Effect of vapor pressure on performance of sulfurized polyacrylonitrile cathodes for Li/S batteries

Abstract

Sulfurized polyacrylonitrile (S@pPAN) composites have been synthesized in a high pressure reactor under certain vapor pressure. According to elemental analysis, Raman spectroscopy and X-ray photoelectron spectroscopy (XPS), the hydrogen content and the degree of graphitization in S@pPAN composites decreases with increasing synthesis vapor pressure, indicating that a suitably high vapor pressure is helpful in promoting the reaction between polyacrylonitrile and elemental sulfur and finally making the conductive molecular structure of sulfurized polyacrylonitrile more complete. The results show that the sulfurized polyacrylonitrile composite prepared under 5 MPa vapor pressure (S@pPAN-5) delivers a high initial discharge specific capacity of 1821 mA h g⁻¹ and the capacity remains at 1357 mA h g⁻¹ (616.8 mA h g_composite⁻¹) after 100 cycles with a retention rate of 88% calculated from the 2^nd discharge capacity.

---

Introduction

Lithium sulfur (Li/S) batteries have been highly considered as one of the most promising high-energy storage systems, owing to high specific capacity (1675 mA h g⁻¹) and theoretical specific energy (2600 W h kg⁻¹).^1–3 Additionally, elemental sulfur also has advantages of natural abundance, low cost, and environmental friendless. However, three adverse factors prevent the practical application of Li–S batteries. First, the poor electrical/ionic conductivity of elemental sulfur and Li₂S₂/Li₂S lowers the utilization of active material for obtaining full capacity of the electrode. Second, the high solubility of intermediate reduction products Li₂S_x (2 < x ≤ 8) in liquid electrolytes and their shuttle between cathode and anode leads to rapid capacity degradation and lower coulomb efficiency. Third, the volume changes between lithiated and non-lithiated sulfur species will damage the structural stability of the electrode and concomitantly decrease the cycle life of the Li/S cells.^4–8 Therefore, many efforts have been made to prepare special structure cathode materials, based on sulfur with various kinds of conductive polymers, metal compounds and carbon materials, in which these conductive agents can reduce polysulfide dissolution, relieve the sulfur cathode volume change and play a role of conductivity enhancers.^9–17

Among the investigated composites, sulfurized polyacrylonitrile (S@pPAN) composite has displayed several amazing advantages: (1) it shows very stable specific discharge capacity, which is close to or even higher than the theoretical value of elemental sulfur; (2) it displays good chemical compatibility with the commercial carbonate-based electrolytes widely used in Li-ion batteries, instead of the ester-based electrolytes preferred in conventional Li–S batteries; (3) the discharge intermediates of high-order polysulfides are almost insoluble in carbonate-based electrolytes, which avoids the shuttle effect and the corrosion of Li anode; (4) it is suitable for the cathode with high active material loading in electrodes; and (5) the cells with sulfurized polyacrylonitrile cathode can retain approximately 100% coulombic efficiency (except for the first cycle) and extremely low self-discharge rate.^17–26,33 Researchers have taken many methods to modify the S@pPAN composites, such grapheme and Mg_0.6Ni_0.4O, for improving their electrochemical performance (seen in Table SI-1†). It is found that there are obvious differences on the electrochemical performance of the S@pPAN composites with similar sulfur content, such as S/PAN/graphene¹⁸ and pPAN–S/GNS¹⁹ materials. The sulfur content of the S@pPAN composite strongly affects its specific capacity: the higher the sulfur content, the larger the capacity of the composite. By carefully comparing some papers, some phenomena can be found that the capacity of the S@pPAN composite can be improved through increasing the sulfur content of the S@pPAN composite, such as SPAN/RGO²⁰ and S/PAN/Mg_0.6Ni_0.4O.²¹ However, the capacity has no greater improvement through increasing the sulfur content of S@pPAN composite, such as S/PAN/graphene¹⁸ and S/PAN/Mg_0.6Ni_0.4O,²¹ and even deteriorate by comparison between S–PAN²³ and PAN–S–VA.²² It can be conclude that the capacity of S@pPAN composite is not always improved by increasing the sulfur content. Therefore, there should have other internal factors to affect the electrochemical performance of S@pPAN composites besides the sulfur content and conductive materials. The synthesis process of S@pPAN composite is generally assumed to be pyrolysis of polyacrylonitrile (PAN) with sulfur as the oxidant. During the reaction process, the sulfur not only facilitates the dehydrogenation and cyclization of PAN, but also can combine with the positively polarized carbon atoms in the polymer matrix. As a reaction with gas involved, including sulfur free radical and hydrogen sulfide, the system vapor pressure will inevitably affect the reaction progress. However, there have been not studies about the effect of reaction vapor pressure on electrochemical performance of sulfurized polyacrylonitrile (S@pPAN) composites at present. The most commonly used method for preparing the S@pPAN composites is heating the sulfur/PAN mixture in a flowing inert atmosphere. However, elemental sulfur would be taken away by the flowing inert gas during heating process, which leads to partly uncompleted reaction between sulfur and polyacrylonitrile (PAN). Finally, the amount of the conductive six-membered rings containing nitrogen atoms in prepared S@pPAN composites will decrease, which lower the conductivity, and the electrochemical performance of S@pPAN composites.

In this work, the S@pPAN composites were prepared through heating the polyacrylonitrile (PAN)/sulfur mixture in a high pressure reactor filled with argon gas under certain vapor pressure. This comparative study assesses the effect of vapor pressure on the morphology, molecular structure and electrochemical performance of the resulting S@pPAN composites, attesting the excellent electrochemical performance of composites prepared in sealed high vapor pressure system.

Experimental

Preparation of the S@pPAN composites

Sulfur (Aladdin-reagent) and polyacrylonitrile (PAN) (Sigma-Aldrich) in a weight ratio of 3 : 1 were mixed by balling for 2 h. Then the mixture was heated at 330 °C for 5 h in a high pressure reactor filled with argon gas under certain vapor pressure of 2 MPa (S@pPAN-2), 5 MPa (S@pPAN-5) and 8 MPa (S@pPAN-8), respectively. For comparison, another sample was heat treated at 330 °C for 5 h in a flowing argon atmosphere to form the sulfurized polyacrylonitrile composite (S@pPAN-0). The above samples were rinsed with toluene for no less than three times to remove any free, non-bound sulfur.

Materials characterization

Elemental composition of the samples was analyzed by means of elemental analysis (CHNS, Vario EL Cube, Elementar). The morphology of the materials was studied by scanning electron microscopy (SEM, JEOL JSM-7401F). X-ray photoelectron spectra (XPS) were recorded on an AXIS Ultra DLD spectrometer (Kratos) to characterize the surface composition. Room-temperature Raman spectra were recorded on a Lab RAM HR Evolution-Horiba Scientific system with a green laser (532 nm). Nitrogen adsorption–desorption was performed using a NOVA1200 instrument (Quantachrome Corporation) at 77 K. Total specific area was calculated by multipoint Brunauer–Emmett–Teller method.

Electrochemical measurements

The electrochemical performance of the composites was assessed using CR2025 type coin cells with a lithium metal anode. The working electrodes were prepared by mixing 70 wt% S@pPAN composite, 20 wt% acetylene black and 10 wt% LA132 (as the binder). The resulting slurry was cast on an Al foil by the doctor blade method. The electrodes were dried at 60 °C under vacuum for 24 h and then were cut into discs with 14 mm in diameter. Sample electrodes with cathode composite loadings of ca. 1.9 mg cm⁻², corresponding to sulfur load of 0.862 mg cm⁻² 1 M LiPF₆ solution in ethylene carbonate/dimethyl carbonate/diethyl carbonate (EC : DMC : DEC, volume ratio = 1 : 1 : 1) as the liquid electrolyte. A Celgard2325 membrane was used as the separator. Constant current discharge/charge tests were performed on a LAND CT2001A battery test system (Wuhan, China) in a voltage range of 1–3 V (vs. Li/Li⁺) at 200 mA g⁻¹ current density. In this paper, if it is not specially marked, the specific discharge capacity and current densities were calculated based on sulfur mass of the S@pPAN composite.

Results and discussion

The photographs of the samples are presented in Fig. 1. The materials prepared in flowing argon atmosphere (S@pPAN-0) and 2 MPa vapor pressure (S@pPAN-2) show the similar morphology of aggregated irregular structures with wide particle size distribution (Fig. 1a and b). This probably due to the interconnection of sulfurized polyacrylonitrile molecular chains to form larger particles in heating process under low vapor pressure. When the vapor pressure increases to 5 MPa and 8 MPa, the materials of S@pPAN-5 and S@pPAN-8 display the morphology of dispersed regular spherical structures with an average particle diameter of about 250 nm (Fig. 1c and d). This demonstrates that the involvement of sulfurized polyacrylonitrile chains will decrease under suitable high vapor pressure, consequently getting smaller particle size and more regular shape materials. As shown in Table SI-2,† the surface area of S@pPAN-0, S@pPAN-2, S@pPAN-5 and S@pPAN-8 is 13.22, 15.47, 22.86 and 20.53 m² g⁻¹, respectively. The higher surface area is not only helpful to shorter diffusion lengths for lithium ions, but also greatly contribute to the electrolyte infiltration and polysulfides adsorption, improving the electrochemical performance of Li/S batteries.

Fig. 1 SEM images of S@pPAN composites: (a) S@pPAN-0, (b) S@pPAN-2, (c) S@pPAN-5, (d) S@pPAN-8.

Information about the structure of prepared sulfurized polyacrylonitrile materials was verified by analysis of the Raman spectra, as shown in Fig. 2. The D band at 1350 cm⁻¹ corresponds to the structural defects and disorders in the carbon matrix. The G band at approximately 1570 cm⁻¹ represents the graphitic order. In the Raman spectra of carbon-based materials, the parameter I_D/I_G is an interesting indicator of the level of defects within the materials.^27,28 In some papers about heteroatom-doped carbon materials,^29,30 a higher intensity of I_D/I_G indicates an increased amount of defects that is proportional to the extent of heteroatom insertion into carbon rings. From Fig. 2, the S@pPAN-5 sample has the highest value of I_D/I_G (0.97), followed by S@pPAN-0 (0.92), S@pPAN-2 (0.93) and S@pPAN-8 (0.95). The higher ratio of I_D/I_G implies that the S@pPAN-5 sample has more defects caused by more conductive six-membered rings containing nitrogen atoms similar to the heteroatom doping.²² The increased conductive six-membered rings containing nitrogen atoms indicate the more complete conductive molecular structure generated by the more thorough reaction between sulfur and polyacrylonitrile under suitable high vapor pressure, which are helpful to the conductivity improvement of the sulfurized polyacrylonitrile composites. Furthermore, Y. Matsumura et al. have proposed model for the interactions between amorphous carbon and lithium ions.³¹ In this model, lithium ions locate not only between graphitic layers, but also at the edge of graphitic structures and on the surface of crystallites. Similarly, lithium ions may interact with various different sites in S@pPAN composites due to the amorphous structure with unique semi-graphitic layers.²² The higher degree of amorphous structure in S@pPAN composites may thus facilitate faster lithium ion diffusion. In addition, the signals of C–S bonds at 790 cm⁻¹ and S–S bonds at 470 cm⁻¹ can be clearly seen.³²

Fig. 2 Raman spectra of S@pPAN-0, S@pPAN-2, S@pPAN-5 and S@pPAN-8.

The difference in the molecular structure of the prepared sulfurized polyacrylonitrile materials was further revealed with X-ray photoelectron spectroscopy (XPS). The main peak S2p_3/2 (marked in blue line in Fig. 3), slightly lower than binding energy (163.86 eV) of elemental sulfur, is consistent with the 2p electrons of S atoms either directly bonded to carbon or in short-chain organosulfide.^22,32–34 The location (marked in blue line in Fig. 3) for S@pPAN-0, S@pPAN-2, S@pPAN-5 and S@pPAN-8 are 163.83 eV, 163.77 eV, 163.11 eV and 163.24 eV, respectively. It can be seen that the peak S2p_3/2 location of S@pPAN-0 and S@pPAN-2 are closer to binding energy of elemental sulfur (163.86 eV) than S@pPAN-5 and S@pPAN-8, which mean that fewer C–S bonds or more relatively longer chain sulfur species are present in S@pPAN-0 and S@pPAN-2. The fewer C–S bonds or more relatively longer chain sulfur species, the fewer polarized carbon atoms or six-membered conductive rings in S@pPAN composite. The lower binding energy location (163.11 eV) of S@pPAN-5 means a more complete structure than other three samples.

Fig. 3 S2p XPS spectra of S@pPAN composites: (a) S@pPAN-0, (b) S@pPAN-2, (c) S@pPAN-5 and (d) S@pPAN-8.

The results of elemental analysis clearly reveal the element content variation of the S@pPAN-0, S@pPAN-2, S@pPAN-5 and S@pPAN-8 (seen in Table 1). The mass percentage of C, N and S in the prepared materials basically has tiny changes. Accordingly, the atomic ratio of C to N and C to S maintains at about 3 : 1 and 2.2 : 1. However, the mass percentage of H gradually decreases from 1.44% in S@pPAN-0 to 1.06% in S@pPAN-5, correspondingly the atomic ratio of C to H increases about from 2.17 : 1 to 2.97 : 1. The great changes on C to H ratio indicates that high vapor pressure within a certain range has promoting effect on the dehydrogenation and cyclization reaction between sulfur and polyacrylonitrile, which may make the molecular structure of sulfurized polyacrylonitrile more complete.³⁵

Table 1 Elemental analysis of the S@pPAN composites

Samples	Mass percentage, wt%				Atomic ratio
	C	N	H	S	C/N	C/S	C/H
S@pPAN-0	37.6	14.67	1.44	45.38	2.98	2.2	2.17
S@pPAN-2	37.66	14.8	1.22	45.3	2.97	2.21	2.58
S@pPAN-5	37.82	14.86	1.06	45.46	2.97	2.22	2.97
S@pPAN-8	37.84	14.82	1.07	45.42	2.98	2.22	2.95

---

The improved degree of dehydrogenation and cyclization in S@pPAN composites significantly affects the electrochemical properties. The performances of prepared S@pPAN composites are investigated and shown in Fig. 4. The initial charge/discharge profiles of four S@pPAN samples are quite different. S@pPAN-5 and S@pPAN-8 show a similar single voltage plateau around 1.8 V with a significant voltage hysteresis (Fig. 4a). It has been widely accepted that the single sloping voltage plateau in discharge profiles suggests a solid–solid phase reaction. Compared with these two samples, the S@pPAN-2 and S@pPAN-0 have much lower discharge potentials around 1.72 V and 1.65 V respectively. The higher initial discharge potentials can be attributed to the more complete molecular structure, which results in smaller electrochemical polarization and higher discharge potential. Moreover, the initial discharge specific capacity is larger than the subsequent charge/discharge cycles. This phenomenon has been observed in micro-porous carbon/sulfur and sulfurized carbon composites and is attributed to possible irreversible reactions of lithium with surface functional groups leading to lithium ion insertion into conjugated backbone. Quantitatively speaking, the initial discharge specific capacity calculated on sulfur mass for S@pPAN-0, S@pPAN-2, S@pPAN-5 and S@pPAN-8 are 1470 mA h g⁻¹, 1628 mA h g⁻¹, 1821 mA h g⁻¹ and 1776 mA h g⁻¹, respectively, and those in the second cycle are 1176 mA h g⁻¹, 1320 mA h g⁻¹, 1542 mA h g⁻¹ (701 mA h g_composite⁻¹) and 1538 mA h g⁻¹, respectively. S@pPAN-5 and S@pPAN-8 consistently show a higher discharge specific capacity than the S@pPAN-2 and S@pPAN-0 samples, indicating that the application of suitable high vapor pressure in material preparing process can significantly improve the electrochemical performance of the S@pPAN composites.

Fig. 4 Electrochemical performance of S@pPAN composites: (a) initial and second discharge/charge curves of S@pPAN composites at a current density of 200 mA g⁻¹. (b) Cycling performance of S@pPAN composites at a current density of 200 mA g⁻¹. (c) Rate performance of S@pPAN composites at various current rates.

The cycling performance of every S@pPAN sample at a current density of 200 mA g⁻¹ is displayed in Fig. 4b. The discharge specific capacity of the S@pPAN-0 composite cathode continuously decreases and reaches about 721 mA h g⁻¹ after 100 cycles, which is only 61% of the second reversible discharge specific capacity. In sharp contrast, the discharge specific capacities of S@pPAN-2, S@pPAN-5 and S@pPAN-8 are significantly improved. When the vapor pressure rises to 2 MPa in the heating process, the S@pPAN-2 composite displays the discharge specific capacity of 1007 mA h g⁻¹ after 100 cycles with a retention rate of 76%. A maximum in the discharge specific capacity of 1357 mA h g⁻¹ (616.8 mA h g_composite⁻¹) after 100 cycles is observed for S@pPAN-5 prepared under 5 MPa vapor pressure, which is 88% of the second reversible discharge specific capacity. Apparently, even under higher vapor pressure in the heating process, the discharge specific capacity has no further improvement. By contrast with S@pPAN-5 composite, the S@pPAN-8 composite only reaches 1277 mA h g⁻¹ after 100 cycles with a retention rate of 83%. The more excellent cycling performance of S@pPAN-5 composite can be attributed to the more complete conductive molecular structure, smaller particle size, higher surface area and electric conductivity than other samples.

In order to further examine the effectiveness of vapor pressure on the electrochemical performance of the S@pPAN cathodes, the rate capability of S@pPAN-0, S@pPAN-2, S@pPAN-5 and S@pPAN-8 is investigated and shown in Fig. 4c, respectively. For S@pPAN-5 composite, relatively stable specific capacities around 1510, 1420, 1233 and 1008 mA h g⁻¹ are obtained at current density of 200, 400, 1000 and 2000 mA g⁻¹, respectively. The specific capacity retention is about 66%, when the rate enhances from 200 mA g⁻¹ to 2000 mA g⁻¹. But as the vapor pressure rises to 8 MPa in heating process, the rate performance of S@pPAN-8 composite has no further improvement and even tiny decrease compared with S@pPAN-5 composite. For a sharp comparison, the specific capacity of S@pPAN-0 and S@pPAN-2 composites drops quickly with increased current density. The specific capacity for S@pPAN-0 and S@pPAN-2 composites at 2000 mA g⁻¹ current density is 201 mA h g⁻¹ and 590 mA h g⁻¹, corresponding to capacity retention of 18% and 46% under the same condition, respectively.

Conclusions

The high performance sulfurized polyacrylonitrile composites have been successfully synthesized in a high pressure reactor filled with argon gas under certain vapor pressure. This preparation method can promote the interaction between sulfur and polyacrylonitrile, resulting in smaller particle size, lower degree of graphitization, higher surface area and electric conductivity of S@pPAN composites than that prepared under flowing argon atmosphere. Moreover, the atomic ratio of C to H gradually increases to 3 : 1 with increasing the vapor pressure, indicating the more thorough reaction between sulfur and polyacrylonitrile and the more complete conductive molecular structure. So when the preparation vapor pressure increases to 5 MPa, the as prepared S@pPAN-5 composite presents a reversible discharge specific capacity of about 1542 mA h g⁻¹ (701 mA h g_composite⁻¹) in the second cycle, and capacity retention of 88% after 100 cycles of this value. The present study provides an effective and feasible approach, which brings the Li/S batteries a step closer to practical realization.

Acknowledgements

This work was financially supported by the Funds from Beijing Science and Technology Project (no. D151100003015002), and the National Key Research and Development Program of China (no. 2016YFB0100200).

Notes and references

1. R. D. Rauh, K. M. Abraham, G. F. Pearson, J. K. Surprenant and S. B. Brummer, J. Electrochem. Soc., 1979, 126, 523.
2. D. Peramunage and S. Licht, Science, 1993, 261, 1029.
3. A. Manthiram, Y. Fu, S.-H. Chung, C. Zu and Y.-S. Su, Chem. Rev., 2014, 114, 11751.
4. L. X. Yuan, J. K. Feng, X. P. Ai, Y. L. Cao, S. L. Chen and H. X. Yang, Electrochem. Commun., 2006, 8, 610.
5. G. Jeong, Y. U. Kim, H. Kim, Y. J. Kim and H. J. Sohn, Energy Environ. Sci., 2011, 4, 1986.
6. S.-E. Cheon, K.-S. Ko, J.-H. Cho, S.-W. Kim, E.-Y. Chin and H.-T. Kim, J. Electrochem. Soc., 2003, 150, A800.
7. S. S. Zhang, J. Power Sources, 2013, 231, 153.
8. A. Hayashi, T. Ohtomo, F. Mizuno, K. Tadanaga and M. Tatsumisago, Electrochem. Commun., 2003, 5, 701.
9. X. Ji, K. T. Lee and L. F. Nazar, Nat. Mater., 2009, 8, 500.
10. M. J. Wang, W. W. Wang, A. B. Wang, K. G. Yuan, L. X. Miao, X. L. Zhang, Y. Q. Huang, Z. B. Yu and J. Y. Qiu, Chem. Commun., 2013, 49, 10263.
11. S. Xiao, S. Liu, J. Zhang and Y. Wang, J. Power Sources, 2015, 293, 119.
12. L. X. Miao, W. K. Wang, A. B. Wang, K. G. Yuan and Y. S. Yang, J. Mater. Chem. A, 2013, 1, 11659.
13. G. Y. Zheng, Y. Yang, J. J. Cha, S. S. Hong and Y. Cui, Nano Lett., 2011, 11, 4462.
14. F. L. Zeng, W. K. Wang, A. B. Wang, K. G. Yuan, Z. Q. Jin and Y. S. Yang, ACS Appl. Mater. Interfaces, 2015, 7(47), 26257.
15. F. Pei, T. H. An, J. Zang, X. J. Zhao, X. L. Fang, M. S. Zheng, Q. F. Dong and N. F. Zheng, Adv. Energy Mater., 2016, 6, 1502539.
16. C. Y. Zhang, W. K. Wang, A. B. Wang, K. G. Yuan, Y. S. Yang and Y. L. Lu, J. Electrochem. Soc., 2015, 162(6), A1067.
17. J. Wang, J. Yang, J. Xie and N. Xu, Adv. Mater., 2002, 14, 963.
18. Y. G. Zhang, Y. Zhao, Z. Bakenov, M.-R. Babaa, A. Konarov, C. Ding and P. Chen, J. Electrochem. Soc., 2013, 160(8), A1194.
19. L. C. Yin, J. L. Wang, F. J. Lin, J. Yang and Y. Nuli, Energy Environ. Sci., 2012, 5(5), 6966.
20. J. Li, K. Li, M. Q. Li, D. Gosselink, Y. Zhang and P. Chen, J. Power Sources, 2014, 252, 107.
21. Y. G. Zhang, Y. Zhao, A. Yermukhambetova, Z. Bakenov and P. Chen, J. Mater. Chem. A, 2013, 1, 295.
22. H. W. Chen, C. H. Wang, C. J. Hu, J. S. Zhang, S. Gao, W. Lu and L. W. Chen, J. Mater. Chem. A, 2015, 3, 1392.
23. K. Jeddi, M. Ghaznavi and P. Chen, J. Mater. Chem. A, 2013, 1, 2769.
24. A. Konarov, D. Gosselink, T. N. L. Doan, Y. G. Zhang, Y. Zhao and P. Chen, J. Power Sources, 2014, 259, 183.
25. J. Wang, Y. J. C. Wan, K. Du, J. Xie and N. Xu, Adv. Funct. Mater., 2003, 13, 487.
26. J. Wang, Y. He and J. Yang, Adv. Mater., 2015, 27, 569.
27. R. Kostecki, B. Schnyder, D. Alliata, X. Song, K. Kinoshita and R. Kotz, Thin Solid Films, 2001, 396, 36.
28. M. Matthews, M. Pimenta, G. Dresselhaus, M. Dresselhaus and M. Endo, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 59, R6585.
29. G. Zhou, E. Paek, G. S. Hwang and A. Manthiram, Nat. Commun., 2015, 6, 7760.
30. X. X. Gu, C. J. Tong, C. Lai, J. X. Qiu, X. X. Huang, W. L. Yang, B. Wen, L. M. Liu, Y. L. Hou and S. Q. Zhang, J. Mater. Chem. A, 2015, 3, 16670.
31. Y. Matsumura, S. Wang and J. Mondori, Carbon, 1995, 33, 1457.
32. X. G. Yu, J. Y. Xie, J. Yang, H. J. Huang, K. Wang and Z. S. Wen, J. Electroanal. Chem., 2004, 573, 121.
33. S. Brauman, J. Polym. Sci., Part A: Polym. Chem., 1989, 27, 3285.
34. S. Wei, L. Ma, K. E. Hendrickson, Z. Tu and L. A. Archer, J. Am. Chem. Soc., 2015, 137, 12143.
35. S. S. Zhang, Energies, 2014, 7, 4588.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra24443b

---