Effects of γ -C 2 S on the Properties of Ground Granulated Blast-Furnace Slag Mortar in Natural and Accelerated Carbonation Curing

: γ -Dicalcium silicate ( γ -C 2 S) is known for its strong carbonation reactivity by which it can capture atmospheric carbon dioxide (CO 2 ), thus, it can be used in construction industries. This paper aims to study the effects of γ -C 2 S on the properties of ground granulated blast-furnace slag (GGBFS) containing cement mortar and paste in natural and accelerated carbonation curing. The compressive strength of 5% γ -C 2 S (G5) added to GGBFS cement mortar is higher compared with the control one in natural carbonation (NC) and accelerated carbonation (AC) up to 14 days of curing, but once the curing duration is increased, there is no signiﬁcant improvement with the compressive strength observed. The compressive strength of AC-cured mortar samples is higher than that of NC. The scanning electron microscopy (SEM) images show that the AC samples exhibited compact, uniform, and regular morphology with less in porosity than the NC samples. X-ray diffraction (XRD) and Fourier transform infra-red (FT-IR) results conﬁrmed the formation of calcium carbonate (calcite: CC) as carbonated products in paste samples, which make the surface dense and a defect-free matrix result in the highest compressive strength. The decomposition of AC samples around 650–750 ◦ C revealed the well-documented and stable crystalline CC peaks, as observed by thermogravimetry analysis (TGA). This study suggests that γ -C 2 S added to concrete can capture atmospheric CO 2 (mostly generated from cement and metallurgy industries), and make the concrete dense and compact, resulting in improved compressive strength. AC exhibited higher compressive strength, compact morphology with lower porosity, and intense CC peaks.


Introduction
Portland cement is the most prevalent form of cement used as an essential ingredient in construction industries worldwide. However, the production of Portland cement requires a huge amount of energy and produces large amounts of greenhouse gases [1,2]. The cement industry has generated a major contribution to emissions of CO 2 , which is an alarming issue for the world. According to 2017 statistical data, it is estimated that the production of Portland cement accounts for 4% of global CO 2 emissions [3,4]. Significant efforts have been made to reduce CO 2 emissions, as well as calcium oxide content, in Portland cement.
Since the 2000s, controlling the emissions of CO 2 from cementitious supplementary materials is required [5]. The accelerated carbonation of cement-based materials exhibited beneficial effects on the properties of cement paste [3]. The carbonation of ordinary Portland cement occurs via diffusion of CO 2 into the pore matrix from the atmosphere, which reacts with calcium silicate hydrate (C-S-H) and calcium hydroxide (Ca(OH) 2 ) to form calcium

Mixture Proportions for the Preparation of Samples
Ordinary Portland cement (OPC), ground granulated blast-furnace slag (GGBFS), and γ-C 2 S were used as binders. The OPC was Type-I grade, having a density of 3.15 g/cm 3 . The densities of the GGBFS and γ-C 2 S were found to be 2.92 and 2.99 g/cm 3 , respectively. The appearance of the binders is shown in Figure 1. The OPC was gray in color (Figure 1a), owing to the presence of some unburnt carbon. The GGBFS showed white in color (Figure 1b), possibly due to the presence of a high content of alumina, which could be confirmed by chemical analysis. γ-C 2 S looked light brown in color (Figure 1c). The synthesized γ-C 2 S and GGBFS was milled and sieved through 150 µm before mixing in the mortar and paste. tainability 2021, 13, x FOR PEER REVIEW 3 of 17 for 2 h (sintering) [24,25]. Finally, the cooling of the samples was carried out at room temperature. In this process, a series of calcium silicates were formed at different temperatures, which decompose from β-C2S to γ-C2S spontaneously upon cooling.

Mixture Proportions for the Preparation of Samples
Ordinary Portland cement (OPC), ground granulated blast-furnace slag (GGBFS), and γ-C2S were used as binders. The OPC was Type-I grade, having a density of 3.15 g/cm 3 . The densities of the GGBFS and γ-C2S were found to be 2.92 and 2.99 g/cm 3 , respectively. The appearance of the binders is shown in Figure 1. The OPC was gray in color (Figure 1a), owing to the presence of some unburnt carbon. The GGBFS showed white in color (Figure 1b), possibly due to the presence of a high content of alumina, which could be confirmed by chemical analysis. γ-C2S looked light brown in color (Figure 1c). The synthesized γ-C2S and GGBFS was milled and sieved through 150 µ m before mixing in the mortar and paste. Two sets of experiments were performed to investigate the effects of γ-C2S on the properties of GGBFS incorporated with mortar and paste. The first set of experiments was the preparation of mortar for compressive strength measurements, while the second was testing the paste for characterization (i.e., X-ray diffraction (XRD), Fourier transform-infra red (FT-IR), scanning electron microscopy (SEM), and thermogravimetric analysis (TGA)) after carbonation. The fine aggregate was sand, having a 2.65 g/cm 3 density, and it was properly washed with distilled water to remove soluble impurities, and then dried prior to mixing into the mortars.
In the present study, three binders (OPC, GGBFS, and γ-C2S) were used. The mixture proportion of mortar and paste samples are shown in Tables 1 and 2, respectively. We took a 0.55 ratio of water/binder (W/B) for the preparation of mortar and paste. The W/B was kept higher to facilitate the carbonation process easily [26]. From these tables, it can be seen that the OPC took 40% in all compositions, but the GGBFS and γ-C2S were varied. The amount of γ-C2S was set to be fixed to 0% (G0), 5% (G5), and 10% (G10), corresponding to the change in the amounts of GGBFS of 60%, 55%, and 50%, respectively. Two sets of experiments were performed to investigate the effects of γ-C 2 S on the properties of GGBFS incorporated with mortar and paste. The first set of experiments was the preparation of mortar for compressive strength measurements, while the second was testing the paste for characterization (i.e., X-ray diffraction (XRD), Fourier transform-infra red (FT-IR), scanning electron microscopy (SEM), and thermogravimetric analysis (TGA)) after carbonation. The fine aggregate was sand, having a 2.65 g/cm 3 density, and it was properly washed with distilled water to remove soluble impurities, and then dried prior to mixing into the mortars.
In the present study, three binders (OPC, GGBFS, and γ-C 2 S) were used. The mixture proportion of mortar and paste samples are shown in Tables 1 and 2, respectively. We took a 0.55 ratio of water/binder (W/B) for the preparation of mortar and paste. The W/B was kept higher to facilitate the carbonation process easily [26]. From these tables, it can be seen that the OPC took 40% in all compositions, but the GGBFS and γ-C 2 S were varied. The amount of γ-C 2 S was set to be fixed to 0% (G0), 5% (G5), and 10% (G10), corresponding to the change in the amounts of GGBFS of 60%, 55%, and 50%, respectively.

Accelerated Carbonation and Curing
Two types of curing (natural carbonation (NC) and accelerated carbonation (AC) curing) for mortar, as well as paste samples, were carried out. The NC curing was performed by keeping the mortar and paste samples in laboratory conditions; the temperature was 20 (±2) • C and the relative humidity was 60 (±5%) where there was approximately 0.03% CO 2 present [28,29]. The accelerated curing happened in a carbonation chamber (Chom Dan Scientific Ind. Co., Seoul, Korea) for up to 28 days. The paste and mortar samples were kept in a carbonation chamber, as shown in Figure 2. The AC was performed by keeping the samples in 5% CO 2 at 20 • C and 60% relative humidity, according to KS F2584 [30]. The details of the curing process and characterizations of the samples are illustrated in Table 3.

Compressive Strength
The compressive strength of 50 × 50 × 50 mm mortar samples was measured according to ASTM C109/C109M [31] by the 30 t class universal testing machine. An average loading rate of 80 kN/min was applied to determine the compressive strength of three consecutive samples after 14 and 28 days in different curing conditions, and the average was taken as the result.

Chemical Analysis of Binders
The oxide composition of all binders was performed by X-ray fluorescence (XRF: Axios, PANalytical, Almelo, The Netherlands).  The compressive strength of 50 × 50 × 50 mm mortar samples was measured according to ASTM C109/C109M [31] by the 30 t class universal testing machine. An average loading rate of 80 kN/min was applied to determine the compressive strength of three consecutive samples after 14 and 28 days in different curing conditions, and the average was taken as the result.

Chemical Analysis of Binders
The oxide composition of all binders was performed by X-ray fluorescence (XRF: Axios, PANalytical, Almelo, The Netherlands).

X-ray Diffraction (XRD)
The mineralogical characterization of synthesized γ-C 2 S, GGBFS powder, and carbonated paste was carried out by XRD (Rigaku D/MAX-2500, Tokyo, Japan) using Cu Kα radiation (λ = 1.54059 Å), generated at 40 kV and 100 mA. The scan was carried out from 2θ = 10 • -70 • with a 0.0167 • step size and 30 s per step. The volume fraction (%) of each was determined by JADE2016 software.

Fourier Transform Infrared Spectroscopy (FT-IR)
The FT-IR (PerkinElmer Spectrum, L160000A, Waltham, USA) was used to quantify the carbonated paste samples after 28 days of curing. The scan range was 4000 cm −1 to 400 cm −1 with a 4 cm −1 resolution.

Scanning Electron Microscope (SEM)
Surface morphology of the paste matrix after 28 days of NC and AC curing was performed by scanning electron microscope (SEM, MIRA3, TESCAN, Brno, Czech Republic), operated at 20 kV.

Thermogravimetric Analysis (TGA)
The weight loss of the carbonated paste sample was measured by TGA (DTG-60, Shimadzu, Japan). The temperature range was from room temperature to 1000 • C in an N 2 atmosphere (20 mL/min) at a 10 • C/min constant heating rate.

Chemical Composition of Binders by XRF
The chemical analysis of binders is shown in Table 4. From this table, it can be seen that the γ-C 2 S mostly contained CaO and SiO 2 (94.73%), whereas the GGBFS and OPC exhibited 80.18% and 83.16%, respectively. However, the GGBFS contained a high amount of Al 2 O 3 impurities (13.32%), due to the by-products of steel industries. The loss of ignition (LOI) in γ-C 2 S and GGBFS was much less, owing to the heating. The γ-C 2 S was produced by calcination up to 1450 • C, while the GGBFS was produced from the blast furnace, where the temperature was around 1300 • C, and thus, most of the carbon was burned. The OPC exhibited a 3.53% LOI, attributed to the presence of unburned carbon.

XRD of Binders
The XRF results (Table 4) show that the synthesized γ-C 2 S mostly contained CaO (68.92wt.%) and SiO 2 (25.81wt.%). Therefore, it was required to confirm the mineralogical composition by XRD. Figure 3 shows the mineralogical composition of the γ-C 2 S and GGBFS. The XRD results confirmed the formation of γ-C 2 S (JCPDS: 86-0397) [32], whereas the GGBFS showed amorphous silica [6,33]. It was earlier reported by Mu et al. that if the Ca/Si ratio was 2:1 and treated at 1400 • C, then pure γ-C 2 S formed [10]. However, the synthesis of γ-C 2 S depends on different parameters, such as the types of CaCO 3 [34] and heating temperatures. The γ-C 2 S formed from the transformation of β-C 2 S. It was observed that, up to 1300 • C, some amount of β-C 2 S still remained, but once the temperature increased up to 1400 • C, it was completely transformed into γ-C 2 S [34]. In the present study, we sintered it up to 1450 • C for 2 h, which lead to forming pure γ-C 2 S. Thus, we can say that our findings are in good agreement with Mu et al.'s work [10]. On the basis of the XRD results, it was important to determine the exact amount (%) of each of the phases present in the γ-C 2 S and GGBFS, using JADE software. It was calculated and found that synthesized γ-C 2 S showed 100%, while in the case of the GGBFS only amorphous silica was present.

Compressive Strength of Mortar
The compressive strength of the mortar samples is shown in Figure 4 at different curing conditions and durations. It can be seen from this figure that the NC mortar samples exhibited a lower compressive strength compared with the AC samples in all curing durations. In NC, there was no effect of carbonation rather than hydration. In this case, the mechanical properties were controlled by the hydration reaction rather than carbonation because the CO2 content was negligible in the open atmosphere. Moreover, once the

Compressive Strength of Mortar
The compressive strength of the mortar samples is shown in Figure 4 at different curing conditions and durations. It can be seen from this figure that the NC mortar samples exhibited a lower compressive strength compared with the AC samples in all curing durations. In NC, there was no effect of carbonation rather than hydration. In this case, the mechanical properties were controlled by the hydration reaction rather than carbonation because the CO 2 content was negligible in the open atmosphere. Moreover, once the 5% γ-C 2 S (G5) was added to the mortar, the compressive strength increased with curing duration in NC and AC conditions. However, it decreased in 10% γ-C 2 S (G10) compared with G0 in NC at 14 days and 28 days of curing. The decrease in compressive strength of 10% γ-C 2 S mixed mortar samples might be attributed either to unreactive CaO in the γ-C 2 S [34] or a higher porosity in the mortar matrix [13]. However, there was no appreciable improvement in the compressive strength after the addition of 5% γ-C 2 S in NC compared with G0, owing to the hydration reaction, rather than carbonation [35]. The concentration of CO 2 was negligible in NC, which was not able to react with the γ-C 2 S and form CaCO 3 . In this case, the pozzolanic or hydration reaction was dominant, and thus, almost identical compressive strength was obtained by the G0 and G5 samples for all curing durations. However, once the mortar was cured in the carbonation chamber (AC), the compressive strength increased (Figure 4). The obtained compressive strength values of mortar samples cured in AC and NC were greater than those in Higuch et al. [14] and Yang et al. [17] works. After 14 days of AC curing, the compressive strength increased by 13.5% (G5) and 25.7% (G10), compared with 28 days of NC, owing to the formation of CaCO 3 , which reduced the pore size [35]. The formed calcium carbonate crystals were inclined to fill the pores of the cementitious matrix. In particular, the G5 samples achieved the highest value of compressive strength after AC, which might be owing to the lowest pore volume. Moreover, γ-C 2 S-incorporated mortar samples (G5 and G10) exhibited higher compressive strength compared with G0 in AC for all curing durations. Nevertheless, as the curing duration extended up to 28 days, there was no significant improvement in compressive strength compared with 14 days. This result suggests that the compressive strength of γ-C 2 S-incorporated mortar exhibit early improvement owing to the reaction of γ-C 2 S with CO 2 , resulting in the formation of CaCO 3 , while a longer duration of curing was dominated by the hydration reaction [35]. It was reported that γ-C 2 S-incorporated samples usually require 10 days to get their final strength [36,37]. Thus, with a longer duration of curing, there was no significant improvement in compressive strength. The hydration reaction occurred via the formation of calcium-silicate-hydrate (C-S-H) gel, while the γ-C 2 S could not react with water, and leaching of Ca 2+ ions was less at a low temperature (i.e., 20 • C), and thus, CaCO 3 was generated [38]. The γ-C 2 S reacted with CO 2 and formed CaCO 3 , which filled the pores of the mortar matrix and increased the compressive strength at an early age of curing [35].
CO2, resulting in the formation of CaCO3, while a longer duration of curing was dominated by the hydration reaction [35]. It was reported that γ-C2S-incorporated samples usually require 10 days to get their final strength [36,37]. Thus, with a longer duration of curing, there was no significant improvement in compressive strength. The hydration reaction occurred via the formation of calcium-silicate-hydrate (C-S-H) gel, while the γ-C2S could not react with water, and leaching of Ca 2+ ions was less at a low temperature (i.e., 20 °C ), and thus, CaCO3 was generated [38]. The γ-C2S reacted with CO2 and formed CaCO3, which filled the pores of the mortar matrix and increased the compressive strength at an early age of curing [35].

SEM of Paste after Carbonation
On the basis of compressive strength, it was required to analyze the morphology of the polished paste samples. The polishing of the paste samples was carried out by emery paper up to a 1200 µ m size. The back-scattered electron (BSE) SEM images of NC and AC paste samples after 28 days of curing are shown in Figures 5 and 6, respectively. It can be seen from Figure 5a that there were many pores, along with unreacted GGBFS, in white

SEM of Paste after Carbonation
On the basis of compressive strength, it was required to analyze the morphology of the polished paste samples. The polishing of the paste samples was carried out by emery paper up to a 1200 µm size. The back-scattered electron (BSE) SEM images of NC and AC paste samples after 28 days of curing are shown in Figures 5 and 6, respectively. It can be seen from Figure 5a that there were many pores, along with unreacted GGBFS, in white color. Some scratch marks were present on the surfaces, owing to the polishing of paste samples. The size of unreacted GGBFS in G0 (Figure 5a) was different. However, once the 5% and 10% γ-C 2 S was added, the number of pores decreased significantly. The porosity was calculated by ImageJ software and was found to be 4.24%, 0.83%, and 1.76% for G0, G5, and G10, respectively. Once the amount of γ-C 2 S increased from 5% to 10%, the unreacted γ-C 2 S led to retain porosity. The porosity was significantly reduced in G5 samples owing to the reaction of γ-C 2 S with CO 2 , and led to forming the CaCO 3 , which filled the porosity. From Figures 5 and 6, it can be illustrated that different color zones were observed. The white areas are mostly unreacted GGBFS, which correlates with Figure 1b, where the GGBFS particles are white in color. The unreacted GGBFS was found to be in all samples of NC ( Figure 5) and AC ( Figure 6). The gray area represents the unreacted γ-C 2 S in G5 ( Figure 5b) and G10 (Figure 5c). The amount of unreacted γ-C 2 S in G10 (Figure 5c) was higher compared with G5 (Figure 5b), owing to the higher added amount. Consequently, it can be seen that the amount of CO 2 in NC was negligible, and thus, the γ-C 2 S remained in an unreacted state, attributed to less solubility in water than in CO 2 in the atmosphere [38]. Thus, the compressive strength of G0 and G5 after 28 days of NC curing was identical, whereas once the amount of γ-C 2 S increased more than 5%, the compressive strength decreased (Figure 4), which was attributed to the unreacted γ-C 2 S in the matrix of the mortar [34]. The medium gray color in Figure 5 is attributed to the presence of carbonated products [10]. The visual observations in the G5 samples (Figure 5b) reveal the greater area of carbonated products compared with others.
Sustainability 2021, 13, x FOR PEER REVIEW 9 of 17 was identical, whereas once the amount of γ-C2S increased more than 5%, the compressive strength decreased (Figure 4), which was attributed to the unreacted γ-C2S in the matrix of the mortar [34]. The medium gray color in Figure 5 is attributed to the presence of carbonated products [10]. The visual observations in the G5 samples (Figure 5b) reveal the greater area of carbonated products compared with others. A noticeable difference is observed in the SEM images of the AC samples ( Figure 6) compared with NC after 28 days of curing. Figure 6a shows the SEM image of G0, where unreacted GGBFS particles were smaller compared with NC, owing to the dissolution in accelerated conditions (CO2 curing). Due to the dissolution of GGBFS, the porosity was found to be 2.07%, which was reduced by more than two times compared with NC. The

XRD of Paste after Carbonation
XRD of the hydration products after 28 days of NC and AC curing are shown in Figures 7a and b, respectively. It can be seen from Figure 7a that G0, G5, and G10 exhibited a presence of calcium-silicate-hydrate (C-S-H: JCPDF: 89-7639), as well as two types of carbonate: stable calcite/calcium carbonate (CC: CaCO3, JCPDF: 72-1937) and metastable Vaterite (V: CaCO3, JCPDF: 72-1616) [39,40]. An amorphous peak was observed in G0 samples cured in NC ( Figure 6a) and AC (Figure 6b) around 2θ = 27-33°, attributed to the presence of unreacted GGBFS. This result corroborates the XRD of bulk GGBFS, as shown in Figure 3. The samples cured in NC conditions exhibited the presence of γ-C2S in G5 and G10 samples, owing to the presence of unreacted parts, which are also observed in the SEM images of Figures 5b and c, respectively. However, the intensity and number of γ-C2S peaks in G5 were lower than in G10 (Figure 7a). From Figure 7a, it can be seen that the relative intensity ratio (RIR) of C-S-H was the highest, and dominated compared with CC and V in the G0 sample, attributed to the proper hydration (slag hydration reaction and pozzolanic reaction) where plenty of amorphous silica was present. Moreover, when comparing the RIR of V with CC, it is greater than CC. However, in the samples prepared with 5% γ-C2S (G5), the RIR of CC was greater, meaning it dominated over V, owing to the consumption of γ-C2S in the formation of CC [41]. Therefore, the XRD results of the G5 samples cured in NC exhibited greater RIR of CC compared with V. Once the amount of γ-C2S increased up to 10% (G10), the unreacted amount of γ-C2S remained higher in samples compared with G5, attributed to less availability of CO2 to react in NC curing, which can be easily seen in the SEM image (Figure 5c). Some studies have reported that γ-C2S requires an alkali environment for its activation process, whereas the GGBFS in the G5 samples partly met the activator demand, which contributed to the mechanical properties [42,43]. Thus, the compressive strength of the mortar decreased ( Figure 4). A noticeable difference is observed in the SEM images of the AC samples ( Figure 6) compared with NC after 28 days of curing. Figure 6a shows the SEM image of G0, where unreacted GGBFS particles were smaller compared with NC, owing to the dissolution in accelerated conditions (CO 2 curing). Due to the dissolution of GGBFS, the porosity was found to be 2.07%, which was reduced by more than two times compared with NC. The medium gray color reveals the carbonated area in all samples [10]. From Figure 6b, it can be seen that the unreacted γ-C 2 S was extremely low, which confirms that most of the γ-C 2 S reacted and formed CaCO 3 , and thus, compact and negligible porosity (0.14%) were observed, which confers the enhancement of compressive strength of mortar after 28 days of AC curing. Due to the high amount of unreacted γ-C 2 S in G10, the porosity slightly increased to 0.49% compared with G5, which led to forming pores in the samples, and thus, the reduction in compressive strength was observed (Figure 4). It is observed from Figures 5 and 6 that if the γ-C 2 S amount increased more than 5%, the unreacted γ-C 2 S remained in the samples and formed the cracks and pores which led to reducing the compressive strength of NC, as well as AC, after 28 days of curing. It can be seen from Figure 6c that unreacted γ-C 2 S was covered by SiO 2 gel and the pores were filled with CaCO 3 [1,2].

XRD of Paste after Carbonation
XRD of the hydration products after 28 days of NC and AC curing are shown in Figures 7a and 7b, respectively. It can be seen from Figure 7a Figure 3. The samples cured in NC conditions exhibited the presence of γ-C 2 S in G5 and G10 samples, owing to the presence of unreacted parts, which are also observed in the SEM images of Figures 5b and 5c, respectively. However, the intensity and number of γ-C 2 S peaks in G5 were lower than in G10 (Figure 7a). From Figure 7a, it can be seen that the relative intensity ratio (RIR) of C-S-H was the highest, and dominated compared with CC and V in the G0 sample, attributed to the proper hydration (slag hydration reaction and pozzolanic reaction) where plenty of amorphous silica was present. Moreover, when comparing the RIR of V with CC, it is greater than CC. However, in the samples prepared with 5% γ-C 2 S (G5), the RIR of CC was greater, meaning it dominated over V, owing to the consumption of γ-C 2 S in the formation of CC [41]. Therefore, the XRD results of the G5 samples cured in NC exhibited greater RIR of CC compared with V. Once the amount of γ-C 2 S increased up to 10% (G10), the unreacted amount of γ-C 2 S remained higher in samples compared with G5, attributed to less availability of CO 2 to react in NC curing, which can be easily seen in the SEM image (Figure 5c). Some studies have reported that γ-C 2 S requires an alkali environment for its activation process, whereas the GGBFS in the G5 samples partly met the activator demand, which contributed to the mechanical properties [42,43]. Thus, the compressive strength of the mortar decreased ( Figure 4).
The XRD pattern of G0, G5, and G10 samples cured in AC are shown in Figure 7b. It can be seen from this figure that the main phases (C-S-H, CC, and V) remained the same, as observed in NC conditions. It can be seen from this figure that the G5 sample did not exhibit any unreacted γ-C 2 S, which infers that during carbonation in AC conditions, almost all γ-C 2 S reacted and formed CC and V. However, it can be seen from the SEM (Figure 6b) that a much less amount of γ-C 2 S was present in the matrix, which cannot be detected by XRD. The RIR of C-S-H in the G0 sample decreased once cured in AC, compared with NC, owing to the formation of CC and V, which participated in overall RIR. This result suggests that in AC conditions, hydration and carbonation reactions occur simultaneously. Thus, the compressive strength of AC samples was greater than NC in the G0 sample (Figure 4), which made the compact matrix of mortar and reduced the porosity. The RIR of CC and V increased in all composition mixtures of samples cured in AC conditions, owing to a greater amount of CO 2 . Therefore, the compressive strength of G5 and G10 in AC was greater than G0 after 28 days of curing ( Figure 4). Due to the addition of a higher amount (10% γ-C 2 S; G10 samples), some unreacted parts remained, as observed in XRD (Figure 7b), which caused the interference in compressive strength. Thus, a lower value was observed compared with G5. This result suggests that anything greater than 5% γ-C 2 S is not able to completely react until 28 days of curing. It is suggested to either increase the concentration of CO 2 or curing duration. However, it was earlier observed that 10 days is the optimum curing duration, where a maximum amount of γ-C 2 S reacts with CO 2 and forms CC or V [36,37]. There is a possibility that if the concentration of CO 2 is increased then the maximum amount of γ-C 2 S reacts and forms CC or V, which fills the pores of concrete or mortar matrix, resulting in higher compressive strength.   The quantitative evaluations of each phase were determined by JADE software, and the results are shown in Table 5. Of the samples cured in NC condition, the C-S-H was found to be the maximum in G0, compared with G5 and G10, owing to the existence of a higher amount of GGBFS, which mostly contained amorphous silica and influenced the pozzolanic activity. The presence of amorphous silica led to the enhancement of the hydration reaction, resulting in a higher amount of hydration products (i.e., C-S-H), rather than carbonation products. However, some CC and V peaks were also found, owing to the atmospheric carbonation, which cannot be neglected during natural curing. It can be seen that once the γ-C 2 S was added, the amount of C-S-H decreased and unreacted γ-C 2 S increased in NC conditions. The CC was found to be the maximum in G5, and thus, the highest compressive strength was observed in this sample, as shown in Figure 4. The unreacted γ-C 2 S found to be remaining in G5 and G10, owing to less availability of CO 2 in NC conditions. The γ-C 2 S was only reactive in the presence of CO 2 rather than water [38].
On the contrary, it can be seen that once the sample cured in AC conditions, the CC volume faction increased up to 50.75% in G5, which suggests that the maximum amount of γ-C 2 S reacted with CO 2 and formed CC. The absence of γ-C 2 S in the G5 sample indicates that all the γ-C 2 S reacted and formed the CC and V, which significantly filled the matrix of mortar, resulting in a reduction of porosity and increased the compressive strength. The G0 sample showed a reduction in C-S-H and an increase in CC and V amounts, compared with NC conditions, attributed to the hydration and carbonation reactions, simultaneously. Once the amount of γ-C 2 S increased up to 10%, the unreacted part remained in cement paste. Therefore, a lower CC amount was obtained compared with G5, but the V was greater than with G5. However, V is metastable and can be transformed into CC, which is very stable if the curing duration is increased [13,44].

FT-IR of Paste After Carbonation
The mineralogical analysis of the paste samples after 28 days of NC and AC curing by FT-IR in the range of 4000-400 cm −1 are shown in Figures 8a and 8b, respectively. It can be seen from Figure 8 that the weak absorption peak at around 3380 cm −1 represents the vibrational bending band of O-H in H 2 O molecules [26]. The peaks around 874 cm −1 and 1400 cm −1 correspond to the C-O bending and stretching vibration of carbonated phases [45][46][47]. However, it can be seen from Figure 8a that the shifts are found to be at 870 cm −1 and 1380 cm −1 , attributed to the lower carbonation activity in NC, while in the case of AC curing, it is shifted towards a higher wave number (875 cm −1 and 1420 cm −1 ) (Figure 8b), owing to the formation of CC, which is in agreement with Lee et al.'s work [26]. Moreover, can clearly be observed that the CO 3 2− peak shifts in AC samples were intense compared with NC, especially the G5 samples, attributed to the complete carbonation. This observation corroborates the SEM ( Figure 6) and XRD (Figure 7b) results, where the surface was compact with the least porosity and intense CC peaks, respectively. This result confirms that 5% γ-C 2 S is the optimum amount, where complete carbonation occurred after 28 days of AC curing and formed CC. It was reported that the band at approximately 950 cm −1 was assigned to Si-O, stretching the vibration of the Q 2 tetrahedron and indicating the silicate vibration regions of C-S-H [47][48][49]. The NC curing samples exhibited C-S-H peaks at 1645 cm −1 (Figure 8a), whereas this peak is absent in AC. In NC curing, both hydration and slightly carbonation (due to atmospheric CO 2 ) participated, while in the case of AC only, the carbonation reaction dominated. Thus, the strongest peak of CO 3 2− was observed in AC conditions. There was a peak shift in Si-O from 955 cm −1 (Figure 8a) to 1012 cm −1 (Figure 8b) after AC curing, being ascribed to the decalcification of C-S-H caused by the difference in the Ca:Si ratio, leading to the higher wave number shift [11].
vibration regions of C-S-H [47][48][49]. The NC curing samples exhibited C-S-H peaks at 1645 cm −1 (Figure 8a), whereas this peak is absent in AC. In NC curing, both hydration and slightly carbonation (due to atmospheric CO2) participated, while in the case of AC only, the carbonation reaction dominated. Thus, the strongest peak of CO3 2− was observed in AC conditions. There was a peak shift in Si-O from 955 cm −1 (Figure 8a) to 1012 cm −1 ( Figure  8b) after AC curing, being ascribed to the decalcification of C-S-H caused by the difference in the Ca:Si ratio, leading to the higher wave number shift [11].

TGA of Paste After Carbonation
The stability of the hydration and the carbonated products could be accessed by TGA, where mass loss occurred at a particular temperature range. The TGA results of the paste samples are shown in Figure 9. The peak between 50 °C and 200 °C can be ascribed to the evaporation of water and dehydration of the C-S-H, whereas, a range of 400 °C -500 °C for the decomposition of portlandite (Ca(OH)2) and the peak between 600 °C -800 °C has a relation to the decarbonization of poorly-crystallized and well-crystallized CaCO3 [50]. Figures 9a and b show the derivative thermogravimetric (DTG) curves of NC and AC samples after 28 days of curing, respectively. An interesting observation was found in NC (Figure 9a), where a mass loss peak at 460 °C attributed to the hydration reaction products of cement (i.e., portlandite), whereas this peak was absent in AC (Figure 9b). This result suggests that the NC sample was dominated by the hydration reaction, where the possibility of forming portlandite was significant. Alternatively, as the temperature increased, there was one peak observed from 650 °C to 750 °C in both conditions (i.e., NC and AC). However, the samples cured in AC, especially G5, exhibited a strong and sharp peak (Figure 9b), owing to the well-documented and stable crystalline CC [10].

TGA of Paste After Carbonation
The stability of the hydration and the carbonated products could be accessed by TGA, where mass loss occurred at a particular temperature range. The TGA results of the paste samples are shown in Figure 9. The peak between 50 • C and 200 • C can be ascribed to the evaporation of water and dehydration of the C-S-H, whereas, a range of 400 • C-500 • C for the decomposition of portlandite (Ca(OH) 2 ) and the peak between 600 • C-800 • C has a relation to the decarbonization of poorly-crystallized and well-crystallized CaCO 3 [50]. Figures 9a and 9b show the derivative thermogravimetric (DTG) curves of NC and AC samples after 28 days of curing, respectively. An interesting observation was found in NC (Figure 9a), where a mass loss peak at 460 • C attributed to the hydration reaction products of cement (i.e., portlandite), whereas this peak was absent in AC (Figure 9b). This result suggests that the NC sample was dominated by the hydration reaction, where the possibility of forming portlandite was significant. Alternatively, as the temperature increased, there was one peak observed from 650 • C to 750 • C in both conditions (i.e., NC and AC). However, the samples cured in AC, especially G5, exhibited a strong and sharp peak (Figure 9b), owing to the well-documented and stable crystalline CC [10]. (Figure 9a), where a mass loss peak at 460 °C attributed to the hydration reaction products of cement (i.e., portlandite), whereas this peak was absent in AC (Figure 9b). This result suggests that the NC sample was dominated by the hydration reaction, where the possibility of forming portlandite was significant. Alternatively, as the temperature increased, there was one peak observed from 650 °C to 750 °C in both conditions (i.e., NC and AC). However, the samples cured in AC, especially G5, exhibited a strong and sharp peak (Figure 9b), owing to the well-documented and stable crystalline CC [10].

Conclusions
In spite of the relatively low hydraulic properties of γ-C2S and a high CO2 uptake, it is expected to utilize CO2 in advanced construction materials. Cement production could be minimized by partial substitution with GGBFS and the addition of γ-C2S, which enhances the compressive strength. The mortar samples cured in AC condition with γ-C2S exhibited higher compressive strength compared with NC, attributed to the reaction of γ-C2S with CO2 and the formation CaCO3, which fill the pores of the mortar matrix. The SEM results show that the addition of 5% γ-C2S filled out the pores of the paste matrix, and that most of the γ-C2S reacted and formed CC in AC conditions. However, the unreacted GGBFS and γ-C2S were observed in NC and AC conditions after 28 days of curing by SEM. The unreacted GGBFS led to forming pores, resulting in lower compressive strength. The XRD results confirmed the formation of C-S-H, CC, and V in NC and AC curing. The G0 sample cured in NC mostly contained C-S-H, owing to the hydration reaction, whereas G5 and G10 contained CC and V, along with unreacted γ-C2S. However, once the samples were cured in AC conditions, the amount of CC and V increased in all samples, leading to high compressive strength. The intense peaks of CO3 2− in FT-IR revealed the carbonation reaction of AC-cured paste samples. The γ-C2S-incorporated samples cured in AC exhibited higher compressive strength, compact morphology with lower porosity, and intense CC peaks.

Conclusions
In spite of the relatively low hydraulic properties of γ-C 2 S and a high CO 2 uptake, it is expected to utilize CO 2 in advanced construction materials. Cement production could be minimized by partial substitution with GGBFS and the addition of γ-C 2 S, which enhances the compressive strength. The mortar samples cured in AC condition with γ-C 2 S exhibited higher compressive strength compared with NC, attributed to the reaction of γ-C 2 S with CO 2 and the formation CaCO 3 , which fill the pores of the mortar matrix. The SEM results show that the addition of 5% γ-C 2 S filled out the pores of the paste matrix, and that most of the γ-C 2 S reacted and formed CC in AC conditions. However, the unreacted GGBFS and γ-C 2 S were observed in NC and AC conditions after 28 days of curing by SEM. The unreacted GGBFS led to forming pores, resulting in lower compressive strength. The XRD results confirmed the formation of C-S-H, CC, and V in NC and AC curing. The G0 sample cured in NC mostly contained C-S-H, owing to the hydration reaction, whereas G5 and G10 contained CC and V, along with unreacted γ-C 2 S. However, once the samples were cured in AC conditions, the amount of CC and V increased in all samples, leading to high compressive strength. The intense peaks of CO 3 2− in FT-IR revealed the carbonation reaction of AC-cured paste samples. The γ-C 2 S-incorporated samples cured in AC exhibited higher compressive strength, compact morphology with lower porosity, and intense CC peaks.