Fe–Fe₃C/C microspheres as a lightweight microwave absorbent

Abstract

As electromagnetic pollution is becoming more and more serious, novel composite microwave absorbents are gaining much attention. In this work, low-cost glucose was used as a carbon source to prepare hydrochar, and Fe–Fe₃C/C microspheres for microwave absorption were successfully synthesized through the hydrothermal synthesis of Fe₃O₄/hydrochar and subsequent high-temperature carbonization at different temperatures. The results showed that the Fe–Fe₃C nanoparticles were uniformly loaded on the carbon microspheres. Resulting from the synergistic effect of Fe–Fe₃C nanoparticles and partially graphitized carbon, a wide region of microwave absorption was achieved due to dual dielectric and magnetic losses. An effective bandwidth of reflection loss less than −10 dB could reach up to 4 GHz with 1.5 mm thickness. Owing to the characteristics of the cost-effective synthetic route, low density and good microwave absorption with thin thickness, the Fe–Fe₃C/C microspheres could be used as a lightweight and highly efficient microwave absorbent.

---

1. Introduction

During the past few years, the electromagnetic (EM) pollution and interference problems are becoming more and more serious, so microwave-absorbing materials (MAMs) have attracted a great deal of interest for solving these problems.^1–3 Among these MAMs, carbon-based composite MAMs containing both carbon materials and magnetic components have received considerable attention because these materials not only have multiform EM losses based on magnetic and dielectric losses, which are quite beneficial to improving microwave absorption, but also can overcome the disadvantage of a high density of magnetic components.^4,5 So far, various carbon materials such as carbon fibers, graphene and carbon nanotubes have been extensively used as main components of carbon-based absorbents.^6–9 Wang et al. synthesized Fe₃O₄/carbon nanocoils using atomic layer deposition (ALD) technique.¹⁰ Tang et al. synthesized carbon-coated cementite (Fe₃C) nanocapsules by utilizing an arc-discharge method in ethanol.¹¹ The preparation of carbon-based composite MAMs with more economical synthesis route is still a challenge.

Biomass is the most abundant and cheap renewable resources. Recently, studies of biomass conversion and application have become a popular topic, receiving more and more attention.^12–14 Glucose is the most representative model compound of biomass and the most important monosaccharide distributed in nature.^15,16 In 2004, Li et al. reported hydrochar by hydrothermal carbonization of glucose.¹⁷ They found that the hydrochar has high carbon content and rich surface oxygen-containing groups, such as hydroxyl, carboxyl. These surface oxygen-containing groups can be combined with a variety of molecules, ions and other materials, providing a favorable condition for the loading of other species. Xu et al. reported that carbon spheres are excellent support for electrocatalyst Pt and Pd.¹⁸ Kim et al. prepared spherical carbon using sucrose as carbon precursor via hydrothermal method, and the spherical carbon can be used as support for PtRu-alloy catalysts in the methanol electro-oxidation.¹⁹ Considering the practical application of MAMs, if we use glucose to prepare hydrochar, and then load magnetic components on the hydrochar to prepare carbon-based absorbents, which will have much significance on the utilization of biomass and wide application of carbon-based composite MAMs. Despite much attention paid to carbon-based absorbents, there were only a few reports on the microwave absorption involved with biomass.

In this paper, we report a simple and cost-effective synthetic route to prepare Fe–Fe₃C/C microspheres, and use these hybrid materials for microwave absorption. We first synthesized Fe₃O₄/hydrochar via a hydrothermal method, and then produced Fe–Fe₃C/C microspheres through subsequent carbothermal reduction. The crystalline structure, morphology, and microwave absorption of these materials were investigated. The novelty of the Fe–Fe₃C/C microspheres lies in its lightweight, strong microwave absorption and wide absorption bandwidth with thin thickness, which makes it attractive for microwave absorption application. The aim of this paper is to prepare carbon-based absorbent using glucose as carbon source and investigate its microwave absorption performance.

2. Experimental section

2.1 Materials preparation

All the reagents were A.R. grade and were used without further purification. Ultra-pure grade water was used throughout the experiments. Ar (high purity, Shanxi Yihong Gas Industry Co.) was used as a protective gas. Fe–Fe₃C/C microspheres were fabricated through a two-step process.

(1) In a typical synthesis, hydrochar was prepared at 180 °C for 10 h in an aqueous glucose solution in Teflon-lined stainless steel autoclave. The Fe₃O₄/hydrochar composites were prepared by hydrothermal method. First of all, a precursor solution was prepared by dissolving 1 g hydrochar into 50 mL double-distilled water, followed by 0.004 mol of FeCl₂·4H₂O and 0.00026 mol of NaH₂PO₄·2H₂O. After the precursor solution was stirred for 0.5 h, the pH value of the mixture was adjusted to 12–13 by the addition of NaOH solution under constant stirring. Subsequently, the mixture was transferred into a 150 mL Teflon-lined stainless steel autoclave and treated for 15 h at 200 °C. After completion of the reaction, the autoclave was naturally cooled down to room temperature. The resultant black precipitate was collected by filtration, washed by water and absolute ethanol for two times and dried at 80 °C.

(2) The product obtained above was put in a ceramic boat and annealed in a horizontal tubular furnace at 700 °C, 800 °C, and 900 °C for 4 h, respectively, with a heating rate of 4 °C min⁻¹ in Ar atmosphere. The obtained samples were respectively called S₇₀₀, S₈₀₀, and S₉₀₀.

2.2 Materials characterization

Crystal phase analysis of the samples was performed by a D8 Advance Bruker AXS diffractometer using Cu Kα radiation (λ = 1.5406 Å), employing a scan step of 0.01° in a 2θ range from 10° to 90°. The morphology of the samples was observed by scanning electron microscope (SEM, JSM-7001F) and transmission electron microscope (TEM, JEM-2100F). Thermogravimetric (TG) analysis was carried out on a SETSYS evolution TGA at a heating rate of 10 °C min⁻¹ from 30 °C to 1000 °C in air atmosphere. The densities of the samples were measured by pycnometer according to national standard GB 4472-84, and the specification of pycnometer was 25 mL (Tianjing Tianbo Glass Instrument Co. Ltd.).

2.3 Microwave absorption measurement

Specimens for microwave absorption measurement were prepared by uniformly mixing the synthesized products in paraffin and pressed into a cylindrical shaped compact (Φ_outer = 7.0 mm and Φ_inner = 3.0 mm). The complex permittivity (ε_r = ε′ − jε′′) and complex permeability (μ_r = μ′ − jμ′′) of the specimen with 25 wt% Fe–Fe₃C/C microspheres were measured using an Agilent N5244A vector network analyzer. The reflection loss (RL) curves were calculated using the measured complex permittivity and complex permeability at given frequency and absorbent thickness based on transmission line theory by the following equations:^20,21

Z_in = (μ_r/ε_r)^1/2 tanh[j(2πfd/c)(μ_rε_r)^1/2], (1)

RL (dB) = 20 log|(Z_in − Z₀)/(Z_in + Z₀)|, (2)

where f is the frequency of electromagnetic wave, d is the thickness of absorbent, c is the velocity of light in free space, Z_in is the input characteristic impedance of absorbent, Z₀ is the characteristic impedance of free space.

3. Results and discussion

3.1 Phase and morphology analysis

It is known that glucose can be converted into hydrochar by hydrothermal carbonization process through dehydration, condensation, polymerization and aromatization reactions.^22,23 Fig. S1a† presents a typical XRD pattern of the synthesized hydrochar and the broad peak at about 13–32° indicates an amorphous nature of the carbon component.²⁴ After the hydrochar was loaded with Fe₃O₄ by hydrothermal method, the XRD pattern was shown in Fig. S1b.† It is clear that all of the diffraction peaks can be well assigned to a spinel structure of Fe₃O₄ (JCPDS 19-0629), indicating the successful loading of Fe₃O₄. Fig. S2a and S2b† show the SEM and TEM images of hydrochar. It is apparent that the hydrochar consists of many well-dispersed microspheres with mean diameter of 250 nm, and the surface of the microspheres is very smooth. Fig. S2c and S2d† show the SEM and TEM images of hydrochar loaded with Fe₃O₄. It is noted that the surface of the microspheres becomes rough, suggesting that Fe₃O₄ nanoparticles are precipitated on the surface of the hydrochar.

Fig. 1a shows typical XRD patterns of the synthesized samples S₇₀₀, S₈₀₀, and S₉₀₀, for these three samples, there exist three main diffraction peaks at 44.67°, 65.02°, and 82.33°, which corresponds to the crystalline planes of (110), (200) and (211) of Fe (JCPDS 06-0696). The diffraction peak at 2θ ≈ 26.5° can be indexed to (002) lattice plane of graphite, indicating the transition from amorphous carbon to nanocrystalline graphite in high temperature.²⁵ Apart from the above characteristic peaks of Fe and graphite, the other obvious diffraction peaks can be indexed as Fe₃C phase (JCPDS 35-0772), revealing that the coexistence of Fe–Fe₃C nanoparticles and carbon phases is achieved via the carbothermal reduction process. It is reported that the oxygen-containing functional groups in hydrochar could be transformed into CO₂ and H₂O during the high temperature carbonization process, then hydrochar could be converted into carbon.²⁶ Therefore it is comprehensible that there is interaction between Fe₃O₄ and hydrochar under high temperature condition, where carbon can induce carbothermic reduction to form Fe, and then Fe can reacts with the carbon to form Fe₃C. The Raman spectra analysis in Fig. 1b was further conducted to confirm the presence of carbon in S₇₀₀, S₈₀₀, and S₉₀₀. In these three samples, two distinguishable broad peaks at around 1350 cm⁻¹ and 1580 cm⁻¹ correspond to the D and G bands of carbon. The D-band can be attributed to the presence of sp³ defects within the carbon, whereas the G-band is characteristic for graphitic sheets.^17,27 Thus, the peak intensity ratio of the D and G bands (I_D/I_G) is indicative of the degree of graphitization. In Fig. 1b, the ratio values for S₇₀₀, S₈₀₀, and S₉₀₀ are calculated to be 0.89, 0.83, and 0.74, respectively, which indicates that the degree of graphitization of the carbon within our samples is strongly enhanced with the increase of calcination temperature.

Fig. 1 (a) XRD patterns and (b) Raman spectra of S₇₀₀, S₈₀₀, and S₉₀₀.

To characterize the morphology of S₇₀₀, S₈₀₀, and S₉₀₀, SEM was performed to provide further information. Fig. 2a and b show typical SEM images of S₇₀₀ and S₈₀₀. It is apparent that the Fe–Fe₃C nanoparticles are evenly loaded on the surface of the carbon microspheres. As shown in Fig. 2c, when the calcination temperature increases to 900 °C for S₉₀₀, the Fe–Fe₃C nanoparticles on the carbon microspheres become bigger, and the sample become agglomeration due to sintering or collapsing during treatment at higher temperature. The detailed structure information of S₈₀₀ is further obtained by the TEM image in Fig. 2d. It is clear that the Fe–Fe₃C nanoparticles on the carbon microspheres are about 20–40 nm, and the carbon microspheres show lamellar structure. Combined with the XRD results, we think the lamellar carbon may be nanocrystalline graphite.

Fig. 2 SEM images of (a) S₇₀₀, (b) S₈₀₀, (c) S₉₀₀; (d) TEM image of S₈₀₀.

Fig. 3 displays the thermogravimetric (TG) curves of S₈₀₀ in air atmosphere at a heating rate of 10 °C min⁻¹. TG analysis shows two weight losses mainly due to the removal of adsorbed water (30–300 °C) and combustion of carbon (300–500 °C), indicating that the weight increase from oxidation of Fe–Fe₃C nanoparticles is well below the weight loss due to the evaporation of adsorbed water and oxidation of carbon. The contents of carbon and iron in the sample can also be calculated from the TG curve. The total weight loss of 86 wt% for S₈₀₀ during the oxidation process suggests that the weight percentage of Fe in Fe–Fe₃C/C composites is approximately 10 wt%, assuming that all Fe has transferred to Fe₂O₃ and all C has been burnt out under air atmosphere.

Fig. 3 TG curves of S₈₀₀ in air atmosphere.

Based on the above XRD patterns, SEM, and TEM images, the schematic illustration of the synthesis procedures for Fe–Fe₃C/C microspheres was shown in Fig. 4. At first, hydrochar was prepared by hydrothermal carbonization of glucose. Then Fe₃O₄/hydrochar was synthesized by hydrothermal method containing hydrochar, and the Fe₃O₄ nanoparticles were deposited on the surface of the hydrochar. Thirdly, compared with bulk materials, nanocrystals have high specific surface area and better flexibility, which make it easier for them to undergo a phase transformation by oxidation/reduction process.^28,29 Therefore, when the Fe₃O₄/hydrochar is annealed at 700 °C or higher temperature for 4 h under constant flow of Ar, it is comprehensible that the Fe₃O₄ nanoparticles on the surface of hydrochar is reduced to Fe and Fe₃C nanoparticles after the carbothermal reduction reaction.

Fig. 4 Illustration of the synthetic protocol for Fe–Fe₃C/C microspheres.

3.2 Microwave absorption analysis

From the above analyses, Fe–Fe₃C nanoparticles with the diameter of about 20–40 nm were successfully loaded on carbon microspheres by the present synthetic method. Recent researches have shown that both Fe–Fe₃C nanoparticles and carbon exhibited an electromagnetic response.^30,31 Thus, the Fe–Fe₃C/C composites are expected to show excellent electromagnetic absorption. The electromagnetic parameters of the paraffin composites containing 25 wt% samples were measured to investigate the microwave absorption.

Fig. 5a shows the frequency dependence of the real part (ε′) and imaginary part (ε′′) of the complex permittivity for samples S₇₀₀, S₈₀₀, and S₉₀₀. ε′ represents the electric field energy capacity and ε′′ stands for the loss of electric field energy.^32,33 For S₇₀₀, S₈₀₀, and S₉₀₀, the values of ε′ are respectively in the range of 13.5–10.9, 20.4–12.5, and 20.3–12.2 in the frequency range of 2–18 GHz. Meanwhile, the ε′′ value of S₇₀₀ decreases from 3.38 to 2.59 first and then increases to 3.69, while the ε′′ values of S₈₀₀ and S₉₀₀ decreases from 10.65 to 5.64 and then slightly increases to 6.05. It is obvious that the ε′ and ε′′ of S₈₀₀ and S₉₀₀ are much higher than that of S₇₀₀, and the high permittivity mainly results from the high conductivity of carbon in high temperature.³⁴ For samples S₇₀₀, S₈₀₀, and S₉₀₀, the real part (μ′) and imaginary part (μ′′) of complex permeability are displayed in Fig. 5b, where μ′ and μ′′ signify the magnetic energy storage capability and magnetic loss ability of absorbent, respectively.^32,33 Because the magnetic components of S₇₀₀, S_800, and S₉₀₀ are similar, so the complex permeability is close to each other. We can see that the μ′′ is more or less constant over a frequency range of 2–12 GHz, and then decreases in the higher frequency range. In addition, μ′′ is found to have negative value in the higher frequency range. As we know, the magnetic behavior may be modulated by the dielectric behavior, which may lead to the coupling between these two parameters.³⁵ According to the Maxwell equations, a magnetic field can be induced by an AC electric field and radiated out. The motion of free charges in S₈₀₀ will produce an AC electric field. So we hypothesize that the negative μ′′ value denotes that the radiated magnetic energy is transferred into the electric energy, which can increase ε′′ and then lead to the negative μ′′. This phenomenon is similar to other reports.^36,37

Fig. 5 Frequency dependence of (a) complex permittivity and (b) complex permeability for S₇₀₀, S_800, and S₉₀₀.

Fig. 6 displays the calculated RL for the absorbents consisting of S₇₀₀, S₈₀₀, and S₉₀₀ dispersed in paraffin matrix. It is well accepted that the thickness of absorbent is a crucial parameter affecting the RL intensity and frequency position of maximum absorption. Therefore, we calculated the RL curve at different thicknesses of 1, 1.5, 2, 2.5 and 3 mm for each sample in the frequency range of 2–18 GHz. Considering the paraffin is an insulator and a nonmagnetic material, so that it is a kind of transparent material to the electromagnetic wave.^38,39 Therefore, the S₇₀₀, S₈₀₀, and S₉₀₀ are the only materials contributing to the electromagnetic response. As seen in Fig. 6, it is apparent that the minimum RL gradually shifts toward lower frequency with the increase of thickness, which can be explained by the “geometrical effect”.³⁶ Generally speaking, a RL value of −10 dB corresponds to 90% microwave absorption and only the absorbent with RL below −10 dB is suitable for practical application. So the effective bandwidth represents the width of frequency range in which the RL is −10 dB. For S₇₀₀, the minimum RL is −16.6 dB while the effective bandwidth achieves 3.4 GHz (from 14.6 to 18 GHz) with a matching thickness of 1.5 mm, shown in Fig. 6a. By contrast, the microwave absorption of S₈₀₀ and S₉₀₀ is evidently improved, as seen in Fig. 6b and c. For S₈₀₀, when the thickness is 1.5 mm, the minimum RL is −17.9 dB, and the effective bandwidth can reach up to 4 GHz (from 12.2 to 16.2 GHz). For S₉₀₀, the minimum RL is as low as −18.8 dB while the effective bandwidth achieves 4 GHz (from 12.3 to 16.3 GHz). Compared with the reported carbon-based composite MAMs such as CoFe₂O₄/grapheme,⁴⁰ and MWNTs/Fe₃O₄,⁴¹ Fe–C nanofibers,⁴² the Fe–Fe₃C/C microspheres have a lower filling rate and thinner thickness, and still exhibit relatively wider effective bandwidth. Moreover, the effective bandwidth of 4 GHz is larger than that of the other reported microwave absorption materials (Table 1).^2,10,11,32 In addition, we have measured that the densities of the Fe–Fe₃C/C composites are 1.4–1.5 g cm⁻³. In comparison with the traditional microwave absorbents such as ferrite, metal, ZnO, and SiC materials, the densities of the Fe–Fe₃C/C composites are very low due to high carbon content. Thus the MAMs based on Fe–Fe₃C/C composites have the advantages of low cost, low density, strong microwave absorption, and wide absorption bandwidth with thin thickness, therefore, they can meet the multiple requirements of microwave absorbent.

Fig. 6 Frequency dependence of the reflection loss curves for (a) S₇₀₀, (b) S₈₀₀, and (c) S₉₀₀ with different thicknesses.

Table 1 Microwave absorption performances of some reported absorbents

Sample	Filling rate (wt%)	Thickness (mm)	Effective bandwidth (GHz)	Ref.
CoFe2O4/graphene	60	2	3.7	40
CoFe2O4/Co3Fe7–Co	80	2	2.5	2
MWCNTs/Fe3O4	30	2	2.5	41
Fe3C/C nanocapsules	40	2	4	11
Fe–C nanofibers	50	2	1.7	42
Fe3O4/carbon nanocoils	25	2	3.5	10
Fe nanocrystals	50	2	3.4	32
Fe–Fe3C/C microspheres	25	1.5	4	This work

---

Dielectric loss and magnetic loss are two main contributions for microwave absorption, which can be expressed as tan δ_e = ε′′/ε′ and tan δ_m = μ′′/μ′, respectively. Fig. 7 shows the frequency dependence of the loss tangent for S₇₀₀, S₈₀₀, and S₉₀₀. For all samples, the value of tan δ_e is remarkably larger than that of tan δ_m in the frequency range of 2–18 GHz, indicating that the dielectric loss plays the main role, which is similar to other reports.^43,44 The tan δ_e of S₈₀₀ and S₉₀₀ is remarkably larger than that of S₇₀₀ in the frequency range of 2–18 GHz, resulting in better microwave absorption. More importantly, because of the difference in complex permittivity between Fe–Fe₃C nanoparticles and carbon, interface scattering would be generated, leading to more microwave attenuation.^2,40 As we know, the attenuation constant α determines the microwave attenuation properties of materials and can be expressed by the following equation:^45,46

where f is the frequency of microwaves and c is the velocity of electromagnetic wave in free space. Using the measured complex permittivity and complex permeability at given frequency, the frequency dependence of attenuation constant was shown in Fig. 8. It is clear that the attenuation constant values of S₈₀₀ and S₉₀₀ are much higher than that of S₇₀₀, therefore, they exhibit better microwave absorption properties.

Fig. 7 Frequency dependence of the loss tangent for S₇₀₀, S_800, and S₉₀₀.

Fig. 8 Frequency dependence of the attenuation constant for S₇₀₀, S₈₀₀, and S₉₀₀.

4. Conclusions

In conclusion, Fe–Fe₃C/C microspheres were successfully fabricated through the hydrothermal synthesis of Fe₃O₄/hydrochar and the subsequent high temperature carbonization. Resulting from the dielectric loss, magnetic loss, and interface scattering, the synthesized Fe–Fe₃C/C microspheres exhibited excellent microwave absorption performance. The effective bandwidth of RL less than −10 dB can reach up to 4 GHz when the thickness of the test sample was as thin as 1.5 mm with an 800 °C carbothermal reduction process. In addition, the synthetic route of Fe–Fe₃C/C microspheres was cost-effective because low cost glucose was used as carbon source. These results demonstrated the potential of Fe–Fe₃C/C microspheres as a lightweight and highly efficient microwave absorbent, and may offer a new way to design promising EM absorption materials.

Acknowledgements

This work was supported by the Startup Foundation of Doctors of Jinzhong University.

References

1. L. Yang, H. L. Fan, J. Liu, Y. Ma and Q. Zheng, Mater. Des., 2013, 50, 863–871.
2. L. Xi, Z. Wang, Y. L. Zuo and X. N. Shi, Nanotechnology, 2011, 22, 045707.
3. G. Z. Wang, Z. Gao, G. P. Wan, S. W. Lin, P. Yang and Y. Qin, Nano Res., 2014, 7, 704–716.
4. Y. C. Du, W. W. Liu, R. Qiang, Y. Wang, X. J. Han, J. Ma and P. Xu, ACS Appl. Mater. Interfaces, 2014, 6, 12997–13006.
5. W. X. Li, B. L. Lv, L. C. Wang, G. M. Li and Y. Xu, RSC Adv., 2014, 4, 55738–55744.
6. Q. M. Su, G. Zhong, J. Li, G. H. Du and B. S. Xu, Appl. Phys. A, 2012, 106, 59–65.
7. G. Li, L. M. Sheng, L. M. Yu, K. An, W. Ren and X. L. Zhao, Mater. Sci. Eng., B, 2015, 139, 153–159.
8. X. G. Meng, Y. Z. Wan, Q. Y. Li, J. Wang and H. L. Luo, Appl. Surf. Sci., 2011, 257, 10808–10814.
9. P. B. Liu and Y. Huang, RSC Adv., 2013, 3, 19033–19039.
10. G. Z. Wang, Z. Gao, S. W. Tang, C. Q. Chen, F. F. Duan, S. C. Zhao, S. W. Lin, Y. H. Feng, L. Zhou and Y. Qin, ACS Nano, 2012, 6, 11009–11017.
11. Y. Tang, Y. Shao, K. F. Yao and Y. X. Zhong, Nanotechnology, 2014, 25, 035704.
12. S. M. Kang, X. L. Li, J. Fan and J. Chang, Bioresour. Technol., 2012, 110, 715–718.
13. A. L. Ma, X. M. Wang, T. B. Li, X. G. Liu and B. S. Xu, Mater. Sci. Eng., A, 2007, 443, 54–59.
14. D. C. Rezan, N. Baccile, M. Antonietti and M. M. Titirici, Chem. Mater., 2009, 21, 484–490.
15. Y. Gao, X. H. Wang, H. P. Yang and H. P. Chen, Energy, 2012, 42, 457–465.
16. C. Falco, N. Baccile and M. M. Titirici, Green Chem., 2011, 13, 3273–3281.
17. X. M. Sun and Y. D. Li, Angew. Chem., Int. Ed., 2004, 43, 597–601.
18. C. W. Xu, L. Q. Cheng, P. K. Shen and Y. L. Liu, Electrochem. Commun., 2007, 9, 99–1001.
19. P. Kim, J. B. Joo, W. Kim, J. Kim, I. K. Song and J. Yi, Catal. Lett., 2006, 112, 213–218.
20. Y. Naito and K. Suetake, IEEE Trans. Microwave Theory Tech., 1971, 19, 65–72.
21. W. M. Zhu, L. Wang, R. Zhao, J. W. Ren, G. Z. Lu and Y. Q. Wang, Nanoscale, 2011, 3, 2862–2864.
22. S. M. Kang, X. L. Li, J. Fan and J. Chang, Ind. Eng. Chem. Res., 2012, 51, 9023–9031.
23. M. Sevilla and A. B. Fuertes, Carbon, 2009, 47, 2281–2289.
24. N. Li, M. H. Cao and C. W. Hu, J. Mater. Chem., 2012, 22, 18426–18432.
25. K. Y. Park, J. H. Han, S. B. Lee and J. W. Yi, Composites, Part A, 2011, 42, 573–578.
26. Q. Wang, H. Li, L. Q. Chen and X. J. Huang, Carbon, 2001, 39, 2211–2214.
27. H. Xia, Y. H. Wan, G. L. Yuan, Y. S. Fu and X. Wang, J. Power Sources, 2013, 241, 486–493.
28. F. Jiao, J. C. Jumas, M. Womes, A. V. Chadwick, A. Harrison and P. G. Bruce, J. Am. Chem. Soc., 2006, 128, 12905–12909.
29. B. L. Lv, Y. Xu, Q. Gao, D. Wu and Y. H. Sun, J. Nanosci. Nanotechnol., 2010, 10, 2348–2359.
30. S. V. Pol, V. G. Pol, G. Seisenbaeva, V. G. Kessler and A. Gedanken, Chem. Mater., 2004, 16, 1793–1798.
31. L. Liu, K. C. Zhou, P. G. He and T. F. Chen, Mater. Lett., 2013, 110, 76–79.
32. Q. C. Liu, Z. F. Zi, D. J. Wu, Y. P. Sun and J. M. Dai, J. Mater. Sci., 2012, 47, 1033–1037.
33. H. B. Yang, T. Ye and Y. Lin, RSC Adv., 2015, 5, 103488–103493.
34. Y. C. Du, T. Liu, B. Yu, H. B. Gao, P. Xu, J. Y. Wang, X. H. Wang and X. J. Han, Mater. Chem. Phys., 2012, 135, 884–891.
35. L. Zhen, J. T. Jiang, W. Z. Shao and C. Y. Xu, Appl. Phys. Lett., 2007, 90, 142907.
36. L. J. Deng and M. G. Han, Appl. Phys. Lett., 2007, 91, 023119.
37. X. L. Shi, M. S. Cao, J. Yuan and X. Y. Fang, Appl. Phys. Lett., 2009, 95, 163108.
38. J. H. Du, C. Sun, S. Bai, G. Su, Z. Ying and H. M. Cheng, J. Mater. Res., 2002, 17, 1232–1236.
39. H. Jiang, J. Guo, L. Zhao and H. Zhu, J. Inorg. Mater., 2010, 25, 73–76.
40. M. Fu, Q. Z. Jiao, Y. Zhao and H. S. Li, J. Mater. Chem. A, 2014, 2, 735–744.
41. Q. C. Liu, J. M. Dai, Z. F. Zi, A. B. Pang, Q. Z. Liu, D. J. Wu and Y. P. Sun, J. Low Temp. Phys., 2013, 170, 261–267.
42. T. Wang, H. D. Wang, X. Chi, R. Li and J. B. Wang, Carbon, 2014, 74, 312–318.
43. X. S. Qi, Y. Yang, W. Zhong, C. Qin, Y. Deng, C. Au and Y. W. Du, Carbon, 2010, 48, 3512–3522.
44. J. Xiang, J. L. Li, X. H. Zhang, Q. Ye, J. H. Xu and X. Q. Shen, J. Mater. Chem. A, 2014, 2, 16905–16914.
45. J. P. Wang, J. Wang, R. X. Xu, Y. Sun, B. Zhang, W. Chen, T. Wang and S. Yang, J. Alloys Compd., 2015, 653, 14–21.
46. Z. J. Wang, L. N. Wu, J. G. Zhou, B. Z. Shen and Z. H. Jiang, RSC Adv., 2013, 3, 3309–3315.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra02787c

---