Experimental study of magnesium phosphate cements modified by metakaolin

Construction and Building Materials 123 (2016) 719–726

Contents lists available at ScienceDirect

Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat

Experimental study of magnesium phosphate cements modified by metakaolin Xuan Lu, Bing Chen ⇑ Department of Civil Engineering, Shanghai Jiaotong University, Shanghai 200240, PR China

h i g h l i g h t s  The MK-MPC mortar yielded a higher one-hour compressive and flexural strength than MPCs.  The introduction of MK largely prolonged the setting time of MPCs up to 52 min.  The addition of MK improved the water resistance of MPCs.  Mechanisms of hydration products of MK-MPC system were proposed.

a r t i c l e

i n f o

Article history: Received 1 June 2016 Received in revised form 27 June 2016 Accepted 15 July 2016

Keywords: Magnesium phosphate cement Metakaolin Setting time Mechanical property Microstructure

a b s t r a c t Magnesium phosphate cements have the characteristics of fast setting and high early strength, which have been extensively used as rapid repair materials in civil engineering. However, Magnesium Phosphate Cement (MPCs) were also reported to have the disadvantage of water instability and short setting time which was difficult to control. The present research was carried out to evaluate the properties of magnesium phosphate cement with high dosage of metakaolin (MK). Different contents of metakaolin were used to replace magnesium to prepare magnesium phosphate cement. The experimental results showed that the successive additions of MK led to a cumulative increase in setting time to a maximum of 52 min for the compositions investigated, and the intensity of the exothermic reactions was also reduced by MK. Moreover, the additions of MK also improved the strengths of the specimens greatly at the early age, and the compressive strength at 1 h can reach 65.7 MPa. Meanwhile, the water resistance of MPC mortar was also investigated by the form of strength retention, and the results revealed that the addition of MK improved the water resistance after curing for 28 days and 56 days. The improvement mechanism was discussed based on the micro-analysis of XRD, SEM along with energy dispersive spectrometer (EDS). The phase analysis revealed the promotion effect of MK on the hydration of MPC, and confirmed the formation of a new gel. From the microstructural and compositional analyses, the gel was identified as aluminum phosphate (AlPO4), which increased the density of the cement. Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction Magnesium Phosphate Cement (MPC) was invented and developed in the 19th century. Since 1970s, it has been investigated as one type of rapid-repair structural materials and through continuous improvements [1,2]. It is well known for its fast setting and high early compressive strength characteristics and attracted much attention in the fields of dental and bone restorations [3], rapid repair to damaged concretes [4–6], and refractories and hazardous waste management [7–9]. However, as an acid-base ⇑ Corresponding author. E-mail address: [email protected] (B. Chen). http://dx.doi.org/10.1016/j.conbuildmat.2016.07.092 0950-0618/Ó 2016 Elsevier Ltd. All rights reserved.

reaction material, certain issues regarding MPCs are still remaining in the engineering applications: (1) the rapid reaction makes the setting time too short to be cast in the construction; (2) the vast heat is hard to disperse timely in the large-scale pouring; (3) the cost is higher than the traditional Portland cement; (4) the water resistance is poor; and (5) unpleasant ammonia gas may be released to cause pollutions to the surroundings during its setting and hardening stages [10–14]. In some common practice, retarders such as borax, boric acid or sodium triphosphate (STP) were normally used to control the setting time and reduce the intensity of the exothermic reactions during the initial setting, however, which may cause a reduction on its mechanical strength, especially in the early period [15,16]. To reduce the high cost, fly ash was

720

X. Lu, B. Chen / Construction and Building Materials 123 (2016) 719–726

added to this type of cement, but it was found that high content of fly ash would cause a dramatic reduction on the early strength of MPC [17,18]. To improve the water resistance, Gai et al. [19] introduced silica and cellulose into this cement system which could enhance its water resistance to some extent, but also had negative effects on the mechanical strength of hardened mortar. In general, although continuous efforts had been done to solve these remaining issues, an effective means to control setting time and resultant properties is still missing. According to previous research [20], pozzolanic material had already been proved as a material which leads to high-early strength, low porosity, high temperature resistance and well chemical corrosion resistance. Among others, metakaolin was a candidate material [21,22]. The metakaolin (MK, Al2O3
2SiO2) is a kind of thermal alteration product made by kaolin calcined and dehydrated at the temperature of 500–980 °C [23], forming the anhydrous aluminum silicate with a strong reactive activity. Some early studies [24,25] found that the mixture of alumina with phosphate acid could yield amorphous gelatinous substance of aluminum phosphate (AlPO4) at temperatures greater than 150 °C, which had a denser microstructure. Moreover, according to Fan [26], small addition of active alumina to MPC system had positive effects on its setting properties and significantly enhanced its mechanical properties and improved its water resistance. Thus, it is reasonable to assume that similar effect could happen in the MK which has active alumina. Also, from the economic and ecological perspective, MK is a natural-sourced mineral which is much cheaper than magnesia, thus this proposed research on the addition of MK to MPC systems should have a broader impact. Therefore, the present paper is concerned with the effects of MK on the MPCs, with particular emphasis on the improvements of strength and performance. For this purpose, the microstructures and compositions of new MK-MPC specimens were studied and the hydration products were examined using X-ray diffraction (XRD) and Scanning Electronic Microscopic-Energy Dispersive Spectrometer (SEM-EDS).

2. Experimental details 2.1. Raw materials The MK-MPC mortar was prepared as a mixture of magnesium oxide (MgO), MK, ammonium dihydrogen orthophosphate (NH4H2PO4), sand, water reducer, and borax (Na2B4O7
10H2O) in diverse proportions. The magnesia powder with average particle size of 1.05 lm was calcined at about 1500 °C for about 6 h, and a purity of 89.5% was achieved in the Taishan Refractory Plant of Shanghai. The ammonium dihydrogen orthophosphate and borax used in the previous research were in the industrial grade provided by Fine Chemical Plant of Wujiang, Jiangsu province, China. Natural river sand with the maximum size of 5 mm was used in the experiments to prepare the MPC mortar. The MK with specific surface area of 18 m2/g was provided by Chaopai Co., Ltd of Hunan province of China. The chemical compositions of the raw materials in the preparation of the mortar are listed in Table 1.

2.2. Specimen preparation For each specimen, the weight ratio of water to solid (W/C) which consists of magnesia, MK, ammonium dihydrogen orthophosphate and sand was fixed at 0.12, and the mass ratio of phosphates to metal oxide (magnesium and MK, P/(P + MK)) (M/P) was always 2.5. The MPC mortar was prepared using a fixed content of borax (3.5% of the weight of metal oxide) with a sand/ binder ratio of 1.0. Table 2 summarizes these mixing parameters. The mixing procedure can be divided into two steps: (i) drymixing of the powders consisting of magnesia, MK, borax, water reducer and sand for 3 min; (ii) further mixing after water introduced into the mixture for 2 min. After mixing, the mixed fresh mortars were casted into the molds immediately and were demoded one hour later.

2.3. Property measurements Considering that MPC is mainly used for repair and as grout material, the fluidity and setting time of the fresh paste were tested. For each mixture, the heat evolution during its exothermic setting and hardening reactions were monitored by the automatic temperature recorder. Meanwhile, the setting time was recorded by using a modified Vicat needle according to the ASTMC187 standard. The flowing table test (standard UNE-EN 1015-3) as demonstrated in Fig. 1 was conducted to check the spread diameter of the MK-MPC slurry. In this test, a platform type mold with an initial diameter of 100 mm was placed on a plate glass, and the slurry was poured into it and then lifted vertically. The spread diameter of the mortar was measured in two perpendicular directions, and the mean values were recorded. The compressive strength and flexure strength of the hardened MK-MPCs mortar were measured using the MTS servo hydraulic testing machine at a speed of 1 mm/min. The specimens were poured into the molds with the dimension of 40  40  40 mm for measuring the compressive strength and 40  40  160 mm for measuring flexure strength. The strengths of mortar samples with three replicates under different curing condition were tested at the ages of 1 h, 1 d, 7 d and 28 d. For the samples cured in water, it was kept in water after demolding. Before strength testing, the water curing samples were taken out of water and dried in air for 2 h (room temperature is around 20 °C). The strength retention (Sr) is defined as the ratio of the compressive strength of the water curing sample to that of the air curing one, which is used to reflect

Table 2 Mixing proportions of MPC mortars. Group

Mix proportions (wt%) MgO

MK

A0 A30 A40 A50 A60

100 70 60 50 40

0 30 40 50 60

P/(M + MK)

W/C

B/O (%)

1/2.5 1/2.5 1/2.5 1/2.5 1/2.5

0.23 0.23 0.23 0.23 0.24

3.5 3.5 3.5 3.5 3.5

Table 1 Chemical compositions of magnesia and MK. Raw material

Magnesia MK

Mass fraction of the sample (%) MgO

Al2O3

SiO2

P2O3

CaO

Fe2O3

Na2O

K2O

SO3

N2O

90.5 –

1.3 43.9.

4.91 49.4

0.11 –

1.44 0.27

1.2 0.51

– –

– 0.23

– 0.14

– 1.52

721

X. Lu, B. Chen / Construction and Building Materials 123 (2016) 719–726

Fig. 1. The flowing table test.

80 70

50

60

300 flow /mm setting time /min

280

50

260

40

250 240

30

230 220

20

210 200

10

190 20

30

40

50

60

Addition of MK /% Fig. 2. Effect of MK on the properties of MPC fresh slurry.

Setting time /min

Flow /mm

270

10

A50

30

A60

20

0

10

20

30

40

50

60

Fig. 3. Calorimetry curves obtained for MPC mortar prepared with various additions of MK.

Fig. 2 illustrates the effects of MK concentration on the fluidity and setting time of the MPC mortar. The fluidity of the mortar increased significantly with the introduction of MK first and then decreased with more MK. The fluidity of mortar without MK was 260 mm, while with 30% MK was 280 mm. After that, the fluidity began to decrease, as observed in Fig. 2. It is generally thought that the fine granules within the MK serve as rolling bearings in the paste and improve the flowability. However, due to the fine nature

0

A40

Time (min)

3.1. Properties of fresh MK-MPC mortar

290

A30

40

10

3. Results and discussion

A0

60

Temperature (°C)

the water stability of the sample. For microstructure and compositional analyses, the reaction products of MK-MPC mortar were crushed and subsequently triturated using an agate mortar and then put into the alcohol for the planned days to avoid the hydrations. Soon after that, these small hydrated samples were analyzed by X-ray diffraction (XRD, D8 ADVANCE) for the mineralogical analyses and diffraction patterns collected between 20° < 2h < 70°. Using JCPDS database, the indexing of the peaks was carried out separately and these phases can be identified by the software. Moreover, the microstructure and micrographs of the reactants were examined by scanning electron microscopy (SEM, NOVA Nano SEM 230) equipped with an energy dispersive spectrometer (EDS).

of MK, more water was need to keep the similar fluidity. The addition of MK prolonged the setting time, and the setting time increased with the increasing MK content. In the case of the mortar without MK, the setting time was only 6 min which is too short to permit casting. When the MK content reached 60%, the setting time was 52 min, which was sufficient for engineering practice requirements. The effect of the MK on the exothermic reaction for the MPC mortar was shown in Fig. 3. It is found that A0, which was prepared without MK, have the highest exothermic reaction intensity. In contrast, the MPC pastes with the addition of MK showed less exothermic reaction intensity. Successive replacements of magnesia by MK led to a cumulative reduction in the peak temperature (TP) during setting of the pastes and an extension at the time that the peak temperature reached. The TP of A0 was 73.5 °C, occurring around 6 min. For the sample of A30 of 30% MK, its TP was 50.0 °C after 22.0 min of reaction. However, for the sample of A60, the temperature showed the slowest increase. The TP was only 27.1 °C after about 52.0 min. In consideration of all, it is generally thought that the addition of MK slowed down the heat evolution of MPCs and also its exothermic peak decreased synchronously with the increasing ratio of MK. It was known that the trivalent metal oxide alumina had a much lower solubility compared to the divalent metal oxide magnesia. For direct comparison, the sample A90 with large content of MK up to 90% was designed to

X. Lu, B. Chen / Construction and Building Materials 123 (2016) 719–726

study the reaction between MK with phosphate. Unfortunately, it was not hardened even waiting for a long time period, indicating that MK didn’t react with phosphate at room temperature. Therefore, alumina in MK can’t be directly used for the formation of phosphate cements but it could be employed using heat treatment. The reactions in the MPC containing MK could be concluded: (a) dissolution of phosphate and magnesia; (b) exothermic reaction between magnesium and phosphate ion raised the temperature; (c) alumina slowly dissolved and participated in the reaction; and (d) the formation of the gel and condensation into the MK-MPCs [27]. The dissolution of alumina was considered much less likely than that of magnesia due to its solubility. Therefore, the setting time of MPC increased with the increasing MK content. Moreover, during the setting reaction of MPC containing MK, it is interesting to find that there is no unpleasant ammonia gas releasing, differing from the previous research [28], which can be extended to indoor applications. Hence, from the viewpoint of controlling of setting time, cost of materials and environmental conservation, the application of MK to MPC mortar has more benefits than conventional MPC products. 3.2. Mechanical properties of MK-MPC mortar The influence of the MK contents on the compressive strength of MPC mortar at different curing ages is shown in Fig. 4. As shown in Fig. 4, the compressive strength at different curing ages initially increased gradually with increasing the MK content, and then decreased as the MK content increased beyond 50%. The compressive strengths of specimens A50 at 1 h, 1 day, 7 days and 28 days were 65.7 MPa, 74.5 MPa, 82.5 MPa and 100.5 MPa, which were 87.7%, 66.2%, 63.3% and 80.1% higher than those of specimens A0, respectively. In contrast, the sample A0 had the lowest strength, and its compressive strength increased from 35 MPa at 1 h to 55 MPa at 28 days, appearing less improvement on the long-term strength relatively. Also, it is known that MPC is well known for its high early-strength characteristic at 1 h from previous researches [17,26,29]. While, the sample A50 containing 50% MK yielded the one-hour strength up to 65.7 MPa which was much higher than the common one, 35 MPa. Overall, the conclusion could be drawn that the addition of MK improved the MPCs mechanical strength in both the early- and the long-term run. It is noted that the increasing trend was not hold when the addition of MK was over 50%, with a decrease instead. It is possible that the sample A60 with a high content of MK may have a higher

14

Flexural strength (MPa)

722

12

1h 1d 7d 28d

10 8 6 4 2 0

A0

A30

A40

A50

A60

MK-MPC mortar samples Fig. 5. The flexural strength of MK-MPC mortar samples.

compressive strength. However, as shown in Table 2, since it is a clay material and it needs more water during mixing for desirable flowability, and thus causing its reduction on the compressive strength. Meanwhile, the insufficient of magnesia was also a cause for the reduction. Therefore, the ratio of MK around 50% maybe the optimal composition for the optimal mechanical strength at the age of 28 days. The results of flexural strength are illustrated in Fig. 5, showing similar trends as the compressive strength. In the case of sample A0 which had the lowest flexural strength, it increased from 3.0 MPa to 10.1 MPa, yielding a low-early flexural strength. While in contrast, the sample A50 with 50% content of MK yielded a higher flexural strength of 11.3 and 13.5 MPa at the age of 1 h and 28 days. Also, when the ratio of MK was greater than 50%, it started decreasing. In general, the addition of MK to MPC mortar yielded a higher flexural strength than the specimen without MK. Furthermore, considering the ratio of flexural strength to compressive strength at the age of 1 h, the sample A0 was just 0.086, while the other specimens A30 to A60 were 0.161, 0.164, 0.172 and 0.193, respectively, showing an increasing trend and were much higher than the initial one. Therefore, the introduction of MK also yielded a better toughness in the early period which may extend rapid repair applications.

3.3. The water resistance of MK-MPC mortar 1h 1d 7d 28d

Compressive strength (Mpa)

100

80

60

40

20

0 A0

A30

A40

A50

A60

MK-MPC mortar samples Fig. 4. The compressive strength of MK-MPC mortar samples.

From the previous research [30], the poor water resistance of MPC resulted in a great loss of the strength of the hardened MPC pastes in water. The strength retention (Sr) is used to evaluate the index of the water resistance and the lager strength retention indicates better water resistance. The strength retentions of MK-MPC mortar samples at the age of 28 days and 56 days are showed in Fig. 6. In contrast, successive replacements of magnesia by MK led to a cumulative increase on the strength retention at the age of 28 days, such as from specimens A0 to A60, theirs retention increased from 78% to 90%, respectively. However, after 56 days water-curing, the sample A0ˈs retention showed a sharply decrease from 78% to 70%, while A30 and A40 exhibited a sustained growth. In particular, the sample A40 increased from 86% to 94% at 56 days, indicating a better water stability among them in the long term. While, the samples A50 along with the A60 which contained a large content of MK and had a higher strength retention at 28 days but dropped a little at 56 days. It is possible that when the ratio of MK is greater than 50%, there may exist a number of unreacted MK in the specimens, including the active silica and alumina particles

X. Lu, B. Chen / Construction and Building Materials 123 (2016) 719–726

28d 56d

Strength retention (%)

100

80

60

40

20

0 A0

A30

A40

A50

A60

MK-MPC mortar samples Fig. 6. The strength retention of MK-MPC mortar samples.

which were much smaller than the magnesia particles. In the aircuring condition, the excess smaller particles could act as a filling effect to improve the mechanical strength, but once placed in water, the water percolated through the mortar and took out these small particles, weakening its water resistance simultaneously. It can be concluded that the presence of MK obviously improves the water resistance of MPC mortar and the ratio of MK around 40% maybe the optimal level which is in consideration of mechanical strength and water resistance in the long term. 3.4. Phase and microstructure analyses MK-MPC mortar To gain insight into the mechanisms of the strength enhancements, the representative specimens cured in the air conditions were researched by X-ray diffraction and scanning electron microscopy equipped with energy dispersive spectrometer. Fig. 7 presents the X-ray diffraction spectra for the MK-MPC pastes cured at 1 day. As observed in Fig. 7, the XRD patterns showed the unreacted magnesia which had the sharp peak intensity and remained in the cement paste matrix. The peak intensity of magnesia decreased gradually with the increasing content of MK in the systems. Apart from the unreacted magnesia, the main hydration product struvite (MgNH4PO4
6H2O) was also clearly

MgO NH4MgPO4•6H2O

c

b

a 20

30

40

50

60

70

Angular Range (2θ) Fig. 7. XRD spectra of MK-MPC samples with various ratios of MK at 1 day (a):A0, (b):A30, (c):A50.

723

observed. The diffraction peak intensity of struvite increased with the addition of MK, implying that it yielded a growing amount of struvite than the initial one. According to Qiao [31], the actual consumption of MgO in the reaction with phosphate (Mg/P = 4) was just 20.8%. It is believed that the massive remaining of the unreacted magnesia was largely due to its rapid reaction, which caused the hydration products coating on the surface of magnesia and retarded its following reaction. While, when MK was introduced to this mortar, it slowed down the hardening process and made more magnesia particle participate in the reaction. Hence, for the promotion effect of MK on the hydration process of MPC, it yielded more products. Further, taking a view of Fig. 7b and c, they are different from Fig. 7a in the presence of broad hump in the range of 25–40°. These broad diffraction peaks represented the formation of gel due to their amorphous nature. Also, because no crystalline phases containing aluminum presented in the mortar can be clearly detected, it is supposed that most of the aluminum introduced from the MK was present in the amorphous phase. Meanwhile, considering the MK belongs to a material that has the gel property, and strong pozzolanic activity can be motivated by alkaline environments (the hydrated MPC system is alkaline, with a PH of 10–11) [32,33]. The amorphous substances containing aluminum could be reasonably thought to be the main hydration products of MK, which play a gelling role in the hardening process of MK-MPC mortar. Fig. 8 shows the microstructures of the MK-MPC mortar specimens cured at one day. Such as sample A0, it could be obviously seen the existence of some lamellar-like particles and identified to be the struvite mentioned by other researchers [34,35]. Meanwhile, the wide and smooth crystals existed in the figure were the unreacted MgO which were reacted incompletely for its rapid reaction, and surrounded by the struvite along the fracture edge [26]. The result showed the presence of much unreacted magnesia at the age of 1 day, which is identical to the phase analysis. For the sample A30, it was covered by lots of fine and dense block crystals and microcracks between them. While compared to A0, these lamellar-like crystals became more intensive, and the wide and smooth crystals above mentioned were less exposed to the outside, indicating gradual decrease of magnesia. Furthermore, considering the samples A40 and A50 together, their fracture surfaces became rougher and more compact, surrounded with a layer of tiny and waved-like substance. It could hardly see the microcracks or pores existing on theirs fracture surfaces, indicating a denser microstructure than before. Also, the wide and smooth unreacted MgO crystals could not be observed any more. In general, the introduction of MK to MPCs makes the structure more compact and possibly due to this effect, it could yield the higher mechanical strength in the early period. Besides, fewer microcracks and pores on the surface of sample A40 and A50 are likely the reason for their good water resistance. The typical microstructures of the specimens A30 and A50 after one day curing observed by using SEM were subsequently scanned by EDS, as shown in Fig. 9. Also, the elemental statistics are complied in Table 3. In area 1, it can be obviously seen that the major elements were O, Mg, Al, Si, N and P. Theirs molar ratios (m.r.) indicated the existence of reaction products struvite, the residual silica and magnesia. Then the remaining molar ratio of Al:P:O was 1: 2.38:9, close to the theoretical value of the molar ratio (1:2:9) of the atoms in AlH3(PO4)2
H2O, which is known as the intermediate phase of the alumina reacted with phosphate acid [36]. In area 2, the molar ratio of major elements indicated the presence of struvite, silica and magnesia, which was identical to area 1 in the final product. The remaining Al:P:O was 1:1.1:4.4 was very close to the molar ratio (1:1:4) of atoms in aluminum phosphate (AlPO4), the final reaction product of alumina with phosphate acid [37,38]. According to Wagh AS [39], once the

724

X. Lu, B. Chen / Construction and Building Materials 123 (2016) 719–726

Fig. 8. SEM photomicrographs of MK-MPC pastes at 1 day: (a) A0; (b) A30; (c) A40; (d) A50.

Fig. 9. Representative SEM photomicrographs with an EDS scan on the window areas for the MK-MPC pastes at 1 day:(a)A50;(b)A30.

alumina dissolves into the phosphate solution, it reacts with dissolved phosphate ions and forms the AlH3(PO4)2
H2O, which would subsequently react with the remaining alumina on the surface of their grains and yields the final product AlPO4. The AlPO4 is known as a gelatinous and amorphous substance and can be formed at room temperature, thus the gelatinous substance mentioned in the phase analysis is generally thought to be the AlPO4. For its gelling effect, it glues the unreacted magnesia or struvite and

coated on their surface, forming a denser microstructure than before. Besides, for the scanned area 3, the major elements were O and Mg, giving evidence that the dark, dense, and smooth parts were the fracture surface of unreacted magnesia, consistent with previous discussions. Finally, in area 4, the major elements were O, Mg, Al and Si, it cannot see the existence of element P, implying that the matrix is mainly a mixture of unreacted magnesia and aluminum-silicon oxide.

X. Lu, B. Chen / Construction and Building Materials 123 (2016) 719–726 Table 3 The elements in the areas for the samples scanned by Energy-dispersive X-ray spectroscopy (EDS). Element Area 1

O Mg Al Si N P

Area 2

Area 3

Atom (%) m.r.

Atom (%) m.r.

Atom (%) m.r.

Atom (%) m.r.

55.19 5.82 5.51 8.39 2.91 22.09

48.73 6.96 7.19 11.44 3.68 17.40

36.36 60.95 0.00 0.00 0.00 0.00

36.02 41.78 9.98 12.22 0.00 0.00

14.08 1.00 0.90 1.20 0.86 3.01

10.5 1.00 0.90 1.40 0.91 2.00

Area 4

0.90 1.00 – – – –

1.30 1.00 0.22 0.25 – –

4. Conclusions This paper conducted an experimental research on the MK-MPC mortar, produced by the original MPC and extra addition of MK, with a focus of the influences of MK on the MPC mortar. The mechanical properties of the mortar including compressive strength, flexural strength, along with its setting time, heat evolution, water resistance were also systematically studied. Based on these observations and results, the following conclusions could be drawn: (1) Increasing the ratio of MK could notably extend the setting time of MPC mortar. When the ratio of MK is up to 60%, the setting time would extend to 52 min which is nearly nine times longer than the case without MK. The prolonged setting time is largely due to the decrease of magnesia and also the reaction of alumina with phosphate requires heat that would retards its reaction to some extent. (2) The addition of MK greatly decreases the intensity of heat evolution of MPC mortar during the exothermic setting and hardening reactions. The peak temperatures reduce dramatically with the increasing MK content. The results imply that the large amount of heat released by MPC contributes a lot to the MK’s hydration. Besides, the insufficient magnesia is also a mainly cause for its reduction. (3) The presence of MK largely improves the mechanical strength of MPC mortar both in short term and long term. It yields a higher compressive strength, flexural strength and better toughness at the age of 1 h. In the present research, when the addition of MK is up to 50%, it attains the maximum value of compressive and flexural strength, independent of curing time. Furthermore, the MK-MPC mortar exhibits better water resistance than that of MPC both at the age of 28 days and 56 days. The water stability of MPC mortar improves gradually with increasing amount of MK after 28 days curing. (4) According to the XRD and SEM analysis, the addition of MK promotes the hydration of MPC mortar and yields more struvite. Meanwhile, the MK-MPC mortar yields the aluminum phosphate gel which is mostly distributed in amorphous phase, coated on the surface of the individual particles and glues each other, making the microstructure more compact than the original MPC.

Acknowledgement This research work was financially supported by the National Natural Science Foundation of China, Grant No. 51378309. References [1] T. Sugama, L.E. Kukacka, Magnesium monophosphate cements derived from diammonium phosphate solutions, Cem. Concr. Res. 32 (3) (1983) 407–416.

725

[2] B.E.I. Abdelrazig, J.H. Sharp, A discussion of the papers on magnesia-phosphate cements by T. Sugama and L.E. Kukacka, Cem. Concr. Res. 15 (5) (1985) 921– 922. [3] M. Kouassi, P. Michailesco, A. Lacoste-Armynot, P. Boudeville, Antibacterial effect of a hydraulic calcium phosphate cement for dental applications, J. Endod. 29 (2) (2003) 100–103. [4] S.S. Seehra, S. Gupta, S. Kumar, Rapid setting magnesium phosphate cement for quick repair of concrete pavements-characterisation and durability aspects, Cem. Concr. Res. 23 (2) (1993) 254–266. [5] Q. Yang, X. Wu, Factors influencing properties of phosphate cement based binder for rapid repair of concrete, Cem. Concr. Res. 29 (4) (1999) 389–396. [6] S. Popovics, N. Rajendran, M. Penko, Rapid hardening cements for repair of concrete, ACI Mater. J. 84 (1987) 64–73. [7] S. Vinokurov, Y.M. Kulyako, O. Slyuntchev, S. Rovny, B. Myasoedov, Low temperature immobilization of actinides and other components of high-level waste in magnesium potassium phosphate matrices, Nucl. Mater. 385 (1) (2009) 189–192. [8] J. Davidovits, Recent progress in concretes for nuclear waste and uranium waste containment, Concr. Int. 16 (1994) 53–58. [9] J. Torras, I. Buj, M. Rovira, J.D. Pablo, Semi-dynamic leaching tests of nickel containing wastes stabilized/solidified with magnesium potassium phosphate cements, Hazard Mater. 186 (2–3) (2011) 1954–1960. [10] Q.B. Yang, S.Q. Zhang, Q.R. Yang, X.L. Wu, Properties of a new-typed repairing material, the rapidly hardened phosphate cement, China Concr. Cem. Prod. 4 (2000) 8–11. [11] Q.B. Yang, B.R. Zhu, S.Q. Zhang, X.L. Wu, Properties and applications of magnesia-phosphate cement mortar for rapid repair of concrete, Cem. Concr. Res. 30 (11) (2000) 1807–1813. [12] F. Qiao, C. Chau, Z. Li, Property evaluation of magnesium phosphate cement mortar as patch repair material, Constr. Build. Mater. 24 (5) (2010) 695–700. [13] D.X. Li, P.X. Li, C.H. Feng, Research on water resistance of magnesium phosphate cement, Build Mater. 12 (5) (2009) 505–510. [14] T. Sugama, L.E. Kukacka, Characteristics of magnesium polyphosphate cements derived from ammonium polyphosphate solutions, Cem. Concr. Res. 13 (4) (1983) 499–506. [15] D.A. Hall, R. Stevens, B. El-Jazairi, The effect of retarders on the microstructure and mechanical properties of magnesia-phosphate cement mortar, Cem. Concr. Res. 31 (3) (2001) 455–465. [16] J. Yang, C. Qian, Effect of borax on hydration and hardening properties of magnesium and potassium phosphate cement pastes, Wuhan Univ. Technol. Mater. Sci. Ed 25 (4) (2010) 613–618. [17] Y. Li, B. Chen, Factors that affect the properties of magnesium phosphate cement, Constr. Build. Mater. 47 (2013) 977–983. [18] N. Vol, High-early-strength magnesium phosphate cement with fly ash, ACI Mater. J. 102 (6) (2005) 375–381. [19] W. Gai, C. Liu, X.Z. Wang, Influence of the compound additives on performances of magnesium phosphate bone cement, J. East China Univ. Sci. Technol. 28 (4) (2002) 393–396. [20] S. Rukzon, P. Chindaprasirt, Strength, porosity, and chloride resistance of mortar using the combination of two kinds of pozzolanic materials, Int. J. Min. Met. Mater 20 (8) (2013) 808–814. [21] M.A. Caldarone, K.A. Gruber, R.G. Burg, High-reactivity metakaolin: a new generation mineral admixture, Concr. Int. 16 (11) (1994) 37–40. [22] M.H. Zhang, V.M. Malhotra, Characteristics of a thermally activated aluminosilicate pozzolanic material and its use in concrete, Cem. Concr. Res. 25 (8) (1995) 1713–1725. [23] A.M. Rashad, Metakaolin as cementitious material: history, scours, production and composition: a comprehensive overview, Constr. Build. Mater. 41 (2) (2013) 303–318. [24] S.J. Lukasiewicz, J.S. Reed, Phase development on reacting phosphoric acid with various bayer-process aluminas, Ceram. Bull. 66 (7) (1987) 1134–1138. [25] M.J. Ohara, J.J. Duga, H.D. Sheets, Studies in phosphate bonding, Am. Ceram. Soc. Bull. 51 (7) (1972) 590–595. [26] S.J. Fan, B. Chen, Experimental research of water stability of magnesium alumina phosphate cements mortar, Constr. Build. Mater. 94 (2015) 164–171. [27] E. Soudée, J. Péra, Mechanism of setting reaction in magnesia-phosphate cements, Cem. Concr. Res. 30 (2) (2000) 315–321. [28] Y.X. Shao, Y.P. Gu, J.M. Yang, Control technology study of early water stability of magnesium potassium phosphate cement, Bull. Chin. Ceram. Soc. 32 (6) (2013) 1141–1145. [29] Z. Ding, Z. Li, High-early-strength magnesium phosphate cement with fly ash, ACI Mater. J. 102 (6) (2005) 375–381. [30] D. Deng, The mechanism for soluble phosphates to improve the water resistance of magnesium oxychloride cement, Cem. Concr. Res. 33 (9) (2003) 1311–1317. [31] F. Qiao, Reaction mechanisms of magnesium potassium phosphate cement and its application Ph.D thesis, 74, Hong Kong University Science Technology, 2010, pp. 64–68. [32] Chao Li, Hengsu Sun, Longtu Li, A review: the comparison between alkaliactivated slag (Si+Ca) and metakaolin (Si+Al) cements, Cem. Conc. Res. 40 (9) (2010) 1341–1349. [33] F. Pacheco-Torgal, J. Castro-Gomes, Alkali-activated binders: a review Part 1. Historical background, terminology, reaction mechanisms and hydration products, Constr. Build. Mater. 22 (2008) 1305–1314.

726

X. Lu, B. Chen / Construction and Building Materials 123 (2016) 719–726

[34] Z. Ding, B. Dong, F. Xing, N. Han, Z. Li, Cementing mechanism of potassium phosphate based magnesium phosphate cement, Ceram. Int. 38 (8) (2012) 6281–6288. [35] C. Chau, F. Qiao, Z. Li, Microstructure of magnesium potassium phosphate cement, Constr. Build. Mater. 25 (6) (2011) 2911–2917. [36] A.S. Wagh, S.Y. Jeong, Chemically bonded phosphate ceramics: Chapter 11.2 Formation of berlinite bonded alumina ceramic, J. Am. Ceram. Soc. 86 (11) (2003) 128–131.

[37] T. Sugama, N. Carciello, Hydrothermally synthesized aluminium phosphate cements, Adv. Cem. Res. 5 (17) (1993) 31–40. [38] D. Hanen, L. Soumaya, B. Samir, et al., Structural, thermal and dielectric properties of phosphoric acid-based geopolymers with different amounts of H3PO4, Mater. Lett. 116 (2014) 9–12. [39] A.S. Wagh, S.Y. Jeong, Chemically bonded phosphate ceramics: Chapter 11.1 Solubility enhancement with temperature and formation of berlinite phase, J. Am. Ceram. Soc. 86 (11) (2003) 1838–1844.