Separator modified by Ketjen black for enhanced electrochemical performance of lithium–sulfur batteries

Abstract

A routine separator modified by a Ketjen black (KB) layer on the cathode side has been investigated to improve the electrochemical performances of Li–S batteries. The KB modified separator was prepared by a facile slurry coating method which offers a low-cost approach to solve the difficulties of Li–S batteries. Li–S cells assembled with this KB coated separator present excellent electrochemical performances in comparison with that of cells with a routine separator. The initial discharge capacity reaches 1318 mA h g⁻¹ at 0.1C, and the reversible discharge capacity is maintained at 815 mA h g⁻¹ after 100 cycles at 1C implying high capacity retention. Meanwhile, it achieves a discharge capacity of 934 mA h g⁻¹ even at 2C demonstrating an excellent rate performance. Furthermore, electrochemical impedance spectroscopy (EIS) shows that the KB separator sample displays a lower charge transfer resistance which is beneficial for the electrochemical kinetics. The improved performance is supposed to be attributed to the porous architecture of the Ketjen black (KB) layer on the routine separator, which served as a physical barrier to block dissolved lithium polysulfides and an upper current collector to facilitate the transition of ions and electrons.

---

Introduction

Li–S batteries have been considered as a potential candidate for next generation electrical energy storage systems since the 1990s due to their ultra-high theoretical energy density of 2600 W h kg⁻¹ which originates from the high theoretical capacity of sulfur (1672 mA h g⁻¹) and lithium anode (3860 mA h g⁻¹).^1–3 Besides, as the active material of cathode, sulfur has many advantages such as natural abundance, low-cost and environmental friendly characteristics.^4,5 However, the commercialization of Li–S batteries has been greatly plagued by its inherent problems, including the insulating nature of sulfur, large volume expansion and soluble intermediates, which can lead to the low utilization of active material, poor coulombic efficiency and great degradation of capacity after a few cycles.⁶ The most sticky obstacle is the shuttle effect of soluble lithium polysulfides (Li₂S_x, 2 < x < 8) by which sulfur active mass will lost through redox reactions of lithium polysulfides at both the cathode and anode surfaces.⁷ On account of these problems, numerous methods were proposed. For example, conductive carbons, such as ordered mesoporous carbon nanoparticles,^8–10 carbon nanotubes,^11,12 carbon nanofibers¹³ and graphene,¹⁴ were introduced to sulfur based cathode to generate essential electrical contact and constrain lithium polysulfides to a certain extent. However, the open architecture cannot absorb polysulfide sufficiently after long cycles. Recently, conductive porous carbon materials including microporous carbon paper¹⁵ and MWCNT interlayer¹⁶ were used as interlayer by Manthiram's group to serve as the upper current collector and polysulfide absorber inserting between the cathode and the separator. However, the loose structure of a lightweight interlayer would affect the specific volumetric capacity of Li–S battery significantly, consequently cannot ensure its normal function in cells during repeated cycling and may reduce the polysulfide-trapping capability.¹⁷ Inspired by the conception of interlayer, modified separator was therefore proposed, which is regarded as a more accessible approach to enhance the electrochemical performance of Li–S batteries with a much thinner coated-layer than interlayer. Various conductive carbon materials such as super P, microporous carbon particles, graphene-sheet, MWCNTs were coated on the surface of conventional separator due to the efficient absorption capability and high electrical conductivity.^18–20 It turns out that carbon-coated separator effectively limits the shuttle effect of lithium polysulfides and shows excellent capacity retention.²¹

In this work, we were motivated to propose a facile and effective approach to prepare a Ketjen black (KB)-coated layer on the surface of separator via a doctor blade. KB (EC600JD) which we used in the experiment is commercially available microporous carbon nanoparticles with high specific surface area and high conductivity. The electrochemical performance of Li–S batteries using KB-coated separator can be significantly improved, the initial discharge capacity reaches 1318 mA h g⁻¹ at the current density of 0.1C and the reversible discharge capacity maintains at 820 mA h g⁻¹ after 100 cycles at 1C (1672 mA h g⁻¹), which is much higher than that of Li–S cells with routine separator. Cyclic voltammogram (CV) of the KB separator sample shows sharper reduction peak than that of the routine separator sample, indicating fast reaction kinetics and rapid decrease of electrode polarization.²² Electrochemical impedance spectroscopy (EIS) of the KB separator sample presents lower R_ct than that of the routine separator sample, implicating the upper current collector function of the KB layer.¹⁸ The enhanced electrochemical performances demonstrate that KB-coated separator is beneficial for further promotion of the practical application of Li–S batteries.

Experimental details

Preparation and characterization of KB-coated separator

Fig. 1(a) schematically illustrates the fabrication processes of the functional KB layer-coated separator. Commercial KB powder and conductive carbon powder (Super P Timcal) were added to the 1.3 wt% polyvinylidene fluoride (PVDF 6020 Solef) solution with N-methyl-2-pyrrolidinone solution (NMP) as the dispersant. The weight ratio of KB/SP/PVDF was fixed at 70/10/20. After vigorous deaerating stirring for 8 min, the mixed slurry was coated evenly on the surface of routine separator (Celgard 2400), followed by a drying process at 40 °C for 15 h in a vacuum oven. Finally the KB coated separator was cut into disc with the diameter of 19 mm. The thickness of the functional KB layer was approximately 10 μm, and the mass is about 0.18 mg cm⁻². For brevity, the modified separator with a functional KB coating layer is referred to as KB separator. Scanning electron microscopy (SEM, JSM-7001F, Japan) and energy-dispersive spectroscopy (EDS) were employed to characterize the morphology and elements distribution of the KB layer before and after discharge/charge processes respectively.

Fig. 1 The schematic diagram of (a) KB separator preparation and (b) a Li–S cell configuration with a KB separator.

Cell assembly and characterization

KB/S composite was uniformly mixed by high energy milling for 4 h at the mass ratio of 1/4 and dried at 60 °C overnight in a circulation oven, followed by traditional heat treatment of heating at 155 °C for 12 h in a sealed reactor. The working cathode was fabricated by mixing 80 wt% KB/S composite, 5 wt% conductive carbon black (Super P Timcal) and 15 wt% PVDF in NMP solution. The mixed slurry was spread on aluminum foil uniformly by an automatic coating machine which can control the layer thickness and coating speed. The slurry coated foil was dried at 60 °C for 15 h in a vacuum oven and cut into disc with the diameter of 12 mm to serve as the cathode electrode. Li–S coin cells (CR2025) were assembled in an argon-filled glove box by using a lithium foil as the anode electrode and routine/KB membrane (Celgard 2400) as the separator. The battery configuration of a Li–S cell with the functional KB separator is displayed in Fig. 1(b). The electrolyte was 1,3-dioxolane and 1,2-dimethoxyethane mixed by the volume ratio of 1 : 1 with 1 M lithium bis(trifluoromethanesulfone)imide and 0.1 M LiNO₃ addition, the amount of the electrolyte in each cell is about 50 μL. The assembled Li–S coin cells were galvanostatically discharged/charged at different current densities between 1.7 and 2.8 V (vs. Li/Li⁺) using a CT2001A cell test instrument (LAND model, Wuhan RAMBO testing equipment, Co. Ltd). The CV measurement were carried out at the scan rate of 0.2 mV s⁻¹ in a voltage range of 1.7–2.8 V. The electrochemical impedance spectroscopy (EIS) analysis were carried out at the frequency range of 1.0 MHz to 0.1 Hz with an alternating current (AC) voltage of 10 mV using a VMP2 electrochemical workstation (DHS Instruments Co. Ltd). All of the measurements of the samples were conducted at 25 °C.

Results and discussion

Characterization of the routine/KB separator

Fig. 2(a) shows the punched KB separator with a diameter of 19 mm. Fig. 2(b) shows robust structure of KB layer, which is the fundamental factor of a physical barrier to block the lithium polysulfides during the cycling processes. The sectional view of the KB layer on pristine separator via SEM is presented in Fig. 2(c), the thicknesses are about 10 μm and 20 μm, respectively. It is worthy to mention that the weight of the KB layer is about 0.18 mg cm⁻², which is much lighter than 1.3 mg cm⁻² of the pristine separator, implicating the introduction of KB layer cannot be a burden of separator. Fig. 2(d) shows the plane view of a pristine separator, it exhibits evenly distributed submicron pore structure which guarantees the fast ion transportation for chemical reactions.²³ In contrast, KB nanoparticles uniformly stacked together forming nano-carbon sponge on the pristine separator, as shown in Fig. 2(e), meaning it can serve as an effective physical barrier to prevent the diffusion of lithium polysufides. Furthermore, the interspace between KB nanoparticles can form interconnected tunnel which is beneficial for the electrolyte infiltration and the electron transportation. All of these advantages promote a better utilization of active material which directly leads to a better electrochemical performance.²⁴ These conclusions are well supported by the SEM image and the corresponding EDS elemental mapping of the KB modified separator before and after cycling. As shown in Fig. 3(a), no elemental sulfur (S) could be discovered before cycling. While Fig. 3(b) illustrates that elemental S is uniformly distributed in the KB barrier layer, demonstrating the effective adsorption of the lithium polysulfides.²⁵ Elemental oxygen (O) and fluorine (F) are recognizable in the KBs after cycles, which should attribute to the abundance of space and good electrolyte absorption ability.²²

Fig. 2 Digital photos of (a) the punched KB separator; (b) the demonstration of the flexibility and mechanical strength of the KB separator; SEM images of (c) a typical sectional view of the KB separator, (d) pristine separator (surface) and (e) KB separator (surface).

Fig. 3 SEM images and elemental mapping of the KB separator (a) before and (b) after 100 cycles at 1C.

The SEM image of the cathode electrode of Li–S cell with KB separator after cycles is shown in Fig. 4(a). The KB nanoparticles can be distinguished and no obvious nonconductive bulks could be found (especially in the area of yellow square), which indicates that the sulfur in the cathode electrode does not agglomerate apparently during dissolution and precipitation. There are two possible explanations for this phenomenon. Firstly, the uniformly stacked conductive KB nanoparticles can transfer electrons to reactivate the trapped active materials successfully during cycling, suppressing the formation of inactive precipitates.²⁶ Secondly, the sponge-like structure of KB layer is unfavourable for the formation of nonconductive agglomerations.¹⁹ Some of the dissolved lithium polysulfides are captured by the pores of KB barrier which can alleviate the sulfur deposition and agglomeration on the surface of cathode. For comparison, Fig. 4(b) shows the SEM picture of the cathode of routine separator sample after cycles presenting a sheet like morphology (especially in the area of yellow square) which is considered as the bulk sulfur agglomeration meaning that during cycling, the redistribution of sulphur is not uniform compared with the sample with KB separator. Based on the above analysis, it can be proved that the active materials of the KB separator sample can be well fixated within the cathode region, which can ensure efficient sulfur utilization and high reversibility.

Fig. 4 SEM images of KB/S cathode after cycling with (a) KB separator and (b) routine separator.

Electrochemical performance of cells with KB separator

The electrochemical performances of Li–S cells with KB separator were evaluated via EIS, CV, and galvanostatic charge–discharge testing. It is reasonable to suppose that the porous sponge-like KB layer can not only promote the transportation of electron between the insulating active materials and the conductive KB nanoparticles, buy also hold active materials during cycling. This may directly result in (I) a low resistance and (II) extraordinarily effective utilization of the trapped active material. To testify these enhancements, EIS analysis was carried out to compare the impedance of KB separator sample with routine separator sample. The Nyquist plots as well as the fitting curves of the cells with KB separator and routine separator before and after cycling are presented in Fig. 5. As illustrated in Fig. 5, the impedance plots are consisted of an intercept at high frequency on the real axis which represents the resistance of the electrolyte (R_e), a depressed semicircle at high frequency region which corresponds to the charge transfer resistance (R_ct) of the sulfur electrode, and an inclined line at low frequency region which reflects the Li⁺ diffusion into the active mass.^27,28 Fig. 5(a) and (b) are the plots of samples before and after cycling, respectively. It can be seen that the R_ct of KB samples are smaller than that of routine samples. In order to understand this result better, equivalent circuit model is used to calculate the values of relevant parameters, which are shown in Table 1.

Fig. 5 Nyquist plots for the Li–S cells (a) before cycling and (b) after cycling, (c) the relationship between Z_re and ω^−0.5 at low frequencies.

Table 1 Impedance parameters of the samples^a

Sample	Re (Ω)	Rct (Ω)	σw (Ω s^−0.5)	D (cm^2 s^−1)
a R-sample: routine separator sample, K-sample: KB separator sample.
R-sample before cycling	1.4	34.6	20.9	5.3 × 10^−11
R-sample after cycling	0.7	147.2	18.3	6.9 × 10^−11
K-sample before cycling	0.79	102.2	17.0	7.9 × 10^−11
K-sample after cycling	3.39	24.5	3.2	2.3 × 10^−9

---

The depressed semicircle at high frequency corresponds to the R_ct of the sulfur electrode. As listed in Table 1, the R_ct of KB samples are 102.2 Ω (before cycling) and 24.5 Ω (after cycling), which are smaller than that of routine samples 147.2 Ω (before cycling) and 34.6 Ω (after cycling) indicating the effects of the KB layer. The inclined line is attributed to the diffusion of the lithium ions into the bulk of the electrode material named as Warburg diffusion.²⁹ The Warburg diffusion coefficient σ_w can be obtained by eqn (1)

Z_re = R_e + R_ct + σ_wω^−0.5 (1)

Both R_e and R_ct are kinetics parameters independent of frequency. So σ_w are the slopes for the plots of Z_re vs. the reciprocal root square of the lower angular frequencies (ω^−0.5), which are presented in Fig. 5(c). The diffusion coefficient values of the lithium ions (D) for its diffusion into the bulk electrode materials have been calculated using eqn (2)³⁰ and are listed in Table 1 too.

D = 0.5(RT/AF²σ_wC)² (2)

where R is the gas constant (8.314 J mol⁻¹ K⁻¹), T is the room temperature (298.5 K), A is the area of the electrode surface (1.13 cm²), F is the Faraday's constant (9.65 × 10⁴ C mol⁻¹) and C is the molar concentration of Li⁺ ions (1.1 × 10⁻³ mol cm⁻³).

The differences between the values of D further prove the function of this KB layer in Table 1, which can greatly promote the transportation of Li ions and electrons. The D value of KB sample after cycling (2.3 × 10⁻⁹) is two orders of magnitude larger than that of routine sample after cycling (6.9 × 10⁻¹¹), proving the outstanding Li⁺ diffusion which will be beneficial for the redox process.

To further investigate the effect of KB barrier layer, CV was conducted to identify the redox reactions for the cells with KB/routine separator. Fig. 6(a) and (b) shows the first five CV cycles of the KB/routine separator samples, respectively. Compared with the routine separator sample shown in Fig. 6(b), Fig. 6(a) exhibits sharp redox peaks, relatively high cathodic peaks, excellent overlap of the first five cycles, and a larger covering area, indicating fast redox process, good stability and high capacity of the KB separator sample. The sharp redox peaks implicate the improvement of reaction kinetics caused by the excellent electrical conductivity and porous structure of the KB layer working as upper current collector. The overlapping CV curves demonstrate that the KB separator sample can obtain improved cycle stability and capacity, owing to the strong polysulfides adsorption and good electrical conductivity of KB. These results are similar to previous studies using super-P as the barrier layer.²¹

Fig. 6 Cyclic voltammogram curves of the Li–S cells with (a) KB separator and (b) routine separator at a scan rate of 0.2 mV s⁻¹.

In order to further illustrate the enhanced electrochemical properties of the KB separator sample, cycle and rate performance test was executed at the current density of 1672 mA g⁻¹ (1C) between 1.7 and 2.8 V. Considering that the content of sulfur in the cell with KB separator decreases from 64% (the mass ratio of the S/KB composite and cathode slurry is 4/1 and 8/1/1, respectively) to 51% (the average weight of KB barrier layer and sulfur cathode is 0.5 mg and 2.0 mg, correspondingly), the cells with 51% sulfur content cathode were also prepared and tested under the same condition for comparison. As shown in Fig. 7(a), it can be seen that the capacity of both cells decreases with an increased number of cycles. However, the KB separator sample shows obvious advantages over that of routine separator sample both in initial capacity and capacity retention overall. The initial discharge capacity of the routine separator sample with 51% sulfur content is approximately 1100 mA h g⁻¹ at 0.1C and decreases from 680 to 435 mA h g⁻¹ after 100 cycles when the current rate increased to 1C. In comparison, the initial discharge capacity of the KB separator sample with 64% sulfur content reaches 1318 mA h g⁻¹ at 0.1C and reduces from 946 to 815 mA h g⁻¹ after 100 cycles when the current rate increased to 1C. The cycle performance tested above is consistent with the results reported by Zhang et al. using N doped porous carbon as the coated layer.³¹ Fig. 7(a) also shows the cycling stability of samples at 2C. The KB separator sample has a discharge capacity of 681 mA h g⁻¹ after 100 cycles, while the routine one only has 386 mA h g⁻¹, almost half of the KB separator sample, further supporting the function of the KB layer. Consequently, we expect this KB layer can greatly suppress the shuttle effect which can lead to high capacity. Besides, the KB separator sample can maintain excellent cycling stability with 0.14% and 0.22% degradation per cycle during 100 cycles at 1C and 2C correspondingly, due to the KB layer coated on routine separator, serving as the conductor of the diffused active material and thus improving the utilization of sulfur which significantly lowered the capacity fade.

Fig. 7 Cycle performance (a) and rate performance (b) of Li–S cells with routine separator and carbon-coated separator. (c) Initial charge–discharge profiles of Li–S cell with routine separator and KB separator.

As shown in Fig. 7(b), the rate performances of the KB/routine separator samples can help to further convince the exceedingly good performance of the KB coated layer through comparison and analysis. The cell with KB separator exhibits good rate capability and excellent reversible capacity retention (934 mA h g⁻¹) even at 2C, while the reversible capacity retention of the cell with routine separator is 550 mA h g⁻¹. When the current density returns to 0.1C, the reversible capacity of the cells with KB/routine separator retains the discharge capacity of 1173 mA h g⁻¹ and 605 mA h g⁻¹, respectively. These results demonstrate excellent rate capability of the KB separator sample, which is caused by the good polysulfides absorption ability, the fast electron conduction and ion transport function of the KB coating.

The discharge/charge curves of the cells equipped with the KB/routine separator is presented in Fig. 7(c). During the process of discharge, there are two separate discharge plateaus indicating the existence of the two complete reduction reactions. It can be seen that the upper discharge plateau and lower discharge plateau are around 2.35 and 2.0 V which match well with the cathodic peaks in CV curves. The former and latter plateau correspond to the reaction of S₈ to discharge mediates known as long-chain polysulfides (Li₂S_x,4 < x < 8), and discharge mediates to discharge end products known as short-chain Li₂S₂/Li₂S, respectively.³² We can also find in the figure that the initial discharge capacity is substantially heightened, implicating a larger active material utilization of Li–S cells with the KB separator. Therefore, it is well validated that the KB separator can successfully restrain the diffusion of polysulfides, limit the loss of the active material and guarantee the integrity of redox reaction. All of the electrochemical performances can be further ground for the effect of the KB separator, which can improve the performance of cells fundamentally due to the effective absorption of diffused polysulfides and fast transportation of ions and electrons of the KB layer.

Conclusions

In conclusion, KB was introduced into Li–S batteries, serving as a barrier on the surface of routine separator. Li–S cells with the KB coated separator exhibit excellent electrochemical performances including a high initial discharge capacity of 1318 mA h g⁻¹ at 0.1C and reversible capacity of 815 mA h g⁻¹ after 100 cycles at 1C. What's more, the cells show extraordinary rate capability which can retain a capacity of 934 mA h g⁻¹ at a discharge current of 2C and maintain 1173 mA h g⁻¹ back to 0.1C. All the excellent electrochemical performances of the KB separator sample should be attributed to the KB layer, which guarantees electrolyte infiltration and serves as a barrier for polysulfides due to its high specific surface area and pore volume. More importantly, the KB layer can affiliate the transportation of ions and electrons leading to high reutilization of active material and excellent discharge capacity retention. Thus, the introduction of the KB separator would be a facile approach to get advanced Li–S batteries and promote the possibility of its commercialization.

Acknowledgements

This work was financially supported by the Start-up Fund of Jiangsu University (Grant No. 14JDG060, 14JDG058), the Postdoctoral Fund of Jiangsu Province (Grant No. 1402196C), open fund of the Laboratory of Solid State Microstructures, Nanjing University (M28035), the National Natural Science Foundation of China (Grant No. 21401081, 51274106, 51474113, 51474037).

Notes and references

1. A. Manthiram, S. H. Chung and C. X. Zu, Adv. Mater., 2015, 27, 1980.
2. W. Y. Li, G. Y. Zheng, Y. Yang, Z. W. Seh, N. A. Liu and Y. Cui, Proc. Natl. Acad. Sci. U. S. A., 2013, 110, 7148.
3. Y. X. Yin, S. Xin, Y. G. Guo and L. J. Wan, Angew. Chem., Int. Ed., 2013, 52, 2.
4. A. Manthiram, Y. Z. Fu and Y. S. Su, Acc. Chem. Res., 2012, 46, 1125.
5. Y. Yang, G. Y. Zheng and Y. Cui, Chem. Soc. Rev., 2013, 42, 3018.
6. D. Bresser, S. Passerini and B. Scrosati, Chem. Commun., 2013, 49, 10545.
7. S. Evers and L. F. Nazar, Acc. Chem. Res., 2012, 46, 1135.
8. X. L. Ji, K. T. Lee and L. F. Nazar, Nat. Mater., 2009, 8, 500.
9. Y. Yang, G. H. Yu, J. J. Cha, H. Wu, M. Vosgueritchian, Y. Yao, Z. A. Bao and Y. Cui, ACS Nano, 2011, 5, 9187.
10. F. G. Sun, J. T. Wang, H. C. Chen, W. C. Li, W. M. Qiao, D. H. Long and L. C. Ling, ACS Appl. Mater. Interfaces, 2013, 5, 5630.
11. J. C. Guo, Y. H. Xu and C. S. Wang, Nano Lett., 2011, 11, 4288.
12. G. Y. Zheng, Y. Yang, J. J. Cha, S. S. Hong and Y. Cui, Nano Lett., 2011, 11, 4462.
13. G. Y. Zheng, Q. F. Zhang, J. J. Cha, Y. Yang, W. Y. Li, Z. W. Seh and Y. Cui, Nano Lett., 2013, 13, 1265.
14. H. L. Wang, Y. Yang, Y. Y. Liang, J. T. Robinson, Y. G. Li, A. Jackson, Y. Cui and H. J. Dai, Nano Lett., 2011, 11, 2644.
15. Y. S. Su and A. Manthiram, Nat. Commun., 2012, 3, 1166.
16. Y. S. Su and A. Manthiram, Chem. Commun., 2012, 48, 8817.
17. R. Singhal, S. H. Chung, A. Manthiram and V. Kalra, J. Mater. Chem. A, 2015, 3, 4530.
18. S. H. Chung and A. Manthiram, Adv. Funct. Mater., 2014, 24, 5299.
19. S. H. Chung and A. Manthiram, J. Phys. Chem. Lett., 2014, 5, 1978.
20. S. H. Chung and A. Manthiram, US provisional application61/988.656, 2014.
21. Z. Y. Zhang, Y. Q. Lai, Z. A. Zhang and J. Li, Solid State Ionics, 2015, 278, 166.
22. Z. Y. Zhang, Y. Q. Lai, Z. A. Zhang, K. Zhang and J. Li, Electrochim. Acta, 2014, 129, 55.
23. J. Q. Huang, Q. Zhang, H. J. Peng, X. Y. Liu, W. Z. Qian and F. Wei, Energy Environ. Sci., 2014, 7, 347.
24. Z. Q. Jin, K. Xie, X. B. Hong, Z. Q. Hu and X. Liu, J. Power Sources, 2012, 218, 163.
25. S. H. Chung and A. Manthiram, Adv. Mater., 2014, 26, 7352.
26. S. H. Chung and A. Manthiram, Electrochim. Acta, 2013, 107, 569.
27. B. Ding, L. F. Shen, G. Y. Xu, P. Nie and X. G. Zhang, Electrochim. Acta, 2013, 107, 78.
28. J. X. Wang, Q. B. Zhang, X. H. Li, Z. X. Wang, K. L. Zhang, H. J. Guo, G. C. Yan, B. Huang and Z. J. He, Electrochem. Commun., 2013, 36, 6.
29. Y. Cui, X. L. Zhao and R. S. Guo, Electrochim. Acta, 2010, 55, 926.
30. A. Y. Shenouda and H. K. Liu, J. Electrochem. Soc., 2010, 157, A1183.
31. Z. A. Zhang, G. C. Wang, Y. Q. Lai, J. Li, Z. Y. Zhang and W. Chen, J. Power Sources, 2015, 300, 157.
32. Z. Yuan, H. J. Peng, J. Q. Huang, X. Y. Liu, D. W. Wang, X. B. Cheng and Q. Zhang, Adv. Funct. Mater., 2014, 24, 6105.

---