Introduction

The energy crisis is a commonly discussed topic, and the direction to solve such problems is the development of renewable new energy, such as solar energy, geothermal energy, wind energy, etc.; the other direction is the exploration of new energy storage media, which can effectively solve the problem of uneven distribution of energy, and can greatly improve the efficiency of energy utilization [1]. Recently, thermal energy storage technology has received more attention due to the fact that thermal energy is the most abundant type of energy in nature. Generally, there are sensible thermal energy storage, latent thermal energy storage and thermochemical energy storage. Among the three thermal energy storage, latent thermal energy storage is highly regarded for its high energy storage density and wide temperature range [2]. This technology uses the change of physical state of the working material to achieve heat storage and release, so it can be divided into four ways: solid–solid, solid–liquid, solid–gas and liquid–gas [3]. The latter two are not suitable for practical applications because of the large volume changes that would take place, so only the first two have attracted interest. In particular, solid–liquid phase change materials (PCMs) are widely used in buildings [4, 5], thermal management [6] and heat recovery systems [7].

There are two main categories of solid–liquid PCMs, namely organic PCMs and inorganic PCMs [8]. Compared with inorganic PCMs, organic PCMs have the merits of low supercooling, no phase separation, good chemical stability, and non-corrosiveness, so they are widely used in current research, such as paraffin [9], fatty acids [10] and alcohols [11]. Since fatty acid PCM possess high phase change enthalpy, self-nucleation [12], non-toxicity, and low cost [13], they are proposed to be potential PCM. Nevertheless, solid–liquid organic PCMs have unavoidable defects, one is the tendency to leak in the molten state and the other is the low heat transfer rate, and these defects greatly suppress their applications. To overcome these drawbacks, combining support material with PCM to form shape-stabilized PCM is an effective strategy [14].

In general, the choice of matrix material directly affects the thermal properties of the PCM. High-density polyethylene (HDPE), polystyrene (PS), SiO2, diatomite, expanded perlite (EP), etc. have been widely considered as matrix materials [15,16,17]. For example, Rahman et al. [18] fabricated paraffin/HDPE shape-stabilized PCM. Ke et al. [19] synthesized a polystyrene-based shape-stabilized PCM with a phase transition temperature tunable to a lower temperature range. Baskar et al. [20] synthesized a composite PCM using eutectic lauric and palmitic acids as the PCM and nano-SiO2 as the support material. The melting enthalpy of the composites reaches 187 J/g, and the thermal conductivity value is increased by 54.4%. Benayache et al. [21] prepared a paraffin/diatomite shape-stabilized PCM by impregnating paraffin into diatomite. This composite exhibits excellent thermal reliability and thermal storage performance. Although the polymer matrix and the clay mineral matrix are able to eliminate leakage well, the thermal conductivity of the composite is not significantly improved.

In contrast, nanometal oxides [22, 23] and hexagonal boron nitride (h-BN) [24, 25] have not only excellent thermal stability but also excellent thermal conductivity, so their addition can both prevent melt leakage and significantly improve the thermal conductivity of PCM. Samiyammal et al. [26] reported that the thermal conductivity of paraffin-based composites is enhanced by the incorporation of magnesium oxide nanoparticles. Xie et al. [27] fabricated composite PCM by stripping h-BN from the native material and using it as a support material. The melting enthalpy of the composite PCM is 136.2 J/g and thermal conductivity is improved to 0.453 W/(m·K). Fang et al. [28] synthesized composite PCM composed of paraffin and h-BN. The results indicate that 10 wt% of h-BN is able to improve the thermal conductivity of composites by 60%. Accordingly, MgO and BN can be served as support materials and also as thermal conductivity enhancers for PCM.

The purpose of this work is to prepare a shape-stabilized PCM that has no significant decrease in latent heat and a significant improvement in thermal stability and thermal conductivity. MA was adopted as a PCM because of its advantages such as high latent heat and its availability for solar water heating systems. Combinations of nano-MgO and h-BN were acted as both matrix materials and thermal conductivity enhancers. Shape-stabilized PCMs were prepared by the melt blending method. The composite PCMs were tested for leakage and their structural characteristics and surface morphology were analyzed. Thermophysical properties of composite PCMs were also determined to evaluate their potential for thermal storage applications.

Results and discussion

Leakage testing analysis

The leakage of the samples can be evaluated by subjecting them to a leakage test. The specific experimental scheme is as follows: place the sample on a clean white paper and place it on the heater with the temperature maintained at 65 °C. After 15 min, the sample is removed. The area of oil stains on the white paper reflects the degree of leakage of the sample. Figure 1 shows nine samples and their leakage test results. It is clear that when the level of h-BN is certain, the area of the oil stain is gradually reduced with the addition of MgO content. When the MgO content is certain, the area of the oil stains gradually decreases with increasing h-BN content. These results indicate that the addition of MgO and h-BN can effectively prevent the leakage of molten MA, and both serve as support materials. In addition, the C6-C9 samples have almost no leakage, so they will be further tested and analyzed.

Figure 1
figure 1

MA/MgO/h-BN composite PCMs and their leakage test results.

Full size image

FT-IR analysis

The FT-IR spectra of MA, MgO, h-BN and C6-C9 are shown in Fig. 2(a). Four representative peaks at 2917, 2850, 1700, and 721 cm−1 can be observed in the spectrum of MA. Among them, the peaks at 2917 and 2850 cm−1 indicate the anti-symmetric and symmetric stretching vibrations of the -CH2 group; the peak at 1700 cm−1 represents the stretching vibrations of the C=O group; and the peak at 721 cm−1 reflects the presence of four or more –CH2 groups in the molecular chain [29]. The spectrum of MgO shows that it does not have a more pronounced characteristic absorption peak at 3500–700 cm−1. The pattern of h-BN shows the characteristic peaks at 1391 and 807 cm−1, which are the bending and stretching vibrations of the B-N bond [27]. The C6-C9 spectra indicate that their absorption peak positions correspond to those of MA and h-BN one by one, and no new characteristic peaks appear, confirming that no chemical reaction occurs during the preparation of the composites. Therefore, the combination of MA, MgO and h-BN is only a physical interaction.

Figure 2
figure 2

The spectra of MA, MgO, h-BN and C6-C9. (a) FT-IR spectra. (b) XRD spectra.

Full size image

XRD analysis

The XRD spectra of MA, MgO, h-BN and C6-C9 are shown in Fig. 2(b), reflecting the corresponding crystal structures. It can be seen that the diffraction curve of MA has a sharp strong peak at 21.6°, a secondary strong peak at 24.4° and a weak peak at 20.3°, which reflects the regular crystallization of MA [29]. The diffraction peak of MgO is located at 42.8°. The h-BN shows a sharp characteristic peak at 26.6°, which is attributed to the typical (0 0 2) crystal face [27]. As observed from the diffraction curves of C6-C9, the position of the characteristic peak of MA remains basically the same and its intensity is significantly weakened, but the peak at 24.1° evolves into a dispersion peak. In addition, the characteristic peaks of MgO and h-BN are also present in C6-C9, and only the peak intensity has been decreased. Furthermore, there are no diffraction peaks other than MA, h-BN and MgO, indicating that the XRD spectra of C6-C9 are essentially composed of the spectra of MA, MgO and h-BN. Therefore, the mixing of MA, MgO and h-BN does not yield new materials.

Microstructure analysis

The microscopic morphologies of MA, MgO, h-BN and C6-C9 are presented in Fig. 3. As observed in Fig. 3(a), MgO appears as smaller particles with more voids. In Fig. 3(b), h-BN is a two-dimensional lamellar structure. Figure 3(c)–(f) correspond to the microstructures of C6-C9. MgO and h-BN show a more uniform dispersion and MA is adsorbed in the voids. Due to the surface tension and capillary force, the bond between MA, MgO and h-BN is stable and can effectively restrict the leakage of liquid phase MA.

Figure 3
figure 3

SEM images. (a) MgO. (b) h-BN. (c)–(f) C6-C9.

Full size image

Thermal performance analysis

The DSC curves of MA and C6-C9 are shown in Fig. 4, where Fig. 4(a) shows the melting process and Fig. 4(b) shows the solidification process. The relevant measurements obtained from these samples are shown in Table 1 in terms of melting/solidification temperature (Tm/Ts), peak temperature (Tpm/Tps), and latent heat of melting/solidification (ΔHm/ΔHs). In Fig. 4(a), pure MA has only one sharp heat absorption peak, which denotes the solid–liquid transition of MA with a latent heat and melting point of 197.50 J/g and 54.83 °C, respectively. However, the melting process of C6-C9 is significantly different from that of pure MA in that they have two heat absorption peaks. The melting points of the first heat absorption peaks of C6-C9 are 51.34, 51.90, 51.23 and 51.14 °C, respectively, which are about 3 °C lower than the melting point of pure MA. The first heat absorption peaks are similar to the absorption peak of pure MA, but they are significantly smaller and the shoulders become slightly broader. The melting points of the second heat absorption peaks of C6-C9 are 70.58, 70.22, 70.87 and 66.77 °C, respectively, which are about 16 °C higher than the melting point of pure MA. The second heat absorption peaks are smoother and their shoulders are much broader. In Fig. 4(b), pure MA also has only one sharp exothermic peak with a latent heat and solidification point of 197.30 J/g and 51.93 °C, respectively. The exothermic peaks of C6-C9 are similar to the exothermic peak of pure MA, but their peaks are obviously smaller and their shoulders become broader. The solidification temperatures of C6-C9 are 40.34, 39.86, 41.01 and 43.60 °C, respectively, which are about 10 °C lower than the solidification temperature of pure MA.

Figure 4
figure 4

DSC curves of MA and C6-C9. (a) Melting process. (b) Solidification process.

Full size image
TABLE 1 DSC results of MA and C6-C9.
Full size table

Composite PCMs exhibit two heat absorption peaks during melting, which are also observed in the works of Li [30] and Yang [31]. In this study, the first heat absorption peak of the composite PCMs is consistent with the usual situation, while the temperature of the second heat absorption peak is significantly elevated. The appearance of this anomalous heat absorption peak can be regarded as a result of superimposed forces based on the Young–Laplace equation. These forces include C-H…π interactions [32], surface tension [30], hydrogen bonding, and van der Waals forces [33]. For composite PCMs, there are large interactions and pressure superposition between the MgO matrix and the confined MA, and this superimposed pressure in the confined space leads to an elevation of the phase transition temperature.

The total melting latent heats of C6-C9 are 140.45, 107.88, 151.41 and 46.45 J/g, respectively. Dividing the total melting latent heat by that of pure MA, the mass fractions of MA in composite PCM are 71%, 55%, 77% and 24%, respectively. As for the formulation of MA/MgO/h-BN, the mass fractions of MA are 78%, 82%, 78% and 74%, respectively. It is clear that the latent heat drop in C9 is large, and this reduction may be caused by the presence of non-melting layers in the composite [34]. Therefore, among the obtained composites, C6 and C8 have more satisfactory high latent heat of phase change.

Thermal stability analysis

Figure 5(a) and (b) show the TGA and DTG curves of MA, MgO, h-BN and C6-C9, respectively, which can reflect the thermal stability of these samples. Table 2 presents the key data of the thermal decomposition process with the peak temperature (Tp) and the specific weight of mass loss (ΔW), as well as the mass residue at 700 °C. It can be seen that MA has an obvious one-step weight loss process in Fig. 5(a), which is more stable below 150 °C. The thermal decomposition rate reaches the maximum at 234.61 °C, and the mass residue is only about 3.5%. MgO shows a very slightly weight loss at 30–300 °C, which can be considered as a moisture-induced loss. There is a significant weight loss at 300–400 °C, and the thermal decomposition rate reaches the maximum at 361.74 °C, with a final weight loss ratio of about 17.4%. The h-BN is very stable, and it almost does not undergo thermal degradation. Unlike the pure MA, the MA/MgO/h-BN composites all show a distinct two-step weight loss process, with two distinct peaks in the Fig. 5(b). The first step is between about 150–250 °C, which can be considered as the lesser MA decomposition process, and the maximum pyrolysis rate of this process is significantly lower than that of pure MA. The weight loss percentage of C6 is about 31%, while that of C7-C9 is much lower. The second step is about 300–500 °C. This process involves decomposition of a small amount of MgO and the remaining MA, and the weight loss ratio of C6 is about 47%, while C7-C9 is over 60%. The final pyrolysis mass residue belongs to MgO and h-BN. Compared with the MgO decomposition curve in Fig. 5(a), the second weight loss process of the composites corresponds to the weight loss process of MgO, indicating that the decomposition process is mainly caused by the strong limiting effect of MgO. Overall, the pyrolysis process of the samples confirms that the composites are almost stable below 150 °C, while the two-step thermal decomposition process can be attributed to the strong limiting effect of MgO, which leads to the secondary weight loss of the composites.

Figure 5
figure 5

Thermogravimetric curves of MA, MgO, h-BN and C6-C9. (a) TGA curve. (b) DTG curve.

Full size image
TABLE 2 Thermogravimetric data of MA, MgO and C6-C9.
Full size table

Thermal conductivity analysis

To evaluate the heat transfer efficiency of the composite PCMs, the thermal conductivity of the nine samples is measured. The thermal conductivity of pure MA is only 0.20 ± 0.01 W/(m·K), while the thermal conductivity of composite PCMs is increased by the addition of the support material. The thermal conductivities of C1-C9 composites are 0.29 ± 0.02, 0.37 ± 0.01, 0.47 ± 0.02, 0.32 ± 0.01, 0.41 ± 0.02, 0.52 ± 0.02, 0.37 ± 0.01, 0.44 ± 0.02 and 0.58 ± 0.03 W/(m·K). Compared with pure MA, the thermal conductivity is increased by 45%, 85%, 135%, 60%, 105%, 160%, 85%, 110% and 190%, respectively. In particular, the increase in thermal conductivity of C6-C9 is more than 85%. For a more visual comparison, the thermal conductivity data are plotted as histograms in Fig. 6. As can be seen, the thermal conductivity of composite PCMs is only slightly increased with the increase of MgO when the h-BN mass fraction is a certain amount. When the mass fraction of MgO is fixed, there is a large increase in the thermal conductivity of composite PCMs with increasing h-BN. In other words, h-BN can be more effective than MgO in the improvement of the thermal conductivity of composite PCMs.

Figure 6
figure 6

Thermal conductivity of C1-C9.

Full size image

In Table 3, the current study is compared with other literature [34,35,36,37,38]. The composite PCMs obtained in this experiment not only have high latent heat but also have good thermal conductivity. Therefore, the composite PCMs prepared in this work have potential applications.

TABLE 3 Comparison of the thermal conductivity between this work and other results.
Full size table

Conclusions

In this paper, a series of MA/MgO/h-BN composite PCMs were prepared by the melt blending method using MA as the PCM and combinations of nano-MgO and h-BN as the support material and thermal conductivity enhancer. Leakage test results show that molten MA leakage is eliminated when the total mass ratio of MgO and h-BN is about 20 wt%, and the higher the MgO content, the less leakage occurs. FT-IR and XRD results indicate that the components of composite PCMs are combined with each other in a physical interaction. SEM results suggest that MA is filled into the voids of MgO and h-BN. Thermal analysis shows two heat absorption peaks for the melting process of C6-C9, which is attributed to the presence of a large superposition pressure between the confined MA and the MgO/BN matrix, leading to a significant increase in the phase transition temperature. Thermogravimetric experiments indicate that the composite PCMs are almost stable below 150 °C. The second weight loss process reflects the limiting effect of the MgO matrix on MA due to the fact that large amounts of MA in the composite PCMs are pyrolyzed in the second weight loss process, whereas originally all MA should have been pyrolyzed in the first weight loss process. Thermal conductivity measurements show that the thermal conductivity of the composite PCM improves with increasing h-BN when the mass fraction of MgO is fixed. Although both MgO and h-BN have the effect of improving the thermal conductivity, the thermal conductivity improved by adding h-BN is significantly better than that of MgO. Therefore, in the prepared MA/MgO/h-BN composite PCMs, both MgO and h-BN are available as matrix materials and thermal conductivity enhancers, but their effects are quite different. MgO mainly affects the phase transition temperature and latent heat of the composite PCMs, while h-BN mainly affects the thermal conductivity of the composite PCMs. In summary, C6 has good thermal properties. Its melting latent heat is 140.45 J/g and its thermal conductivity is 0.52 W/(m·K). The thermal conductivity of C8 is 0.44 W/(m·K), which is slightly lower than that of C6. However, its melting latent heat is higher than that of C6, which is 151.41 J/g.

Experimental methods

Materials

Myristic acid (MA, C14H28O2, molecular weight: 228.37, content ≥ 98.0%, chemically pure) was sourced from Sinopharm Chemical Reagent Co., Ltd. It shows a white blade-like crystalline powder with a melting point of approximately 52.5 to 54.5 °C. Magnesium oxide nanoparticles (MgO, purity: 98%, specific surface area (SSA): 42.05 m2/g) and high-purity hexagonal nitride boron (h-BN, purity ≥ 99.5%) nanoplates were provided from Jiangsu XFNANO Materials Tech Co., Ltd. MgO is a white powder with a particle size of about 20 nm. h-BN is also a white powder with a sheet length of about 0.3–2 µm.

Preparation of MA/MgO/h-BN composites

In this work, MA/MgO/h-BN composite PCMs were prepared by melt blending technique. The weights of the components in these composites are listed in Table 4. The detailed preparation procedures for the MA/MgO/h-BN composite PCM are as follows: first, each substance is weighed according to the masses assigned in Table 4 and put into a beaker. Then, the beaker containing MA is placed on a heater and the temperature is maintained at 60 °C so that the MA is completely melted; when the MA is completely melted, the corresponding h-BN is added and stirred at 1000 rpm for 10 min by a cantilevered mechanical stirrer; next, the corresponding MgO is gradually added and stirred at 1000 rpm for 1 h to ensure uniform dispersion. Finally, the molten composite is poured into a glass dish and allowed to cool naturally. The above preparation process is repeated to obtain nine products, which are labeled in sequence as C1, C2, C3, C4, C5, C6, C7, C8 and C9, as shown in Fig. 1.

TABLE 4 Constitution of MA/MgO/BN composites.
Full size table

Instruments and characterization techniques

Structural characterization

Fourier transform infrared (FTIR, Nicolet Nexus 870) spectroscopy is adopted to measure the chemical groups of the composites. Wavenumbers range from 4000 to 400 cm−1 with 2 cm−1 resolution. X-ray diffraction (XRD, D/MAX-Ultima III) is employed to characterize the crystal structure of the composites. The working parameters of XRD are: scanning speed 5° (2θ)/min, voltage 40 kV, and current 40 mA. Ultra-high resolution scanning electron microscope (SEM, Gemini 500, Carl Zeiss, Germany) is utilized to observe the microscopic morphology of composites.

Thermal behavior characterization

Differential scanning calorimetry (DSC) is used to determine the relevant parameters (phase change point and phase change enthalpy) of composite PCM in the phase change process with an accuracy of ± 0.2 °C for temperature and ± 5% for enthalpy. The operating temperature and heating/cooling rate of the DSC are set to 10–100 °C and 5 °C /min, respectively. A thermogravimetric analyzer (TGA, Perkin-Elmer Pyris 1) is implemented to evaluate the thermal stability of the composites. The working temperature and heating rate of the TGA were configured to 30–700 °C and 20 °C/min, respectively. The TC3020 transient hot-wire conductivity meter is employed to measure the thermal conductivity of the composites with an accuracy of ± 2%.