Characterization of MgCl2·6H2O-Based Eutectic/Expanded Perlite Composite Phase Change Material with Low Thermal Conductivity

The melting points of the phase change materials (PCMs) incorporated into the walls of buildings should be within the human thermal comfort temperature range. In this paper, 15 wt.% of MgCl2·6H2O was mixed with CaCl2·6H2O to obtain the eutectic with a melting point of 23.9 °C. SrCl2·6H2O suppresses the supecooling of the eutectic. The combination with expanded perlite (EP) via the impregnation method overcomes the phase separation and liquid leakage of the CaCl2∙6H2O-MgCl2∙6H2O mixture. The composite PCM is form-stable with the maximum loading mass fraction up to 50 wt.% and latent heat of 73.55 J/g. EP also significantly reduces the thermal conductivity of the CaCl2∙6H2O-MgCl2∙6H2O from 0.732 to 0.144 W/(m·K). The heating-cooling cycling test reveals that the composite PCM is thermally stable. The cheap eutectic salt hydrate, with little supercooling, no phase separation and liquid leakage, low thermal conductivity and good thermal reliability, show great potential as envelope materials to save energy consumption in buildings.


Introduction
A greater thermal mass and better thermal insulation of building envelopes provides improved thermal comfort, thus reducing the heating/cooling load. Phase change materials (PCMs) have large thermal mass because they can store and release a large amount of heat within a small temperature range during the phase change process between liquid and solid [1,2]. Therefore, incorporating PCMs with high heat capacity into building envelopes will definitely save energy consumptions in buildings [3,4].
The perfect PCM for building envelopes should be safe and cheap. They should also have high thermal storage density and a suitable phase change temperature within the human thermal comfort zone, which is between 22-26 • C. Compared to fatty acids or other organic PCMs [5], inorganic PCMs-such as salt hydrates, molten salts, etc.-offer the advantages of nonflammability and larger thermal storage density [6]. Hydrated salts are quite cheap due to rich resources in salt lakes [7,8]. Therefore, inorganic PCMs have recently gained intensive attention in building energy savings [9][10][11]. CaCl 2 ·6H 2 O is an inorganic hydrate with a melting point of 28 • C [12]. As reported by Li et al. [13], the melting point of CaCl 2 ·6H 2 O can be regulated to 21-26 • C if it is mixed with First, the relatively high thermal conductivity gives the hydrated eutectic a poorer thermal insulation performance than the organic PCMs. Thus, the extra heat exchange between the building and the ambient will cause higher cooling/heating loads. Second, eutectics have inherent shortcomings such as supercooling and phase separation. More energy is consumed to extract the heat stored in the PCM if it is supercooled. Phase separation may cause the salt hydrate to stop functioning once its composition has changed. Unless the problems of high thermal conductivity, high supercooling degree and phase separation can be solved, the CaCl 2 ·6H 2 O-MgCl 2 ·6H 2 O eutectic cannot be integrated into building envelopes for energy saving efforts.
Nucleating agents such as SrCl 2 ·6H 2 O have been found effective to reduce the supercooling degree of CaCl 2 ·6H 2 O and its eutectics [14]. Ways of alleviating the phase separation include adding thickening agents such as Carboxymethyl cellulos (CMC) and absorbing the PCMs into porous supporting materials such as expanded graphite [15][16][17], silica [18] or expanded perlite [19,20]. The porous matrix is a tank to hold the melted PCM that is in liquid state with the capillary force, so that the liquid leakage of the PCM can be avoided. In the meantime, the hydrated salts are constrained into macro or micro pores, which prevents the salts from separating from the hydrates. If the porous matrix has low thermal conductivity like silica or expanded perlite, the thermal conductivity of the salt hydrates can be reduced as well. Therefore, if CaCl 2 ·6H 2 O-MgCl 2 ·6H 2 O is composited with SrCl 2 ·6H 2 O and a porous matrix with low thermal conductivity, the problems that hinder its application in building envelopes will be solved simultaneously.
However, the preparation of CaCl 2 ·6H 2 O-MgCl 2 ·6H 2 O composite PCM has not yet been reported. In this work, we prepare a series of CaCl 2 ·6H 2 O-MgCl 2 ·6H 2 O eutectics with different compositions and different melting points. Then the mixture with the optimal melting point for use in building is selected. After that, the eutectic is composited with SrCl 2 ·6H 2 O and expanded perlite (EP) to reduce the supercooling degree, ease phase separation, and prevent liquid leakage. We study the microstructure, thermal properties and thermal reliability of the obtained composite PCM. The aim of this paper is to prepare a composite PCM with good thermo-physical properties for building envelopes.

Characterization
A differential scanning calorimeter (DSC, Q20, TA Instruments, New Castle, DE, USA) was used to analyze the phase change characteristics of the pristine eutectic, and the composite PCMs. The DSC tests were carried out under the atmosphere of nitrogen at a flow rate of 50 mL per minute. The testing temperature rose up from −5 to 50 • C at a rate of 3 • C/min.
Under the nitrogen atmosphere (flowrate: 100 mL/min), a thermoanalyzer (STA409PC, Netzsch, Waldkraiburg, Germany) tested the mass change of each sample during the heating process from 25 • C to 600 • C at a rate of 10 • C per minute. Thermal gravimetric analysis (TGA) characterizes the thermal stability of those materials.
The thermal conductivity of the materials in solid state were measured with a thermal constant analyzer (TPS 2500, Hot Disk, Gothenburg, Sweden). The composite PCM was compressed into two cylinders 40 mm in diameter and 10 mm in height. The sensor (type: 7577, 2.01 mm in diameter, Hot Disk, Gothenburg, Sweden) was sandwiched between the two cylinders. The thermal conductivity of the EP and eutectic were also measured, respectively.
The supercooling degree of the PCMs was characterized by the cooling curves. The schematic diagram is shown in Figure 1. Two tubes that contained the eutectic and the composite PCM composites were put into a thermal bath in which the temperature was controlled by a thermal oil (Phenyl silicone oil, flash point: 315 • C, Clearco Products Co., Inc., Willow Grove, PA, USA). Two K-type thermocouples (NR-81530, accuracy: ±0.5 • C, HILA, Taipei, China) were placed at 20 mm from the bottom of each tube. First, the sample was heated to 50 • C and kept at this temperature for more than 10 min. Then, the thermal bath temperature was cooled down to −5 • C at a rate of 1 • C/min. After 10 min at −5 • C, the oil was heated to 50 • C again. The two thermocouples were connected to a data logger (34970A, Agilent, Santa Clara, CA, USA), which recorded the data of the temperature variations for each sample.

Characterization
A differential scanning calorimeter (DSC, Q20, TA Instruments, New Castle, DE, USA) was used to analyze the phase change characteristics of the pristine eutectic, and the composite PCMs. The DSC tests were carried out under the atmosphere of nitrogen at a flow rate of 50 mL per minute. The testing temperature rose up from −5 to 50 °C at a rate of 3 °C /min.
Under the nitrogen atmosphere (flowrate: 100 mL/min), a thermoanalyzer (STA409PC, Netzsch, Waldkraiburg, Germany) tested the mass change of each sample during the heating process from 25 °C to 600 °C at a rate of 10 °C per minute. Thermal gravimetric analysis (TGA) characterizes the thermal stability of those materials.
The thermal conductivity of the materials in solid state were measured with a thermal constant analyzer (TPS 2500, Hot Disk, Gothenburg, Sweden). The composite PCM was compressed into two cylinders 40 mm in diameter and 10 mm in height. The sensor (type: 7577, 2.01 mm in diameter, Hot Disk, Gothenburg, Sweden) was sandwiched between the two cylinders. The thermal conductivity of the EP and eutectic were also measured, respectively.
The supercooling degree of the PCMs was characterized by the cooling curves. The schematic diagram is shown in Figure 1. Two tubes that contained the eutectic and the composite PCM composites were put into a thermal bath in which the temperature was controlled by a thermal oil (Phenyl silicone oil, flash point: 315 °C , Clearco Products Co., Inc., Willow Grove, PA, USA). Two Ktype thermocouples (NR-81530, accuracy: ±0.5 °C , HILA, Taipei, China) were placed at 20 mm from the bottom of each tube. First, the sample was heated to 50 °C and kept at this temperature for more than 10 min. Then, the thermal bath temperature was cooled down to −5 °C at a rate of 1 °C/min. After 10 minutes at −5°C , the oil was heated to 50 °C again. The two thermocouples were connected to a data logger (34970A, Agilent, Santa Clara, CA, USA), which recorded the data of the temperature variations for each sample. To determine whether the composite PCM was thermally reliable, the composite PCMs underwent 500 heating and cooling thermal cycles. In each cycle, the composite PCM was cycled between 50 °C and −5 °C via a temperature-controlled instrument (BPHJS-060A, Bluepard, Shanghai, China). In each stage of the thermal cycling, the composite PCM sample was maintained at the specified temperature for 30 min to ensure that the phase change completed. The results from SEM, Fourier transform infrared spectra, and DSC were compared before and after the thermal cycling.
The X-ray diffractometer (XRD, D8-ADVANCE, Bruker, Billerica, MA, USA, Cu Kα radiation λ = 1.5406 Å ) collected the diffraction patterns of CaCl2•6H2O, MgCl2•6H2O, and the eutectics in the 2θ ranging from 5° to 80°. The FT-IR of the pristine salts and the composite PCMs whose wavenumber ranged from 4000-400 cm −1 were obtained with the spectrophotometer (Tensor 27, Bruker, Billerica, To determine whether the composite PCM was thermally reliable, the composite PCMs underwent 500 heating and cooling thermal cycles. In each cycle, the composite PCM was cycled between 50 • C and −5 • C via a temperature-controlled instrument (BPHJS-060A, Bluepard, Shanghai, China). In each stage of the thermal cycling, the composite PCM sample was maintained at the specified temperature for 30 min to ensure that the phase change completed. The results from SEM, Fourier transform infrared spectra, and DSC were compared before and after the thermal cycling.
The X-ray diffractometer (XRD, D8-ADVANCE, Bruker, Billerica, MA, USA, Cu Kα radiation λ = 1.5406 Å) collected the diffraction patterns of CaCl 2 ·6H 2 O, MgCl 2 ·6H 2 O, and the eutectics in the 2θ ranging from 5 • to 80 • . The FT-IR of the pristine salts and the composite PCMs whose wavenumber ranged from 4000-400 cm −1 were obtained with the spectrophotometer (Tensor 27, Bruker, Billerica, MA, USA) at room temperature. The microstructures of EP and the CaCl 2 ·6H 2 O-MgCl 2 ·6H 2 O/EP composite PCM were observed using a scanning electron microscope (SEM, Quanta, FEI, Hillsboro, OR, USA).  Table 1. The increase in the mass fraction of MgCl 2 ·6H 2 O leads to a decrease in the melting point and specific phase change enthalpy. Note that the suitable phase change temperature of the PCM applied into the walls of buildings has to be between 22 • C and 26 • C. Moreover, the peak temperature of the mixture containing 15% of MgCl 2 ·6H 2 O is lower than that of the one containing 10%. Therefore, the mixture containing 15% of MgCl 2 ·6H 2 O was selected as a suitable PCM for use in buildings, exhibiting a latent heat value of 151.9 kJ/kg.  Figure 2 shows the DSC curves of the CaCl2•6H2O-MgCl2•6H2O eutectic mixtures with different mass fractions of MgCl2·6H2O; the details are listed in Table 1. The increase in the mass fraction of MgCl2·6H2O leads to a decrease in the melting point and specific phase change enthalpy. Note that the suitable phase change temperature of the PCM applied into the walls of buildings has to be between 22 °C and 26 °C. Moreover, the peak temperature of the mixture containing 15% of MgCl2·6H2O is lower than that of the one containing 10%. Therefore, the mixture containing 15% of MgCl2·6H2O was selected as a suitable PCM for use in buildings, exhibiting a latent heat value of 151.9 kJ/kg.   Figure 3a. The CaCl2•6H2O sample has a melting point of 28.3 °C , slightly higher than the comfortable temperature range for human beings. Significantly, the addition of MgCl2•6H2O with a phase change temperature of around 117 °C makes the obtained CaCl2•6H2O-MgCl2•6H2O mixture exhibit a melting point of 23.9 °C , lower than that of the CaCl2•6H2O sample. Consequently, the CaCl2•6H2O-MgCl2•6H2O is more suitable for use in walls of buildings for energy conservation. Moreover, Figure 4b    similar XRD pattern to that of the CaCl2•6H2O sample; no diffraction peaks of MgCl2•6H2O are observed, due to its low content in this mixture. Before the mixture containing 15 wt.% of MgCl2•6H2O is employed for practical applications, its supercooling degree needs to be evaluated. Consequently, its cooling curve has been measured by the T-history method. As displayed in Figure 4, no obvious supercooling is found for the CaCl2•6H2O-MgCl2•6H2O mixture, which is attributed to the addition of SrCl2•6H2O as the nucleating agent. With the help of MgCl2·6H2O and SrCl2·6H2O, the suitable melting point, along with low supercooling degree, is achieved.

Characteristics of the Composite PCM
The disadvantages of phase separation and liquid leakage for the eutectic mixture have been overcome by combining it with EP to prepare a composite PCM. In order to determine the adsorptive capacity of EP for the eutectic mixture, the liquid leakage tests have been conducted to the four composites containing different mass fractions of the eutectic. As shown in Figure 5, the composite containing 50% of the eutectic exhibits no liquid leakage, while some leakage is found on the filter of the one with PCM mass fraction of 55%. Therefore, the suitable mass fraction of the eutectic PCM Before the mixture containing 15 wt.% of MgCl 2 ·6H 2 O is employed for practical applications, its supercooling degree needs to be evaluated. Consequently, its cooling curve has been measured by the T-history method. As displayed in Figure 4, no obvious supercooling is found for the CaCl 2 ·6H 2 O-MgCl 2 ·6H 2 O mixture, which is attributed to the addition of SrCl 2 ·6H 2 O as the nucleating agent. With the help of MgCl 2 ·6H 2 O and SrCl 2 ·6H 2 O, the suitable melting point, along with low supercooling degree, is achieved. similar XRD pattern to that of the CaCl2•6H2O sample; no diffraction peaks of MgCl2•6H2O are observed, due to its low content in this mixture. Before the mixture containing 15 wt.% of MgCl2•6H2O is employed for practical applications, its supercooling degree needs to be evaluated. Consequently, its cooling curve has been measured by the T-history method. As displayed in Figure 4, no obvious supercooling is found for the CaCl2•6H2O-MgCl2•6H2O mixture, which is attributed to the addition of SrCl2•6H2O as the nucleating agent. With the help of MgCl2·6H2O and SrCl2·6H2O, the suitable melting point, along with low supercooling degree, is achieved.

Characteristics of the Composite PCM
The disadvantages of phase separation and liquid leakage for the eutectic mixture have been overcome by combining it with EP to prepare a composite PCM. In order to determine the adsorptive capacity of EP for the eutectic mixture, the liquid leakage tests have been conducted to the four composites containing different mass fractions of the eutectic. As shown in Figure 5, the composite containing 50% of the eutectic exhibits no liquid leakage, while some leakage is found on the filter of the one with PCM mass fraction of 55%. Therefore, the suitable mass fraction of the eutectic PCM

Characteristics of the Composite PCM
The disadvantages of phase separation and liquid leakage for the eutectic mixture have been overcome by combining it with EP to prepare a composite PCM. In order to determine the adsorptive capacity of EP for the eutectic mixture, the liquid leakage tests have been conducted to the four composites containing different mass fractions of the eutectic. As shown in Figure 5, the composite containing 50% of the eutectic exhibits no liquid leakage, while some leakage is found on the filter of the one with PCM mass fraction of 55%. Therefore, the suitable mass fraction of the eutectic PCM adsorbed into EP is determined to be around 50%, and the corresponding sample is thus taken as the CaCl 2 ·6H 2 O-MgCl 2 ·6H 2 O/EP composite PCM. adsorbed into EP is determined to be around 50%, and the corresponding sample is thus taken as the CaCl2•6H2O-MgCl2•6H2O/EP composite PCM. The microstructure of the composite PCM containing 50% CaCl2•6H2O-MgCl2•6H2O has been observed by SEM, together with that of EP. As observed from Figure 6a, EP consists of thin sheets and pores, thereby possessing good adsorptive capacity. EP has been shown to be hydrophilic, thus having a good compatibility with hydrated salts. As shown in Figure 6b, the adsorbed hydrated salt is uniformly attached to the sheets of EP and filled into the pores. The microstructure of the composite PCM containing 50% CaCl 2 ·6H 2 O-MgCl 2 ·6H 2 O has been observed by SEM, together with that of EP. As observed from Figure 6a, EP consists of thin sheets and pores, thereby possessing good adsorptive capacity. EP has been shown to be hydrophilic, thus having a good compatibility with hydrated salts. As shown in Figure 6b, the adsorbed hydrated salt is uniformly attached to the sheets of EP and filled into the pores. adsorbed into EP is determined to be around 50%, and the corresponding sample is thus taken as the CaCl2•6H2O-MgCl2•6H2O/EP composite PCM. Figure 5. Photographs of the composite PCM containing 50% (a) and 55% (b) of the PCM, along with their corresponding filter paper.
The microstructure of the composite PCM containing 50% CaCl2•6H2O-MgCl2•6H2O has been observed by SEM, together with that of EP. As observed from Figure 6a, EP consists of thin sheets and pores, thereby possessing good adsorptive capacity. EP has been shown to be hydrophilic, thus having a good compatibility with hydrated salts. As shown in Figure 6b, the adsorbed hydrated salt is uniformly attached to the sheets of EP and filled into the pores.  The thermal gravimetric analysis curves shows that the onset decomposition temperature of the composite PCM is slightly higher that of the pristine eutectic mixture in Figure 7b, which implies that the EP improves the stability of the eutectic mixture. The weight loss of the composite PCM is equivalent to the mass fraction of the eutectic mixture in it.  Figure 8a shows the DSC curve of the composite PCM, along with that of the CaCl2•6H2O-MgCl2•6H2O mixture for comparison. The melting point of the composite PCM is 23.5 °C , very close to the CaCl2•6H2O-MgCl2•6H2O mixture, suggesting that the composite PCM can be applied into the walls of buildings for energy conversation. The specific phase change enthalpy of the composite PCM is 73.55 J/g, which agrees with the eutectic mass fraction in the composite PCM. Thanks to the low thermal conductivity of the EP, the thermal conductivity of the composite PCM is far lower than that of the eutectic, as shown in Figure 8b. Both the relatively high specific phase change enthalpy and the thermal conductivity guarantees that the composite PCM is suitable for use in energy-saving efforts for buildings.  The thermal gravimetric analysis curves shows that the onset decomposition temperature of the composite PCM is slightly higher that of the pristine eutectic mixture in Figure 7b, which implies that the EP improves the stability of the eutectic mixture. The weight loss of the composite PCM is equivalent to the mass fraction of the eutectic mixture in it.  The thermal gravimetric analysis curves shows that the onset decomposition temperature of the composite PCM is slightly higher that of the pristine eutectic mixture in Figure 7b, which implies that the EP improves the stability of the eutectic mixture. The weight loss of the composite PCM is equivalent to the mass fraction of the eutectic mixture in it.  Figure 8a shows the DSC curve of the composite PCM, along with that of the CaCl2·6H2O-MgCl2·6H2O mixture for comparison. The melting point of the composite PCM is 23.5 °C, very close to the CaCl2·6H2O-MgCl2·6H2O mixture, suggesting that the composite PCM can be applied into the walls of buildings for energy conversation. The specific phase change enthalpy of the composite PCM is 73.55 J/g, which agrees with the eutectic mass fraction in the composite PCM. Thanks to the low thermal conductivity of the EP, the thermal conductivity of the composite PCM is far lower than that of the eutectic, as shown in Figure 8b. Both the relatively high specific phase change enthalpy and the thermal conductivity guarantees that the composite PCM is suitable for use in energy-saving efforts for buildings.  Figure 8a shows the DSC curve of the composite PCM, along with that of the CaCl 2 ·6H 2 O-MgCl 2 ·6H 2 O mixture for comparison. The melting point of the composite PCM is 23.5 • C, very close to the CaCl 2 ·6H 2 O-MgCl 2 ·6H 2 O mixture, suggesting that the composite PCM can be applied into the walls of buildings for energy conversation. The specific phase change enthalpy of the composite PCM is 73.55 J/g, which agrees with the eutectic mass fraction in the composite PCM. Thanks to the low thermal conductivity of the EP, the thermal conductivity of the composite PCM is far lower than that of the eutectic, as shown in Figure 8b. Both the relatively high specific phase change enthalpy and the thermal conductivity guarantees that the composite PCM is suitable for use in energy-saving efforts for buildings. Materials 2018, 11, x FOR PEER REVIEW 8 of 11 Figure 8. DSC data (a) and thermal conductivity values (b) of the materials.

Thermal Reliability of the Composite PCM
Thermal reliability is critical to PCMs in practical applications. After experiencing 500 heatingcooling cycles, the physical properties of the composite PCM are compared with the sample without cycling. We find no obvious changes in the FT-IR spectra (Figure 9a) of the composite PCM before and after the heating-cooling cycling. These results suggest that composite PCM possess good structure stability. Moreover, the DSC curves of the composite PCM before and after the heatingcooling cycling almost overlaps, as shown in Figure 9b, indicating its excellent thermal reliability.

Thermal Reliability of the Composite PCM
Thermal reliability is critical to PCMs in practical applications. After experiencing 500 heatingcooling cycles, the physical properties of the composite PCM are compared with the sample without cycling. We find no obvious changes in the FT-IR spectra (Figure 9a) of the composite PCM before and after the heating-cooling cycling. These results suggest that composite PCM possess good structure stability. Moreover, the DSC curves of the composite PCM before and after the heating-cooling cycling almost overlaps, as shown in Figure 9b, indicating its excellent thermal reliability. Materials 2018, 11, x FOR PEER REVIEW 9 of 11 Figure 9. FT-IR spectra (a) and DSC curves (b) of the CaCl2•6H2O-MgCl2•6H2O/EP composite PCM that experiences (or does not) the heating-cooling cycling test.

Conclusions
MgCl2•6H2O has been mixed with CaCl2•6H2O at different mass fractions to obtain a series of eutectic mixtures. It was found that the increase in mass fraction of MgCl2•6H2O leads to reductions in melting point and latent heat. The mixture containing 15% of MgCl2·6H2O exhibits a melting point of 23.9 °C , latent heat of 151.9 J/g, and little supecooling, making it a suitable PCM for use in walls of buildings. The combination with expanded perlite has been employed for overcoming the phase separation and liquid leakage of the CaCl2•6H2O-MgCl2•6H2O mixture. The adsorptive capacity of EP for the mixture was determined to be around 50 wt.%, and the corresponding composite PCM shows a melting point of 23.5 °C and latent heat of 73.55 J/g. The thermal conductivity of the CaCl2•6H2O-MgCl2•6H2O/EP composite PCM is just slightly higher than that of EP, remarkably lower than that of

Conclusions
MgCl 2 ·6H 2 O has been mixed with CaCl 2 ·6H 2 O at different mass fractions to obtain a series of eutectic mixtures. It was found that the increase in mass fraction of MgCl 2 ·6H 2 O leads to reductions in melting point and latent heat. The mixture containing 15% of MgCl 2 ·6H 2 O exhibits a melting point of 23.9 • C, latent heat of 151.9 J/g, and little supecooling, making it a suitable PCM for use in walls of buildings. The combination with expanded perlite has been employed for overcoming the phase separation and liquid leakage of the CaCl 2 ·6H 2 O-MgCl 2 ·6H 2 O mixture. The adsorptive capacity of EP for the mixture was determined to be around 50 wt.%, and the corresponding composite PCM shows a melting point of 23.5 • C and latent heat of 73.55 J/g. The thermal conductivity of the CaCl 2 ·6H 2 O-MgCl 2 ·6H 2 O/EP composite PCM is just slightly higher than that of EP, remarkably lower than that of the CaCl 2 ·6H 2 O-MgCl 2 ·6H 2 O mixture. The interaction between EP and the CaCl 2 ·6H 2 O-MgCl 2 ·6H 2 O mixture has been verified to be physical. The heating-cooling cycling test reveals that the composite PCM possesses good thermal reliability. Suitable phase change temperature, little supercooling, no phase separation and liquid leakage, low thermal conductivity, and good thermal reliability make the CaCl 2 ·6H 2 O-MgCl 2 ·6H 2 O/EP composite PCM show great promise for use in walls of buildings. The application of this composite PCM in buildings is under investigation.