Compensation for shrinkages with three kinds of lightly burnt MgO expansive agent (LBMEA) is used in a reinforced concrete wall poured in the summer. Influences of the internal temperature history on the expansion of concrete and the microstructure of cement paste containing LBMEA were investigated. The results showed that LBMEA exhibited significant expansion around the end of the fall temperature stage; then, the expansion rate declined obviously, and concrete containing LBMEA with low hydration reactivity (140 s and 220 s) showed larger expansion than LBMEA with high hydration reactivity (60 s). Microstructural analysis indicated that brucite preferentially forms in the pores in cement paste containing LBMEA with high reactivity, but brucite mainly grows on the surface of the MgO particles in cement paste containing LBMEA with low reactivity during the early age. Paste containing LBMEA with low reactivity showed a larger volume of single brucite crystal than LBMEA with high reactivity, which further led to larger expansion in the latter than the former. The results revealed the expansion process of LBMEA and can help engineers select suitable LBMEA for application to actual engineering.

1. Introduction

Cementitious material usually undergoes volume shrinkages because of chemical reaction, moisture, and temperature exchange [1, 2]. Stress occurs in concrete structures when the volume shrinkages are restrained. If concrete does not develop enough tensile strength to withstand this stress, cracks are generated, and then ions are diffused into the inner concrete, which accelerates obviously the deterioration of concrete [3]. For the early age reinforced concrete members, the autogenous shrinkage, thermal shrinkage, and dry shrinkage play an important role in cracking problems. Mitigating those shrinkages of cementitious material through several traditional methods is complicated and has a limited effect, and it can even lead to some negative influences on concrete performance [46]. Using expansive agents is an effective measure to compensate for volume shrinkages [7]. However, those traditional expansive agents depended mightily on the curing condition and expansion mainly at a very early age, usually being within 14 days, leading to the problem that some of the shrinkages generated at a later age might not be compensated effectively [7, 8]. Compared with those expansive agents, an MgO expansive agent (MEA) has significant advantages, including the chemically stable hydration production, namely, Mg(OH)2, a relatively low water requirement, and designable expansion characteristics [912].

The hydration reactivity of MEA was controlled by adjusting the calcination conditions that determine its microstructure [9]. When calcined under a higher temperature and longer residence time, resulting in less hydration reactivity and a higher reactivity value, the MEA generated slower and less expansion at an early age but more rapid and a large expansion at a later age, which has been proved effective in compensating for the late shrinkage of mass concretes, such as dams and diversion tunnels [13, 14]. When calcined under a lower temperature and shorter residence time, namely, lightly burnt MgO, the MEA produced a faster expansion rate at a relatively early age, which may be suitable for industrial and civil building [15, 16].

The expansive property of MEA is not only influenced by its hydration reactivity but also depends strongly on the curing temperature [1719]. Many studies focused on the effect of the constant curing temperature on MEA in the laboratory. However, the temperature field of concrete in the actual structure changes with age and ambient temperature, and the expansive characteristics of MEA cannot be forecasted accurately by the results of constant temperature. Although studies have been focused on the expansive deformation and stress of cementitious material containing MEA under fluctuant temperature in the laboratory [10, 20], there has been no research about the expansive properties and microstructural diversities of cementitious materials with different kinds of MEA curing in actual structures. Understanding correctly the expansion and hydration characteristics of different kinds of MEA under variable temperature conditions helps engineers design or select the proper kind of MEA for a specific environmental condition and concrete structures, to reduce or even avoid the occurrence of mismatch between the expansion caused by MEA and shrinkages of cementitious materials.

In this research, three kinds of LBMEA with different reactivity values were used to compensate for shrinkages of the reinforced concrete walls poured in the summer. The expansive deformation of LBMEA under the temperature history of the reinforced concrete walls was measured. The microstructures of the cement pastes containing LBMEA embedded in field concrete were also investigated to explicate the expansive process and diversity of LBMEA with different reactivity values.

2. Materials and Methods

2.1. Materials

In this project, three kinds of industry-made LBMEA with different reactivity values of 60 s, 140 s, and 220 s tested according to a standard neutralization method were studied [9], designated as MEA60, MEA140, and MEA220. The chemical compositions and particle-size distributions of LBMEA are shown in Table 1 and Figure 1, respectively. The mean particle sizes of MEA60, MEA140, and MEA220 were 14.073 μm, 13.080 μm, and 16.450 μm, respectively. There was almost no difference in the particle-size distribution among three kinds of LBMEA.

P.O. 42.5 Portland cement, class I fly ash, and class S95 slag were used as cementitious material. Natural sand with a fineness modulus of 2.8 and crushed basalt with a diameter ranging from 5 to 30 mm were used as an aggregate. The mix proportions of field concrete are shown in Table 2, in which the dosages of LBMEA are 0% and 8% (weight percentage of cementitious materials). In general, reaction rate of fly and slag is much slower compared to cement hydration, reducing the 8% dosage of slag, and fly ash may increase slightly the shrinkage of cementitious material when compared to reference concrete (REF) [21].

The mix proportions of three paste samples for microstructural investigation are shown in Table 3. While workers began to pour field concrete, cement paste was prepared in a laboratory close to the field. LBMEA was first mixed into cementitious material homogeneously; then, water was added into the mixture and mixed for 4 min to obtain consistent mixtures. The consistent mixtures were cast into 20 × 20 × 20 mm molds and vibrated for 1 min. Molds containing cement paste were separately put into PVC pipes preinserted into a wood formwork, and then one side of the PVC pipes was sealed. When the field concrete approached initial setting, the PVC pipes were drawn out from the reinforced concrete wall, and then the molds were taken out from the PVC pipes and demolded, pastes were stuffed into holes, and finally it was sealed by grout and sealant.

2.2. Methods
2.2.1. Deformation of Field Concrete

A Type-VWS strain transducer was used to monitor the deformation of field concrete. The vibration of the strain transducer, F, is closely related to the length of the strain transducer, varying with the length change of the transducer brought about by the deformation of concrete. The temperature of the field concrete was measured by a thermocouple included in the transducer. The strain of the specimens was calculated by the following equation:where is a constant of the strain transducer; is the change of vibrating modulus; is the temperature correction coefficient of the strain transducer; is the thermal dilation coefficient of concrete; and is the temperature variation.

2.2.2. Embedded Instruments Inside of Reinforced Concrete Walls in Test Section

The reinforced concrete walls were 1.85 m high, 0.8 m thick, and 3 m long. Every wall was built on the concrete foundation. Each concrete foundation was 0.2 m high, 1.2 m thick, and 3 m long. Strain transducers were set inside of the wall, as shown in Figure 2. Figure 3 shows that two strain gauges were located in the same thickness plane (0.4 m) and altitude plane (1.2 m), and the distances from the edges were, respectively, 0.5 m (site 2) and 1.5 m (site 1).

2.2.3. Field Concrete Pouring

The ambient temperature was 35°C when the concrete was poured into the moldboard. Four reinforced concrete walls were poured continuously by workers. The whole casting time was performed within two hours to ensure that there were almost no differences in the temperature histories on the same site.

2.2.4. Microstructure Characterization

The paste samples used for the microstructural investigation were embedded in the concrete walls. Three paste samples with different kinds of MEAs were removed at prespecified times (28 days and 90 days), clamped into pieces, and then soaked in alcohol for 24 h and dried at 50°C for 12 h and tested. The morphologies of three paste samples were investigated by a scanning electron microscope (SEM) coupled with energy dispersive X-ray (EDX). In addition, three paste samples were sliced, dried, epoxy impregnated, and polished for investigating with backscattered electronic microscopy (BSEM). The porosity of pastes ranging from 7 nm to 200 μm was examined by mercury intrusion porosimetry (MIP). 2–4 pieces with a sample size of 2 mm were used for the MIP test for each paste sample.

3. Results and Discussion

3.1. Temperature History of Field Concrete in Test Section

Figure 4 shows the temperature histories of the field concrete in two sites. The maximum temperature occurred at approximately 1 day, and then temperature dropped from 1 day to 10 days. Then, the internal temperature changed along with the ambient temperature. In the same site, the temperature history showed almost no difference whether LBMEA was added to the field concrete or not, which indicated there was less effect on temperature history when LBMEA with dosages of 8% (by mass as substitutions of fly ash and slag) was added into the concrete. Taking the temperature history of reference concrete (REF) as an example, the average temperature of concrete in site 1 and site 2 was 50.2°C and 45.0°C in the first 10 days, while the average temperature was 31.6°C and 31.7°C from 10 to 90 days, respectively. The differences of temperature histories between site 1 and site 2 were mainly concentrated in the first 10 days.

3.2. Deformations of Field Concrete in Test Section

Figure 5 displays the self-deformation of field concrete poured in summer. There was one inflexion point occurring at an age of around 10 days for each concrete. Two different stages were separated by the inflexion point. All the concretes containing LBMEA with different hydration reactivity showed significant expansion to compensate the drastic shrinkages of concrete because of the relatively high internal temperature in the first 10 days; then, expansive deformation showed a trend of slowing down and even turned into shrinkage at a later stage. This phenomenon may result from the growth of the strength of concrete during continuous stiffening process and the drop of temperature and humidity from 10 to 90 days, which hindered the expansive deformation of concrete containing LBMEA and decreased obviously the hydration rate of LBMEA. Owing to the insufficient hydration of LBMEA under the condition of low temperature and humidity at a later age, expansion caused by the hydration of LBMEA had difficulty in compensating for shrinkage of the concrete.

Figure 6 shows that the expansive deformation caused by LBMEA with different hydration reactivity under the internal condition of reinforced concrete wall. At the first stage (0–10 days), the expansive deformations were different among the three kinds of field concrete containing LBMEA, which depended on the hydration activity of LBMEA and temperature history. Concrete containing MEA140 and MEA220 showed the larger expansion than that containing MEA60, which indicated that LBMEA with low hydration reactivity showed a larger expansive deformation than LBMEA with a high hydration reactivity when cured at a high temperature. The above results indicate that the hydration activity of LBMEA should be properly selected according to the specific temperature history.

At the second stage (10–90 days), concrete containing MEA60 showed a small expansion of 35 με but concrete containing MEA140 and MEA220 exhibited more considerable expansions of 62 με and 78 με in site 2. The hydration of MEA60 tends to stop from 45 to 90 days. In comparison with the concrete in site 2, the expansive deformation caused by MEA140 and MEA220 in site 1 decreased by 15 με and 63 με, respectively, and a tendency was obviously observed that expansive deformation nearly stopped in site 1 but expansive deformation increased continuously in site 2 in the second stage, as shown in Figure 6. This result may indicate that expansive deformation generated by LBMEA in a later stage was affected by the hydration of LBMEA in the early stage. When the temperature history of concrete in site 1 was compared with that of site 2, the average temperatures of concrete were similar in the second stage, but the average temperature of concrete in site 1 was 5°C higher than that at site 2 in the first stage. The hydration degree of LBMEA was accelerated by elevated temperature in the first stage, which may lead to the insufficient hydration of LBMEA in the later stage.

3.3. Porosity of Cement Paste Embedded in Reinforced Concrete Wall

Figure 7 displays the pore structure of the cement pastes embedded in the reinforced concrete wall. As shown in Figure 7(a), most of the pores of cement pastes made with LBMEA had a pore diameter range of 0.007–0.1 μm at 28 days, and sample P3 had more pores with a size range of 0.1–0.3 μm than the others. When the curing time increased from 28 to 90 days (Figure 7(b)), the number of porosities from 0.007–0.1 μm for all samples decreased, and the main porosities of samples were 0.007–0.05 μm. The result may be related to the hydraulic activity of fly ash and slag at the later age. The gap and microvoids of cement matrix were filled with reaction products generated by a pozzolanic reaction [21, 22], which refined the pore structure of cement paste when the curing time increased from 28 to 90 days.

Figure 7(c) shows the cumulative volume of the pores of cement pastes. The specimen with a lower reactivity of LBMEA exhibited a larger total porosity at the curing time of 28 days. In general, MEAs with lower hydration activity have a denser pore structure [11], and this phenomenon may be inferred from Figure 5. MgO particles were surrounded by the hydration products of cementitious material, MgO reacted with water, and then Mg(OH)2 formed. When crystals formed and grew in a restricted area, the volume expansion occurred. The larger expansive deformation generated by MEA with low reactivity leads to microcracks of the cement matrix, which may lead to a larger total porosity in samples P2 and P3. The total porosity decreased when the curing time increased from 28 to 90 days for all samples, and the decreases were 6.8%, 27.9%, and 22.3% for samples P1, P2, and P3, respectively. The paste with a lower reactivity tended to show a higher decrement in porosity.

3.4. Morphology of Cement Paste Embedded in Reinforced Concrete Wall

Figure 8 shows typical BSE images of the hydrated cement pastes embedded in the concrete wall. There was an obvious boundary between the MEA and cement matrix. Hydrated MEAs were surrounded by the hydration product of cementitious materials. There was a denser structure in MEA with low hydration reactivity, which indicated that less space could be occupied when MgO began to react with water. As shown in Figures 8(b) and 8(c), cracks occurred in the paste containing MEA140 and MEA220. Some cracks appeared at the cement matrix and others formed in the MEA particle.

Figure 9 displays the morphologies of cement pastes embedded in the reinforced concrete wall. The shapes of the brucite crystals were similar and showed a rod-like morphology, but there was a discrepancy in the locations of the brucite crystals among three samples. As shown in Figure 9(a), MEA60 exhibited a faster hydration rate because of more crystal defects and the specific surface area [9]; more MgO was dissolved, and then Mg2+ was diffused into the pore near the MgO grain under a relatively high temperature. With the increase of concentration of Mg2+ and OH in the pore solution, some brucite crystals formed and grew in the pore, and this part of the brucite crystals did not contribute to the macroscopic expansion. In Figures 9(c) and 9(e), brucite crystals mainly formed on the surface of the MgO grain and overlapped each other with the growth of Mg(OH)2 to generate pressure and then expanded the surrounding cement matrix, which led to more effective expansive deformation. In addition to the difference in the forming location of the brucite crystals, the volume of a single brucite crystal in sample P3 was larger than that in sample P1, which further increased the expansive deformation of cementitous materials with low hydration reactivity of LBMEA.

3.5. Discussion

The difference and expansive process of the volume expansion of paste with LBMEA could be explicated from the microstructural analysis. During the early age, high temperature accelerated the hydration rate of MgO and, therefore, accelerated the expansion process. MEA60 showed a faster hydration rate and generated more brucite to fill the interior pores of MgO particles or microvoids inside the cement matrix because of the diffusion of Mg2+ at a high temperature, which led to a lower total porosity and macroscopic expansion. MEA140 and MEA220 had a denser pore structure of the MEA particle, and the single brucite crystal exhibited the larger crystal volume. More brucites were distributed at the location of the MgO surface and confined zone nearby MgO particle. When the growing pressure of brucite exceeds the tensile strength of the paste at an early age, microcracks form and volume expansion occur, increasing the total porosity and volume deformation to some extent. Sustained hydration of LBMEA and pozzolanic hydration of fly ash and slag decreased the total porosity from 28 to 90 days. Sample containing low reactivity of LBMEA showed a high decrease of total porosity indicating a high hydration degree of LBMEA with low reactivity from 28 to 90 days, leading to a phenomenon in which expansive deformation increased continuously. This matched well with the volume deformation caused by LBMEA, as shown in Figure 6(b).

4. Conclusions

Three kinds of LBMEA with different hydration reactivity were, respectively, added into the reinforced concrete wall poured in the summer. The main conclusions drawn are as follows:(1)In this project, three kinds of LBMEA showed significant expansion during the first 10 days, and then it slowed down obviously and even stopped, which may be ascribed to the fall of temperature in the later stage and inhibitory effect caused by the high hydration degree of LBMEA caused by a high temperature in the early age. The results show that using LBMEA compensated effectively for early shrinkages of field concrete but did not have the ideal compensation effect on shrinkages at a later age.(2)Field concrete containing MEA140 showed the largest expansion among three kinds of LBMEA during the first 10 days. LBMEA with a low hydration reactivity compensated more effectively for shrinkages of the reinforced concrete wall. Concrete containing MEA140 was ranked the best in cracking resistance, which indicated that MEA with a low hydration reactivity was ranked better in cracking resistance than MEA with a high hydration reactivity when LBMEA was applied to field concrete poured at a high temperature.(3)SEM analysis showed that the growth of brucite crystals occurred mainly on the surface of MgO particles in MEA140 and MEA220, but brucite crystals grow better in the microvoids in MEA60, and the volume of a single brucite crystal in the latter was smaller than that in the former. The sample containing MEA60 showed the smallest total porosity at 28 days, which indicated that the more brucite crystals filled into voids than for MEA140 and MEA220. The microstructural analysis explicated that larger brucite crystals and more effective expansion lead to a larger macroscopic expansion of concrete containing MEA with low reactivity than that containing MEA with high reactivity. Compared with the sample containing MEA60 from 28 to 90 days, total porosity of the samples containing MEA140 and MEA220 decreased obviously.

Data Availability

All data generated or analysed in this research were included in this published article. And readers can access all data used to support conclusions of the current study from the corresponding author upon request.

Conflicts of Interest

The authors declare no conflicts of interest.

Authors’ Contributions

Lanqing Yu designed and conducted the experimental program. Min Deng provided the project. Liwu Mo gave many suggestions about experiment. Feifei Jiang and Jinxin Liu helped in conducting experiment. All authors contributed to analysis and conclusion.


This research was financially supported by the National Key Research and Development Plan of China (2017YFB0309903-01), the Science and Technology Plan Project of Shandong Province Communications (2018B37-02), and the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD).