A focused review of short electrospun nanofiber preparation techniques for composite reinforcement

3.1 Mechanical cutting

Mechanical cutting is a facile approach that offers varying degree of control over nanofiber length and relative effectiveness in retaining morphological features. Xu et al. electrospun polyacrylonitrile (PAN) nanofiber to prepare PI composites [52]. Small flakes of nanofibers were ground for 1 h, ultrasonicated (10 min) in ethanol, and dried (60°C) for 4 h. Microscopic characterization revealed short nanofiber length in 3–7 μm range (Figure 1(a)) with aspect ratio between 7.5 and 23.3. Grinding of nanofibers is likely to damage morphology as milling action has been reported to induce flake-like morphology of nanofiber [24], a fact author left unverified. Yoshikawa et al. fabricated brush-like symmetry of hydrophilic poly(styrene sodium sulfonate) on electrospun nanofibers of a polystyrene-based hydrophobic copolymer to assess suitability of the solvent system for hydrophobic polymers [53]. Copolymer had reactive sites for surface-initiated atom transfer radical polymerization (SI-ATRP). Mechanical cutting of nonwoven mat, in water and phase-separated systems of water and hexane, using a homogenizer was pursued after 3 h of polymerization. Generally, the fiber length became shorter and more uniform with increment in cutting time. The mean length was 11 ± 17 µm after 3 h of cutting. Homogenization did not affect the morphology or fiber diameter as confirmed from SEM images. Nanofiber cutting facilitated by using a mixture of water and hexane as nanofiber collection at interface of two liquids helped them encounter the blades. However, polystyrene nanofibers without brush-like symmetry escaped the blade action owing to their tendency to float on water surface, hence offering poor dispersion even after long hours of cutting. Overall, nanofiber cutting through homogenizer provides advantages of short processing time, ease of operation, high yield (80%), and control over length (few to tens of microns). This report claimed that length can be controlled through both batch size and processing time without any attempt to prepare different batch sizes. Huang et al. developed nanofiber recipe, with grafted poly(sodium styrene sulfonate) on initiating moiety, with an aim to shorten them and act as cell growth scaffold [54]. A water–hexane mixture was used to disperse nanofibers and homogenization was carried out for 3 h. An average length of 11 ± 17 µm was attained, with higher dispersion, when poly (sodium styrene sulfonate) was grafted as brushes. This behavior was induced by hydrophilicity of the grafted material. Further studies require to investigate the effect of grafted materials content on the fiber cutting efficiency. Kriha et al. prepared composite nanofiber by adding cobalt nanoparticles in P(MMA-c-VA) polymer solution [55]. The aligned electrospun nanofibers (1–3 µm diameter) were cooled under liquid nitrogen and cut to a length of 50–100 µm, using standard razor blade. Distinctive aspect of the adopted procedure was the retrieval of rod-like non-aggregated short nanofiber. There was no report of fiber length distribution to estimate the efficiency of the cutting procedure. Zhao et al. homogenized PI nanofibers to obtain short nanofibers with average length of 300 µm [56,57]. There was no attempt to establish parametric correlation to short nanofiber generation in the reported work.

Figure 1

Short nanofiber of (a) brittle PAN with average length range of 3–7 µm [52]; Copyright 2013, reproduced with permission from Elsevier Ltd, (b) composite nanofibers (GO/Iron oxide-reinforced PAN) with average length of 49.73 µm [59]; Copyright 2013, reproduced with permission from Elsevier Ltd, (c) ductile irradiated poly(lactic acid) (PLLA) with average length of 5 µm [24], and (d) pitting of PLLA nanofiber [24]; Copyright 2013, reproduced with permission from Elsevier Ltd.

Deuber and Adlhart reported a facile method of controlling the aspect ratio of water-soluble pullulan/PVA nanofibers [58]. Fiber length of 40 µm and aspect ratio of 140 were achieved after homogenization (13,000 rpm) of PVA nanofiber, in 1,4-dioxane solution, for 20 min. Wetting of nanofiber by the solvent was believed to be crucial to facilitate cutting. Feng et al. reported graphene and iron oxide-reinforced composite short nanofiber (d = 300 nm) preparation using homogenizer [59]. Nanofiber pieces of 1 cm × 1 cm were immersed in tert-butanol to agitate under higher shear forces (16,000 rpm) for 40 min. Average nanofiber length was of 49.73 µm (Figure 1(b)). Li et al. cut GLA/PVA nanofibers by mixing small fragments in ter-butanol with homogenizer (IKA T-18) operated for 15 min at 13,000 rpm [60]. Short nanofiber length of 27 µm was reported without standard deviation. Silica and PAN mats were cut and dispersed in camphene (100 mL) at 70°C by Si et al. [61]. Solution was homogenized for 30 min at 13,000 rpm to retrieve short nanofibers of 50 µm length. There was no report of attempt to measure statistical average diameter of the nanofibers through a range of 100–300 nm. In an identical study, Si et al. fabricated aerogels by dispersing silica nanofibers in 200 mL of pure water to homogenize it at 13,000 rpm for 20 min [62]. Short nanofiber length was not reported. Similarly, short PAN nanofibers (decorated with gold) of length 650 ± 218 μm were retrieved by mixing them with dioxane at 5,000 rpm and duration of 35 s [63]. Jiang et al. added electrospun PI nanofibers, in small pieces, to solvents (dioxane/water) [64]. After cooling with liquid nitrogen, the resultant paste was cut by operating the mixer at 3,500 rpm for 2 min. Average length was 14 ± 6 μm when the average diameter was 75 nm. Identical methodology was adopted to prepare short PI nanofiber in length distribution range of 20–40 µm [57].

Duan et al. irradiated non-woven nanofiber mat for 5 h from 15 cm to cross-link poly(MA-co-MMA-co-MABP) and PAN [65]. Then, these were mixed with dioxane and cut with razor blade for 45 s at 5,000 rpm. Different weight percentages of nanofibers were used to prepare aerogel although without length characterization. Adopting identical technique, another study from same group reported 150 ± 30 µm short nanofiber length [66]. Blender cutting was utilized to attain short nanofibers of PLA/PCL by mixing them with H₂O/tBuOH (500 mL) solvent. Short nanofiber diameter to length ratio of 1/150 μm was obtained [67]. Drummer employed same equipment to prepare short PAN nanofiber in water mixture [68]. Cutting was performed for 10 min at 23,000 rpm to retrieve short nanofiber of length 442 ± 171 µm. A similar work reported 416 ± 83 μm short PAN nanofiber length after nanofiber yarns were mixed with ethanol/water and cut in liquid nitrogen-assisted solidified condition [69]. Short yarns were sonicated to disperse nanofibers. Jiang et al. irradiated cross-linkable PNIPAM fibrous mat before mixing them with dioxane and mechanical cutting through mixer for 1 min at 5,000 rpm [70]. Reported short nanofiber length distribution was 100 ± 60 μm.

Xu et al. blended cellulose acetate (CA)/ polycaprolactone (PCL) nanofiber mats with deionized water and ethanol. Operation lasted 10 min at 19,600 rpm [71]. A uniform suspension of short nanofibers with length of 95.2 µm was obtained. Since a mixture of CA and PCA was used to spin nanofibers, it would have offered additional insight if the effect of either polymer on the fracture of nanofiber had been explained. Same author treated PAN and PI nanofibers with oxygen-plasma to improve hydrophilicity [72]. Nanofibers were blended in water/ethanol to attain 113 and 133 μm nanofiber length for PAN and PI, respectively. Brittleness of PAN was believed to facilitate mat fragmentation and lower length range. To prove the effect of hydrophilicity in enhancing the cutting efficiency, further investigation requires comparing the results with non-treated nanofiber mats. Another study by the same author combined features of earlier works by blending PAN and PI nanofiber chunks in ethanol [73,74]. Again, blending continued for 10 min at 19,600 rpm. Average length of short nanofibers was reported to be 123 µm though without considering the difference of PAN and PI nanofibers and their comparison. Using identical materials, equipment, and parameters, nanofiber length from tens to hundreds of micrometer was reported [75]. Cellulose acetate was electrospun and hydrolyzed by the same group to prepare short nanofiber suspension [76]. Blending was done using identical parameters with water as carrying medium. Short nanofibers with average length of 50 ± 15 µm were obtained. Although intact morphology was reported after treatment, author did not comment how treatment can affect nanofiber generation. Zhu et al. fabricated poly(bis(benzimidazo)benzo-phenanthroline-dione) nanofiber for sponge preparation [77,78]. Nanofiber mats were put in electric mixer along with 1,4 dioxane to be stirred 10 times, each time for 30 s. A wide length distribution of 50–500 µm was obtained.

This author’s attempts to shorten nanofibers using homogenizer or blender led to clogging of the chunk in drive shaft head of the equipment. This has not been reported in studies using homogenizer except a study by Xu et al. [79] where large chunks were used again to prepare nanofibers. This phenomenon limits the technique’s application in studies requiring pre-determined nanofiber mass fraction.

Thieme et al. synthesized triblock copolymers made up of poly(ethylene oxide) (PEO) and polylactide (PLA) and transformed it into nanofibers by the electrospinning process [7]. Mechanical cutting (motor-driven blade) of liquid nitrogen solidified ethanol (200 mL) and nanofiber (40 mg) suspension was carried out at 26,000 rpm for 10 min. Short Nanofibers were separated by centrifugation at 35,000 rpm. The obtained fiber had a length in the range of 5–15 µm that satisfied the criterion of being below 15 µm, acceptable for inhalation therapy, although length distribution was not reported. As differential scanning calorimetry results confirmed the presence of amorphous PLA and crystalline PEO phases, this can be further investigated to assess effect on nanofiber breakage. Jiang et al. produced nylon 6 mat with mean fiber diameter of 163 nm to be cut mechanically into short fibers [80,81]. Nylon 6 dispersion in dimethylformamide was prepared by cutting with rotation blades (RPM: 15000) at −13°C for 5 min. Nanofiber length range of staple nanofibers was few to tens of micron. Langner and Greiner cut poly(acrylic acid) (PAA) nanofiber (2.4 g) in a solution (2-propanol and demineralized water), at −18°C, using a blender for 1 min operating at 3,500 rpm [82]. Short nanofibers with an average length distribution of 87 ± 53 μm were retrieved. Boda et al. prepared collagen-gelatin (PCG) nanofiber of average diameter 249 nm [83]. Subsequently, membranes were mineralized (NaCl [1,000 mM], CaCl₂ [25 mM], and NaH₂PO₄·2H₂O [10 mM] in deionized water). Mineralized nanofibers were, first, frozen at −80°C and then cryo-cut at −20°C. As the diameter increased to 1,758 nm after mineralization, heavy mineral deposition offered aggregates of nanofibers instead of dispersion of individual fibers. Short nanofiber fragments (20 µm) were obtained after freeze drying. Apparently, higher diameter and mineral coating might have facilitated the fracture of fibers. Further studies focusing on changing mineral content and mineralization time will reveal whether agglomeration can be controlled.

Zhang et al. aimed to shorten ductile PLA nanofiber (625 ± 278 nm diameter) length, using mechanical stirring and ultrasonication in toluene/petroleum ether (T/P) dispersant media, for composite reinforcement [84]. Sonication was carried out (at amplitude: 40%; Pulse mode: 2 s/2 s) for 2 min. For mechanical cutting, solution was stirred for 24 h at 1,500 rpm using a magnetic stirrer. Short nanofibers could settle as sediment and the supernatant was decanted. Effectiveness of sonication in breaking the ductile polymer nanofibers was minimal. Additionally, yield of short PLA fibers was higher with higher toluene concentration probably owing to partial swelling of PLA in toluene. The surface of nanofibers became rougher and pitted, without fracture, at low volume fraction of toluene (T/P = 30/70). At higher concentration (T/P = 80/20), morphology change was noticed. To correlate toluene concentration (10–100% toluene in petroleum ether) with ease in fragmentation, mechanical stirring method was adopted. At a toluene concentration approaching 80%, stirring produced fiber with lengths of 50–700 µm range. The average length of the PLLA fibers was 220 ± 112 µm with most of the cut fibers’ lengths lying in 100–200 µm range. Since toluene concentration of 80% produced damaged and intact morphology during sonication and stirring, respectively, effect of temperature rise, during sonication, on morphology needs to be carefully sorted out. Chen et al. received different weight fraction of gelatin/PLA nanofiber (1 cm² × 1 cm²) membranes in ter-butanol solvent to prepare short nanofibers for 3D-scaffolding [85,86,87,88]. The mixture was homogenized for 15 min at 12,000 rpm. At least 50 random measurements of nanofiber length were taken using optical microscope. Average fiber length and diameter were 86 µm and 1,114 nm, respectively, although nanofiber length range was 4–600 µm. Adopting same nanofiber material, nanofiber homogenization fractured brittle gelatin into groove fibers of 14.6 µm length [89]. Similarly, Deuber et al. electrospun pullulan/poly(vinyl alcohol) nanofibers (d _avg = 240 nm) to be used in aerogel preparation [90]. Nanofiber mat pieces (1 cm² × 1 cm²) were dispersed in 1,4-dioxane and homogenized (IKA T25) for 20 min at 13,000 rpm. Fiber length distribution of 48.8 ± 30 µm was obtained. Duan et al. electrospun poly(2VP-co-MABP), poly(MA-MMA-MABP), and PAN nanofibers and did curing using UV light [63]. Gold nanoparticles were anchored on mats after improving their hydrophilicity through ammonia treatment. Short nanofibers in dioxane were retrieved using rotating mixer operated at 5,000 rpm for 35 s. Average fiber length of 650 ± 218 µm was used to fabricate sponges using freeze drying method. Si et al. homogenized PAN and silica nanofiber in camphene at 70°C [61]. Homogenization time and rpm were 30 min and 13,000, respectively. Uniform nanofiber dispersion with an average length of 50 µm was attained. Another study utilized pre-oxidized electrospun PAN nanofiber to retrieve short nanofibers with length in the range of 40–60 µm [91]. For that, nanofiber piece was homogenized in water/ethanol mixture at 13,000 rpm for 30 min.

Xu et al. utilized blending operation to retrieve short nanofibers of PAN and PI [73]. First, nanofibers were treated with oxygen plasma, to make the surface hydrophilic, and then cut into 1 cm × 1 cm pieces to be immersed in deionized water. The suspension was blended for 10 min at 19,600 rpm to get nanofibers with an average length of 123 µm. Another related route employing mechanical/physical force was reported by Deniz et al. [92]. Titanium nanoparticle solution with polyvinylpyrrolidone was prepared and electrospun to get bead-free nanofibers (d _avg = 200 nm). These nanofibers underwent calcination to attain pure titania nanofibers. Mortar and pestle-assisted crushing of the nanofibers produced nanofibers in 1–10 µm range. Crystalline structure of the nanofibers was confirmed through XRD, and the brittleness was believed to facilitate fracture of the nanofibers in such a short length range. These short nanofibers were successfully dispersed in thermoplastic polymers to electrospun composite nanofibers. Despite setting the scope of preparing short titania nanofiber with different polymers, no correlation between nature (soft/rigid) of polymer and ease/shortening of nanofiber was drawn. Yao et al. spun PCL/PLA nanofiber and cut them into flakes to be mixed with ethanol [74,93,94]. Soaked pieces were solidified using liquid nitrogen and ground using mortar and pestle. The product was sieved through 1 mm mesh to sift larger pieces and to be returned for grinding. There was no attempt to measure the length of short nanofibers.

In addition to the works summarized above, there has been attempts to prepare short nanofibers from homogenization, primarily, and through other mixing methods but without reporting or measuring nanofiber length [62,65,74,93,95,96,97,98]. This might have been owing to insignificance of length characterization for the application intended.

3.2 Ultrasonication

Mechanism of sonication involves bubble generation in fluid that grows and collapses thereby releasing energy. Bubble growth, initiating at cavities, can be up to 50 mm in diameter under negative pressure [99]. Time order of bubble generation and collapse are in µs and ns, respectively [100]. This method has been effective for the cutting of carbon nanotubes [101]. Sawawi et al. conducted study on several nanofibers using mechanical- and ultrasonic-assisted arrangements [24]. Sonication was carried out using a sonicator (power: 750 W; probe dia: 13 mm) working at 20 kHz. Pulse mode of 2 s/2 s and amplitude of 80% were adopted and experiments were conducted in a sonication time of 1–35 min. Short nanofibers of poly(methyl methacrylate) (PMMA) and polystyrene (PS) were detected after sonication time of 40 and 60 s, respectively. Measured lengths were 10.3 ± 5.6 and 10.5 ± 6.2 µm for PMMA and PS, respectively. Whitish color of the supernatant indicated a degree of homogeneity achieved and scission of nanofibers. Distinctive aspect of the study was breakage of the ductile fibers such as PLLA as brittle PAN nanofiber fracture was possible through sonication alone. Realizing the higher ductility of these nanofibers through tensile tests, PLLA membrane was irradiated for 12 min with the intensity of 14.75 mW/cm² by UV Ozone Procleaner prior to sonication. PLLA short nanofiber with length in the range of 5 ± 5 µm were obtained (Figure 1(c)) with localized fiber etching evident in SEM images (Figure 1(d)). It seemed that the micro jetting and erosion created rougher weaker zones. Further investigation on PS nanofibers revealed that average fiber length (6 ± 2 µm) of random nanofibers was higher than that of aligned nanofibers (3 ± 1 µm). Also, aligned nanofiber sonication at temperatures of 30 and 90°C showed no significant effect on nanofiber lengths. This work utilized turbidity measurements to assess degree of nanofiber shortening which might be erroneous as erosion of material from surface, as reported, might change the solution color.

Liu et al. tested morphological durability of the SiO₂ (∼10 mg) nanofibers by having trace amount of nanofiber placed in a 20 mL glass vial with 10 mL EtOH [102]. The suspension was then subjected to vigorous ultrasonic vibrations (200 W) for three cycles, each lasting 5 min. An important finding was to correlate morphology of the nanofibers with pyrolyzing temperatures (600, 800, 1,000, 1,200, and 1,400°C). In 600–1,000°C temperature window, fiber morphology was retained, and the average aspect ratio of the broken nanofibers was larger than 100. However, above 1,000°C, short fiber bundles were witnessed because of fusion and melting caused at cross over points of the nanofiber mats. Additionally, it can be concluded that the fiber fracture for brittle materials such as silica is possible although fusing phenomenon induce greater fiber–fiber interaction. Chen et al. prepared 5 wt% glass nanofiber suspension in water to be subjected to vigorous ultrasonication using an ultrasonic probe (100 W) [103]. Short nanofibers with few to tens of micrometers in length were obtained. Proper characterization of nanofiber length and distribution was missing. Wang et al. sonicated titania nanofibers for 1 h in a mixture of water to attain short nanofiber of 3 µm length [104].

Similarly, Mulky et al. conducted comprehensive experiments to investigate the effect of sonication time, temperature, and nanoparticle addition on staple PLLA nanofiber length (Figure 2) [105]. Nanofiber flakes (0.5 cm × 1 cm) immersed in relevant dispersant media were sonicated (400 W, pulse on/off 2 s/2 s). Ice bath was arranged to (ice/acetone) maintain −80°C temperature. Titanium nanoparticle and fiber ratio was 1:1. Average diameter of the electrospun nanofiber was 244 nm and the dispersant media were water, water/ethanol (4:1), and hexane. Key factor was interaction between PLLA nanofiber and dispersant media: surface energy and hydrophilic/hydrophobic character. Water, water/ethanol, and hexane suspensions offered lumps, disintegration of non-woven mat, and disentanglement/fracture of nanofibers, respectively. In water-based suspensions, fiber–fiber interaction was believed to be stronger than water–nanofiber interaction. On the other hand, low vapor pressure of hexane was conducive to cavitation resulting in higher degree of implosion to cut nanofibers. For hydrophilic and hydrophobic TiO₂ nanoparticles, average staple fiber lengths were 63 and 51 µm, respectively, whereas this value for neat nanofiber was 39 µm. As the scope of the study involved investigating the effect of solvent and nanoparticle addition on short nanofiber preparation, there were different outcomes from both the approaches. A hexane dispersion provided complete fragmentation whereas nanofiber agglomerates were excluded from nanoparticle suspension to take length measurement. That makes the nanoparticle addition inefficient route to prepare pre-determined mass fraction featured samples. Effect of temperature was investigated, at 0 and −80°C, in hexane. At −80°C, average lengths obtained were 77, 63, 44, and 35 µm after 10, 20, 30, and 60 min of sonication, respectively. At 0°C, nanofiber lengths were 54 and 47 µm in ice bath and ice-acetone bath, respectively. Additionally, it was concluded that below a threshold aspect ratio, induced stresses in nanofiber may not be high enough to fracture them.

Figure 2

Fiber length distribution in relation to (a) sonication time and (b) nanoparticle type and (c) sonication bath temperature [105]; Copyright 2014, reproduced with permission from Elsevier Ltd.

3.3 Electrical spark

Basic idea of this technique is to interrupt the electrospinning jet through electric spark periodically to induce thermal energy that will cut the nanofiber. Fathona and Yabuki utilized electric spark to interrupt electrospinning jet to produce short nanofibers (Figure 3(b)) [106]. Aluminum electrodes were placed just next to needle to apply square wave voltage, (f: 30 Hz; duty cycle: 90%), to generate electric spark. The fibers that flew through electrode gap were shortened whereas those evading electrode gap were collected as continuous fibers. The length of the short fiber was approximately 100 µm with one edge curled although a length range of 22–400 µm was recorded in a full experiment. Yield was reported by counting the number of fibers in a unit area. This value was 1–5 short nanofibers per 0.12 mm × 0.2 mm area. Due to the limited spark-controlled area (1.5 mm × 0.5 mm), yield of continuous nanofibers was much higher. It was obvious that through periodic activation of electrical field and thermal energy from the spark, electrospun nanofiber was stretched and cut, respectively (Figure 3(a)). Diameter of short nanofiber at edges was smaller compared to diameter measured at center owing to compressive force exerted by the electric field. Spark-activated cutting was demonstrated to be a function of polymer solution velocity, solution jet trajectory, and frequency of spark activation. More studies are required to explore whether flow rate, frequency of spark, and trajectory can help enhance yield and narrow length distribution of short nanofibers.

Figure 3

(a) Short nanofiber and (b) schematic of the electric spark mechanism utilized for obtaining short nanofiber [18]; Copyright 2013, reproduced with permission from Elsevier Ltd.

3.4 Entanglement loss

Solution entanglement number is a fundamental concept considered when selecting polymer–solvent combination. It is a measure of minimum amount of polymer chains, in entangled form, that are necessary to make a continuous polymer jet traveling in electrostatic field. Fathona and Yabuki exploited the idea of altering solution dynamics under electrostatic forces, by changing polymer concentration, to prepare short nanofiber [39]. Acetone/dimethylacetamide (DMAc) (solvent) and cellulose acetate (polymer) were used to prepare dope solution. By lowering polymer concentration below the threshold value required for continuous spinning, short fibers, spun at 13 wt% concentration, with length in the range of 50–150 µm were obtained. However, polymer concentration ranges of 9–12 and 16–18 wt% produced beaded and continuous fibers, respectively. Although it was concluded that the polymer concentration was the most important factor in shortening nanofibers, it missed to relate that fiber length distribution increased with polymer concentration within the concentration window offering short nanofibers. Tungprapa et al. reported that 2:1 volume ratio of acetone/DMAc has a low surface tension that had bearing on solvent–polymer interaction [40]. That would alter the tensile strength of the polymer solution to achieve facile segmentation. The average length of nanofiber at 13, 14, and 15 wt% were 130, 180, and 230 µm, respectively. It was expected that the increment in solution viscosity may stretch the jet for a longer distance before breaking. Additionally, fiber length distribution narrowed down when the polymer concentration dropped from 15 to 13 wt%. Once optimum polymer concentration, 13 wt%, was determined, effect of flow rate and voltage was investigated. A flow rate of 0.02, 0.04, and 0.4 µL/min offered average lengths of 65, 120, and 150 µm, respectively, at a voltage of 5.5 kV. Above 0.5 µL/min, continuous fibers were retrieved. When the voltage values were 4.5, 4.7, 5.5, and 7 kV, average fiber lengths were 670, 250, 170, and 40 µm, respectively. An interplay of the following factors led to the formation of short nanofibers: (1) tensile strength of the polymer solution, (2) the lateral perturbation of the surface charge on solution, and (3) longitudinal force. Another insightful study by the same group incorporated titania nanoparticle in cellulose acetate (13 wt%) solution to prepare short nanofibers [107]. Objective was to investigate how nanoparticle surface charge and concentration affect composite nanofiber length. Concentration was 0.5–17 wt% in polymer solution that translated into 4–56 wt% after solvent evaporation. For analysis, 0, 4, and 38 wt% nanoparticle concentrations were considered representative of concentration range adopted with appreciation of change in length after addition of every 5 wt% of nanoparticle. Diameters range for 0, 4, and 38 wt% nanofibers were 110–150, 250–270, and 310–370 nm, respectively. Average lengths for same concentrations were 120, 112, and 47 µm as shown in Figure 4(a)–(c), respectively. Apparently, nanoparticle addition increased viscosity of the solution thus avoiding jet interruption and breakage to produce nanofibers, the reason why wider length distribution was offered at lower concentration.

Figure 4

Fiber length distribution of (a) 0, (b) 4, and (c) 38 wt% TiO₂ nanoparticles-reinforced nanofibers [107]; Copyright 2014, reproduced with permission from Elsevier Ltd.

An increment of repulsive force, F _R, induced thinning of polymer jet and eventual break (Figure 5). At low concentrations and pH = 4, repulsive force was generated by agglomerates of positively charged nanoparticle and negatively charged polymer thereby stretching the jet (Figure 5(a)). However, at pH = 9, polymer was repelled by negatively charged nanoparticles thus improving dispersion and reducing chances of jet breakage (Figure 5(a)). At higher nanoparticle concentration, either positively or negatively charged agglomerates will induce breakage.

Figure 5

Breakage mechanism and repulsive force generation in polymer jet at (a) pH = 4 and (b) pH = 9 [107]; Copyright 2014, reproduced with permission from Elsevier Ltd.

In an another study, Fathona et al. altered the inner diameter of electrospinning needle to control nanofiber length [38]. Initial experiments confirmed that the 4.5 kV voltage and flow rate of 0.1 μL/min were conducive to short nanofiber generation. It was reported that nanofiber diameter increased with increment in needle diameter. And for 0.26, 0.16, and 0.11 mm needle diameter, average nanofiber lengths were 123, 80, and 50 µm, respectively. An imbalance of forces of surface tension, coulomb attraction, and net cohesive force of polymer led to the breakage of nanofibers. Thinning and stretching of nanofiber edges was caused by longitudinal force exerted by electric field. An ancillary study can investigate the effect of these parameters on the yield of short nanofiber by counting them per unit area.

Greenfeld and Zussman conducted 76 experiments on PMMA nanofiber to understand entanglement loss relation to short nanofiber production [41]. Electric field intensity and polymer concentration were used as free variables. The polymer concentration “c” was identified with semi-dilute entangled regime with a relative concentration c/c _e (c _e is the entanglement concentration) lying between 1 and 2. This range denotes transition between beads and continuous nanofiber production. In a fluid jet, electrostatic force applied on charged ionic species of solution generates stress. The stress in the jet is directly proportional to s·E, where E is the electrostatic field intensity, and “s” is the surface charge density. If the stress exceeds the jet tensile strength, the jet will break at weak points. As the effect of the break is no longer significant on the stresses in the jet, the next segmentation occurs at a distance from the previous break point. This mechanism will generate a short nanofiber. Evolution of this phenomenon is shown in Figure 6. In general, the fragment length gets shorter with: (1) increase in the electric field intensity, (2) decrease in the polymer concentration, and (3) decrease in the flow rate. PMMA nanofiber ranging from 1 to 1,000 µm in length and diameters ranging from 50 nm to 3 µm were collected. More studies are required to confirm the reliability of the technique and whether entanglement loss is possible for both high and low molecular weight polymers, since the solution entanglement number is directly related to number average molecular weight of the polymer [108].

Figure 6

(a) Diameter variation; (b) fragment; (c) neck production; (d) breaking; (e) individual fragments; (f) round end; (g) fractured face; and (h) fractured cross-section [41]; Copyright 2013, reproduced with permission from John Wiley and Sons.

3.5 Chemical treatment

This method, primarily, considers chemical composition (crystalline-amorphous phases) of polymer to exploit it for short fiber preparation. Kim and Park did aminolysis of PLA for 1, 3, and 5 h to retrieve the short form of nanofibers [23]. For aminolysis, a 10 wt% of PLA nanofiber suspension (distilled 1,6-hexanediamine (Aldrich)/2-propanol) was prepared and allowed to react at 37°C. Application of gentle shear stress (orbital shaker: 500 rpm) fragmentated the nanofiber into nanocylinders that were centrifuged and washed, subsequently. Aminolysis degraded amorphous region of the nanofibers and helped in concomitant lamellar crystallization thereby producing semi-crystalline nanocylinders (Figure 7). An increase in aminolysis time (1–5 h) improved fragmentation with mean length decreasing from 15.8 to 5.2 µm; corresponding aspect ratio varied from 48.6 to 16.0. Treatment for an hour was reported to induce dents and cracks on surface but prolonged treatment for up to 5 h was associated with morphology loss. To investigate the effect of fiber diameter on treatment induced nanofiber length, nanofibers of diameter 958, 325, and 152 nm were fabricated and treated for 3 h. Corresponding lengths and aspect ratio were reported to be 5.1, 7.7, and 11.4 µm and 75, 23.7, and 5.3, respectively. Apparently, with diameter reduction, an easy fragmentation was achieved when all other parameters were constant. This might hint at offering higher yield if the total mass fraction of the nanofiber is controlled by keeping diameter low, a relation which requires further investigations to be established.

Figure 7

Aminolysis-induced fragmentation and crystallization of PLA nanofibers [23]; Copyright 2018, reproduced with permission from John Wiley and Sons.

Celebioglu and Uyar reported successful spinning of γ-cyclodextrin/PEO nanofibers from homogeneous solutions [109]. PEO was chosen as sacrificial matrix to obtain short nanofibers that were believed to enhance volatile organic compound removal. γ-cyclodextrin concentration was kept at 20% by weight but the PEO concentration was changed from 1 to 3 wt%. Non-beaded nanofibers were obtained at all concentrations except 1 wt% PEO. Diameter of the nanofibers increased with the increment in PEO concentration. Chloroform rinsing was conducted for 3 h to dissolve PEO matrix and retrieve short nanofibers. Nanofiber generation was explained by breaking down of long PEO polymeric chains. Author did not report the average length or aspect ratio.

3.6 UV irradiation

Selective crosslinking is another tailorable method to exploit heterogenous composition of nanofiber for short fiber preparation. Stoiljkovic and Agarwal shortened poly(butadiene) nanofiber utilizing ultraviolet irradiation [110]. Two types of masks were used: (1) mask with slit span between 100 and 150 µm and (2) glass plate having narrow palladium strips of 50 µm. Distance between two consecutive palladium strips was 20 µm that determined length of the cut fibers. A UV photoreactor was used to irradiate nanofiber thereby inducing crosslinking. Removal of non-cross-linked polymer was achieved by immersing substrates in suitable solvent for 60–120 s. Nanofiber length range was 100–150 µm in case of random fibers, in correlation with slit span. The length of the aligned fibers covered with 100 µm mask was in the range of 100–160 µm (140 ± 30 µm), whereas the fibers covered with 20 µm mask had length from 20–40 µm (35 ± 5 µm). Microscopy confirmed that fiber cannot be prevented from swelling during solvent treatment. Further studies are required to correlate morphology of the nanofiber with UV exposure time and nature of solvent used for removal of non-cross-linked sections. Basic requirements for this technique were that polymer should be (1) photo-cross linkable and (2) amenable for electrospinning.

5.1 Bulk matrix modification

Short nanofiber preparation techniques such as mechanical cutting and sonication do involve solvent to facilitate shortening. Selection of a suitable solvent makes them compatible with principles of solvent casting method for nanocomposite fabrication [48]. Here solvent is likely to serve dual purpose as facilitating medium for nanofiber fracture and dispersant. Such a dispersion will be mixed, eventually, with matrix for bulk matrix modification (Figure 9(a)) and solvent will be evaporated using vacuum oven or rotary evaporator. Since homogenization and sonication are conventional mixing techniques for nanocomposite preparation, matrix modification using short nanofibers will incur no additional cost in terms of equipment and materials.

Nature of electrospun nanofiber polymer is another aspect that makes this approach promising. Rigid materials such as PAN [52] and silica [103] are easy to be fractured and collected in powder form for them to be directly mixed in resin. However, heat treatment of these high strength nanofibers and surface treatment such as plasma etching might be associated with surface defects and remaining structural defects, respectively [111,112]. These defects are likely to initiate fracture of short nanoparticles during shear mixing or sonication [113,114]. Shortening of reinforcement will reduce aspect ratio below the critical value required for efficient stress transfer. Such complications with brittle materials might be responsible for degradation of composite properties.

Shortening, however, is difficult to practice in case of soft polymers such as nylon nanofibers. Solvent-assisted dispersion and casting help tackle short nanofibers of soft polymers to improve bulk matrix properties. In addition to nanofiber-reinforced matrix composites, such a recipe can be conveniently adopted to prepare laminates with modified matrix, also called fiber-reinforced nanocomposites, through vacuum bagging.

Other techniques such as UV irradiation and chemical treatment are reported to incur nanofiber morphology loss and alteration in phase structures. This entails poor mechanical properties, hence making them unsuitable for composite reinforcements.

Short nanofibers for bulk matrix modification has been successfully utilized to improve mechanical properties [48,115]. It was concluded that 2 wt% of short nanofiber was as efficient as 38 wt% continuous nanofibers in improving modulus (Figure 8(a)). Adopting similar approach, dielectric properties of PI matrix were improved at an optimum concentration (Figure 8(b)) [52]. Short nanofiber reinforcement has been effective too in improving thermo-mechanical properties of polymer matrix composites and in tailoring interphase responses [116]. There are reports demonstrating the effectiveness of these reinforcement in introducing ductile fracture modes and mechanisms when they are incorporated at optimum concentration [117].

Figure 8

(a) Mechanical reinforcing potential of short nanofiber [48] and (b) improvement in dielectric properties at optimum short nanofiber concentration [52]; Reproduced with permission from Elsevier Ltd.

5.2 Spray deposition

A dispersion of short nanofibers, attained through sonication or mechanical cutting, can be used directly to alter interphase or interlaminar properties of fiber-reinforced nanocomposites (Figure 9(b)). A combination of suitable solvent and preparation technique has been highlighted to deposit short nanoparticle on fiber surface [118]. A major difference here, when compared to bulk matrix modification, is that solvent removal step is not required through advance equipment. As the purpose of solvent was to carry the nanofiber to fabric reinforcement surface, it can be allowed to evaporate at room temperature or in an oven, subsequently. Selection of solvent is crucial here as a volatile solvent will shorten evaporation time. Application of short nanofibers, using this technique, offers a window of tailoring interfaces to enhance interaction with matrices. Also, spray deposition is likely to improve the mechanical adhesion of nanofiber with micro-fiber reinforcement. Once the deposition is complete, lamina can be stacked to fabricate laminate using hot press or vacuum infusion.

Figure 9

Short nanofiber application in (a) bulk matrix modification, (b) spray deposition, and (c) direct spinning.

5.3 Direct spinning

Another application area is to use prepreg as collector or substrate of nanofibers to enhance interlaminar fracture properties of fiber-reinforced nanocomposites [119]. This involves wrapping collecting drum with prepreg and spinning nanofibers (Figure 9(c)). Such an approach excludes the requirement of handling nanofiber veils for laminate fabrication since the lamina with deposited nanofibers will be stacked to cure resin through hot press. This approach has two major complications: (1) spinning time should not outlast prepreg cure time and (2) nanofiber melting temperature should be higher than curing temperature of prepreg.

More mechanical adhesion is possible since nanofibers are directly deposited on prepreg (Figure 10(a) and (b)). In situ short nanofiber techniques such as electric spark and entanglement loss are suitable for direct spinning. This is particularly so since short nanofiber cannot be collected from substrate, a feature that can restrict their applications.

Figure 10

(a) Direct deposition of nanofibers on prepreg and (b) improvement in interlaminar fracture toughness [120]; Reproduced with permission from John Wiley and Sons.

This approach has been successfully utilized to improve mechanical properties of fiber-reinforced composites, primarily, that involves tensile [121,122], flexural [123,124,125], impact [126,127,128,129], indentation [130,131,132], interlaminar fracture toughness [133,134,135], fatigue [136,137], and viscoelastic [138] properties of laminated composites.