, the dissolution of C3A increased again and a second hydration exothermic peak was observed. The amount of sulfate ions hence determined the duration of this stage. Adsorbed sulfate ions on the active site of C3A could reduce the dissolution rate of C3A as well.

3.1.3. Transformation Stage

After the gypsum depletion, C3A could be consumed by AFt and the reaction in Equation (3) would proceed during the transformation stage, which included the second hydration exothermic peak. As depicted in Figure2, with the increases of sulfate ions, the second exothermic peak of samples was delayed and reduced when gypsum was exhausted. Regarding this stage, the work of Quennoz et al. [1] reported that the remaining site for AFm nucleation on the surface of C3A controlled the acceleration part, and the space available for growth was the main factor impacting the slope of the deceleration part. Also, the result of Kirchheim et al. [19] proved that only the reaction Equation (3) existed during this stage instead of the concurrence of Equations (1) and (3).

3.2. Hydration of C3A-Gypsum in the Presence of PCs Two kinds of functional units are discussed in this study. PC2 contains AM with an amide unit as the ﬁrst functional unit. PC3 contains AMPS with the same molar ratio of the amide unit and sulfo unit. The sulfo unit is the second functional unit. PC1 was synthesized with only carboxylate monomers.

3.2.1. Effects of PC with Only Carboxylate Monomers

The hydration heat of C3A-gypsum mixed with different concentration of PC1 solutions within 24 h is demonstrated in Figure4. As depicted in Figure4, low concentration of PC1 could enhance the transformation of AFt not only by cutting the induction stage but by increasing the second heat flow peak compared to the sample without PCs. With the increased concentration of PC1, the second heat flow peak first increased and then decreased, and the induction stage was shortened as well. Samples with 0.12% and 0.24% concentration of PC1 had a subtle difference on the second heat flow peak, but the slopes of the deceleration part were not the same. The XRD patterns of phases after the 24-h hydration are indicated in Figure5a. It was noticed that the 2 θ value of hydrates formed via samples mixed with PC1 solution had little shift compared with the same hydrates produced using the blank sample. Carboxylate units could be adsorbed on the surface of positive charge particles, which included C3A and hydrates such as hydroxy-AFm [10,14,16]. Therefore, the peak shift may reveal that the PC intercalation [13,14] occurred during C3A-gypsum hydration. The area of phase peaks in Figure5a is shown in Figure5b. With the increased dosage of PC1, more ettringite remained with a lower gypsum cost. Samples with 0.06% concentration of PC1 had the highest production of AFm, but the product of hydroxy-AFm was the least. Materials 2019, 12, x FOR PEER REVIEW 8 of 17

Materialswith a 2019lower, 12 ,gypsum x FOR PEER cost. REVIEW Samples with 0.06% concentration of PC1 had the highest production8 of 17of AFm, but the product of hydroxy-AFm was the least. with a lower gypsum cost. Samples with 0.06% concentration of PC1 had the highest production of AFm,Materials but2019 the, 12 product, 1132 of hydroxy-AFm was the least. 8 of 18

400 C3A_1G 50 C3A_0.06PC1 400 A) 3 A) C C33A_1GA_0.12PC1 40

3 50 300 C C33A_0.06PC1A_0.24PC1 30 A) 3 A) C3A_0.12PC1 40 3 300 C3A_0.24PC1 20 30

200 (mW/g C Heat flow

10 20

200 (mW/g C Heat flow 0 Heat flow (mW/gC 100 10 0246810 Time (h) 0 Heat flow (mW/gC 100 0246810 Time (h) 0 0 5 10 15 20 25

0 Time (h) 0 5 10 15 20 25 .

Time (h) Figure 4. Heat evolution curves of C3A-gypsum with different concentrations of PC1 solution. during

Figurethe 24-h 4. hydration.Heat evolution curves of C3A-gypsum with different concentrations of PC1 solution during Figurethe 24-h 4. hydration.Heat evolution curves of C3A-gypsum with different concentrations of PC1 solution during the 24-h hydration.

1 - Ettringite 2 - Katoite

8.8 8.9 9.0 9.1 9.2 9.3 9.4 39.0 39.1 39.2 39.3 39.4 39.5 39.6 2θ (°) 2θ (°) 38.8 - Kuzelite 8.9 9.0 9.1 9.2 9.3 9.4 39.0 39.1 39.2 39.3 39.4 39.5 39.6 4 - Gypsum C3A_1G

2θ (°) 2θ (°) C3A_0.06PC1

3 - Kuzelite C3A_0.12PC1 4 - Gypsum C3A_1G C3A_0.24PC1 C3A_0.06PC1

C3A_0.12PC1

C3A_0.24PC1

9.6 9.7 9.8 9.9 10.0 10.1 10.2 11.4 11.5 11.6 11.7 11.8 11.9 12.0 2θ (°) 2θ (°) 9.6 9.7 9.8 9.9 10.0 10.1 10.2 11.4 11.5 11.6 11.7 11.8 11.9 12.0 (a) 2θ (°) 2θ (°)

Figure 5. Cont. (a)

Materials 2019, 12, 1132 9 of 18 Materials 2019, 12, x FOR PEER REVIEW 9 of 17

2500

2.35E3 C3A_1G

C3A_0.06PC1

C3A_0.12PC1

C3A_0.24PC1 2000

845

654

Peak area (a.u.) Peak 547 523 500 477 444

271 245

113 99.3 101 105 61.2 55.1 32.8 0 1 - Ettringite 2 - Katoite 3 - Kuzelite 4 - Gypsum

Phases (b)

FigureFigure 5. 5.( (aa)) XRDXRD patternspatterns ofof phasesphases obtainedobtained fromfrom C C33A-gypsumA-gypsum withwith differentdifferent concentrationsconcentrations ofof PC1PC1 solutionsolution after after the the 24-h 24-h hydration. hydration. ( b(b)) XRD XRD peak peak area area of of each each phase phase in in (a). (a).

1.1. Dissolution–crystallizationDissolution–crystallization stage stage PC1 molecules could be adsorbed onto the surface of C3A particles after the mixing step. The PC1 molecules could be adsorbed onto the surface of C3A particles after the mixing step. The more more PC1 molecules that were introduced, the more they were adsorbed onto the C3A particles before PC1 molecules that were introduced, the more they were adsorbed onto the C3A particles before reachingreaching thethe saturatedsaturated adsorption.adsorption. WithWith thethe increasedincreased concentrationconcentration ofof PC1,PC1, moremore andand moremore PC1PC1 molecules occupied the dissolution sites on the surface of C3A, and the size effect of adsorbed PC1 molecules occupied the dissolution sites on the surface of C3A, and the size effect of adsorbed PC1 moleculesmolecules maymay havehave reducedreduced thethe nucleationnucleation andand growthgrowth ofof hydrateshydrates suchsuch thatthat thethe precipitationprecipitation ofof hydroxy-AFmhydroxy-AFm at at the the beginning beginning of hydrationof hydration was was hindered. hindered. On the On other the other hand, hand, EO chains EO couldchains attract could moreattract water more molecules water molecules to enhance to enhance the ion exchangethe ion exchange between between solution solution and solids and [8 solids], which [8], was which due was to thedue characteristics to the characteristics of the surfactant. of the surfactant. Both factors Both are thefactors reasonable are the cause reasonable of the similarcause productionof the similar of hydroxy-AFmproduction of inhydroxy-AFm the blank sample in the and blank sample sample with and a 0.24%sample concentration with a 0.24% of concentration PC1. of PC1. 2. Induction stage 2. InductionThe dosage stage of PC1 may have effected the induction stage, which was reflected in the duration betweenThe dosagethe two of exothermic PC1 may have peaks. effected More the PC1 induction advanced stage, the which transformation was reﬂected stage, in theprobably duration by betweeninhibiting the the two dissolution exothermic of gypsum, peaks. More which PC1 coul advancedd be concluded the transformation from the remains stage, of probably gypsum, by as inhibitingshown in theFigure dissolution 5b. The ofconcentration gypsum, which of uncons couldtrained be concluded PC1 molecules from the determine remains of the gypsum, release asof shownsulfate in ions Figure with5b. negative The concentration charge. Despite of unconstrained the fact that PC1free PC1 molecules mightdetermine have expended the release Ca2+ in of solution sulfate ions[20], with the original negative equilibrium charge. Despite between the the fact formation that free PC1 of AFt might and have the dissolution expended Caof 2+gypsumin solution may [have20], thebeen original broken equilibrium by PC1 and between the new the dynamic formation equilibrium of AFt and about the dissolution the dissolution of gypsum of C3A, may the have release been of brokensulfate byions, PC1 and and the newconsumption dynamic equilibriumof free PC1 about molecules the dissolution was possibly of C3A, established. the release ofWith sulfate the ions,exhaustion and the of consumption PC1, the dissolution of free PC1 of molecules C3A increased was possibly again and established. AFm was With formed the exhaustion because the of PC1,activation the dissolution energy of Equation of C3A increased (1) was higher again andthan AFmthat of was Equation formed (2) because [1] such the that activation gypsum energyremained. of Equation3. Transformation (1) was higher stage than that of Equation (2) [1] such that gypsum remained. PCs had an influence on the morphology of AFt via hindering the growth of crystals and 3. Transformation stage increasing the surface area [15]. After depletion of the dissociative sulfate ions, the interaction betweenPCs hadthe ansmall inﬂuence ettringite on the particles morphology with substa of AFtntial via hinderingspecific surface the growth area ofand crystals the unhydrated and increasing C3A thewas surface enhanced area under [15]. After the influence depletion of of adsorbed the dissociative PC molecules sulfate on ions, both the surfaces. interaction This between was probably the small the reason why the heat flow peak was higher and narrower than the blank sample. Samples with less PC1 dosage may have produced more AFt before the transformation stage, while the dissolution of

Materials 2019, 12, x FOR PEER REVIEW 10 of 17 gypsum was restrained. Hence, the highest AFm production, from the sample with 0.06% concentration of PC1, could be possible.

3.2.2. Effects of PC with AM Monomers

The hydration heat of C3A-gypsum mixed with a different concentration of PC2 solutions within 24 h is demonstrated in Figure 6. It is apparent from Figure 6 that with the increased PC2 dosage, the major differences were concentrated upon the induction stage and the transformation stage. A 0.06% concentration of PC2 could prolong the induction stage, but more PC2 might shorten that. No matter howMaterials many2019 PC2, 12 molecules, 1132 were introduced in the hydration of C3A-gypsum, exothermic peaks in 10the of 18 transformation stage had a minor difference, yet higher than the blank sample. XRD patterns of C3A- gypsum hydration with different PC2 concentration are revealed in Figure 7a. Furthermore, the area of phaseettringite peaks particles in Figure with 7a substantial is shown specificin Figure surface 7b. The area amount and the of unhydrated ettringite was C3 Adecreased was enhanced first and under thenthe increased influence in ofthe adsorbed peak area PC with molecules the increa onsed both PC2 surfaces. dosage. The This amount was probably of hydroxy-AFm the reason created why the viaheat samples flow with peak PC2 was was higher more and than narrower that of the than blank the sample. blank With sample. the increased Samples withPC2 concentration, less PC1 dosage themay production have produced of AFm morewas nearly AFt before the same, the transformation which was still stage, less than while that the of dissolution the blank sample. of gypsum was restrained.Based on Hence,articles thepublished highest in AFm public production, [21], the fromamide the unit sample is nonionic with 0.06% unit that concentration can hardly of be PC1, could be possible. adsorbed on the surface of particles. Therefore, the adsorption amount of PC2 on the surface of C3A particles was less than that of PC1 under the same PC concentration. This was owed to the lower 3.2.2. Effects of PC with AM Monomers electrostatic charge of the PC2 backbone. Fewer adsorbed PC2 molecules on the clinker surface may have releasedThe hydration more dissolution heat of C3A-gypsum sites such mixedthat the with precipitation a different concentrationof hydroxy-AFm of PC2 was solutions promoted. within Furthermore,24 h is demonstrated the hydrogen in Figure bond6 .between It is apparent the water from Figuremolecule6 that and with amide the increasedunit also PC2enhanced dosage, the the hydrationmajor differences of C3A. On were the other concentrated hand, the upon residual the gypsum induction increased stage and as thea result transformation of more unconstrained stage. A 0.06% PC2concentration molecules in of the PC2 media. could That prolong means the that induction the induction stage, but stage more was PC2 shortened might shorten with increased that. No PC2 matter concentration.how many PC2The moleculesAFm production were introduced of samples in with the hydrationPC2 was less of C than3A-gypsum, that of the exothermic blank sample, peaks in wherethe transformationthis may be attributed stage had to the a minor lack of difference, C3A. PC2 with yet higher a lower than electrostatic the blank charge sample. than XRD PC1 patterns could of haveC3 A-gypsuma weaker inhibition hydration effect with differenton the dissolution PC2 concentration of gypsum are such revealed that in the Figure AFt 7formationa. Furthermore, rate of the samplesarea of with phase PC2 peaks was in higher Figure than7a is shownthat of insamples Figure7 ofb. PC1. The amount The precipitation of ettringite of was hydroxy-AFm decreased first and and AFtthen consumed increased more in the C3 peakA with area PC2 with addition the increased during PC2 the dosage. dissolution–crystallization The amount of hydroxy-AFm stage and created the inductionvia samples stage, with and PC2 as a was conseque more thannce, thatthe formation of the blank of sample.AFm in Withthe transformation the increased PC2 stage concentration, was less thanthe the production blank. of AFm was nearly the same, which was still less than that of the blank sample.

400

350 C3A_1G

C3A_0.06PC2 40 300 C3A_0.12PC2 A) A) 3 3 C3A_0.24PC2 30 250

20 200 Heat C (mW/g flow 150 10

0 Heat flow (mW/g C (mW/g flow Heat 100 0246810

Time (h) 50

0 0 5 10 15 20 25

Time (h)

FigureFigure 6. Heat 6. Heat evolution evolution curves curves of C of3A-gypsum C3A-gypsum with with different different concentrat concentrationsions of PC2 of PC2 solution solution during during the the24-h 24-h hydration. hydration.

Based on articles published in public [21], the amide unit is nonionic unit that can hardly be adsorbed on the surface of particles. Therefore, the adsorption amount of PC2 on the surface of C3A particles was less than that of PC1 under the same PC concentration. This was owed to the lower electrostatic charge of the PC2 backbone. Fewer adsorbed PC2 molecules on the clinker surface may have released more dissolution sites such that the precipitation of hydroxy-AFm was promoted. Furthermore, the hydrogen bond between the water molecule and amide unit also enhanced the hydration of C3A. On the other hand, the residual gypsum increased as a result of more unconstrained PC2 molecules in the media. That means that the induction stage was shortened with increased PC2 concentration. The AFm production of samples with PC2 was less than that of the blank sample, where Materials 2019, 12, 1132 11 of 18

this may be attributed to the lack of C3A. PC2 with a lower electrostatic charge than PC1 could have a weaker inhibition effect on the dissolution of gypsum such that the AFt formation rate of samples with PC2 was higher than that of samples of PC1. The precipitation of hydroxy-AFm and AFt consumed Materialsmore 2019 C3A, 12 with, x FOR PC2 PEER addition REVIEW during the dissolution–crystallization stage and the induction stage,11 of 17 and as a consequence, the formation of AFm in the transformation stage was less than the blank.

1 - Ettringite 2 - Katoite

8.8 8.9 9.0 9.1 9.2 9.3 9.4 39.0 39.1 39.2 39.3 39.4 39.5 39.6 2θ (°) 2θ (°)

C3A_1G 3 - Kuzelite 4 - Gypsum C3A_0.06PC2

C3A_0.12PC2

C3A_0.24PC2

9.6 9.7 9.8 9.9 10.0 10.1 10.2 11.4 11.5 11.6 11.7 11.8 11.9 12.0 2θ (°) 2θ (°)

(a)

700

654 C3A_1G 616 C3A_0.06PC2 600 C3A_0.12PC2 547 C A_0.24PC2 511 3 500 446 408 411 410 404 400 365 340 300

Peak area (a.u.) 245

200 149

100 84.5 61.3 32.8 0 1 - Ettringite 2 - Katoite 3 - Kuzelite 4 - Gypsum

Phases

(b)

FigureFigure 7. ( 7.a)( XRDa) XRD patterns patterns of ofphases phases obtained obtained from from C3A-gypsum C3A-gypsum with with different different concentrations concentrations of ofPC2 PC2 solutionsolution after after the the 24-h 24-h hydration. hydration. (b) ( bXRD) XRD peak peak area area of ofeach each phase phase in in(a). (a).

3.2.3. Effects of PC with AMPS Monomers

The hydration heat of C3A-gypsum mixed with different concentration of PC3 solutions within 24 h is demonstrated in Figure 8. As can be seen from Figure 8, although AMPS contained the amide unit, the hydration heat evolution of C3A-gypsum with PC3 varied from that with PC2. With the increased PC3 concentration, the second exothermic peak of each sample rose, but the order of appearance time of the transformation heat flow peak was different. XRD patterns of phases after the

Materials 2019, 12, x FOR PEER REVIEW 12 of 17

24-h hydration are shown in Figure 9a. The peak area of phases in Figure 9a is shown in Figure 9b. The sample with a 0.12% concentration of PC3 had the shortest induction stage, with the most gypsum remaining. Although samples with 0.12% and 0.24% concentration of PC3 released more heat during the transformation stage, the produced AFm and remaining AFt were less than the blank sample and the sample with a 0.06% concentration of PC3. MaterialsUnlike2019 the, 12 amide, 1132 unit, the sulfo unit is easier to ionize; furthermore, its hydrophilicity is better12 of 18 than the carboxylate unit [8]. Although the adsorbed PC3 molecules occupy lots of active sites on the surface3.2.3. of Effects C3A, oflike PC PC1, with AMPStheir excellent Monomers hydrophilicity resulted in an attraction for more water molecules such that the ion transport between the clinker and solution was accelerated. Therefore, a low concentrationThe hydration of heatPC3 ofmay C3 A-gypsumenhance the mixed prod withuction different of hydroxy-AFm concentration without of PC3 solutionsreducing withinthe formation24 h is demonstratedof AFt and AFm. in Figure When8. Asthe can concentrat be seen fromion of Figure PC3 8increased,, although more AMPS unconstrained contained the amidePC3 moleculesunit, the hindered hydration the heat dissolution evolution ofof Cgypsum,3A-gypsum the with induction PC3 varied stage from was that shortened, with PC2. and With the the productionincreased of PC3 AFt concentration,and AFm decreased the second as well. exothermic peak of each sample rose, but the order of appearanceAs the concentration time of the of transformation PC3 continually heat rose, ﬂow th peake sample was different.with a 0.24% XRD concentration patterns of phases of PC3 after had the a longer24-h hydration induction are stage shown than in the Figure sample9a. The with peak a 0.12% area of concentration phases in Figure of 9PC3.a is shownFurthermore, in Figure the9b. residualThe sample gypsum with after a 0.12% the 24-h concentration hydration of decrease PC3 hadd. the This shortest phenomenon induction might stage, be with explained the most with gypsum a properremaining. hypothesis Although about samples the critical with micelle 0.12% and conc 0.24%entration concentration (CMC) of of PC3. PC3 releasedActive sites more of heat created during hydratesthe transformation were mostly stage,covered the with produced PC3 molecules AFm and such remaining that the AFt induction were less stage than would the blank be prolonged sample and untilthe PC3 sample molecules with a were 0.06% depleted concentration via intercalation of PC3. and complexation with Ca2+.

400 60 C3A_1G

C3A_0.06PC3 50 A) A) C A_0.12PC3 3 3 3 300 40 C3A_0.24PC3

200 20 Heat flow (mW/g C

0 Heat flow (mW/g C 100 0246810 Time (h)

0 0 5 10 15 20 25

Time (h)

FigureFigure 8. Heat 8. Heat evolution evolution curves curves of C of3A-gypsum C3A-gypsum with with different different concentrat concentrationsions of PC3 of PC3 solution solution during during the the24-h 24-h hydration. hydration.

Unlike the amide unit, the sulfo unit is easier to ionize; furthermore, its hydrophilicity is better than the carboxylate unit [8]. Although the adsorbed PC3 molecules occupy lots of active sites on the surface of C3A, like PC1, their excellent hydrophilicity resulted in an attraction for more water molecules such that the ion transport between the clinker and solution was accelerated. Therefore, a low concentration of PC3 may enhance the production of hydroxy-AFm without reducing the formation of AFt and AFm. When the concentration of PC3 increased, more unconstrained PC3 molecules hindered the dissolution of gypsum, the induction stage was shortened, and the production of AFt and AFm decreased as well. As the concentration of PC3 continually rose, the sample with a 0.24% concentration of PC3 had a longer induction stage than the sample with a 0.12% concentration of PC3. Furthermore, the residual gypsum after the 24-h hydration decreased. This phenomenon might be explained with a proper hypothesis about the critical micelle concentration (CMC) of PC3. Active sites of created hydrates were mostly covered with PC3 molecules such that the induction stage would be prolonged until PC3 molecules were depleted via intercalation and complexation with Ca2+. MaterialsMaterials 20192019, 12, x12 FOR, 1132 PEER REVIEW 13 of13 17 of 18

1 - Ettringite 2 - Katoite

8.88.99.09.19.29.39.4 39.0 39.1 39.2 39.3 39.4 39.5 39.6 2θ (°) 2θ (°) 3 - Kuzelite C A_1G 4 - Gypsum 3 C3A_0.06PC3

C3A_0.12PC3

C3A_0.24PC3

9.6 9.7 9.8 9.9 10.0 10.1 10.2 11.4 11.5 11.6 11.7 11.8 11.9 12.0 2θ (°) 2θ (°)

(a) 800 731 C3A_1G 700 675 654 C3A_0.06PC3

C3A_0.12PC3 600 C A_0.24PC3 547 3 504 500 459 428 403 401 400 385 340 300 Peak area (a.u.) Peak area 245 200 156 130 110 100 32.8 0 1 - Ettringite 2 - Katoite 3 - Kuzelite 4 - Gypsum

Phases (b)

FigureFigure 9. (a 9.) XRD(a) XRD patterns patterns of phases of phases obtained obtained from from C3A-gypsum C3A-gypsum with with different different concentrations concentrations of PC3 of PC3 solutionsolution after after the the24-h 24-h hydration. hydration. (b) (XRDb) XRD peak peak area area of each of each phase phase in (a). in (a).

3.2.4.3.2.4. Effect Effect of the of the PCs PCs with with Different Different Functional Functional Units Units

TheThe hydration hydration heat heat of ofC3A-gypsum C3A-gypsum mixed mixed with with a 0.12% a 0.12% concentration concentration of PCs of PCs with with different different functionalfunctional units units within within 24 24h is h revealed is revealed in inFigure Figure 10. 10 According. According to Figure to Figure 10, 10 PCs, PCs shortened shortened the the inductioninduction stage stage of C of3A-gypsum C3A-gypsum hydration. hydration. PC PCcopolymerized copolymerized with with AMPS AMPS could could tremendously tremendously cut cut downdown on onthe the induction induction stage. stage. The The second second heat heat flow ﬂow peaks peaks of the of the samples samples with with PCs PCs was was higher higher than than thatthat of ofthe the blank blank sample. sample. XRD XRD patterns patterns of of phases phases after after the the 24-h 24-h hydration areare illustratedillustrated in in Figure Figure 11 a. 11a. All the hydrates formed by samples with PCs displayed a shift of the intensity peaks, which

Materials 2019, 12, x FOR PEER REVIEW 14 of 17 indicated that PC intercalation occurred during C3A-gypsum hydration. Different shifts caused by different kinds of PCs probably meant the conformation of intercalated PC molecules was different. This agreed well with results reported by Plank et al. [9]. The area of the XRD phase peaks in Figure 11a is shown in Figure 11b. Except for PC1, PCs clearly promoted the formation of hydrogarnet. The dissolution of gypsum was inhibited by PCs such that the formation of AFt was restrained. Gypsum remained in the system after the 24-h hydration, which was consistent with the duration of the induction stage of samples. Although all the functional units mentioned may reduce the adsorption amount of PCs on the surface of clinkers, different functional units had varying influences on the hydration of C3A-gypsum. PC1 molecules could be primarily adsorbed onto the surface of clinkers because of electrostatic interaction when C3A-gypsum mixed with the PCs–water solution [22]. Active sites on the surface of C3A were occupied such that the precipitation of hydroxy-AFm of the sample with PC1 was slowed down. Other PC molecules might have mainly been still in solution due to the lack of carboxylate units. The formation of hydroxy-AFm was advanced with other PCs because the adsorbed PC moleculesMaterials 2019could, 12 ,attract 1132 more water molecules to accelerate the dissolution of C3A [8]. PC molecules14 of 18 with negative electrostatic charge in solution prevented gypsum from dissolving and occupying dissolution sites on the surface of clinkers that were originally for sulfate ions. Therefore, the formationAll the hydratesof AFt during formed the by induction samples with stage PCs is displayedhindered, athe shift duration of the intensity of the induction peaks, which stage indicated was prolongedthat PC as intercalation well. PC molecules occurred with during different C3A-gypsum amounts hydration.of electrostatic Different charge shifts had different caused by abilities different to kindsinhibit of sulfate PCs probably ions from meant releasing. the conformation PC3 with of sulfo intercalated units PChad molecules the most was amount different. of negative This agreed electrostaticwell with charge results among reported them. by Plank Hence, et the al. [sample9]. The with area PC3 of the had XRD the phase shortest peaks induction in Figure stage 11a during is shown C3A-gypsumin Figure 11 hydration.b. Except forThe PC1, produced PCs clearly AFt particles promoted were the small formation but had of hydrogarnet. higher specific The surface dissolution area of [15],gypsum which waswas inhibitedowed to bythe PCs steric such hindrance that the formationand repulsive of AFt interaction was restrained. [23,24] Gypsumof adsorbed remained PCs. in Furthermore,the system this after is thea possible 24-h hydration, reason for which the high waser exothermic consistent withpeak the during duration the transformation of the induction stage stage thanof the samples. blank sample.

400 C3A_1G 60

C3A_0.12PC1 50 C A_0.12PC2 A) 3 A) 3 3 300 C3A_0.12PC3 40

200 20 Heat flow (mW/g C (mW/g flow Heat

0 Heat (mW/g C Heat flow 100 0246810 Time (h)

0 0 5 10 15 20 25

Time (h)

FigureFigure 10. Heat 10. Heat evolution evolution curves curves of C of3A-gypsum C3A-gypsum with with a 0.12% a 0.12% concentrat concentrationion of different of different kinds kinds of PCs of PCs solutionsolution during during the the24-h 24-h hydration. hydration.

Although all the functional units mentioned may reduce the adsorption amount of PCs on the surface of clinkers, different functional units had varying inﬂuences on the hydration of C3A-gypsum. PC1 molecules could be primarily adsorbed onto the surface of clinkers because of electrostatic interaction when C3A-gypsum mixed with the PCs–water solution [22]. Active sites on the surface of C3A were occupied such that the precipitation of hydroxy-AFm of the sample with PC1 was slowed down. Other PC molecules might have mainly been still in solution due to the lack of carboxylate units. The formation of hydroxy-AFm was advanced with other PCs because the adsorbed PC molecules could attract more water molecules to accelerate the dissolution of C3A[8]. PC molecules with negative electrostatic charge in solution prevented gypsum from dissolving and occupying dissolution sites on the surface of clinkers that were originally for sulfate ions. Therefore, the formation of AFt during the induction stage is hindered, the duration of the induction stage was prolonged as well. PC molecules with different amounts of electrostatic charge had different abilities to inhibit sulfate ions from releasing. PC3 with sulfo units had the most amount of negative electrostatic charge among them. Hence, the sample with PC3 had the shortest induction stage during C3A-gypsum hydration. The produced AFt particles were small but had higher speciﬁc surface area [15], which was owed to the steric hindrance and repulsive interaction [23,24] of adsorbed PCs. Furthermore, this is a possible reason for the higher exothermic peak during the transformation stage than the blank sample. MaterialsMaterials 20192019, 12,, 12x FOR, 1132 PEER REVIEW 15 of15 17 of 18

1 - Ettringite 2 - Katoite

8.8 8.9 9.0 9.1 9.2 9.3 9.4 39.0 39.1 39.2 39.3 39.4 39.5 39.6 2θ (°) 2θ (°) 3 - Kuzelite 4 - Gypsum C3A_1G

C3A_0.12PC1

C3A_0.12PC2

C3A_0.12PC3

9.6 9.7 9.8 9.9 10.0 10.1 10.2 11.4 11.5 11.6 11.7 11.8 11.9 12.0 2θ (°) 2θ (°)

(a)

700 654 C3A_1G

C3A_0.12PC1 600 C3A_0.12PC2 547 523 C3A_0.12PC3 500 459 446 444 411 401 400 385 365

300

Peak area (a.u.) Peak area 245

200 156

99.3 100 100 84.7

32.8 0 1 - Ettringite 2 - Katoite 3 - Kuzelite 4 - Gypsum

Phases

(b)

FigureFigure 11. 11.(a) (XRDa) XRD patterns patterns of ofphases phases obtained obtained from from C3A-gypsum C3A-gypsum with with the the same same concentrations concentrations of of differentdifferent kinds kinds of PCs of PCs solution solution after after the the 24-h 24-h hydration. hydration. (b) ( bXRD) XRD peak peak area area of ofeach each phase phase in in(a). (a).

4. Conclusions4. Conclusions

First,First, hydration hydration of C of3A C in3A the in presence the presence of various of various amounts amounts of gypsum of gypsum was studied. was studied. Based Basedon the on hydrationthe hydration heat evolution heat evolution curves, curves, three threestages stages were were proposed proposed in an in attempt an attempt to elaborate to elaborate the the early early hydrationhydration behavior behavior of ofC3A-gypsum: C3A-gypsum:

Materials 2019, 12, 1132 16 of 18

• In the dissolution–crystallization stage, the precipitation of hydroxy-AFm and AFt coexisted, although there may be a time interval between them. The dosage of gypsum could not have a signiﬁcant impact on the duration of this stage.

• Unhydrated C3A remained after the sulfate depletion was the precondition of the induction stage, and the speciﬁc G/C ratio allowing the induction stage could exist was hard to obtain because of the uncertain consumption of C3A or gypsum during the dissolution–crystallization. The duration of this stage depended on the number of soluble sulfate ions. • The transformation stage started after the sulfate exhaustion when AFt was converted into AFm. Remaining C3A after the induction stage determined the duration of this stage.

Inﬂuences of PCs with different functional units on hydration behavior of C3A-gypsum was investigated using isothermal calorimetry and XRD analysis:

• A low concentration of PC, which copolymerized only with carboxylate unit in solution, could promote the transformation stage of C3A-gypsum hydration significantly. With the increased concentration of PC1, the appearance of second heat flow peak advanced, the amount of remained gypsum increased, and the amount of remained AFt increased. • The concentration of PC copolymerized with an amide unit influenced the duration of the induction stage. A low concentration of PC2 could prolong it. PC2 may have accelerated the formation of hydrogarnet. Although the amount of AFm created was less than the blank sample, a higher transformation heat flow peak could be found during the hydration of samples with PC2. • As the concentration of PC that copolymerized with AMPS monomers increased, the duration of the induction stage was shortened first, and then prolonged. A low concentration of PC3 had a minor influence on the hydration of C3A-gypsum, except for the exotherm during the transformation stage. Also, the dosage of PC3 could enhance the hydration of C3A such that the production of hydrogarnet was more than the blank sample.

In general, the induction stage of C3A-gypsum hydration may have been shortened with the PCs addition, and the amount of remaining gypsum was related to the duration of the induction stage. Moreover, the exothermic peak in the transformation stage was higher than the blank sample. PC intercalation existed during the hydration.

Author Contributions: K.H. conceived and designed the experiments, analyzed the data, and wrote the paper; Z.S. supervised the research. Funding: This research was funded by the National and Technology Program During the 13th Five-Year Plan Period (grant number 2016YFC0701004), National Natural Science Joint Foundation of China (grant number U1534207), National Natural Science Foundation of China (grant number 51678441), and Opening Test Fund for Large-Scale Instruments and Equipment of Tongji University (grant number 0002016025). Acknowledgments: The first author would like to especially thank Weixiang Zhang, Haijing Yang, Yihe Zhao, Yanliang Ji, Yunzi Mao, Deyu Fu, Jingbin Yang, and Chao Chen for helpful suggestions. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Quennoz, A.; Scrivener, K.L. Hydration of C3A-gypsum systems. Cem. Concr. Res. 2012, 42, 1032–1041. [CrossRef] 2. Taylor, H.F.W. Cement Chemistry, 2nd ed.; Thomas Telford: London, UK, 1997. 3. Collepardi, M.; Baldini, G.; Pauri, M.; Corradi, M. Tricalcium aluminate hydration in presence of lime, gypsum or sodium-sulfate. Cem. Concr. Res. 1978, 8, 571–580. [CrossRef] Materials 2019, 12, 1132 17 of 18

4. Brown, P.W.; Liberman, L.O.; Frohnsdorff, G. Kinetics of the early hydration of tricalcium aluminate in solutions containing calcium-sulfate. J. Am. Ceram. Soc. 1984, 67, 793–795. [CrossRef]

5. Scrivener, K.L.; Pratt, P.L. Microstructural studies of the hydration of C3A and C4AF independently and in cement paste. Br. Ceram. Proc. 1984, 35, 207–219. 6. Minard, H.; Garrault, S.; Regnaud, L.; Nonat, A. Mechanisms and parameters controlling the tricalcium aluminate reactivity in the presence of gypsum. Cem. Concr. Res. 2007, 37, 1418–1426. [CrossRef]

7. Pourchet, S.; Regnaud, L.; Perez, J.P.; Nonat, A. Early C3A hydration in the presence of different kinds of calcium sulfate. Cem. Concr. Res. 2009, 39, 989–996. [CrossRef] 8. Hu, K.; Sun, Z.; Yang, H. Effects of polycarboxylate superplasticizers with different functional units on the early hydration behavior of cement paste. J. Mater. Civ. Eng. 2019, in press. [CrossRef] 9. Winnefeld, F.; Becker, S.; Pakusch, J.; Götz, T. Effects of the molecular architecture of comb-shaped superplasticizers on their performance in cementitious systems. Cem. Concr. Res. 2007, 29, 251–262. [CrossRef] 10. Zingg, A.; Winnefeld, F.; Holzer, L.; Pakusch, J.; Becker, S.; Gauckler, L. Adsorption of polyelectrolytes and its inﬂuence on the rheology, zeta potential, and microstructure of various cement and hydrate phases. J. Colloid Interface Sci. 2008, 323, 301–312. [CrossRef][PubMed] 11. Kong, F.R.; Pan, L.S.; Wang, C.M.; Xu, N. Effects of polycarboxylate superplasticizers with different molecular structure on the hydration behavior of cement paste. Constr. Build. Mater. 2016, 105, 545–553. [CrossRef] 12. Liu, X.; Guan, J.; Lai, G.; Wang, Z.; Zhu, J.; Cui, S.; Lan, M.; Li, H. Performances and working mechanism of a novel polycarboxylate superplasticizer synthesized through changing molecular topological structure. J. Colloid Interface Sci. 2017, 504, 12–24. [CrossRef] 13. Merlini, M.; Artioli, G.; Cerulli, T.; Cella, F.; Bravo, A. Tricalcium aluminate hydration in additivated systems. A crystallographic study by SR-XRPD. Cem. Concr. Res. 2008, 38, 477–486. [CrossRef] 14. Plank, J.; Zhimin, D.; Keller, H.; Hössle, F.V.; Seidl, W. Fundamental mechanisms for polycarboxylate

intercalation into C3A hydrate phases and the role of sulfate present in cement. Cem. Concr. Res. 2010, 40, 45–57. [CrossRef] 15. Dalas, F.; Pourchet, S.; Rinaldi, D.; Nonat, A.; Sabio, S.; Mosquet, M. Modiﬁcation of the rate of formation

and surface area of ettringite by polycarboxylate ether superplasticizers during early C3A–CaSO4 hydration. Cem. Concr. Res. 2015, 69, 105–113. [CrossRef] 16. Alonso, M.M.; Puertas, F. Adsorption of PCE and PNS superplasticisers on cubic and orthorhombic C3A. Effect of sulfate. Constr. Build. Mater. 2015, 78, 324–332. [CrossRef] 17. Myers, R.J.; Geng, G.; Li, J.; Rodríguez, E.D.; Ha, J.; Kidkhunthod, P.; Sposito, G.; Lammers, L.N.; Kirchheim, A.P.; Monteiro, P.J. Role of adsorption phenomena in cubic tricalcium aluminate dissolution. Langmuir 2016, 33, 45–55. [CrossRef][PubMed] 18. Sun, Z.; Yang, H.; Shui, L.; Liu, Y.; Yang, X.; Ji, Y.; Hu, K.; Luo, Q. Preparation of polycarboxylate-based grinding aid and its inﬂuence on cement properties under laboratory condition. Constr. Build. Mater. 2016, 127, 363–368. [CrossRef] 19. Kirchheim, A.P.; Rodriguez, E.D.; Myers, R.J.; Gobbo, L.A.; Monteiro, P.J.M.; Dal Molin, D.C.C.; de Souza, R.B.; Cincotto, M.A. Effect of Gypsum on the Early Hydration of Cubic and Na-Doped Orthorhombic Tricalcium Aluminate. Materials 2018, 11, 568. [CrossRef] 20. Zingg, A.; Winnefeld, F.; Holzer, L.; Pakusch, J.; Becker, S.; Figi, R.; Gauckler, L. Interaction of

polycarboxylate-based superplasticizers with cements containing different C3A amounts. Cem. Concr. Compos. 2009, 31, 153–162. [CrossRef] 21. Liu, Y. Research on Synthesis, Properties and Mechanism of Different Carboxyl Density and Functionallizing Polycarboxylate Superplasticizer. Ph.D. Thesis, China University of Mining and Technology, Xuzhou, China, June 2013. 22. Yoshioka, K.; Tazawa, E.; Kawai, K.; Enohata, T. Adsorption characteristics of superplasticizers on cement component minerals. Cem. Concr. Res. 2002, 32, 1507–1513. [CrossRef] Materials 2019, 12, 1132 18 of 18

23. Zingg, A.; Holzer, L.; Kaech, A.; Winnefeld, F.; Pakusch, J.; Becker, S.; Gauckler, L. The microstructure of dispersed and non-dispersed fresh cement pastes—New insight by cryo-microscopy. Cem. Concr. Res. 2008, 38, 522–529. [CrossRef] 24. Holzer, L.; Gasser, P.; Kaech, A.; Wegmann, M.; Zingg, A.; Wepf, R.; Muench, B. Cryo-FIB-nanotomography for quantitative analysis of particle structures in cement suspensions. J. Microsc.-Oxf. 2007, 227, 216–228. [CrossRef]

© 2019 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/).