Quicklime and Calcium Sulfoaluminate Cement Used As Mineral Accelerators to Improve the Properties of Cemented Paste Backﬁll with a High Volume of Fly Ash

materials

Article Quicklime and Calcium Sulfoaluminate Cement Used as Mineral Accelerators to Improve the Properties of Cemented Paste Backﬁll with a High Volume of Fly Ash

Hangxing Ding 1,2 and Shiyu Zhang 1,2,*

1 Key Laboratory of Ministry of Education on Safe Mining of Deep Metal Mines, Northeastern University, Shenyang 110819, China; [email protected] 2 Science and Technology Innovation Center of Smart Water and Resource Environment, Northeastern University, Shenyang 110819, China * Correspondence: [email protected]

Received: 21 August 2020; Accepted: 8 September 2020; Published: 10 September 2020

Abstract: In order to reduce the CO2 emission and cost of binders used in cemented paste backfill (CPB) technology, new blended binders with a large amount of fly ash (FA) were fabricated. Different doses of quicklime and calcium sulfoaluminate cement (CSA)˙ were used as mineral accelerators to improve the early workability of CPB. The effects of CSA˙ and quicklime on flowability, compressive strength, pore structure, hydration heat, and hydration evolution were investigated experimentally. The results showed that the addition of quicklime and CSA˙ reduced the spread diameter of the fresh backfill and improved the mechanical performance of the hardened CPB. With increasing quicklime and CSA,˙ the cumulative hydration heat of the blended binder distinctly increased in the first 6 h. CSA˙ improved the initial hydration by increasing the reactivity, and quicklime increased the hydration rate by activating FA. The blended binder (15% quicklime + 10% CSA)˙ with the lowest CO2 emission and cost had potential application in filling technology.

Keywords: calcium sulfoaluminate cement; quicklime; ﬂy ash; hydration evolution; CO2 emission

1. Introduction Cemented paste backfill (CPB) technology is widely accepted in the mining industry, addressing the issue of instability in underground stopes and mine tailings [1–5]. This technology correspondingly solves many issues (low ore recovery, surface subsidence, acid mine drainage, groundwater pollution, tailing dam failure, etc.) and brings economic and environmental benefits [6–11]. CPB is an engineered mixture composed of thickened and filtered tailings, water, and hydraulic binder, with a total solid concentration (tailings + binder) of 70–85% [7,12–19]. The function of the binder is to maintain a uniform suspension of tailings during pipe transportation, allowing the hardened CPB to achieve the required strength [4,7]. The ratio is in the range of 2–7% (sometimes up to 10%), and the binder accounts for about 75% of the total filling cost [3,20,21]. In recent years, discussions on saving filling costs have become mainstream. Much attention has been directed to the potential use of aluminosilicate source of materials (i.e., fly ash (FA), blast-furnace slag, metakaolin, and silica fume) as a partial replacement for cement in backfill. FA, a by-product of power plants, is commonly used as supplementary cementitious material for fabricating cement-based materials. The addition of FA reduces the rheological properties and improves the flowability of the fresh backfill due to the lubrication effect of spherical morphology [22]. Besides, FA, rich in aluminum and silicon, can be involved in cement hydration and form hydrate products (e.g., calcium

Materials 2020, 13, 4018; doi:10.3390/ma13184018 www.mdpi.com/journal/materials Materials 2020, 13, 4018 2 of 20 silicate hydrate, calcium aluminate hydrate, and ettringite) that affect strength development [23–26]. Deschner et al. [27] documented that with an early hydration time of up to 2 days, cement hydration is mainly affected by the so-called “filler effect” of FA. At the same time, the FA reaction is detected by the consumption of portlandite after a hydration time of 2 days. Schöler et al. [28] showed that low-calcium FA slightly reduced the formation of portlandite and calcium silicate hydrate but increased the formation of monocarbonaluminate/hemicarbonaluminate. Addition of less than 30% FA has a limited effect on strength development at later ages. However, the mechanical performance of the backfill with high volume (>30%) of FA is significantly reduced due to low activity, especially at an early age. To overcome this issue, several methods have been used to improve the mechanical properties of CPB in the early hydration time. Strong alkaline activators, which are generally composed of sodium/potassium hydroxide (NaOH/KOH) and sodium/potassium water glass (nSiO Na O) in 2· 2 fixed proportions, are an effective option. The high pH of the alkaline solution can break down covalent Si–O–Si and Al–O–Si bonds, providing more dissolved silica and aluminum to for hydration reaction [29,30]. However, the production of NaOH/KOH and nSiO Na O consumes a lot of energy 2· 2 and can release an immense amount of CO2, so it does not lead to environmental protection. The use of quicklime, a weakly alkaline mineral promotor, is a relatively simple and economical alternative. Shi [31] first investigated the effect of quicklime on natural pozzolan/cement systems in 2001, with promising results regarding the early-stage promoting effect of quicklime. Antiohos et al. [32,33] experimentally explored the influence of industrially produced quicklime on the pozzolanic reaction rates of various fly ash/cement systems. An addition of 5% quicklime was found effective only during the initial hydration stage, and almost no accelerating effect was detected with an increased dose. However, these studies have primarily focused on a small addition of FA. Quicklime activation is not enough for strength development of a cementitious system, with FA content exceeding 30%. Calcium sulfoaluminate (CSA)˙ cement has been widely used in concrete repair engineering due to its high initial compressive strength, microexpansion, low-temperature adaptability, and poor permeability [34,35]. Ye’elimite is the main mineral component of CSA.˙ After hydration, it forms monosulfoaluminate (AFm) or ettringite (AFt) based on the concentration of sulfate [34–36]. The rapid hydration of ye’elimite shortens the setting time and increases the initial strength rapidly. All these advantages support the utilization of CSA˙ and large amounts of fly ash as supplementary cementitious materials in CPB. However, previous studies failed to evaluate the synergistic effect of quicklime and CSA˙ on the properties of CPB with high amounts of FA. Therefore, in this paper, CSA˙ and quicklime were used as mineral promotors to increase the hydration rate of cementitious systems. Slag cement (SC), which consists of cement clinker, granulated blast-furnace slag, and gypsum, was selected as a basic cementitious material because it is lower in cost and emits less CO2 than Ordinary Portland cement (OPC). Fluidity and unconfined compressive strength (UCS) tests were conducted to evaluate the workability of CPB. The hydration evolution of solid phases was determined using several laboratory techniques to suggest a mechanism for experimental findings.

2. Materials and Methods

2.1. Raw Materials The basic cementitious material used in this study was PSA 3.25 SC (Shandong Juzhou Cement Co., Ltd., Rizhao, China). The particle size distribution and chemical composition of SC were measured by Malvern Mastersizer 2000 (Malvern Instruments Co. Ltd., Malvern, UK) and X-Ray Fluorescence Spectrometer, respectively. Figure1a shows that the SC of D 20,D50, and D80 are 6.72, 18.7, and 43.2 µm, respectively. D20,D50, and D80 represent the critical diameters when the cumulative volume of solid particles reaches 20%, 50%, and 80%, respectively. The summarized results of chemical composition in Table1 show that SC contains 55.24% calcium oxide (CaO), 26.15% silicon dioxide (SiO 2), and Materials 2020, 13, x FOR PEER REVIEW 3 of 21 and are shown in Figure 2. As shown in the figure, SC is rich in crystalline calcite, quartz, mullite, tricalcium silicate, and dicalcium silicate. Despite its high cost, CṠA (Shandong Juzhou Cement Co., Ltd., Rizhao, China) was used as supplementary cementitious material (limited addition) due to its superior gelling property. The physical and chemical properties were characterized and shown in Figure 1a and Table 1, respectively. The D20, D50, and D80 of CṠA are 8.97, 18.36, and 31.1 μm, respectively. It can be concluded that CṠA has a relatively narrow particle size distribution compared to SC. The main chemical composition of CṠA is 53.91% CaO, 13.6% SiO2, 15.13% Al2O3, and 10.35% SO2. The identified mineral phases of CṠA shown in Figure 2 are ye’elimite, anhydrite, calcite, tricalcium silicate, and dicalcium silicate. In addition, quicklime (Liaoning Bangka Calcium Industry Co., Ltd., Benxi, China) was also adopted here as a mineral activator. FA was used as a supplementary cementitious material to replace 40% SC. FA has a coarser particle size with the D20, D50, and D80 of 12.06, 34.73, and 82.35μm, respectively. It contains 55.51% SiO2, 30.82% Al2O3, and only 4.08% CaO. XRD analysis results indicate that FA is composed of crystallineMaterials 2020 quartz,, 13, 4018 mullite, and magnetite. In order to eliminate the influence of tailing minerals3 of on 20 the properties of CPB, silica tailings (ST) with 96.46% SiO2 was used as aggregate. ST has the D20, D50, and D80 of 28.45, 58.09, and 93.21 μm, respectively. The morphology of SC, CṠA, quicklime, and FA was8.83% detected aluminum using oxide SEM. (Al As2 Oshown3). The in mineral Figure 3, phases FA has of a SCspherical were identiﬁed shape, and by SC, XRD CṠ andA, and are quicklime shown in haveFigure irregular2. As shown shapes. in the ﬁgure, SC is rich in crystalline calcite, quartz, mullite, tricalcium silicate, and dicalcium silicate.

(a)100 (b)100 10 SC Cumulative volume 80 CṠA Volume density FA 80 8

60 60 6

40 40 4

Volume density (%)

Cumulative volume (%) Cumulative volume (%) 20 20 2

0 0 0 0.1 1 10 100 1000 0.1 1 10 100 1000 Particle size (mm) Particle szie (mm) Figure 1. Particle size distribution of (a) Slag cement (SC), CSA,˙ and ﬂy ash (FA); (b) silica tailings (ST). Figure 1. Particle size distribution of (a) Slag cement (SC), CṠA, and fly ash (FA); (b) silica tailings (ST). Table 1. Chemical composition of FA, CSA,˙ SC, and ST (%).

˙ ElementTable 1. Chemical SC compositionC ofSA FA, CṠA, SC, FAand ST (%). ST CaO 55.24 53.91 4.08 0.03 Element SC CṠA FA ST SiO2 26.15 13.60 55.51 96.46 Al2O3 CaO 8.8355.24 15.3153.91 4.08 30.820.03 1.87 MgOSiO2 2.9326.15 3.0113.60 55.51 0.6396.46 – Fe2O3 Al2O3 3.028.83 2.0315.31 30.82 4.021.87 0.07 Na O 0.02 0.34 0.04 0.01 2 MgO 2.93 3.01 0.63 – K2O 1.27 0.81 2.49 1.30 Fe2O3 3.02 2.03 4.02 0.07 SO2 1.77 10.35 0.79 0.03 SSA a (m2/kg) Na2O528.42 0.02 496.930.34 0.04 297.910.01 227.27 K2O a 1.27 0.81 2.49 1.30 Materials 2020, 13, x FOR PEER REVIEW speciﬁc surface area. 4 of 21 SO2 1.77 10.35 0.79 0.03 SSA a (m2/kg) 528.42 496.93Q 297.91 227.27 FA CṠA a specificQ surface area. SC M M M M a M M Q A Y C2S, C3S

C Y C3S Y C

C3S

5 10 15 20 25 30 35 40 45

2-Theta Figure 2. 2. XRDXRD patterns patterns of ofFA, FA, CṠ CA,SA,˙ and and SC (Q SC— (Q—quartz,quartz, A—Anhydrite, A—Anhydrite, C—calcite, C—calcite, Y—ye Y—ye’elimite,’elimite, M—

mullite,M—mullite, Ma— Mmagnetite).a—magnetite).

Despite its high cost, CSA˙ (Shandong Juzhou Cement Co., Ltd., Rizhao, China) was used (a) 10μm (b) 10μm as supplementary cementitious material (limited addition) due to its superior gelling property. The physical and chemical properties were characterized and shown in Figure1a and Table1,

Figure 3. SEM images of (a) SC, (b) CṠA, (c) quicklime, and (d) FA (Mag = 3000×).

2.2. Sample Preparation Six groups of pure blended binders were prepared according to the detailed mixing ratios shown in Table 2. The first sample series was prepared to explore the influence of CṠA on hydration reactions. The mass ratios of SC:CṠA:FA:quicklime were 10:0:8:2, 9:1:8:2, and 7:3:8:2, and the sample numbers were defined as SCA0, SCA5, and SCA15, respectively. A second sample series was prepared to confirm the effect of quicklime on the hydration reaction. The mass ratios of SC:CṠA:FA:quicklime were 9:2:8:1, 8:2:8:2, and 7:2:8:3, and the sample numbers were defined as SCC5, SCC10, and SCC15. A control sample was also prepared and named SC100 for comparison. The water/binder ratio was 0.43. Accordingly, seven groups of CPB samples were prepared using blended binders. The binder/ST ratio was 0.2, and the solid concentration was 70% by mass. After mixing for 5 min, a portion of the fresh backfill was subjected to the fluidity test, and the remaining was poured into plastic cylinder molds (5 cm in diameter × 10 cm in height). It should be noted that both of the upper and lower ends of the molds are sealed. All samples were cured at room temperature (20 ± 2 °C) with a relative humidity of 95 ± 2%.

Materials 2020, 13, x FOR PEER REVIEW 4 of 21

Materials 2020, 13, 4018 4 of 20 Q FA CṠA Q SC respectively. The D ,D , and D of CSA˙ are 8.97, 18.36,M and 31.1 µm, respectively. It can be 20 50 80 M M M a M Q ˙ M concluded that CSA has a relatively narrow particleA size distribution compared to SC. The main Y C S, C S chemical composition of CSA˙ is 53.91% CaO, 13.6% SiO2, 15.13%2 3 Al2O3, and 10.35% SO2. The identiﬁed mineral phases of CSA˙ shown in Figure2 are ye’elimite,C anhydrite, calcite, tricalcium silicate, and Y C3S dicalcium silicate. In addition, quicklime (Liaoning Bangka Calcium Industry Co., Ltd., Benxi, China) Y C was also adopted here as a mineral activator. FA was used as a supplementary cementitious material to replace 40% SC. FA has a coarser particle C3S size with the D20,D50, and D80 of 12.06, 34.73, and 82.35µm, respectively. It contains 55.51% SiO2, 30.82% Al2O3, and only 4.08% CaO. XRD analysis results indicate that FA is composed of crystalline quartz, mullite, and magnetite. In order to eliminate the inﬂuence of tailing minerals on the properties 5 10 15 20 25 30 35 40 45 of CPB, silica tailings (ST) with 96.46% SiO2 was used as aggregate. ST has the D20,D50, and D80 of 28.45, 58.09, and 93.21 µm, respectively. The2-Theta morphology of SC, CSA,˙ quicklime, and FA was ˙ detectedFigure using 2. XRD SEM. patterns As shown of FA, inCṠ FigureA, and 3SC, FA (Q— hasquartz, a spherical A—Anhydrite, shape, and C— SC,calcite, C SA, Y— andye’elimite, quicklime M— have irregular shapes. mullite, Ma—magnetite).

(a) 10μm (b) 10μm

FFigureigure 3. SEMSEM images images of of ( (aa)) SC, SC, ( (bb))C CṠSA,˙A, ( (cc)) quicklime, quicklime, and and ( (d)) FA FA (Mag = 30003000×).). × 2.2.2.2. Sample Sample Preparation Preparation SSixix groups groups of of pure pure blended blended binders binders were were prepared prepared according according to to the the detailed detailed mixing mixing ratios ratios shown shown ˙ inin Table2 . 2. The The first first sample sample series series was was prepared prepared to explore to explore the influence the influence of C SA ofon C hydrationṠA on hydration reactions. ˙ reactions.The mass The ratios mass of SC:C ratiosSA:FA:quicklime of SC:CṠA:FA:quicklime were 10:0:8:2, were 9:1:8:2, 10:0:8:2, and 9:1:8:2, 7:3:8:2, and and 7:3:8:2, the sample and the numbers sample numberswere defined were as defined SCA0, SCA5,as SCA0, and SCA5, SCA15, and respectively. SCA15, respectively. A second sample A second series sample was prepared series was to ˙ preparedconfirm the to eff confirmect of quicklime the effect on of the quicklime hydration onreaction. the hydration The mass reaction.ratios of SC:C The SA:FA:quicklime mass ratios of SC:CwereṠ 9:2:8:1,A:FA:quicklime 8:2:8:2, and were 7:2:8:3, 9:2:8:1, and 8:2:8:2, the sample and numbers 7:2:8:3, and were the defined sample as numbers SCC5, SCC10, were anddefined SCC15. as SCC5,A control SCC10, sample and was SCC15. also prepared A control and sample named was SC100 also for prepared comparison. and named The water SC100/binder for ratiocomparison. was 0.43. TheAccordingly, water/binder seven ratio groups was of 0.43. CPB Accordingly, samples were seven prepared groups using of blendedCPB samples binders. were The prepared binder/ST using ratio blendedwas 0.2, andbinders. the solidThe binder/ST concentration ratio was was 70% 0.2, byand mass. the solid After concentration mixing for 5 min, was a70% portion by mass. of the After fresh mixingbackfill for was 5 min subjected, a portion to the of fluidity the fresh test, backfill and the was remaining subjected was to th pourede fluidity into test, plastic and cylinderthe remaining molds (5 cm in diameter 10 cm in height). It should be noted that both of the upper and lower ends of the was poured into plastic× cylinder molds (5 cm in diameter × 10 cm in height). It should be noted that molds are sealed. All samples were cured at room temperature (20 2 C) with a relative humidity of both of the upper and lower ends of the molds are sealed. All samples± were◦ cured at room temperature 95 2%. (20 ± 2 °C) with a relative humidity of 95 ± 2%.

Materials 2020, 13, 4018 5 of 20

Table 2. Mixing ratios of blended binders (%).

Blended Binder SC CSA˙ FA Quicklime SC100 100 0 0 0 SCA0 50 0 40 10 SCA5 45 5 40 10 SCA15 35 15 40 10 SCC5 45 10 40 5 SCC10 40 10 40 10 SCC15 35 10 40 15

2.3. Testing Methods

2.3.1. Spread Diameter The slump test is convenient and is often used to assess the workability of fresh concrete. Recently, this method has been increasingly adopted to measure the fluidity of fresh backfills. According to the literature [11], the spread diameter (SD) obtained from the slump test is determined by the yield shear stress, an important rheological parameter of fluids. Hence, in this study, a mini-cone (5 cm in top diameter, 10 cm in bottom diameter, and 15 cm in height) was selected to measure the diameter of the spread.

2.3.2. Unconfined Compressive Strength After reaching the pre-determined hydration time (3, 7, and 28 days), CPB samples were subjected to the UCS test based on ASTM C39 standard test procedure [37]. A computer-controlled loading machine (Humboldt HM-5030(Raleigh, NC, USA)) was used herein. The load capacity was 50 kN, and the load-deformation rate was 1 mm/min. Both ends of the cylindrical specimen were kept as flat as possible to reduce edge effects. During the compressive test, the peak stress corresponds to the mechanical strength of the sample. If the measurements of the two samples differ by more than 15%, a third sample needs to be measured to verify its accuracy. Hence, at least two samples were measured for each recipe, and only the mean value was considered the UCS for CPB samples.

2.3.3. Pore Structure The pore structure of the samples was analyzed by the MIP method using Micromeritics’ AutoPore IV 9500 (Atlanta, GA, USA). A piece of sample less than 15 15 15 mm in volume was taken from × × the crushed specimen whose location must be far enough from the shear plane [19]. Samples were immersed in isopropanol to stop the hydration for 12 h, and then dried in vacuum drying oven at 45 ◦C for 24 h. The testing range for pore diameter was 3–1000 µm.

2.3.4. Hydration Heat Approximately 6 g of binder paste with a water/binder ratio of 0.43 was loaded into a glass vial and mixed evenly using a slow stirring mixer. In this study, isothermal calorimetry was conducted using a TAM Air 8-channel microcalorimeter (New Castle, DE, USA) to confirm the heat flow of the blended binder. The ambient temperature was 20 2 C, and 72 h of experimental data was recorded. ± ◦ 2.3.5. Hydration Evolution As discussed in the literature [25,26,38,39], the hydration process of pure binder pastes cured for 3, 7, and 28 days, respectively, was stopped by isopropyl alcohol. Samples were pulverized after drying in a vacuum drying oven to a constant weight at 45 ◦C. Several techniques were employed to characterize the hydration evolution: (1) XRD patterns of binder pastes were recorded using a XRD 7000 diffractometer (Tokyo, Japan). The test interval angle was 5–50◦, and the scan step size was 5◦/min. Materials 2020, 13, 4018 6 of 20

(2) Phase composition analysis of binder pastes was performed using thermogravimetry analysis STA409PC (Selb, Germany) coupled with diﬀerential thermogravimetry (DTG). About a 30 mg sample was loaded into an alumina crucible and heated to 1000 ◦C in a nitrogen atmosphere (15 ◦C/min) [25,26]. The amount of hydrate water (H) and portlandite (CH) can be expressed as:

M M H = 50 − 550 100% (1) M550 ·

M M 74 CH = 400 − 550 100% (2) M550 ·18· where M50, M400, and M550 correspond to the residual mass of samples at temperatures of 50 ◦C, 400 ◦C, and 550 ◦C, respectively. (3) A SEM (Hitachi S-3400N (Tokyo, Japan)) was used to observe the morphology of samples at an accelerating voltage of 15 keV. Energy-dispersive X-ray (EDX) spectroscopy was used to analyze the elemental composition. More details can be found in reference [40].

3. Results

3.1. Spread Diameter Figure4 shows the e ffect of the addition of quicklime and CSA˙ on the flowability of fresh CPB. The addition of CSA˙ slightly increased SD, but when SC was replaced with quicklime, the fluidity decreased to some extent at the initial setting time of 5 min. For instance, SD increased from 29.47 cm to 30.86 cm with an increment of 1.39 cm for the samples with the addition of CSA˙ increasing from 0 to 15%, but decreased from 32.41 cm to 26.73 cm with a reduction of 5.68 cm for the samples with addition of quicklime increasing from 5% to 15%.The SD gain of fresh sample is attributed to the large amount of ultrafine particles with a diameter of less than 20 µm in CSA.˙ Partial replacement of the SC increased the packing density of the fresh sample. According to the literature [41], the amount of free water that imparts fluidity to a fresh slurry is positively related to its packing density. In this case, the addition of CSA˙ increased the amount of free water for flowability. Additionally, the decrease in SD with increasing doses of quicklime is mainly associated with water-consuming chemical reactions and irregular shapes, as shown in Figure3. Regardless of the e ffects of CSA,˙ quicklime, or both, the fluidity performance of fresh samples with 40% FA is better than with pure SC. This is due to the lubrication Materials 2020, 13, x FOR PEER REVIEW 7 of 21 effect of spherical FA (Figure3).

(a)32 (b)35

SC100 SC100 SCA0 30 SCC5 28 SCA5 SCC10 SCA15 SCC15 25 24 20

20 Spread diameter (cm) Spread diameter (cm) 15

16 10 5 10 30 60 5 10 30 60 Time (min) Time (min) Figure 4. Spread diameter of fresh cemented paste backﬁll (CPB) with diﬀerent doses of (a)CSA˙ and Figure 4. Spread diameter of fresh cemented paste backfill (CPB) with different doses of (a) CṠA and (b) quicklime. (b) quicklime.

3.2. Strength Development Figure 5 shows the strength development of CPB hardened with different doses of CṠA and quicklime. It was found that when the amount of FA (40%) that replaces SC was large, the mechanical performance noticeably decreased irrespective of the curing time. In contrast, the addition of CṠA and quicklime had a positive effect on the mechanical properties. As shown in Figure 5a, compared with SCA0, the addition of 5% CṠA (SCA5) caused a 0.21 MPa increase over a 3-day curing time, and a 15% SC replacement (SCA15) caused a 0.39 MPa increase. This inconsistent improvement in strength (nearly twice) relative to the amount of CṠA added (three times) is potentially attributed to insufficient drainage due to more ultrafine particles (Figure 1a) [3]. As shown in Figure 5b, the UCS value of samples with 10% SC replacement (SCC10) increased from 0.53 MPa to 0.66 MPa with an increase of 0.13 MPa in the hydration time of 3 days, while the addition (SCC15) of 15% quicklime caused an improvement of 0.1 MPa. When quicklime reacts with water, it releases a large amount of hydroxyl ions, leading to an alkaline environment and activating FA [32]. Amorphous silicon and aluminum liberated by activated FA particles consume portlandite to anticipate in hydration reaction, producing ettringite and C–S–H/C–A–S–H. The addition of a 10% quicklime can accelerate the pozzolanic reaction of FA and compensate for the loss of hydration products due to the replacement of 10% SC. When the replacement of SC with quicklime is increased to 15%, the limited solubility of calcium hydroxide in water restricts the acceleration effect of FA on the pozzolanic reactions. It was also found that the improvement in curing time from 7 to 28 days compared to the increase in UCS at seven days of hydration time was approximately similar regardless of binder types. In SC100 samples, SC consists of Portland cement clinker, 20–50% granulated blast furnace slag, and an appropriate amount of gypsum. Coupled with the filler effect of granulated blast-furnace slag particles, ettringite and gel products (e.g., C–S–H), which result from the hydration of aluminates and silicates, strengthen the hardening process at the curing time of 7 days. Although slag is highly active and can quickly participate in the hydration reaction, it does not perform well in the subsequent hydration process compared to the hydration rate of cement clinker. This results in a slow increase in strength in the next 7 to 28 days of curing time. By adding CṠA and quicklime in the blended binders, ettringite develops as the main hydration product and fills the voids that exist in solid skeletons. With extra sulfate ions, quicklime can be hydrated to form dihydrate gypsum, which also leads to the hardening of CPB [42]. However, when sulfate ions are consumed, monosulfoaluminate is formed instead of ettringite [43,44].As a result of low crystallinity, the filler effect of AFm on CPB is low [25]. Furthermore, FA activated by quicklime hydration participates in the hydration reaction and decreases pH with increasing curing time. This is not desirable for

Materials 2020, 13, 4018 7 of 20

The SD of all fresh samples clearly decreased with increased curing time. This is associated with the initial hydration reaction. Tricalcium silicate, tricalcium aluminate, and calcium sulfoaluminate quickly dissolved in water and hydrated to form AFt/AFm and gel products [36]. The reduction in SD is based on the doses of CSA˙ and quicklime added. In the SC100 sample, the SD experienced a drop of 5.23 cm with the hydration time increasing from 5 to 60 min. With the addition of 15% CSA˙ and 10% quicklime (SCA15), SD reduced by 14.42 cm. Ye’elimite, the most reactive component of CSA,˙ reacts with water to produce cementitious materials in a short time, promoting the solidiﬁcation of fresh CPB signiﬁcantly. Increasing the amount of quicklime from 5% to 15% resulted in similar reductions in SD for SCC5, SCC10, and SCC15 with increased setting time, higher than the SC100 samples. This is because the dissolution of the quicklime that is associated with releasing heat can partially increase the hydration rates of silicates, aluminates, and sulfoaluminates.

3.2. Strength Development Figure5 shows the strength development of CPB hardened with di fferent doses of CSA˙ and quicklime. It was found that when the amount of FA (40%) that replaces SC was large, the mechanical performance noticeably decreased irrespective of the curing time. In contrast, the addition of CSA˙ and quicklime had a positive effect on the mechanical properties. As shown in Figure5a, compared with SCA0, the addition of 5% CSA˙ (SCA5) caused a 0.21 MPa increase over a 3-day curing time, and a 15% SC replacement (SCA15) caused a 0.39 MPa increase. This inconsistent improvement in strength (nearly twice) relative to the amount of CSA˙ added (three times) is potentially attributed to insufficient drainage due to more ultrafine particles (Figure1a) [ 3]. As shown in Figure5b, the UCS value of samples with 10% SC replacement (SCC10) increased from 0.53 MPa to 0.66 MPa with an increase of 0.13 MPa in the hydration time of 3 days, while the addition (SCC15) of 15% quicklime caused an improvement of 0.1 MPa. When quicklime reacts with water, it releases a large amount of hydroxyl ions, leading to an alkaline environment and activating FA [32]. Amorphous silicon and aluminum liberated by activated FA particles consume portlandite to anticipate in hydration reaction, producing Materials 2020, 13, x FOR PEER REVIEW 8 of 21 ettringite and C–S–H/C–A–S–H. The addition of a 10% quicklime can accelerate the pozzolanic reaction ofsubsequent FA and compensate pozzolanic for reactions the loss [42, of hydration45–47]. Hence, products inadequate due to thehydration replacement of blended of 10% binders SC. When causes the replacementa slow increase of SCin mechanical with quicklime performance is increased within to 15%, the hydration the limited time solubility of 7–28 of days calcium. hydroxide in water restricts the acceleration effect of FA on the pozzolanic reactions.

(a)2.5 (b)2.5 SC100 SCA0 SCA5 SCA15 SC100 SCC5 SCC10 SCC15

2.0 2.0

1.5 1.5

1.0 1.0

UCS (MPa)

0.5 0.5

0.0 0.0 3 7 28 3 7 28 Curing time (d) Curing time (d) Figure 5. Unconfined compressive strength (UCS) variation of hardened CPB with different doses of Figure 5. Unconfined compressive strength (UCS) variation of hardened CPB with different doses of (a)CSA˙ and (b) quicklime. (a) CṠA and (b) quicklime. It was also found that the improvement in curing time from 7 to 28 days compared to the 3.3. Pore Structure increase in UCS at seven days of hydration time was approximately similar regardless of binder types. In SC100Figures samples, 6 and SC 7 show consists variations of Portland in the cement incremental clinker, pore 20–50% volume granulated curves blast of CPB furnace at various slag, and doses an appropriateof CṠA and amountquicklime of gypsum.over a 28 Coupled-day curing with time. the fillerAs shown effect ofin granulatedthe figures, blast-furnace the pore size slag distribution particles, of the SCA0-15 and SCC5-15 samples is relatively scattered compared to the SC samples. This is mainly associated with the filler effect of FA and leads to the refinement of the pore structure [48]. However, large amounts of FA cannot react completely, resulting in lower content of hydration products (e.g., AFt, AFm, and C–S–H/C–A–S–H) and higher pore volumes. This can be used to explain the weak mechanical performance of CPB with 40% FA, as shown in Figure 5. In addition, it was clearly observed that the increased content of CṠA and quicklime contributed to the improvement of the pore structure and decreased its total volume. This is because CṠA hydrates rapidly, and hydrated quicklime activates the pozzolanic reaction of FA, as explained in Section 3.2. Additional hydration products can be formed to fill the voids. Based on the references [49–52], pores with diameter smaller than 500 nm are micropores and mespories, and pores with diameter larger than 1000 nm are large capillary pores. Hence, pores are classified into three types according to the pore size in this study: (1) type I: < 500 nm; (2) type II: 500– 1000 nm; (3) type III: > 1000 nm. The volumes of the three categories of pores are also listed in Figures 6 and 7. As shown in the figures, SC100 samples had the largest volume of type I pores and the smallest volume of type III pores. Pores with diameters greater than 1000 nm are called “harmful pores” that significantly influence strength development [53–55]. This verifies the result that SC100 backfill has the best mechanical performance. As the amount of CṠA and quicklime increased, the volume of the type III pores decreased dramatically, which was associated with the hydration reaction rate. The addition of CṠA and quicklime accelerates the hydration and thus improves the mechanical properties of hardened backfill. In order to investigate the relationship between pore volume and mechanical strength, a linear regression between strength and type I pore volume, type II pore volume, type III pore volume, and total pore volume was calculated (Figure 8). It could be found easily that UCS of samples had a positive relationship with the volume of type I pores, while a negative relationship existed between the strength and type III pore volume. This can be explained by the influence of “harmful pores”. Their corresponding correlation coefficients are 0.521, 0.235, 0.654, and 0.887, indicating that the total pore volume (porosity) is responsible for the mechanical performance of hardened CPB.

Materials 2020, 13, 4018 8 of 20 ettringite and gel products (e.g., C–S–H), which result from the hydration of aluminates and silicates, strengthen the hardening process at the curing time of 7 days. Although slag is highly active and can quickly participate in the hydration reaction, it does not perform well in the subsequent hydration process compared to the hydration rate of cement clinker. This results in a slow increase in strength in the next 7 to 28 days of curing time. By adding CSA˙ and quicklime in the blended binders, ettringite develops as the main hydration product and fills the voids that exist in solid skeletons. With extra sulfate ions, quicklime can be hydrated to form dihydrate gypsum, which also leads to the hardening of CPB [42]. However, when sulfate ions are consumed, monosulfoaluminate is formed instead of ettringite [43,44].As a result of low crystallinity, the filler effect of AFm on CPB is low [25]. Furthermore, FA activated by quicklime hydration participates in the hydration reaction and decreases pH with increasing curing time. This is not desirable for subsequent pozzolanic reactions [42,45–47]. Hence, inadequate hydration of blended binders causes a slow increase in mechanical performance within the hydration time of 7–28 days.

3.3. Pore Structure Figures6 and7 show variations in the incremental pore volume curves of CPB at various doses of CSA˙ and quicklime over a 28-day curing time. As shown in the figures, the pore size distribution of the SCA0-15 and SCC5-15 samples is relatively scattered compared to the SC samples. This is mainly associated with the filler effect of FA and leads to the refinement of the pore structure [48]. However, large amounts of FA cannot react completely, resulting in lower content of hydration products (e.g., AFt, AFm, and C–S–H/C–A–S–H) and higher pore volumes. This can be used to explain the weak mechanical performance of CPB with 40% FA, as shown in Figure5. In addition, it was clearly observed that the increased content of CSA˙ and quicklime contributed to the improvement of the pore structure and decreased its total volume. This is because CSA˙ hydrates rapidly, and hydrated quicklime activates the pozzolanic reaction of FA, as explained in Section 3.2. Additional hydration products can be formed to fill the voids. Based on the references [49–52], pores with diameter smaller than 500 nm are micropores and mespories, and pores with diameter larger than 1000 nm are large capillary pores. Hence, pores are classified into three types according to the pore size in this study: (1) type I: < 500 nm; (2) type II: 500–1000 nm; (3) type III: > 1000 nm. The volumes of the three categories of pores are also listed in Figures6 and7. As shown in the figures, SC100 samples had the largest volume of type I pores and the smallest volume of type III pores. Pores with diameters greater than 1000 nm are called “harmful pores” that significantly influence strength development [53–55]. This verifies the result that SC100 backfill has the best mechanical performance. As the amount of CSA˙ and quicklime increased, the volume of the type III pores decreased dramatically, which was associated with the hydration reaction rate. The addition of CSA˙ and quicklime accelerates the hydration and thus improves the mechanical properties of hardened backfill. In order to investigate the relationship between pore volume and mechanical strength, a linear regression between strength and type I pore volume, type II pore volume, type III pore volume, and total pore volume was calculated (Figure8). It could be found easily that UCS of samples had a positive relationship with the volume of type I pores, while a negative relationship existed between the strength and type III pore volume. This can be explained by the influence of “harmful pores”. Their corresponding correlation coefficients are 0.521, 0.235, 0.654, and 0.887, indicating that the total pore volume (porosity) is responsible for the mechanical performance of hardened CPB. Materials 2020, 13, x FOR PEER REVIEW 9 of 21 Materials 2020, 13, x FOR PEER REVIEW 9 of 21

Materials 20202020, 1313, 4018x FOR PEER REVIEW 9 of 21 20 (a)0.08 (b)0.25 (a)0.08 SC100 SCA0 SCA5 SCA15 (b)0.25 SC100 SCA0 SCA5 SCA15 SC100 SCA0 SCA5 SCA15 SC100 SCA0 SCA5 SCA15 0.20 (a)0.08 (b)0.25 0.06 ) SC100 SCA0 SCA5 SCA15 0.20 SC100 SCA0 SCA5 SCA15

0.06 )

ml/g ( 0.200.15

ml/g

( 0.040.06 ) 0.15 0.04

ml/g 0.10 ( 0.15 0.10 0.04 0.02 Porevolume

0.02 Porevolume 0.100.05 Incremental pore volume (ml/g) 0.05

Incremental pore volume (ml/g) 0.000.02 Porevolume 10 100 1000 10000 0.00 0.00 0.05 <500 500-1000 >1000

Incremental pore volume (ml/g) 10 100 1000 10000 0.00 Pore diameter (nm) <500 Pore diameter500-1000 (nm) >1000 0.00 Pore diameter (nm) Pore diameter (nm) 10 100 1000 10000 0.00 Figure 6. Pore size distribution results of hardened CPB with different<500 doses500-1000 of CṠA at the>1000 hydration Pore diameter (nm) Pore diameter (nm) timeFigure of 6.28 Pore days size: (a) distributionthe incremental results pore of volumehardened, and CPB (b )with the volumedifferent of doses different of C poreṠA at types the hydration. Figuretime of 6.28 days: (a) the incremental pore volume, and (b) the volume of different pore˙ types. Figure 6. Pore size distribution results of hardened CPB with different diﬀerent doses of CSAṠA at the hydration time of 28 days: (a) the incremental pore volume, and (b) the volume of diﬀerent pore types. time of 28 days: (a) the incremental pore volume, and (b) the volume of different pore types. (a)0.08 (b)0.25 SC100 SCC5 SCC10 SCC15 (a)0.08 (b)0.25 SC100 SCA0 SCA5 SCA15 SC100 SCC5 SCC10 SCC15 SC100 SCA0 SCA5 SCA15 0.20 (a)0.060.08 (b)) 0.25 SC100 SCC5 SCC10 SCC15 0.20 0.06 SC100 SCA0 SCA5 SCA15

)

ml/g ( 0.15

ml/g 0.20 0.040.06 ( ) 0.15 0.04

ml/g 0.10 ( 0.15 0.10

0.020.04 Pore volume 0.05 0.02 Pore volume 0.10 Incremental pore volume (ml/g) 0.05

Incremental pore volume (ml/g) 0.000.02 Pore volume 0.00 10 100 1000 10000 0.05 <500 500-1000 >1000 0.00 0.00 Incremental pore volume (ml/g) 10 Pore 100diameter (nm)1000 10000 <500 Pore diameter500-1000 (nm) >1000 Pore diameter (nm) 0.00 Pore diameter (nm) 0.00 10 100 1000 10000 <500 500-1000 >1000 Figure 7. Pore size distribution results of hardened CPB with different doses of quicklime at the Figure 7. Pore size distributionPore diameter (nm) results of hardened CPB with diﬀerentPore diameter doses of (nm) quicklime at the hydrationFigure 7. Poretime of size 28 days distribution: (a) the incrementalresults of hardened pore volume, CPB and with (b different) the volume doses of different of quicklime pore types. at the hydration time of 28 days: (a) the incremental pore volume, and (b) the volume of diﬀerent pore types. hydration time of 28 days: (a) the incremental pore volume, and (b) the volume of different pore types. Figure 7. Pore size distribution results of hardened CPB with different doses of quicklime at the 2.4 hydration time of 28 days: (a) the incremental Total pore volume pore volume, Pore volume andof < 500 (b nm) the volume of different pore types. 2.4 Pore volume of 500-1000 nm Pore volume of >1000 nm Total pore volume Pore volume of < 500 nm Pore volume of 500-1000 nm Pore volume of >1000 nm 2.4 R2=0.235 2.0 Total pore volume Pore volume of < 500 nm 2 2.0 PoreR =0.235 volume of 500-1000 nm Pore volume of >1000 nm R2=0.887 R2=0.235 2.0 R2=0.887 1.6

1.6 2 R =0.521 R2=0.887

UCS (MPa)UCS R2=0.521

UCS (MPa)UCS 1.6 1.2 1.2 R2=0.521

UCS (MPa)UCS R2=0.654 1.2 2 0.8 R =0.654 0.80.00 0.07 0.14 0.21 0.28 0.35 0.00 0.07 0.14 R2=0.6540.21 0.28 0.35 Pore volume (ml/g) 0.8 Pore volume (ml/g) 0.00 0.07 0.14 0.21 0.28 0.35 FFigureigure 8. Relationship between UCS and pore volume. Figure 8. RelationshipPore between volume UCS (ml/g) and pore volume. 3.4. Hydration Heat 3.4. Hydration Heat Figure 8. Relationship between UCS and pore volume. 3.4. Hydration Heat FigureFigure9 9shows shows the the normalized normalized heat heat flow flow of blended of blended binder binder pastes, pastes, and SC100 and (control) SC100 (control) is included is forincluded3.4. comparison.HydrationFigure for 9comparison. Heatshows It can the be clearlynormalizedIt can be seen clearly that heat the seen flow first that of peak blendedthe of first the peak binder heat flowof pastes,the of heat binder and flow pastes SC100 of binder appeared (control) pastes in is theincluded first 30 for min comparison. (Figure9a). It This can isbe mainly clearly attributed seen that tothe the first dissolution peak of the of soluble heat flow particles, of binder including pastes appeaFigurered in 9 theshows first the 30 minnormalized (Figure heat9a). Thisflow is of mainly blended attributed binder pastes, to the and dissoluti SC100on (control) of soluble is cementappeared clinker, in the ye’elimite, first 30 andmin quicklime (Figure 9a [27). ]. This The is SC100 mainly paste attributed released about to the 0.014 dissoluti W/g ofon heat of soluble mostly includedparticles, forincluding comparison. cement It clinker,can be clearlyye’elimite, seen and that quicklime the first peak [27]. of The the SC100 heat flow paste of released binder pastesabout dueparticles, to the including wetting of cement tricalcium clinker, aluminate ye’elimite, (C A) and and quicklime the initial [27]. precipitation The SC100 of paste ettringite. released When about SC appeared in the first 30 min (Figure 9a). This3 is mainly attributed to the dissolution of soluble was replaced with 40% FA, the exothermic peak value showed a significant decrease. Coupled with particles, including cement clinker, ye’elimite, and quicklime [27]. The SC100 paste released about

Materials 2020, 13, x FOR PEER REVIEW 10 of 21

0.014 W/g of heat mostly due to the wetting of tricalcium aluminate (C3A) and the initial precipitation of ettringite. When SC was replaced with 40% FA, the exothermic peak value showed a significant decrease. Coupled with the reaction of quicklime and water, increasing the dose of CṠA from 0% to 15% increased the heat peak from 0.0095 to 0.013 W/g. This increase results from faster dissolution and hydration of ye’elimite. Increasing the dose of quicklime from 5% to 15% increased the first exothermic peak from 0.01 to 0.022 W/g with the accelerating effect of CṠA. This improvement is due to the more vigorous reaction of quicklime forming portlandite. Then, as the concentration of calcium, sulfate, and hydroxyl ions increased, a decrease in heat flow occurred [42]. The initial alkaline environment can be strengthened by the increased amount of portlandite generated, promoting the activity of FA [56]. The dissolution of amorphous silicon and aluminum from active FA contributes to the heat flow. Hence, the SCC15 sample exhibited the highest rate of heat release in the first 1 h. An exothermic acceleration phase appeared after 4h, which is associated with the dissolution of tricalcium silicate (C3S) and anhydrite, synchronous with the precipitation of portlandite and C–S–H phases [57]. With the increased formation rate of ettringite, the concentration of sulfate ions decreased, leading to the release of sulfate ions originally attached to tricalcium aluminate. This leads Materialsto a further2020, 13dissolution, 4018 and hydration reaction of tricalcium aluminate [57]. None of the heat10 flow of 20 curves for the blended binders containing additional FA showed sharp peaks, but wider shoulders. This different behavior is associated with the retarding effect of FA reported in the literature [58,59]. ˙ theCumulative reaction ofheat quicklime release from and water,the first increasing 6 h of hydration the dose was of C higherSA from for 0% all toblended 15% increased binders than the heat the peakSC100 from paste, 0.0095 excluding to 0.013 W the/g. SCA0 This increase sample. results This from indicates faster that dissolution the hydration and hydration rates of of ye’elimite.CṠA and Increasingquicklime cover the dose the insufficient of quicklime hydration from 5% due to 15%to the increased replacement the of first cement exothermic with FA peak. The from cumulative 0.01 to ˙ 0.022heat release W/g with of the the SC100 accelerating sample e ffshowedect of C theSA. most This noticeable improvement increase is due with to the increasing more vigorous hydration reaction time offrom quicklime 6h to 12 formingh. This is portlandite. associated with Then, a higher as the concentrationcontent of cement of calcium, clinker sulfate,(e.g., C3A and and hydroxyl C3S). With ions a increased,curing time a decreaseof 72 h, inthe heat SC100 flow samp occurredle released [42]. The the initial highest alkaline amount environment of heat, and can the be strengthenedSCC15 paste byranked the increased second. The amount additional of portlandite heat of SCC15 generated, compared promoting to other the activityblended of binders FA [56 ].is Thederived dissolution from a oflarge amorphous amount of silicon activated and FA aluminum particles from involved active in FA the contributes hydration toreaction. the heat flow. Hence, the SCC15 sample exhibited the highest rate of heat release in the first 1 h.

(a) 0.025 (b) 14 SC100 SCA0 SCA5 SCA15 SC100 SCA0 SCA5 SCA15 SCC5 SCC10 SCC15 SCC5 SCC10 SCC15 First peak 12 0.020 10 0.015 8

0.010 6

Heat flow (W/g) 4 0.005 Cumulative heat (J/g) 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 2 0.000 0 0 12 24 36 48 60 72 0 12 24 36 48 60 72 Time (h) Time (h) Figure 9. The heat flow of binder pastes: (a) first exothermic peak, and (b) cumulative heat. Figure 9. The heat flow of binder pastes: (a) first exothermic peak, and (b) cumulative heat. An exothermic acceleration phase appeared after 4h, which is associated with the dissolution of 3.5. Hydration Evolution tricalcium silicate (C3S) and anhydrite, synchronous with the precipitation of portlandite and C–S–H phases [57]. With the increased formation rate of ettringite, the concentration of sulfate ions decreased, 3.5.1. XRD Patterns leading to the release of sulfate ions originally attached to tricalcium aluminate. This leads to a further dissolutionThe XRD and patterns hydration of blended reaction binders of tricalcium and control aluminate samples [57]. were None measured of the heat at flowhydration curves times for the of blended3, 7, and binders28 days containing (Figure 10) additional. The quartz FA phase showed of sharpall blended peaks, binders but wider was shoulders. higher than This the di ffSC100erent behaviorsample. This is associated is mainly withrelated the to retarding replacing e ffSCect with of FA large reported amounts in theof FA, literature which [is58 rich,59]. in Cumulative crystalline heatquartz release (Figure from 2). As the the first amount 6 h of hydrationof CṠA increased was higher from for0% allto blended15%, the intensity binders than of the the ettringite SC100 paste, peak excludingbecame more the pronounced. SCA0 sample. Ye’elimite This indicates (C4A3 thatṠ) reacts the hydration with water rates in the ofC presenceSA˙ and quicklimeof gypsum cover to form the insuettringite.fficient The hydration Hc/Mc peak due was to the detected, replacement and the of cementintensity with of this FA. peak The cumulativestrengthened heat with release increasing of the SC100 sample showed the most noticeable increase with increasing hydration time from 6h to 12 h. This is associated with a higher content of cement clinker (e.g., C3A and C3S). With a curing time of 72 h, the SC100 sample released the highest amount of heat, and the SCC15 paste ranked second. The additional heat of SCC15 compared to other blended binders is derived from a large amount of activated FA particles involved in the hydration reaction.

3.5. Hydration Evolution

3.5.1. XRD Patterns The XRD patterns of blended binders and control samples were measured at hydration times of 3, 7, and 28 days (Figure 10). The quartz phase of all blended binders was higher than the SC100 sample. This is mainly related to replacing SC with large amounts of FA, which is rich in crystalline quartz (Figure2). As the amount of C SA˙ increased from 0% to 15%, the intensity of the ettringite peak became more pronounced. Ye’elimite (C4A3S)˙ reacts with water in the presence of gypsum to form ettringite. The Hc/Mc peak was detected, and the intensity of this peak strengthened with increasing curing time. C3A and C4A3S˙ react with calcium carbonate to form Mc [43,44]. Due to the consumption of calcium Materials 2020, 13, x FOR PEER REVIEW 11 of 21

Materialscuring time.2020, 13 C,3 4018A and C4A3Ṡ react with calcium carbonate to form Mc [43,44]. Due to the consumption11 of 20 of calcium carbonate, insufficient carbonate leads to the formation of Hc instead of Mc. In addition, the value of ettringite peak increased with increasing curing time from 3 to 28 days. The formed Hc/Mc carbonate, insuﬃcient carbonate leads to the formation of Hc instead of Mc. In addition, the value of inhibits the conversion of AFt to Ms. It was impossible to detect the Ms phase by XRD due to poor ettringite peak increased with increasing curing time from 3 to 28 days. The formed H /M inhibits the crystallinity. The peaks of C–S–H and calcium carbonate in the SC100 sample werec higherc than all conversion of AFt to M . It was impossible to detect the M phase by XRD due to poor crystallinity. blended binders, irrespectives of curing time. This implies thats SC100 has a faster hydration rate than The peaks of C–S–H and calcium carbonate in the SC100 sample were higher than all blended binders, blended binders. This is consistent with the strength development shown in Figure 5. When the irrespective of curing time. This implies that SC100 has a faster hydration rate than blended binders. dosage of quicklime increased from 5% to 15%, the portlandite phase was intensified. In addition, This is consistent with the strength development shown in Figure5. When the dosage of quicklime dicalcium silicate (C2S) was also detected in all samples. This is because less reactive C2S hydrates in increased from 5% to 15%, the portlandite phase was intensiﬁed. In addition, dicalcium silicate (C S) the late age. 2 was also detected in all samples. This is because less reactive C2S hydrates in the late age.

(a) SC100 SCA0 SCA5 SCA15 (b) SC100 SCA0 SCA5 SCA15 SCC5 SCC10 SCC15 SCC5 SCC10 SCC15 CH CH CH M CH M c Q C,CSH c Q C,CSH E H ,M E H ,M c c E c c E H E Q Hc C S E Q c C S M 2 M 2

5 10 15 20 25 30 35 40 45 5 10 15 20 25 30 35 40 45 2-Theta 2-Theta

5 10 15 20 25 30 35 40 45 2-Theta

Figure 10. XRD patterns of binder pastes at hydration times times of of ( (aa)) 3 3 days, days, ( (bb)) 7 7 days, days, and and ( (cc)) 28 28 days. days. (E = ettringite, H = hemicarbonaluminate, M = monocarbonaluminate, CH = portlandite, Q = quartz, (E = ettringite, Hc = hemicarbonaluminate, Mc = monocarbonaluminate, CH = portlandite, Q = quartz, C = Calcite, M = mullite, C S = dicalcium silicate). C = Calcite, M = mullite, C22S = dicalcium silicate). 3.5.2. TG Analysis 3.5.2. TG Analysis To characterize the transformation of hydration products formed, TG analysis was performed To characterize the transformation of hydration products formed, TG analysis was performed on all binder pastes in curing times of 3, 7, and 28 days. DTG curves were obtained by the ﬁrst on all binder pastes in curing times of 3, 7, and 28 days. DTG curves were obtained by the first derivative of the TG data, as shown in Figure 11. Weight loss dehydration of ettringite almost occurred derivative of the TG data, as shown in Figure 11. Weight loss dehydration of ettringite almost at 75–120 ◦C[34]. Thus, it is clearly seen that all blended binders, except the SCA0 (0% addition of occurred at 75–120 °C [34]. Thus, it is clearly seen that all blended binders, except the SCA0 (0% CSA)˙ and SCA5 (5% addition of CSA)˙ samples, showed higher peak values than SC100 samples after addition of CṠA) and SCA5 (5% addition of CṠA) samples, showed higher peak values than SC100 hydration for three days. This indicates that more AFt was generated early in hydration. The formation samples after hydration for three days. This indicates that more AFt was generated early in hydration. of AFt probably results from the fast hydration of ye’elimite in the presence of gypsum. This contributes

Materials 2020, 13, x FOR PEER REVIEW 12 of 21

The formation of AFt probably results from the fast hydration of ye’elimite in the presence of gypsum. This contributes to the filler effect and improves the strength of the hardened backfill. Besides, the Hc/Mc peak value for the SC100 sample was found to be higher than blended binders, regardless of curing time. This may be related to the high content of C3A in neat SC. According to the literatures [40,60,61], the decomposition temperatures of hydration products (AFt, AFm, and C–(A)–S–H) and portlandite are 50–550 °C and 400–550 °C, respectively, while the decomposition of carbonate mainly appears between 550–1000 °C . The amount of hydrate water and portlandite obtained is shown in Figure 12. As can be seen in the Figure 12a, the SC100 sample contained the most chemically bound water (hydrate water) during the hydration process and, therefore, the highest hydration products, regardless of the curing time. As the dose of CṠA increased from 0 to 15%, the amount of hydrate water increased dramatically. This is because ye’elimite is highly reactive. When the amount of quicklime added was increased from 5% to 15%, changes in hydrate water showed a similar upward trend. This increase in hydration products is a result of the addition of more quicklime to create a highly alkaline environment. The high concentration of hydroxide ions promotes the breakdown of covalent Si–O–Si and Si–O–Al bonds in FA. At the same time, negatively charged tetrahedral anions (SiO44− and AlO45−) absorb positively charged ions (e.g., Ca2+) to drive polycondensation under the action of hydration [56]. With sufficient portlandite and sulfate, ettringite and C–S–H/C–A–S–H gelling products are formed. In addition, the increase in hydration products is also partially due in part to the hydration of quicklime, which produces portlandite. The SC100 sample presented an increase in portlandite formed with longer curing time, indicating a higher degree of hydration. For blended binders, the content of portlandite tended to increase and then decreased from the hydration time of 3 to 28 days. With a 7-day curing time, a small fraction of FA is activated and participates in the hydration reaction. The newly formed portlandite Materialsfrom cement2020, 13 hydration, 4018 exceeds the amount consumed in the FA reaction, while after a 7-day curing12 of 20 time, a large amount of FA participates in the hydration reaction and consumes a considerable amount of calcium hydroxide. The SCC15 sample had more portlandite than SCC10 paste. Combined towith the the filler TG e ffresultsect and and improves the strength the strength development of the hardened results, it backfill. can be Besides,inferred thethat H thec/M additionc peak value of 15% for theof quicklime SC100 sample reached was the found limit. to be higher than blended binders, regardless of curing time. This may be related to the high content of C3A in neat SC.

(a) 0.0 (b) 0.0

-0.5 -0.5

H /M -1.0 c c -1.0 Hc/Mc SC100 SC100 CH SCA0 SCA0 CH -1.5 SCA5 -1.5 SCA5 SCA15 SCA15 SCC5 SCC5

DTG (%/°C) DTG (%/°C) -2.0 SCC10 -2.0 SCC10 SCC15 SCC15 AFt and C-S-H AFt and C-S-H -2.5 -2.5

Materials-3.0 2020, 13, x FOR PEER REVIEW -3.0 13 of 21 50 150 250 350 450 550 50 150 250 350 450 550 Temperature (°C) Temperature (°C) (c) 0.0

-0.5

-1.0 SC100 SCA0 Hc/Mc CH -1.5 SCA5 SCA15 SCC5

DTG (%/°C) -2.0 SCC10 SCC15 -2.5 AFt and C-S-H

-3.0 50 150 250 350 450 550 Temperature (°C) Figure 11. Diﬀerential thermogravimetry (DTG) curves of blended binder pastes at hydration times of Figure 11. Differential thermogravimetry (DTG) curves of blended binder pastes at hydration times (a) 3 days, (b) 7 days, and (c) 28 days. of (a) 3 days, (b) 7 days, and (c) 28 days. According to the literatures [40,60,61], the decomposition temperatures of hydration products (AFt,(a)21 AFm, and C–(A)–S–H) and portlandite are 50–550(b)21 ◦C and 400–550 ◦C, respectively, while the decomposition of SC100 carbonate SCA0 mainly SCA5 appears SCA15 between 550–1000 ◦SC100C. The SCC5 amount SCC10 of hydrate SCC15 water and portlandite obtained is shown in Figure 12. As can be seen in the Figure 12a, the SC100 sample contained18 the most chemically bound water (hydrate water)18 during the hydration process and, therefore, the highest hydration products, regardless of the curing time. As the dose of CSA˙ increased from 0 to 15%, the amount of hydrate water increased dramatically. This is because ye’elimite is highly 15 15 reactive. When the amount of quicklime added was increased from 5% to 15%, changes in hydrate water showed a similar upward trend. This increase in hydration products is a result of the addition of more12 quicklime to create a highly alkaline environment.12 The high concentration of hydroxide ions promotes the breakdown of covalent Si–O–Si and Si–O–Al bonds in FA. At the same time, negatively 4 5 2+ charged of hydrate water (%) The amount tetrahedral anions (SiO4 − and AlO4 −) absorb of hydrate water (%) The amount positively charged ions (e.g., Ca ) to drive 9 9 polycondensation3 under the7 action of hydration28 [56]. With3 suﬃcient portlandite7 and sulfate, ettringite28 and C–S–H/C–A–S–HCuring gelling time products (d) are formed. In addition, the increaseCuring intime hydration (d) products is also partially due in part to the hydration of quicklime, which produces portlandite. (c)21 (d)21 SC100 SCA0 SCA5 SCA15 SC100 SCC5 SCC10 SCC15

18 18

15 15

12 12

The amount of portlandite (%) of portlandite The amount (%) of portlandite The amount

9 9 3 7 28 3 7 28 Curing time (d) Curing time (d)

Figure 12. The content of hydrate water of binder pastes with different doses of (a) CṠA and (b) quicklime, and the content of portlandite of binder pastes with different doses of (c) CṠA and (d) quicklime.

Materials 2020, 13, x FOR PEER REVIEW 13 of 21

-0.5

-1.0 SC100 SCA0 Hc/Mc CH -1.5 SCA5 SCA15 SCC5

DTG (%/°C) -2.0 SCC10 SCC15 -2.5 AFt and C-S-H

-3.0 50 150 250 350 450 550 Temperature (°C)

Figure 11. Differential thermogravimetry (DTG) curves of blended binder pastes at hydration times Materialsof (2020a) 3 ,days,13, 4018 (b) 7 days, and (c) 28 days. 13 of 20

(a)21 (b)21 SC100 SCA0 SCA5 SCA15 SC100 SCC5 SCC10 SCC15

18 18

15 15

12 12

The amount of hydrate water (%) The amount of hydrate water (%) The amount 9 9 3 7 28 3 7 28 Curing time (d) Curing time (d)

18 18

15 15

12 12

The amount of portlandite (%) of portlandite The amount (%) of portlandite The amount

9 9 3 7 28 3 7 28 Curing time (d) Curing time (d) ˙ Figure 12. TheThe content content of of hydrate hydrate water water of of binder binder pastes pastes with with different diﬀerent doses doses of ( ofa) ( CaṠ)CA SA and and (b) (b) quicklime, and the content of portlandite of binder pastes with diﬀerent doses of (c)CSA˙ and quicklime, and the content of portlandite of binder pastes with different doses of (c) CṠA and (d) (d) quicklime. quicklime. The SC100 sample presented an increase in portlandite formed with longer curing time, indicating 3.5.3. SEM Observation a higher degree of hydration. For blended binders, the content of portlandite tended to increase and thenS decreasedEM observation from the of hydrationthe microstructure time of 3 toof 28blended days. Withbinders a 7-day was performed, curing time, and a small the results fraction are of FAshown is activated in Figure and 13.participates A control sample in the hydration(SC100) was reaction. also included The newly for comparison. formed portlandite As shown from in cementFigure hydration13, the SC100 exceeds sample the has amount a relatively consumed coarser in the microstructure FA reaction, whilethan blended after a 7-day binders, curing which time, reveals a large a amount of FA participates in the hydration reaction and consumes a considerable amount of calcium hydroxide. The SCC15 sample had more portlandite than SCC10 paste. Combined with the TG results and the strength development results, it can be inferred that the addition of 15% of quicklime reached the limit.

3.5.3. SEM Observation SEM observation of the microstructure of blended binders was performed, and the results are shown in Figure 13. A control sample (SC100) was also included for comparison. As shown in Figure 13, the SC100 sample has a relatively coarser microstructure than blended binders, which reveals a higher degree of hydration. Increasing the amount of CSA˙ from 0 to 15% caused the morphology of samples to be rough, which is related to the acceleration eﬀect of CSA.˙ Increasing the amount of quicklime from 5 to 10% caused a clear change in the microstructure of SCC10. As the dose of quicklime continued to increase to 15%, the morphology did not continue to coarsen, and there was a large amount of calcium hydroxide. This is consistent with TG experimental results shown in Figure 12. Materials 2020, 13, 4018 14 of 20 Materials 2020, 13, x FOR PEER REVIEW 15 of 21

SC100

C-S-H

SCA0 SCC5 C-S-H C-S-H

SCA5 SCC10

C-S-H C-S-H

SCA15 SCC15

C-S-H C-S-H

FFigureigure 13. 13. SEMSEM observation observation of of binder binder pastes pastes after after 28 28 days days of of hydration hydration (Mag (Mag == 50005000×).). × EDX analysis was also conducted to evaluate the C–S–H composition of binder pastes, and (a)0.6 (b)0.6 the results are shown in Figure 14. According to the literatures [60,62,63], the Al/Ca atomicSCA0 ratio of SCA5 AFm AFm portlandite,0.5 AFt, and AFm are 0, 0.33, and 0.5, respectively.0.5 To avoid the inﬂuence of SCA15 AFt and AFm, a straight line was drawn through the points with the lowest Al/Ca ratio to obtain the Al/Si atomic ratio. As0.4 seen in Figure 14, the SC100 sample had an0.4 Al/Si atomic ratio of about 0.14, and a Si/Ca AFt AFt ˙ atomic ratio0.3 of about 0.3–0.4. When incorporated into0.3 FA, CSA, and quicklime (blendedAl/Si=0.27 binder), the data points were dispersed. This is mainly attributed to the variable composition of active FA 0.2 0.2 involved in the hydration reaction. In addition,Al/Si=0.14 if the Si/Ca ratio is extremely high, it may be caused

Al/Ca atomic ratio

Al/Ca atomic ratio by the measurement0.1 error due to wrong dot position.0.1 There is a trend toward an increase in the overall Si/CaCH ratio for FA-cement cementitious system, whichCH is consistent with the results of previous 0.0 0.0 studies [64,650.0]. Blended0.2 0.4 binders0.6 had0.8 higher1.0 Al/1.2Si atomic0.0 ratios0.2 (0.270.4 and 0.43)0.6 than0.8 the1.0 control1.2 sample. This means that more FASi/Ca is activatedatomic ratio and hydrated to form C–S–A–H.Si/Ca C–S–H atomic ratio absorbs dissolved Si to form longer silicate chain lengths, allowing the bridging tetrahedra of silicate chains to absorb more aluminum [66,67].

Materials 2020, 13, x FOR PEER REVIEW 15 of 21

SC100

C-S-H

SCA0 SCC5 C-S-H C-S-H

SCA5 SCC10

C-S-H C-S-H

SCA15 SCC15

C-S-H C-S-H

Materials 2020,F13igure, 4018 13. SEM observation of binder pastes after 28 days of hydration (Mag = 5000×). 15 of 20

(a)0.6 (b)0.6 SCA0 SCA5 AFm AFm 0.5 0.5 SCA15

0.4 0.4 AFt AFt

0.3 0.3 Al/Si=0.27

0.2 0.2 Al/Si=0.14

Al/Ca atomic ratio

0.1 0.1

CH CH Materials 20200.0, 13, x FOR PEER REVIEW 0.0 16 of 21 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 Si/Ca atomic ratio Si/Ca atomic ratio (c)0.6 SCC5 SCC10 AFm 0.5 SCC15

Al/Si=0.43 0.4 AFt 0.3

0.2

Al/Ca atomic ratio

0.1

CH 0.0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 Si/Ca atomic ratio Figure 14. Plot of Al/Ca atomic ratio vs. Si/Ca atomic ratio of (a) SC100; (b) SCA0—SCA15 and (c) Figure 14. Plot of Al/Ca atomic ratio vs. Si/Ca atomic ratio of (a) SC100; (b) SCA0—SCA15 and (c) SCC5—SCC15 samples. SCC5—SCC15 samples. 3.6. Discussion 3.6. Discussion This study developed novel blended binders composed of SC, FA, quicklime, and calcium sulfoaluminateThis study cement. developed SC provides novel blended major bindersgel products composed as basic of cementitious SC, FA, quicklime, material. and FA calcium has the sulfoaluminate cement. SC provides major gel products as basic cementitious material. FA has the potential to reduce CO2 emissions and costs as a source of aluminosilicate material. The filler effect of FApotential provides to reduce additional CO2 nucleationemissions and sites costs for cement as a source hydration, of aluminosilicate increasing the material. packing The density filler ofeffe thect cementitiousof FA provides system, additional and refiningnucleation the sites pore for structure cement [ 40hydration,,68,69]. Under increasing the influence the packing of the density alkaline of environment,the cementitious active system, FA particles and refining participate the pore in hydration structure to[40,6 form8,6 C–S–H9]. Under/C–A–S–H, the influence improving of the strengthalkaline developmentenvironment, during active and FA after particles the curing participate stage [44 in, 70hydration]. When adding to form large C–S amounts–H/C–A of–S FA–H, to improving a blended binder,strength mineral development accelerators during (e.g., and CafterSA˙ andthe curing quicklime) stage are [44 essential,70]. When to adding improve large the amounts initial strength of FA performanceto a blended binder, of the hardenedmineral accelerators backfill. C (e.g.,SA˙ hydrates CṠA and rapidly quicklime) to form are essential large AFt to phases improve to the fill initial voids. Quicklimestrength per hydratesformance to of increase the hardened the concentration backfill. CṠ ofA hydroxidehydrates rapidly ions that to promoteform large the AFt activity phases of to FA. fill voids. Quicklime hydrates to increase the concentration of hydroxide ions that promote the activity According to the manufacturer of raw materials, the CO2 emissions of SC, CSA,˙ FA, and quicklime productsof FA. are 0.72, 0.69, 0.07, 0.18 kg/kg, and the costs are 71, 136.8, 23.4, and 32.1 $US/ton, respectively. According to the manufacturer of raw materials, the CO2 emissions of SC, CṠA, FA, and The CO2 emission and cost of binders can be normalized per compressive strength with a 28-day curingquicklime time products (Equations are (3) 0.72, and 0.69, (4)), 0.07, and the0.18 results kg/kg, are and shown the costs in Figure are 71, 15 136.8,. It can 23.4, be concludedand 32.1 $US/ton, that the respectively. The CO2 emission and cost of binders can be normalized per compressive strength with SCC10 and SCC15 binders have lower CO2 emissions and costs than pure SC sample after 28 days of hydration.a 28-day curing These time experimental (Equations findings (3) and help (4)), to promote and the the results practical are shown application in Figure of blended 15. It binders can be inconcluded filling technology that the SCC10 and optimize and SCC15 solid binders waste disposal. have lower CO2 emissions and costs than pure SC sample after 28 days of hydration. These experimental findings help to promote the practical application of blended binders in filling technology COand2 optimizeemission solid waste disposal. BCO2 emission = (3) 퐶푂UCS2 푒푚푖푠푠푖표푛 value 퐵 = (3) CO2 emission cost푈퐶푆 푣푎푙푢푒 B = (4) cost UCS푐표푠푡 value 퐵 = (4) 푐표푠푡 푈퐶푆 푣푎푙푢푒

Materials 2020, 13, 4018 16 of 20 Materials 2020, 13, x FOR PEER REVIEW 17 of 21

0.5 60 E E CO2emission cost 50 0.4

40 0.3

(kg/kg/MPa) 30 0.2

emission

(dollar/ton/MPa) 2 20

cost

E 0.1 E 10

0.0 0 SC100 SCA0 SCA5 SCA15 SCC5 SCC10 SCC15 Binder pastes

FigureFigure 15. Normalized CO 2 emissionemission and and cost cost of of binder binder pastes pastes..

4. Conclusion Conclusionss IInn this study, a large amount of FA (40%)(40%) waswas usedused toto lowerlower thethe costcost ofof thethe binder.binder. Calcium sulfoaluminate cement and quicklime quicklime were were used used to to accelerate accelerate the the early early hydration hydration of of blended blended binders. binders. The flowability, flowability, mechanical mechanical performance, performance, and and pore porestructure structure of cemented of cemented paste backfill paste were backfill examined. were examined.In addition, In theaddition, heat ofthe hydration heat of hydration and hydration and hydration evolution evolution of pure of blended pure blended binder binder pastes pastes were wereinvestigated. investigated. Based Based on the on experimental the experimental results, results, the conclusion the conclusion can be can drawn be drawn as: as: (1) AddingAdding FAFA significantly significantly improves improves the the flowability flowability of theof the fresh fresh CPB CPB due todue the to lubrication the lubrication effect ˙ effectof spherical of spherical shape. shape. However, However, this increasethis increase was graduallywas gradually suppressed suppressed by the by increasethe increase of C SAof C andṠA ˙ andquicklime. quicklime. Due Due to the to rapidthe rapid reaction reaction of C SA,of C theṠA, SDthe decreased SD decreased the fastest the fastest with with increasing increasing curing curing time ˙ timein the in SCA15 the SCA15 sample. sample. As the As amount the amount of CSA of and CṠ quicklimeA and quicklime increased, increased, CPB samples CPB samples with blended with blendedbinders showed binders improved showed improvedmechanical mechanical performance performance associated with associated the formation with the of ettringite formation and of ettringiteC–S–H/C–A–S–H. and C–S– IncreasingH/C–A–S– theH. Increasing dosage of quicklime the dosage from of quicklime 10% to 15% from slightly 10% to reduced 15% slightly the UCS reduced value theof the UCS hardened value of backfill. the hardened backfill. (2) The volume of pores with diameters less than 500 nm and 500 500–1000–1000 nm did not show a clear ˙ trend with with increasing increasing amounts amounts of C ofSA C andṠA quicklime.and quicklime. In contrast, In contrast, those larger those than larger 1000 than nm decreased 1000 nm decredramatically.ased dramatically. Linear fitting Linear reveals fitting that total reveals pore that volume total was pore closely volume related was to closely CPB strength. related to Due CPB to strength.the high contentDue to ofthe cement high content clinker, of the cement cumulative clinker, heat the flow cumulative for the SC100heat flow sample for the was SC100 high, andsample the ˙ wasaddition high, of and CSA the and addition quicklime of CṠ increasedA and quicklime the cumulative increased heat the flow cum inulative the first heat 6 h. flow in the first 6 h. ˙ (3) Addition of C ṠSAA and quicklime promoted the formation of ettringite and portlandite. From day 7 to day 28 28 of of hydration, hydration, FA FA was involved in hydration, thus reducing the amount of portlandite produced in blended binders binders.. Furthermore, Furthermore, due due to to hydration hydration of of quicklime, quicklime, FA FA was activated by the alkaline environment.environment. Then,Then, AlAl and and Si Si released released from from FA FA were were hydrated hydrated to form to form C–S–H C–/SC–A–S–H–H/C–A– withS–H withhigh high Al/Si Al/Si ratios. ratios. (4) Blended binder SCC15 showed the least normalized CO 2 emissionemission and and lower lower cost. cost. Therefore, Therefore, itit had the potential for application in fillingfilling technology.technology.

Author Contributions: S.Z.; writing—original draft preparation, H.D.; writing—review and editing. All authors Author Contributions: S.Z.; writing—original draft preparation, H.D.; writing—review and editing. All authors have read and agreed to the published version of the manuscript. have read and agreed to the published version of the manuscript. Funding: This work was supported by the National Science Foundation of China (Grant No. 51534003 and Funding:51774066), This the Ministry work was of Science supported and by Technology the National of the Science People’s Foundation Republic of of China China (Grant (Grant No. No. 2016YFC0801601), 51534003 and 51774066),National Natural the Ministry Science Foundation of Science of China and Technology (Grant No. 51904055) of the People’sand the Fundamental Republic of Research China Funds (Grant for No. the 2016YFC0801601),Central Universities National of China Natural (Grant No.Science N 2001010). Foundation of China (Grant No. 51904055) and the Fundamental ResearchAcknowledgments: Funds for theThe Central Authors Universities would like of toChina thank (Grant the Science No. N 2001010 Compass). Testing Center for their help with SEM and TG. Acknowledgments: The Authors would like to thank the Science Compass Testing Center for their help with SEMConﬂicts and TG of Interest:. The authors declare no conﬂict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to Conflictspublish the of results. Interest: The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

Materials 2020, 13, 4018 17 of 20

References

1. Behera, S.K.; Ghosh, C.N.; Mishra, D.P.; Singh, P.; Mishra, K.; Buragohain, J.; Mandal, P.K. Strength development and microstructural investigation of lead-zinc mill tailings based paste backfill with fly ash as alternative binder. Cem. Concr. Compos. 2020, 109, 103553. [CrossRef] 2. Benzaazoua, M.; Fall, M.; Belem, T. A contribution to understanding the hardening process of cemented pastefill. Miner. Eng. 2004, 17, 141–152. [CrossRef] 3. Fall, M.; Benzaazoua, M.; Ouellet, S. Experimental characterization of the influence of tailings fineness and density on the quality of cemented paste backfill. Miner. Eng. 2005, 18, 41–44. [CrossRef] 4. Fall, M.; Célestin, J.; Pokharel, M.; Touré, M. A contribution to understanding the effects of curing temperature on the mechanical properties of mine cemented tailings backfill. Eng. Geol. 2010, 114, 397–413. [CrossRef] 5. Yang, L.; Yilmaz, E.; Li, J.; Liu, H.; Jiang, H. Effect of superplasticizer type and dosage on fluidity and strength behavior of cemented tailings backfill with different solid contents. Constr. Build. Mater. 2018, 187, 290–298. [CrossRef] 6. Li, W.; Fall, M. Strength and self-desiccation of slag-cemented paste backfill at early ages: Link to initial sulphate concentration. Cem. Concr. Compos. 2018, 89, 160–168. [CrossRef] 7. Fall, M.; Pokharel, M. Coupled effects of sulphate and temperature on the strength development of cemented tailings backfills: Portland cement-paste backfill. Cem. Concr. Compos. 2010, 32, 819–828. [CrossRef] 8. Pokharel, M.; Fall, M. Combined influence of sulphate and temperature on the saturated hydraulic conductivity of hardened cemented paste backfill. Cem. Concr. Compos. 2013, 38, 21–28. [CrossRef] 9. Qi, C.; Fourie, A.; Chen, Q.; Zhang, Q. A strength prediction model using artificial intelligence for recycling waste tailings as cemented paste backfill. J. Clean. Prod. 2018, 183, 566–578. [CrossRef] 10. Qi, C.; Liu, L.; He, J.; Chen, Q.; Yu, L.-J.; Liu, P. Understanding Cement Hydration of Cemented Paste Backfill: DFT Study of Water Adsorption on Tricalcium Silicate (111) Surface. Minerals 2019, 9, 202. [CrossRef] 11. Qiu, J.; Guo, Z.; Yang, L.; Jiang, H.; Zhao, Y. Effects of packing density and water film thickness on the fluidity behaviour of cemented paste backfill. Powder Technol. 2020, 359, 27–35. [CrossRef] 12. Benzaazoua, M.; Bussière, B.; Demers, I.; Aubertin, M.; Fried, É.; Blier, A. Integrated mine tailings management by combining environmental desulphurization and cemented paste backfill: Application to mine Doyon, Quebec, Canada. Miner. Eng. 2008, 21, 330–340. [CrossRef] 13. Fall, M.; Belem, T.; Samb, S.; Benzaazoua, M. Experimental characterization of the stress-strain behaviour of cemented paste backfill in compression. J. Mater. Sci. 2007, 42, 3914–3922. [CrossRef] 14. Jiang, H.; Qi, Z.; Yilmaz, E.; Han, J.; Qiu, J.; Dong, C. Effectiveness of alkali-activated slag as alternative binder on workability and early age compressive strength of cemented paste backfills. Constr. Build. Mater. 2019, 218, 689–700. [CrossRef] 15. Qiu, J.; Yang, L.; Sun, X.; Xing, J.; Li, S. Strength characteristics and failure mechanism of cemented super-fine unclassified tailings backfill. Minerals 2017, 7, 58. [CrossRef] 16. Qiu, J.-P.; Yang, L.; Xing, J.; Sun, X.-G. Analytical solution for determining the required strength of mine backfill based on its damage constitutive model. Soil Mech. Found. Eng. 2018, 54, 371–376. [CrossRef] 17. Yang, L.; Qiu, J.; Jiang, H.; Hu, S.; Li, H.; Li, S. Use of cemented super-fine unclassified tailings backfill for control of subsidence. Minerals 2017, 7, 216. [CrossRef] 18. Yilmaz, E.; Belem, T.; Benzaazoua, M. Specimen size effect on strength behavior of cemented paste backfills subjected to different placement conditions. Eng. Geol. 2015, 185, 52–62. [CrossRef] 19. Yilmaz, E.; Belem, T.; Bussière, B.; Benzaazoua, M. Relationships between microstructural properties and compressive strength of consolidated and unconsolidated cemented paste backfills. Cem. Concr. Compos. 2011, 33, 702–715. [CrossRef] 20. Fall, M.; Benzaazoua, M. Modeling the effect of sulphate on strength development of paste backfill and binder mixture optimization. Cem. Concr. Res. 2005, 35, 301–314. [CrossRef] 21. Fall, M.; Benzaazoua, M.; Saa, E. Mix proportioning of underground cemented tailings backfill. Tunn. Undergr. Space Technol. 2008, 23, 80–90. [CrossRef] 22. Zhang, S.; Yang, L.; Ren, F.; Qiu, J.; Ding, H. Rheological and mechanical properties of cemented foam backfill: Effect of mineral admixture type and dosage. Cem. Concr. Compos. 2020, 112, 103689. [CrossRef] Materials 2020, 13, 4018 18 of 20

23. De Belie, N.; Soutsos, M.; Gruyaert, E. Limestone Powder. In Properties of Fresh and Hardened Concrete Containing Supplementary Cementitious Materials; Springer: Cham, Switzerland, 2018; Volume 25, pp. 123–151. [CrossRef] 24. Fernandez, A.; Calvo, J.G.; Alonso, M. Ordinary Portland Cement composition for the optimization of the synergies of supplementary cementitious materials of ternary binders in hydration processes. Cem. Concr. Compos. 2018, 89, 238–250. [CrossRef] 25. Zhao, Y.; Qiu, J.; Xing, J.; Sun, X. Chemical activation of binary slag cement with low carbon footprint. J. Clean. Prod. 2020, 267, 121455. [CrossRef] 26. Zhao, Y.; Qiu, J.; Zhang, S.; Guo, Z.; Ma, Z.; Sun, X.; Xing, J. Effect of sodium sulfate on the hydration and mechanical properties of lime-slag based eco-friendly binders. Constr. Build. Mater. 2020, 250, 118603. [CrossRef] 27. Deschner, F.; Winnefeld, F.; Lothenbach, B.; Seufert, S.; Schwesig, P.; Dittrich, S.; Goetz-Neunhoeffer, F.; Neubauer, J. Hydration of Portland cement with high replacement by siliceous fly ash. Cem. Concr. Res. 2012, 42, 1389–1400. [CrossRef] 28. Schöler, A.; Lothenbach, B.; Winnefeld, F.; Zajac, M. Hydration of quaternary Portland cement blends containing blast-furnace slag, siliceous fly ash and limestone powder. Cem. Concr. Compos. 2015, 55, 374–382. [CrossRef] 29. Criado, M.; Fernández-Jiménez, A.; Palomo, A. Alkali activation of fly ash: Effect of the SiO2/Na2O ratio. Microporous Mesoporous Mater. 2007, 106, 180–191. [CrossRef] 30. Pacheco-Torgal, F.; Castro-Gomes, J.; Jalali, S. Alkali-activated binders: A review. Constr. Build. Mater. 2008, 22, 1305–1314. [CrossRef] 31. Shi, C. Studies on Several Factors Affecting Hydration and Properties of Lime-Pozzolan Cements. J. Mater. Civ. Eng. 2001, 13, 441–445. [CrossRef] 32. Antiohos, S.; Papageorgiou, A.; Tsimas, S. Activation of fly ash cementitious systems in the presence of quicklime. Part II: Nature of hydration products, porosity and microstructure development. Cem. Concr. Res. 2006, 36, 2123–2131. [CrossRef] 33. Antiohos, S.; Tsimas, S. Activation of fly ash cementitious systems in the presence of quicklime: Part I. Compressive strength and pozzolanic reaction rate. Cem. Concr. Res. 2004, 34, 769–779. [CrossRef] 34. Li, G.; Zhang, J.; Song, Z.; Shi, C.; Zhang, A. Improvement of workability and early strength of calcium sulphoaluminate cement at various temperature by chemical admixtures. Constr. Build. Mater. 2018, 160, 427–439. [CrossRef] 35. Gastaldini, A.L.G.; Isaia, G.C.; Saciloto, A.P.; Missau, F.; Hoppe, T.F. Influence of curing time on the chloride penetration resistance of concrete containing rice husk ash: A technical and economical feasibility study. Cem. Concr. Compos. 2010, 32, 783–793. [CrossRef] 36. Zhang, G.; Yang, Y.; Yang, H.; Li, H. Calcium sulphoaluminate cement used as mineral accelerator to improve the property of Portland cement at sub-zero temperature. Cem. Concr. Compos. 2020, 106, 103452. [CrossRef] 37. ASTM C39. Standard Test Method for Compressive Strength of Cylindrical Concrete Specimens; ASTM International: West Conshohocken, PA, USA, 2010. 38. Zhao, Y.; Qiu, J.; Xing, J.; Sun, X. Recycling of quarry dust for supplementary cementitious materials in low carbon cement. Constr. Build. Mater. 2020, 237, 117608. [CrossRef] 39. Zhang, S.; Ren, F.; Guo, Z.; Qiu, J.; Ding, H. Strength and deformation behavior of cemented foam backfill in sub-zero environment. J. Mater. Res. Technol. 2020, 9, 9219–9231. [CrossRef] 40. De Weerdt, K.; Haha, M.B.; Le Saout, G.; Kjellsen, K.O.; Justnes, H.; Lothenbach, B. Hydration mechanisms of ternary Portland cements containing limestone powder and fly ash. Cem. Concr. Res. 2011, 41, 279–291. [CrossRef] 41. Ye, H.; Gao, X.; Wang, R.; Wang, H. Relationship among particle characteristic, water film thickness and flowability of fresh paste containing different mineral admixtures. Constr. Build. Mater. 2017, 153, 193–201. [CrossRef] 42. Nguyen, H.-A.; Chang, T.-P.; Shih, J.-Y.; Chen, C.-T. Influence of low calcium fly ash on compressive strength and hydration product of low energy super sulfated cement paste. Cem. Concr. Compos. 2019, 99, 40–48. [CrossRef] 43. Baquerizo, L.G.; Matschei, T.; Scrivener, K.L.; Saeidpour, M.; Wadsö, L. Hydration states of AFm cement phases. Cem. Concr. Res. 2015, 73, 143–157. [CrossRef] Materials 2020, 13, 4018 19 of 20

44. Matschei, T.; Lothenbach, B.; Glasser, F.P. The AFm phase in Portland cement. Cem. Concr. Res. 2007, 37, 118–130. [CrossRef] 45. Jeon, D.; Yum, W.S.; Jeong, Y.; Oh, J.E. Properties of quicklime(CaO)-activated Class F fly ash with the use of CaCl2. Cem. Concr. Res. 2018, 111, 147–156. [CrossRef] 46. Wang, X.-Y. Effect of fly ash on properties evolution of cement based materials. Constr. Build. Mater. 2014, 69, 32–40. [CrossRef] 47. Wang, Y.; Shui, Z.; Wang, L.; Gao, X.; Huang, Y.; Song, Q.; Liu, K. Alumina-rich pozzolan modification on Portland-limestone cement concrete: Hydration kinetics, formation of hydrates and long-term performance evolution. Constr. Build. Mater. 2020, 258, 119712. [CrossRef] 48. Yingliang, Z.; Jingping, Q.; Zhengyu, M.A.; Zhenbang, G.; Hui, L. Effect of superfine blast furnace slags on the binary cement containing high-volume fly ash. Powder Technol. 2020, 375, 539–548. [CrossRef] 49. Everett, D.H. IUPAC Manual of Symbols and Terminology, appendix 2, Part 1, Colloid and Surface Chemistry. Pure Appl. Chem. 1972, 31, 578–621. [CrossRef] 50. Ho, L.S.; Nakarai, K.; Duc, M.; Le Kouby, A.; Maachi, A.; Sasaki, T. Analysis of strength development in cement-treated soils under different curing conditions through microstructural and chemical investigations. Constr. Build. Mater. 2018, 166, 634–646. [CrossRef] 51. Nakarai, K.; Ishida, T.; Kishi, T.; Maekawa, K. Enhanced thermodynamic analysis coupled with temperature-dependent microstructures of cement hydrates. Cem. Concr. Res. 2007, 37, 139–150. [CrossRef] 52. Horpibulsuk, S.; Rachan, R.; Chinkulkijniwat, A.; Raksachon, Y.; Suddeepong, A. Analysis of strength development in cement-stabilized silty clay from microstructural considerations. Constr. Build. Mater. 2010, 24, 2011–2021. [CrossRef] 53. Berodier, E.; Scrivener, K. Evolution of pore structure in blended systems. Cem. Concr. Res. 2015, 73, 25–35. [CrossRef] 54. Chindaprasirt, P.; Jaturapitakkul, C.; Sinsiri, T. Effect of fly ash fineness on compressive strength and pore size of blended cement paste. Cem. Concr. Compos. 2005, 27, 425–428. [CrossRef] 55. Jin, S.; Zhang, J.; Han, S. Fractal analysis of relation between strength and pore structure of hardened mortar. Constr. Build. Mater. 2017, 135, 1–7. [CrossRef] 56. Fernando, P.-T.; Joao, C.-G.; Said, J. Alkali-activated binders: A review Part 1. Historical background, terminology, reaction mechanisms and hydration products. Constr. Build. Mater. 2008, 22, 1305–1314. 57. Jansen, D.; Goetz-Neunhoeffer, F.; Stabler, C.; Neubauer, J. A remastered external standard method applied to the quantification of early OPC hydration. Cem. Concr. Res. 2011, 41, 602–608. [CrossRef] 58. De Rojas, M.I.S.; Luxan, M.P.; Frias, M.; Garcia, N. The influence of different additions on portland cement hydration heat. Cem. Concr. Res. 1993, 23, 46–54. [CrossRef] 59. Dyer, T.D.; Dhir, R.K. Hydration reactions of cement combinations containing vitrified incinerator fly ash. Cem. Concr. Res. 2004, 34, 849–856. [CrossRef] 60. Bullard, J.W.; Jennings, H.M.; Livingston, R.A.; Nonat, A.; Scherer, G.W.; Schweitzer, J.S.; Scrivener, K.L.; Thomas, J.J. Mechanisms of cement hydration. Cem. Concr. Res. 2011, 41, 1208–1223. [CrossRef] 61. Briendl, L.G.; Mittermayr, F.; Baldermann, A.; Steindl, F.R.; Sakoparnig, M.; Letofsky-Papst, I.; Galan, I. Early hydration of cementitious systems accelerated by aluminium sulphate: Effect of fine limestone. Cem. Concr. Res. 2020, 134, 106069. [CrossRef] 62. Lothenbach, B.; Scrivener, K.; Hooton, R.D. Supplementary cementitious materials. Cem. Concr. Res. 2011, 41, 1244–1256. [CrossRef] 63. Damidot, D.; Lothenbach, B.; Herfort, D.; Glasser, F.P. Thermodynamics and cement science. Cem. Concr. Res. 2011, 41, 679–695. [CrossRef] 64. Duchesne, J.; Be’rubé, M.A. Effect of supplementary cementing materials on the composition of cement hydration products. Adv. Cem. Based Mater. 1995, 2, 43–52. [CrossRef] 65. Girão, A.V.; Richardson, I.G.; Taylor, R.; Brydson, R.M.D. Composition, morphology and nanostructure of

C–S–H in 70% white Portland cement–30% ﬂy ash blends hydrated at 55 ◦C. Cem. Concr. Res. 2010, 40, 1350–1359. [CrossRef] 66. Richardson, I.G.; Groves, G.W. The incorporation of minor and trace elements into calcium silicate hydrate (C S H) gel in hardened cement pastes. Cem. Concr. Res. 1993, 23, 131–138. [CrossRef] 67. Chen, J.J.; Thomas, J.J.; Taylor, H.F.W.; Jennings, H.M. Solubility and structure of calcium silicate hydrate. Cem. Concr. Res. 2004, 34, 1499–1519. [CrossRef] Materials 2020, 13, 4018 20 of 20

68. Hwang, C.; Shen, D. The effects of blast-furnace slag and fly ash on the hydration of Portland cement. Cem. Concr. Res. 1991, 21, 410–425. [CrossRef] 69. Rahhal, V.; Talero, R. Influence of two different fly ashes on the hydration of Portland cements. J. Therm. Anal. Calorim. 2004, 78, 191–205. [CrossRef] 70. Zhao, X.; Liu, C.; Zuo, L.; Wang, L.; Zhu, Q.; Wang, M. Investigation into the effect of calcium on the existence form of geopolymerized gel product of fly ash based geopolymers. Cem. Concr. Compos. 2019, 103, 279–292. [CrossRef]

© 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/).