Effects of Lightly Burnt MgO Expansive Agent on the Deformation and Microstructure of Reinforced Concrete Wall

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. +e results showed that LBMEA exhibited significant expansion around the end of the fall temperature stage; then, the expansion rate declined obviously, and concrete containing LBMEAwith 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 LBMEAwith 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. +e results revealed the expansion process of LBMEA and can help engineers select suitable LBMEA for application to actual engineering.


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 [4][5][6].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 [9][10][11][12].
e 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].
e expansive property of MEA is not only influenced by its hydration reactivity but also depends strongly on the curing temperature [17][18][19].Many studies focused on the e ect of the constant curing temperature on MEA in the laboratory.However, the temperature eld 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 uctuant temperature in the laboratory [10,20], there has been no research about the expansive properties and microstructural diversities of cementitious materials with di erent kinds of MEA curing in actual structures.Understanding correctly the expansion and hydration characteristics of di erent kinds of MEA under variable temperature conditions helps engineers design or select the proper kind of MEA for a speci c 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 di erent reactivity values were used to compensate for shrinkages of the reinforced concrete walls poured in the summer.e expansive deformation of LBMEA under the temperature history of the reinforced concrete walls was measured.e microstructures of the cement pastes containing LBMEA embedded in eld concrete were also investigated to explicate the expansive process and diversity of LBMEA with di erent reactivity values.

Materials.
In this project, three kinds of industry-made LBMEA with di erent 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.e chemical compositions and particle-size distributions of LBMEA are shown in Table 1 and Figure 1, respectively.e mean particle sizes of MEA60, MEA140, and MEA220 were 14.073 μm, 13.080 μm, and 16.450 μm, respectively.ere was almost no di erence in the particle-size distribution among three kinds of LBMEA.P.O.42.5 Portland cement, class I y ash, and class S95 slag were used as cementitious material.Natural sand with a neness modulus of 2.8 and crushed basalt with a diameter ranging from 5 to 30 mm were used as an aggregate.e mix proportions of eld 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 y and slag is much slower compared to cement hydration, reducing the 8% dosage of slag, and y ash may increase slightly the shrinkage of cementitious material when compared to reference concrete (REF) [21].
e mix proportions of three paste samples for microstructural investigation are shown in Table 3.While workers began to pour eld concrete, cement paste was prepared in a laboratory close to the eld.LBMEA was rst mixed into cementitious material homogeneously; then, water was added into the mixture and mixed for 4 min to obtain consistent mixtures.e 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 eld 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 stu ed into holes, and nally it was sealed by grout and sealant.

Deformation of Field Concrete.
A Type-VWS strain transducer was used to monitor the deformation of eld concrete.e 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.
e temperature of the eld concrete was measured by a thermocouple included in the transducer.e strain of the specimens was calculated by the following equation: where K is a constant of the strain transducer; ΔF is the change of vibrating modulus; b is the temperature correction coe cient of the strain transducer; α is the thermal dilation coe cient of concrete; and ΔT is the temperature variation.

Embedded Instruments Inside of Reinforced Concrete
Walls in Test Section.e reinforced concrete walls were  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).

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

Microstructure Characterization.
e paste samples used for the microstructural investigation were embedded in the concrete walls.ree paste samples with di erent kinds of MEAs were removed at prespeci ed 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.e 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).e 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.

Temperature History of Field Concrete in Test Section.
Figure 4 shows the temperature histories of the eld concrete in two sites.
e maximum temperature occurred at approximately 1 day, and then temperature dropped from 1 day to 10 days.en, the internal temperature changed along with the ambient temperature.In the same site, the temperature history showed almost no di erence whether LBMEA was added to the eld concrete or not, which indicated there was less e ect on temperature history when LBMEA with dosages of 8% (by mass as substitutions of y 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 rst 10 days, while the average temperature was 31.6 °C and 31.7 °C from 10 to 90 days, respectively.e di erences of temperature histories between site 1 and site 2 were mainly concentrated in the rst 10 days.

Deformations of Field Concrete in Test Section.
Figure 5 displays the self-deformation of eld concrete poured in summer.ere was one in exion point occurring at an age of around 10 days for each concrete.Two di erent stages were separated by the in exion point.All the concretes containing LBMEA with di erent hydration reactivity showed signi cant expansion to compensate the drastic shrinkages of concrete because of the relatively high internal temperature in the rst 10 days; then, expansive deformation showed a trend of slowing down and even turned into shrinkage at a later stage.is phenomenon may result from the growth of the strength of concrete during continuous sti ening 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 insu cient hydration of LBMEA under the condition of low temperature and humidity at a later age, expansion caused by the hydration of LBMEA had di culty in compensating for shrinkage of the concrete.
Figure 6 shows that the expansive deformation caused by LBMEA with di erent hydration reactivity under the internal condition of reinforced concrete wall.At the rst stage (0-10 days), the expansive deformations were di erent among the three kinds of eld 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.e above results indicate that the hydration activity of LBMEA should be properly selected according to the speci c 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. e hydration of MEA60 tends to stop from 45 to 90 days.In comparison 4 Advances in Materials Science and Engineering 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. is result may indicate that expansive deformation generated by LBMEA in a later stage was a ected 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 rst stage.e hydration degree of LBMEA was accelerated by elevated temperature in the rst stage, which may lead to the insu cient hydration of LBMEA in the later stage.

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 0.1-0.3μm 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 0.007-0.05μm. e result may be related to the hydraulic activity of y ash and slag at the later age.
e gap and microvoids of cement matrix were lled with reaction products generated by a pozzolanic reaction [21,22], which re ned 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.
e 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.e 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. e 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.e paste with a lower reactivity tended to show a higher decrement in porosity.

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.
ere was an obvious boundary between the MEA and cement matrix.Hydrated MEAs were surrounded by the hydration product of cementitious materials.ere 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.e 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 speci c surface area [9]; more

Advances in Materials Science and Engineering
MgO was dissolved, and then Mg 2+ was di used into the pore near the MgO grain under a relatively high temperature.With the increase of concentration of Mg 2+ 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 e ective expansive deformation.In addition to the di erence 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.

Discussion.
e di erence 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 ll the interior pores of MgO particles or microvoids inside the cement matrix because of the di usion of Mg 2+ 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 con ned 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,

Conclusions
ree kinds of LBMEA with di erent hydration reactivity were, respectively, added into the reinforced concrete wall poured in the summer.e main conclusions drawn are as follows: (1) In this project, three kinds of LBMEA showed signi cant expansion during the rst 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 e ect caused by the high hydration degree of LBMEA caused by a high temperature in the early age.e results show that using LBMEA compensated e ectively for early shrinkages of eld concrete but did not have the ideal compensation e ect on shrinkages at a later age.(2) Field concrete containing MEA140 showed the largest expansion among three kinds of LBMEA during the rst 10 days.LBMEA with a low hydration reactivity compensated more e ectively 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 eld 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.e sample containing MEA60 showed the smallest total porosity at 28 days, which indicated that the more brucite crystals lled into voids than for MEA140 and MEA220.
e microstructural analysis explicated that larger brucite crystals and more e ective 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.

Figure 4 :Figure 5 :
Figure 4: Temperature histories of eld concrete in two sites.(a) Strain in site 1.(b) Strain in site 2.

Figure 7 :Figure 8 :
Figure 7: Pore structure of cement paste embedded in reinforced concrete wall at (a) 28 days and (b) 90 days.(c) Cumulative porosity curve at 28 days and 90 days.

Table 3 :
Mix proportions of paste samples embedded in concrete wall (%).