Electrospun Nanofibers for Integrated Sensing, Storage, and Computing Applications

Abstract

Electrospun nanofibers have become the most promising building blocks for future high-performance electronic devices because of the advantages of larger specific surface area, higher porosity, more flexibility, and stronger mechanical strength over conventional film-based materials. Moreover, along with the properties of ease of fabrication and cost-effectiveness, a broad range of applications based on nanomaterials by electrospinning have sprung up. In this review, we aim to summarize basic principles, influence factors, and advanced methods of electrospinning to produce hundreds of nanofibers with different structures and arrangements. In addition, electrospun nanofiber based electronics composed of both two-terminal and three-terminal devices and their practical applications are discussed in the fields of sensing, storage, and computing, which give rise to the further integration to realize a comprehensive and brain-like system. Last but not least, the emulation of biological synapses through artificial synaptic transistors and additionally optoelectronics in recent years are included as an important step toward the construction of large-scale, multifunctional systems.

1. Introduction

One-dimensional (1D) nanostructures have attracted growing interest in the past few decades due to their fascinating properties and unique applications in areas ranging from physics, chemistry, biology, and material science [1,2,3,4,5]. Among different kinds of preparation methods for 1D materials, such as arc discharge, chemical vapor deposition (CVD), laser sputtering, and liquid-phase synthesis, the electrospinning method has been a standout for its simplicity and cost-effective characteristics for producing nanofibers. Currently, hundreds of polymers, inorganic compounds, and composites with one-dimensional nanostructures have been successfully developed based on electrospun materials. With the continuous research, it is believed that more and more types of 1D nanomaterials will be fabricated by the electrospinning method, which will play an increasingly important role in the field of science and technology shortly.

Electrospinning evolves from electrospray technology, and the key technique of electrospray is to destroy the equilibrium of the droplet: the electric field force must be increased to exceed the droplet surface tension so that the bigger droplet will split into smaller charged droplets, which is also known as “Rayleigh instability” phenomenon [6]. Based on this principle, electrospinning produces nanofibers by volatilization or solidification of the jet stream. Because of high aspect ratio, high specific surface area, and porosity of obtained nanofibers, together with effective control of the fiber arrangement and size, electrospinning would have great prospects for applications in electronic devices, energy storage, and biomedicine. This review paper aims to introduce the fundamentals of electrospinning and the current development of electrospun nanofiber-based electronics consisting of field-effect transistors, memory devices, and sensors, which suggests great possibilities for further research. State-of-the-art fiber producing methods, including coaxial electrospinning, near-field electrospinning, and corresponding complex fiber structure, are presented carefully. Furthermore, electrospun nanofiber-based electronic devices consisting of field-effect transistors, memory devices, sensors, and other electronic applications are discussed in detail. Finally, an outlook on the potential for comprehensive systems and some drawbacks are provided.

2. Different Types of Electrospun Nanofiber Process and Structures

As one of the most convenient and efficient ways to prepare nanofiber in today’s scientific research, the electrospinning technique emerged in the experiment of German scientist Anton Formhals in the 1930s, and it was originally used to prepare artificial silk-like fibers [7]. This ancient nanotechnology has been rediscovered for modern demands by Reneker’s group, who officially introduced the professional term “electrospinning” in the 1990s [8]. Today, blowout research has made this technique one of the most important methods for the multi-functional, large-scale, and low-cost preparation of continuous nanofibers with diameters as small as nanometers. Generally, the composition of this technology is very simple, mainly including a high-voltage power source, a syringe pump, a spinneret, and a conductive collector (Figure 1) [9]. In the electrospinning process, the droplets ejected by the spinneret generate jets, which are then stretched into nanofibers with the help of the electric field and deposited on the collector. These systems can be easily assembled to realize a spinning system in any laboratory. Electrospun nanofibers have a cross-linked ultra-fine fiber structure, which gives the material the characteristics of large specific surface area, light weight, high porosity, excellent thermal stability, and mechanical flexibility [10,11,12,13]. They have broad applications in filtration membranes, catalytic supports, energy storage, biomedical research, and photonic and electronic devices [14,15,16,17,18,19,20,21,22,23,24]. At present, the applications of electrospinning have not only been studied in the laboratory, but have also made some achievements in the field of commercialization. The first commercial product based on electrospun nanofibers, namely the air filtration technology Ultra-Web^®, appeared in 1981 with the work of the Donaldson company [25]. Recently, the international market for electrospun fiber-based applications has been brisk, and the total value reached 927 million dollars in 2018 [26]. It is believed that in the future, the 1D nanomaterials prepared by electrospinning will continue to be developed and this will give full play to their potential to be used in various industries.

Electrospinning was originally used to spin polymer solutions into polymetric nanofibers. Organic polymers can be directly applied to electrospinning as long as they can be dissolved in appropriate solvents to obtain a viscous solution. Combined with sol-gel chemistry, electrospinning is enabled to prepare composite and ceramic nanofibers [27]. At present, a variety of organic polymers and metal oxides, such as TiO₂, SnO₂, ZnO, and CeO₂ [28,29,30,31], can be prepared by electrospinning. One typical process of electrospinning includes the following three steps: (i) preparation of spinning solution from a sol-gel precursor, a polymer, and a solvent, (ii) fabrication of primary nanofibers by electrospinning under appropriate conditions, and (iii) calcination of the primary nanofiber at high temperature to remove organic components and obtain ceramic nanofibers. It was demonstrated that a cross-linked InTiO nanofiber network can be obtained using the InTiO precursor solution, composed of indium chloride tetrahydrate (InCl₃·4H₂O), titanium tetrachloride (TiCl₄), and N, N-Dimethylformamide (DMF), by electrospinning [32]. The diameter of the nanofiber was around 220 nm and its randomly oriented distribution on the substrate was caused by the bending instability of the whipping process. Figure 2 is the SEM image of the above nanofiber prepared by the electrospinning. This approach is usually in the far-field mode with the use of a hollow needle as the spinneret to achieve solid unordered nanofibers. However, new methods comprising coaxial electrospinning, near-field electrospinning, and pretreatment with the precursor solution bring about changes in the traditional technology. Consequently, nanofibers with a complex structure, such as porous, hollow, or aligned arrangement, and adjustable characteristics are prepared.

2.1. Porous Nanofibers by Pretreatment

Nanofibers prepared by traditional electrospinning methods ought to have a solid structure, but porous fibers can have more advantages due to the dramatically increased specific surface area ratio of the material. In the previous study, there were several ways to generate porous nanofibers. First of all, different material concentrations in the solution can easily affect the morphology and pore formation of nanofibers. It has been proven that varying weight percentages of zeolitic imidazolate framework-8 (ZIF-8) can effectively manipulate the formation of pores in the poly (lactic acid) (PLA)/ZIF-8 nanofibers. Increasing concentrations of ZIF-8 (0, 1, 2, 3, 5 wt% based on the weight of PLA) were added to the precursor solution to produce porous fibers with rising pore size from 0.018 μm² to 0.080 μm² as exhibited in Figure 3a–e [33]. Besides, selectively removing one of the components from the as-spun fibers is another way to the formation of the porous structure. The calcination process in the fabrication of nanofibers will result in the removal of the organic components. When adding a highly volatile solvent into the spinning solution, pores would appear on the surface of nanofibers after the annealing process. This proper treatment of selective removal of the sacrificial components in the fiber is considered to be the most straightforward approach to generating nanofibers with pores, and several experiments have clarified the contributions of the solvent with volatility such as dichloromethane (DCM) to the porous nanofibers [9]. Finally, electrospinning the nanofibers directly into a cryogenic liquid causes polymer solvent phase separation, providing another route for porous structure formation. Polystyrene (PS), Poly (acrylonitrile) (PAN), and Poly (vinylidene fluoride) (PVDF) nanofibers with pores made by rapidly cooling the fibers before complete solidification can be seen in Figure 4f–h [34].

2.2. Hollow Nanofibers by Coaxial Spinneret

The special core-shell structure of electrospun nanofibers can bring high conductivity of several orders magnitude higher than those of common solid nanofibers [35]. Hence, a novel technique of coaxial electrospinning has attracted more attention for fabricating continuous core-shell composite nanofibers or further hollow nanofibers. As shown in Figure 4, the key component of the coaxial electrospinning system is the coaxial needle, with one smaller capillary inside and another larger one in coaxial geometry [36]. These two capillaries separately drive different fluids with tunable flow rates. By pumping both inner and outer precursor solution in the two needles, a compound Taylor cone consisting of the core fluid wrapped by the shell fluid would emerge at the end of the tip to generate continuous core-sheath nanofibers on the receiving plate. In this process, adjustable factors, including the physical properties of both inner and outer materials, as well as electrospinning parameters, altogether play important roles in successfully preparing nanofibers of coaxial morphology. Specifically, the immiscibility of the two electrospun solutions ensures the detached structure of the core and shell fiber instead of mixing or inversion. Sufficient viscosity is another solution factor influencing the production of long and continuous fibers. Moreover, two optimal flow rates of the coaxial fluids should be obtained to keep the inner fluid fully wrapped by the outer one.

Different kinds of materials, such as metal salts, eupolymers, and oligomers, have been made in pairs to realize the core-shell structure nanofibers [37,38,39,40]. Figure 5a shows a regular polymer nanofiber blend of poly (vinylidene fluoride) (PVDF) and polyaniline (PANI), and the production of the core-shell nanofibers is also observed by transmission electron microscopy in Figure 5b. Experiment results have proven that a core-shell electrospun fibrous mat had much better conductivity of 2.75 × 10⁻⁹ S cm⁻¹ than 4.08 × 10⁻¹⁴ S cm⁻¹ conductivity for a simple electrospun fibrous mat [41]. Notably, coaxial electrospinning is of great help to produce nanofibers by unspinnable liquids in that these liquids serve as the inner fluid while the outer fluid guides it to spin. For instance, ultrafine Polydimethylsiloxane (PDMS) fibers were prepared by the coaxial electrospinning method using a PDMS and polyvinylpyrrolidone (PVP) as core and sheath materials with the final removal of PVP [42] (Figure 5c–f). Based on core-shell nanofibers, hollow nanofibers can be achieved by the removal of the core materials. In a classic instance, PVP hollow nanofibers derived from PVP/mineral oil core-shell fibers were obtained by being immersed in hexane for 24 h to remove the mineral oil and the hollow structure of these electrospun fibers could be seen in Figure 5g,h [43]. Besides, more complex structures could be constructed in this way with the aid of more spin needles. In one study, multichannel tubes were created by three spinnerets in the layout of an equilateral triangle [44] (Figure 6). A mixture of PVP and Ti(OiPr)₄ was added to the outer spinneret while the paraffin oil was fed into the inner three needles. The specific three-channel structure of hollow nanofibers was supposed to form after removing the core oil of the core-shell nanofibers by calcination.

2.3. Aligned Nanofibers by Special Collectors

Compared with the original unordered nanofibers, highly aligned nanofibers elicit considerable fascination among researchers on account of the improved performance resulting from their directionality. For example, one recent study indicated that the photoelectric devices made by aligned nanofibers tended to exhibit higher sensitivity than devices fabricated by the common messy nanofibers [45]. Among diverse approaches to realizing the aligned structure of fibers, near-field electrospinning stands out due to the ability to precisely control the positions of deposited nanofibers. Conventional electrospinning is usually conducted in the far-field mode with the distance between the needle and the collector larger than 5 cm, while near-field electrospinning works for a corresponding distance of 500 μm–5 cm, causing the thorough change of the stage at which the fibers will be deposited on the collector. A typical scheme of near-field electrospinning with a movable collector can be seen in Figure 7a [46]. With the decrease in spinning distance, the straight jet was put onto the collector in the absence of whipping instability. On this condition, the electric field between the spinneret and the collector is more concentrated, and the applied voltage and the accordingly flow rate could be reduced. In addition, near-field electrospinning allows for the desired fiber distribution by adjusting the average motion speed of the needle (S_n) and the relative movement speed of the collector (S_c), and this was demonstrated by using an electrospun solution blend of PEO, water, and ethanol [47]. However, when the S_c was smaller than S_n, the oscillation motion of the jet was inevitable, and wild coil fibers were deposited on the collector (Figure 7b). With the increase in the moving speed of the collector to 0.36 m/s, one aligned straight fiber of 0.49 μm diameter was achieved because of the vanishment of the unstable oscillation (Figure 7c). By carefully matching the motion speed of the jet and collector, a specified pattern of the nanofibers could be obtained (Figure 7d,e) [48,49].

Apart from near-field electrospinning, change of the collectors is also an efficient approach to producing aligned nanofibers. Researchers came up with an intriguing method utilizing a Cu wire-framed drum acting as a collector, different from the conventional electrospinning setup [50]. As shown in Figure 8a,b, the drum had two circular nonconducting disks with a hole cut in the center. Each disk had many saw-cut notches and two disks were mounted on the end of a pipe separately. Using this kind of peculiarly designed collector, aligned poly (methyl methacrylate) (PMMA) fibers were fabricated, and the fiber diameter ranged from 2.5 to 3.5 μm [51] (Figure 8c,d). Moreover, adopting collectors consisting of a pair of electrodes with an insulating gap (Figure 8e) was the most common method to obtain aligned nanofibers by exploiting the electrostatic forces associated with the electric field [52]. The separate collectors usually had applied voltages applied to let the electrostatic repulsion forces act in opposite directions and force the jet to align across the gap. Figure 8f shows a representative SEM image of uniaxially aligned polycaprolactone (PCL) nanofibers produced using this technique.

2.4. Parameters Affecting Electrospinning

Electrospinning gives the corresponding nanofibers many advantages over conventional film-based planar materials, such as better mechanical characteristics, more flexibility, better uniformity, enhanced electrical properties, more porosity, larger surface area, and ease of synthesis and processing. Nevertheless, its large-scale commercial production still has a long way to go. This sophisticated method involves many technological procedures and appropriate control over the whole electrospinning process is necessary to promote efficiency and fiber performance. The features of electrospun nanofibers are dependent on several parameters falling into three main categories: inner properties of the solution, conditions of the electrospinning process, and surrounding environmental factors.

Firstly, solution variables generally contain the molecular weight, viscosity, polarity, concentration, surface tension, electrical conductivity, and elasticity of the solvent. Among them, concentration is a typical factor affecting the morphology and characteristics of the electrospun nanofibers. To form the continuous fibers, a critical concentration should be reached for the occurrence of the electrospinning process. At low concentration, the diameter of nanofibers is smaller on average [53] and what is even worse is the formation of droplets and beaded nanofibers caused by insufficient chain entanglement to stabilize the surface tension of the jet. However, a much higher concentration could also harm the construction of fibers due to the relatively higher viscosity of the solution [54]. Consequently, proper concentration is the key to the formation of uniform and smooth nanofibrous materials, letting the electric field fully override the surface tension. The viscosity and surface tension of the solution impact the electrospinning process and fiber fabrication in the same manner as the concentration: In the solution with too high or low viscosity and surface tension, the formation of fibers will be limited. Thus, suitable parameters are crucial for the production of high-quality fibers. Particularly, the electrospinning process would be applied at lower applied voltages when lower surface tension is achieved [55]. Furthermore, the molecular weight of the targeted material in the electrospun solution plays a critical role in the formation of good nanofibers. With a lower molecular weight, a relatively increased diameter of the fibers appears [10]. Opposite to the anti-dependence of the molecular weight and fiber diameter, increased electrical conductivity of the solution would lead to increased fiber diameter, but a slight decrease would occur later [56].

Conditions of electrospinning process, such as applied voltage, feed rate of the solution, and the distance between the spinneret and the collector, are other influencing factors that belong to the following part to achieve tunable fiber morphology and quality after getting the electrospun solution ready. Particularly, electrospinning variables have a significant connection with the diameter of the fibers. For example, a smaller size of fibers will be collected on the plate with a small inner diameter of the spinneret as a result of the increased surface tension of the charged jet [57]. Then, the acceleration of the jet will be weakened for the need for more energy to overcome the interfacial tension, and more time for the jet to be stretched into a smaller diameter is needed. Besides, a growing flow rate and smaller distance between the needle and the collector will bring increased fiber diameter despite the extreme circumstances [58].

Environmental factors, including ambient humidity and temperature, influence the quality of the jet and its deposited fibers at any time. Studies have reported that high humidity will cause shorter solidification time and in such cases prolong the produced jet, whereas low humidity will lead to too thorough evaporation of the solvents, and the formation of the continuous fibers would be impeded [59]. Environmental temperature is another parameter that impacts the diameter of the nanofibers by controlling the viscosity of the solution [60]. Problematic beads or flat fibers will arise at low or high temperature. Summarized parameters that influence the nanofibers’ characteristics in the electrospinning process are presented in

Table 1. Parameters influencing the nanofibers’ characteristics in the electrospinning process.

Category	Detailed Factors	Changes in Nanofibers
Solution Variables	Target material concentration	Fiber diameter increases with the rising concentration
	Electrical conductivity of the solution	Fiber diameter increases with the rising conductivity.
	Viscosity and surface tension of the solution	Too high or too low factors lead to poor fiber morphology.
	The molecular weight of the target material	Fiber diameter decreases with increasing molecular weight.
Conditions of Electrospinning Process	Spinneret inner diameter	Fiber size is smaller with a smaller spinneret inner diameter.
	Flow rate	High flow rate causes increased fiber diameter.
	Distance between the needle and the collector	Smaller distance results in larger fiber diameter.
Environmental Factors	Ambient humidity	High humidity prolongs the produced jet.
	surrounding temperature	Temperature controls the viscosity of the solution and the fiber size.

.

Based on the discussion above, it is easy for us to conclude that current electrospun fibers are to a great extent affected by complicated parameters throughout the whole production process. Hence, the final products cannot be precisely controlled for targeted characteristics. Significantly, on account of the above properties of being highly susceptible to environmental influences, minor changes could affect the final electrospun fibers and they are still far from industrial production.

3. Electrospun Nanofibers in Transistors: Computation and Multifunctional Electronics

Compared with other low-dimensional materials, one-dimensional (1D) nanostructures have emerged as competing candidates for the active materials for field-effect transistors with their various advantages of large specific surface area, better uniformity, and favorable carrier mobility. In this way, the directional charge transport channel confined by the nanofibers avoids electron scattering and enhances the efficiency of charge transmission. The creative application of electrospun fibers with the above benefits to the active layers of the field-effect transistors (FETs) will bring about new changes for these conventional electronic devices. Commonly, the study of electrospun nanofibers for transistors can be separated into two categories, namely metal oxide semiconductor transistors and polymer semiconductor transistors. Representative materials are zinc oxide (ZnO), indium oxide (In₂O₃), nickel oxide (NiO), stannic oxide (SnO₂), copper oxide (CuO), and Polyethylene oxide (PEO), polyaniline (PANI), poly(3,3‴-didodecylquarterthiophene) (PQT-12), and poly (3-hexylthiophene-2,5-diyl) (P3HT), respectively.

3.1. Nanofiber Materials and Transistor Performance

There is no denying that the conductive polymer is the earliest material introduced to the channel layer for the classic transistor in the early stages of the development of the electrospinning technique. As early as 2003, camphorsulfonic-acid-doped PEO prepared by electrospinning was creatively used as the semiconductor layer instead of the common use for the substrate [61]. Corresponding FET was demonstrated to have the typical I–V characteristics and its mobility was measured to be 1.43 × 10⁻⁴ cm² V⁻¹ s⁻¹. Since then, more and more organic polymers were brought to this new field. P3HT is one of the most widely used polymers in the organic thin-film transistor (OTFT) and displays new vigor when transformed into nanofiber by electrospinning. The mobility of transistors with a P3HT active layer has been increased from 4 × 10⁻⁴ to 2 in recent years, experiencing several orders of magnitude of improvement [62,63,64]. The main reason for the unsatisfactory mobility of the polymer organic field-effect transistor (OFET) is the molecular self-organization and structural defects that provide trapping or scattering sites. Later, a high-performance polymer, OFET, based on PANI nanofibers was put forward with super-high charge mobility of 11.6 cm² V⁻¹ s⁻¹, for the reason that the quasi-1D nanostructure of electrospun fibers promoted the charge transfer by reducing grain boundary and doped gold nanoparticles provide “conducting bridges” between the semiconducting domains of PANI [65]. Moreover, combined with its portable nature, polymer FET exhibited high bending stability and flexibility. Aligned organic nanofibers composed of PEO, which assisted in the processing and semiconducting polymer PQT-12 of semiconducting properties, were spread on the stretchable poly(dimethylsiloxane) (PDMS)/poly(ethylene terephthalate) (PET) composite substrate to fabricate flexible FET, shown in Figure 9a, and these transistors showed remarkably higher mechanical stability than the ones deposited on the PDMS-only or film-type PET substrates [66]. Under strict bending conditions with a bending radius of 0.75 mm and repeated tests over 1000 cycles, its on-current retention could reach 82.3 (±6.7)%, while retention could hold to be 32.6 (±11.4)% and 14.6 (±20.0)% for transistors on PET and PDMS substrates, respectively (Figure 9b,c).

Besides the application of organic FET, metal oxide semiconducting nanofibers produced by electrospinning technology open up new possibilities for the fabrication of the field-effect transistor. In 2006, p-type CuO nanofiber was successfully made to further contribute to the transistor with a conductivity of 3 × 10⁻³ S/cm [67]. Two kinds of hybrid metal oxide materials (InZnO and HfInZnO) were adopted separately to achieve transistors with superior device performance, and their structures are shown in Figure 9e [68]. These transistors demonstrated negligible hysteresis, a decent electron mobility of 5.49 cm² V⁻¹ s⁻¹, and a good on/off current ratio of 10⁷.

3.2. Advanced Electrospinning Methods for Transistor Production

The study on the electrospun nanofiber transistors mainly focused on upgrading process technology to improve device characteristics, and impressive results have been achieved in the past few decades. The typical fabrication procedure of transistors with electrospun nanofiber semiconducting layer consists of solution preparation, fibers electrospinning, annealing and other treatments, electrode deposition, and final FET formation. Any details of these five steps directly impact the performance of the transistors, and thus great efforts have been made to adjust the process parameters to find the optimal ones.

For the solution preparation part, either the concertation or the ratio of the targeted element would change the properties of the fabricated fibers according to transistor performance. The transfer characteristics of poly(stearyl acrylates)(PSA)/poly(3-hexylthiophene)(P3HT) nanofiber-based FET for various blending ratios (Figure 10) were studied and most excellent electrical performance was obtained with 1:0.2 P3HT/PSA mixture electrospun nanofibers of 3.21 × 10⁻² cm² V⁻¹ s⁻¹ mobility, which was two orders of magnitude more than that of pure P3HT nanofibers (1.92 × 10⁻⁴ cm² V⁻¹ s⁻¹) due to the P3HT edge-on orientation, and a more compact structure of P3HT nanofibers were promoted by crystalline PSA to accelerate the transportation of the hole [69]. High-performance zinc tin oxide (ZnSnO) electrospun nanofibers with a variety of composition ratios were made [13], and optimized Zn_0.3Sn_0.7O nanofiber FET based on SiO₂ dielectric showed a high on/off current ratio (I_on/I_off) of 2 × 10⁷ and field-effect mobility (μ_FE) of 0.17 cm² V⁻¹ s⁻¹ (Figure 11a–c). In addition, P3HT nanofiber-based organic field-effect transistors produced by the coaxial electrospinning method were prepared, in which the P3HT served as the core polymer and poly(methyl methacrylate) (PMMA) served as the shell polymer, followed by the extraction of PMMA(Figure 11g) [70]. Three types of solvent (chlorobenzene (CB), chloroform (CF), and 1,2,4-trichlorobenzene (TCB)) were used to pick the greatest one and CF was found to achieve nanofibers with the smallest fiber diameter, maximum FET mobility of 3.57 × 10⁻¹ cm² V⁻¹ s⁻¹, and strongest flexibility. The reason why CF stood out among those three solvents is that stretching of the P3HT chain stretching along the long axis of the nanofiber during the electrospinning process was greatly promoted by a higher solubility in the chloroform system, which in turn enabled the formation of highly oriented crystalline grains after thermal annealing and resulted in an improvement of both the ductility of P3HT nanofibers and the performance of P3HT electrospun nanofiber-based transistors. As mentioned in the previous section, morphology and characteristics can be managed by changing the conditions of electrospinning. Aligned nanofibers were received by two parallel conducting Si strips as electrospinning collectors (Figure 11d), and corresponding transistors (Figure 11e) had greater mobility than disordered nanofiber-based FET, as shown in Figure 11f [71]. Besides, it was demonstrated that nanofiber networks with various channel coverage produced by altering the nanofiber collecting time during the electrospinning would affect the device performance. A higher density of nanofibers generally led to more conduct points for the carrier movement and thus larger on current (I_on), higher field-effect mobility, and greater on/off current ratio. However, when the NiO network coverage hit about 85%, the device’s current modulation capability deteriorated dramatically, with a tiny I_on/I_off ratio and a low μ_FE value. This could be explained by the transition from a two-dimensional network plane to a three-dimensional spatial structure when the density of NiO nanofiber increases. The irregular channel thickness and morphologies were caused by the overlaying of a large number of individual fibers. Meanwhile, the higher overlapped fibers were unable to directly contact the bottom dielectric, but they were able to raise bulk resistance, resulting in poor carrier modulation under the bias voltage.

To address this issue, controllable doping to obtain binary-cation MOS was adopted to parallelly improve metal oxide properties for a variety of applications and fibers with a mixture of metal elements, such as Sn-doped ZnO (ZnSnO), Gd-doped In₂O₃ (InGdO), Mg-doped In₂O₃, Sr-doped In₂O₃, and Al-doped In₂O₃ [72,73,74], have appeared over the past period. Recently, Al doping was used to improve the morphology of ternary-cation InAlZnO (IAZO), nanofiber-based field-effect transistors [75]. As we all know, InZnO nanofibers are representative of the doped metal oxide nanofibers for transistor application, but InZnO nanofibers exhibit a coarse surface and are discrete as disconnected segments. By adding the Al elements to the InZnO fibers, morphology can be greatly improved and the IAZO nanofibers exhibit a smooth surface with invisible broken fibers, as well as an improved on/off current ratio of 10⁷ and high electron mobility of ~10 cm² V⁻¹ s⁻¹. The comparison among the transfer characteristics of the In₂O₃, InAlZnO, and InZnO FET is shown in Figure 11h.

Annealing and other treatments represent the primary procedure for easy and direct control, and different annealing temperature, time, and atmosphere have been confirmed to affect the mobility and current ratio of the FET by electrospinning. Water-induced (WI) In₂O₃ nanofibers with varied annealing dependence were prepared and experiments revealed that annealing temperature (280 °C, 310 °C, 350 °C, 380 °C) and hold time (0.5 h, 1 h, 1.5 h, etc.) would influence the traps in the bulk channel and/or at the interface between the channel and the dielectric, consequently changing the FETs performance in parallel [76]. Among changing annealing conditions, FET based on WI In₂O₃ nanofibers annealed at 380 °C for 1 h exhibited field-effect mobility (µ_FE) of 1.53 cm² V⁻¹ s⁻¹, and an ON/OFF current ratio (I_ON/I_OFF) of 8 × 10⁴ (Figure 11i–l). Similar research on ZnO and SnO₂ transistors has proven the impact of annealing temperature and time on the grain size and continuity of the nanofibers and the consequent electrical performance of the transistors [77,78]. After fibers are fabricated by electrospinning on the collector, plasma treatment can selectively be applied to enhance the FET performance and stability. Using various ratios of Ar/O₂ gas blends, the effects of plasma surface treatment on electrospun In-Ga-ZnO (IGZO) nanofibers and electrospun FETs were examined systematically [79]. As can be seen in Figure 12, because of the ability of Ar plasma to etch the nanofibers, Ar/O₂ mixed plasma treated IGZO nanofibers showed a further reduction in diameter with an increase in the Ar ratio. In addition, when the O₂ ratio in the Ar/O₂ mixed-gas flow grew, the drain current increased. However, as the Ar ratio in the Ar/O₂ mixed-gas flow increased, the drain current of the nanofiber-based FET was reduced. As a result, the highest performance of FET, which exhibited mobility of 5.32 cm² V⁻¹ s⁻¹ and an on/off current ratio (I_ON/I_OFF) of 4.35 × 10⁶, was obtained with the maximin O₂ concentration in the Ar/O₂ mixed-gas flow without the Ar gas. In summary, electrospun, nanofiber-based field-effect transistors have gained continuous improvement in performance through the persistent exploration of process factors, mobility, and the on/off current ratio for the current nanofiber transistors to become more competitive in the field of electronic devices based on 1D-nanostructure materials.

3.3. Typical Applications of Nanofiber-Based Transistors

Even though development is mainly focused towards promoting the process and technique for field-effect transistors with the nanofiber active layer to obtain more stable and reliable devices, several applications have emerged to demonstrate the value of introducing electrospinning to the fabrication of FETs. Preliminary gas and humidity sensing, light-emitting devices, and basic logic circuit successively show up, caused by the special feature of better homogeneity, more porosity, stronger mechanical strength, and larger surface areas. In 2011, the single PANI nanofiber field-effect transistor produced by near field electrospinning (Figure 13a,b) first appeared and this gas sensing transistor offered tunable sensitivity and excellent signal-to-noise ratio with applied field effect [80]. Compared with traditional polymer film and CNT FET sensors, this sensor had higher sensitivity (7% versus 1 ppm of NH₃), greater reliability, and requires much less energy. The main gas sensing mechanism of the PANI nanofiber was the deprotonation of PANI by NH₃, which caused a resistance rise. The response of this FET to the NH₃ gas with varying concentrations can be seen in Figure 13c. Moreover, a field-effect transistor could be used to boost the sensitivity of a humidity sensor in an appropriate gate voltage region. For example, a kind of Ga-doped ZnO nanostructure transistor (Figure 13d) was fabricated and its voltage region was in the positive range because of its n-type semiconducting nature [81]. When the environmental relative humidity (RH) changed between 15% and 85%, the current response of the FET (I_ds) changed accordingly, and cyclic behavior and consistency in its functionality were explicit in this way (Figure 13e). Low cost and flexible one-dimensional light-emitting transistors represent another potential application for nanofiber-based devices. As early as 2010, a single nanofiber transistor made by the light-emitting polymer poly(2-methoxy-5-(2-ethylhexoxy)-1,4-phenylenevinylene (MEH-PPV) (Figure 13f) was prepared and it exhibited controllable photoluminescence intensity through applied bias [82]. As shown in Figure 13g–j, the intensity of photoluminescence (PL) from single fibers within FETs can be electrically manipulated in a luminist mode, reducing by a factor of 1.5 when the gate voltage (V_GS) is varied from 0 to −40 V. Textile-based OFETs containing PQT-12:PEO nanofibers also showed superior photosensitivity under wavelengths of 470 nm compared to those under 670 nm, indicating that they might be used as photosensors with targeted wavelength detection [66] (Figure 13k,l). Further photo-sensing study for this FET was performed using a 10 × 10 photodetector array and a star-shaped pattern achieved using a laser source can be exactly recognized with this high resolution, fabricated, textile-based device (Figure 13m,n).

Besides the form capability of sensing, electrospun nanofiber-based FETs can act as chemical sensors. For instance, at an unprecedentedly low concentration of 1 fM of the Bisphenol A(BPA), a field-effect transistor sensor using aptamer-modified multichannel carbon nanofibers (MCNFs) to detect the chemical showed remarkable sensitivity and specificity for BPA [83] (Figure 14a,b). It was also proven that an MCNFs FET sensor responds directly in seconds when exposed to BPA. Furthermore, typical and simple logic circuits are realized by electrospun nanofiber-based transistors, and their fundamental application of building blocks for hardware circuit implementation to execute complicated computation offers these kinds of transistors great opportunities to truly reform the future integrated circuits. For instance, FETs based on ZnSnO nanofiber mat were connected in series with a 2 MΩ external resistor to create a resistor-loaded inverter [13]. Voltage transfer characteristic (VTC) and a sharp inverter response with a maximum gain (defined as −∂V_OUT/∂V_IN) of 11 at V_DD = 4 V are shown in Figure 14c,d. This transistor had a good noise tolerance ability and wonderful dynamic behavior of the inverter under an ac square wave signal. Subsequently, transistors with different electrospun nanofibers, such as InGdO, Sr-doped In₂O₃, Al, Ga, Cr doped In₂O₃, and IGZO, were put into the resistor-loaded inverter circuit with different operating voltage and gain.

In conclusion, the combination of 1D materials and transistors offers more chances for the cost-efficient electrospinning method to expand its application field. The number of varied fiber materials, including the majority of organic materials and metal-oxide semiconductors, is not inferior to the kinds of normal film channel materials, and special structures, such as hollow and core-shell fibers, for high-performance FETs have been successfully produced in this simple way. Nanofiber-based transistors coupling electrical, sensing and optical functionalities will play a significant role in innovative electronic devices in the near future and a summary of information on the electrical properties and applications of FETs with different kinds of nanofibers is shown in

Table 2. Electrical Properties and Applications of FETs with Different Kinds of Nanofibers.

Year	Nanofiber Material	Mobility (μ, cm2 V−1 s−1)	Ion/Ioff	Characteristics	Applications	References
2003	Camphorsulfonic-acid-doped PEO	1.43 × 10−4			—	[61]
2005	RRP3HT	0.03	103	—	—	[84]
2005	P3HT	4 × 10−4	7			[64]
2006	CuO(p-type)	—	—	Conductivity~3 × 10−3 S/cm	—	[67]
2009	P3HT	0.017	102			[62]
2010	P3HT	2	105	Flexible, high frequency operation		[63]
2010	MEH-PPV	10−3	780	Single nanofiber	Photoluminescence/Light-emitting	[82]
2011	P4TDPP/PMMA	0.305	1.3 × 105	Crystalline-induced procedure		[85]
2011	PANi(p-type)	2 × 10−3		Single nanofiber	Gas sensor with tunable sensitivity	[80]
2011	Au-doped PANi	11.6				[65]
2012	P(NDI2OD-T2)/PEO	0.1		Single fiber, Stable		[86]
2012	GZO		104		Gas sensing including humidity	[81]
2013	ZnO	4.6 × 10−3			Different calcined atmosphere	[87]
2013	P3HT/PSA	3.21 × 10−2	104	Excellent air resistive ability		[69]
2014	LiNiO	3.3 × 10−4		Transparent		[88]
2014	P3DT:P3HT	5.0 × 10−3		High bending stability		[89]
2016	ZnO	0.018	103	Aligned	Photoconductivity	[90]
2016	PQT-12/PEO	1.18 × 10−3	1.86 × 103	Flexible, highly stable under extreme bending conditions	Real-time photosensing	[66]
2016	MCNFs			High sensitivity, specificity and stability	BPA dection	[83]
2016	AZO	3.3 × 10−3	103			[91]
2017	PEDOT:PSS(p-type)	5.5	730			[92]
2017	CuO(p-type)	3.5	3 × 105	Low operating voltage, fast switching speed	Driving unit for LED	[93]
2017	PS/DPP	0.22	105	Floating-gate transistor memory		[94]
2018	NiO(p-type)	2.8	104	Excellent dynamic behavior	Light-emitting	[71]
2018	ZnO	0.02	106	Low-operating voltages		[78]
2018	ZnSnO	0.17	2 × 107	Low operating voltage	Resistor-loaded inverter	[13]
2018	InGdO	17.4	108	Low operating voltage	Resistor-loaded inverter	[95]
2018	Mg-doped In2O3	5.3	1.7 × 105			[73]
2019	Sr-doped In2O3	4.8	108	Enhanced stability	Resistor-loaded inverters	[96]
2019	PEDOT-PSS(p-type)	0.5			Non-volatile charge storage	[97]
2019	InZnO	5.49	107			[68]
	HfInZnO	0.06	105
2019	TiO2			Arranged nanofibers, Mixed anatase-rutile TiO2		[98]
2019	P3HT	0.357	4.57 × 103	Hollow (Coaxial electrospinning)		[70]
2020	In2O3	1.27	5 × 107	Excellent reliability and reproducibility		[99]
2020	In2O3	1.53	8 × 104	Nontoxic and low-temperature preparation		[76]
2020	ZnSnO	21.58	108	Different annealing atmosphere and temperature	Resistor-loaded inverters	[74]
2020	IGO	35.1	107	Dual-gate structure		[100]
2020	InYbO	6.18	107	Outstanding selectivity towards DMF	Gas sensing	[101]
2020	Al, Ga, Cr doped In2O3	2.0	108	Low operating voltage	Resistor loaded inverter	[72]
2020	Al functionalized β-Bi2O3	64.4	106	Excellent sensitivity, stability and reproducibility	Biosensor for detection of serotonin	[102]
2020	IGZO	5.32	4.35 × 106	Good electrical properties and reliability	Resistor-loaded inverters	[79]
2021	IGZO	2.45	3.8 × 102	Double gate, high pH sensitivities	Chemical sensors	[103]
2021	In2O3	1.51	1.5 × 107	Ar/O2 plasma treatment		[104]
2021	IAZO	30	5.6 × 106			[75]
2021	In2O3	27.7	107	Great stability	Resistor-loaded inverter	[105]
2021	SnO2	2.3	106	Adjustable grain size		[77]
2021	PVA (with evaporated CuPc)	8.77 × 10−3	1.37 × 103	Short response and recovery time for detecting NO2	Gas monitoring	[106]

.

4. Applications of Electrospinning in Memory: Storage Electronics

In recent years, the usage of the semiconducting device as a memory device for storage purposes has seen huge growth. On the other hand, electrospinning is one of the most cutting-edge methods for producing fine fibers from a wide range of materials and can synthesize ultrafine fibers ranging from micrometers to nanometers with simplicity in preparation and the potential for large-scale production. Several decades of development have brought benefits to the morphology and performance of the electrospun nanofibers. Combing the memory device and electrospun nanofibers will undoubtedly be highly prospective for high-performance memory electronics at a low cost. As one of the key tasks of modern computers, a memory device’s basic objective is to provide an approach for storing and accessing binary digital data sequences of ‘1’s and ‘0’s. According to this premise, an electrospun nanofiber-based memory device stores data based on different electrical conductivity states (ON and OFF states) in response to an applied electric field, which entirely differs from the traditional silicon-based memory devices in which data are stored based on the amount of charge stored in the memory cells. However, there have not been many reports of electrospun materials being applied for memory devices and the number of published papers is limited.

Many studies for memory materials or devices involving the electrospun nanofiber mats are still in their infancy. Non-woven bead-free highly crystalline T-Nb₂O₅ nanofibers with a length of 100 μm and a diameter of 80–200 nm were produced by the sol-gel aided electrospinning method [107] (Figure 15a). Incorporated with nitrogen-doped diamond-like carbon (DLC) and Pt electrode, these Nb₂O₅ fibers exhibited memristive properties. Such as reproducible resistive switching with an on/off resistance ratio as high as 2 × 10⁴, as shown in Figure 15b. I–V characteristics for the embryonic memory device had a bipolar nature which began with a threshold voltage of 0.8–1.7 V and progressed to continuous conductivity development. Besides, a vertical sandwich Pt/erbium (Er)-doped sodium-potassium niobite (Er:NKN) fibers/Au structure was observed to have memristor-type current-voltage (I–V) characteristics with the changing resistance ratio of 10² from low-resistance to high-resistance state [108] (Figure 15c). The electrospun Er:NKN nanofibers (Figure 15d) provided the channel for the electrochemical migration of oxygen vacancies which started at the n-type Er:NKN oxide/high work function Pt cathode interface, and in turn brought about electric field-induced resistance switching characteristic for the devices. In addition, negative differential resistance was the key feature for the perovskite (CH₃NH₃PbI₃) decorated electrospun TiO₂ nanofibers (Figure 15e,f) due to the quantum-hybridization mechanism in Figure 15g and the “Write–read–erase–read” sequence of the electrical response of the composites was probed to determine the memory applications [109]. The voltage scan range of −2 V and +2 V was applied to “write” the high-conducting state and then to “erase” it back to a low-conducting state, respectively. Figure 15h depicts the current value that corresponds to the applied voltage value based on whether the following “write” or “erase” operation is performed, demonstrating the simple memory behavior. Furthermore, novel conductive polymer aligned nanofibers made of electrospun polycaprolactone (PCL) with a complex made of multiwalled carbon nanotubes (MCNT) and fullerene C₆₀ (Figure 15i) displayed electrical switching behavior due to charge buildup of fullerene C₆₀ based on electrical stimulation [110]. These conveniently prepared nanocomposites on interdigitated coplanar electrodes could be applied to write-once-read-many memory devices (WORM), which is one type of non-volatile memory device. With further electrical and thermal cycling, the devices could be thermally reset, reprogrammed, and wiped. Figure 15j indicates that the devices could be cycled at least six times using sequential thermal and electrical stimuli, with the exact annealing temperature condition of 60 °C.

Interest in one of the types of electronic memory devices, resistive memory devices, has increased markedly during the last few years. The most striking feature of this kind of device is the hysteresis loop in the current–voltage plane, which cannot be duplicated by other three-layer circuit elements, and the device resistance can be reversibly changed by an applied voltage while being maintained in its absence. It is well-known fact that the abrupt transition between two resistive switching states is caused by the formation and breaking of the conductive filament of the active layer. Initially, the device is in the high resistance state (HRS), with very little current flowing through it. The top electrode is biased while the bottom electrode is grounded during the electrical measurements. In this case, metal atoms from the top electrode are ionized and drift towards the bottom electrode with a positive voltage given to the top electrode. When the magnitude of the applied voltage (V_SET) reaches a certain level of value, a conductive filament made up of ions originating from the top electrode forms a bridge between the top and bottom electrodes, allowing current to flow freely through the device, and the resistance of the device drops, leading to the low resistance state (LRS). The negative bias mode exhibits a similar pattern of behavior concerning electrochemical breakage of the conductive filament and the corresponding state-changing voltage is termed as the RESET voltage [111]. In 2020, Aditya A. Patil et al. innovatively came up with a memory device based on 1D functionalized multi-walled carbon nanotubes (f-MWCNTs)-TiO₂ composite electrospun nanofibers [112]. The device structure can be seen in Figure 16a. This three-layer random-access memory (RAM) showed obvious bipolar resistive switching (RS) behavior with clear evidence of the hysteresis loop (Figure 16b). The composite nanofiber-based device had excellent endurance and memory retention qualities, as demonstrated in Figure 16c,d respectively, which suggested that this memory device could easily toggle between LRS and HRS 500 times and could also keep these two states until 104 s without any significant degradation. Among other kinds of organic fibers, cellulose fibers with biodegradability, biocompatibility, and high tensile strength provided good charge transport pathways as well as support for the formation of the conductive filament of the bipolar resistive switching operation [113]. The “Sandwich” structure of Ag/cellulose fibers/Al memory device (Figure 16e) was then fabricated using the electrospinning technique and worked on low resistive switching voltage (±1 V) instead of an additional electroforming step (Figure 16f). It also had excellent endurance for 6 × 10³ resistive switching cycles with a 10× memory window and great data retention capability for the 3 × 10² s without any observable degradation in the resistance states (Figure 16g,h). Moreover, cellulose–gelatine micro composite fibers (CGMFs) were fully used to realize the organic resistive switching devices and they are different from the previous ones in that test results proved that they had a pretty large current (~4 mA) when the RAM was in the LRS due to the formation of the conductive filament and the low resistance value of the CGMFs composite at SET voltage [114] (Figure 16i). These exploratory studies present the fact that cellulose fiber-based memristive devices represent a sensible option for the organic non-volatile memory application.

All in all, memory devices with electrospun nanofibers unlock the potential for electrospinning techniques with low cost and universality. Memory application with a nanofiber-based middle layer provides a directional transmission channel for the carriers and would offer an excellent chance to develop a more stable and larger capacity memory. Moreover, electrospun nanofibers can also be extended to the applications of electrodes of random-access memory (RRAM) [115] or the dielectric layer of field-effect transistors toward nonvolatile memory [116], indicating the versatility for the electrospun one-dimensional fibers.

5. Applications of Electrospinning in Sensors: Various Sensing Electronics

As the most popular two-terminal devices, sensors have been widely used in the field of industrial production, smart homes, environmental protection, health monitoring, and so on. Lately, an emerging development trajectory in sensors focuses on their flexibility and stretchability, and the electrospinning technique precisely has potential in this field as a result of its large surface-to-volume ratio, high tortuosity, high permeability, high porosity, and portable nature, especially providing strain and pressure sensors. Compared with other metal oxide materials, conductive and piezoelectric polymers prevail because they can exploit deformations induced by small forces. For instance, Someya and his group fabricated a nanomesh skin-strain gauge with breathable, ultrathin, and soft characteristics to continuously monitor skin strain without disturbing [117], and the electrospinning method was used for this nanomesh to produce the compact polyurethane (PU) nanofibers with extremely thin PDMS to reinforce the whole structure. The sensor was incredibly thin (430 nm) and ultralightweight (0.012 mg/cm²), which enable the sensor to stick to the skin and detect minute skin strains (Figure 17a). Distinguishable sweat pores cam be seen from the microscopic images of nanomeshes on the surface of artificial skin in Figure 17b and the as-fabricated sensor showed high mechanical durability (resistance degradation lower than 3% after 5000 repeating stretching and releasing cycles) and excellent sustainability (smaller than 5.3% electrical drift error under 40% strain for 12 h). With the unique properties of monitoring signals affecting the natural movement of the skin, the nanomesh sensor was harmlessly attached to the human face and successfully measured the facial strains during speech as shown in Figure 17c. The skin strain distribution on the right side of the face appeared to be symmetrical to the left side of the face with the black markers, indicating that these nanomesh devices accurately reflected actual skin deformations, and allowing the skin to be strained and compressed freely during speech. Furthermore, this device was proven to be capable of detecting stable full-range facial movements, e.g., joy, anger, sadness, fear, and other facial micro-expressions in real-time. This work demonstrated the great potential for the electrospinning technique to produce cutting-edge electronic skins (e-skins) and limitless possibilities to create more and more practical wearable electronic devices simply by electrospinning. Additionally, through the electrospinning technique, researchers prepared an aligned thermoplastic polyurethane (TPU) fibrous mat, then induced CNTs to wrap around the TPU mat through ultrasonication [118] (Figure 17d,e). This aligned CNTs/TPU fibrous strain sensor had a vertically aligned wave-like structure with ultra-high stretchability (900%) and excellent durability (10,000 cycles at the strain of 200%) (Figure 17f). The ultra-low detection limit (0.5%) and fast response time of 70 ms were also achieved, exhibiting a favorable sensitivity. Above fine sensing, performances were mainly due to the planned wave-like and joint fiber structure. Thus, the strain sensors were promising candidates for wearable devices to monitor real-time human activities.

Except for strain sensors, pressure sensors are also critical for the accurate measurement of small normal pressures to precisely monitor the external force on the dynamic human skin. There are various types of well-developed pressure sensors, e.g., capacitive [119], resistive [120], triboelectric [121], and piezoelectric [122]. Among all these sensors, resistive ones are the most easily fabricated and cost-effective types and hundreds of thousands of studies have been made in the past few decades. For example, stretchable resistive carbon nanotubes-embedded nanofiber sensing yarns were successfully prepared by a facile electrospinning technique and proved mechanically robust enough to be woven to a wearable textile sensor capable of distinguishing numerous mechanical inputs, such as pressure, stretching, and bending [120]. Delicate sensing micro/nanostructures and efficient conductive networks endow the sensors with high-pressure sensitivity (12.3 N⁻¹), a wide sensing range (around 0.001–5 N), high durability over more than 10,000 cycles, as well as high gauge factors (114–720) over a large stretch range (0.1–220%). Besides, a piezoelectric sensor based on electrospun aligned nanofibrous arrays of poly(vinylidene-fluoride-co-trifluoroethylene) (P(VDF-TrFE)) was reported with an average diameter of 260 nm [123] (Figure 17g). With overall porosity of 65%, the three-dimensional nanofibers had excellent flexibility and were easy to bend (Figure 17i). The ultimate sensors offered exceptional piezoelectric characteristics, allowing for ultra-high sensitivity for pressure measurement, even at extremely small values (0.1 Pa).

Electrospun nanofibers with large specific surface area, good biocompatibility, and the ability to specifically detect biological molecules or chemical compounds have been widely researched for sensor applications, such as biosensing or gas sensing. For example, a new electrochemical biosensor based on functional composite nanofibers was created for the sensitive hybridization detection of the p53 tumor suppressor using methylene blue (MB) as an electrochemical indicator [124]. Electrospinning technology was applied to prepare carboxylated multi-walled carbon nanotubes (MWNTs) doped nylon 6 (PA6) composite nanofibers (MWNTs-PA6) to serve as the backbone for pyrrole (Py) electropolymerization (Figure 18a,b). Reproductivity and specificity of the fabricated sensor were tested by investigation after 10 days of storage at 4 °C and further being hybridized with the target wtp53. Results showed that this sensor remained highly sensitive and provided a stable platform as an electrochemical DNA biosensor. For the application of electrospun nanofibers for gas sensors, a conductive polymer with inherent porous nature for the thorough absorption of gas molecules is better than inorganic materials. A novel approach for highly efficient conducting polymers-based gas sensors was proposed by the introduction of polymeric sensitizers (polymeric adsorbent) within the conducting polymeric nanostructures to form 1D polymeric adsorbent/conducting polymer core-shell nanocomposites via electrospinning and solution-phase polymerization [125]. It was possible to observe stable and distinct responses for this gas sensor, as shown in Figure 18c, and high sensitivity was exhibited when an extremely low concentration of NH₃ (20 ppb) stimulated the response as high as 3.8. Figure 18d shows that the response of SPEEK/PPy and PAN/PPy nanofibers increased rapidly under exposure to NH₃ and then gradually decreased when the gas was absent. It was also obvious that the PAN/PPy nanofiber-based sensors had a recovery time of less than 25 s, which is significantly smaller than that of SPEEK/PPy nanofibers (180 s).

In summary, nanofibers prepared by the electrospinning technique always take a place in the applications of sensors for the fast and accurate detection of various signals, such as pressure, gas, chemical substances, biological materials, and so on, which mainly benefit from the universality of electrospinning based on solution preparation to produce the most variety of materials. Moreover, electrospun nanofibers with large specific surface area and porosity enable the exploration of e-skin and further research would prompt the development of electrospun nanofiber-based sensors towards higher performance and more flexibility for wearable and functional systems.

6. Artificial Synapses and Other Electronic Devices Prepared by Electrospinning

In addition to the electrospun nanofiber applications including transistors, memories, and sensors, mentioned above, electrospinning has been involved in many other fields, such as optoelectrical and biological devices. In particular, brain-inspired electronic devices have emerged as efficient approaches to reform traditional computing. With the scaling down of the critical channel length of the transistor in integrated circuits, the number of transistors per unit chip is growing exponentially following Moore’s Law, triggering many quantum effects by the small size of the FET and leading the improvement of integrated circuit performance into a bottleneck. Further, traditional von Neuman computers with separate storage and computing blocks could not effectively deal with the massive amount of information. Neuromorphic electronic devices as creative components with brand new architecture appear to be comprehensive innovations of the common computer chips, with advantages of extremely high efficiency, ultra-low power consumption, and flexible adaptive capability.

For the neural network in the human brain, neurons are the basic units for cognition and computation, and synapses that connect neurons and transmit information are also important elements for the formal function of nerve cells. Solid-state synaptic devices represent the underlying foundation for the implementation of brain-like computing systems and are as important as transistors for modern circuits. To successfully fabricate synaptic devices, careful attention should be paid to the basic functions and principles of biological synapses in neural networks. The action potential of a synapse is a process for generating impulses by a neuron, and the signal transmission between neurons utilizing synapses is the basis for information flow and data processing in the human brain. When an action potential is transmitted to an axon terminal, the presynaptic membrane will send out a pulse, and a change in the cell membrane potential on the postsynaptic membrane will follow. If the pulse on the presynaptic membrane causes positive ion channels on the postsynaptic membrane to open, such as Na+ ion channels, the cell membrane potential will rise and prompt the release of an action potential. Consequently, the potential is referred to as the excitatory postsynaptic potential (EPSP), and the corresponding rise in current is referred to as the excitatory postsynaptic current (EPSC). Conversely, while a pulse on the presynaptic membrane causes the opening of a negative ion channel, such as the Cl- ion channel, the post-synaptic cell membrane potential is referred to as the inhibitory postsynaptic potential (IPSP) and the corresponding decrease in current is referred to as the inhibitory postsynaptic current (IPSC) (Figure 19a,b). In this case, the most representative artificial electronic synapses, synaptic transistors, utilize the gate as the presynaptic pulse applied terminal while source and drain are combined to detect the postsynaptic current (I_ds) (Figure 19c). As a result, nearly all synaptic behaviors could be imitated by the simple structure of synaptic transistors for the further construction of complicated neuromorphic networks.

Due to the enormous specific surface area, 1D electrospun nanofibers for artificial synapse applications can be in full contact with the surrounding dielectric, which significantly improves the device’s operating sensitivity and reduces power consumption. However, electrospun nanofiber-based FETs have not been deeply explored and few studies have been conducted to introduce 1D-nanostructure materials into brain-like computing. Ultralow power synaptic simulation through solid-ion gated organic field-effect transistors (OFETs) with a 3D interface conducting channel was realized in 2020 [128] and this single fiber field-effect transistor was made by the core-shell structure of PEO/PQT12 nanofiber joint with liquid electrolyte (PEDOT:PSS) as a gate dielectric. Figure 20a depicts a schematic picture of a biological synapse and PQT-12 fiber-based artificial synapse with a 3D ionic electrolyte as the transistor gate. The classic EPSC behavior of synaptic transistors induced by a presynaptic spike (−1 V, 0.1 s) applied on the gate electrode can be seen in Figure 20b. Moreover, paired-pulse facilitation (PPF) was proven by applying two consecutive impulses with a time interval of 0.2 s on the gate and the later EPSC was three times more than the former EPSC (Figure 20c). By changing the read voltage and amplitude of the pulse, minimum power consumption was calculated to be around 3.9 fJ per spike (Figure 20d,e), which was attributed to the large contacting interface area between the organic polymer and the solid ionic dielectric, offering a large number of ions to attach. Usually, the function of a high-pass filter can be realized by a synapse having a low probability of vesicle release in a biological neuromorphic network, which was also found for this single nanofiber FET with a gain of EPSC increasing from 1.2 to 6.6 when the frequency rising from 1 to 10 Hz (Figure 20f), reflecting the potential for this simple-fabricated device to realize the artificial synapse and further development in artificial intelligence and neuromorphic computing system applications.

In 2021, organic electrochemical transistors with PEDOT:PSS/polyacrylamide (PAAm) nanofiber as channel layer took the stage for comprehensive neuromorphic computing and sensing platforms to monitor the biological signals in real-time (Figure 21a,b) [129]. In the whole electrical system, the added PAAm served as an ion-blocking layer to enable the device to have long-term plasticity and to avoid too much energy consumption, which would happen with only PEDOT:PSS material. Typical transient current response and long-term potentiation and depression (LTD) concerning pulse number are shown in Figure 21c,d. This FET exhibited an ultralow energy consumption of 113 fJ, as shown in Figure 21e. Originated from electrospun nanofibers, this transistor possessed tunable electrical conductance by controlling the density of fibers, and the variation for the EPSC could be three orders of magnitude with electrospun volumes ranging from 0.5 to 30 µL (Figure 21f). Further software simulation performance based on the array arrangement of FET devices offered possible chances toward smart platforms comprised of neuromorphic devices for interacting with biology.

Apart from the organic-fiber-based field-effect transistors, metal oxide semiconductor FET with electrospun nanofibers open a path for mimicking neural components, such as the ZnSnO nanofiber-based FET with ion-gel dielectric (Figure 21g,h) [130]. The amplitude of EPSC in this device rose linearly with the duration time and finally reached a saturating value of 71 nA (Figure 21i–j), and the shift in synaptic response generated by paired-pulse stimulation at the same intensity was referred to as PPF (Figure 21k). Moreover, this one-dimensional fiber-based FET was proven to mimic the high-pass dynamic filtering property when the amplitude of the EPSC increased from 40 nA to 76 nA with the changing of the spike frequency from 1 Hz to 100 Hz (Figure 21l). It was also found that EPSC could not decay back to the original current value following intense stimulation in a brief amount of time, demonstrating the long-term synaptic plasticity of the device. Low power consumption of 0.44 pJ demonstrated its potential applications for energy-efficient artificial synapses for sophisticated neuromorphic networks.

Simultaneously, electrospun nanofibers also give full play to their advantages for various electronic devices. For instance, photoelectric detectors could have sensitive photoresponse and incredible electrical performance when they are fabricated by 1D electrospun nanofibers with a large specific area and increased photosensitive area of the materials. A novel UV photodetector with a ZnO nanofiber array was realized in this way. Aligned pure n-type ZnO nanofibers and Ag-doped p-type ZnO nanofibers formed a crossed array that functioned as homojunction to change the material conductivity under optical radiation and the device was fully transparent on the quartz substrate [131] (Figure 22a,b). The performance of this photodetector was considerably improved to have a high on/off ratio of 2.5 × 10⁴ together with a responsivity of 1 × 10³ μm W⁻¹ in the p-n junction form under zero bias voltage. (Figure 22c,d). Electrospun nanofibers were able to produce a heterojunction that acted as photodetector as well and the obtained SnO₂/NiO nanofiber array (Figure 22e) was proven to have a high detectivity of 2.33 × 10¹³ Jones and responsivity up to 8.67 A W⁻¹, as exhibited in Figure 22f, outperforming most of the state-of-art SnO₂-based photodetectors [132]. Besides, electrospinning also brought new possibilities to the portable and wearable power source as conductive electrodes. A uniform, free-standing, Sn-based (Sn@CFC) electrode was fabricated by the electrospinning technique to serve in the sodium-ion batteries (SIBs) for large-scale energy storage instead of conventional lithium-ion batteries (LIBs) (Figure 22g) [133]. Ultra-small Sn nanoparticles were decorated around the three-dimensional carbon nanofibers, leading to the shortened pathway for sodium ions and electrons, mitigated volume expansion, and improved electrical conductivity. This fiber-based electrode showed high reversible areal capacity (1.68 mA h cm⁻²) and a long-life cycle of 1000 cycles with above 80% capacity retention in Figure 22h,i. Apart from electrode materials of SIBs and LIBs, electrospun fiber-based mats have gradually extended their applications to many other energy storage devices, such as supercapacitors, Li–S batteries, Li–CO₂ batteries, Li–O₂ batteries, and so on [134,135,136,137,138,139]. These studies demonstrated the fact that nanomaterials produced by electrospinning have become a potential option for high-efficiency photodetectors and electrodes, and inspired a new way to design other optoelectronic devices with various 1D semiconductor materials.

7. Conclusions and Future Insights

The electrospinning technique has experienced the process of rapid development, gaining widespread usage as a simple and convenient method to quickly produce fibrous nanomaterials from a broad spectrum of materials, and its fiber-based applications have extended to every corner of the electronic products, whether the devices are two-terminals, such as sensors and RRAM, or three-terminal ones, such as field-effect transistors. Combining the capability of logic computation for FETs, sensing for sensors, and storage for memories, an innovative system consisting of devices prepared only by electrospinning can act as a practical computer in which all these function modules are fabricated separately by a different process. In addition, electrospun, nanofiber-based, neuromorphic electronic devices inspired by the human brain with ultra-low power consumption provide new chances for the superior design for hardware integrated circuits, which will reform the current information processing technology. Moreover, integrated systems, including as-developed soft electronics, such as flexible sensors, energy supply, memory devices, and optoelectronics, have huge potential for revolutionary electronic skin products and this kind of assembly will offer tremendous opportunities for both future research and commercial products.

In this review, we collectively presented the development of the electrospinning technique, various approaches for preparing nanofibers with different properties, and their representative electronic applications. First, the introduction of the electrospinning technique and three different types of ways, including pretreatment of the solution, usage of the coaxial spinneret, and near-field electrospinning, that were taken to fabricate all types of nanofiber structures, such as aligned arrays, hollow tubes, core-sheath, porous structure, and multichannel tubes, were put forward. Afterward, three main categories of factors that had a powerful influence on the electrospun nanofibers, composed of inner properties of solutions, conditions of the electrospinning process, and surrounding environmental factors, were discussed in detail. Based on the advanced technology of the electrospinning process, various electronic applications involving most of the two and three-terminal devices were explored, e.g., field-effect transistors with electrospun fiber-based channel layer, memristor featured by 1D active layer, and stretchable and flexible sensors, as the most valuable applications for electrospun nanofibers. Especially, artificial synapses and optoelectronics based on electrospun fibers were studied to reveal the potential of the electrospinning technique for developing multifunctional systems with low energy consumption.

Nevertheless, to truly approach cost-effective, daily life applications or large-scale and efficient commercial production for electrospinning, there are still several remaining challenges. Firstly, considering the characteristics easily affected by the surrounding environmental parameters, the electrospun nanofibers tend to be varied in different production batches, which leads to the poor performance and inferior quality of the prepared devices. Consequently, the inconsistency and instability of electrospinning cannot ensure the reliability and efficiency of industrialization production. Moreover, the low output of the electrospinning process (0.01 g–2 g/h) owing to the use of single needle nozzles hinders the further development of the electrospun fiber-based applications. Although the volume of production will be increased through the use of needle-free methods [140], the needs of industrial production are still unsatisfied.