Introduction

In electrospinning, ultrafine polymer nanofibers may be obtained by the application of a strong electrical field between a grounded target and a polymer solution that is pumped from a storage chamber through a small capillary orifice. Fibers are collected as a nonwoven mesh or membrane on a collector plate that acts as the counter electrode.1,2 These fibers are of increasing interest to the food industries due to their potential application as novel packaging materials, membranes, or controlled release systems in ingredients.

To date, a wide variety of polymers have been electrospun with synthetic polymers yielding the best results. Biopolymers such as chitosan derived from agricultural raw materials and used in foods are of great interest because they may be nontoxic, edible and digestible, biocompatible and biodegradable, and renewable and sustainable, giving rise to a broader utilization especially in fields such as biomedical sciences, pharmaceuticals, cosmetics, and other related fields.3–5 However, electrospinning of biopolymer solutions has proven to be difficult due their polycationic nature, low chain flexibility opposing chain entanglement, and poor solubility in organic solvents. In particular, the repulsive interaction encountered in similarly charged polycationic chitosan chains may prevent sufficient chain entanglement required for continuous jet ejection and, therefore, fiber formation.

One approach to circumvent this problem has been to add cospinning agents such as salts or other polymers. These cospinning agents may either provide for additional entanglement and/or decrease electrostatic repulsive interactions. For example, chitosan has been electrospun from mixtures of chitosan and polymers that facilitate entanglement such as poly(vinyl alcohol),6,7 poly(ethylene oxide) (PEO),8,9 and others.10 Pure chitosan has only been successfully electrospun from highly concentrated solutions of chitosans of relatively low molecular weight and/or in solvents such as trifluoroacetic acid,11 1,1,1,3,3,3-hexafluoro-2-propanol12 where further purification and removal of the solvent from fibers was subsequently required.13

In this study, we suggest that addition of surfactants may be used to improve spinnability of solutions to yield natural nanofibers. Surfactants are amphiphilic molecules that readily absorb at surfaces and thereby lower surface or interfacial tension of the medium in which they are dissolved; a key parameter that influences electrospinning. Surfactants above a critical concentration self-assemble to form a variety of colloidal structures. These colloidal structures have different properties from those of the dissolved surfactant monomers, e.g., solubility, surface hydrophilicity, charge density. We further suggest that surfactants with charged headgroups may exhibit strong electrostatic interactions with charged biopolymers in the solution (e.g., chitosan). This may have a pronounced effect on solution properties, such as conductivity and surface tension, which in turn influences the occurrence of bending instabilities yielding fibers with different structures, thickness, and mechanical strength. For example, it has been reported that the addition of a small amount of nonionic surfactant to polymer solutions reduced the onset voltage required to induce spinnability and improved reproducibility of the electrospinning process.14 Others found that, in systems of nonionic polymers, nonionic surfactants did not stop bead formation but greatly reduced it, while cationic surfactants prevented beaded fibers and led to fibers with smaller diameters.15 In a previous study, we demonstrated that addition of nonionic surfactants to solutions of chitosan or chitosan–PEO blends improved fiber characteristics. When ionic micelles were added to these systems, conductivity was significantly improved and bead on string structures on the filaments were suppressed, allowing for the formation of smaller bead defect-free fibers.

Based on our previous studies and the above cited published literature, we hypothesize that addition of surfactants below, at, or above the critical micellar concentration significantly influences fiber morphologies of electrospun fibers since polymer–surfactant and subsequently polymer–polymer interactions are possibly altered. The objective of this study was, therefore, to determine the influence of surfactant type and concentration on solution properties, as well as morphologies and compositions of electrospun fibers from chitosan–PEO–surfactant solutions.

Materials and Methods

Materials

Chitosan with a medium molecular weight was obtained from Primex (Reykjavik, Iceland) in the form of flakes. As stated by the manufacturer, the viscosity in 1% acetic acid was 569 cP (M w ~ 1,000 kDa) and the degree of deacetylation was 80%. PEO (Cat. # 343) with a molecular weight of 900 kDa was purchased from Scientific Polymer Products (Ontario, NY, USA). Glacial acetic acid (CAS # 64197, UN 2789) was purchased from Acros Organics (Morris Plains, NJ, USA). Anionic sodium dodecyl sulfate (SDS; #71729) was obtained from Fluka, nonionic polyoxyethylene glycol (23) lauryl ether (Brij 35; #P1254) from Sigma, and cationic dodecyl trimethyl ammonium bromide (DTAB; CAS # 1119944) from Acros Organics (see above). The critical micellar concentrations were 2.3, 0.09, and 14 mM for SDS, Brij 35, and DTAB, respectively.16 All reagents were used as received from the manufacturer without further purification.

Methods

Preparation of Polymer Solutions

Solutions were prepared with distilled and deionized water and reagent grade glacial acetic acid. Chitosan, PEO, and chitosan–PEO dispersions (3:1) at an overall polymer concentration of 1.6% (w/w) were prepared by dispersing chitosan and PEO in 50% and 90% acetic acid. Based on preliminary experiments on the electrospinnability of chitosan–PEO blend solutions, a 3:1 mass ratio of chitosan to PEO was selected. Solutions containing surfactants at three different concentrations below, at, and above the critical micellar concentration in water (0.01, 0.09, and 2 mM Brij 35; 10.9, 14, or 36 mM DTAB; 1, 2.3, and 10 mM SDS) were obtained by dissolving the surfactants first in distilled deionized water or acetic acid followed by dispersion of chitosan and PEO. Additional controls consisted of 0.4% PEO and of 1.2% chitosan—the concentrations of individual components used in the chitosan–PEO blend. Concentrations of all polymer solutions were expressed in percent weight per weight, except where noted otherwise. All the solutions were prepared under constant magnetic stirring for a minimum of 8 h at 25 °C to ensure complete dissolution of the polymers. All solutions were then used immediately for electrospinning and analysis of solution properties.

Electrospinning

A previously described electrospinning setup was used.17 Briefly, a syringe (Micro-Mate, Popper & Sons, New Hyde Park, NY, USA) fitted with a 0.69-mm diameter stainless steel capillary (No. 91019; Hamilton, NE, USA) with a blunt tip was filled with 20 mL of polymer solutions. The syringe was placed in a syringe pump (Harvard apparatus; 11plus, Holliston, MA, USA) which permitted adjustment and control of solution flow rates. The metal capillary of the syringe was connected to the positive lead of a high-voltage power supply (Gamma High Voltage; ES 30P-5W, Ormond Beach, FL, USA) operated in positive DC mode that could generate voltages up to 30 kV. A grounded copper plate wrapped in aluminum foil and mounted on two polypropylene blocks was used as the target for collection of fibers and/or beads. The target was placed 10 cm away from the capillary tip. The syringe pump delivered polymer solution at a controlled flow rate of 0.02 mL/min. The electrospinning was carried out at an applied voltage of 20 kV and a temperature at 25 °C. These conditions were kept constant throughout all experiments. Electrospun fiber mats were stored in desiccators prior to fiber characterization to prevent moisture uptake.

Solution Viscosity

Solution viscosity was measured with an oscillatory rheometer equipped with a double coaxial cup and bob measurement system (length = 40 mm, diameter = 26.66 mm, gap width = 0.225 mm; MCR 300, Paar Physica, NJ, USA). The shear stress σ (in pascals) was recorded as a function of shear rate \( \dot \gamma \) (per second) from 10−3 to 103 s−1. Solutions were equilibrated to 25 °C prior to all measurements using a Peltier system. Reported results are averages of triplicate measurements. Measurements were fitted to the power law model:17

$$ \sigma = K{\dot \gamma^n} $$
(1)

where K is the consistency coefficient and n is the flow behavior index. If the flow behavior index n equals 1, the solution behaves as a Newtonian fluid; if the index n is smaller than 1, the solutions is a shear-thinning fluid; and if the index is larger than 1, the solution is a shear-thickening fluid.

Solution Conductivity

Electrical conductivity of the polymer solutions was determined using a Zetasizer (Nano series, Malvern Instruments, Worcestershire, UK). The temperature was adjusted to 25 °C prior to the measurements. Reported results are averages of triplicate measurements.

Surface Tension of Solutions

Surface tensions of solutions were determined using a digital tensiometer (Model K10ST, Kruss USA, Nazareth, PA, USA) based on the Wilhelmy plate method.17 Forty grams of the tested solutions was poured into a 70-mm diameter glass beaker which had been previously rinsed with absolute ethanol and deionized and double-distilled water and dried at 70 °C overnight to remove any surface-active material. Solutions were then equilibrated to 25 °C and the platinum plate was lowered into the dispersion to form a meniscus. The surface tension σ was calculated in millinewtons per meter from the force F acting on the platinum plate using Eq. 2 where L is the length of the total meniscus (2 × length + thickness of the plate) and θ is the contact angle:

$$ \sigma = \frac{F}{{L\cos \theta }}. $$
(2)

Results shown are averages of duplicate measurements and duplicate samples.

Electron Microscopy of Nanofibrous Structures

The morphology of electrospun nanofibers was observed with a field emission scanning electron microscope (FESEM 6320 FXV, JEOL, MA, USA) operating at 5 kV. Nanofibers were electrospun directly onto aluminum scanning electron microscope stubs which were mounted on the grounded collector plate. After collection of the fibers, samples were sputter coated with Au in a sputter coater (Model 108, Cressington, Watford, UK) for 60 s to reduce electron charging effects. Additional coatings were necessary for samples containing DTAB. The average fiber diameter was determined by image analysis (ImageJ, NIH, USA) from >50 randomly selected fibers for each sample.

Fourier Transform Infrared Spectroscopy

Compositional and chemical characteristics were evaluated by recording infrared spectra of electrospun fibers using a Fourier transform infrared (FTIR) spectrophotometer (Model IR Prestige 21, Shimadzu Corporation, Columbia, MD, USA) with an attenuated total reflection unit attached. Samples were mounted on the mirror and each specimen was scanned between 4,000 and 700 cm−1. Each measurement consisted of an average of 32 scans at a resolution of 4 cm−1 using a Bessel apodization. IR-Solution (Shimadzu, Columbia, MD, USA) and Peakfit 4.12 (SeaSolve Software, San Jose, CA, USA) were used to analyze and deconvolute infrared spectra.

Differential Scanning Calorimetry

Thermal analysis of electrospun fibers was carried out with a differential scanning calorimeter (Model Q100 DSC, TA Instruments, DA, USA). Samples of approximately 5 mg were loaded in DSC pans that were closed using a crimping tool. The specimens were equilibrated to 25 °C for 5 min and then heated from 25 to 180 °C at a heating rate of 2 °C/min. The temperatures of the sample and reference pans were measured as a function of oven temperatures to determine heat flows. Results represent an average of four measurements.

Statistical Analysis

All experiments were performed in duplicates or triplicates and repeated twice (see above). Data are expressed as the mean + standard deviations. Mean separations were determined using the Student’s t test with a level of significance of p < 0.05 (SAS Ver. 9.0, SAS Institute, Cary, NC).

Results and Discussion

Physical Properties of Composite Solutions

Conductivity

Electrical conductivity of 50% aqueous acetic acid was 0.98 mS/cm whereas that of 90% acetic acid was 0.03 mS/cm (Tables 1 and 2). At lower solvent concentrations, more molecules could be ionized and thus higher conductivities were observed. Pure PEO solutions had approximately the same conductivity as the solvents, i.e., 50% and 90% acetic acid. Conductivities of chitosan solutions were considerably higher than those of PEO and increased with increasing chitosan concentration, e.g., conductivities of 1.2% chitosan were 1.34 and 0.25 mS/cm in 50% and 90% acetic acid, respectively.

Table 1 Electrical conductivity in millisiemens per centimeter and surface tension in millinewtons per meter of polymer–surfactant solutions dispersed in 50% acetic acid
Full size table
Table 2 Electrical conductivity in millisiemens per centimeter and surface tension in millinewtons per meter of polymer–surfactant solutions dispersed in 90% acetic acid
Full size table

As PEO was added to the system, conductivity of chitosan–PEO solutions decreased compared to that of pure chitosan solution. This is because the solutions contain less of the polyionic polymer (chitosan) and more of the uncharged PEO. Solutions containing Brij 35 had conductivities ranging between 1.39–1.47 and 0.20–0.26 mS/cm in 50% and 90% acetic acid, respectively. With increasing concentration of Brij 35, the conductivities slightly decreased possibly due to Brij 35 interfering with the electron transport between electrodes. Generally, polymer dispersions containing 50% aqueous acetic acid had significantly higher conductivities compared to polymer dispersions in 90% acetic acid due to increased ionization. Conductivities of solutions containing SDS ranged from 1.43 to 1.63 mS/cm in 50% acetic acid, depending on SDS concentration, and 0.20, 0.22, and 0.38 mS/cm in 90% acetic acid for 1, 2.3, and 10 mM SDS, respectively. Conductivities of DTAB containing polymer dispersions were significantly higher than those containing either Brij or SDS, e.g., conductivities rose from 1.69 to 1.79 and 2.55 mS/cm for 10.9, 14, and 36 mM DTAB, respectively. When DTAB and chitosan–PEO were dispersed in 90% acetic acid, polymers were not completely dissolved; and no measurements could be performed on these samples. Moreover, in neat acetic acid, not enough polymers could be dissolved to allow for electrospinning.

Surface Tension

As acetic acid concentration increased from 50% to 90%, surface tension decreased from 38.6 to 30.8 mN/m without significant changes in viscosity (Tables 1 and 2). Upon addition of surfactant, a decrease in surface tension was observed. In 50% acetic acid, surface tension of samples containing 0.01, 0.09, and 2 mM Brij 35 were 38.7, 38.3, and 37.2 mN/m, respectively, while addition of 10.9, 14, or 36 mM DTAB resulted in a decrease from 38 to 37.7 and 36.9 mN/m, respectively, which was lower than the effect observed for Brij 35. SDS led to the most pronounced decline in surface tension in 50% acetic acid. When the anionic surfactant was used in the polymer blend, surface tensions of 37.7 to 34.8 and 29.0 mN/m were recorded for 1, 2.3, and 10 mM SDS, respectively. In 90% acetic acid, surface tensions were 30.4, 31.7, and 30.9 mN/m in solutions containing 0.01, 0.09, and 2 mM Brij 35, respectively, and 31.9, 30.7 and 31.8 mN/m in solutions containing 1, 2.3, and 10 mM SDS, respectively. It is interesting to note that the effect of acetic acid was more influential on the surface tension and that addition of surfactants only had a slight effect. Overall, the surface tension was significantly lower in samples containing 90% acetic acid compared to solutions containing 50% acetic acid.

Generally, surface tension has been suggested to be one of the most important solution properties in the electrospinning of chitosan.13 Surface tension and conductivity have counteracting effects, that is, the decreased surface tension will result in a lower electric field strength required for jet initiation,14 while the decreased conductivity will require a higher electric field strength. Furthermore, surface tension has a direct influence on the formation of beads and bead defects.18

Viscosity

The apparent viscosity of chitosan at a shear rate of 100 s−1 (η a,100) increased from 0.56 to 1.15 Pa s when the chitosan concentration increased from 1.2% to 1.6% in 50% acetic acid. The solvent itself had an apparent viscosity of 0.01 Pa s. In PEO solutions, an approximately 18-fold increase in apparent viscosity (from 0.01 to 0.18 Pa s) was observed when the polymer concentration was increased from 0.4% to 1.6%. The composite solution of PEO and chitosan had an apparent viscosity of 0.88 Pa s. While the addition of PEO decreased the viscosity of the solution, PEO has been successfully electrospun under a variety of conditions and thus is an ideal cospinning agent known to promote entanglement.18–20

With increasing concentrations of SDS (1, 2.3, and 10 mM) and Brij 35 (0.01, 0.09, and 2 mM), an increase in apparent viscosity from 0.85 to 0.92 Pa s and from 0.84 to 0.89 Pa s for SDS and Brij 35, respectively, of composite polymer solutions in 50% acetic acid was observed (Table 3). Table 3 also shows the results of a fit of the flow curve of polymer solutions measured at shear rates varying from 10−3 to 103 s−1 to the power law (Eq. 1). From the fits, the power law consistency coefficient K and the power law flow behavior index n were determined. K increased from 5.42 ± 0.65 to 13.04 ± 0.32 when the concentration of SDS increased to 10 mM and from 7.81 ± 0.26 to 9.63 ± 0.29 when the concentration of Brij 35 increased to 2 mM. The flow behavior index n decreased significantly from 0.55 ± 0.01 to 0.46 ± 0.01 upon addition of 10 mM SDS. Addition of Brij on the other hand did not significantly impact the flow behavior index n (n = 0.51 ± 0.01 at 0 mM Brij 35 and n = 0.48 ± 0.01 at 2 mM Brij 35). When DTAB was added to the polymer solutions, the apparent viscosity and K were slightly decreased from 0.90 to 0.83 Pa s and from 10.15 ± 0.84 to 8.47 ± 1.33, respectively. The flow behavior index n, however, remained virtually unchanged (0.48 ± 0.01 versus 0.49 ± 0.02).

Table 3 Apparent viscosity at a shear rate of 100 s−1 and power law indices K and n of chitosan–PEO–surfactant dispersions in systems containing 50% acetic acid as the solvent
Full size table

The rheological properties of polymer solutions are indicative of a pronounced shear-thinning effect which can be related to polymer entanglement. Entanglement of polymers in the electrospinning solution is critical to the formation of fibers. If polymers are not or insufficiently entangled, beads or droplets instead of fibers are typically deposited on the collector plate due to the low viscoelastic force which does not counterbalance the higher Coulombic stretching force that causes jet instability.8,12,13,21–26 The critical chain overlap concentration c*, which is the crossover concentration between the dilute and the semidilute concentration regimes, is thus a critical parameter for electrospinning. The overlap concentration can be determined from the intrinsic viscosity as:27

$$ c* \sim {\left[ \eta \right]^{ - 1}}. $$
(3)

Gupta et al. studied the relationship between viscosity, polymer concentration, and fiber formation and found a good correlation between solution regimes and the occurrence of beads and droplets or bead defect-free and uniform fibers in electrospinning of poly(methyl methacrylate).27 For example, polymer droplets together with beaded fibers were observed in the semidilute unentangled regime, at concentrations of 1 < c/c* < 3. However, polymer concentration had to be increased to c ~ 6c* for uniform fiber formation to occur with polymers of a narrow molecular weight distribution (MWD) and even higher concentrations were required in case of broad MWD. Thus, the extent of polymer entanglement depends on the polymer as well as the solvent and polymer concentrations needed to substantially exceed c* for a successful electrospinning experiment and nanofiber formation to occur. The critical chain overlap concentration in our system was previously determined to be 0.103 g/dL. Thus, we were working well above this critical concentration with c/c* being >10.

A similar behavior was observed when 90% acetic acid was used as the solvent. Generally, viscosities of polymer blend solutions were higher using the more concentrated solvent. The apparent viscosity at a shear rate of 100 s−1 (η a,100) increased from 0.61 to 1.15 Pa s in solutions containing 1.2% to 1.6% chitosan, respectively. The solvent had an apparent viscosity of 0.01 Pa s. One hundred percent PEO samples of 0.4% and 1.6% concentration dissolved in 90% acetic acid had apparent viscosities of 0.02 and 0.27 Pa s, respectively, while the composite solution had an apparent viscosity of 1.04 Pa s. Apparent viscosities decreased with increasing concentration of SDS or Brij 35 from 1.17 to 1.07 Pa s and from 1.30 to 0.99 Pa s, respectively. The consistency coefficient remained unchanged at 14.24 ± 2.12 to 14.54 ± 2.80 upon addition of SDS concentration and increased from 13.1 ± 0.11 to 14.50 ± 0.10 when increasing concentrations of Brij 35 were added. The flow behavior index n decreased for SDS-containing samples from 0.45 ± 0.02 to 0.32 ± 0.17 at 10 mM SDS but remained unchanged at 0.45 ± 0.01 and 0.43 ± 0.01 for 0 and 2.3 mM Brij 35, respectively. A flow index of 0.32 in SDS-containing solutions is indicative of a very strong shear-thinning behavior. Results of both viscosity and conductivity measurements highlight the strong influence that SDS appears to have on polymer–polymer interactions (Table 4).

Table 4 Apparent viscosity at a shear rate of 100 s−1 and power law indices K and n of chitosan–PEO–surfactant dispersions in systems containing 90% acetic acid as the solvent
Full size table

Morphology and Fiber Diameter

Pure chitosan did not form fibers in either solvent. Instead, beads or drops with a very limited degree of fiber formation were deposited. Due to the limited solubility of chitosan in common organic solvents, the concentration of chitosan was likely too low to lead to a sufficiently high degree of entanglement required for electrospinning. Jets thus broke up into individual droplets instead of forming continuous fibers.

FESEM images showed that nanofibers had diameters ranging from of 10 to 250 nm. At 0.4 % PEO (the concentration of PEO in the composite solution), the solution was too diluted in either solvent to form continuous jets and only droplets were deposited. However, when the PEO concentration was increased to 1.6%, which equals the total polymer concentration in all the remaining samples, relatively uniform nanofibers were fabricated in 50% acetic acid with virtually no bead defects (Figure 1a). In 90% acetic acid, the fiber surfaces were smooth and fiber similarly lacked defects (Figure 1b). Fiber diameters ranged from 70 to 110 nm and from 60 to 240 nm for PEO fibers spun in 50% and 90% acetic acid, respectively. The conductivity of the solution containing 50% acetic acid was higher compared to that of PEO in 90% acetic acid. The higher conductivity introduces more whipping and bending instabilities that yielded fibers with a smaller diameter and a somewhat rougher surface.

Fig. 1
figure 1figure 1figure 1figure 1

Scanning electron microscopic images (magnification ×10,000) of depositions of electrospun polymer dispersions. a 1.6% PEO electrospun in 50% acetic acid; b 1.6% PEO in 90% acetic acid; c chitosan–PEO in 50% acetic acid; d chitosan–PEO in 90% acetic acid; chitosan–PEO with e 0.01 mM Brij 35, f 0.09 mM Brij 35, and g 2 mM Brij 35 in 90% acetic acid; chitosan–PEO with h 1 mM SDS, i 2.3 mM SDS, and j 10 mM SDS in 90% acetic acid; chitosan–PEO with k 2 mM Brij 35 in 50% acetic acid; chitosan–PEO with l 1 mM SDS in 50% acetic acid; as well as chitosan–PEO with m 10.9 mM DTAB, n 14 mM DTAB, and o 36 mM DTAB in 50% acetic acid

Full size image

When chitosan was added to the PEO solution, the fiber diameter dramatically decreased to 30 to 80 nm in the case of 50% acetic acid and 50 to 170 nm in 90% acetic acid (Figure 1c, d). This may be attributed to an increased chitosan and a decreased PEO fraction in the solution. Amongst the tested acetic acid concentrations, 90% acetic acid appeared to be best suited to initiate electrospinning and form nanofibers without bead defects (Figure 1d–j). Bead defects can be an indication of inconsistent polymer flows. The flow rate and the strength of the applied electric field must both be properly adjusted so as to prevent that the supplied flow rate from the reservoir to the tip of the needle exceeds the flow rate of the polymer jet. Otherwise, excess polymer solution will accumulate at the tip of the capillary and form a droplet that will eventually detach from the tip of the capillary to interrupt the continuous flow of the jet. On the other hand, if too little solution is supplied, the solution at the tip of the capillary will be depleted causing an interruption of the ejection of a polymer jet.

When Brij 35 was added to the polymer solutions, a sufficiently high amount of nonionic surfactant had to be added in order to initiate continuous fiber production. At insufficient concentrations of Brij, numerous bead defects and few small fibers were observed. For example, at 0 mM Brij, numerous bead defects were observed, while at 0.09 mM Brij, the number of bead defects decreased and more fibers with larger diameters were formed. At a concentration of 2 mM Brij 35, smooth nanofibers with diameters of 70 to 120 nm were deposited on the collector plate (Figure 1k). Long and branched fibers were observed, which is indicative of the split of a single fiber during the electrospinning process. Addition of 1 mM SDS yielded nanofibers of 10–60 nm, the smallest size of nanofibers that was collected in these sets of experiments, and some bead defects were observed (Figure 1l). The smaller diameters can be associated with an increased charge density along the polymer jet, which can lead to higher elongational forces stretching the polymer jet. Increasing the concentration of the anionic surfactant to 2.3 mM drastically increased the number and size of bead defects, but at the same time, reduced the amount of fibers and their size. When 10 mM SDS was incorporated in the polymer solutions, no fiber formation was observed and only beads or drops were deposited. Nanofibers were formed from all polymer dispersions supplemented with DTAB independent of the concentration (Figure 1m–o). The average fiber size did not significantly vary among the three samples and fiber diameter ranged from 40 to 130 nm. The changes in the macroscopic appearances of the collected material are especially noteworthy. Nanofibers containing DTAB, the cationic surfactant, formed open, low-density networks comparable to spider webs rather than the dense paper-like deposition observed with all other surfactants.

Using 90% acetic acid, nanofibers were formed in polymer dispersions supplemented with Brij 35 or SDS regardless of the surfactant concentration. However, addition of DTAB led to the inability to dissolve the polymers in the solution; and polymers instead remained as dispersed powders. The viscosity of the solution was similar to that of the solvent, a further indication that polymers did not dissolve. Thus, nanofiber production was not possible at all. However, when the water content was increased (50% acetic acid), polymers could be dissolved and an increase in viscosity and polymer entanglement was seen. Addition of up to 0.09 mM Brij 35 resulted in nanofibers having diameters ranging from 40 to 170 nm, while nanofibers produced from polymer solutions containing 2 mM Brij 35 had a broader diameter (60–250 nm).

Solutions containing 1 mM SDS produced fibers with sizes ranging from 30 to 200 nm. The diameter distribution was very broad, but fibers had very smooth surfaces. Higher concentrations of SDS caused an increase in fiber diameter. When 2.3 mM SDS was added, the diameter ranged from 30 to 220 nm. At 10 mM SDS, fibers had average diameters of 70–200 nm.

Compositional Analysis

Fourier Transform Infrared Spectroscopy

The infrared spectra of cast films and electrospun nanofibers are shown in Figure 2. Spectra of pure chitosan were obtained by analyzing solution cast films prepared in the same solvents used for electrospinning followed by drying overnight under vacuum to obtain a transparent polymer film. This is because chitosan could not be spun into fibers. The most characteristic absorption bands for chitosan consisted of a broad band at 3,400–3,100 cm−1, which can be attributed to N–H and OH···O stretching vibration and intermolecular hydrogen bonding of the polysaccharide molecules,6,28–30 and two strong bands at 1,655 and 1,550 cm−1 attributed to the carbonyl C=O–NHR or amide I band and the amine –NH2 or amide II absorption band, respectively.8,28,29

Fig. 2
figure 2

Infrared absorbance spectra of electrospun chitosan–PEO–surfactant fibers at wave numbers ranging from 4,000 to 700 cm−1. a Spectra of PEO (1), chitosan–PEO film (2), chitosan–PEO nanofibers (3), and chitosan powder (4); b spectra of pure Brij 35 powder (1), composite fibers with 2 mM (2), 0.09 mM (3), and 0.01 mM Brij 35 (4); c spectra of pure SDS powder (1) and composite nanofibers incorporating 10 mM (2), 2.3 mM (3), and 1 mM SDS (4); d spectra of pure surfactant powder of DTAB (1) as well as chitosan–PEO nanofibers containing 36 mM (2), 14 mM (3), and 10.9 mM DTAB (4)

Full size image

Typical absorption bands for PEO can be detected at 2,885 cm−1 attributed to CH2 stretching,31 which can overlap with the band observed in chitosan samples. Other characteristic bands are observed at 1,470 cm−1 assigned to CH bending, at 1,340 cm−1 assigned to CH deformation of the methyl group, and absorption bands around 1,100 cm−1 assigned to C–O–C stretching vibration.8 Since PEO with a high molecular weight (900 kDa) was used in the experiments, the OH absorption bands may be neglected. The OH absorption band in the range of 3,400 and 3,100 cm−1 is thus only attributed to OH within chitosan.32 With the addition of PEO, the absorbance intensity of CH2 stretching at 2,885 cm−1 increased while the absorbance intensity of –NH2 stretching at around 1,550 cm−1 decreased.

PEO and the nonionic surfactant Brij 35 experience similar structural features as evidenced by similar FTIR spectra and absorption bands. Typical bands for DTAB include two strong absorption bands at 2,915 and 2,850 cm−1 attributed to CH2 methylene stretching vibrations and four medium strong peaks in the range of 1,488 to 1,462 cm−1. In case of SDS, characteristic bands are observed at 2,930, 2,847, 1,469, and 1,184 cm−1 assigned to symmetric CH stretching, asymmetric CH stretching, CH bending, and SO stretching, respectively. Generally, with increasing content of surfactant of any kind, the CH2 stretching vibration around 2,900 cm−1 as well as bands characteristic for each specific surfactant increased (in case of DTAB at 1,500, 960, and 842 cm−1) while bands typical for chitosan polysaccharide such as amide bands at 1,655 and 1,550 cm−1 gradually decreased, indicating that the content of chitosan decreased relative to the other components, suggesting that chitosan is less preferential ejected from the polymer surface.

Deconvolution of the FTIR spectra using single-component spectra of pure chitosan, PEO, Brij 35, SDS, and DTAB and calculation of peak areas by integration was used to determine the composition of nanofibers (IR-Solution, Shimadzu, Columbia, MD, USA and Peakfit 4.12, SeaSolve Software, San Jose, CA, USA). Intensities at 1,550 cm−1 were used to calculate relative concentrations of chitosan in the nanodepositions (Figure 3). Chitosan concentration in electrospun nanofibers was 71% when 50% acetic acid was used and 73.9% when 90% acetic acid was used. These data were in close agreement of data obtained from thermogravimetric analysis (data not shown). Here, chitosan concentration was determined to be 75% in 90% acetic acid and 72% in 50% acetic acid. When Brij 35 was added to the solution to be electrospun, chitosan concentration in the fibers decreased to 57.1% using 50% acetic acid and from 71.5 to 21% with increasing surfactant concentration in 90% acetic acid. A similar trend was observed for the other two surfactants. Addition of SDS to the polymer dispersions in 50% acetic acid resulted in a decrease of the chitosan concentration to 53.5% and to 39% chitosan 90% acetic acid at the highest concentration of the anionic surfactant (10 mM). Using the cationic DTAB as an additive, the chitosan concentration decreased from 63.5% to 28% upon an increase in surfactant concentration from 10.9 to 14 and 36 mM. Thus, generally less chitosan was present in electrospun nanofibers upon addition of surfactants, a possible indication of a disruption of the wrapping entanglement of PEO with chitosan.

Fig. 3
figure 3

Compositional analysis of composite nanofibers. a Composition of samples obtained in 50% acetic acid, b composition of samples spun in 90% acetic acid. Concentrations were calculated from analysis of infrared absorbance peaks at 1,550 cm−1

Full size image

Thermal Analysis

DSC thermograms of fibers are shown in Figure 4 with analysis of thermograms shown in Tables 5 and 6. Pure chitosan powder, pure PEO electrospun nanofibers, as well as pure surfactant powders were used as controls in the DSC experiments. The majority of polysaccharides including chitosan do not experience melting but rather degradation upon heating above a certain temperature. This is due to their extensive associations through hydrogen bonding. Thus, rather than showing sharp transitions, their thermograms have broad endotherms with associated water evaporation below the degradation temperature.8

Fig. 4
figure 4

Thermograms of electrospun composite fibers measured at a scanning rate of 2 °C/min from 25 to 180 °C. a Thermograms of chitosan–PEO fibers (3), chitosan powder (2), and PEO fibers (1) in 50% acetic acid; b thermograms of chitosan–PEO fibers (3), chitosan powder (2), and PEO fibers (1) in 90% acetic acid; c thermograms of composite nanofibers containing 2 mM Brij electrospun in 50% acetic acid, d fibers with 1 mM SDS solubilized in 90% acetic acid prior to nanofiber manufacture, and e nanofibers incorporating DTAB at 10.9 mM (1), 14 mM (2), and 36 mM (3); f thermograms of nanofibers with Brij 35 at 0.01 mM (1), 0.09 mM (2), and 2 mM Brij 35 (3) and g composite nanofibers with SDS of 1 mM (1), 2.3 mM (2), and 10 mM SDS (3)

Full size image
Table 5 Heat flow data of nanofibers of chitosan–PEO–surfactant systems in 50% acetic acid
Full size table
Table 6 Heat flow data of nanofibers of chitosan–PEO–surfactant systems in 90% acetic acid
Full size table

In the DSC thermogram of chitosan films prepared with 90% acetic acid, a broad peak was observed at 75.92 ± 1.90 °C. A similar broad peak also appeared in the thermogram of chitosan–PEO at a slightly higher temperature of 78.46 ± 3.50 °C. A second peak was observed in the composite samples from PEO at 60.65 ± 0.43 °C which represents a shift to a lower temperature from its original melting point at 64.71 ± 0.76 °C in samples of pure PEO fibers. The peak temperature determined for the PEO film in 90% acetic acid was 65.27 °C ± 0.26 °C, which is somewhat higher than the melting point for PEO fibers, suggesting that the overall crystallinity of PEO fibers is lower than in cast films. This can be associated with the electrically forced orientation of the polymer chains during electrospinning,20 transforming the coiled polymer chains into oriented entangled structures, which are also present in the solidified electrodepositions. The melting point of PEO is even lower in the composite fibers, possibly due to interactions between PEO and chitosan chains obstructing the crystallization of PEO.

When 50% acetic acid was used to dissolve the polymers, a broad transition peak was observed at 82.23 ± 3.50 °C for pure chitosan film, a sharp and intense peak at 60.09 ± 0.57 °C for melting of pure PEO fiber, and a T m at 52.34 ± 0.25 and 81.93 ± 1.94 °C in the composite fiber. The melting point of PEO in the PEO film was 65.35 ± 0.27 °C. Using 50% acetic acid as the solvent caused an even more pronounced shift towards lower melting temperatures of PEO in both pure and composite fibers, which may correlate to a low crystallinity of PEO in these samples. This observation is of particular importance to possible applications of electrospun nanofibers.

Upon addition of increasing concentrations of the nonionic surfactant Brij 35 using 90% acetic acid as a solvent, the melting point for PEO and chitosan shifted from 53.57 ± 0.74 and 80.59 ± 0.04 °C at the lowest concentration of surfactant (0.01 mM Brij 35) for PEO and chitosan, respectively, to 57.74 ± 0.93 and 81.65 ± 0.32 °C at 0.09 mM Brij 35. A broad transition peak of chitosan was observed in both cases. When 2 mM Brij 35 was present in the polymer dispersion, the endotherm from chitosan decreased significantly compared to the other two endothermic melting transitions at 37.92 ± 0.58 and 51.93 ± 1.20 °C were observed in Brij 35 and PEO, respectively. Pure Brij 35 powder had two melting endotherms at 36.56 ± 0.54 and 139.25 ± 1.60 °C.

Increasing the concentration of the anionic surfactant SDS resulted in a shift of PEO and the broad chitosan endotherms from 57.22 ± 0.84 and 79.43 ± 2.57 °C at 1 mM SDS to 56.63 ± 0.12 and 86.42 ± 1.24 °C at 2.3 mM SDS to 57.16 ± 1.07 and 75.17 ± 3.02 °C at 10 mM SDS. Pure SDS had melt transitions at 99.76 ± 0.10 and 106.35 ± 0.44 °C.

Fibers containing 2 mM Brij 35 electrospun in 50% acetic acid had melting temperatures of 36.56 ± 0.27, 50.76 ± 1.10, as well as 74.56 ± 2.16 °C, which may be attributed to Brij 35, PEO, and chitosan, respectively. Here, the melting point for PEO was reduced compared to fibers spun in 90% acetic acid. In the case of the anionic surfactant, only the lowest concentration of SDS yielded electrodepositions in 50% acetic acid. Here, T m for both components shifted to lower transition temperatures when compared with the system produced in 90% acetic acid, e.g., 54.64 ± 1.10 and 76.26 ± 2.83 °C for PEO and chitosan, respectively.

When the cationic DTAB was added to the test system and 50% acetic acid was used as the solvent, two sharp peaks were present. The melting point of PEO was 56.08 ± 0.89, 54.52 ± 1.20, and 53.41 ± 0.01 °C when the concentration of DTAB was increased to 36 mM. A second peak was observed at 96.48 ± 0.15, 87.05 ± 4.68, and 87.86 ± 1.54 °C at a concentration of 10.9, 14, and 36 mM DTAB, respectively. Pure DTAB had a transition at 99.1 ± 0.31 °C.

Conclusions

Ultrafine fibers could not be formed via electrospinning from chitosan solutions in 50 and 90 wt.% acetic acid. Instead, beads were deposited on the collector plate. However, fibers with diameters ranging from 10 to 240 nm were obtained from composite dispersions at a chitosan–PEO ratio of 3:1. When nonionic, anionic, and cationic surfactants were added to the polymer blend, smooth fibers of various sizes were obtained. Composite solutions containing DTAB at 90 wt.% acetic acid were not electrospinnable due to insufficient solubility. Measurement of solution properties such as viscosity, conductivity, and surface tension revealed that the polymer–solvent and polymer–polymer interactions are key factors in the electrospinning process and results suggest that surfactants are able to modulate these interactions via electrostatic and/or hydrophobic interactions as well as hydrogen bonding.

The obtained results are relevant to the food and agricultural industries due to the fact that they offer a new means to produce nanofibers that are solely or partially composed of food biopolymers and that are, therefore, suitable for use in food or food packaging. Rather than using organic solvents that are not suitable for food applications, food-approved emulsifiers may be used to induce electrospinning in solutions of biopolymer–polymer blend solutions. Food-grade nanofibers could find use in a wide variety of food applications ranging from packaging materials to filtration aids and new ingredients systems due to their unique properties.