Effect of Curing Conditions on the Hydration of MgO in Cement Paste Mixed with MgO Expansive Agent

Using the volume expansion generated by the hydration of the MgO expansive agent to compensate for the shrinkage deformation of concrete is considered to be an effective measure to prevent concrete shrinkage and cracking. Existing studies have mainly focused on the effect of the MgO expansive agent on the deformation of concrete under constant temperature conditions, but mass concrete in practical engineering experiences a temperature change process. Obviously, the experience obtained under constant temperature conditions makes it difficult to accurately guide the selection of the MgO expansive agent under actual engineering conditions. Based on the C50 concrete project, this paper mainly investigates the effect of curing conditions on the hydration of MgO in cement paste under actual variable temperature conditions by simulating the actual temperature change course of C50 concrete so as to provide a reference for the selection of the MgO expansive agent in engineering practice. The results show that temperature was the main factor affecting the hydration of MgO under variable temperature curing conditions, and the increase in the temperature could obviously promote the hydration of MgO in cement paste, while the change in the curing methods and cementitious system had an effect on the hydration of MgO, though this effect was not obvious.


Introduction
Although concrete has been used as a building material for a long time, some basic problems have not been completely solved. The shrinkage deformation of concrete leading to the decline of durability is one of the most important problems [1][2][3][4]. The shrinkage deformation of concrete is a non-external deformation caused by the combined action of physics and chemistry. In actual structures, this volume deformation is often limited by external constraints, resulting in stresses within the concrete [5]. The cracking of concrete not only reduces the mechanical properties of the structure and affects the beauty of the concrete building but also provides a convenient channel for harmful ions to enter the concrete interior, which could further affect the durability of the concrete and reduce the service life of the concrete structures [6]. When the durability of the concrete decreased seriously, a large amount of production materials was spent on either the maintenance or even the reconstruction of the concrete structure, which is contrary to the concept of energy conservation and the emissions reduction proposed today.
In order to reduce the negative effects caused by the shrinkage cracking of concrete, researchers have attempted to use a variety of anti-cracking measures in engineering practice and achieved better anti-cracking effects. The anti-cracking measures for concrete mainly were considered in the following three directions: (1) Reduce the shrinkage deformation generated by the concrete itself: For example, covering the surface of concrete in a timely manner after it has been formed and spraying water on the surface periodically as a means The raw materials used in this experiment included Portland cement, secondary fly ash, S95 slag powder and the MgO expansive agent (i.e., MEA). Among them, Portland cement was P·II52.5 cement produced by Jiangnan Onoda Cement Co., Ltd., Nanjing, China, and secondary fly ash and S95 mineral powder was provided by Nanjing Pudi Concrete Company, Nanjing, China. Four kinds of MgO expansive agents were provided by Wuhan Sanyuan Special Building Materials Company and Jiangsu Sobute Company. According to the citric acid method [31], the reactivity values of four kinds of MgO expansive agent were 120s, 180s, 240s and 330s. Additionally, they were named MEA-120, MEA-180, MEA-240 and MEA-330, respectively. Table 1 shows the chemical composition of Portland cement, secondary fly ash, S95 slag powder and the four kinds of magnesium oxide expansive agent used in the experiment. In this experiment, the simulation of the temperature change process of mass concrete was based on the actually measured temperature data of C50 mass concrete in engineering, and the curing box and external temperature acquisition module were used to realize the stage temperature change.
The temperature simulation process was as follows: the temperature change process of concrete was broken down into several small temperature change stages. Additionally, the temperature was adjusted every 4 h during the heating phase until it increased to the maximum temperature; the temperature was adjusted every 8 h during the cooling phase until it dropped to an ambient temperature. The heating and cooling rate of the temperature was determined according to the concrete temperature data measured in the engineering. Figure 1 shows the temperature variation process inside two groups of concrete walls that were measured in a C50 mass concrete engineering project. As can be seen from Figure 1, the internal temperature of concrete reached its maximum value around 24-48 h (taking its time when concrete began to be poured as the initial zero point), and the internal temperature of concrete was basically consistent with the ambient temperature at 14 d, i.e., the cooling process was basically over. The simulation of the temperature variation process in this study was based on the temperature variation in the data of mass concrete obtained from engineering practice (as shown in Figure 1), and Figure 2 shows the variable temperature curing environment with peak temperatures at 65 • C and 85 • C, which were simulated by a curing box based on the temperature change data obtained in engineering practice. (a) 65 °C (peak temperature) ( b) 85 °C (peak temperature)

Preparation and Curing of Cement Pastes
The tests involved a total of two cementitious systems, and the raw material for lation of the formed cement paste is shown in Table 2. The MEA content in the tw mentitious systems accounted for 8% of the total mass of the cementitious materials. cement paste test block used to test the hydration of MgO was molded in Φ25 mm mm columnar rigid PVC molds with a water-cement ratio of 0.32. Before forming, the materials were evenly mixed by a mixer, and the evenly mixed cement paste was put the test mold by a cement paste purifying mixer with water for mixing. Then, the test m was placed on a shaking table for 60 s to eliminate the air inside the paste. The prep cement paste specimens were maintained in the curing environment shown in Figure observe the hydration of MgO in the cement paste.

Preparation and Curing of Cement Pastes
The tests involved a total of two cementitious systems, and the raw material formulation of the formed cement paste is shown in Table 2. The MEA content in the two cementitious systems accounted for 8% of the total mass of the cementitious materials. The cement paste test block used to test the hydration of MgO was molded in Φ25 mm × 30 mm columnar rigid PVC molds with a water-cement ratio of 0.32. Before forming, the raw materials were evenly mixed by a mixer, and the evenly mixed cement paste was put into the test mold by a cement paste purifying mixer with water for mixing. Then, the test mold was placed on a shaking table for 60 s to eliminate the air inside the paste. The prepared cement paste specimens were maintained in the curing environment shown in Figure 2 to observe the hydration of MgO in the cement paste. The experiment involved two curing methods: water curing and non-wet curing. The prepared cement paste specimens were cured directly with a mold (the mold could be sealed with a cover), and the mold was wrapped with plastic wrap to isolate water so as

Preparation and Curing of Cement Pastes
The tests involved a total of two cementitious systems, and the raw material formulation of the formed cement paste is shown in Table 2. The MEA content in the two cementitious systems accounted for 8% of the total mass of the cementitious materials. The cement paste test block used to test the hydration of MgO was molded in Φ25 mm × 30 mm columnar rigid PVC molds with a water-cement ratio of 0.32. Before forming, the raw materials were evenly mixed by a mixer, and the evenly mixed cement paste was put into the test mold by a cement paste purifying mixer with water for mixing. Then, the test mold was placed on a shaking table for 60 s to eliminate the air inside the paste. The prepared cement paste specimens were maintained in the curing environment shown in Figure 2 to observe the hydration of MgO in the cement paste. The experiment involved two curing methods: water curing and non-wet curing. The prepared cement paste specimens were cured directly with a mold (the mold could be sealed with a cover), and the mold was wrapped with plastic wrap to isolate water so as to realize the not-wet curing. The specimens of the cement paste cured in water were removed from the mold and placed in water for curing after 20 h of curing with the mold. In addition, paste specimens with curing ages of 1 d, 3 d, 7 d and 14 d were selected to complete the characterization of MgO's hydration degree under the whole temperature course (take the preparation time of ready cement paste as the initial zero point).

Determination of the Remaining MgO Content in Cement Paste
The K-value method (XRD) was used to determine the content of MgO in the cement paste for hydration. ZnO was chosen as the internal standard substance (the dosage is 2 wt.%). The scanning range was 35 • -45 • , and the scanning speed was 1 • /min. ZnO and MgO powders were, respectively, weighed according to the mass ratio of 1:1, and the powders were ground with an agate mortar to obtain a homogeneous mixed powder sample. The milled powder sample was subjected to X-ray diffraction analysis, and the integrated area of the strongest diffraction peaks of ZnO and MgO in the diffraction pattern was calculated using MDI Jade software. The ratio of the integrated area of the strongest diffraction peaks of ZnO and MgO was the K value, which was calculated to be 0.57. The position of the strongest diffraction peak of ZnO was d 101 = 2.47, and the position of the strongest diffraction peak of MgO was d 200 = 2.11. The remaining content of MgO in the cement paste was calculated by Equation (1): where ω MgO is the remaining content of MgO in the cement paste. ω ZnO is the doping amount of the internal standard substance ZnO. I ZnO is the integrated intensity of the strongest diffraction peak of the internal standard substance ZnO, and I MgO is the integrated intensity of the strongest diffraction peak of MgO. K is the characteristic constant (K = 0.57).

Effect of Curing Temperature on MgO Hydration in Cement Paste
In this section, the effect of the curing temperature on MgO hydration in cement paste was studied. The cement paste was prepared according to the raw material ratio of "system a" in Table 2. Figure 3 shows the hydration process of MgO when mixed with four types of MEA cement paste under a 65 • C variable temperature with water curing conditions. It can be seen from Figure Figure 4 shows the hydration process of MgO mixed with four MEA cement pastes under 85 °C variable temperature water curing conditions. As can be seen from Figure 4, the trend in the change in the MgO content in cement paste mixed with four kinds of MEA at 65 °C and 85 °C and under a variable temperature water curing condition was basically similar. The hydration rate of MgO in the cement paste was the fastest at 1d, and then the hydration rate started to decrease gradually. However, compared to the 65 °C variable temperature water curing condition, the residual content of MgO in the cement paste under an 85 °C variable temperature water curing condition at a curing age of 1d was significantly reduced. Moreover, the reaction rate of MgO in the cement paste became slower during the cooling stage, and the content of MgO in the cement paste of multiple MEAdoped groups appeared as a "plateau period"., i.e., the content of MgO remained essentially unchanged.   Figure 4 shows the hydration process of MgO mixed with four MEA cement pastes under 85 • C variable temperature water curing conditions. As can be seen from Figure 4, the trend in the change in the MgO content in cement paste mixed with four kinds of MEA at 65 • C and 85 • C and under a variable temperature water curing condition was basically similar. The hydration rate of MgO in the cement paste was the fastest at 1 d, and then the hydration rate started to decrease gradually. However, compared to the 65 • C variable temperature water curing condition, the residual content of MgO in the cement paste under an 85 • C variable temperature water curing condition at a curing age of 1 d was significantly reduced. Moreover, the reaction rate of MgO in the cement paste became slower during the cooling stage, and the content of MgO in the cement paste of multiple MEA-doped groups appeared as a "plateau period"., i.e., the content of MgO remained essentially unchanged.   Figure 4 shows the hydration process of MgO mixed with four MEA cement pastes under 85 °C variable temperature water curing conditions. As can be seen from Figure 4, the trend in the change in the MgO content in cement paste mixed with four kinds of MEA at 65 °C and 85 °C and under a variable temperature water curing condition was basically similar. The hydration rate of MgO in the cement paste was the fastest at 1d, and then the hydration rate started to decrease gradually. However, compared to the 65 °C variable temperature water curing condition, the residual content of MgO in the cement paste under an 85 °C variable temperature water curing condition at a curing age of 1d was significantly reduced. Moreover, the reaction rate of MgO in the cement paste became slower during the cooling stage, and the content of MgO in the cement paste of multiple MEAdoped groups appeared as a "plateau period"., i.e., the content of MgO remained essentially unchanged.    water curing condition showed a "plateau period" during the cooling stage (3-14 d). This was not significantly observed for other cement pastes when mixed with other active MEAs under the same curing condition. In comparison, it was found that the MgO content in the cement paste of these four groups started to show signs of a "plateau" mostly at the age of 3 d (i.e., the cooling stage). At this time, the MgO content in the cement paste was mixed with MEA-120 under a 65 • C variable temperature water curing condition, MEA-120, MEA-180 and MEA-240 was under an 85 • C variable temperature water curing condition and were 2.33 wt.%, 2.01 wt.%, 2.17 wt.% and 2.38 wt.%, respectively. The other four groups of cement pastes without this situation had a higher MgO content at a curing age of 3 d compared to the MgO content values at the time of plateauing. In addition, it could also be found that the increase in the curing temperature had a slightly different effect on the promotion of the hydration reaction of different active MEAs. The low activity of MEA seemed to be more sensitive to the temperature, and the increase in the temperature was more effective when promoting the hydration of low-activity MEA. This may also be due to the rapid reaction of the highly reactive MEA, resulting in a premature approach to the "plateau content" of MgO in the cement paste mixed with a highly reactive MEA. and MEA-120, MEA-180 and MEA-240 under an 85 °C variable temperature water curing condition showed a "plateau period" during the cooling stage (3-14d). This was not significantly observed for other cement pastes when mixed with other active MEAs under the same curing condition. In comparison, it was found that the MgO content in the cement paste of these four groups started to show signs of a "plateau" mostly at the age of 3d (i.e., the cooling stage). At this time, the MgO content in the cement paste was mixed with MEA-120 under a 65 °C variable temperature water curing condition, MEA-120, MEA-180 and MEA-240 was under an 85 °C variable temperature water curing condition and were 2.33 wt.%, 2.01 wt.%, 2.17 wt.% and 2.38 wt.%, respectively. The other four groups of cement pastes without this situation had a higher MgO content at a curing age of 3d compared to the MgO content values at the time of plateauing. In addition, it could also be found that the increase in the curing temperature had a slightly different effect on the promotion of the hydration reaction of different active MEAs. The low activity of MEA seemed to be more sensitive to the temperature, and the increase in the temperature was more effective when promoting the hydration of low-activity MEA. This may also be due to the rapid reaction of the highly reactive MEA, resulting in a premature approach to the "plateau content" of MgO in the cement paste mixed with a highly reactive MEA.

Effect of Curing Methods on the Hydration of MgO in Cement Paste
In order to clarify the effect of water curing and non-wet curing on MgO hydration in the cement paste under variable temperature conditions, the hydration of MgO in the cement paste was compared under two curing methods according to "system a" forming cement paste. Figure 6 shows the comparison of MgO residual content in cement paste mixed with the same active MEA by different curing methods under variable temperature conditions of 65 • C and 85 • C. In order to clarify the effect of water curing and non-wet curing on MgO hydration in the cement paste under variable temperature conditions, the hydration of MgO in the cement paste was compared under two curing methods according to "system a" forming cement paste. Figure 6 shows the comparison of MgO residual content in cement paste mixed with the same active MEA by different curing methods under variable temperature conditions of 65 °C and 85 °C. As can be seen in Figure 6, for cement paste mixed with the same active MEA curing in the same variable temperature environment, the effect of non-wet curing and water curing on the hydration of MgO in the cement paste was not as great as the effect of temperature on the hydration of MgO. The effect of moisture on MgO hydration in the MEA mixed cement paste was mainly reflected in the period before the curing age of 3d, and the promotion effect of moisture on MgO hydration in the cement paste in the subsequent process seemed not to be not obvious. It can be seen from the changing curve of MgO content in the cement paste underwater with non-wet curing at a 65 °C variable temperature that the difference between non-wet and water curing methods on MgO hydration was very small at the curing age of 1-3d, and the remaining MgO content in the cement paste under the two curing methods of MEA-120, MEA-180, MEA-240 and MEA-330 was basically the same. In addition, in the subsequent cooling process, no significant difference was observed in the content of MgO on the paste under two curing conditions, even with As can be seen in Figure 6, for cement paste mixed with the same active MEA curing in the same variable temperature environment, the effect of non-wet curing and water curing on the hydration of MgO in the cement paste was not as great as the effect of temperature on the hydration of MgO. The effect of moisture on MgO hydration in the MEA mixed cement paste was mainly reflected in the period before the curing age of 3 d, and the promotion effect of moisture on MgO hydration in the cement paste in the subsequent process seemed not to be not obvious. It can be seen from the changing curve of MgO content in the cement paste underwater with non-wet curing at a 65 • C variable temperature that the difference between non-wet and water curing methods on MgO hydration was very small at the curing age of 1-3 d, and the remaining MgO content in the cement paste under the two curing methods of MEA-120, MEA-180, MEA-240 and MEA-330 was basically the same. In addition, in the subsequent cooling process, no significant difference was observed in the content of MgO on the paste under two curing conditions, even with the different curing methods. As can be seen from the variation curves of the content of MgO in cement paste under 85 • C variable temperature conditions for water and non-wet curing, compared with non-wet curing, the remaining content of MgO in the cement paste mixed with MEA under water curing significantly decreased at this stage before the curing age of 3 d, where no significant difference was found in the change in the MgO content in the paste under the two curing methods of water and non-wet curing in the cooling stage after 3 d.
When comparing the hydration process of MgO in the cement paste in Figure 6c, it was found that the content of MgO in the cement paste mixed with MEA was significantly reduced when the temperature peak of variable temperature curing increased from 65 • C to 85 • C under the same non-wet condition. Compared to 65 • C variable temperature non-wet curing, the residual content of MgO in the cement paste at 85 • C variable temperature non-wet curing ages of 1 d, 3 d, 7 d and 14 d decreased by 0.50 wt.%, 0.69 wt.%, 0.76 wt.% and 0.21 wt.%, respectively. However, under the same 65 • C variable temperature curing condition, when the curing method changed from non-wet curing to water curing, there was no significant difference in the content of MgO in the cement paste when mixed with MEA-240 during the whole age. In addition, it could also be found that the MgO content in the cement paste mixed with MEA-240 was basically the same in this stage before the maintenance age of 3 d under the condition of water curing and non-wet curing at a 65 • C variable temperature curing. However, when the temperature peak of variable temperature curing from 65 • C increased to 85 • C, compared with the non-wet curing, the content of MgO in the water-curing cement paste mixed with MEA-240 decreased significantly by 0.35 wt.% at the curing age of 3 d. The variation in the MgO content in the cement paste when mixed with other active MEA in Figure 6 was also similar to the results shown in Figure 6c.

Effect of Cementitious System on MgO in Cement Paste
In order to clarify the effect of the cementitious system on MgO hydration in the cement paste under variable temperatures and non-wet curing conditions, the hydration of MgO in the cement paste was formed according to the cementitious system a and b was compared. Figure 7 shows the comparison of the MgO residual content in cement paste mixed with the same active MEA under the condition of variable temperature and non-wet curing at 65 • C and 85 • C.
As can be seen from Figure 7, for the cement paste formed according to the cementing systems a and b and mixed with the same active MEA, there was little difference in the variation in the MgO content in the cement paste under the same curing condition. The content of MgO in cement paste formed according to cementing system b was lower than that of the cement paste formed according to system a at all ages. This may be caused by the addition of an 8 wt.% S95 slag powder in system a to replace part of the cement without changing the yield of fly ash. According to Figure 7d, under variable temperature and nonwet curing at 65 • C, the residual content of MgO in the cement paste was formed according to the cementitious system a and was 0.12 wt.%, 0.14 wt.%, 0.10 wt.% and 0.04 wt.% higher than that in the cementitious system b at the age of 1 d, 3 d, 7 d and 14 d, respectively. Therefore, it can be considered that the addition of the 8 wt.% slag powder with the same fly ash content could inhibit the hydration of MgO in the cement paste; however, the inhibition effect was not obvious.