Polyethylene glycol/Cu/SiO₂ form stable composite phase change materials: preparation, characterization, and thermal conductivity enhancement

Abstract

Novel form-stable composite phase change materials (FS-CPCMs) of polyethylene glycol (PEG)/Cu/SiO₂ were prepared by adding Cu powder to PEG and SiO₂ via the ultrasound-assisted sol–gel method. This method ensured the uniform distribution of Cu powder in the FS-CPCMs, thus providing an important method to develop composite phase change materials (CPCMs) with a high thermal conductivity. The FS-CPCMs were characterized by various techniques. The results showed that the FS-CPCMs remained in the solid state without leakage above the melting point of PEG. The XRD and FTIR results indicated that no new chemical bond was formed between the constituents of FS-CPCMs: Cu, PEG, and SiO₂. The DSC and TGA analyses showed that the FS-CPCMs had an optimum phase-change temperature, a high enthalpy of phase change, an excellent thermal stability, and a good form-stable performance. The thermal conductivity was 0.431 W m⁻¹ K⁻¹ for 3.45 wt% Cu powder in the FS-CPCMs, an increase of 49.13% compared to pure PEG.

---

Introduction

Phase change materials (PCMs) can store and release large amounts of energy by changing the temperature slightly because of a high energy storage density and the isothermal nature of the storage process.^1,2 In recent years, PCMs have been widely used in solar engineering, solar heating systems, building energy conservation, smart textiles, air-conditioning systems, heat pumps, waste heat recovery, and thermal insulation.^1,3–8

According to their chemical composition, PCMs are mainly divided into two types: inorganic and organic.^9,10 Compared to inorganic materials, a large number of organic materials have been studied because of their wide melting temperatures for convenient, slight supercooling, chemical and thermal stability, and moderate phase change enthalpies.^8–11 Among organic materials, polyethylene glycol (PEG) is undoubtedly an ideal choice because of its suitable phase change temperatures, high latent heat storage capacity, nontoxicity, and low-cost.^1,12 However, the main drawbacks of PEG are the phase instability in the melting state,¹³ low thermal conductivity,¹⁴ and weak interfacial combination with the supporting materials,¹⁵ limiting its further applications. To solve the main drawbacks of PEG, form-stable organic–inorganic hybrid PCMs mixed with materials of high thermal conductivity prepared via the ultrasound-assisted sol–gel method have become popular because they are easy to prepare and can be used directly without additional encapsulation.^16,17

Several studies have been carried out to improve the thermal conductivity of PCMs. Carbon-based nanostructures (nanofibers, nanoplatelets, and graphene flakes), metals (Ag, Fe, Au, and Cu), metal oxides (Al₂O₃, CuO, MgO, and TiO₂), and metallic nanowires (Ag, Au, and Cu) have been investigated as the materials to enhance the thermal conductivity.^18–24 Cu powder is inexpensive than metallic nanowires, graphene flakes, and Ag and Au powders. Moreover, Cu powder has a higher thermal conductivity (400 W m⁻¹ K⁻¹) than graphite, Fe powder, and metal oxides.^18,20,22 Therefore, Cu powder was selected to improve the thermal conductivity of PCMs. A comparison of the PEG/SiO₂–Al₂O₃ FS-PCMs reported in ref. 23, the tetraethoxylsilane and ultrasound-assisted sol–gel method were the same. However, there were more simple experiment process and higher thermal conductivity^20,22 in this study.

Fig. 1 shows the schematic of light-to-heat conversion, storage, and release. The ability of Cu powder to transfer the heat rapidly enhanced the use of solar energy in PCMs. During the preparation of FS-CPCMs, the long chains of PEG embed into the fluffy network structure of SiO₂, forming an interpenetrating network, crosslinking, and cladding structure;^25,26 thus, Cu powder can be distributed uniformly in the composite. Therefore, SiO₂ acts as the carrier matrix and can strengthen the interfacial combination with PEG, thus providing structural strength. The PEG in the FS-CPCMs could not easily leak from the fluffy network structure of SiO₂ during the solid–liquid phase transition,⁷ and the FS-CPCMs exhibited a high conductance.

Fig. 1 Schematic of light-to-heat conversion, storage, and release.

The resulting FS-CPCMs are potential candidates for applications in the fields of cooling/heating of buildings, such as PCM walls and wallboards, trombe walls, shutters, tiles, building blocks, and air-based heating systems,^{3,5,6,8,27,28} whose temperature is often as high as 40–60 °C.

Experimental section

Materials

Tetraethoxylsilane (TEOS), Cu powder of 625 mesh, PEG with an average molecular weight of 4000, distilled water, HCl solution, and absolute ethanol were used in this experiment. All the chemical reagents were analytically pure and purchased from Beijing Chemical Reagent Ltd.

Characterization

The X-ray diffraction analysis (XRD, Model XD-3) was carried out using Cu-Kα radiation. The scanning step size was 0.021, and the 2θ was varied from 10° to 70°. The chemical compatibilities of FS-CPCMs was analysed by Fourier transform infrared spectroscopy (FT-IR, Model Frontier). The morphology of SiO₂ and the prepared FS-CPCMs samples was observed using a scanning electron microscope (SEM, JSM-5610LV, and JEOL). The thermal property of the FS-CPCMs was determined using a differential scanning calorimeter (Q2000) calibrated with an indium standard and using a temperature program of 0–100–0 °C at 10 °C min⁻¹ heating/cooling rate. The thermal stabilities of the FS-CPCM samples and PEG were investigated by thermogravimetric analysis (TGA, Q50) in the range from 30 °C to 500 °C with a scanning rate of 10 °C min⁻¹ under nitrogen atmosphere. The thermal conductivity (K) was determined using the formula: K = α × C_p × ρ where C_p is the specific heat capacity, α is the thermal diffusivity, and ρ is the density. The thermal conductivity of the cylindrical die specimens with 10 mm in diameter and 1–2 mm in thickness was measured at 30 °C using a DRL-III tester.

Preparation of FS-CPCMs

Fig. 2 shows the schematic diagram of the preparation of FS-CPCMs. A certain amount of TEOS and distilled water were added to an appropriate amount of absolute ethanol. The mixture was subjected to ultrasound under 300 W power for 15 min at 70 °C to obtain a well-dispersed suspension A. PEG and Cu powder were magnetically stirred for 15 min at 70 °C to obtain a well-dispersed suspension B that was added dropwise to suspension A under magnetic stirring. At the same time, an aqueous solution of HCl (C_HCl = 0.5 mol L⁻¹, V_HCl = 0.8 mL) was added dropwise. A silica sol containing PEG and Cu powder was formed rapidly. Next, a silica gel containing PEG and Cu powder was obtaining after stirring for an appropriate time. Then, the gel was dried in an oven at 70 °C for 24 h, and the PEG/Cu/SiO₂ composites were cooled to room temperature. Thus, the PEG/Cu/SiO₂ composite phase change materials were prepared.

Fig. 2 Schematic diagram of the preparation of FS-PCMs.

In this study, the weights of Cu powder in FS-CPCMs-1, FS-CPCMs-2, FS-CPCMs-3, and FS-CPCMs-4 samples were 0 g, 1 g, 2 g, and 3 g, respectively. The mass fraction of Cu powder in the FS-CPCMs is shown in Table 1.

Table 1 Mass fraction of Cu powder in the FS-CPCMs

Samples	Weight of PEG/SiO2 (g)	Weight of Cu powder (g)	Cu powder loading (wt%)	PEG loading (wt%)
FS-CPCMs-1	28	0	0	80
FS-CPCMs-2	28	1	3.45	77.24
FS-CPCMs-3	28	2	6.67	74.67
FS-CPCMs-4	28	3	9.68	72.26

---

Results and discussion

FTIR characterization of FS-CPCMs

Fig. 3a shows the FTIR spectra of SiO₂, pristine PEG, and PEG/Cu/SiO₂ composites. In the SiO₂ spectrum, the peak at 472 cm⁻¹ can be attributed to the vibration of Si–O group; the peak at 793 cm⁻¹ can be assigned to the vibration of O–H group; the peak at ∼1108 cm⁻¹ can be attributed to the stretching vibration of Si–O–Si group. All these characteristic peaks can be assigned to SiO₂.^11,29,30 Typical pristine PEG bands are visible at 3419 cm⁻¹ (O–H stretching), 2881 cm⁻¹ and 950 cm⁻¹ (CH₂ stretching), and 1096 cm⁻¹ (C–O–C symmetrical stretching).^31,32

Fig. 3 FTIR spectra (a) and XRD (b) of SiO₂, pristine PEG, and PEG/Cu/SiO₂ composites.

Moreover, all the characteristic absorption peaks of SiO₂ and PEG also appeared in the spectrum of the composite PEG/SiO₂ and PEG/Cu/SiO₂ without the appearance of any new peaks. This indicates a good chemical compatibility between SiO₂ and PEG.

Crystalline and form-stable properties of FS-CPCMs

The XRD patterns of SiO₂, pristine PEG, and the PEG/Cu/SiO₂ composites are shown in Fig. 3b. The XRD pattern of SiO₂ shows a broad peak in the range 18–32°, indicating a typical noncrystalline structure.³³ The XRD patterns of PEG, PEG/SiO₂ and PEG/Cu/SiO₂ show two main similar diffraction peaks at approximately 19° and 23°. The XRD patterns of Cu and PEG/Cu/SiO₂ show two main similar diffraction peaks at approximately 42° and 50°. These results indicate that the PEG on the support was in the crystalline state, and the introduction of SiO₂ and Cu to PEG did not affect its crystal structure in the composite.

Comprehensive analysis of FTIR and XRD, the interactions in the composites involve simple physical loading between Cu and PEG/SiO₂, rather than chemical interactions.^11,23,25

The form-stable property of the FS-CPCMs was evaluated using the hot-stage-digital camera technology.^23,34 The FS-CPCMs and pristine PEG were pressed into wafers each with a diameter of 10 mm and a height of 2 mm. The wafers were placed on a hot stage at 80 °C for 2 h. The changes in shape were observed via tracking the photographs using the digital camera. The results are shown in Fig. 4. The pristine PEG melted, but no leakage of PEG was observed on the surfaces of the FS-CPCMs even when the temperature was higher than the melting point of PEG. PEG has long chains, and SiO₂ gel has an intercross-linked network.²³ The long chains of PEG can completely or partially interpenetrate the network.²⁶ The capillary force and surface tension induced by the SiO₂ gel network hindered the leakage of liquid PEG. Thus, the FS-CPCMs could retain its original shape after the phase transition despite no chemical interactions between SiO₂ and PEG.

Fig. 4 Photos of FS-CPCMs and PEG taken in a hot stage: (a) room temperature and (b) 80 °C.

Surface morphology of FS-CPCMs

The SEM images of SiO₂, FS-CPCMs-1, and FS-CPCMs-2 are shown in Fig. 5. The SiO₂ network structure (Fig. 5a and b) was filled with PEG (Fig. 5c) and Cu (Fig. 5d). The results show that PEG was well-dispersed throughout the solid SiO₂ used as the supporting material. Hence, the composites maintained their form-stable structure during the phase transition without any leakage of melted PEG (Fig. 4). Moreover, the energy spectrum analysis (Fig. 6)¹⁹ and energy dispersive spectrometer mapping (Fig. 7) within the scope of view shown in Fig. 5d supported the formation of FS-CPCMs-2. The elemental composition of Cu powder was confirmed to be 3.13%. The results clearly confirm that FS-CPCMs-2 contains Cu powder. Meanwhile, the distribution of Cu and SiO₂ in the FS-CPCMs-2 were uniform. The FTIR, XRD, SEM, and EDS analyses show that PEG/Cu/SiO₂ CPCMs with a good crystal performance were successfully prepared.

Fig. 5 SEM images of the samples: the surface of SiO₂ gel (a), the cross-section of SiO₂ gel (b), FS-CPCMs-1 (c), and FS-CPCMs-2 (d).

Fig. 6 Energy-dispersive X-ray spectrum of FS-CPCMs-2.

Fig. 7 Energy dispersive spectrometer mappings of Si (a) and Cu (b) of FS-CPCMs-2.

Enhancement of thermal conductivity

The results of thermal conductivity at different Cu mass fractions are shown in Fig. 8. The thermal conductivities of PEG and FS-CPCMs with 0 wt%, 3.45 wt%, 6.67 wt%, and 9.68 wt% of Cu were 0.289 W m⁻¹ K⁻¹, 0.387 W m⁻¹ K⁻¹, 0.431 W m⁻¹ K⁻¹, 0.478 W m⁻¹ K⁻¹, and 0.513 W m⁻¹ K⁻¹, respectively. The thermal conductivity increases with the Cu powder amounts. A relative enhancement of more than 49.13% in the thermal conductivity was observed in the FS-CPCMs containing 3.45 wt% of Cu powder. The results show that Cu powder was very effective in improving the thermal conductivity of the FS-CPCMs.

Fig. 8 Thermal conductivity of PEG and FS-CPCMs.

Melting and freezing temperature curves of FS-CPCMs

The increase in thermal transfer was also investigated by comparing the melting and solidifying processes of the FS-CPCMs with those of pure PEG.^35,36 First, 20 g pristine PEG, FS-CPCMs-1, FS-CPCMs-2, FS-CPCMs-3, and FS-CPCMs-4 were added to five test tubes. Thermometers with a temperature accuracy of ±0.1 °C were placed in the center of the five tubes. Two water baths were used to complete the melting and solidifying processes. First, the five test tubes were placed inside a water bath at 100 °C (until 90 °C) for complete melting, and then the five test tubes were immediately placed inside another water bath maintained at 20 °C until 24 °C for complete solidification. The temperature variations of pure PEG and FS-CPCMs during the melting and solidifying processes were measured and recorded every 30 s.

The results (Fig. 9a) show that to achieve the same temperature of 62 °C from 20 °C, the heating times were 1430 s for pure PEG, 1110 s for FS-CPCMs-1, 1020 s for FS-CPCMs-2, 870 s for FS-CPCMs-3, and 700 s for FS-CPCMs-4, i.e., the melting times of the FS-CPCMs with different mass fractions of Cu were 22.38%, 28.67%, 39.16%, and 51.05%, respectively, less than that of pristine PEG. In the solidifying process (Fig. 9b), to achieve the same temperature of 24 °C from 77 °C, the solidifying times were 2010 s for pure PEG, 1800 s for FS-CPCMs-1, 1470 s for FS-CPCMs-2, 1260 s for FS-CPCMs-3, and 1050 s for FS-CPCMs-4, i.e., the solidifying times of the FS-CPCMs with different mass fractions of Cu were 10.45%, 26.87%, 37.31%, and 47.76%, respectively, less than that of pristine PEG. It is well known that the thermal conductivity of SiO₂ is significantly higher than that of PEG,³⁷ and Cu powder has a high thermal conductivity.²⁰ Therefore, the reduction in the melting and solidifying times of the FS-CPCMs can be attributed to the enhanced heat transfer through Cu addition. Therefore, in practical thermal energy storage applications, the FS-CPCMs can absorb or release latent heat more rapidly.

Fig. 9 Melting (a) and solidifying (b) temperature curves of PEG and FS-CPCMs.

Phase change behavior of FS-CPCMs

The thermal energy storage capacity and phase change temperature were determined using DSC technique. Fig. 10a shows the melting and solidifying DSC curves of PEG and the FS-CPCMs. The phase change parameters obtained from the DSC evaluation are shown in Table 2, including the onset melting temperature (T_M), latent heat during the heating (H_M), onset freezing temperature (T_S), and latent heat during the cooling (H_S).

Fig. 10 PEG and FS-CPCMs: (a) DSC curves, (b) enthalpies of melting and solidifying, (c) extent of supercooling, (d) percentage of heat loss.

Table 2 Phase change behaviour of pristine PEG and FS-CPCMs

Samples	Melting process		Solidifying process
	TM (°C)	HM (J g^−1)	TS (°C)	HS (J g^−1)
PEG	56.4	183.4	42.1	158.5
FS-CPCMs-1	52.9	136.7	45.6	120.6
FS-CPCMs-2	53.1	129.6	45.6	117.9
FS-CPCMs-3	53.1	126.4	45.8	112.8
FS-CPCMs-4	53.7	125.2	43.9	110.7

---

A comparison of the thermal properties between the previously reported CPCMs and those obtained in this study (Table 3) shows that the phase change latent heat of the latter was lower than the former,^12,32,34 and the thermal conductivity of the latter was also lower than the former.¹⁸ However, the prepared CPCMs had a relatively higher enthalpy compared to the other reported PCMs^18,19,23,36 as well as a relatively higher thermal conductivity than the reported PCMs.^{12,19,23,34,36} Deng¹⁸ used Ag nanowires to increase the thermal conductivity of CPCMs. Compared to Ag nanowires, the Cu powder used in the CPCMs is inexpensive, easier to produce, and has great potential for thermal energy storage applications in energy-efficient buildings.

Table 3 Thermal property comparison between CPCMs prepared in this work and some CPCMs previously prepared in related references

PCMs	Melting process		Solidifying process		Thermal conductivity (W m^−1 K^−1)	References
	TM (°C)	HM (J g^−1)	TS (°C)	HS (J g^−1)
PEG (80 wt%)/SiO2	—	137.7	—	—	0.3615	12
PEG (58.8 wt%)/expanded vermiculite (21.9 wt%)/Ag nanowires (19.3 wt%)	59.96	110.0	43.22	108.1	0.680	18
Paraffin (53.2 wt%)/expanded vermiculite	48.85	101.14	53.01	103	0.452	19
PEG/SiO2/Al2O3 (12.6 wt%)	57.1	123.8	42.0	126.4	0.435	23
PEG/SiO2	61.11	131.90	46.00	121.44	—	32
Palmitic acid (79.9 wt%)/polypyrrole/graphene nanoplatelets (1.6 wt%)	—	151	—	—	0.43	34
Capric–myristic acid (50 wt%)/expanded perlite (40 wt%)/expanded graphite (10 wt%)	21.70	85.40	—	—	0.076	36
PEG (72.26 wt%)/SiO2/Cu (9.68 wt%)	53.7	125.2	43.9	110.7	0.513	This work

---

According to the literature,¹⁶ the theoretical latent heat value of PCMs can be determined using eqn (1):

H_Theo = εH_PCM (1)

where H_Theo is the theoretical latent heat of the PCMs; H_PCM is the latent heat of PEG; ε is the mass fraction of PEG in the PCMs.

As shown in Fig. 10b, the phase change enthalpies of the FS-CPCMs are lower than their theoretical enthalpy. A similar phenomenon was observed by Min et al.¹¹. This is probably because SiO₂ acts as an impurity, preventing the complete crystallization of PEG.³⁰ Moreover, the PEG segments are confined by the silica gel network, and it is difficult for a part of those to form stable crystals.^3,29 Fig. 10b and Table 2 shown that the latent heat of the FS-CPCMs-1 was lower than that of PEG, due to the mass fraction of SiO₂. The latent heat of FS-CPCMs was decreases with the increase of the Cu power. However, the enhancement of thermal conductivity was observed in the FS-CPCMs.

The supercooling degree of FS-CPCMs is an important parameter for practical applications. The extent of supercooling (ΔT) was evaluated using eqn (2):

ΔT = T_M − T_S (2)

The extent of supercooling shown in Fig. 10c is based on the DSC analysis data shown in Table 2. Compared to the supercooling degree of pristine PEG, the supercooling degrees of the FS-CPCMs with different mass fraction of Cu decreased by 48.95%, 47.55%, 48.95%, and 31.47%, respectively. This indicates that the extent of supercooling of PEG can be favorably reduced by blending with SiO₂.

The percentage of heat loss (η)¹¹ was evaluated using eqn (3):

η = (H_M − H_S)/H_M × 100% (3)

Fig. 10d shows the percentage of heat loss and phase change enthalpy for pristine PEG and FS-CPCMs. The percentages of heat loss of pristine PEG and the FS-CPCMs with different mass ratios of Cu were 13.58%, 11.78%, 9.03%, 10.76%, and 11.58%, respectively. The heat loss percentage of pristine PEG between the endothermic and exothermic cycles was higher than that of the FS-CPCMs.

As shown in Table 2, the melting temperatures of the FS-CPCMs were lower than that of pristine PEG, probably caused by the physical interactions and confined effect of PEG and SiO₂ such as surface tension, capillary forces, and hydrogen bonding.^38,39 Moreover, Li⁴⁰ reported that the increase in thermal conductivity caused a rapid temperature response; therefore, the melting temperature decreased.

Thermal stability of FS-CPCMs

Thermal stability is an important factor in PCMs research and applications.⁴¹ The TGA curve and DTG thermograms of PEG and FS-PCMs-2 are shown in Fig. 11. The weight loss processes of both pristine PEG and FS-PCMs-2 were carried out in only one step. For FS-PCMs-2, no apparent decomposition reaction and weight loss were observed before 247 °C. Therefore, FS-PCMs-2 has a good thermal stability when the temperature is below 247 °C. This property is needed for a material in heat storage applications. Moreover, only 0.37 wt% unknown residues remained for pristine PEG, and the residual mass of FS-PCMs-2 was 18.54 wt%, indicating that the prepared FS-CPCMs-2 was homogeneous. The loading content (77.24 wt%) of PEG in FS-PCMs-2 from the TGA measurements is in good agreement with that obtained in the experiment.

Fig. 11 TGA curves and DTG thermograms of PEG and FS-PCMs-2.

To determine the excellent thermal reliability of the FS-CPCMs, a 200-cycle experiment was carried out. Fig. 12a shows that no clear change was observed in the endothermal curve after the cycles, indicating that the FS-CPCMs had excellent thermal reliability, indicating a longer life cycle. As shown in Fig. 12b, the peak in the spectrum of FS-PCMs-2 before and after 200 thermal cycles did not change, indicating that no chemical reaction occurred after 200 thermal cycles in FS-PCMs-2. The results indicate that the chemical structure of FS-PCMs-2 was retained during the thermal cycling.

Fig. 12 (a) DSC curves of the prepared CPCM before and after 200 thermal cycles, (b) FT-IR spectrum of FS-CPCMs-2 before and after 200 thermal cycles.

Conclusions

Novel form-stable CPCMs of PEG/Cu/SiO₂ were prepared by adding Cu powder to PEG and SiO₂ via the ultrasound-assisted sol–gel method. PEG was successfully encapsulated by the fluffy network structure of SiO₂. The FTIR and XRD analyses show that the physical blending of Cu, PEG, and SiO₂ was homogeneous. A comparison of the thermal properties between the previously reported and these CPCMs, the PEG/Cu/SiO₂ CPCMs showed an optimum phase change temperature and a high enthalpy of phase change. The thermal conductivities of FS-CPCMs-2, FS-CPCMs-3, and FS-CPCMs-4 were 49.13%, 65.40%, and 77.51%, respectively, higher than that of pure PEG. Thus, the problems associated with the low thermal conductivity of PEG/SiO₂ can be solved by adding Cu powder. Therefore, the PEG/Cu/SiO₂ CPCMs are suitable for applications in building envelopes during peak hot summers.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (Grant No. 51472222 and 51372232) and the Research Fund for the Doctoral Program of Higher Education of China (Grant No. 20130022110006).

Notes and references

1. A. Sharma, V. V. Tyagi, C. R. Chen and D. Buddhi, Renewable Sustainable Energy Rev., 2009, 13, 318–345.
2. D. Zhou, C. Y. Zhao and Y. Tian, Appl. Energy, 2012, 92, 593–605.
3. A. de Gracia, L. Rincón, A. Castell, M. Jiménez, D. Boer, M. Medrano and L. F. Cabeza, Energy and Buildings, 2010, 42, 1517–1523.
4. E. Mettawee and G. Assassa, Energy, 2006, 31, 2958–2968.
5. S. Mondal, Appl. Therm. Eng., 2008, 28, 1536–1550.
6. T. Nomura, N. Okinaka and T. Akiyama, Resour., Conserv. Recycl., 2010, 54, 1000–1006.
7. T. Qian, J. Li, H. Ma and J. Yang, Sol. Energy Mater. Sol. Cells, 2015, 132, 29–39.
8. F. L. Tan and C. P. Tso, Appl. Therm. Eng., 2004, 24, 159–169.
9. D. P. Bentz and R. Turpin, Cem. Concr. Compos., 2007, 29, 527–532.
10. D. Zhang, Z. Li, J. Zhou and K. Wu, Cem. Concr. Res., 2004, 34, 927–934.
11. X. Min, M. Fang, Z. Huang, Y. Liu, Y. Huang, R. Wen, T. Qian and X. Wu, Sci. Rep., 2015, 5, 12964.
12. W. Wang, X. Yang, Y. Fang and J. Ding, Appl. Energy, 2009, 86, 170–174.
13. T. Qian, J. Li, X. Min, Y. Deng, W. Guan and H. Ma, Energy, 2015, 82, 333–340.
14. L. Xia, P. Zhang and R. Z. Wang, Carbon, 2010, 48, 2538–2548.
15. Y. Qian, P. Wei, P. Jiang, Z. Li, Y. Yan and J. Liu, Appl. Energy, 2013, 106, 321–327.
16. M. K. A. Sharif, A. A. Al-Abidi, S. Mat, K. Sopian, M. H. Ruslan, M. Y. Sulaiman and M. A. M. Rosli, Renewable Sustainable Energy Rev., 2015, 42, 557–568.
17. H. Zhang, X. Wang and D. Wu, J. Colloid Interface Sci., 2010, 343, 246–255.
18. Y. Deng, J. Li, T. Qian, W. Guan, Y. Li and X. Yin, Chem. Eng. J., 2016, 295, 427–435.
19. W. Guan, J. Li, T. Qian, X. Wang and Y. Deng, Chem. Eng. J., 2015, 277, 56–63.
20. J. M. Khodadadi, L. Fan and H. Babaei, Renewable Sustainable Energy Rev., 2013, 24, 418–444.
21. Y. Liu, Y. Zhou, M. Tong and X. Zhou, Microfluid. Nanofluid., 2009, 7, 579–584.
22. R. V. Seeniraj, R. Velraj and N. Lakshmi Narasimhan, J. Sol. Energy Eng., 2002, 124, 243.
23. B. Tang, C. Wu, M. Qiu, X. Zhang and S. Zhang, Mater. Chem. Phys., 2014, 144, 162–167.
24. J. L. Zeng, Z. Cao, D. W. Yang, L. X. Sun, L. Zhang and J. Therm, Anal. Calorim., 2009, 101, 385–389.
25. L. He, J. Li, C. Zhou, H. Zhu, X. Cao and B. Tang, Sol. Energy, 2014, 103, 448–455.
26. C. Oh, Y. Lee, T. Choi, C. Jon and S. Oh, Colloids Surf., A, 2009, 349, 145–150.
27. A. Sarı, Energy and Buildings, 2014, 69, 184–192.
28. R. K. Sharma, P. Ganesan, V. V. Tyagi, H. S. C. Metselaar and S. C. Sandaran, Energy Convers. Manage., 2015, 95, 193–228.
29. L. Feng, W. Zhao, J. Zheng, S. Frisco, P. Song and X. Li, Sol. Energy Mater. Sol. Cells, 2011, 95, 3550–3556.
30. B. Tang, M. Qiu and S. Zhang, Sol. Energy Mater. Sol. Cells, 2012, 105, 242–248.
31. M. Catauro, F. Bollino, F. Papale, M. Gallicchio and S. Pacifico, Mater. Sci. Eng., C, 2015, 48, 548–555.
32. C. Chen, W. Liu, Z. Wang, K. Peng, W. Pan and Q. Xie, Sol. Energy Mater. Sol. Cells, 2015, 134, 80–88.
33. R. S. Dubey, Y. B. R. D. Rajesh and M. A. More, Materials Today: Proceedings, 2015, 2, 3575–3579.
34. M. Silakhori, H. Fauzi, M. R. Mahmoudian, H. S. C. Metselaar, T. M. I. Mahlia and H. M. Khanlou, Energy and Buildings, 2015, 99, 189–195.
35. Y. Cai, X. Zong, J. Zhang, Y. Hu, Q. Wei, G. He, X. Wang, Y. Zhao and H. Fong, Sol. Energy Mater. Sol. Cells, 2013, 109, 160–168.
36. A. Karaipekli and A. Sarı, Renewable Energy, 2008, 33, 2599–2605.
37. F. He, X. Wang and D. Wu, Energy, 2014, 67, 223–233.
38. T. Qian, J. Li, H. Ma and J. Yang, Polym. Compos., 2014, 37, 854–860.
39. C. Wang, L. Feng, W. Li, J. Zheng, W. Tian and X. Li, Sol. Energy Mater. Sol. Cells, 2012, 105, 21–26.
40. J. Li, L. He, T. Liu, X. Cao and H. Zhu, Sol. Energy Mater. Sol. Cells, 2013, 118, 48–53.
41. L. Feng, J. Zheng, H. Yang, Y. Guo, W. Li and X. Li, Sol. Energy Mater. Sol. Cells, 2011, 95, 644–650.

---