Ecological Upgrade of Normal-Strength Mortars by Using High Volume of GGBS

Normal-strength concrete is widely used in construction sites considering the cost, technology, and structural safety. )e ecological upgrade of such materials is more meaningful for the sustainable development in a greener way. To this end, the feasibility of ecological upgrade of normal-strength mortars (NSM) by using high volume of ground granulated blast furnace slag (GGBS) (70%–90%) was evaluated in this paper. Comprehensive experiments were conducted to investigate the influences of experimental variables such as content of cement, curing temperature, and mass ratio of water to binder (w/b) on the fresh properties, compressive strength, hydration products, microstructure, and pore structure of NSM. Ecoefficiency evaluation was conducted based on the energy requirement for the whole production of cement and GGBS. Experimental results showed that ecological upgrade of NSMwas viable and feasible. When substituting 70% to 90% cement by GGBS, the energy requirement of 1 t binder can be accordingly saved by 67% to 86%, and the performance energy can be reduced from 25.4 (kWh/t)/MPa to 6 to 8 (kWh/t)/MPa. With proper contents of GGBS (70% or 75%), the 28 d compressive strengths were acceptable with reductions less than 10%. Evaluated curing temperatures and decreased w/b were viable methods to promote the early-age compressive strength of NSM incorporating high volume of GGBS. For instance, raising the curing temperature to 40°C can help achieve higher earlyage compressive strength than that of the control group. In addition, the pore sizes within ZII (<100 nm) of GGBS-incorporated NSM were refined by over 20 nm.


Introduction
Nowadays, the requirements of low carbon and environmental protection on the construction industry have prompted great motivations to seek alternative materials to cement since the emission of about 7% global CO 2 , the consumption of 5% global primary industrial energy, and the release of large amounts of NO x , SO 2 , particulate matter, mercury, and other compounds attributed to the cement production [1][2][3]. One of the promising strategies is the substitution of cement by low-carbon pozzolanic mineral admixtures, i.e., ground granulated blast furnace slag (GGBS) [4], a by-product from blast furnaces during the production of pig iron [5,6]. GGBS containing calcium silicoaluminate as its major component is highly vitreous in phases and is reactive to hydroxides released from the cement self-hydration to yield hydration products, i.e., CSH gel and hydrocalcium aluminate [7,8].
e employment of GGBS in cementitious materials brings two conspicuous benefits. e first one is the direct environment effect in terms of energy saving, pollutant discharge reduction, natural resources protection, and waste recycling. e manufacture of GGBS mainly involves the processes of cooling of molten slag through high-pressure water jets, drying, and grinding in a rotating ball mill [5,9]. ese operations consume less energy as compared to the clinkering process of cement manufacturing [10]. Shi and Qian [11] estimated that the required energy for grinding GGBS was less than 10% of the total energy requirement for producing cement. Prominently, greenhouse emission and waste or toxic matters release related to the cement production are reduced [12,13]. It was reported that the incorporation of 50% GGBS in mortar or concrete can help reduce about 0.5 t CO 2 emission [10]. e second benefit can be classified into the material effect. It is generally agreed that cementitious materials blended with GGBS develop refined pore sizes in their service life [14,15]. Li et al. [16] confirmed that the pore sizes of paste or mortar moved from RII size range (10 nm-100 nm) to RI size range (<10 nm) when 70% GGBS was incorporated. is finer pore structure of GGBScontaining cementitious composites contributes to barriers for the ingression of harmful substance, i.e., chloride ion and sulfate [5,15]. Furthermore, GGBS contains no C 3 A, making GGBS-incorporating cementitious composites less reactive to sulfates [5]. GGBS itself or its crystalline phases provides additional nucleation sites for the hydration products of cement, contributing to accelerated hydration of cement [17,18]. Other reported benefits cover the fire resistance, improved workability, increased pumpability, densified ITZ microstructure, and low hydration heat [19][20][21][22].
is is attractive in construction sites with dominating requirements of normal-strength concrete (NSC), which receives less attention in the recent years in the concrete studies compared to (ultra) high-strength concrete. Actually, NSC is widely used in construction sites considering the cost, technology, and structural safety. e ecological upgrade of NSC by incorporating high volume of GGBS may provide a better opportunity to develop in a more environmentally friendly manner. To achieve a maximized environmental or material benefit, the substitution rates of GGBS for cement are thus pivotal. By now, the replacement quantity of GGBS to cement has been investigated from 30% to 85% by weight; however, 50% is a common value in most applications [5,7,9,23,24]. Although abundant studies regarding the properties of concrete or mortar incorporating GGBS up to 60% of the total binder have been carried out, there are still lack of sufficient concerns on higher substitution rates larger than 60% [14]. is may be mainly because incorporating high volume of GGBS into cementitious composites brings about low early-age strength [25,26]. Nowadays, the defect of low early-age strength can be overcome by means of elevated curing temperatures [27,28], increasing GGBS fineness [24,29], and using finer mineral admixtures, i.e., nano-silica [20,30], nano-CaCO 3 [31], silica fume [32], and metakaolin [33]. In addition, it is also largely agreed that, with proper proportion, GGBS-incorporating cementitious composites can develop comparable or superior compressive strength than that incorporating cement only at late ages. erefore, this paper intends to ecologically upgrade normal-strength mortars (NSM) by incorporating high volume of GGBS. e investigated substitution ratios of cement by GGBS range from 70%-90% with an interval of 5%. e feasibility of mechanical compensation methods that higher curing temperatures or lower w/b will be examined. Comprehensive experiments were conducted to investigate the content of cement, curing temperature, and w/b on the fresh properties, compressive strength, hydration products, microstructure, and pore structure of NSM. TGA-DTG, SEM, XRD, and MIP were applied to study the underlying mechanisms on multiscales. e ecoefficiency of NSM incorporating various volume of GGBS was evaluated based on the energy requirement for the whole production of cement and GGBS.

Materials and Methods
2.1. Materials. P.O. 42.5 ordinary Portland cement was purchased from Anhui Conch Cement Co., Ltd., China. GGBS corresponding to Grade S105 was used according to JC/T2238 [34]. e apparent density and chemical composition of cement and GGBS are listed in Table 1. eir particle size distributions are presented in Figure 1. It can be found that GGBS is finer in particle sizes as the mean particle sizes of cement and GGBS are 28.19 μm and 16.18 μm, respectively. Sand was ordinary river sand with a finesse of 2.6 and an apparent density of 2550 kg/m 3 . Water used throughout this study was city tap water.
e applied characterization analysis for cement and GGBS was performed by a XRF-1800 sequential X-ray fluorescence spectroscopy (XRF) for the chemical composition investigation and an ASALD-7101 laser particle size analyzer (LPSA) for the particle size analysis.

Mixing
Proportions. Ten mixtures containing one control group were designed, as presented in Table 2. e 28 d compressive strength of the control group was 42.9 MPa, belonging to the context of NSM that normally less than 60 MPa. ree variables were considered encompassing (1) contents of cement in the blend binders (0%, 10%, 15%, 20%, 25%, and 30%), (2) curing temperatures (23°C, 40°C, and 60°C), and (3) w/b (0.35, 0.45, and 0.55). All mixtures were prepared following the same procedure. Cement and/or GGBS and sand were mixed firstly and stirred by a harbor mixer in low speeds for 5 min. en, water was added, and the mixture was stirred for another 3 min. Fresh mixture was casted into 50 mm cubic mortar for the compressive strength testing. Mortars were firstly sealed and cured in specified environments within modules for 24 h, as indicated in Table 2. Afterwards, they were demoulded and continuously cured in previous curing environment. Average value of mortar was reported by quadruplicate specimens. Mixtures were coded in the formula of C xx T yy W zz to directly indicate the content of cement in the blend binder, curing temperature, and w/b in sequence.

Setting Time and Density Tests.
e densities of fresh mixtures were simply measured by a container with fixed volume. Setting times of fresh pastes were tested by using a Vicat needle according to ASTM C191 [35].

Hydration Products Analysis.
Paste samples were soaked in isopropanol to stop the hydration process. ey were vacuum-dried and grounded into particle sizes less than 74 μm for further study. A TTR-III h/h rotating anode X-ray diffractometer (XRD) with Cu-Ka radiation at 40 kV and 200 mA was applied to study the crystal compositions. e scan speed and scope of each sample was 8°/min and 2θ � 10°-60°, respectively. A TGA/DSC 1/1600 high-temperature thermogravimetric analyzer was also applied to further study the quantitative composition of the hydration products of each sample. e grounded pastes were heated from 50°C to 800°C with the speed of 10°C/min.

Microstructure and Pore Structure Analysis.
Microstructure analysis of 28 d mortars was performed by a FlexSEM1000 scanning electron microscope (SEM). SEM   Advances in Civil Engineering samples were collected from the compressive strength testing and were stored in isopropanol to cease the hydration process. Prior to SEM observation, vacuum drying was applied on these samples. A layer of palladium particles was sprayed on the fresh fracture surface for 120 s in the frontage and 30 s in the side direction to achieve well electroconductibility. e pore structure of 28 d mortars was studied by an AutoPoreIv 9510 mercury intrusion porosimetry (MIP).

Ecoefficiency Evaluation.
Ecoefficiency evaluation was carried out for a full-life cycle environmental assessment of the proposed technology. e ecoefficiency of the blend cement was evaluated by simply calculating the energy required for cement manufacture based on the energy requirement for the whole production of cement and GGBS, as proposed by Hamidi et al. [36,37]. e required energy is given by the following equation: where E is the energy consumption for producing 1 t binder (kWh/t); C and G are the mass contents of cement and GGBS, respectively; E process cement and E grinding cement are the energy consumptions of the production process and the grinding of cement, respectively (kWh/t); E grinding GGBS is the energy consumption of the grinding of GGBS (kWh/t); E performance is the energy consumption for producing 1 t binder, relative to the 28 d compressive strength of mortar s ((kWh/t)/MPa); and f cm is the 28 d compressive strength of mortar. In the case of that, GGBS is a by-product of the steel production, and no energy is specially needed for its production. erefore, the value of E process GGBS is zero. Table 3 presents the setting times and densities of fresh mixtures. In general, both the initial and final setting times of pastes are found to be prolonged with increased contents of GGBS. e substitutions of 70% to 90% cement by GGBS resulted in increased initial and final setting times by times of 1.09 to 1.31 and 1.12 to 1.68, respectively. is is more prominent in C20T23W55 but vanishes in C20T23W35, in which the initial and final setting times are 39 min and 6 min less than that of Con. Moreover, elevated curing temperatures accelerated significantly the set of paste. For example, the initial and final setting times of C20T40W45 are about half that of C20T23W45. In addition, the densities of GGBS-incorporating groups are lower than that of Con as a result of the lower density of GGBS than cement, as shown in Table 1.

Hydration Products of Pastes
Analyzed by XRD. XRD analysis was conducted to investigate the crystalline compositions of 3 d and 28 d pastes, as presented in Figures 2  and 3, respectively. A total of ten substance, i.e., CH (Ca(OH) 2 ), calcite, C 4 AH 13 , C 3 S, C 2 S, CSH (I), gehlenite, quartz, hydrotalite, and AFt (ettringite), were identified out as marked in the figures according to their JCDP files. Figure 2(a) shows the XRD patterns of 3 d pastes incorporating 0% or 70% to 90% GGBS cured at 23°C with the fixed w/b � 0.45. Among the ten identified substance, hydrocalumite with the characteristic peak at 2θ � 10.8°and gehlenite with the characteristic peak at 2θ � 30.9°were only found in the GGBS-incorporating pastes other than Con.
ese new crystals are thus speculated as a result of the activation reaction of GGBS by the free CH released from cement hydration. In fact, the hydration of cement-GGBS cementitious system primarily involves two associated hydration processes, which are the self-hydration of cement and the activation of GGBS by alkalis, i.e., CH [7,11]. e activation reaction of GGBS is highly depended on the released CH by the self-hydration of cement in both the content and the formation rate [19]. e quantity of CH within paste is a critical indicator to understand the underlying hydration mechanism. Figure 2(a) reveals that the characteristic peak heights of CH at 2θ � 18°and 33°increase in a saltatory tendency with the increasing contents of cement from 10% to 30%. Tiny peaks of CH were detected in C15T23W45 rather than in C10T23W45. is may indicate that the released CH was consumed almost completely by the activation reaction of GGBS.
is, on the one hand, confirms the dilution effect of cement by GGBS leading to less CH production and, on the other hand, suggests an initial equilibrium between the demand of GGBS for the activation reaction and the supply of CH by the cement hydration at early age. 15% cement is sufficient to provide the required CH for 85% GGBS at 3 d. Figure 2(b) presents the XRD patterns of 3 d pastes considering the variables curing temperature and w/b. As for curing temperature, the characteristic peak heights of CH within pastes located at 2θ � 18°and 33°firstly increase from 23°C to 40°C but then decrease or even tend to disappear from 40°C to 60°C. is may be the competitive result of both the enhanced self-hydration of cement and activation of GGBS by elevated curing temperatures. From 40°C to 60°C, the activation reaction of GGBS seems to be far accelerated than that of the self-hydration of cement, leading to most CH consumption. In addition, the XRD patterns of C20T23W55, C20T23W45, and C20T23W35 display no distinct difference in this figure. Figure 3(a) presents the XRD patterns of 28 d pastes with the content of cement as the variable. A similar tendency that the characteristic peak heights of CH increase gradually with the content of cement is found in Figure 2(a). e equilibrium proportion of the blend cement for sufficient activation of GGBS has shifted from 85% GGBS-15% cement at 3 d to 90% GGBS-10% cement at 28 d since C10T23W45 has the lowest peak heights of CH at 2θ � 18°and 33°. Figure 3   Advances in Civil Engineering 5 further confirms that the consumption rates of CH by the activation reaction with GGBS surpassed the formation rates of CH by the self-hydration of cement at elevated curing temperatures. Figures 4(a) and 5(b) present the TGA-DTG results of 3 d and 28 d pastes with cement content as the variable, respectively. Four major endothermic peaks were distinctly identified in these DTG curves as marked, which correspond to the dehydration of CSH as well as decomposition of ettringite and gehlenite ranging from 80°C to 200°C, dehydroxylation of CH around 450°C, and decarbonation of CaCO 3 (CC(-)) crystals in amorphous or crystalline phases around 700°C [38][39][40][41]. In general, the endothermic peak heights or their enveloping areas are observed visually to increase when more cement was incorporated. e amounts of CH and CC(-) were calculated based on the mass losses ranging from 400°C to 500°C for CH and 600°C to 800°C for CC(-), as presented in Figure 4(c). e formation of CC(-) was attributed to the carbonation of CH during the moisture curing. erefore, the total CH encompasses the content of CH and the carbonated CH with the same stoichiometric mole to that of CC(-). For 3 d pastes, the amounts of CH or CC(-) or the total CH of these pastes are less than half of that of Con (i.e., 30%-50% for the total CH). is is due to the dilution effect of cement by GGBS at early ages in particular. For 28 d pastes, the contents of CH and total CH of GGBS-incorporating groups are far lower than that of Con. For instance, the contents of total CH of these GGBS-containing groups are just 15%-25% of that of Con. is decreased CH and total CH from 3 d to 28 d reveals the consumption of CH by GGBS. Figures 5(a) and 5(b) show the TGA-DTG curves of 3 d and 28 d pastes prepared by different curing temperatures and w/b. e endothermic peaks of DTG curves are similar to that in Figure 4. In general, the contents of CH and total CH of pastes increase, and the content of CC(-) decreases with increasing curing temperatures, as further studied in Figure 5(c). is confirms that elevated curing temperatures were beneficial for the improvement of cement hydration, contributing to the generation of more hydration products. e amounts of CH or total CH of pastes with different w/b values decrease with the decreasing w/b at 3 d and 28 d. Since the contents of cement in pastes with lower w/b were actually higher as less water was used, it is indicated that lower w/b promoted the consumption rates of CH. Figure 6(a) presents the compressive strength of mortars with different contents of cement. In general, the compressive strength of these mortars increases with increasing contents of cement at three ages. e early-age compressive strengths of GGBS-incorporating mortars are markedly lower than that of Con and are 30%-50% at 3 d and 40%-79% at 7 d of Con, respectively. Nevertheless, the 28 d compressive strengths of those GGBSincorporating mortars developed much quickly, and the values of C25T23W45 and C30T23W45 are 90.7% and 95.3% of that of Con. e subsequent increments of the compressive strengths of GGBS-incorporating mortars are primarily attributed to the activation reaction of GGBS with the free CH. rough this activation reaction, supplementary stress-bearing C-S-H can be generated [7,10]. is also implies important achievement that one can replace 75% or 70% cement by GGBS to prepare NSM with comparable longer compressive strength to that of mixtures prepared by cement only. anks to this, great benefits can be gained in terms of environment protection, cost reduction, waste utilization, as well as energy saving. Figure 6(b) presents the compressive strength of mortars prepared by different curing temperatures and w/b. It can be found that 40°C is the optimal curing temperature to achieve desired compressive strength. Compared to Con, the compressive strengths of C20T40W45 are higher by 7.8% at 3 d and 5.9% at 7 d than that of Con. e 28 d compressive strength of C20T40W45 is lower but still gained 91% of that of Con at 39.0 MPa. From the point of chemical reaction, curing temperature is one of the methods to facilitate the dissolution of GGBS and cement and crystallization of hydration products. However, too high temperatures may cause degraded properties of the hydration products as well as the mechanical strength, i.e., 60°C. Another method to gain desired compressive strength is well adopted reducing the w/b. As can be seen from this figure, the compressive strength normally increases with decreased w/b mainly due to formation of a denser microstructure. e compressive strengths of C20T23W35 are 74.2% at 3 d, 95.2% at 7 d, and 93.2% at 28 d of that of Con.

Microstructure Observation by SEM.
e microstructure evolution of GGBS-incorporating mortars was examined by SEM, as shown in Figure 7. Figure 7(a) presents the SEM image of 3 d C10T23W45 mortar. It displays a much porous microstructure constituted by adhesive particles and tiny schistose crystals that most possibly are CH. In the case of that 90% of the binder is GGBS, the particles are GGBS that have reacted with the released CH, nevertheless with low reaction degree. When incorporating 20% cement into the mortar (Figure 7(c)), a larger porous matrix formed instead of part of the adhesive particles, as observed in Figure 7(a). More prominently, as shown in Figure 7(e), the morphology of the 3 d C30T23W45 mortar is much denser as irregular particles are bonded compactly by porous binders as further observed from the inset in this figure. e differences in the 3 d morphology among these mortars vanished at 28 d as they display a similar microstructure. As presented in Figures 7(b), 7(d), and 7(f ), a dense matrix formed in the three samples despite the contents of GGBS varying from 70% to 90%. is confirms that a longer curing duration has transformed GGBS into the binder that structured the dense stress-bearing matrix. Figure 7(g) shows the SEM images of 3 d C20T40W45 mortar. Large dense matrix and fine particles are found, and ribbon-like CSH formed. is figure is much different from that of C20T23W45 with the same proportion but lower curing temperature at 23°C. is confirms the role of a higher curing temperature in enhancing both the self-hydration of cement and activation reaction of GGBS by CH. Figure 7(h) presents the SEM image of 28 d C20T40W45 mortar. e noteworthy feature of this figure is the existence of fibrous CSH intertwining together. Figure 7(i) presents the SEM images of 3 d C20T23W35. It displays a porous morphology and nevertheless is still denser than that of C20T23W45, as proved in Figure 7(j). is is one of the reasons for the higher compressive strength of mortars with a lower w/b.

Microstructure Analysis by MIP.
e pore structures of representative mortars were studied by MIP, as presented in Figure 8. e pore sizes of all mortars primarily distribute within two zones: ZI (>50 μm) and ZII (<100 nm). e former commonly suggests the existence of voids or gas inclusion. Within ZI, the pore volume of Con is the lowest while that of GGBS-incorporating mortars cured under 23°C  Advances in Civil Engineering are much higher. Within ZII, the pore sizes of GGBS-incorporating groups are finer than that of Con. It can be seen that the pore sizes of Con mainly concentrate around 50 nm as a peak. When 90% GGBS and 10% cement were blended, two peaks of pore sizes appear in C10T23W45. e first one is higher but shifts left to the sole peak of Con. e second intense peak of pore sizes formed around 10 nm, as further expounded by the inset. When blending more cement in mortars, the first peak of pore sizes observed in C10T23W45 disappears in C20T23W45 and C30T23W45, and consequently, broader peaks of pore sizes formed. is suggests that the incorporation of GGBS into NSM results in promoted porosity but refined pore sizes. e formation of the binding system in these mortars incorporating GGBS as the major cementing material highly depended on the activation reaction of GGBS with the free CH. e activation reaction generated hydration products which filled the voids or pores within the microstructure of mortars. Nevertheless, the produced hydration products among these GGBS-incorporating mortars are less than that of Con, as calculated in Figures 4 and 5, leading to less pores being occupied or filled. Elevated curing temperatures or low w/b are also helpful to refine the pore sizes of GGBS-incorporating mortars. e major difference between the two compressive strength compensation methods on the pore sizes is that elevated curing temperatures increase the refined pore sizes from 10 nm to 20 nm within ZII and nevertheless decrease the pore volumes within ZI, as found in Figure 8 and the inset. Figure 9 presents the calculated cumulative pore volume based on Figure 8. is figure clearly confirms that the incorporation of GGBS promoted the pore volumes of NSM as the ultimate cumulative pore volume of Con is the lowest. Nevertheless, the pore sizes of GGBS-incorporating groups are finer than that of Con since most ascending curves of GGBS-incorporating groups shift left to that of Con. is refinement in pore sizes is beneficial to improve the durability performance of NSM with respect to external environmental actions, i.e., chloride migration and sulfate attack [16].

Ecoefficiency Evaluation of the Binary Cement.
e energy consumption for producing and grinding 1 t cement has been reported as E process cement � 1039 kWh/t [42] and E grinding cement � 50 kWh/t [37]. According to DB31 581 [43], a comprehensive electricity consumption of GGBS encompasses the whole production process of GGBS, i.e., raw material conveying, drying, grinding, powder separating, product storage and delivering, and the supporting processes, i.e., electricity consumption for illumination and temperature control. It has been specified that the norm of comprehensive electricity consumption of GGBS should be less than 50 kWh/ t. erefore, the energy consumption for grinding 1 t GGBS is determined as E grinding GGBS � 50 kWh/t in this paper. Figure 10 presents the calculated energy requirements of 1 t binder E and performance energy E performance . e ecoefficiency of NSM incorporating high volume of GGBS is obvious. In general, incorporating 70% to 90% GGBS into binders contributes to decreased E by 67% to 86% as well as lessened E performance by 65% to 75%. Among the ten mixtures, the three lowest values of E performance are 6.28 (kWh/t)/ MPa (C15T23W45), 6.44 (kWh/t)/MPa (C20T23W35), and 6.61 (kWh/t)/MPa (C20T40W45). From the point of performance energy, it is thus suggested to replace 80% to 85% cement by GGBS to gain optimal ecoefficiency. Low w/b and high curing temperature at 40°C are effective methods to gain a lower E performance . erefore, the ecological upgrade of NSM by using high volume of GGBS provides a better choice for the sustainable development in a greener way. It should be noted that the calculation of E performance of C20T40W45 and C20T60W45 did not involve the extra energy for high temperature curing; thus, the ecoefficiency was underestimated from a practical point.

Conclusions
is paper has evaluated the feasibility of ecological upgrade of NSM by using high volume of GGBS (70%-90%). e following conclusions can be drawn based on the experimental results: (i) Ecological upgrade of NSM is viable and obvious by substituting 70% to 90% cement by GGBS. Accordingly, the energy requirements of 1 t binder can be saved by 67% to 86%, and the performance energy can be reduced from 25.4 (kWh/t)/MPa to 6 to 8 (kWh/t)/MPa. (ii) e setting times of pastes were prolonged with increased GGBS. e substitution of 70% to 90% cement by GGBS resulted in increased initial and final setting times by times of 1.09 to 1.31 and 1.12 to 1.68, respectively. (iii) NSM with proper contents of GGBS can develop acceptable 28 d compressive strengths than that of the control group. e early-age compressive strengths of GGBS-incorporating NSM were far lower but can be mitigated by methods of evaluated temperature curing or decreasing w/b. (iv) e incorporation of high volume of GGBS increased the porosities but refined the pore structures of mortars. Mortars with GGBS displayed very porous microstructures at 3 d with visible bonded GGBS particles. Nevertheless, raising the curing temperature to 40°C or decreasing w/b at 0.35 helped dense the microstructures. At 28 d, all mortars displayed a similar morphology consisted by a dense matrix. Moreover, the pore sizes of mortars were significantly refined by incorporating GGBS.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that they have no conflicts of interest. Advances in Civil Engineering 11