High-performance dye-sensitized solar cell based on an electrospun poly(vinylidene fluoride-co-hexafluoropropylene)/cobalt sulfide nanocomposite membrane electrolyte

Abstract

Electrospun poly(vinylidene fluoride-co-hexafluoropropylene) (PVdF-HFP) nanocomposite membranes incorporated with different weight percentages (1, 2 and 3 wt%) of cobalt sulfide (CoS) were prepared by an electrospinning technique. The surface morphology, crystallinity, porosity and electrolyte uptake of the electrospun nanocomposite membranes were examined. The prepared electrospun PVdF-HFP/CoS nanocomposite membranes (esCPM) were activated by an ionic liquid electrolyte containing 0.5 M LiI, 0.05 M I₂, 0.5 M 4-tert-butylpyridine and 0.5 M 1-butyl-3-methylimidazolium iodide in acetonitrile to obtain electrospun PVdF-HFP/CoS nanocomposite membrane electrolytes (esCPMEs). The uniformly dispersed CoS nanoparticles increased charge transport and facilitated the diffusion of the redox couple in the electrolyte system. The electrochemical characteristics of the dye-sensitized solar cells depended on the amount of CoS incorporated into the esCPMEs. The photovoltaic performance of the dye-sensitized solar cell assembled using the esCPME incorporated with 1 wt% of CoS was measured and found to be 7.34%, which is higher than that of the dye-sensitized solar cell assembled using the esPME without CoS (6.42%).

---

Introduction

Dye-sensitized solar cells (DSSCs) are promising for use as a mainstream renewable energy resource as a result of their high energy efficiency, clean energy source and low cost. Since the first report of these devices by Grätzel,¹ overall efficiencies as high as 11% have been achieved using liquid electrolytes. However, technical problems such as electrolyte leakage, solvent evaporation, instability at high temperatures and flammability limit their long-term performance. Among the various alternatives to liquid electrolytes, inorganic or organic hole conductors,² ionic liquids³ and polymer gel electrolytes⁴ stand out because of their excellent properties, which include easy fabrication, low cost and good stability. However, the gel network in gel electrolytes hinders charge transport and gel electrolytes suffer from a low electrolyte loading per unit volume and, consequently, a low ionic conductivity. To rectify this problem in polymer gel electrolytes, electrospun polymer membrane electrolytes have been investigated for use in DSSC applications. These polymer electrolytes have several advantages over their liquid counterparts, such as no internal shorting, no leakage of the electrolyte and the production of non-combustible reaction products at the electrode surface. Polymer hosts such as poly(ethyleneoxide), poly(propylene oxide), poly(acrylonitrile), poly(methyl methacrylate), poly(vinyl chloride), poly(vinylidene fluoride) (PVdF) and poly(vinylidene fluoride-hexafluoro propylene) (PVdF-HFP) have been used in DSSC applications.⁵ Recent advances have proved that nanostructured metal chalcogenides such as FeS₂, CoS and NiS have great potential in energy conversion and storage because of their unique physical and chemical properties.^6–12 The advantages of using metal chalcogenides are their higher electrical conductivity and shorter path length for the transport of electrons.

We report here the development of electrospun PVdF-HFP/CoS nanocomposite membrane electrolytes (esCPMEs). CoS has a high electrocatalytic activity. The electrospinning technique is a top-down, simple, versatile and cost-effective approach to fabricating nanofibers with a high degree of reproducibility for various applications such as DSSCs, supercapacitors, Li ion batteries and nano-filters.^13–20 The electrospun polymer membranes have unique properties, such as a high porosity, a large surface area, a fully interconnected pore structure with sufficient mechanical strength and a three-dimensional network structure. They offer the advantage of the cohesive properties of a gel polymer electrolyte together with the diffusive nature of a liquid electrolyte as a result of their interconnected pores, which can easily entrap a large quantity of liquid electrolyte; this improves the interfacial contact between the two electrodes and the electrolyte.²¹ Room temperature ionic liquids are promising alternatives to volatile electrolytes as a result of their unique properties, such as negligible vapor pressure, excellent electrochemical and thermal stability, and high ionic conductivity.²² Fluorinated polymers are known to be photochemically stable even in the presence of TiO₂ and Pt. The higher conductivity of PVdF-HFP films is attributed to their higher amorphicity because the two randomly mixed monomers provide more free mobile ions.²³ CoS nanoparticles were introduced into the electrospun membrane to improve the interfacial and ionic conductivity of the electrospun polymer membrane electrolyte. The encapsulation of the nanoparticles addressed a vast range of problems, such as the low ionic conductivity, interfacial stability, dimensional mechanical strength and thermal stability. The addition of CoS nanoparticles will further increase the amorphous content, thereby increasing uptake of the electrolyte. The CoS nanoparticles also enhance the diffusion coefficient of I⁻₃ and reduce charge recombination at the TiO₂/electrolyte interface. The significance of this study lies in improving the kinetics of ion transport and thereby the electrochemical activity of the electrospun membrane electrolyte. Using this approach, the electrospun polymer membrane electrolyte can be used to increase the photoconversion efficiency of DSSCs.

Experimental

Materials

Acetone and N,N′-dimethylacetamide were obtained from Merck India Ltd. Lithium iodide, iodine, 4-tert-butylpyridine and acetonitrile were obtained from Sigma-Aldrich and the PVdF-HFP was obtained from Arkema (Kynar Flex 2801). All these chemicals were of analytical-reagent grade and were used without further purification. The cobalt sulfide nanoparticles were prepared by a simple hydrothermal method. CoCl₂·6H₂O solution (3 mmol) in deionized water was added to a 3 mmol L-cysteine solution in deionized water with constant and vigorous stirring. After 15 min the resulting mixture was heated and maintained at 190 °C for 6 h in an autoclave. The resulting product was then washed and vacuum-dried overnight at 50 °C.²⁴

Preparation of electrospun PVdF-HFP/CoS composite membranes

Different weight percentages of CoS (1, 2 and 3 wt%) were added to a 16 wt% solution of PVdF-HFP in a mixture of acetone/N,N′-dimethyl acetamide (7 : 3 wt%). The mixture was electrospun at 19 kV and the distance between the collector and the tip of the syringe was kept at 12 cm. The polymer solution was supplied to the stainless-steel needle (27 G) using a syringe pump at a flow-rate of 0.5 ml h⁻¹. The thickness of the membrane was controlled to 30 μm. The prepared nanofibrous composite membrane was vacuum-dried at 80 °C for 12 h to remove any residual solvent. The thickness of the esCPMs was reduced from about 30 to 20 μm by hot pressing.²¹

Characterization of electrospun PVdF-HFP/CoS composite membranes

The surface morphology of the esCPMs was examined by field-emission scanning electron microscopy (FESEM; Model JSM-7600F microscope). The structural characterization of the esCPM incorporated with different wt% of CoS was studied by X-ray diffraction (XRD; Rigaku, Ultima IV spectrometer) with nickel-filtered Cu Kα radiation in the range 20–80° with an increment of 0.05°. Fourier-transform infrared (FTIR) spectra were recorded for the electrospun polymer membrane (esPM), CoS and esCPM (Thermo Nicolet, Model 6700 spectrometer). The thermal behavior of the esCPMs was studied by differential scanning calorimetry (DSC) at a heating rate of 10°C min⁻¹ under a nitrogen atmosphere over the temperature range 30–280 °C (TA Instruments; Model Q600 SDT instrument).

The crystallinity (X_c) of the esCPMs was calculated as follows:²⁵

where ΔH^sample_m is the heat of melting of the sample and ΔH^*_m is the crystalline melting heat of PVdF (104.7 J g⁻¹).

The porosity (%) of the esCPMs was measured by immersing the nanocomposite membranes in 1-butanol for 2 h. The porosity (P) was calculated using the following equation:²⁶

where m_a is the weight of the esCPM after impregnation with 1-butanol, m_p is the weight of the esCPM before impregnation with 1-butanol, and ρ_a and ρ_p are the densities of 1-butanol and the dried esCPM, respectively.

To measure the uptake of the electrolyte by the esCPMs, the nanocomposite membranes were soaked in the ionic liquid electrolyte for 24 h. After activation, the electrolyte membranes were removed from the electrolyte solution and any excess of electrolyte on the electrospun membrane was wiped off using a Whatman filter paper. The electrolyte uptake (U) was estimated using the formula⁵

U(%) = [(m − m₀)/m₀] × 100 (3)

where m and m₀ are the mass of the wet and dry esCPMs, respectively.

The leakage of the electrolyte was calculated using the equation⁴

where R is the relative absorption ratio of the liquid electrolyte, M_PE,saturated is the mass of the polymer electrolyte when the electrospun membrane was fully saturated with the liquid electrolyte and M_PE is the mass of the electrospun membrane electrolyte after a time interval during which the saturated electrospun membrane electrolyte had been squeezed by pressing it between two filter papers.

The electrochemical impedance spectra and Tafel polarization measurements were used to verify the electrochemical activities. The esCPMs containing different weight percentages of CoS were soaked in the ionic liquid electrolyte to form the corresponding esCPMEs. The ionic conductivity of the resultant esCPMEs (σ) was measured by sandwiching the nanocomposite membrane electrolyte between two stainless-steel blocking electrodes using an AC impedance technique (Biologic, Model VSP) at 25 °C. The ionic conductivity (σ) of the esCPMEs was calculated using the equation²³

σ = [small script l]/RA (5)

where [small script l] is the polymer membrane thickness, A is the area of the esCPME and R is the bulk resistance. The bulk resistance was obtained from the complex impedance plot. The frequency limit was set between 1 mHz and 100 kHz with an AC amplitude of 10 mV. The thickness ([small script l]) of the esCPM was determined to be 20 μm using a digital micrometer. The area (A) of the polymer membrane was 1 cm².

The Tafel polarization measurements were carried out using symmetrical cells consisting of FTO/Pt/esCPMEs/Pt/FTO to reconfirm the electrocatalytic activity of the esCPMEs.²⁷

Fabrication of DSSCs

The DSSCs were assembled as reported previously by sandwiching a slice of esCPME between a dye-sensitized TiO₂ photoanode and a Pt counter electrode.²³ The dye-adsorbed titania photoanode and the Pt counter electrode were assembled using a 60 μm thick hot melt thermoplastic sealer (Surlyn). One drop of the electrolyte solution containing 0.5 M LiI, 0.05 M I₂, 0.5 M 4-tert-butylpyridine and 0.5 M 1-butyl-3-methylimidazoliun iodide in acetonitrile²¹ was introduced into the clamped electrodes through one of two small holes drilled in the counter electrode. DSSCs based on electrospun polymer membrane electrolytes (esPME) were also assembled using the same procedure for comparison.

Photovoltaic performance of DSSCs

The photovoltaic performance of DSSCs were measured using a calibrated AM 1.5 solar simulator (Oriel Instruments; Model 67005) with a light intensity of 100 mW cm⁻² and a computer-controlled digital source meter (Keithley; Model 2420). I–V measurements were carried out on the DSSCs after an aging period of 24 h. The assembled DSSCs were stored in a desiccator and electrochemical measurements were conducted every 48 h to study their long-term stability. The photoelectrochemical parameters, i.e. the fill factor (FF) and the light-to-electricity conversion efficiency (η) were calculated with the following equations:²⁸where J_sc is the short circuit current density (mA cm⁻²), V_oc is the open-circuit voltage (V), P_in is the incident light power (mW cm⁻²), and J_max and V_max are the current density (mA cm⁻²) and voltage (V) in the J–V curves, respectively, at the point of maximum power output. All the fabrication steps and characterization measurements were carried out in an ambient environment without a protective atmosphere. The values of the photovoltaic parameter were determined by taking the average values of three DSSCs for each system.

Results and discussion

Characterization of esCPMs and esCPMEs

The FESEM images of the prepared esPM and esCPM (1 wt%) are shown in Fig. 1a and b, respectively. It can be seen from Fig. 1a that the esPM has a three-dimensional network with a fully interconnected pore structure that is capable of entrapping large amounts of ionic liquid electrolyte, which is favorable for the transportation of the redox couple and results in an improvement in the ionic conductivity. A bubble-like structure is observed in Fig. 1b as a result of the CoS nanoparticles incorporated into the host esCPM. The esCPM has good mechanical strength as a result of its three-dimensional network structure with cross-linking points.²³ The esCPM (average diameter of fibers = 300–350 nm) has a maximum porosity of 94%, an electrolyte uptake of 410% and a leakage of 0.2. The increase in the uptake of the electrolyte on the addition of the CoS nanoparticles is mainly due to the increase in the porosity as a result of the increase in surface roughness. The CoS fillers decrease the degree of crystallinity, which provides a large free volume and many free surface sites capable of trapping the ionic liquid electrolyte. This further increases the ionic conductivity and the electrochemical performance. However, beyond the optimum concentration of CoS, a blocking effect is caused by the aggregation of CoS nanoparticles, which decreases the porosity and reduces the uptake of the ionic liquid electrolyte.

Fig. 1 FESEM images of (a) esPM and (b) esCPM with 1 wt% CoS.

The FTIR spectra of esPM, CoS and esCPM (1 wt%) are shown in Fig. 2. The peaks at about 1395, 867 and 485 cm⁻¹ are due to C–F₂ bending, wagging and stretching vibrations and the peak at 1185 is due to the C–C bond of PVdF. It is clear that the spectrum of the esPM (PVdF-HFP) contains peaks at 611, 872, 1064, 1180 and 1400 cm⁻¹ that correspond to vinylidene, the CH₂ wagging of the vinylidene band, and –C–F– stretching, scissoring and bending vibrations of the vinyl group, respectively. The peaks at 680 and 550 cm⁻¹ correspond to twisting in the vinyl group. For the CoS nanoparticles, the appearance of bands at 474.4 and 1115 cm⁻¹ corresponds to the S–O modes. The peak at 1527 cm⁻¹ is due to –OH stretching. No shift in the peaks was observed for the esCPM, indicating that there were no conjugate bonds or interactions between the PVdF-HFP and the CoS nanoparticles. This reveals that the CoS nanoparticles were incorporated into the host esCPM and there was no complex formation between the PVdF-HFP and the CoS nanoparticles.

Fig. 2 FTIR spectra of esPM, CoS and esCPM with 1 wt% CoS.

The variation in the crystallinity of PVdF-HFP as a result of the addition of different weight percentages of CoS nanoparticles was studied by XRD. Fig. 3 shows the XRD patterns for PVdF-HFP powder, CoS, esPM and esCPM. The diffraction peaks of the CoS nanoparticles could be matched with the standard XRD pattern (ICSD no. 029305). No impurity peak was observed, indicating the formation of pure CoS nanoparticles.²⁹ In the XRD pattern of PVdF-HFP, the major peaks at 18.2 and 20.0° correspond to the (100) and (020) crystalline planes, respectively. The intensity of the peaks was weaker for the electrospun PVdF-HFP, in which the migration of the polymer chain was more free.³⁰ The intensity of these peaks reduced considerably with the addition of CoS, indicating that the addition of CoS changed the partial crystallinity to a more amorphous state. This increase in the degree of amorphicity will help the migration of ions in the electrolyte system, resulting in an enhancement of the ionic conductivity. The decrease in the crystallinity suggested that the CoS nanoparticles were better dispersed in the electrospun membrane. The lowest crystallinity was found for 1 wt% CoS incorporated into the esCPM. However, on increasing the CoS content beyond 1 wt%, the nanoparticles tended to agglomerate and create a blocking effect on the mobility of the redox couple, increasing the crystallinity and decreasing the movement of the polymer chain. The incorporation of 1 wt% CoS into the esCPME provided a high ionic conductivity and increased the photovoltaic performance.

Fig. 3 XRD patterns of PVdF-HFP powder, CoS and esCPM with different wt% of CoS.

Fig. 4 shows the DSC thermograms of esPM and esCPM with 1 wt% CoS. The addition of CoS nanoparticles increased the amount of amorphous phase in the host. It was observed that esCPM with 1 wt% CoS had a higher amorphous content than the esPM, as seen in the XRD results. In the esCPME, the solar conversion efficiency depends strongly on the ionic conductivity. The amorphous content increased considerably with the incorporation of 1 wt% CoS. The amorphous content had a high ionic mobility and therefore enhanced the ionic conductivity. The percentage of crystallinity was calculated using eqn (1). The esCPM with 1 wt% CoS had a lower crystallinity (22.84%) than the esPM (31.97%), which ensured a higher uptake of electrolytes, leading to an enhanced charge carrier mobility and increased ionic conductivity.

Fig. 4 DSC analysis of esPM and esCPM with 1 wt% of CoS.

Table 1 and Fig. 5 give the ionic conductivity of esCPMEs with different weight percentages of CoS as calculated from the bulk resistance obtained from complex impedance measurements at ambient temperature. The inclusion of CoS nanoparticles improved the decree of amorphicity and provided a diffusion path in the electrospun membrane. Thus the improvement in ionic conductivity in the esCPMEs is due to its amorphous structure, which encourages the dissociation of charge carriers from the interactive bonds and provides favorable conditions for the rapid migration of iodide and triiodide ions, resulting in an improvement in the conductivity of the ionic liquid. The esCPME is thought to consist of a solid fibrous phase, an amorphous swollen fibrous phase and a liquid phase in its pores. This indicates that the partial swelling of the polymer fibers with a large surface area significantly contributes to increasing the stability of the electrolyte in electrochemical environments, although ionic conduction mainly occurred through the entrapped liquid electrolyte in a fully interconnected pore structure. The liquid electrolyte that accumulated between the interconnected pores present in the esCPM acted as a gel as well as liquid-like manner.

Table 1 Effect of different amounts of CoS incorporated into the esCPMEs on the ionic conductivity

Electrolyte	Ionic conductivity(× 10^−3 S cm^−1)	U (%)	P (%)
esPME	5.83	340	84
esCPME (1 wt% CoS)	36.43	410	94
esCPME (2 wt% CoS)	29.17	392	90
esCPME (3 wt% CoS)	16.22	376	87

---

Fig. 5 Nyquist plots of ionic conductivity for esCPMEs with different wt% CoS.

The increase in ionic conductivity with the addition of CoS was related to both the generation of a free volume at the interface of the inorganic materials and surface interactions with the inorganic iodide. The CoS provides an increase in molecular ordering and ionic conductivity as well as the formation of a diffusion path for the redox species, thereby increasing the electrochemical activity of the esCPMEs with CoS. The reduction in the ionic conductivity on increasing the CoS content above 1 wt% is a result of the agglomeration of the CoS nanoparticles. The agglomeration makes the polymer membrane stiffer and obstructs the segmental motion of the polymer chains, which hinders the transport of the charge carriers. This blocking effect decreases the ionic conductivity. Fig. 6 illustrates the formation of the esCPME and the blocking effect resulting from the agglomeration of CoS beyond its optimum concentration. The ionic conductivity of esCPME with 1 wt% CoS had a maximum value of 36.43 × 10⁻³ S cm⁻¹. The power conversion efficiency of the DSSCs depends on the ionic conductivity of the esCPME. Thus the esCPME with 1 wt% CoS may have a high power conversion efficiency.

Fig. 6 Formation of esCPM and blocking effect resulting from the agglomeration of CoS at higher than optimum concentration.

Fig. 7 shows the Tafel curves of esCPMEs with different wt% of CoS. The slope of the polarization zone represents the charge transfer resistance (R_ct), which is inversely proportional to the exchange current density (J₀). This is an important indicator for assessing the electrochemical reaction kinetics. The steep slope of the Tafel zone has a large J₀ and small R_ct values. It can be seen that the esCPME with 1 wt% CoS has a larger slope and higher J_lim than the esCPMEs with other weight percentages of CoS. Values for J₀ can be obtained from the intersection of the linear cathodic and anodic regions and values of J_lim can be obtained by drawing a straight line toward the current density axis from a point in both anodic and cathodic regions in the Tafel polarization curves. This results indicate the higher catalytic activity toward the triiodide couple and the higher diffusion velocity of the facial redox reaction in the esCPME.³¹

Fig. 7 Tafel polarization curves of the symmetrical cells assembled using esCPME with different wt% CoS.

Photovoltaic performance of DSSCs

Fig. 8 (Table 2) shows the photocurrent density–voltage (J–V) curves obtained at a light intensity of 100 mW cm⁻² under standard global AM 1.5 irradiation. The DSSCs assembled using esCPME (1 wt% CoS) had a power conversion efficiency of 7.34%, higher than that of the esPME (6.42%). The increasingly amorphous nature of the esCPME on the addition of CoS nanoparticles improved the penetration of the esCPME into the porous TiO₂ electrode, thereby decreasing the interfacial recombination of electrons and leading to higher J_sc and FF values. The electrocatalytic activity of the esCPME reduces the over-potential, which, in turn, increases the V_oc value. Fig. 9 illustrates the mechanism for the increased reaction kinetics of the DSSCs assembled using the esCPME. At the optimum concentration of CoS (1 wt%), the reaction kinetics increased and hence the photovoltaic performance was also increased. This may be a result of the direct conductive pathway for the electrons provided by the CoS nanoparticles and their electrocatalytic activity. However, further increases in the concentration of CoS result in poor penetration of the esCPME into TiO₂ photoanode due to stiffness caused by agglomeration. The direct contact between the CoS nanoparticles may form a short circuit, leading to a decrease in the photovoltaic performance. The increase in the J_sc values from 13.10 to 14.42 is mainly attributed to the low charge transfer resistance and higher electrocatalytic activity in the reduction of the I₃⁻/I⁻ electrolyte. Thus the esCPME with incorporated CoS ensured effective charge transfer at the CE–electrolyte interface, with a lower recombination rate in the DSSC, which ultimately increased the power conversion efficiency.

Fig. 8 Photocurrent density–voltage (J–V) curves for DSSCs based on (a) esPME and (b) esCPME with 1 wt% CoS.

Table 2 Photovoltaic performance of DSSCs based on esPME and the esCPME with 1 wt% CoS

Electrolyte	Jsc (mA cm^−2)	Voc (V)	FF	Efficiency (%)
esPME	13.10	0.71	69	6.42
esCPME (1 wt% CoS)	14.42	0.73	70	7.34

---

Fig. 9 Mechanism of increased reaction kinetics of DSSCs assembled using esCPME.

The stability of the DSSCs assembled using the esPME and esCPME (1 wt%) was studied over a period of 30 days (Fig. 10). Both the DSSCs retained about 99% of their initial value. Notably, no decay was observed in the overall power conversion efficiency of either of the DSSCs. This might be a result of their three-dimensional network with interconnected pores. These interconnected pores are able to trap the liquid electrolyte and, as a consequence, the resulting esPME/esCPME acts as both a liquid and a gel electrolyte. Thus both the esPME and esCPME appear to promote interfacial contact between the dye-adsorbed TiO₂ electrode and the Pt counter electrode, which helps to give a more stable photovoltaic performance.

Fig. 10 Normalized efficiency of (a) esPME and (b) esCPME with 1 wt% CoS.

Conclusions

The esCPMs with incorporated CoS were successfully prepared by an electrospinning technique and their properties compared with those of the esPM. Their surface morphology, crystallinity, thermal behavior, percentage porosity and percentage electrolyte uptake were studied. They were then activated by a liquid electrolyte and the ionic conductivity and Tafel polarization were measured. The esCPME (with 1 wt% CoS) had a higher ionic conductivity of 36.43 × 10⁻³ S cm⁻¹. The esCPME with 1 wt% CoS showed a superior photovoltaic performance. The increase in the active surface area of esCPME with the incorporation of CoS could promote the electrocatalytic activity of the I₃⁻ reduction and the total current for the I₃⁻/I⁻ redox reaction, resulting in an enhancement of the J_sc value of the DSSC. Thus the DSSC assembled using esCPME with 1 wt% CoS showed an improved power conversion efficiency of 7.34%, which is higher than that of the DSSC assembled using esPME (6.42%).

Acknowledgements

The authors gratefully acknowledge the CSIR (Ref. no. 01/2359/10/EMR-II) New Delhi for financial support and also the CIF of Pondicherry University for extending the instrumentation facilities.

References

1. B. O'Regan and M. Grätzel, Nature, 1991, 353, 737–740.
2. I.-K. Ding, J. Melas-Kyriazi, N.-L. Cevey-Ha, K. G. Chittibabu, S. M. Zakeeruddin, M. Grätzel and M. D. McGehee, Org. Electron., 2010, 11, 1217–1222.
3. P. Wang, S. M. Zakeeruddin, J.-E. Moser and M. Grätzel, J. Phys. Chem. B, 2003, 107, 13280–13285.
4. N. Angulakshmi and A. M. Stephan, Electrochim. Acta, 2014, 127, 167–172.
5. J. Zhao, S.-G. Jo and D.-W. Kim, Electrochim. Acta, 2014, 142, 261–267.
6. Q.-H. Huang, T. Ling, S.-Z. Qiao and X.-W. Du, J. Mater. Chem. A, 2013, 1, 11828.
7. Y.-C. Wang, D.-Y. Wang, Y.-T. Jiang, H.-A. Chen, C.-C. Chen, K.-C. Ho, H.-L. Chou and C.-W. Chen, Angew. Chem., Int. Ed. Engl., 2013, 52, 6694–6698.
8. H. Sun, D. Qin, S. Huang, X. Guo, D. Li, Y. Luo and Q. Meng, Energy Environ. Sci., 2011, 4, 2630.
9. Z. Ku, X. Li, G. Liu, H. Wang, Y. Rong, M. Xu, L. Liu, M. Hu, Y. Yang and H. Han, J. Mater. Chem. A, 2013, 1, 237.
10. J.-Y. Lin, J.-H. Liao and T.-C. Wei, Electrochem. Solid-State Lett., 2011, 14, D41.
11. J.-Y. Lin and J.-H. Liao, J. Electrochem. Soc., 2012, 159, D65.
12. X. Miao, K. Pan, G. Wang, Y. Liao, L. Wang, W. Zhou, B. Jiang, Q. Pan and G. Tian, Chem.–Eur. J., 2014, 20, 474–482.
13. X. Zhang, P. Suresh Kumar, V. Aravindan, H. H. Liu, J. Sundaramurthy, S. G. Mhaisalkar, H. M. Duong, S. Ramakrishna and S. Madhavi, J. Phys. Chem. C, 2012, 116, 14780–14788.
14. X. Zhang, H. Liu, S. Petnikota, S. Ramakrishna and H. J. Fan, J. Mater. Chem. A, 2014, 2, 10835–10841.
15. Z. Xiang, B. R. S. Rajaraman, H. Liu and S. Ramakrishna, RSC Adv., 2014, 28987–29011.
16. X. Zhang, V. Thavasi, S. G. Mhaisalkar and S. Ramakrishna, Nanoscale, 2012, 4, 1707.
17. S. Homaeigohar and M. Elbahri, Materials, 2014, 7, 1017–1045.
18. S. Sundarrajan, K. Luck, S. Huat and S. Ramakrishna, Procedia Eng., 2014, 75, 159–163.
19. K. Yoon, K. Kim, X. Wang, D. Fang, B. S. Hsiao and B. Chu, Polymer, 2006, 47, 2434–2441.
20. S. Ramakrishnan, M. Dhakshnamoorthy, E. J. Jelmy, R. Vasanthakumari and N. K. Kothurkar, RSC Adv., 2014, 4, 9743.
21. E. Vijayakumar, A. Subramania, Z. Fei and P. J. Dyson, J. Appl. Polym. Sci., 2015, 132, 42032.
22. Y.-L. Lee, Y.-J. Shen and Y.-M. Yang, Nanotechnology, 2008, 19, 455201.
23. A. R. S. Priya, A. Subramania, Y.-S. Jung and K.-J. Kim, Langmuir, 2008, 24, 9816–9819.
24. S.-J. Bao, C. M. Li, C.-X. Guo and Y. Qiao, J. Power Sources, 2008, 180, 676–681.
25. A. Subramania, N. T. K. Sundaram, A. R. S. Priya and G. V. Kumar, J. Membr. Sci., 2007, 294, 8–15.
26. J. Hao, Q. Xiao, G. Lei, Z. Li and L. Wu, Electrochim. Acta, 2014, 125, 450–456.
27. S. Yuan, Q. Tang, B. He and P. Yang, J. Power Sources, 2014, 254, 98–105.
28. A. Subramania, E. Vijayakumar, N. Sivasankar, A. R. Sathiya Priya and K.-J. Kim, Ionics, 2013, 19, 1649–1653.
29. G. Panthi, N. A. M. Barakat, K. Abdelrazek, A. Yousef, K. Jeon and H. Yong, Ceram. Int., 2013, 39, 1469–1476.
30. R. Miao, B. Liu, Z. Zhu, Y. Liu, J. Li, X. Wang and Q. Li, J. Power Sources, 2008, 184, 420–426.
31. Q. Li, X. Chen, Q. Tang, H. Cai, Y. Qin, B. He, M. Li, S. Jin and Z. Liu, J. Power Sources, 2014, 248, 923–930.

---