Electrospinning of Nanofibers for Energy Applications

With global concerns about the shortage of fossil fuels and environmental issues, the development of efficient and clean energy storage devices has been drastically accelerated. Nanofibers are used widely for energy storage devices due to their high surface areas and porosities. Electrospinning is a versatile and efficient fabrication method for nanofibers. In this review, we mainly focus on the application of electrospun nanofibers on energy storage, such as lithium batteries, fuel cells, dye-sensitized solar cells and supercapacitors. The structure and properties of nanofibers are also summarized systematically. The special morphology of nanofibers prepared by electrospinning is significant to the functional materials for energy storage.

In this review, we focus specifically on the application of electrospun NF materials in lithium-based batteries, fuel cells, dye-sensitized solar cells and supercapacitors.

Electrospinning
The electrospinning technique is a versatile, efficient, and low-cost method for NFs' fabrication, which traces its roots back to an initial patent on the fabrication of textile yarns by Formhals in 1934 [28]. However, electrospinning did not become a popular technique. This technique was not revived until

Li-Ion Batteries
At present, for the commercial LIBs, the electrode materials usually are based on powder materials. The long migration pathways for the Li + of powder materials may lead to the large volume expansion occurring during cycling, resulting in the low rate of performance and poor cyclability [38]. NF materials are promising materials for LIBs because of their good electrochemical activity, high mechanical strength, and large specific surface area. In this section, recent advances in the areas of electrospun NFs as cathodes, anodes, and separator materials for LIBs are briefly summarized.

Cathode Materials
LiCoO2 is the first commercial cathode material for LIBs [40]. Jiao et al. [41] synthesized LiCoO2 NFs via the sol-gel and electrospinning technique. The LiCoO2 NFs delivered first cycle discharge capacities of 182 mAh/g (powder-based electrodes with first cycle discharge capacities of 140 mAh/g). However, LiCoO2 NF electrodes exhibited a poor cyclability. In order to solve this problem, the coreshell LiCoO2-MgO NFs were prepared through a coaxial electrospinning [42]. The first cycle discharge capacity of core-shell LiCoO2-MgO NFs was close to that of the bare LiCoO2 NFs. After 40 cycles, the discharge capacity of the core-shell LiCoO2-MgO NFs electrode still remained at 90.0% of the initial value (bare LiCoO2 NFs with a capacity retention of 52.0%). LiCoO2-MgO NFs can effectively avoid the increasing of the impedance result from passive surface film formation during cycling.
As a ternary layered material, Li-Ni-Co-Mn-O has attracted considerable interest as the cathode material for LIBs to replace the commonly-used material LiCoO2 due to its high capacity, good thermal stability and excellent rate capability [43]. Currently, most prepared Li-Ni-Co-Mn-O have the Li + /Ni 2+ cation mixing structure, resulting in depressing of the Li + mobility, which leads to the poor cyclability of the cells [44]. In order to overcome this obstacle, electrospun LiNi1/3Co1/3Mn1/3O2 was fabricated, which was proven to be able to largely improve the cell performance [45][46][47][48]. Xiong et al. [45] prepared LiNi0.5Co0.2Mn0.3O2 with a particle size of 300 nm by the electrospinning method and LiNi0.5Co0.2Mn0.3O2 with a particle size of 1 um by the sol-gel method. The electrospinning material exhibited better performance than the sol-gel material, which is mainly attributed to the shorter Li + and electron migration distance of the electrospinning material. LiNi1/3Co1/3Mn1/3O2 NFs were obtained by electrospinning poly(vinylpyrrolidone) (PVP) and MNO3 (M = Li, Ni, Co, Mn). The resulting LiNi1/3Co1/3Mn1/3O2 NFs showed a high first charge and discharge capacity of 217.93 mAh/g and 172.81 mAh/g, respectively [46]. Electrospun PVP and M(CH3COO)2 (M = Li, Ni, Co, Mn) fibers can yield a Li-rich Li1.2Mn0.54Ni0.13Co0.13O2-encapsulated carbon nanofiber (CNF) network (Li1.2Mn0.54Ni0.13Co0.13O2/CNF) that exhibited a high coulombic efficiency of 83.5% and a discharge capacity of 263.7 mAh/g [47]. The Li-rich Li1.2Ni0.17Co0.17Mn0.5O2 were also synthesized by the electrospinning technique, with PVP and polyvinyl acetate (PVAc) as different polymer sources [48]. The morphology of Li1.2Ni0.17Co0.17Mn0.5O2 NFs has depended on the different heat temperatures and

Li-Ion Batteries
At present, for the commercial LIBs, the electrode materials usually are based on powder materials. The long migration pathways for the Li + of powder materials may lead to the large volume expansion occurring during cycling, resulting in the low rate of performance and poor cyclability [38]. NF materials are promising materials for LIBs because of their good electrochemical activity, high mechanical strength, and large specific surface area. In this section, recent advances in the areas of electrospun NFs as cathodes, anodes, and separator materials for LIBs are briefly summarized.

Cathode Materials
LiCoO 2 is the first commercial cathode material for LIBs [40]. Jiao et al. [41] synthesized LiCoO 2 NFs via the sol-gel and electrospinning technique. The LiCoO 2 NFs delivered first cycle discharge capacities of 182 mAh/g (powder-based electrodes with first cycle discharge capacities of 140 mAh/g). However, LiCoO 2 NF electrodes exhibited a poor cyclability. In order to solve this problem, the core-shell LiCoO 2 -MgO NFs were prepared through a coaxial electrospinning [42]. The first cycle discharge capacity of core-shell LiCoO 2 -MgO NFs was close to that of the bare LiCoO 2 NFs. After 40 cycles, the discharge capacity of the core-shell LiCoO 2 -MgO NFs electrode still remained at 90.0% of the initial value (bare LiCoO 2 NFs with a capacity retention of 52.0%). LiCoO 2 -MgO NFs can effectively avoid the increasing of the impedance result from passive surface film formation during cycling.
As a ternary layered material, Li-Ni-Co-Mn-O has attracted considerable interest as the cathode material for LIBs to replace the commonly-used material LiCoO 2 due to its high capacity, good thermal stability and excellent rate capability [43]. Currently, most prepared Li-Ni-Co-Mn-O have the Li + /Ni 2+ cation mixing structure, resulting in depressing of the Li + mobility, which leads to the poor cyclability of the cells [44]. In order to overcome this obstacle, electrospun LiNi 1/3 Co 1/3 Mn 1/3 O 2 was fabricated, which was proven to be able to largely improve the cell performance [45][46][47][48]. Xiong et al. [45] prepared LiNi 0.5 Co 0.2 Mn 0.3 O 2 with a particle size of 300 nm by the electrospinning method and LiNi 0.5 Co 0.2 Mn 0.3 O 2 with a particle size of 1 um by the sol-gel method. The electrospinning material exhibited better performance than the sol-gel material, which is mainly attributed to the shorter Li + and electron migration distance of the electrospinning material. LiNi 1/3 Co 1/3 Mn 1/3 O 2 NFs were obtained by electrospinning poly(vinylpyrrolidone) (PVP) and MNO 3 [47]. The Li-rich Li 1.2 Ni 0.17 Co 0.17 Mn 0.5 O 2 were also synthesized by the electrospinning technique, with PVP and polyvinyl acetate (PVAc) as different polymer sources [48]. The morphology of Li 1.2 Ni 0.17 Co 0.17 Mn 0.5 O 2 NFs has depended on the different heat temperatures and the polymer used. The result shows that the porous nanostructure morphology can enhance the initial discharge capacity of the electrospun samples. LiFePO 4 was discovered by Goodenough in 1997, which possesses high discharge potential, large specific capacity (170 mAh/g) and good thermal stability [49]. However, LiFePO 4 has low conductivity (~10´9 S/cm) and low rate capability due to the one-dimensional twisted channel of Li + [50]. Particle nanocrystallization is supposed to be an effective method to shorten the diffusion channels. Thus, the LiFePO 4 /C composite NFs were prepared via electrospinning, with polyacrylonitrile (PAN) as polymer additives, which showed a good cyclability [51]. Interestingly, a slight increase of capacity was exhibited rather than the capacity fading after several cycles, which could be attributed to slow electrolyte penetration into the electrode or the crack formation on the amorphous carbon layer during cycling that increases the surface area of NFs and improves the electrode-electrolyte interaction. LiFePO 4 /carbon nanotube (CNT)/C composite NFs also were obtained via the combination of electrospinning and sol-gel methods, with PAN as the electrospinning media and carbon source [52]. The result shows that electrospinning is an effective method for minimizing the aggregation of LiFePO 4 particles and promoting the formation of a conducting carbonaceous layer on the LiFePO 4 surface. CNTs were well-dispersed in the carbonaceous matrix, which can help increase the electrochemical performance of the LiFePO 4 cathodes by forming conducting bridges between LiFePO 4 particles and enhancing the electron transport of the system. In another format, triaxial LiFePO 4 nanowires (NWs) with a multiwall CNT core column and an outer shell of amorphous carbon were successfully synthesized with the electrospinning method [53]. The model image of the material is demonstrated in Figure 3. The CNT core oriented in the direction of the wire played an important role in the conduction of electrons during the cycling; the outer amorphous carbon shell suppressed the Fe 2+ oxidation. the polymer used. The result shows that the porous nanostructure morphology can enhance the initial discharge capacity of the electrospun samples. LiFePO4 was discovered by Goodenough in 1997, which possesses high discharge potential, large specific capacity (170 mAh/g) and good thermal stability [49]. However, LiFePO4 has low conductivity (~10 −9 S/cm) and low rate capability due to the one-dimensional twisted channel of Li + [50]. Particle nanocrystallization is supposed to be an effective method to shorten the diffusion channels. Thus, the LiFePO4/C composite NFs were prepared via electrospinning, with polyacrylonitrile (PAN) as polymer additives, which showed a good cyclability [51]. Interestingly, a slight increase of capacity was exhibited rather than the capacity fading after several cycles, which could be attributed to slow electrolyte penetration into the electrode or the crack formation on the amorphous carbon layer during cycling that increases the surface area of NFs and improves the electrode-electrolyte interaction. LiFePO4/carbon nanotube (CNT)/C composite NFs also were obtained via the combination of electrospinning and sol-gel methods, with PAN as the electrospinning media and carbon source [52]. The result shows that electrospinning is an effective method for minimizing the aggregation of LiFePO4 particles and promoting the formation of a conducting carbonaceous layer on the LiFePO4 surface. CNTs were well-dispersed in the carbonaceous matrix, which can help increase the electrochemical performance of the LiFePO4 cathodes by forming conducting bridges between LiFePO4 particles and enhancing the electron transport of the system. In another format, triaxial LiFePO4 nanowires (NWs) with a multiwall CNT core column and an outer shell of amorphous carbon were successfully synthesized with the electrospinning method [53]. The model image of the material is demonstrated in Figure 3. The CNT core oriented in the direction of the wire played an important role in the conduction of electrons during the cycling; the outer amorphous carbon shell suppressed the Fe 2+ oxidation. At the current stage, for a commercial battery, the cathode materials usually participate in a oneelectron reaction, which inevitably limits the practical capacity [54]. Therefore, one promising approach to increase the delivered capacity is to develop a material that involves a two (or more)electron reaction. Vanadium pentoxide (V2O5) has attracted particular attention because of the availability of the two-electron V 5+ -V 3+ couple and its low cost, ease of synthesis, rich abundance, as well as its relatively high theoretical capacity (about 294 mAh/g with 2 Li + insertions/extractions per unit formula) [55]. V2O5 NFs were synthesized by a hydrothermal treatment of electrospun PMMA and vanadium oxide NFs [56]. The delivered discharge capacity for the first cycle in an LIB with V2O5 NFs was 350 mAh/g. In addition, the capacity was maintained above 240 mAh/g after 25 cycles. The hierarchical V2O5 NFs with diameters of 200-400 nm were obtained via electrospinning combined with annealing, with PVP as the electrospinning media [57]. After 100 cycles, the specific capacity of the V2O5 NFs retains 133.9 mAh/g at a current density of 800 mA/g, corresponding to a high capacity At the current stage, for a commercial battery, the cathode materials usually participate in a one-electron reaction, which inevitably limits the practical capacity [54]. Therefore, one promising approach to increase the delivered capacity is to develop a material that involves a two (or more)-electron reaction. Vanadium pentoxide (V 2 O 5 ) has attracted particular attention because of the availability of the two-electron V 5+ -V 3+ couple and its low cost, ease of synthesis, rich abundance, as well as its relatively high theoretical capacity (about 294 mAh/g with 2 Li + insertions/extractions per unit formula) [55]. V 2 O 5 NFs were synthesized by a hydrothermal treatment of electrospun PMMA and vanadium oxide NFs [56]. The delivered discharge capacity for the first cycle in an LIB with V 2 O 5 Nanomaterials 2016, 6, 129

of 30
NFs was 350 mAh/g. In addition, the capacity was maintained above 240 mAh/g after 25 cycles. The hierarchical V 2 O 5 NFs with diameters of 200-400 nm were obtained via electrospinning combined with annealing, with PVP as the electrospinning media [57]. After 100 cycles, the specific capacity of the V 2 O 5 NFs retains 133.9 mAh/g at a current density of 800 mA/g, corresponding to a high capacity retention of 96.05%. The ultra-long hierarchical V 2 O 5 NWs with diameters of 100-200 nm were also prepared using PVA and low-cost raw materials by electrospinning combined with annealing [58]. The initial discharge capacities of the ultra-long hierarchical V 2 O 5 NWs were up to 390 within a potential range of 1.75-4.0 V. In another study, electrospun PVP and a low-cost inorganic vanadium precursor can yield porous V 2 O 5 NTs that showed good cycling performance; the capacity retention is 97.4% after 200 cycles at 50 C [59].
LiMn 2 O 4 [60], Li 3 V 2 (PO 4 ) 3 [61,62] and LiNi 0.5 Mn 1.5 [63][64][65] NFs as cathode materials were also prepared by electrospinning, which exhibited high capacity and good cyclability. This indicates that the NFs with a shorter Li + and electron migration distance can provide fast Li + intercalation and de-intercalation channels. The overview of some electrospun fibers as cathode materials of LIBs and their electrochemical performances is presented in Table 1. Anode Materials Electrospun CNFs are popular anode materials for LIBs due to their low cost, easy availability and long cycle life [38]. However, LIBs with electrospun CNFs as the anode exhibited relatively low specific capacity and rate capability [66,67]. Therefore, flexible highly porous CNFs (HPCNFs) [68] and hollow CNFs (HCNFs) [69] were prepared by the electrospinning method. The HPCNFs delivered a reversible capacity as high as 1780 mAh/g after 40 cycles at current densities of 50 mA/g [68]. Controlling the mesopores in HCNFs is an effective means to improve the electrochemical performance of HCNFs [69]. HPCNFs and HCNFs with excellent electrochemical and mechanical properties result from the novel porous structure, which can provide the good access of the electrolyte to the electrode surface. Moreover, CNFs complemented with the high lithium storage capacity transition metals and transition metal oxides were used as anode materials for LIBs, which can result in a higher electrochemical performance; for example, Ag/CNFs composite by electrospinning PMMA/PAN and AgNO 3 solution in DMF, where Ag NPs were embedded into the hollow CNFs [70]. The LIBs with Ag/CNFs composite as the anode delivered a capacity of 739 mAh/g at 50 mA/g, with 85% capacity retention after 100 cycles. Similarly, GeO x NPs were encapsulated into hollow carbon shells using co-axial electrospinning [71]. This core-shell structure can alleviate the volume change of GeO x during cycling, decrease the contact area between electrolyte and GeO x to form a stable solid electrolyte interface film and increase the electrical conductivity. Cho et al. [72] prepared carbon-coated Fe 3 O 4 hollow NFs (Fe 3 O 4 /C hNFs) and Fe 3 O 4 hNFs via the electrospinning method, annealing and hydrothermal processing using PVP and Fe(NO 3 ) 3 . The Li + diffusion coefficient of the Fe 3 O 4 /C hNFs (8.0ˆ10´1 4 cm 2 /s) is 60-times higher than that of Fe 3 O 4 hNFs (1.33ˆ10´1 5 cm 2 /s). The Fe 3 O 4 /C hNFs exhibit superior electrochemical performance compared to that of the Fe 3 O 4 hNFs. CNF and transition metal oxide composites were also reported, such as Co 3 O 4 /CNF core-shell NWs [73], Mn 3 O 4 /CNF [74], GeO 2 /CNFs [75] and γ-Fe 3 O 4 /CNFs [76]. All of the composites are promising materials, which have important implications for developing high performance anodes for next-generation LIBs.
Among anode materials for LIBs, silicon (Si) is the most promising candidate because of its high theoretical lithium storage capacity (~4200 mAh/g) [77]. Nevertheless, the pure Si anode, with the format of a nanoparticle (NP), NW or membrane, has faced adverse issues on electrode pulverization and capacity loss during frequent charging/discharging cycles [78]. Carbon/silicon composite was proved to exhibit a synergistic effect [79]. For example, Si/CNFs with mesopores were prepared by electrospinning PVA, aluminum acetylacetonate (AACA) and Si in DMF solution [80]. The mesoporous structure was generated by foaming NFs via AACA sublimation. The LIBs with Si/CNFs as the anode showed a superior high first discharge capacity of 1639 mAh/g and a good capacity retention of 83.6%. Xing et al. [81] synthesized pyrolytic carbon-coated Si/C NF (Si/C-CNF) composites by electrospinning and carbonization, which exhibited a high retention capacity of 1215.2 mAh/g after 50 cycles. An amorphous carbon layer covering the surface of Si/CNFs (prepared by electrospinning and sintering PAN and Si) can limit the volume expansion of the exposed Si NPs and prevent Si/CNF;s direct contact with the electrolyte. It also can create connections between the fibers that are conducive to Li + ion transport. In other formats, a core-shell silicon/carbon fiber (Si/po-C@C) was prepared by simultaneously electrospinning polystyrene (PS)/PAN/Si as the core and electrospinning PAN as the shell then carbonization [82]. The structural design of the Si/po-C@C composite fiber is demonstrated in Figure 4. The LIBs with Si/po-C@C composite fiber as the anode exhibited a high initial discharge capacity of 997 mAh/g and a capacity retention of 71% after 150 cycles at a current density of 0.2 A/g. A novel flexible 3D Si/C fiber paper electrode was also fabricated via simultaneously electrospraying nano-Si/PAN clusters and electrospinning PAN fibers then carbonization [83]. The as-prepared Si/PAN paper and the carbonized Si/PAN paper (3D Si/C fiber paper) exhibited good flexibilities, as shown in Figure 5a,b, respectively. Figure 5c-f is the SEM micrographs of 3D Si/C fiber paper. When the novel structure of the 3D Si/C fiber paper was used as an anode for LIB, the LIB exhibited a high capacity (1600 mAh/g), an excellent rate capability performance and a low capacity loss (0.079% after 600 cycles).
via simultaneously electrospraying nano-Si/PAN clusters and electrospinning PAN fibers then carbonization [83]. The as-prepared Si/PAN paper and the carbonized Si/PAN paper (3D Si/C fiber paper) exhibited good flexibilities, as shown in Figures 5a,b, respectively. Figure 5c-f is the SEM micrographs of 3D Si/C fiber paper. When the novel structure of the 3D Si/C fiber paper was used as an anode for LIB, the LIB exhibited a high capacity (1600 mAh/g), an excellent rate capability performance and a low capacity loss (0.079% after 600 cycles).   As an anode material for LIBs, tin (Sn) has attracted considerable interest due to its high theoretical capacity (990 mAh/g) [84]. However, the volume changes during cycling and the tin NPs aggregate during the discharging process, resulting in the Sn-based materials' poor cyclability [85]. To address these problems, Sn NP-loaded CNF (Sn NP/CNF) [86] and Sn NP-loaded PCNF (Sn/PCNF) [87] composites were fabricated. The discharge capacity for LIB with Sn/PCNF composite was 774 mAh/g at a high current density of 0.8 A/g after 200 cycles [87]. Wang et al. [88] prepared Sn@carbon NPs encapsulated in bamboo-like hollow CNFs (see Scheme 1) by pyrolysis of coaxial tributyltin (TBT)/PAN electrospun NFs, which contained Sn and C with a weight ratio of 7:3. The LIBs with the composite material as the anode displayed a high discharge capacity of 737 mAh/g after 200 cycles at 0.5 C. The high cyclability and capacity retention ratio could be attributed to providing the appropriate void volume of the composite to respond to the large volume change and to prevent pulverization of the Sn NPs. The 3D Sn/C hybrid core/shell NFs with micro-/nano-channels were also prepared by electrospinning, annealing and glucose-hydrothermal processing, with PAN as the electrospinning media and carbon source [85]. The LIBs with the obtained Sn/C hybrid NFs as the anode displayed a high initial discharge capacity of 1090.8 mAh/g at a current density of 40 mA/g and a high coulombic efficiency of nearly 100%. The advantages of this particular 3D core/shell As an anode material for LIBs, tin (Sn) has attracted considerable interest due to its high theoretical capacity (990 mAh/g) [84]. However, the volume changes during cycling and the tin NPs aggregate during the discharging process, resulting in the Sn-based materials' poor cyclability [85]. To address these problems, Sn NP-loaded CNF (Sn NP/CNF) [86] and Sn NP-loaded PCNF (Sn/PCNF) [87] composites were fabricated. The discharge capacity for LIB with Sn/PCNF composite was 774 mAh/g at a high current density of 0.8 A/g after 200 cycles [87]. Wang et al. [88] prepared Sn@carbon NPs encapsulated in bamboo-like hollow CNFs (see Scheme 1) by pyrolysis of coaxial tributyltin (TBT)/PAN electrospun NFs, which contained Sn and C with a weight ratio of 7:3. The LIBs with the composite material as the anode displayed a high discharge capacity of 737 mAh/g after 200 cycles at 0.5 C. The high cyclability and capacity retention ratio could be attributed to providing the appropriate void volume of the composite to respond to the large volume change and to prevent pulverization of the Sn NPs. The 3D Sn/C hybrid core/shell NFs with micro-/nano-channels were also prepared by electrospinning, annealing and glucose-hydrothermal processing, with PAN as the electrospinning media and carbon source [85]. The LIBs with the obtained Sn/C hybrid NFs as the anode displayed a high initial discharge capacity of 1090.8 mAh/g at a current density of 40 mA/g and a high coulombic efficiency of nearly 100%. The advantages of this particular 3D core/shell architecture result from the greatly enhanced lithium storage from enlarging the electrode-electrolyte contact area, shortening the Li + migration pathways and enhancing the Li + and electrolyte diffusion of the active materials during cycling processes.
As an anode material for LIBs, tin (Sn) has attracted considerable interest due to its high theoretical capacity (990 mAh/g) [84]. However, the volume changes during cycling and the tin NPs aggregate during the discharging process, resulting in the Sn-based materials' poor cyclability [85]. To address these problems, Sn NP-loaded CNF (Sn NP/CNF) [86] and Sn NP-loaded PCNF (Sn/PCNF) [87] composites were fabricated. The discharge capacity for LIB with Sn/PCNF composite was 774 mAh/g at a high current density of 0.8 A/g after 200 cycles [87]. Wang et al. [88] prepared Sn@carbon NPs encapsulated in bamboo-like hollow CNFs (see Scheme 1) by pyrolysis of coaxial tributyltin (TBT)/PAN electrospun NFs, which contained Sn and C with a weight ratio of 7:3. The LIBs with the composite material as the anode displayed a high discharge capacity of 737 mAh/g after 200 cycles at 0.5 C. The high cyclability and capacity retention ratio could be attributed to providing the appropriate void volume of the composite to respond to the large volume change and to prevent pulverization of the Sn NPs. The 3D Sn/C hybrid core/shell NFs with micro-/nano-channels were also prepared by electrospinning, annealing and glucose-hydrothermal processing, with PAN as the electrospinning media and carbon source [85]. The LIBs with the obtained Sn/C hybrid NFs as the anode displayed a high initial discharge capacity of 1090.8 mAh/g at a current density of 40 mA/g and a high coulombic efficiency of nearly 100%. The advantages of this particular 3D core/shell architecture result from the greatly enhanced lithium storage from enlarging the electrode-electrolyte contact area, shortening the Li + migration pathways and enhancing the Li + and electrolyte diffusion of the active materials during cycling processes.

Scheme 1. Preparation of Sn@carbon nanoparticles encapsulated in hollow carbon nanofibers.
Reproduced with permission from [88]. Copyright Wiley, 2009. Scheme 1. Preparation of Sn@carbon nanoparticles encapsulated in hollow carbon nanofibers. Reproduced with permission from [88]. Copyright Wiley, 2009. As a promising candidate to replace graphite, metal oxide is believed to contribute to increasing energy density and decreasing cost [89]. TiO 2 is one of the most exciting anode candidates for LIBs due to its high abundance, nontoxicity and structural integrity during cycling [90]. Nevertheless, TiO 2 possesses low intrinsic electronic conductivity, which limits its application [91]. In order to improve the shortcoming, the TiO 2 NFs were synthesized by electrospinning [91][92][93][94][95]. Sigmund et al. [91] obtained the electrospun TiO 2 NFs with anatase and anatase-rutile mixed phase crystallites without carbon coating and then calcined the electrospun TiO 2 NFs under different atmospheres. The results show that argon-calcined TiO 2 NFs exhibited several orders of magnitude higher electronic conductivity. The anatase TiO 2 NF with a fiber-in-tube structure was also reported, with the formation mechanism as shown in Scheme 2, which exhibited high structural stability during cycling [92]. Ramakrishna et al. [96] studied electrospun TiO 2 nanostructures and their composites with multiwalled CNTs for rechargeable LIBs. The TiO 2 /CNT composite showed enhancement in the capacity retention (10-800 cycles) by increasing the retention from 81% to 92%. To improve the rate capability of TiO 2 NFs, the metal oxide/TiO 2 was prepared via electrospinning, such as Co 3 O 4 /TiO 2 [93], which exhibited high discharge capacity and an excellent rate capability, resulting from the synergetic effect between Co 3 O 4 and TiO 2 , as well as the unique feature of hierarchical heterostructures.
The Nb 2 O 5 was also used as the anode material for LIBs [97]. The polymorphs of Nb 2 O 5 , including pseudohexagonal (H-Nb 2 O 5 ), orthorhombic (O-Nb 2 O 5 ) and monoclinic (M-Nb 2 O 5 ), were synthesized by electrospinning [98]. After annealing, the H-Nb 2 O 5 and O-Nb 2 O 5 phases maintained the fibrous morphology, whereas the M-Nb 2 O 5 phase adopted a distorted nugget structure. The M-Nb 2 O 5 delivered the highest capacity and exhibited the highest capacity retention compared to the other two phases. The Li + diffusion coefficients of H-Nb 2 O 5 , O-Nb 2 O 5 and M-Nb 2 O 5 were also studied, which were in the ranges of 10´1 7 -10´1 6 , 10´1 5 -10´1 4 and 10´1 3 -10´1 2 cm 2 /s, respectively [99]. The overview of some electrospun fibers as anode materials of LIBs and their electrochemical performances is presented in Table 2.
the fibrous morphology, whereas the M-Nb2O5 phase adopted a distorted nugget structure. The M-Nb2O5 delivered the highest capacity and exhibited the highest capacity retention compared to the other two phases. The Li + diffusion coefficients of H-Nb2O5, O-Nb2O5 and M-Nb2O5 were also studied, which were in the ranges of 10 −17 -10 −16 , 10 −15 -10 −14 and 10 −13 -10 −12 cm 2 /s, respectively [99]. The overview of some electrospun fibers as anode materials of LIBs and their electrochemical performances is presented in Table 2.  Scheme 2. Formation mechanism of the TiO 2 nanofiber with a fiber-in-tube nanostructure from the tributyltin (TBT)-poly(vinylpyrrolidone) (PVP) composite nanofiber. Reproduced with permission from [92]. Copyright Wiley, 2015. Separator Materials The traditional commercial separators for LIBs are microporous membranes, whose low porosity, unsatisfactory thermal stability and poor wettability in liquid electrolytes inevitably limit the battery performance [100]. In this regard, fiber-based membranes and fiber-coated membranes have been developed [101]. Electrospun poly(vinylidene fluoride) (PVdF)-based membranes as separators for LIBs have been reported and show advantages in electrochemical performances [102,103]. Owing to the softening of PVdF fibers at high temperatures, the electrospun PVdF-based membranes can form an interconnected web structure by annealing. The interconnected web structure greatly improves the physical properties and electrochemical performance [102]. Porous PVdF fiber-based membranes possess good electrochemical properties, such as a high ionic conductivity of 1.0ˆ10´3 S/cm and a wide electrochemical stability window up to 4.5 V vs. Li + /Li at room temperature [103]. Furthermore, the electrospun poly(vinylidene fluoride-co-hexafluoropropylene) (PVdF-HFP) membranes possessed high electrolyte uptake and ionic conductivities (10´3 S/cm) [104]. However, PVdF and PVdF-HFP separators also have their disadvantages, such as the inferior thermal stability and intrinsically hydrophobic character. Wei et al. [105] prepared cellulose/PVdF-HFP membranes by coaxial electrospinning with superior thermal stability and a hydrophilic property. Other electrospun PVdF-based materials, such as P(VdF-HFP)/SiO 2 [106,107], PVdF-HFP/PMMA [108], P(VdF-co-HFP)/PAN [109] and PVdF/SiO 2 -Poly(acrylic acid) lithium (PAALi) [110] as separators for LIBs, all delivered high ionic conductivity, good tensile strength and high uptake/leakage. These results show that the physical properties and electrochemical performance of the membranes can be improved by adding SiO 2 , PMMA and PAN [106][107][108][109][110]. Electrospun PAN-based NF membranes are another common host separator for LIBs due to their high ionic conductivity, electrolyte uptake and electrochemical stability window [111,112]. Sui et al. [113] prepared PMMA/PAN NF membranes with a core-shell structure by electrospinning. The LIBs with the PMMA/PAN NF membranes as the separator exhibited excellent initial discharge capacities, as well as remarkable cycle performance, resulting from the unique electrospun core-shell structure of PMMA/PAN NF membranes. Electrospun SiO 2 /PAN [114,115], PAN/Al 2 O 3 -TEGDA-BA [116] (triethylene glycol diacetate-2-propenoic acid butyl ester (TEGDA-BA)), PAN/TEGDA-BA [117], PAN/PEO [118] (polyethylene oxide (PEO)), oxy-PAN [119] and LLTO/PAN (lithium lanthanum titanate oxide (LLTO)) [120] NF membranes were used as separators for LIBs, which showed remarkably good physical properties and electrochemical performance. The overview of electrospun fiber membranes as separators of LIBs mentioned above is presented in Table 3.

Li-S Batteries
Sulfur possesses the characteristics of high natural abundance, low cost and environmental friendliness. Equally, as a promising electrode material for lithium secondary batteries, sulfur has attracted attention because of its high theoretical specific capacity (1675 mAh/g) and high theoretical specific energy (2600 Wh/kg) [122]. Sulfur, working as the cathode of Li-S batteries, has an insulating property [123]. In recent years, though some progress has been made in LSBs, the dissolution of intermediates (Li 2 S x , 3 ď x ď 8) and the rapid capacity decay limit the future application of LSBs [124,125]. In order to overcome the above disadvantages, various kinds of carbon materials were used to composite with sulfur. Electrospinning is an effective method to obtain carbon materials with a special morphology for the storage of sulfur. The porous carbon NF (PCNF)/S nanocomposites were fabricated by carbonization of electrospun PAN/PMMA NFs and the subsequent chemical reaction-deposition strategy [126]. The as-prepared nanocomposites have a high electrical conductivity and extremely high surface area. The LSBs with the PCNF/S nanocomposites as cathodes showed high reversible capacity, good discharge capacity retention and an excellent rate capability because the sulfur was dispersed and immobilized on the porous structures of CNFs, alleviating the polysulfide shuttle phenomenon. A hierarchical structure carbon/sulfur composite is fabricated based on electrospun CNF matrices [123]. The CNF matrices were obtained by removing Ni NPs of Ni/CNF, which was synthesized by electrospinning PAN and Ni(AC) 2 H 2 O. The cell with the carbon/sulfur composite as the cathode exhibited an initial discharge capacity of 845 mAh/g at 0.25 C, with a retention of 77% after 100 cycles. Guo et al. [127] prepared the ordered mesoporous carbon fiber/sulfur (OMCF)/S composite via electrospinning. The OMCF/S composite with a weight ratio of 3:7 delivered a stable discharge capacity of 690 mAh/g after 300 cycles at 0.3 C. In situ synthesis is the most preferred preparation method of cathode materials for LSBs, particularly in synthesizing precursor mixture materials by electrospinning [128]. PAN NFs were fabricated in situ firstly with a porous structure by electrospinning, and then, the final product, PCNFs, was obtained by the following carbonization [129]. The prepared PCNFs were both mesoporous and microporous with large surface areas, which can store much sulfur and were utilized as conductive matrices for sulfur to form PCNFs/S nanocomposites. The LSB with the PCNFs/S nanocomposites as the cathode showed an initial discharge capacity of 1155 mAh/g at 0.02 C.

Lithium-O 2 Batteries
Recently, a rechargeable Li-O 2 battery has attracted significantly attention due to its high theoretical energy density (11,680 Wh/kg) [130]. The Li-O 2 battery involves a distinct operating mechanism (typically, 2Li + + O 2 + 2e´Ø Li 2 O 2 , E 0 = 2.96 V vs. Li/Li + ) on the O 2 electrode, which relates to the Li 2 O 2 formation during the discharging process (oxygen reduction reaction = ORR) and the Li 2 O 2 decomposition during the charging process (oxygen evolution reaction = OER) [131,132]. To date, the O 2 electrode is porous, which can provide the solid-liquid-gas tri-phase and electron migration regions for the ORR and OER. Therefore, the performances of the Li-O 2 battery depend on the material and architecture of the O 2 electrode [133]. Electrospinning is an attractive method for fabricating the materials of the O 2 electrode for Li-O 2 batteries [134].
In recent years, numerous electrospun electrocatalyst have been explored to enhance the performance of Li-O 2 batteries. For example, Kim et al. [135] prepared the 1D Co 3 O 4 NFs (obtained by electrospinning and calcination of PVP and Co precursor) immobilized on both sides of 2D non-oxidized graphene nanoflakes as a bifunctional composite catalyst for an O 2 electrode in Li-O 2 batteries. It showed the first discharge capacity of 10,500 mAh/g and a stable cyclability of 80 cycles with a limited capacity of 1000 mAh/g. 1D porous La 0.5 Sr 0. 5  The O 2 electrode architecture also influences the performance of Li-O 2 batteries. To avoid the negative influence of the binder on the long-term stability, a novel "binder-free" design has been proposed [137]. Typically, electrospun PAN and Co(NO 3 ) 2¨6 H 2 O can yield the Co 3 O 4 /CNF composites, that as the O 2 electrode exhibited a high initial discharge capacity of 760 mAh/g [137]. In another study, a Co 3 O 4 -based binder-free O 2 electrode was reported for Li-O 2 batteries [138]. Scheme 3 illustrates a typical procedure for the fabrication of the Co 3 O 4 -based binder-free O 2 electrode. The Co 3 O 4 nanoparticles formed nanoflakes by the aggregation in O 2 electrode, which can provide a large amount of catalytic active sites for ORR/OER. The Co 3 O 4 spheres embedded in the TiO 2 fiber mesh may function as anchors to prevent the detachment of the Co 3 O 4 layer from the current collector, resulting in excellent structural and cycling stability.
The O2 electrode architecture also influences the performance of Li-O2 batteries. To avoid the negative influence of the binder on the long-term stability, a novel "binder-free" design has been proposed [137]. Typically, electrospun PAN and Co(NO3)2·6H2O can yield the Co3O4/CNF composites, that as the O2 electrode exhibited a high initial discharge capacity of 760 mAh/g [137]. In another study, a Co3O4-based binder-free O2 electrode was reported for Li-O2 batteries [138]. Scheme 3 illustrates a typical procedure for the fabrication of the Co3O4-based binder-free O2 electrode. The Co3O4 nanoparticles formed nanoflakes by the aggregation in O2 electrode, which can provide a large amount of catalytic active sites for ORR/OER. The Co3O4 spheres embedded in the TiO2 fiber mesh may function as anchors to prevent the detachment of the Co3O4 layer from the current collector, resulting in excellent structural and cycling stability. Scheme 3. Schematic illustration of the typical preparation procedure for the Co3O4-based binder-free (CCTN) electrode. Reproduced with permission from [138]. Copyright Royal Society of Chemistry, 2015.

Fuel Cells
Scheme 3. Schematic illustration of the typical preparation procedure for the Co 3 O 4 -based binder-free (CCTN) electrode. Reproduced with permission from [138]. Copyright Royal Society of Chemistry, 2015.

Electrode Materials
The catalyst is the key material in fuel cells. It has been proven that the employment of a catalyst can largely influence the fuel cell performance [145]. Currently, electrospun NF materials have been used as catalyst and catalyst supporting materials for fuel cells because of their high poisoning resistance, high activity and good durability [146]. The PVP/Pt composite fibers were synthesized via electrospinning. After calcining, the Pt NWs with a high surface area were obtained, which displayed large electrochemically-active surface areas (ESA) of 6.2 m 2 /g and a high specific activity toward a methanol oxidation reaction (MOR) of 1.41 mA/cm 2 in DMFCs [140]. Pt/Rh NWs [145] and Pt/Ru NWs [147] were further synthesized, which exhibited higher catalytic activities in fuel cells than that of the traditional NP catalyst. It has been also reported that the bimetallic NWs electrocatalysts showed better stability than the bimetallic NPs [147]. Nanoporous Pt/Fe alloy NWs [148] and Pt/Co alloy NWs [149] were synthesized and studied. The Pt/Fe alloy NWs' overall wire diameters are about 10-20 nm, which showed a high specific activity (2.3-times that of conventional Pt/C catalyst) [148]. Pt/Co alloy NWs' average diameters are 28 nm, which showed excellent stability activity to the ORR in acidic electrolytes [149].
It has been proven that the catalytic activity of the catalyst depends not only on the shape, size and distribution of the catalyst materials, but also on the categories and properties of the catalyst support [150]. The catalyst particles are dispersed on various supports, including CNFs, SnO 2 and TiO 2 . Among them, CNFs are the most widely used catalyst supports. For example, Pt/CNF mats were prepared by electrospinning [146]. The catalytic peak current on the optimum Pt/CNF mat electrode reaches 420 mA/mg (the catalytic peak current of methanol oxidation on a commercial catalyst of 185 mA/mg). The results suggest that the electrospun CNF mats are favorable to enhance the performance of the catalyst. However, the cost and scarcity of Pt limit its application in fuel cells. Therefore, the low-cost N-doped CNFs were synthesized to replace Pt/CNFs [151][152][153]. Yu et al. [151] investigated the effect of different nitrogen sources (including melamine, aniline, urea and polyaniline) and nitrogen content on the activity of electrocatalyst, indicating that the best electrocatalytic activity was obtained with an aniline/PAN with a mass ratio of 20:3. You et al. [152] and Huang et al. [153] used NH 3 as the nitrogen source to prepare the N-doped CNFs by carbonizing electrospun polyacrylonitrile (PAN) NFs in NH 3 . Metal-/alloy-doped CNFs were also used as catalysts for fuel cells, such as Ag [154], Co/N [155,156], Cu/Co [150], Fe/Co [157,158], Ni/Cu alloy [159], Fe/N [160,161] and Ni/Co [162] -doped CNFs. For example, Smirnova et al. [154] fabricated Ag/CNFs with different Ag loading amounts of 11, 15 and 25 wt %. The well-dispersed Ag NPs were attached to the surface of the electrospun CNFs. The results show that the mass activity of the 15 wt % Ag/CNFs system was the highest (119 mA/mg).
The transition metal oxides possess high chemical stability, good corrosion resistance and strong metal support interaction, like SnO 2 and TiO 2 , which can enhance the activity of the catalysts. As catalyst supports, they are able to stabilize and disperse adequately a number of active phases [163]. Electrospun Pt/SnO 2 NFs were used as the anode electrocatalyst material for PEMFCs [164], which exhibited a high electrochemically-active surface area (81.17 m 2 /g-Pt). A diffusion-limited current was achieved at 0.07 V. Cavaliere et al. [165] prepared the Pt supported on the electrospun Nb-SnO 2 with a loose-tube (fiber-in-tube) morphology ( Figure 6), which exhibited a higher electrochemical stability than conventional Pt/C electrodes. Cavaliere et al. [166] also synthesized Pt supported on electrospun Nb-SnO 2 with a loose-tube structure as the cathode material for fuel cells, who studied in situ fuel cell operation under accelerated stress tests and confirmed that the voltage loss was negligible degradation. Electrospun Pt/TiO 2 NFs were also synthesized by electrospinning and the reductive impregnation method, which were used as electrocatalysts and exhibited high activities for methanol oxidation [167]. Electrospun Nb-TiO 2 NFs were investigated as electrocatalyst supports for PEMFCs [168]. The Pt/Nb-TiO 2 NFs showed a high stability after accelerated stress tests, retaining 73% of the electroactive Pt area (the value of Pt/CNFs was 8%).
The membrane electrode assembly (MEA) is one of the important parts for fuel cells, which consists of the membrane and electrocatalyst [169]. Electrospinning is a straightforward method to prepare NFs of MEA for fuel cells. Electrospun nafion/polyphenylsulfone (PPSU) mats were synthesized for the H 2 /Br 2 fuel cell [169], which exhibited good selectivity and high ionic conductivity. Shul et al. [170] fabricated electrospun SiC webs as membranes. The functionalized SiC webs (by introducing the hydroxyl group and phosphoric acid) showed a 70% better ion-exchange capability than that of the conventional membrane. The Pt NWs deposited on electrospun carbon film with a hierarchical structure were also used as MEA for fuel cells [171], which can induce the catalytic reaction toward formic acid electrooxidation. oxidation [167]. Electrospun Nb-TiO2 NFs were investigated as electrocatalyst supports for PEMFCs [168]. The Pt/Nb-TiO2 NFs showed a high stability after accelerated stress tests, retaining 73% of the electroactive Pt area (the value of Pt/CNFs was 8%).
(a) (b) The membrane electrode assembly (MEA) is one of the important parts for fuel cells, which consists of the membrane and electrocatalyst [169]. Electrospinning is a straightforward method to prepare NFs of MEA for fuel cells. Electrospun nafion/polyphenylsulfone (PPSU) mats were synthesized for the H2/Br2 fuel cell [169], which exhibited good selectivity and high ionic conductivity. Shul et al. [170] fabricated electrospun SiC webs as membranes. The functionalized SiC webs (by introducing the hydroxyl group and phosphoric acid) showed a 70% better ion-exchange capability than that of the conventional membrane. The Pt NWs deposited on electrospun carbon film with a hierarchical structure were also used as MEA for fuel cells [171], which can induce the catalytic reaction toward formic acid electrooxidation.

Electrolyte Membrane
The proton exchange membrane (PEM) is an electrolyte that behaves as an important indicator of the performance of fuel cells [172]. Nafion has been regarded as the most commonly-used polymer electrolyte membrane, which showed excellent chemical and physical stabilities and high proton conductivity [173]. However, the pure nafion is not a good candidate for electrospinning due to its low viscosity [174]. Aiming at the enhancement of the mechanical strength of the composite membrane, matrix polymers, such as PEO/Nafion [175] and PVA/Nafion [174], have been studied. To improve the proton conductivity, Na et al. [176] fabricated the sulfonated poly(ether ether ketone ketone) (SPEEKK) membrane with a spherical structure by electrospinning. The sulfonated SPEEKK membrane exhibited a high proton conductivity (0.37 S/cm), which is 37-times higher than that of the common SPEEKK membrane. The Nafion and PVdF composite membranes were fabricated by electrospinning [177]. They studied the Nafion NFs embedded in a PVdF matrix (N(fibers)/PVdF) and PVdF NFs embedded in a Nafion matrix (N/PVdF(fibers)). Both membrane structures exhibited similar in-plane conductivity, where the conductivity scaled linearly with the Nafion volume fraction. Hybrid organic-inorganic membranes were also prepared by electrospinning, which exhibited good thermal stability and a high proton conductivity of 15 ms/cm at 120 °C [178]. Shanmugam et al. [179] fabricated TiO2, CeO2 and ZrO1.95 NTs with mesoporous structures as fillers in the composite membranes, which were then incorporated in a Nafion ionomer to increase the water retention capability in the Nafion membrane.

Electrolyte Membrane
The proton exchange membrane (PEM) is an electrolyte that behaves as an important indicator of the performance of fuel cells [172]. Nafion has been regarded as the most commonly-used polymer electrolyte membrane, which showed excellent chemical and physical stabilities and high proton conductivity [173]. However, the pure nafion is not a good candidate for electrospinning due to its low viscosity [174]. Aiming at the enhancement of the mechanical strength of the composite membrane, matrix polymers, such as PEO/Nafion [175] and PVA/Nafion [174], have been studied. To improve the proton conductivity, Na et al. [176] fabricated the sulfonated poly(ether ether ketone ketone) (SPEEKK) membrane with a spherical structure by electrospinning. The sulfonated SPEEKK membrane exhibited a high proton conductivity (0.37 S/cm), which is 37-times higher than that of the common SPEEKK membrane. The Nafion and PVdF composite membranes were fabricated by electrospinning [177]. They studied the Nafion NFs embedded in a PVdF matrix (N(fibers)/PVdF) and PVdF NFs embedded in a Nafion matrix (N/PVdF(fibers)). Both membrane structures exhibited similar in-plane conductivity, where the conductivity scaled linearly with the Nafion volume fraction. Hybrid organic-inorganic membranes were also prepared by electrospinning, which exhibited good thermal stability and a high proton conductivity of 15 ms/cm at 120˝C [178]. Shanmugam et al. [179] fabricated TiO 2 , CeO 2 and ZrO 1.95 NTs with mesoporous structures as fillers in the composite membranes, which were then incorporated in a Nafion ionomer to increase the water retention capability in the Nafion membrane.

Dye-Sensitized Solar Cells
Dye-sensitized solar cells (DSSCs) as the third generation of solar cells have attracted much attention due to their high energy efficiency, clean energy source and low cost, which can be applied to many portable and mobile ubiquitous power sources [180]. DSSCs consists of a photoanode, counter electrodes, dye and electrolyte [181]. NFs have been considered to be the most important material for DSSC photoanodes, counter electrodes and electrolytes.

Photoanode
In the DSSCs, the photoanode is a transparent conducting oxide coated with a film of a high band gap metal oxide (TiO 2 , ZnO, SnO 2 and Nb 2 O 5 ) [182,183]. Electrospun metal oxide NFs as the film coated with photoanode have been studied due to their high specific area [184][185][186]. Leung et al. [187] fabricated an innovative bilayer TiO 2 NF photoanode by electrospinning. The bilayer TiO 2 NF photoanode can offer excellent dye-loading, light harvesting and electron-transport properties, which improved the power conversion efficiency. Figure 7 shows the schematic diagram of the fabrication procedure for the bilayer TiO 2 NF photoanode. In general, the photoanode with high specific surface area can improve light-scattering effects and facilitate electrolyte diffusion [188]. In this regards, Jang et al. [189] synthesized the multiscale porous TiO 2 NFs via electrospinning, which were used as photoanodes of DSSCs. In another study, CNT/TiO 2 NF [190] was fabricated by electrospinning. When 5 wt % of hollow CNTs was added into TiO 2 NFs, the energy conversion efficiency of DSSCs reached 3.39%. Ko et al. [191] prepared a novel nanostructured PVdF NFs/TiO 2 NP composite by electrospinning. The composite showed an outstanding bending stability because the PVdF NFs can reinforce the photoelectrode. Electrospun hollow TiO 2 NFs were also obtained to enhance the conversion efficiency of ZnO-based DSSCs [192]. The results show that the weight of TiO 2 influences the performance of the TiO 2 /ZnO composite photoanode. The maximal energy conversion efficiency reached 4.59% by adding 10 wt % of hollow TiO 2 NFs, which is 62% higher than that of pure ZnO NPs-based DSSCs.

Photoanode
In the DSSCs, the photoanode is a transparent conducting oxide coated with a film of a high band gap metal oxide (TiO2, ZnO, SnO2 and Nb2O5) [182,183]. Electrospun metal oxide NFs as the film coated with photoanode have been studied due to their high specific area [184][185][186]. Leung et al. [187] fabricated an innovative bilayer TiO2 NF photoanode by electrospinning. The bilayer TiO2 NF photoanode can offer excellent dye-loading, light harvesting and electron-transport properties, which improved the power conversion efficiency. Figure 7 shows the schematic diagram of the fabrication procedure for the bilayer TiO2 NF photoanode. In general, the photoanode with high specific surface area can improve light-scattering effects and facilitate electrolyte diffusion [188]. In this regards, Jang et al. [189] synthesized the multiscale porous TiO2 NFs via electrospinning, which were used as photoanodes of DSSCs. In another study, CNT/TiO2 NF [190] was fabricated by electrospinning. When 5 wt % of hollow CNTs was added into TiO2 NFs, the energy conversion efficiency of DSSCs reached 3.39%. Ko et al. [191] prepared a novel nanostructured PVdF NFs/TiO2 NP composite by electrospinning. The composite showed an outstanding bending stability because the PVdF NFs can reinforce the photoelectrode. Electrospun hollow TiO2 NFs were also obtained to enhance the conversion efficiency of ZnO-based DSSCs [192]. The results show that the weight of TiO2 influences the performance of the TiO2/ZnO composite photoanode. The maximal energy conversion efficiency reached 4.59% by adding 10 wt % of hollow TiO2 NFs, which is 62% higher than that of pure ZnO NPs-based DSSCs.

Counter Electrodes
In the DSSCs, the counter electrode is a transparent conducting oxide coated with a thin layer of Pt, which plays an important role on transferring the electron to the external circuit and regenerating the I3 − into I − by reduction in the electrolyte system [193]. However, Pt is an expensive metal, and the corrosive I − /I3 − redox couple can reduce its catalytic activity [194]. To address these issues, many

Counter Electrodes
In the DSSCs, the counter electrode is a transparent conducting oxide coated with a thin layer of Pt, which plays an important role on transferring the electron to the external circuit and regenerating the I3´into I´by reduction in the electrolyte system [193]. However, Pt is an expensive metal, and the corrosive I´/I3´redox couple can reduce its catalytic activity [194]. To address these issues, many studies have been carried out on the alternatives to replace the Pt. For example, Qiao et al. [195] fabricated CNFs by electrospinning using an electrocatalyst and low-cost alternative to Pt for I 3ŕ eduction in DSSCs. The short circuit current density and open circuit voltage of CNF-based cells were comparable to Pt-based cells, while the conversion efficiency was slightly lower than that of Pt-based cells. The hollow activated CNFs with a mesoporous structure were reported, which can promote the electron and ion transfer, decrease the resistance of charge transfer and increase the contact area between the liquid electrolyte and the electrode [196,197]. The resulting overall conversion efficiency of hollow activated CNF-based cells is comparable to that of Pt-based cells. To enhance the electrocatalytic activity, some of the teams prepared transition metals/CNFs [198][199][200][201][202] and transition metal carbides/CNFs [203,204]. The overview of electrospun NFs as a counter electrode material of DSSCs is summarized in Table 4.

Electrolytes of DSSCs
The electrolyte is an indispensable component in the DSSCs, which plays an important role in reduction of the oxidized dye by a suitable redox couple and the assistance of the charge transfer from the counter electrode to the photoanode [205]. In the DSSCs, the traditional organic liquid electrolytes and ionic liquid electrolytes exhibit a poor long-term stability because of their being prone to leakage and volatility [206]. Therefore, polymer electrolytes have been developed due to their high ionic conductivity, high thermal stability and good permeability [207]. Kim et al. [208] prepared the PVdF-HFP via the electrospinning method, which showed high electrolyte uptake and ionic conductivity of 653˘50% and 4.53˘1.3ˆ10´3 S/cm, respectively. Electrospun PVdF-HEP NFs with different diameters were further discussed [206]. The PVDF-HFP NFs prepared from a 15 wt % spinning solution showed high ionic conductivity (1.295 S/cm) and electrolyte uptake (947%). Dyson et al. [209] soaked electrospun PVdF-HFP membrane (es-PM) in the ionic liquid electrolyte containing 0.5 M 1-butyl-3-methylimidazolium iodide (BMImI), 0.5 M LiI, 0.05 M I 2 and 0.5 M 4-tert-butylpyribine to obtain the electrospun PVdF-HFP membrane electrolyte (esPME). The 0.5 M BMImI-esPME exhibited better performance than 0.5 M BMImI containing liquid electrolyte (LE), as shown in Figure 8. The poly(vinylidenefluoride-cohexafluoropropylene)/polystyrene (PVdF-HFP/PS) [210] and PVdF-HFP/cobalt sulfide (CoS) [211] were also prepared by the electrospinning method, which showed the good photovoltaic performance in DSSCs. SiO 2 NF/PEO-(PVdF-HFP) was also synthesized via electrospinning, which possesses high ionic conductivity of 9.9ˆ10´4 S/cm [212]. In addition, SiO 2 NF/PEO-(PVdF-HFP) as the polymer electrolyte can enhance the stability of the interface between the electrolyte and the semiconductor electrodes. The photovoltaic performances of electrospun NFs electrolytes are summarized in Table 5.

Supercapacitors
As a promising novel energy storage device, electrochemical capacitors (supercapacitors) have been used in various fields because of their high power density, fast charging and discharging times and long cycle life [215]. According to different energy storage mechanisms, supercapacitors are classified into pseudocapacitors and electric double-layer capacitors (EDLCs) [216]. The pseudocapacitors involve a faradic charge storage process that was facilitated by a redox reaction at the interface between electrode and electrolyte [217]. The EDLCs store electrical energy via the accumulation of electric charges at the electrical double layer formed at the electrode-electrolyte interface [218].

Supercapacitors
As a promising novel energy storage device, electrochemical capacitors (supercapacitors) have been used in various fields because of their high power density, fast charging and discharging times and long cycle life [215]. According to different energy storage mechanisms, supercapacitors are classified into pseudocapacitors and electric double-layer capacitors (EDLCs) [216]. The pseudocapacitors involve a faradic charge storage process that was facilitated by a redox reaction at the interface between electrode and electrolyte [217]. The EDLCs store electrical energy via the accumulation of electric charges at the electrical double layer formed at the electrode-electrolyte interface [218].
Besides, transition metal oxides are also promising pseudocapacitors electrode materials, due to their high electrical conductivity, several possible oxidation states and electrochemical stability. Kim et al. [238] fabricated electrospun RuO 2 NF mats to support the RuO 2 NPs. RuO 2 NF mats with RuO 2 NPs displayed high specific capacitance of 886.9 F/g at a scan rate of 10 mV/s and a low capacity loss of only 30% from 10 to 2000 mV/s. Electrospun V 2 O 5 NFs [134] and Mn 3 O 4 NFs [239] were also demonstrated to be promising electrode materials for pseudocapacitors. Wang et al. [240] fabricated the NiO NFs modified by citric acid (NiO/CA) with a hollow tube via electrospinning techniques. The pseudocapacitors with NiO/CA as electrodes displayed superior capacitive performance with large capacitance (336 F/g) and excellent cyclic performance (the capacitance decreases only by 13% of the initial capacitance after 1000 cycles). α-Fe 2 O 3 with porous fiber and α-Fe 2 O 3 with nanograin structures were prepared by the electrospinning technique [241], which were used as electrode materials in supercapacitors. The supercapacitors with both showed high capacitance retentions even after 3000 cycles. Electrospun 1D porous bimetallic ZnCo 2 O 4 NTs showed a much higher specific capacitance of 770 F/g at 10 A/g and good cycling stability of only a 10.5% loss after 3000 cycles [242]. Lin et al. [243][244][245] reported La x Sr 1´x Cu 0.1 Mn 0.9 O 3´δ (0.1 ď x ď 1) NFs, La x Sr 1´x CoO 3´δ (0.3 ď x ď 1) NFs and La x Sr 1´x NiO 3´δ (0.3 ď x ď 1) NFs, which exhibited excellent electrochemical performance. The supercapacitors with La x Sr 1´x CoO 3´δ (0.3 ď x ď 1) NFs as the electrode exhibit the highest specific capacity of 747.75 F/g at a current density of 2 A/g.

Conclusions
In this review, we summarized some progress in electrospun NFs for energy applications, such as lithium-based batteries (LIBs, LSBs and Li-O 2 batteries), fuel cells, dye-sensitized solar cells and supercapacitors. As the materials for electrodes and separators in lithium-based batteries, electrospun NFs exhibited high power capabilities and good kinetic properties because of the decreasing length of Li + diffusion pathways, which is mainly attributed to their high specific surface areas and high porosities. In fuel cells, electrospun NFs were used as the electrode materials and electrolyte membranes. Being electrocatalysts for electrode materials, electrospun NFs showed high activities and good durabilities. As electrolyte membranes, electrospun NFs exhibited high electrolyte uptakes and ionic conductivities. Equally, in terms of the application in dye-sensitized solar cells, the photoelectrodes prepared by electrospun NFs demonstrated high photoelectric conversion efficiencies due to their unique fiber morphologies. Supercapacitors with NFs exhibited high specific capacities and good electrochemical performances, which result from the high specific surface areas, good thermal stabilities and high electrical conductivities of NFs. As a promising synthesis technology for NFs with various morphologies, electrospinning has been largely explored. Nonetheless, the improvement of the accuracy of the experimental parameters and the development of NFs with multiple structures are currently being undertaken.