Electrospun Nanomaterials for Energy Applications: Recent Advances

Abstract

Electrospinning is a simple, versatile, cost-effective, and scalable technique for the growth of highly porous nanofibers. These nanostructures, featured by high aspect ratio, may exhibit a large variety of different sizes, morphologies, composition, and physicochemical properties. By proper post-spinning heat treatment(s), self-standing fibrous mats can also be produced. Large surface area and high porosity make electrospun nanomaterials (both fibers and three-dimensional fiber networks) particularly suitable to numerous energy-related applications. Relevant results and recent advances achieved by their use in rechargeable lithium- and sodium-ion batteries, redox flow batteries, metal-air batteries, supercapacitors, reactors for water desalination via capacitive deionization and for hydrogen production by water splitting, as well as nanogenerators for energy harvesting, and textiles for energy saving will be presented and the future prospects for the large-scale application of electrospun nanomaterials will be discussed.

1. Introduction

Electrospinning is one of the most widely adopted techniques for the production of high aspect-ratio nanostructures featured by high porosity, great surface area-to-volume ratio, and cross-sectional diameters ranging from tens to hundreds of nanometers. It is a very simple, scalable, and cost-effective technique, particularly suitable for the manufacturing of fibrous nanomaterials on an industrial scale since the production process, involving two well distinct steps (namely, fibrous film deposition and subsequent thermal treatment, if any), can progress without any interruption.

The wide variety of intriguing nanomaterials with outstanding physical and chemical properties that can be produced by electrospinning (natural and synthetic polymers, ceramic oxides, carbon and composite carbon-based fibers, and fibrous flexible membranes with hierarchical porosity) find application in a plethora of both well-assessed and emerging applications, spanning from healthcare and biomedical engineering (e.g., wound dressing, biological sensing, drug and therapeutic agent delivery, tissue regeneration, and engineering) to environmental protection and remediation (e.g., chemical and gas sensing, photocatalytic abatement of pollutants, water treatment), from energy harvesting, generation, and storage (e.g., solar cells, hydrogen production and storage, supercapacitors, batteries, and fuel cells) to electronics (e.g., field-effect transistors, diodes, photodetectors, and electrochromic devices) [1,2,3,4,5,6,7,8,9,10].

The huge number of potential applications in which electrospun nanomaterials (ENMs) can have significant impact hinders from exhaustively describing all of them in a single review. On the other hand, several excellent reviews illustrating the current status, challenges, and emerging development of nanofiber technology for specific application fields are already available in the literature (see f.i. [3,4,5,6,8,9,10,11,12,13,14,15,16,17,18], just to cite a few of them).

This paper deals with a specific theme—the energy-related applications of ENMs, which are increasingly gathering interest by the scientific community, as demonstrated by the impressive rate at which the number of published papers dealing with this subject is growing. Figure 1 displays the numbers relative to a single database. Among them, those concerning rechargeable alkali metal-ion batteries, supercapacitors, reactors for water desalination via capacitive deionization and for hydrogen production by photo-assisted water splitting (which presently represent the field on which the scientific interest of the author’s research group is mainly focused), as well as redox flow batteries, metal-air batteries, nanogenerators for energy harvesting, and textiles for energy saving through personal thermal management are selected and discussed, trying to report aspects, which can attract the interest and stimulate the inventiveness of the readers to produce further remarkable advances not only in the presented application fields, but also in others.

A brief description of the principle on which electrospinning is based and of the parameters to be tuned to control the physicochemical properties (size, morphology, texture, alignment, porosity, composition, etc.,) of the produced ENMs is first given. Aiming at facilitating the comprehension of the subject to readers experienced in different fields, each section begins with a very short introductive outline on the discussed application. Finally, the concluding section discusses the future prospects for the commercialization of ENM-related applications and their entrance in our daily life.

2. Principle of Electrospinning, Experimental Setup, and Production of Fibrous Nanomaterials

The history of the amazing technique that nowadays allows producing a wide variety of functional nanostructured materials endowed with unique physicochemical properties and outstanding characteristics for applications in the most diverse fields starts in 1902 [19,20] and proceeds through a sequence of patents [21,22,23,24,25,26,27,28] and related literature “milestones” [29,30,31,32,33]. Nonetheless, it only truly gained attention since the 1990–2000s thanks to the studies conducted to deepen the knowledge of the process [34,35,36,37,38,39,40,41,42,43,44]. For the interested readers, a comprehensive illustration of all the advancements in the electrospinning technique, from its invention in 1900 until nowadays, can be found in reference [5].

The electrospinning technology exploits electric forces to draw charged threads of a viscoelastic fluid (polymeric solution). The needed experimental setup is very simple and widely accessible. Basically, it consists of a syringe equipped with a needle and a pump to feed the solution, a collector, and a high-voltage power supply (Figure 2). Due to surface tension, a small amount of the fluid pumped out through the syringe needle tends to form a spherical droplet. If a high voltage is applied between the tip of the needle and the collector, the charged liquid droplet elongates assuming a conical shape (Taylor cone) due to electrostatic repulsion between the charges inside it. When the applied voltage is sufficiently high, the electric field force overcomes the surface tension and, if the molecular cohesion of the liquid is sufficiently high, a stable charged liquid jet is ejected from the apex of the cone. The jet continues to decrease in diameter until it starts to bend and enters an instability regime. As a result, it undergoes a continuous elongating-and-whipping process leading to the formation of a long and thin thread that finally deposits onto the collector. Since the solvent almost completely evaporates during the motion of the jet towards the collector, a non-woven mat of solid fibers is obtained by this process [5,12,14,15,45].

Nowadays, industrial spinners are commercially available and able to produce a continuous spun layer having an effective width of 1.6 m. Depending on the specifics required by the application for which the fibers are designed, spinners with different nozzle number and/or symmetry (single- or multi-nozzle, uni-, multi-, or co-axial nozzle spinners), as well as collectors with different shapes (plate-, disk-, or drum-collectors), can be utilized [15].

Multi-needle and needle-less systems allow for higher productivity with respect to the conventional electrospinners, whereas by the use of co-axial systems, in which precursors are fed through separate co-axial channels, hollow, or composite core-shell nanofibers with additional functionalities can be produced [4]. Mats consisting of uni-axially aligned rather than randomly-oriented fibers can be obtained by utilizing rotating drum-collectors in place of the conventional plate, or by the aid of suitably applied magnetic or electric fields [3]. Besides, through rational design and/or patterning of the collectors, nanofibers with very complex orientation patterns can be prepared [3]. Two separate polymer solutions can simultaneously and independently be spun on a same rolling collector thanks to the horizontally-opposed blending electrospinning method to achieve a bi-component fibrous mat.

Also, manifold are the ambient, instrument, and solution parameters that can be tuned in order to obtain the desired fiber properties (temperature, atmosphere, humidity, needle gauge, needle-tip/collector distance, solution feeding-rate, applied voltage, nature and molecular weight of the polymer(s), polymer mass-ratio in the case of blend, type and relative amount of the solvent, addition of precursors, preformed nano-sized components or sacrificial templating agents and their relative-to-polymer amount, use of conductive additives). At last, a single-step (calcination) or multi-step post-spinning heat treatment (stabilization and carbonization, sometimes followed by activation) upon different atmospheres (air, inert gases, gas mixture, activating gases) can be necessary to finalize the production process of the various fiber typologies (Figure 3). Polymer fibers do not require any treatment [46,47]; pure (and doped) oxide fibers are obtained by removing the organic component of the precursor(s) via an oxidative process [48,49,50,51], whereas stabilization of the polymer enabling its subsequent processing at higher temperature for graphitization is needed to obtain carbon and composite carbon-based fibers [52,53,54,55,56,57,58]. Activation upon gaseous atmosphere [53] or chemical etching to remove sacrificial templating agents [59] are utilized to enhance the fiber porosity and surface area. The produced ENMs can be in the form of powders (oxide fibers), self-standing non-woven membranes (polymer fibers), or self-standing flexible paper-like membranes (carbon and composite carbon-based fibers).

For the sake of completeness, it has to be mentioned that, in the latest years, several novel alternative strategies to fabricate nanofibers are emerging, such as plasma-induced synthesis, CO₂ laser supersonic drawing, electro-hydrodynamic direct writing, solution blow spinning, and centrifugal jet spinning. The readers interested in this subject can refer to reference [5].

In the following, an overview of some energy-related applications of ENMs is given. Recent advances in the field of rechargeable alkali metal-ion batteries, redox flow batteries, metal-air batteries, and supercapacitors are illustrated. Also, reactors for water desalination via capacitive deionization (based on energy storage) and for hydrogen production by water splitting, as well as nanogenerators for energy harvesting and textiles for energy saving through personal thermal management, exploiting EMNs, are briefly presented.

3. Rechargeable Lithium-Ion and Sodium-Ion Batteries

Nowadays, energy and environment represent key factors for the social development and long-term survival of human beings. The increasing energy demand associated with the growth of the world population and the enhancement of living standards, coupled to the need for decreasing greenhouse gas (GHG) emissions to limit effects on environmental pollution, global warming, and climate changes and to the progressive depletion of fossil-based resources, has strongly encouraged the utilization of clean and renewable energy sources. However, the stochastic and discontinuous nature of these sources calls for the use of efficient energy storage systems (EESs) to be settled in the electrical grid distribution [60]. The development of cost-effective, efficient, and large-scale EESs for the expected coming growth of renewable energy production is one of the key challenges of the 21st century.

Among the different EESs, those based on electrochemical methods, such as the rechargeable electric batteries, represent one of the most interesting alternatives because they are suitable to meet the different needs of power supply, while offering other intrinsic advantages, such as the compact size and the lack of emissions associated with their operation. Globally, electric batteries account for a total power of ~2.2 GW [61]. Among them, lithium-ion batteries (LIBs), which deliver 1.7 GW [61], presently represent the primary EES systems, with a production higher than 100 million cells/month and about 1500 tons/month of electrode materials. They are widely utilized to power small/portable electronic devices, as well as hybrid and fully electric vehicles; thanks to the recent advances, LIBs might also represent a viable solution for bulk EES at power stations and load leveling of renewable sources.

3.1. Lithium-Ion Batteries

As sketched in Figure 4, a LIB cell consists of anode (e.g., graphite), cathode (e.g., a Li-containing transition metal oxide, LiMO₂, with M standing for Mn, Fe, Co, or Ni), and a separator permeable to the ions and the electrolyte (e.g., LiPF₆ in an organic solvent). It is able to supply energy thanks to the spontaneous oxidation-reduction reactions occurring at the electrodes. During the charging process (de-lithiation), energy is supplied by an external source; Li⁺ ions are extracted from the cathodic material, diffused in the electrolyte, and enter the anodic material (according to the reaction C_x + LiMO₂ → Li_(1−y)MO₂ + C_xLi_y, in the case of a traditional LIB [62,63]), while electrons are transferred to the positive electrode through the external circuit. In the discharge process (lithiation), the opposite process takes place (i.e., Li_(1−y)MO₂ + C_xLi_y → C_x + LiMO₂, in the considered example): Li⁺ ions, extracted from the anodic material, are re-inserted into the lithiated cathodic material, and the cell provides energy.

The gravimetric capacities of the commonly used electrodes (372 mAh/g for graphite and 274–286 mAh/g for LiMO₂ [64]) are inadequate to meet the ever-increasing requirements for ESS devices. The search for alternative highly performing, durable, low-cost, and eco-friendly electrode materials is continuously progressing. As for the anodes, the focus is mainly on transition metal oxides MO_x [56,65,66,67,68], which store Li⁺ ions through the conversion reaction MO_x + 2x Li⁺ + 2x e⁻ ↔ M⁰ + x Li₂O, and IV group elements M⁰ [69,70,71,72,73] that are able to store larger amount of Li⁺ ions through the alloying reaction M⁰ + y Li⁺ +y e⁻ ↔ Li_yM [74], whereas transition metal oxides with olivine structure seem to be promising materials for cathodes [75]. However, some problems limiting the performance and life of the new electrode materials and thus hampering their commercialization still need to be solved. Various strategies have been proposed to enhance electronic conductivity and to reduce pulverization of the active material, to limit the use of inactive battery components (binder and current collectors), as well as to meet the additional requirements imposed by the development of flexible electronics. It is widely recognized that carbon incorporation and nano-sizing improve the electrochemical and mechanical performance of the electrode materials [72,76,77,78,79], providing them with enhanced electronic conductivity and increased resistance against the large volume changes, from which they suffer during the lithiation/de-lithiation process. It has been further shown that the use of binder-free self-supporting flexible carbon-based membranes as electrode materials [56,57,58,66,80,81,82,83,84] translates into a beneficial reduction of the inactive weights with a consequent gain in the effective gravimetric and volumetric capacities of the batteries. These benefits are of crucial importance in manifold applications including transportation, aerospace, and portable and wearable electronics.

Electrospinning is a viable route to simultaneously implement all these ameliorative strategies. Therefore, it has been increasingly and extensively utilized for the production of all the battery components—not only anodes [12,56,57,58,72,74,81,82,84,85,86,87,88,89,90,91] and cathodes [12,92,93,94,95], but also electrolytes [12,96,97,98] and separators [12,99,100], which makes it feasible to manufacture secondary batteries by assembling components produced exclusively by electrospinning (Figure 5).

Actually, the increasing number of published papers focused on this subject witnesses the growing interest of the scientific community for the use of this technique in the manufacturing of LIB components. Figure 6 displays the numbers relative to a single database. Several exhaustive review papers are available in the literature that report the outstanding electrochemical properties that suitably engineered ENMs exhibit compared to the commercial references (see f.i. [10,11,12,14,15,18,101], just to cite a few of them).

Indeed, the LIB-technology exploiting ENMs continuously progresses thanks to the incessant efforts made to enhance their performance and to render easier and more efficient their production process. For instance, Zhao et al. [80] have recently proposed to utilize polyimide (PI), in place of the commonly selected polyacrylonitrile (PAN), to fabricate flexible self-standing paper-like electrospun carbon membranes. According to Zhao et al., the oxidative stabilization, through which PAN is converted to an infusible stable ladder polymer [102], is a slow and not easily controllable step, whereas PI, containing aromatic heterocyclic rings with semi-ladder structure, can be easily obtained through a thermal or chemical imidization process from the polyamic acid (PAA) precursor [80]. Flexible self-standing paper-like electrospun nitrogen-doped carbon fibrous membranes produced via the multi-step process schematically illustrated in Figure 7a, using PI with biphenyl and pyrimidine rings (Figure 7b), are able to deliver a reversible specific capacity of 695 mAh/g at a rate of 0.1 A/g. This value is higher than that reported, at the same rate, by Dufficy et al. [78] for composite anodes consisting of PAN-derived carbon fibers enriched with thermally reduced graphene oxide. Flexible anodes derived from heterocyclic PI further exhibit good stability, retaining 245 mAh/g after 300 cycles at a rate of 1.5 A/g [80].

Owing to the very high theoretical specific capacity of silicon (4200 mAh/g), carbon fibers with embedded Si nanoparticles (NPs) have received great attention as a possible solution to the rapid capacity fading caused by the pulverization of the bare silicon electrode due to its mechanical failure upon repeated Li insertion and extraction. In particular, Xu et al. [82] have designed flexible non-woven membranes consisting of electrospun silicon/carbon@carbon fibers (Si/C@CNFs). The fibers, fabricated by the use of a co-axial electrospinning configuration (Figure 8a), exhibit a core-shell structure (Figure 8b–e), where the external shell is constituted by PAN-derived carbon, and the internal core consists of well-dispersed Si NPs wrapped to PVP-derived carbon. The Si/C core acts as an efficient conductive network (Figure 8f), enabling the membranes with 40 wt% Si loading to deliver up to 1507 and 555 mAh/g at current densities of 0.1 and 2 A/g, respectively [82]. Thanks to the buffering action of the PVP-derived carbon protecting Si NPs from the mechanical damage, after 500 cycles at a rate of 0.2 A/g, the electrode retains 86% of the initial capacity.

Although flexible self-supporting paper-like fibrous membranes synthesized via electrospinning need neither binders nor metal substrates as current collectors, conversely, their open and loose structure, affording Li⁺ ion diffusion and strain accommodation, results in low-density and, hence, in a low volumetric specific capacity of the electrode. Recently, Pantò et al. [57] have demonstrated that the hollow space within the fibers can be noticeably reduced by cold pressing the as-membrane. This extremely simple procedure that allows for obtaining rigid self-standing membranes having a more compact and tight structure (compare Figure 9a,b). In the case of nitrogen-doped carbon fibers encapsulating GeO_x NPs (bottom of Figure 9), the anodes are able to deliver initial discharge volumetric capacities up to 3580 mAh/cm³ [58].

Although some unsolved problems persist and further ameliorative actions are still needed [58], the efforts spent by the Scientific Community in this research field ensure the full exploitation of the huge ENM potential in the near future.

Nonetheless, different aspects also deserve consideration. The advancement of LIB technology, alone, is not sufficient for a LIB-based economy. It is of fundamental importance that the future systems for EES, besides exhibiting outstanding electrochemical performance to meet the ever-increasing market requirements, also be sustainable. Since their first commercialization by Sony in 1991, the pricing of LIBs has continuously been dropping due to the expansion of worldwide production capacity, increased by 11% between 2015 and 2016 [103], and oversupply of the produced power (only 40% of available 53 GWh produced in 2015 was used [104]). However, due to the LIB penetration in automotive and on-grid storage markets and the consequent increasing consumption of lithium, this trend is destined to reverse. Moreover, at the average 5% yearly growth rate in lithium mining necessary to pace up the demand increase, reserves are expected to encounter a severe shortage in less than 65 years [105], with consequent serious concerns for the future production costs of lithium raw material. This is the reason why, in the latest years, a growing interest has been focused on secondary batteries based on different alkali metal ions (i.e., sodium [105,106,107,108,109,110] and, more recently, even potassium [109,111,112,113,114,115]) with similar chemistry to lithium.

3.2. Sodium-Ion Batteries

The use of sodium would remarkably reduce storage costs since its resources are more abundant and better distributed on Earth. A study based on a prospective life-cycle-assessment has demonstrated that sodium-ion batteries (SIBs), showing, per kWh of storage capacity, impacts at the lower end of the range reported for current LIBs [116], represent a more sustainable, cost-effective, and eco-friendly alternative to LIBs in the energy- and climate-related challenges of the 21st century. Of course, many efforts are still required to meet the increasing demands of the market on zero-emission electric vehicles and on-grid energy storage and to fill the gap between LIB and SIB technology. The development of SIB technology requires the identification/design and preparation of new electrode materials with reversible Na⁺ intercalation/conversion/alligation reaction to reduce the technological gap with LIBs.

In principle, in spite of the higher weight of sodium compared to lithium (23 g/mol against 6.9 g/mol), its higher standard electrode potential (−2.71 V vs. SHE against −3.02 V vs. SHE for lithium) and the larger size of its ions (1.02 Å against 0.76 Å), the transition from LIBs to SIBs should not cause any consequence in terms of energy density, since the capacity primarily depends on the characteristics of the nanostructures acting as electrodes [105].

Various materials have been evaluated as cathode materials for SIBs obtaining interesting results, in terms of stability and charge retention [105,117,118,119]. They include layered sodiated transition metal oxides (Na_1−xMO₂, where M stands for Ti, V, Cr, Mn, Fe, Co and Ni, Ru) with O3- and P2-type structure (for x close to 0 or in the range 0.3–0.7, respectively) and their derivatives, layered transition metal oxides and fluorides, transition metal polyanion materials, phosphates, fluorophosphates, pyrophosphates, mixed phosphates, carbonophosphates, sulfates, fluorosulfates, cyanides, and organic compounds, with theoretical (measured) specific capacities ranging between 79 and 710 mAh/g (30 and 300 mAh/g) [105].

Table S1 reports some examples of ENMs for SIB cathodes and their preparation conditions; the main parameters describing their performance are reported in Table S2. Polyvinyl pyrrolidone (PVP) [120,121,122,123] and polyvinylacetate (PVAc) [124,125] are frequently used as polymers for the preparation of the spinnable solution. The fibers of sodiated transition metal oxides, obtained after thermal removal of the organic component of the as-spun fibers, generally consist of interconnected building blocks, whose shape may evolve with changing the heating temperature (Figure 10) [120,122,124].

As in the case of LIBs, the preparation of the working electrode by utilizing these kinds of materials (in the form of powders) requires the use of binder and carbonaceous additive (Table S2). Besides, a current collector is necessary to complete the SIB assembly. On the contrary, paper-like flexible membranes, consisting of carbon fibers encapsulating very small-sized oxide particles (Figure 11), can be directly used as self-supporting working electrodes—they do not need any binder, conductive additive, and collector [123,125,126,127], which improves the overall energy density of batteries. These cathode materials generally offer a variety of advantages and exhibit relevant electrochemical performance for Na-storage. For instance, the self-standing membranes formed by a non-woven network of composite NaVPO₄F/C fibers synthesized by Jin et al. [123] are able to deliver a reversible capacity of 103 mAh/g at a rate of 1 C, and to retain 56 mAh/g even at a very high rate (50 C), further exhibiting a ultra-long cyclability with 96.5% capacity retention after 1000 cycles at 2 C. The shortened length of Na-ion transport associated to the very small size of the oxide particles, together with the fiber porosity, favoring the permeation of the electrolyte, and the material structure consisting of a three-dimensional network of interconnected fibers that results in improved ionic and electronic transport, is responsible for this behavior [123].

Different from cathode materials, for which satisfactory results in terms of stability and charge retention have been achieved [105], anode materials still represent a challenge [128]. Graphite, the standard anode material for LIBs, exhibits poor performance in SIBs since Na⁺ ions hardly intercalate between its layers because of the energetic instability of the Na-graphite intercalation compounds [105,129]. Non-graphitic disordered anode materials, such as hard carbons, where the enlarged interlayer spacing and/or the smaller size of the parallel graphitic domains allows for the sodium insertion, are currently utilized. Recently, self-standing membranes consisting of fulvic acid-based electrospun hard-carbon nanofibers have been evaluated as binder- and collector-free anode materials [130]. They have shown a reversible sodium intercalation capacity of 248 mAh/g with a capacity retention ratio of 91% after 100 cycles at a current density of 100 mA/g. The proper interlayer distance, coupled with the abundance of oxygenated functional groups and benefits deriving from the high aspect-ratio morphology of the electrospun fibers (i.e., short ion diffusion distance and excellent stress tolerance) account for their good storage performance.

However, since the production of hard carbons generally has great environmental impact [116], different, more eco-friendly anode materials have been recently investigated for rechargeable SIBs [117,128,129,131].

Conversion materials, such as transition metal oxides (TMOs) and transition metal sulfide (TMSs), conversion-alloying materials, such as transition metal phosphide (TMPs), and alloying materials, such as elements of the 14th and 15th groups in the periodic table, have received particular attention [105,129,131]. The highest theoretical capacity (~2600 mAh/g) pertains to phosphorus. Nevertheless, the large volume changes (up to ~400%) undergone by the alloying materials during the sodiation/de-sodiation process, with consequent electrode pulverization and rapid capacity fading, is still an unresolved problem, which hampers their use in practical applications [105,131]. Conversely, TMSs and TMOs exhibit still high theoretical capacities (up to ~900 and ~1000 mAh/g, respectively) and the main limitations of their use as anode materials, namely structural damage leading to pulverization and poor electronic conductivity, can be effectively tackled with nano-sizing [132,133], confinement in a carbonaceous matrix [131,134,135,136,137,138,139], and doping with aliovalent elements [140].

Tables S3 and S4 respectively report the preparation conditions of some ENMs for SIB anodes and the main parameters describing their performance. As in the case of LIB anodes, PAN [130,133,135,136,140,141] and polyvinyl PVP [134,137,138,139] are the polymers most frequently selected to prepare the spinnable solution, whereas polymethyl methacrylate (PMMA) generally acts as a sacrificial pore-forming agent [136,141].

As for anode materials consisting of electrospun TMO-fibers, the high aspect ratio fiber morphology is responsible for the interestingly high initial specific capacity (983 mAh/g) recently reported for spinel cobalt oxide (Co₃O₄) fibers [133]. However, after the first sodiation/de-sodiation cycle, their capacity lowers down to 576 mAh/g as an effect of the formation of CoO as a stable intermediate. The CoO-based anode retains 71% of its initial capacity over 30 cycles [133].

Anodes based on electrospun iron(III) oxide fibers doped with silicon exhibit better stability [140]. They are able to deliver 350 mAh/g after 70 cycles retaining 85% of the initial capacity—a result never achieved for iron oxide-based electrode standard formulation. The doping with an aliovalent element induces changes in the crystalline phase of the oxide and in its morphology, as well. The undoped fibers consist of interconnected polycrystalline hematite (α-Fe₂O₃) grains (Figure 12a), whereas the Si-doped fibers are constituted by elongated maghemite (γ-Fe₂O₃) nanostructures developing mainly along the longitudinal fiber axis (Figure 12b).

The latter change is responsible for the improved cyclability, while the formation of γ-Fe₂O₃, an intrinsically more conductive phase, accounts for the enhanced electrochemical performance with a specific capacity (400 mAh/g at a rate of C/20) higher by a factor of ~4 compared to the α-Fe₂O₃-based electrode [140]. Electrodes based on γ-Fe₂O₃ NPs with standard formulation (Table S4) exhibit worse performance too, with lower stability and smaller delivered capacity (~150 mAh/g after 50 cycles) [136]. Instead, in free-standing γ-Fe₂O₃/C fibrous membranes, prepared by using PMMA as a sacrificial agent to generate pores in the nitrogen-doped PAN-derived carbon fibers (Figure 13), the electrochemical behavior benefits from γ-Fe₂O₃ NPs being embedded in high aspect-ratio carbon-fibers [136], often reported as one-dimensional (1D) nanostructures. After 50 cycles at 0.1 C, the binder-free electrode still delivers 290 mAh/g.

By far, greater stability pertains to anodes based on electrospun carbon fibers containing anatase NPs (Table S4). TiO₂/C fibers are able to deliver 240 mAh/g at 0.2 A/g with capacity retention close to 100% over 1000 cycles [134]. Over the same number of cycles, thanks to the synergistic interaction between anatase and tin NPs (Figure 14), TiO₂-Sn/C fibers deliver a lower reversible capacity (191 mAh/g) at a higher current density (1 A/g) with comparable capacity retention (95.4%) [137].

Very recently, Xia et al. [139] have proposed a four-step process to prepare composite tin disulfide/N- and S-doped carbon nanofibers. As schematically depicted in Figure 15a, the process consists of (i) electrospinning, followed by (ii) stabilization in the air, (iii) sulfidation, and (iv) carbonization in a vacuum. It allows for the production of dual-doped carbon fibers encapsulating crystalline SnS₂ NPs and decorated by crystalline SnS₂ nanosheets (Figure 15b–f). In Na⁺ ion half-cell, the electrode prepared by the use of these fibers delivers reversible capacities of 380 mAh/g at 0.5 A/g after 200 cycles and of 310 mAh/g at a remarkably higher current density (4 A/g). Moreover, at a low rate (50 mA/g), the Na⁺ ion full-cell prepared by coupling pre-sodiated SnS₂/N,S-C fibers with Na_2.3Cu_1.1Mn₂O_7-δ-based cathode (with 100 mAh/g specific capacity) is able to power 24 red LEDs (light emitting diodes) operating at 2 V, since it delivers an initial discharge capacity of 350 mAh/g, and after 50 cycles it exhibits a capacity retention of 61% (corresponding to a delivered capacity of 213 mAh/g) [139]. The good electrochemical performance of the SnS₂/N,S-C fibers originates from the synergy between inherent advantages of their high aspect-ratio morphology (short diffusion length and ability to buffer volume changes) and doping with heteroatoms, which further enhances the electrical conductivity of the carbonaceous matrix.

Liu et al. have reported better results, in terms of cycling life, for binder- and collector-free membranes consisting of electrospun porous N-doped carbon fibers encapsulating Sn [141] and MnFeO₄ [135] nanodots with typical sizes of 1–2 nm and ~3.3 nm, respectively. The former electrodes deliver reversible capacities of 633 and 450 mAh/g at rates of 0.2 and 10 A/g, respectively; besides, they retain 483 mAh/g over 1300 cycles at 2 A/g [141]. Conversely, the latter exhibit slightly lower capacities (504 and 305 mAh/g at rates of 0.1 and 10 A/g, respectively), but impressively longer cyclability, delivering 360 mAh/g (corresponding to capacity retention of 90%) after 4200 cycles [135].

4. Secondary Redox Flow Batteries

As mentioned above, the predicted severe shortage of the lithium reserves in less than 65 years [105] will result in serious concerns for the future production costs of LIBs, whereas the SIB technology has not reached the needed maturity yet. In this scenario, the secondary redox flow batteries (RFBs) gather great interest among the traditional large-scale EESs. They can convert chemical energy into electricity by the oxidation-reduction reactions of positive and negative reactants [142]. Thanks to their excellent performance in terms of efficiency, life cycle, environmental friendliness, and low toxicity, particular attention is devoted to vanadium-RFBs, considered prospective future EESs since vanadium is more abundant in nature than lithium and its cost is much lower. In a first-generation vanadium-RFB [143], positive (VO²⁺/VO₂⁺ sulfate) and negative (V²⁺/V³⁺ sulfate) electrolytes are stocked in tanks, and a proton-exchange membrane separates the electrodes in the reaction cell (Figure 16). The energy conversion in the cell occurs through the chemical reaction VO²⁺+ H₂O + V³⁺ ↔ VO₂⁺ + 2H⁺ + V²⁺ (E⁰ = 1.26 V), with VO²⁺+ H₂O ↔ VO₂⁺ + 2H⁺ + e⁻ (E⁰ = 1 V) and V³⁺ + e⁻ ↔ V²⁺ (E⁰ = −0.26 V) reactions taking place at the positive and negative electrode, respectively [142,144,145].

Since the mid-1980s, when the research on vanadium-RFBs began [143], new generations of systems have been proposed to improve the stability and the solubility of vanadium ions over a wider range of temperature [142]. Second-generation vanadium-RFBs utilize a mixture of vanadium chloride and vanadium bromide as the electrolyte, whereas third-generation systems operate with mixed H₂SO₄/HCl electrolyte solutions in both half-cells [142].

Advances and prospects in the field of RFBs have exhaustively been illustrated in the literature (see f.i. [142,146]). The strength point of such systems is that the battery capacity can be increased by scaling up the tank volume to contain larger amounts of electrolytes, and the power output can be increased by adding more reaction cells. However, the performance of a battery ultimately depends on its components.

Various types of membranes are used in the RFBs. Among them, commercial (cation exchange) Nafion membranes are widely utilized. They exhibit high chemical stability and excellent electrochemical properties, but, conversely, suffer from high permeability and cost [142]. The poor electrochemical activity of graphite felt (GF), commonly employed as an electrode material, limits the practical application of vanadium-RFBs [145,147]. Although research is still addressed toward carbon-based electrodes due to their low cost, new strategies have been thought up to enhance the electrochemical activity of electrode materials for vanadium-RFBs. Increasing their effective surface area through nanostructuring and accelerating the electrochemical kinetics for redox reactions thanks to the development of novel electrocatalysts represent the usually followed approaches [144]. In all the cases, the use of ENMs appears to be a viable route to improve the RFB performance [148,149,150]. Just to cite some examples, better performance of non-aqueous vanadium acetylacetonate-based RFBs, in terms of coulombic efficiency (CE), energy efficiency (EE), and voltage efficiency (VE), have been obtained at all of the tested current densities thanks to selective membranes fabricated by depositing Nafion/polyvinyl alcohol (PVA) fibers directly on a rotating drum collector wrapped with a porous Celgard 2400 membrane [151]. Compared to the commercial dispersion-cast Nafion NR 212 membrane, the fibrous Nafion/PVA membrane exhibits lower swelling ratio and lower vanadium acetylacetonate permeability.

As for the electrode materials, several interesting results have been achieved by the use of electrospun carbon-based fibers as the negative electrode in RFBs [145,152,153]. Generally, the electrode performance of woven electrospun fibrous membranes is limited by low porosity and poor permeability that hampers the mass-transport during the RFB operation. To tackle this problem in vanadium-RFBs, Xu et al. [152] have designed a freestanding carbon electrode consisting of an electrospun fiber web with ultra large pores. The electrode has been fabricated via the horizontally-opposed blending electrospinning method, involving the simultaneous and independent spinning of two separate polymer solutions (PAN/DMF and PVP/deionized water) on a same rolling collector, followed by PVP fiber removal through a selective dissolution approach and standard two-step (stabilization/carbonization) heat treatment. At 60 mA/cm² current density, the CE, VE, and EE of the single cell using the so-obtained freestanding electrode reach 95.4%, 81.6%, and 77.9%, respectively. These values are higher than the ones achieved by the use of non-woven fibrous webs.

To improve the mass-transport of active electrode species in vanadium-RFBs, Sun et al. [153] have proposed a novel method to fabricate carbon-based electrospun fibrous electrodes endowed with internal channels in each fiber. By a proper tuning of the viscosity of the polyacrylonitrile/polystyrene (PAN/PS) precursor solution leading to self-assembling of the carbon fibers into large bundles, they have prepared porous carbon fiber bundle mats (Figure 17).

The produced bundle structure results in significantly enlarged pore size compared with the traditionally electrospun fiber electrode, whereas the nanochannels formed inside the fibers due to the PS thermal decomposition upon carbonization provide additional specific surface areas. As a result, at current densities of 100 and 200 mA/cm², the single cell with the as-prepared bundled electrodes achieves nearly doubled discharge capacity and electrolyte utilization, as well as EE values of ~88% and 80%, respectively.

As is widely recognized, nitrogen doping of the active material is beneficial for the wettability and electrocatalytic activity of electrodes [6]. He et al. [145] have synthesized PAN-based nitrogen-doped carbon fibers via electrospinning followed by fiber dispersion in the liquid phase (urea solution, with urea acting as an external nitrogen source) and carbonization at different temperatures (800–1000 °C). Due to the strong electronegativity of nitrogen, doping increases the negative charge density on the carbon fiber surface, facilitating the adsorption of V²⁺ and V³⁺ ions and providing active sites for V²⁺/V³⁺ couples. Therefore, at 90 mA/cm², the cell using fibers carbonized at 900°C exhibits a discharge capacity of 98.0 mAh and an EE close to 73%.

In recent years, various noble metal species have been evaluated as alternative materials to the commonly used carbonaceous ones. However, in spite of their good electrocatalytic activity, their cost severely limits any practical applications [154]. Metal oxide-carbon composites, combining electrical conductivity peculiar to carbonaceous materials with the enhanced catalytic properties of the metal oxides, are very promising candidates for electrodes in RFBs. Interesting results have recently been presented by Antonucci’s research group [154,155,156]. At 80 mA/cm², this group has reported high EE values (81%) for the cell assembled with electrodes based on flexible PAN-derived electrospun carbon nanofibers (ECNFs) incorporating spinel Mn₃O₄ NPs [156], which exhibits better performance with respect to the cells exploiting electrodes based on bare ECNFs as well as commercial carbon felt (73%). The improvement is ascribed to the increase in electron conductivity originating from the synergistic action between the two electrode material components, while the high utilization degree of the active species in the Mn₃O₄/ECNF-based electrode is attributed to the high reversibility of the Mn₃O₄ component.

Also, electrodes based on mesoporous nickel/ECNFs exhibit higher catalytic activity toward both V²⁺/V³⁺ and VO²⁺/VO₂⁺ reactions with respect to the electrode based on bare ECNFs [154]. At a current density of 160 mA/cm², an EE value of 68% is obtained for the single cell thanks to the contribution of the Ni NPs in favoring electron transfer and vanadium ion accessibility.

5. Metal-Air Batteries

A single review cannot cover in detail all kinds of EESs utilizing ENMs. In the following, metal-air batteries (MABs) are briefly mentioned, even if, for sure, they are not the unique EES that would deserve to be presented, besides LIBs, SIBs, and RFBs.

Thanks to the affordable cost, extremely high theoretical energy density, environmental friendliness, and safety, MABs are considered as promising next-generation EESs for portable, mobile, or stationary applications [157]. They employ a negative electrode composed of a light metal such as lithium, sodium, aluminum, iron, or zinc, and a positive electrode, which takes oxygen from atmospheric air to generate energy. Several exhaustive review papers on MABs are available in the literature (see f.i. [158,159,160] just to cite a few of them).

In spite of their huge potential, some drawbacks, such as large overpotentials associated with charge and discharge processes, still hinder their commercialization. Relevant efforts have been devoted to the development of inexpensive and highly effective bifunctional cathode catalysts able to promote both the oxygen evolution reaction (OER) and the oxygen reduction reaction (ORR), which directly govern the MAB rechargeability. Up to now, materials based on precious metals (platinum or iridium) supported on carbon are still regarded as the best ORR and OER catalysts [161], but their high prices represent a serious concern in view of the use on a large-scale. In the last few years, several alternative materials including graphene [162], transition-metal oxides (e.g., perovskites and spinels) [163,164], and porous carbon-based materials [165] have been evaluated as bifunctional catalysts.

Electrospinning that allows for the preparation of porous graphite-like carbon fibers embedding or supporting metal and metal oxide NPs strongly interacting with the carbonaceous matrix has received great attention and has become a popular and efficient technique widely adopted to produce bifunctional catalysts [161,166,167,168,169,170,171,172,173]. Just to cite some examples, Alegre et al. [173] have recently designed highly stable bifunctional catalysts for air electrodes consisting of a combination of metallic cobalt and cobaltous oxide NPs finely dispersed on the surface of ECNFs. The Co/CoO/ECNFs outperform the state-of-the-art catalysts in the O₂ evolution. Also, self-standing Co₃O₄/ECNF mats prepared by Song et al. [167], by using PAN and zeolite imidazolate frameworks as precursors, exhibit superior performance as a cathode for MABs. Besides, neither binder nor conductive metal foam is required for their use. Churros-like metal-free nitrogen-doped ECNFs, featured by internal nanochannels, have been obtained by Park et al. [166] from a bicomponent polymer blend, with PAN and PS respectively acting as carbon-source and pore-forming sacrificial agent, as frequently reported also for different applications [153]. The ORR activity of these fibers strongly depends on the temperature at which carbonization is operated (800–1100 °C), which affects not only the specific surface area but also the concentration of beneficial surface graphitic-N species. Electrospinning has then been utilized to produce the perovskite component of a hybrid bifunctional catalyst consisting of La_0.5Sr_0.5Co_0.8Fe_0.2O₃ porous nanorods (LSCF-PR) and nitrogen-doped reduced graphene oxide (NRGO), active towards both ORR and OER [161]. As schematically depicted in Figure 18a, the LSCF-PR/NRGO composite is obtained by dispersing in Nafion/ethanol solution LSCF-PR synthesized via electrospinning followed by calcination and GO reduced and doped with nitrogen through exposure to NH₃ atmosphere at high temperature. LSCF-PR (Figure 18a,e), uniformly distributed throughout the NRGO sheets (Figure 18d), is responsible for the most of OER activity, whereas the NRGO component (Figure 18c) primarily contributes to the ORR activity.

6. Supercapacitors

A wide branch of research on energy storage is focused on the electrochemical capacitors [174,175], better known as supercapacitors and also including pseudocapacitors. Supercapacitors and pseudocapacitors differ for the mechanism through which the charge is stored. In the former, the charge storage is based on the fast electrostatic interactions at the electrode surface in contact with the electrolyte (Figure 19)—during the charging process, the ions from the electrolyte are reversibly adsorbed on the high-surface-area electrodes, finally leading to an electrical double layer (EDL). In the latter, fast and reversible redox (Faradaic) processes mainly take place at the surface of the electrodes.

Graphene and other carbonaceous materials show EDL behavior and are featured by high power densities [175]. Transition metal oxides (such as RuO₂, MnO₂ Fe₃O₄, V₂O₅, and Fe₃O₄ [48,176,177]) show pseudocapacitive behavior and possess high energy density [143]. Hybrid capacitors, consisting of a pseudocapacitive electrode and an EDL electrode, try to efficiently combine the two characteristics.

In spite of the higher specific capacitance generally delivered by materials with pseudocapacitive behavior [48,176,177,178], in recent years, the research efforts are being increasingly focused on sp²-carbon-based materials, which, with their tunable pore structure, abundant surface functionalities, and active defective sites, offer appealing advantages for all applications concerning energy conversion and storage [179] and, in the cases of supercapacitors, also the possibility of fabricating stretchable devices [180].

Since good performance of capacitors with EDL behavior requires appropriate pore size distribution (micro-, meso-, and macro-pores), high electrochemically accessible surface area, and high electrical conductivity [175], porous carbon materials have received great attention. Among them, electrospun carbon fibers, which couple short diffusional paths and flexibility to high conductivity, large aspect ratio, and surface area, represent the most promising candidates.

Table S5 reports some examples of electrode ENMs and their preparation conditions; the main parameters describing their performance are reported in Table S6. PAN [181,182,183,184,185,186,187] or a blend of PAN and PVP [52,53,54,59] is the most frequently utilized polymeric component of the spinnable solution, whereas N,N-dimethylformamide (DMF) commonly acts as a solvent [52,53,54,59,181,182,183,184,185,186,187]. A variety of different additives is employed to generate carbon fibers with specific textural, compositional, or conductive properties [46,47,59,178,182,183,184,185,186,187].

By the use of a blend of PAN and PVP dissolved in DMF, Li et al. [54] have obtained non-woven membranes consisting of porous nitrogen-doped carbon fibers, whose porosity and nitrogen-content can be easily controlled by adjusting the relative-to-PAN PVP content (150–240/100 wt%/wt%) of the solution. As binder-free electrodes, the membranes are able to deliver specific capacitance of 148–198 F/g and exhibit excellent cyclic stability with a retention rate of 104% after 5000 cycles at a rate of 1 A/g [54].

Fan et al. [59] have produced binder-free flexible electrodes for supercapacitors consisting of nitrogen-enriched carbon fiber networks, by adding silica (SiO₂) NPs to the spinnable PAN+PVP/DMF solution (Figure 20a–c). SiO₂ NPs act as a as a sacrificial templating agent and, after carbonization (Figure 20d–f), are removed by chemical NaOH etching (Figure 20g), finally giving rise to the formation of meso-macroporous fibers with a pretty morphology featured by conjugated surfaces (Figure 20h,i). The specific capacitance delivered by these electrodes is 242 F/g at a rate of 0.2 A/g. They are able to retain 99% of the initial capacitance after 5000 cycles [59].

The size of the electrodes tested in the research works is generally small (their surface area does not usually exceed 1 cm² [54,59,183,184]). Kim et al. [182] have built supercapacitor cells by assembling two 2.25 cm² electrodes and, more remarkably, have demonstrated the possibility of depositing an electrospun thin A3-sized (297 mm × 420 mm) fibrous membrane. The addition of (1−5 wt%) zinc chloride to the PAN/DMF solution, enhancing its conductivity, brings about a fiber diameter shrinking during the spinning stage (from 350 to 250 nm). During the subsequent heat treatments (stabilization and carbonization), producing a further fiber thinning, ZnCl₂ exerts a catalytic function, accelerating the oxidative stabilization rate and creating micropores on the outer surface of the fibers by etching C atoms [182].

The specific capacitance of the electrodes based on electrospun carbon fibers (Table S6) usually does not exceed 250 F/g [52,53,54,59,181,182,183,184,188]. Various additives, such as graphene nanoplatelets [184], biochar [183], and carbon nanotubes (CNTs) [46,47,185], have been used to improve their performance.

By adding 12 wt% CNTs to the polyaniline/polyethylene oxide (PANI/PEO) spinnable solution, Simotwo et al. [46] have fabricated free-standing, binder-free supercapacitor electrodes by a single-step process requiring no heat treatment. The PANI/CNTs-based electrodes are able to deliver a capacitance of 385 F/g at a current density of 0.5 A/g with a retention capability of 81.4% after 1000 cycles. The three-dimensional electrode structure characteristic of the non-woven interconnected fiber network is responsible for their good electrochemical performance since it facilitates efficient electron conduction, while the inter- and intra-fiber porosity enables excellent electrolyte penetration within the polymer matrix, allowing fast ion transport to the active sites [46].

Tebyetekerwa et al. [47] have used a similar single-step process to fabricate self-supporting supercapacitor electrodes by electrospinning a CNT-enriched polyindole-based polymeric solution directly on stainless steel current collectors (Figure 18a,b). The flexible symmetric all-solid-state supercapacitor fabricated by interposing PVA/H₂SO₄ gel electrolyte between the two so-prepared electrodes (Figure 21c,d) is able to deliver 476 F/g at a current density of 1 A/g, with an excellent retention capability (95% after 2000 cycles) [47].

At the same rate, much higher specific capacitance (1119 F/g) has been obtained by Agyemang et al. [185] by coating CNT-enriched carbon fibers (produced by stabilization and carbonization of as-spun PAN/CNTs fibers under conditions reported in Table S5) with aniline monomer, through in situ chemical polymerization, to form PANI-coated C/CNTs fibers (Figure 22). At very high current density (10 A/g), these composite fibers are still able to deliver 748 F/g. Their high specific capacitance is attributed to the synergic effects of the combination of the high surface area resulting from CNT incorporation, the pseudocapacitive property of PANI, which increases the electrical conductivity, the interaction between the CNT-enriched C fibers and PANI, and the uniform distribution of CNTs within the fibers [185]. PANI-coated C/CNTs fibers are able to retain 98% of the initial capacitance even after 2000 cycles at a current density of 10 A/g [185].

Outstanding stability (86% capacitance retention after 10000 cycles) and excellent flexibility (but associated to much lower specific capacitance values (~93 F/g)) has been reported by Chen et al. [186] for fiber-shaped all-solid-state supercapacitors fabricated, as shown in Figure 23, by using a binder-free carbon nanofibers yarn (CNY) with a core-shell structure (Figure 23e). CNY has been prepared by carbonizing twisted electrospun PAN nanofibers (Figure 23a,b) and subsequently depositing polypyrrole (PPy) on the surface of the CNY (Figure 23c) and reduced graphene oxide (rGO) on the surface of the resulting PPy@CNY (Figure 23d) to finally form a core-shell structure (CNY@PPy@rGO). The high porosity of the flexible CNY results in a large specific surface area for PPy deposition. The deposition of PPy on CNY improves the specific capacitance due to the high pseudocapacitance of PPy, while rGO layer contributes to enhance the electrical conductivity and act as a physical buffering layer. Thanks to the core-shell structure of CNY@PPy@rGO and synergistic effects of the three components, the electrode exhibits a length capacitance of 111.5 mF/cm [186].

7. Capacitive Deionization of Water

The increase in the world population is reflected in a growing demand for fresh water, further boosted by the increasing level of environmental contamination of the commonly exploited aquifer sources. This situation has encouraged the development of the water desalination technologies, and particularly of those addressed to cost-effectiveness and energy-efficiency. Among them, the capacitive deionization (CDI) is one of the most attractive for its simplicity and eco-friendliness, and it has recently received great attention [189,190,191].

CDI, also known as electrosorptive desalination, is based on the formation of an EDL between the parallel electrodes forming the CDI cell, upon application of a voltage between them [192,193,194,195]. In the presence of the EDL, the Na⁺ and Cl⁻ ions, present in the salty water inlet into the cell, are reversibly adsorbed onto the porous surface of the negatively and positively charged electrodes, respectively (Figure 24), and the so-obtained desalted water goes out from the cell. During this process, no secondary contaminant is released. Once their surface is saturated by salt ions, the electrodes can be regenerated by reducing/reversing/shorting the cell voltage. This causes the stored ions to be desorbed and released into a wastewater solution.

The low operating voltage of CDI (0.8–1.6 V) can be of great help in remote areas. In addition, the energy required to desalinate one cubic meter of brackish waters by this method is lower than for the reverse osmosis (only 0.1 against 0.2 kWh) [194].

The capacitance of the porous electrode material and, hence, its desalination performance depends on many parameters, including its specific surface area and hydrophilicity, the size and distribution of its pores, as well as its ionic conduction properties [189,191,192,193,194,195]. Based on the electrosorption mechanism and kinetics, porous carbons with good electrical conductivity, hierarchical pore distribution, and high specific surface area are commonly regarded as the optimal materials for the achievement of high CDI performance. A great variety of carbon-based electrode materials, including carbon aerogel [196,197], activated carbons [198], carbon nanotubes [199,200], graphene [201], and graphene-enriched carbonaceous nanostructures [202], have been developed in order to reach this goal. However, some of these materials (e.g., nanotubes) are produced through time-expensive and/or sophisticated processes and have high costs, while others (e.g., activated carbons) involve the use of harsh chemical oxidants, such as KOH or piranha-solution, for activation before their use as CDI electrodes. Electrospinning is a simple and cost-effective technique; moreover, electrospun carbon nanofibers (ECNFs) do not necessarily require activation treatments to exhibit good electrosorptive performance [55].

In the last decade, electrosorption capacities ranging between 1.9 and 6.4 mg/g have been reported for ECNFs and ECNF-based composite materials [203,204,205,206,207,208,209]. In particular, Dong et al. [206] have utilized electrospinning followed by CO₂ activation to prepare a hybrid nanomaterial (Figure 25a–c) consisting of CNTs embedded in activated PAN-based carbon fibers (ACF). For this purpose, CNTs have been dispersed in a PAN/DMF solution, and the resulting homogeneous CNT/PAN suspension has been electrospun by the use of the spinning set-up schematized in Figure 25d. After stabilization in air at 280 °C for 2 h, the as-obtained web-like CNT/PAN fibers, have been carbonized in N₂ at 800 °C for 2 h and finally activated for 1 h in CO₂ at the same temperature, obtaining a composite material (CNT/ACF) with high specific surface area (651 m²/g) and mesopore ratio (64%). The test of the composite nanostructures by means of the setup schematically depicted in Figure 25e has led to a desalination capacity of 6.4 mg/g at an initial NaCl concentration of 400 mg/L and a working potential of 1.2 V. The great improvement (39%) with respect to the CNT-free nanomaterial, prepared and tested under the same conditions, has been ascribed to the increase of conductivity brought about by the incorporation of the high aspect-ratio morphology conductive additive coupled with the high porosity due to the CO₂-mediated activation [206].

Indeed, activation by means of different agents is a strategy followed by various groups to enhance the specific surface area of the ECNFs and improve their NaCl removal capacity [203,207,210]. Liu et al. [210] have reported an electrosorption capacity of 10.52 mg/g for PAN-based ECNFs obtained by stabilization in air at 250 °C for 1.5 h, followed by carbonization in N₂ at 800 °C for 1 h and activation by means of ZnCl₂. Activation consists of impregnation of ECNFs by immersion in a ZnCl₂ solution, subsequent heat-treatment in N₂ atmosphere at 500 °C, immersion in 1 mol/L hydrochloric acid, and removal of the residual zinc compounds through thorough washing. The consequent increase in specific surface area, micro- and meso-pore volume, and electrode surface wettability by water is thought of as responsible for the relevant improvement in the CDI performance.

Very recently, remarkably higher electrosorption capacities (>15 mg/g at initial NaCl concentrations of 500–600 mg/L) have been reached by the use of graphene-enriched nanohybrids as CDI electrodes [55,211]. Graphene oxide (GO), incorporated at any stage during their preparation and transformed to thermally reduced graphene oxide (TRGO) during carbonization, acts as a bi-dimensional (2D) conductive additive, thus improving the conduction properties of the TRGO-containing hybrid.

Luo et al. [211] have produced a carbon aerogel with a structure consisting of 1D and 2D carbon nanomaterials, namely ECNFs coupled to graphene (Figure 26a,b), following the three-step procedure sketched in Figure 26c. Briefly, the membrane consisting of PAN-based ECNFs stabilized in air at 250 °C for 2 h (s-PAN), is cut into pieces and, after homogenization in a mixed solvent, is filtered. The resulting short fibers are redispersed, mixed with a GO suspension and homogenized at a very fast rotating speed to assure the sufficient contact and interactions between s-PAN fibers and GO sheets. s-PAN/GO composite aerogel is obtained via freeze-drying of the resulting material, whereas subsequent carbonization at 800 °C in N₂ for 2 h of the s-PAN/GO composite aerogel leads to PG carbon aerogel. With respect to the pristine s-PAN/GO composite aerogel, PG carbon aerogel is endowed with higher specific capacitance and remarkably improved electrosorption capacity (Figure 26d). The enhanced electrochemical performance is due to the strong interconnection between the 1D and 2D aerogel components, which favors the electron transfer within the porous structure.

Belaustegui et al. [55] have proposed an unconventional approach to improve the electrochemical performance of ECNFs in water desalination by CDI. As mentioned above, increasing the surface area of the ECNFs via different methods is a commonly followed strategy to improve their salt removal capacity from water. However, as pointed out by Huang et al. [212], the use of extremely high-surface area carbons may be unprofitable since micropores, which contribute to enhancing the material specific surface area, are hardly accessible for the ions and, hence, not active in CDI process. On the other hand, water wettability is known to be very beneficial to the diffusion and adsorption of ions in the solution to the fiber surface and results in improved surface utilization [212]. It is widely recognized that a high N-doping level of the active material is beneficial for the wettability of the electrode surface and pseudo-capacitance, as well [6,195]. Based on these concepts, Belaustegui et al. [55] have demonstrated that ECNFs self-doped with very high nitrogen concentrations (around 20 wt%) are able to remove a relevant amount of NaCl (17.0 mg/g) from a salty solution with an initial concentration of 585 mg/L, and that the electrosorption capacity outstandingly increases (up to 27.6 mg/g) if the N-doped fibers are enriched with graphene (Figure 27). The N-doped ECNFs have been produced (Figure 27a) by choosing a polymer (PAN) with a high N-content (26.4 wt%) and a low carbonization temperature (500 °C). The latter condition results in lower energy consumption and limited release of volatile N-containing by-products, with beneficial effects on the cost-effectiveness, safety, and environmental friendliness of the ECNF production process [55]. Moreover, the superior performance of these fibers with relatively small specific surface areas (17–20 m²/g) with respect to other nanomaterials (e.g., N-doped graphene sponge [201]) featured by smaller N-contents and larger specific surface area demonstrates the crucial role of surface wettability and pore accessibility.

8. Hydrogen Production by Photo-Assisted Electro-Chemical Water Splitting

The need for limiting the effects of using fossil fuels on climate changes and the increasing energy demand associated with the growth of the world population and the enhancement of living standards call for developing clean energy sources. Hydrogen is widely recognized to be a carbon-free, environmentally friendly, renewable, and efficient fuel with high energy density. It has a calorific value nearly three times greater than hydrocarbon fuels. Currently, the production of H₂ mainly comes from steam reforming of natural gas [213], consuming natural resources and generating carbon dioxide as an undesired by-product. Solar energy conversion for H₂ production via photo-induced water splitting (WS) represents one of the most promising alternative approaches, but presently still contributes to a very small extent (<4%) [213].

The process is schematically depicted in Figure 28. Essentially, photo-electrochemical (PEC) cells act as “artificial photosynthetic devices” that mimic the process of hydrogen production through the light-assisted water splitting occurring in nature [214]. In the cell (Figure 28a), the anode (based on a photoactive n-type semiconductor) is the oxygen-evolving electrode, whereas hydrogen is generated at the cathode, which acts as the counter-electrode [216].

When a light source illuminates the n-type semiconductor acting as photo-anode, the electrons (e⁻) in the valence band (VB) are excited to the conduction band (CB), and holes (h⁺) are generated in the VB (Figure 28b). The photo-generated e⁻ and h⁺ that do not undergo recombination act as the reducing and oxidizing agents, respectively. Electrons participate in the reduction reaction and produce H₂, whereas holes carry out oxidation reaction and produce O₂ [213].

To achieve an overall WS in hydrogen and oxygen, both the reduction and oxidation potentials of water should lie within the band gap (E_g) of the photo-catalyst. The bottom of the CB must be more negative than the reduction potential of H⁺/H₂ (0 V vs. normal hydrogen electrode, NHE), while the top of the VB must be more positive than the oxidation potential of O₂/H₂O (1.23 V vs. NHE) [216,217,218].

Figure 29 displays the energy gap (in eV) and band edge positions of some representative semiconductor materials relative to water redox potentials. Only some of them are suitable for overall WS in hydrogen and oxygen. Moreover, since the intensity of the solar spectrum dramatically falls off below 400 nm, to absorb the visible-light irradiation from sunlight, E_g should be lower than 3.1 eV [216].

Titanium dioxide (TiO₂) has been the pioneering semiconductor for the photo-catalysis of water [220]. It is environmentally friendly, easily available, cheap, and stable. Nevertheless, absorbing in the near UV region, it can use only an exiguous fraction (5%) of the solar light. In principle, the narrower band gap (2.1–2.3 eV) of hematite enables the absorption of about 40% of the solar spectrum. Thanks to its favorable properties (VB edge lower than the water oxidation potential, chemical stability in a oxidative environment, low cost, and non-toxicity) α-Fe₂O₃ might represent an attractive alternative to TiO₂ for the application in PEC WS on a large scale [214]. However, poor conductivity (of the order of 10⁻⁶ S cm⁻¹ for high purity hematite), short hole diffusion length (typically 2–4 nm) and fast electron-hole recombination (10⁻¹² s) at the grain boundaries mainly limit the PEC activity of polycrystalline hematite films.

The design of efficient and stable photo-catalysts for the harvesting of solar energy has become one of the primary challenges in the development of solar H₂ economy [221,222]. Many efforts have been made to increase the optical absorption of large-E_g photo-catalysts by varying their chemical composition with metal or non-metal doping or self-impurity [223,224,225]. Also, nano-structuring has allowed obtaining even 1000 h stable photo-anodes [226].

1D-nanostructures, such as nanorods, nanowires, or nanofibers, offer advantages over their bulk counterpart [227]. Besides, minimizing the length scale through which the minority carriers must diffuse and, hence, reducing the probability of recombination losses, 1D-nanostructures may contribute to enhance the photo-conversion efficiency thanks to their higher surface area. In this scenery, the electrospun nanofibres, featured by high porosity, aspect ratio, and surface area, represent very promising candidates for the efficient production of H₂ by PEC WS. Table S7 reports some examples of electrospun photo-catalysts and summarizes their preparation conditions. Figure 30 shows the morphology of some of them.

Under UV light irradiation, by the use of electrospun TiO₂ nanofibers, hydrogen evolution increases by a factor of ~2.8 with respect to TiO₂ prepared by a hydrothermal method, and PEC activity is higher than for reference commercial TiO₂ nanoparticles [228]. Lee et al. have reported an H₂ evolution rate of 62.7 µmol/h and up to 200 µmol/h by using TiO₂/CuO [229] and TiO₂/SnO₂ (Figure 30a) [230] composite fibers, respectively.

After optimization of the preparation conditions, photo-current densities ranging between 150–200 μA/cm² and 0.6–1.4 mA/cm² have been measured for pure and N-doped electrospun TiO₂ anodes, respectively [231,232]. Dispersions of the sintered TiO₂ fibers produced via ES have been employed to obtain photo-anodes for WS in the form of homogenous films on FTO/glass substrates by the use of doctor-blade, sol-gel dip coating, or by combining the two techniques. In this case, the highest photocurrent density obtained has been 0.5 mA/cm² [233].

Al-Enizi et al. [234] reported an improvement in the generation of hydrogen gas through the hydrolysis of ammonia borane complex by the use of carbon fibers decorated with Co-TiC nanoparticles-like superficial protrusions (Figure 30b) prepared through a one-step electrospinning approach.

Interesting results have also been obtained by combining electrospinning with different synthesis techniques [235,236]. For instance, under optimized conditions, C-coated TiO₂/WO₃ fibers prepared by electrospinning combined with hydrothermal method have shown remarkable light absorption in the visible region and enhanced H₂-generation rate thanks to the multichannel-improved charge-carrier photosynthetic hetero-junction system with the C layer on the surface of TiO₂ as an electron collector and WO₃ as a hole collector, leading to effective charge separation on these components [236].

Gao et al. [237] have greatly enhanced the H₂ production rate using electrospun TiO₂/WO₃/Au nanofibers. The enhanced PEC activity of TiO₂/WO₃/Au with respect to TiO₂ and TiO₂/WO₃ nanofibers has been attributed to the synergistic effect of Schottky and surface plasmon resonance (SPR) due to Au NPs decorating the surface of the composite fibers. In the Z-scheme photosynthetic heterojunction system, WO₃ and Au respectively act as h⁺-collector and e⁻-collector, promoting effective charge separation, whereas the SPR effect favors the absorption of visible light.

Recently, Saveh-Shemshakia et al. [238] have demonstrated the possibility of fabricating efficient photo-anodes by direct deposition of electrospun hematite fibrous mats over FTO/glass plates. The physicochemical properties of the oxide fibers can be effectively modified by the insertion of elemental impurities in their lattice [51,239,240]. A preliminary study has shown that the doping electrospun of Fe₂O₃ fibers with aliovalent impurities is a viable route to improve their conductive properties and, hence, their PEC performance [241]. The photocurrents measured in these preliminary experiments (0.94 μA/cm² and 1.42 μA/cm² for pure and Si-doped Fe₂O₃ fibers)—comparable to those reported for hematite films produced by different technique and precursors [242]—are by far below the benchmark for the optimized electrospun hematite photo-anodes (0.5 mA/cm² [238]). Nonetheless, the nearly threefold increase in the donor concentration (from 1.51∙10¹⁹ to 4.35∙10¹⁹ cm⁻³) brought about by doping is responsible for the remarkable enhancement (53%) of photocurrent [241]. In the presence of hole scavengers, such as H₂O₂, the surface recombination almost disappears, and photocurrent density increases by one order of magnitude (from 1.42 to 15 μA/cm² at 1.23 V [241]).

Aiming at outlining the larger number of energy-related applications as possible, only the photo-assisted approach to WS has been reviewed here. Of course, also electrospun catalysts able to catalyze the hydrogen and oxygen evolution reactions without the aid of light deserve to be mentioned, as well as ENMs utilized in other technologies for clean energy production, such as fuel cells (see f.i. [243,244,245,246]) and solar cells (see f.i. [247,248,249,250]).

9. Nanogenerators for Energy Harvesting

Energy harvesting is another emerging application field of the ENMs worthy of mention. In recent years, the number of portable or wearable devices available on the electronics market has rapidly multiplied. Owing to the constant enhancement of the living standard of the world population, this trend is not destined to be reversed in the future. Currently, the large majority of these electronic systems uses batteries as external power sources. This implies limited utilization time of the devices and, in second place, great impact on the environment due to the consumption of resources for the battery fabrication and the production of waste at the end of their life cycle. Therefore, besides reducing their weight and size to increase their portability, it is desirable to provide electronic devices with sustainable, self-sufficient power sources to ensure the independence and continuity of their operation.

In recent years, harvesting energy from the surrounding environment (e.g., from acoustic waves, mechanical loads, wind or machine vibrations, human body movements) and converting it into electricity to power these systems has been established as a viable route to provide autonomous portable or wearable electronic devices. This has encouraged the development of various energy-harvesting techniques and micro- and nano-generator technology following the tendency toward the progressive device miniaturization [251]. The majority of energy generators harvest mechanical energy and exploit piezoelectric or triboelectric effect to convert it into electrical energy to supply the self-powered devices into which they are integrated [251,252,253,254].

As schematically illustrated in Figure 31, a non-centrosymmetric crystal structure features piezoelectric materials; lattice distortion produced by the application of stress/strain induces polarization charges, which generate an electrical potential. The transient electron flow driven across the electrode pair applied to the material by this potential can supply low-power electronics. Based on this working mechanism, piezoelectric nanogenerators (PENGs) harvest waste mechanical energy from walking, foot- and finger-tapping, talking, breathing, bending, twisting, and other body movements and transform it to useful electric power [251,255].

The working mechanism of triboelectric nanogenerators (TENGs) is different. They are based on static electrification induced through frictional contact between different dielectric materials, with the polarity of the induced charge depending on the relative polarity of the two materials [251,256]. Figure 32 sketches the four chief kinds of triboelectric-harvesting device geometry utilized for the fabrication of different types of TENGs [251,256]. A detailed description of their operating modes, device design, and performance enhancement, as well as an accurate illustration of the fundamentals of this technology, can be found elsewhere [252,253,254,256,257,258,259].

Polyvinylidene fluoride (PVDF), a semicrystalline polymer with alternating hydrogen and fluorine units attached to the carbon chain, is one of the most efficient piezoelectric materials [251,255]. It is biocompatible, flexible, chemically resistant, and endowed with high mechanical strength. Its piezoelectricity is due to the electroactive β-phase of the polymer that possesses spontaneous polarization. Electrospun PVDF and PVDF-based composites are extensively utilized in PENGs [251,255,260], since, during electrospinning, the other phases of the polymer can become electroactive (by poling or mechanical stretching) [255]. In addition, with respect to the piezoelectric inorganic semiconductors, PVDF and its copolymers have a greater potential towards flexible and wearable applications because they are lightweight and can be easily processed.

Electrospun PVDF membranes are also utilized in the fabrication of TENGs [261,262,263]. PVDF membranes coated with polydimethylsiloxane to improve their mechanical robustness and optimize their porosity properties have also been employed to manufacture a wearable TENG capable of converting human biomechanical energy into electricity for next-generation wearables [261]. A humidity-resisting TENG, designed to adapt to the environmental humidity caused by human perspiration during sport, has been constructed by the use of PVDF and surface amino-modified cellulose acetate/polyurethane electrospun membranes [262].

Most recently, EESs and energy-harvesting devices have combined into self-charging power systems for wearable and portable electronics of daily use [264]. Sun et al. [265] have designed an ultra-light and flexible self-charging power system entirely based on electrospun paper-like membranes. The system consists of a TENG as an energy harvester and supercapacitors as EESs. The former is arch-shaped and exploits conductive PAN-derived carbon paper as an electrode and nonconductive stabilized-PAN membrane as a triboelectric layer. In the latter, the nonconductive PAN paper acts as a separator, whereas the conductive one is used as a capacitive electrode material.

10. Textiles for Energy Saving through Personal Thermal Management

In the latest years, the need for limiting GHG emissions has greatly boosted the development and utilization of renewable carbon-free energy sources. Cooling and heating systems take up an appreciable fraction of the global energy demand and significantly contribute to enhancing the GHG emissions [266,267]. Hence, another important action against climate changes and extreme weather can be the reduction of these energy consumptions through the improvement of energy efficiency. Taking into consideration the residential and commercial buildings, the principal energy demand originates from indoor cooling and heating. However, a large portion of the energy spent on this purpose is wasted for empty spaces and inanimate objects. For a long time, the approach commonly followed to cope with this problem has consisted of trying to improve the building thermal isolation through the use of innovative materials and suitable design.

In recent years, a new and complementary approach has emerged, intended to profit from the smaller thermal mass of human beings compared with buildings. This approach is centered on providing cooling or heating only in the immediate environment around the human body, to minimize wasted energy and, thus, enhance efficiency [268]. Cui’s group has devoted several efforts in the field of personal thermal management (PTM) [266,267,268,269]. The group has developed wearable textiles with a superior cooling [266,270] or heating effect [269] with respect to common commercially available textiles, able to mitigate the energy demand for outdoor cooling or indoor heating. In particular, Chen et al. [271,272] have utilized the co-axial electrospinning technique to prepare phase-change polyethylene glycol/cellulose acetate (PEG/CA) fibers with great potential in PTM. In the ultrafine composite core-shell fibers, CA and PEG act as a supporting fiber-template material and as a model phase-change material, respectively. The electrospun fibrous material exhibits good thermo-regulating capability. For further information on electrospun phase-change fibers see ref. [8,273].

11. Conclusions

Electrospinning is a very simple, widely accessible, scalable, and cost-effective technique, particularly suitable for the manufacturing of engineered fibrous nanomaterials on an industrial scale. In fact, different from the other synthetic techniques, the production process via electrospinning can proceed continuously, by collecting the as-spun fibers on a suitably designed system using rotating drums. Any thermal treatment needed to modify the fiber physicochemical properties can be operated separately, with no necessity of interrupting the production process. Obviously, this constitutes a singular inherent advantage in view of the large-scale production. Besides, in an industrial context, multi-needle and needleless systems allow facing up to the moderate productivity of the conventional electrospinning.

The outstanding physical and chemical properties of the large variety of intriguing nanostructured materials produced via this fascinating synthetic approach are suitable to the development of a huge number of advanced applications involving the most different fields, exhaustively illustrated in several excellent reviews. The focus of this review is on energy-related applications of the electrospun nanomaterials. In particular, the recent advances produced by their use in secondary alkali metal-ion batteries, redox flow batteries, metal-air batteries, supercapacitors, reactors for water desalination via capacitive deionization, hydrogen production by solar-driven water splitting, nanogenerators for energy harvesting, and textiles for energy-saving personal thermal management are reviewed, trying to report aspects believed to be of importance to attract the interest and stimulate the inventiveness of the readers.

Even if outstanding results have been achieved, many are the challenges still to be faced, bearing in mind the ultimate goal of pushing electrospun nanomaterials from the laboratory to industry. For instance, although doping with aliovalent elements has enhanced their conductive properties, the mechanical strength of ceramic nanofibers needs to be improved, and solutions for the production of flexible mats for their use as free-standing electrodes have still to be found. In the case of carbonaceous fibrous membranes, despite the progress recently made, the tendency towards the aggregation of the dispersed very small-sized nanoparticles responsible for the electrode instability still needs attention. Besides, the lack of full control of their pore size and distribution hinders from having carbon fibers with finely tunable porosity and hierarchical porosity, thus limiting the performance of all applications benefiting from them. A similar necessity concerns fiber diameter and mechanical robustness, as well as orientation. Another absolutely non-negligible aspect is represented by the impact on human health and the environment. The solvent (N,N-dimethylformamide) utilized to disperse the polymers (polyacrylonitrile and polyvinyl pyrrolidone) most frequently selected for the fabrication of carbon-, carbon-based, and ceramic fibers are not free of toxic effects.

In light of these considerations, some future research directions in the field of energy-related applications of electrospun nanomaterials could be:

• Search for safer polymer/solvent combinations and abundant, non-toxic precursors to limit the impact of the fiber production on human health as well as the environment.
• Development of ceramic fibers with improved mechanical strength for the production of flexible mats to be used as free-standing electrodes.
• For the composite fibers, improvement of the existing approaches to contrast the tendency towards the aggregation of the very small-sized nanoparticles embedded into the carbon matrix.
• Search for new additives and development of new fiber surface-modification methods to generate electrospun nanomaterials endowed with new functionalities.
• Introduction of new methodologies for the full control of the fiber diameter to be able to produce ultrathin fibers, as well as of the fiber pore size and distribution to enable the fine tunability of their (even hierarchical) porosity.
• Design of new (e.g., suitably patterned) collectors to realize fibrous mats endowed with specific shapes and fiber orientations.
• Invention of new spinning geometries allowing for the integration of different functionalities in a single-layered multi-component fibrous mat.

For sure, all the challenging open questions will motivate the scientific community, leading them to make significant breakthroughs in the future. This will permit the move beyond the current state of the electrospun nanomaterial technology, towards their commercialization and penetration in daily life.