© 2020 Elsevier

Electrospinning of Octenylsuccinylated Starch-Pullulan Nanofibers from Aqueous Dispersions

Songnan Li – University of Alabama

Lingyan Kong – University of Alabama Gregory R. Ziegler – Pennsylvania State University

Deposited 08/31/2020

Citation of published version:

Li, S., Kong, L., Ziegler, G. (2020): Electrospinning of Octenylsuccinylated Starch-Pullulan Nanofibers from Aqueous Dispersions. Carbohydrate Polymers.

DOI: https://doi.org/10.1016/j.carbpol.2020.116933

1

1

Electrospinning of Octenylsuccinylated Starch-Pullulan Nanofibers from 2

Aqueous Dispersions 3

4

Songnan Li a,b, Lingyan Kong b,∗, Gregory R. Ziegler a,∗ 5

6

a Department of Food Science, Pennsylvania State University, 341 Food Science Building, 7

University Park, PA, 16802, USA 8

b Department of Human Nutrition and Hospitality Management, The University of Alabama, 9

Tuscaloosa, AL, 35487, USA 10

11

12

13

* Corresponding authors.14

Address: 15

341 Rodney A. Erickson Food Science Building, University Park, PA 16802, USA (G. Ziegler) 16

482 Russell Hall, 504 University Blvd, Tuscaloosa, AL 35487, USA (L. Kong) 17

E-mail address:18

lkong@ches.ua.edu (L. Kong) 19

grz1@psu.edu (G. Ziegler) 20

For published full text, send request to lingyan.kong@ua.edu

2

Abstract 21

We aimed to develop a greener process for dry-electrospinning food-grade modified starch 22

through the elimination of organic solvents. The rheological properties and electrospinnability of 23

aqueous dispersions of commercial octenylsuccinylated (OS) starches with various molecular 24

weight (Mw) were investigated, yet only nanofibers with beads or defects could be obtained from 25

OS starch with the highest Mw, i.e., Purity Gum@ Ultra (PGU). Further improvement in the fiber 26

morphology was achieved by adding pullulan (PUL) as a minor component in the spinning dope. 27

Smooth, continuous, and bead-free nanofibers (147-250 nm) were obtained from the PGU-PUL 28

dispersions. Shown on an electrospinnability map, the successful electrospinning of 12%, 15%, 29

and 20% (w/v) aqueous PGU dispersions required a minimum addition of 6%, 5%, and 3% (w/v) 30

of PUL, respectively. The addition of PUL contributed to establishing sufficient molecular 31

entanglement for electrospinning. This study provides a promising green process to produce 32

starch-based nanofibers for use in various applications, e.g., drug delivery, wound dressing, and 33

tissue engineering. 34

Keywords 35

Modified starch; octenylsuccinylated starch; pullulan; rheology; electrospinning; nanofiber 36

3

1. Introduction 37

Electrospinning, an electrohydrodynamic processing technique, employs a high electrostatic 38

force to stretch a viscoelastic jet derived from a polymeric solution or melt into fine fibers. Both 39

synthetic polymers and natural biopolymers can be fabricated into micro- and nano-scale fibers 40

via electrospinning. Because it is cost-effective, applicable to a large variety of materials, capable 41

of controlling fiber size and morphology, and easily scaled, electrospinning has garnered much 42

attention (Xue, Wu, Dai, & Xia, 2019). A variety of natural biopolymers, such as chitosan, 43

hyaluronic acid, collagen, cellulose, starch, and soy proteins, have been electrospun into micro- 44

and nanofibers with potential uses in food technology, packaging, filtration, controlled drug 45

delivery, tissue engineering, wound dressing, and other biomedical applications (Ameer, Pr, & 46

Kasoju, 2019; Mendes, Stephansen, & Chronakis, 2017). 47

Starch is the second most abundant renewable biopolymer on earth after cellulose and widely 48

used in paper products, bioplastics, engineered products (e.g., composite fibers and films), 49

agricultural products (e.g., water release material and seed carrier), and biomedical products (e.g., 50

in drug delivery and tissue scaffold compositions) (Masina et al., 2017; Sweedman, Tizzotti, 51

Schafer, & Gilbert, 2013). The electrospinning of micro- to nano-scale starch fiber mats with 52

random and aligned orientation has been demonstrated by a number of researchers (Kong & 53

Ziegler, 2012, 2014a; Vasilyev, Vilensky, & Zussman, 2019; Wang, Kong, & Ziegler, 2019a). Kong 54

4

and Ziegler (2012) developed a wet-electrospinning technique using an ethanol coagulation bath 55

to fabricate pure starch fibers from a dimethyl sulfoxide (DMSO)/water solvent medium. There 56

are drawbacks to this technique: the drying step after wet electrospinning is time-consuming (＞6 57

h), the fiber mat was fragile, and the fiber diameter was in the micron range (Kong & Ziegler, 2013, 58

2014a). The plasticizing effect of water during storage (Wang, Kong, & Ziegler, 2018), 59

incorporating nanocellulose-cationic starch (Wang, Kong, & Ziegler, 2019b), and achieving fiber 60

alignment with a rotating drum collector (Wang et al., 2019a) were all found to improve the tensile 61

strength of starch fiber mats by a modest amount. Starch-based nanofibers (~146 nm) were 62

electrospun from aqueous solvent with the addition of sodium palmitate and pullulan (PUL) (Wang 63

& Ziegler, 2019). However, this was still a wet-spinning process. 64

Lancuski, Vasilyev, Putaux, and Zussman (2015) used aqueous formic acid to dissolve starch 65

and obtained starch-formate fibers with diameters ranging from 80 to 300 nm using a dry-66

electrospinning process. Though formic acid can chemically gelatinize starch granules, it may also 67

decrease the molecular weight and cause the formylation of starch (Lancuski et al., 2015). 68

Furthermore, there were kinetic constraints, as the solutions were spinnable for a short window in 69

time only (Lancuski et al., 2015). Current techniques relying on organic solvents may limit the 70

commercialization and utilization of starch fibers, especially for food and biomedical applications. 71

Dry-electrospinning of starch nanofibers using green solvents, e.g., aqueous solvents, would lower 72

the cost of chemicals and waste treatment, minimize chemical residues, and ensure non-irritability 73

5

and safety. 74

Native starches possess poor solubility in aqueous solvents resulting from strong hydrogen 75

bonding between starch chains (Zhang et al., 2017). Various modification techniques (i.e., physical, 76

chemical, enzymatic, genetic, and blended treatments) can alter the gelatinization, swelling, 77

solubility, pasting and retrogradation characteristics of starch (Masina et al., 2017). Among 78

modified starches, octenylsuccinate (OS) starch, i.e., starch after chemical modification with 79

octenyl succinic anhydride, has been used as a high-performance, cost-effective replacement for 80

gum Arabic in food applications. Most commercial OS starches are derived from waxy maize (e.g., 81

Purity GumTM, CAPSULTM, and Hi-CAPTM) or normal maize (e.g., DRYFLOTM) starches through 82

octenylscuccinate substitution and enzymatic hydrolysis (Sweedman, Tizzotti, Schafer, & Gilbert, 83

2013). OS starches have better water solubility or dispersibility making them suitable for 84

formulations such as liquid beverage emulsions. Here we used OS starch, a GRAS (Generally 85

Recognized as Safe) additive in food and cosmetics, to obtain a relatively stable dispersion of 86

starch in water at concentrations suitable for electrospinning. 87

PUL was added to aid the electrospinning of OS starch. PUL is an extracellular and water-88

soluble microbial polysaccharide produced by Aureobasidium pullulans, and currently exploited 89

in food and pharmaceutical industries due to its unique characteristics (non-ionic, non-toxic, non-90

immunogenic and non-carcinogenic) (Prajapati, Jani, & Khanda, 2013). The unique glycoside 91

linkages of α-1,6 and α-1,4 in PUL endow this polymer with distinctive physical traits, including 92

6

adhesive properties and the ability to form films and fibers. Electrospun PUL fiber mats with and 93

without other polymers or additives have been reported (Karim et al., 2009; Kong & Ziegler, 2014b; 94

Li et al., 2017). PUL addition made the wet-electrospinning of nanofibers from aqueous 95

dispersions of high-amylose starch (HAS) sodium palmitate inclusion complexes feasible (Wang 96

& Ziegler, 2019). 97

Based on our previous work electrospinning starch-sodium palmitate inclusion complexes, we 98

hypothesized that OS starch would be electrospinnable but that this would depend on the degree 99

of substitution and molecular weight. In this study, we attempted to dry-electrospin commercial 100

OS starches of various molecular weight (Mw) from their aqueous dispersions, and further 101

improve the nanofiber morphology by blending with PUL. Both sufficient molecular entanglement 102

and shear viscosity were found to be determining factors dictating the electrospinnability. An 103

electrospinnability map as a function of OS starch and PUL concentration was obtained. Nano-104

sized OS starch-PUL fibers were successfully fabricated by dry-electrospinning and could be a 105

promising nanofiber material for various applications in the food, pharmaceutical, cosmetic, and 106

biomedical fields. 107

2. Materials and methods 108

2.1. Materials 109

Three commercial OS starches, including Purity Gum@ Ultra (PGU), Purity Gum@ 2000 110

7

(PG2000), and Hi-Cap@ 100 (HC100), were kindly supplied by Ingredion (Bridgewater, NJ). Food 111

grade pullulan (PUL) was purchased from Hayashibara Biochemical Laboratories Inc. (Okayama, 112

Japan). Deionized water (DI water) was used to prepare all aqueous solutions. 113

Physical characteristics of the starches and PUL, including viscosity average molecular weight 114

(Mw), degree of substitution (DS), and moisture content (MC), are summarized in Table 1. The 115

viscosity average molecular weight of the OS starch samples was determined by viscometric 116

measurements and the Mark’s variation of Staudinger’s formula (Dokić, Dokić, Dapčević, & 117

Krstonošić, 2008; Kerr, Cleveland, & Katzbeck, 1951). The DS of OS starches was determined 118

using excess alkali saponification and back titration with HCl, according to Chiu et al. (2017). The 119

MC of OS starch and PUL samples were determined using an air-oven method (AACC method 120

44-15.02) (AACC International, 1999). Measurements were made in triplicate and data are 121

reported as mean ± standard deviation. 122

Table 1. Physical characteristics of biopolymer samples. 123

Samples PGU PG2000 HC100 PUL

Mw (×104 g/mol) a 12.4 3.8 2.3 10~20

DS (×10-2) b 0.16 ± 0.04 4.72 ± 0.10 4.77 ± 0.02 n.a.

MC (%) c 5.45 ± 0.13 5.58 ± 0.16 4.61 ± 0.10 6.15 ± 0.21

a Mw, molecular weight; Mw of PUL was obtained from manufacturer’s datasheets. 124

b DS, degree of substitution. 125

c MC, moisture content. 126

8

2.2. Dope preparation 127

To prepare the spinning dopes, an appropriate amount (0.1-30%, w/v) of OS starch, PUL or 128

their mixtures were dispersed into DI water. These dispersions were heated in a boiling water 129

bath with continuous stirring on a magnetic stirrer hotplate for 1 h, and then allowed to cool to 130

room temperature (20 °C). The surface tension of the dispersions was measured via an automated 131

tensiometer (Rame-hart model 260, Succasunna, NJ, USA) using pendant drop mode. The 132

conductivity of the dispersions was measured using a calibrated EXTECH conductivity meter 133

(EC100, FLIR Systems, Inc., Waltham, MA, USA). Measurements were made in triplicate and 134

data are reported as mean ± standard deviation. 135

2.3. Rheology 136

The apparent shear viscosity of spinning dopes differing in polymer concentration was 137

measured using a strain-controlled rheometer (ARES, TA Instrument, New Castle, DE, USA) with 138

cone and plate geometry (50 mm), a gap of 0.043 mm, and cone angle of 0.04 radians. Flow curves, 139

i.e., apparent shear viscosity versus shear rate increasing from 0.1 to 100 s-1, were collected at 140

20 °C. 141

In this study, the relative viscosity at 100 s-1, close to the shear rate of the electrospinning 142

process, was plotted against polymer concentration to illustrate the scaling relationship. The log-143

log plot of relative viscosity, ηr (Eq. (1)) of each aqueous dispersion at 100 s-1, against polymer 144

9

concentration was fitted by linear equations to calculate the overlap and entanglement 145

concentrations. 146

𝜂𝑟 =𝜂

𝜂𝑠 (1) 147

Where ηs is the measured viscosity of DI water and η is the shear viscosity of polymer 148

dispersions at 100 s-1, which is the approximate shear rate at the tip of spinneret (Greiner & 149

Wendorff, 2008). The overlap concentration (C*) and entanglement concentration (Ce) were taken 150

as those concentrations where abrupt changes in the power law exponent for the concentration 151

dependence of viscosity were observed. 152

2.4. Electrospinning 153

The above spinning dopes were loaded into 12 mL syringes (Becton, Dickinson and Company, 154

Franklin Lakes, NJ, USA) with 22-gauge blunt needles (Hamilton Company, Reno, NV, USA) as 155

the spinneret. The electrospinning apparatus comprised a high voltage power supply (ES40P, 156

Gamma High Voltage Research, Inc., Ormond Beach, FL), a syringe pump (81,620, Hamilton 157

Company, Reno, NV), and a piece of grounded aluminum foil as the collector. Electrospinning of 158

each OS starch dispersion was conducted by varying three spinning parameters (flow rate, voltage, 159

and spinning distance) until the onset of a Taylor cone was observed at the tip of the spinneret. The 160

electrospinning process was performed at room temperature (20 °C) in an enclosed Plexiglas box 161

(Kong & Ziegler, 2014a; Wang & Ziegler, 2019). The electrospinnability was determined by visual 162

10

inspection, and the resulting mat was collected for further analysis. 163

2.5. Characterization 164

Observation of as-spun fiber mats was performed using a Phenom G2 Pro scanning electron 165

microscope (SEM, Phenom-World, Eindhoven, the Netherlands) at an accelerating voltage of 5 166

keV. The samples were not coated with metal, the support was a standard SEM aluminum stub and 167

the sample was adhered to the stub using double-sided carbon tape. Open software, Image J 168

(National Institutes of Health, Bethesda, MD), was used for analyzing fiber diameters from the 169

SEM images (Angel, Guo, Yan, Wang, & Kong, 2020). Five images were used for each sample 170

and at least 150 different segments were randomly measured. 171

3. Results and discussion 172

3.1. Rheological properties and electrospinnability of OS starches 173

Previous studies have revealed that rheological properties of a polymer dispersion are crucial 174

for determining its electrospinnability. For a polymer dispersion to form a coherent viscoelastic jet 175

and be elongated by the electrostatic force, sufficient molecular entanglements need to be 176

established to resist the breakup of the jet. Meanwhile, the resistance to deformation, i.e., shear 177

viscosity, should be within an appropriate range for sufficient stretching of the jet that leads to a 178

reduction in fiber diameter. To explore the electrospinnability of OS starches, the rheological 179

11

properties of their aqueous dispersions were characterized. 180

Flow curves of OS starch (HC100, PG2000 and PGU) aqueous dispersions varying in 181

concentration (0.1-30%, w/v) are shown in Figure S1. Unreliable data, i.e., data lower than the 182

detection limit of the rheometer, were not plotted. 183

Both relative viscosity (ηr) and specific viscosity have been used to distinguish unentangled 184

from entangled regimes (Kong & Ziegler, 2014b; Wang & Ziegler, 2019). Here log ηr at 100 s-1 185

was plotted against log starch concentration (Figure 1A-C). Three regions of concentration 186

dependence can be clearly distinguished: the dilute regime, semidilute unentangled regime, and 187

semidilute entangled regime, respectively. In the dilute regime, the concentration dependences of 188

the three OS starches were all very low, i.e., 0.05, 0.11, and 0.17, for HC100, PG2000 and PGU, 189

respectively, indicating that no polymer chain overlap occurred. Beyond the overlap concentration 190

(C*), i.e., in the semidilute unentangled regime, the overlap is insufficient to cause any significant 191

degree of entanglement (Colby, 2009). The concentration dependencies in the semidilute 192

unentangled regime were 0.64, 1.03, and 1.06 for HC100, PG2000, and PGU, respectively. These 193

values were lower than theoretically predicted (~c1.25) (McKee, Wilkes, Colby, & Long, 2004) or 194

observed for neutral linear polymers (de Gennes, 1979), including HAS in pure DMSO (ηsp~ c1.40) 195

(Kong & Ziegler, 2012), chitosan in 80% aqueous acetic acid solution (ηsp~ c1.30) (Klossner, Queen, 196

Coughlin, & Krause, 2008), and for branched polymers, e.g., poly (ethylene terephthalateco-197

ethylene isophthalate) in chloroform/dimethylformamide mixture solvent (ηsp~ c1.39) (McKee et 198

12

al., 2004). This suggests that the OS starches, which were modified from waxy maize starch (i.e., 199

highly branched amylopectin), had a less extended or a rather spherical conformation in the 200

aqueous dispersions at low concentrations (Sweedman et al., 2013). 201

In the semidilute entangled regime, the concentration dependencies were 1.70, 2.25, and 2.65 202

for HC100, PG2000, and PGU, respectively. These values were higher than the reported value (ηr~ 203

c1.60) for HAS-palmitate-PUL in DI water (Wang & Ziegler, 2019), comparable to that of starches 204

with various amylose content in 95% (v/v) DMSO aqueous solution (ηr~ c2.14-2.87) observed by 205

Kong and Ziegler (2012), and lower than the theoretical predictions for linear polymers in a good 206

solvent (ηsp~ c4.80) (de Gennes, 1979). This implies that although the OS starches begin to interact 207

above the entanglement concentration, they could not entangle as sufficiently as linear polymers. 208

In both the semidilute unentangled and semidilute entangled regimes, the concentration 209

dependence increased with the relative molecular weight. 210

Entanglement concentrations (Ce) of OS starches in aqueous dispersions were obtained as the 211

intercept of the fitted lines of the semidilute unentangled and semidilute entangled regimes (Figure 212

1A-C). The Ce values were determined to be 12.68% (w/v), 12.36% (w/v), and 6.43% (w/v) for 213

HC100, PG2000, and PGU, respectively. As would be expected, these values decreased with the 214

increasing relative molecular weight of the OS starches (HC100＜PG2000＜PGU). Higher 215

molecular weight polymers may occupy greater hydrodynamic volume and thus require a lower 216

concentration to entangle (Gupta, Elkins, Long, & Wilkes, 2005). Moreover, these results were far 217

13

higher than those of HAS palmitate-PUL in DI water (3.63%, w/v) (Wang & Ziegler, 2019) and 218

HAS in aqueous DMSO (3.61%, w/v) (Kong & Ziegler, 2012). It appears that in DI water these 219

OS starches need to be more concentrated to establish significant molecular entanglement as might 220

be expected for a branched conformation. 221

In addition to the fulfillment of molecular entanglement, the apparent viscosity of the polymer 222

dispersion at a high shear rate must fall into an appropriate range for electrospinnability. According 223

to Kong and Ziegler (2012), the electrospinnable viscosity range (at 100 s-1) for pure HAS in 224

aqueous DMSO dispersions was 0.2-2.2 Pa·s. Figure 1D presents the shear viscosities of OS starch 225

aqueous dispersions at 100 s-1 as a function of starch concentration (0.1-30%, w/v). The shear 226

viscosities at the same concentration followed the order of PGU＞PG 2000＞HC100, which would 227

be expected from their relative molecular weights. Polymers of higher molecular weight would 228

possess a larger surface area, facilitating the entanglement of polymer chains, and thus greater 229

intermolecular friction and resistance to flow (Colby, Fetters, & Graessley, 1987). PGU of 8-20% 230

(w/v) and PG2000 of 20-30% (w/v) demonstrated shear viscosities at 100 s-1 in the range from 0.2-231

2.2 Pa·s, suggesting possible working conditions. 232

14

233

Figure 1. Plots of relative viscosity (A-C: HC100, PG2000 and PGU) and shear viscosity at 100 s-234 1 (D) of OS starch aqueous dispersions as a function of starch concentration (0.1-30%). The overlap 235

concentration (C*) and entanglement concentration (Ce) and slopes of fitted lines in three regimes 236

are illustrated. 237

Electrospinning of PGU dispersions (15-30%, w/v) was attempted using predetermined 238

parameters (fixed spinning distance of 5~10 cm, feed rate of 0.2~0.6 mL/h, and voltage of 15~25 239

kV), and the products are shown in Figure S2. OS starch nanofibers with beads start to form from 240

30% (w/v) PGU, that is the onset concentration for successful electrospinning of PGU aqueous 241

dispersions was at least 4.7 times of its Ce (6.43%, w/v, as shown in Figure 1C). This value is 242

higher than the ranges observed for HAS in DMSO/water solutions (1.2-2.7 times), PUL (1.9-2.3 243

15

times), and HAS-palmitate-PUL (2.8 times) aqueous solutions (Kong & Ziegler, 2012, 2014b; 244

Wang & Ziegler, 2019). This phenomenon could be explained by the insufficient chain 245

entanglement between molecular chains of the OS starch derived from waxy maize starch 246

(Sweedman et al., 2013). In contrast, HC100 and PG2000 dispersions could not be spun into fibers 247

even at 30% (w/v), but could only produce nano- and micro-size particles (Figure S3). This could 248

be attributed to their low shear viscosity or much higher Ce. During the electrospinning process, 249

with insufficient molecular entanglement and low viscosity, the jets of polymer dispersion are 250

unstable and tend to break up and form droplets, resulting in electrosprayed particles (Niu, Shao, 251

Luo, & Sun, 2020). Gupta et al. (2005) investigated the electrospinnability of poly (methyl 252

methacrylate) with Mw ranging from 12,470 to 365,700 g/mol, and found that in the semidilute 253

unentangled regime, polymer droplets were formed at lower Mw (12,470 and 17,710 g/mol), some 254

limited fiber formation connected by polymer droplets for increased Mw (125,900 g/mol), beaded 255

fibers for higher Mw (205,800 and 365,700 g/mol), while uniform and bead-free fibers were 256

observed for all the samples in the semi-dilute highly entangled regime. Tao and Shivkumar (2007) 257

reported that at constant concentration the structure of electrospun poly (vinyl alcohol) changed 258

from beads, to beaded fibers, to complete fibers and to flat ribbons as the Mw increased from 9,500 259

g/mol to 155,000 g/mol. 260

In addition to the dope solution, other electrospinning parameters, including voltage, flow rate 261

and distance, are also key factors that have significant influence on the as-spun fiber morphology 262

16

(Jafari, 2017). Further trials with 30% of PGU were carried out by varying the spinning distance 263

(5-11 cm), voltage (21-26 kV), and flow rate (0.2-1 mL/h), and their results are shown in Figure 264

S4. Even the optimized electrospinning parameters (distance of 9 cm, voltage of 24 kV and flow 265

rate of 0.2 mL/h) could not produce bead-free fibers (Figure 2). The result suggested that sufficient 266

chain entanglement in the polymer solution is necessary for electrospinning fibers, otherwise 267

electrospraying will take place, and instead of fibers, particles, beads, or droplets are formed on 268

the collector. Fong, Chun, and Reneker (1999) reported that the formation of beaded nanofibers 269

results from capillary breakup of the electrospinning jets by surface tension, altered by the presence 270

of electrical forces. In addition, starches containing high amylopectin content (≥75%) were not 271

electrospinnable at any concentration in 95% DMSO, despite their high Mw, due to the molecular 272

conformation of amylopectin, a dense spheroid, resulting from a high degree of branching (Kong 273

& Ziegler, 2012). These results implied that sufficient chain entanglement in the polymer 274

dispersion plays a more important role than Mw per se in the electrospinning process. Approaches 275

for electrospinning OS starch-based fibers by increasing chain entanglement might include further 276

increasing polymer concentration, decreasing the surface tension by adding salts (Li et al., 2017), 277

or blending with other water-soluble polymers, such as PUL (Wang & Ziegler, 2019) and 278

hyaluronic acid (Sun, Perry, & Schiffman, 2019). 279

17

Figure 2. SEM image of electrospun OS starch nanofibers produced from 30% of PGU aqueous

dispersion with optimized electrospinning parameters (spinning distance of 9 cm, voltage of 24

kV, and flow rate of 0.2 mL/h).

3.2. Electrospinnability of OS starch with PUL 280

PUL is an excellent fiber-forming polymer and can be added to enhance molecular 281

entanglement and thus aid in electrospinning of a non-fiber-forming polymer (Aceituno-Medina, 282

Mendoza, Lagaron, & López-Rubio, 2013; Blanco-Padilla, López-Rubio, Loarca-Pina, Gómez-283

Mascaraque, & Mendoza, 2015; Drosou, Krokida, & Biliaderis, 2018; Wang & Ziegler, 2019; Yang, 284

Xie, Liu, Kong, & Wang, 2020). Flow curves of aqueous PUL dispersions as a function of 285

concentration (0.1-30%, w/v) and its entanglement concentration determination are shown in 286

Figure S5. The Ce of PUL was determined to be 6.08% (w/v), lower than that of PGU (6.43%, 287

w/v). Successful electrospinning of PUL in water requires a minimum concentration of 14% (w/v) 288

(Figure S6), where fair fibers with slight beading were formed. The transition of beaded-fiber to 289

bead-free fiber for electrospinning of PUL was observed during the increase of PUL concentration 290

18

from 8% to 15% in aqueous dispersions (Li et al., 2017). The minimum electrospinnable 291

concentration of PUL is 2.3 times Ce, lower than that of PGU (4.7 times), despite its lower 292

molecular weight (1-2 ×105 g/mol). The more linear and flexible chain structure of PUL facilitated 293

molecular entanglement. 294

Aqueous dispersions of PUL (1-13%, w/v) and OS starch (12%, 15%, 20%, w/v) were blended, 295

tested for shear viscosity, and electrospinning attempted. Figure 3 shows the SEM images of 296

electrospun PGU-PUL mats at varying polymer concentrations. The addition of PUL enabled the 297

formation of smooth, continuous, randomly oriented, and bead-free nanofibers. The lowest PUL 298

addition levels for achieving good nanofibers from 12%, 15%, and 20% (w/v) PGU were 6%, 5%, 299

and 3%, respectively. This result indicated that a relatively small addition of PUL played a vital 300

role in establishing sufficient molecular entanglement for electrospinning. A similar effect was 301

seen in our previous study on the electrospinning of aqueous HAS-palmitate-PUL dispersions 302

(Wang & Ziegler, 2019). PUL could hinder starch association, modify dispersion properties (e.g., 303

reduced conductivity and decreased shear viscosity), as well as promote molecular entanglement, 304

which were suggested to be responsible for the successful electrospinning.305

19

Figure 3. SEM images of electrospun PGU-PUL nanofibers from varying polymer concentrations.

20

Table 2. Properties of OS starch and spinnable PGU-PUL dispersions at room 306

temperature (≈20 °C). 307

Samples Shear viscosity at 100 s-1

(Pa·s)

Conductivity

(uS/cm)

Surface tension

(nN/m)

20% PGU 2.07 1712 ± 29 57.81 ± 0.62

20% PG2000 0.20 2519 ± 26 34.56 ± 0.14

20% HC100 0.06 2498 ± 28 36.04 ± 0.62

12%PGU+6%PUL 1.52 1103 ± 22 59.99 ± 0.38

15%PGU+5%PUL 1.77 1307 ± 25 55.36 ± 0.54

20%PGU+3%PUL 3.02 1581 ± 22 55.86 ± 1.20

308

From these electrospinning attempts, a map illustrating regions of 309

electrospinnability at varying concentrations of PGU and PUL was constructed (Figure 310

4A). PGU-PUL dispersions with good fiber forming ability are marked above the blue 311

dotted line. When reading this map, it is important to note that the electrospinnable 312

region was determined using the parameter ranges defined in this study, i.e., feed rate 313

from 0.1 to 0.4 mL/h, spinning distance from 5 to 10 cm, and voltage from 5 to 15 kV. 314

Expanding the parameter ranges, if possible, could alter the electrospinnable area. It is 315

also worth noting that the voltage applied for PGU-PUL aqueous dispersions (5-15 kV) 316

was lower as compared to that of PGU (21-26 kV), implying a more stable jet was 317

formed at a lower voltage. The shear viscosities at 100 s-1 of these PGU-PUL 318

dispersions were plotted in Figure 4B as a function of polymer concentration. The 319

viscosity of PGU dispersions increased with the increasing PUL addition. 320

Electrospinnable dispersions were above the blue dotted line. When PGU concentration 321

increased from 12% to 20%, the lowest viscosity required for the PGU-PUL dispersions 322

increased from 1.52 to 3.02 Pa·s (Table 2). The shear viscosity of all electrospinnable 323

dispersions was below 7.50 Pa·s. A conclusion could be drawn, from a rheological point 324

21

of view, that molecular entanglement and shear viscosity together influenced the 325

electrospinnability of PGU-PUL dispersions, which is consistent with our previous 326

findings (Wang & Ziegler, 2019). 327

328

Figure 4. Evaluation of electrospinnability (A) and apparent viscosity at 100 s-1 (B) of 329

PGU-PUL dispersions as a function of polymer concentration 330

3.3. Effect of polymer concentration on nanofiber diameter 331

Figure 5 shows SEM images with diameter histograms of fibers electrospun from 332

12-20% (w/v) PGU blended with a fixed PUL concentration of 12% (w/v), i.e. below 333

that where PUL alone is spinnable. Continuous smooth fibers were obtained from all 334

PGU-PUL dispersions. As PGU concentration increased from 12% to 20%, the mean 335

diameter of electrospun PGU-PUL fibers increased from 150±34 nm to 250±41 nm. 336

22

Figure 6 presents SEM images with diameter histograms of fibers electrospun from 15% 337

(w/v) PGU with varying PUL concentrations (8-13%, w/v). There was also an 338

increasing trend in mean fiber diameter from 147±26 nm to 209±57 nm with increasing 339

PUL concentration which is expected since more total polymer would lead to thicker 340

fibers. Previously reported diameters of HAS fibers were around a few micrometers by 341

wet-electrospinning from aqueous DMSO solutions (Kong & Ziegler, 2014a) and 80 to 342

300 nm by dry-electrospinning from formic acid solutions (Lancuski et al., 2015). The 343

diameters of PUL fibers from wet-electrospinning and dry-electrospinning were 300 344

nm-15 μm (Kong & Ziegler, 2014b) and 100-700 nm (Li et al., 2017; Sun, Jia, Kang, 345

Cheng, & Li, 2013), respectively. Starch-based fibers at nanoscale (~146 nm) were 346

obtained from aqueous HAS-palmitate-PUL dispersions by wet-electrospinning (Wang 347

& Ziegler, 2019). 348

23

Figure 5. SEM images and diameter histograms of fibers electrospun from 12-20% (w/v) 349

PGU blended with a fixed PUL concentration of 12% (w/v). 350

100 200 300 400 5000

10

20

30

40

5012%PGU-12%PUL

Fre

qu

ency

(%

)

Diameter (nm)

Mean diameter: 150±34 nm

100 200 300 400 5000

10

20

30

40

5015%PGU-12%PUL

Fre

qu

ency

(%

)

Diameter (nm)

Mean diameter: 197±44 nm

100 200 300 400 5000

10

20

30

40

5020%PGU-12%PUL

Fre

qu

ency

(%

)

Diameter (nm)

Mean diameter: 250±41 nm

24

Figure 6. SEM images and diameter histograms of fibers electrospun from 15% (w/v) 351

PGU with varying PUL concentrations (8-13%, w/v). 352

353

Conclusion 354

In this work, three types of commercial OS starches were studied for the 355

relationship between rheological properties in aqueous dispersions and their 356

electrospinnability. The OS starches were all modified from waxy maize starch and 357

their relative Mw followed the order of PGU＞PG2000＞HC100. Their concentration 358

dependences follow the same order in the semidilute unentangled and entangled 359

100 200 300 400 5000

10

20

30

40

5015%PGU-8%PUL

Fre

qu

ency

(%

)

Diameter (nm)

Mean diameter: 147±26 nm

100 200 300 400 5000

10

20

30

40

5015%PGU-10%PUL

Fre

qu

ency

(%

)

Diameter (nm)

Mean diameter: 181±37 nm

100 200 300 400 5000

10

20

30

40

5015%PGU-13%PUL

Fre

qu

ency

(%

)

Diameter (nm)

Mean diameter: 209±57 nm

25

regimes, while their entanglement concentration was inversely related to apparent Mw 360

(Table 3). 361

Table 3. Rheological characteristics of biopolymer samples 362

Sample Exponent n in relationship ηr∝Cn

C* Ce

C<C* C*<C<Ce Ce<C

HC100 0.05 0.64 1.70 3.39 12.68

PG2000 0.11 1.03 2.25 3.19 12.36

PGU 0.17 1.06 2.65 0.90 6.43

PUL 0.11 1.07 3.40 0.95 6.08

C*: overlap concentration; Ce: entanglement concentration. 363

364

The shear viscosity at 100 s-1 was also positively related with apparent Mw. 365

Electrospinning of OS starch dispersions at varying concentrations up to 30% (w/v) was 366

attempted, yet only the PGU dispersions of the highest concentration showed a 367

promising hint of nanofiber formation. The HC100 and PG2000 dispersions were not 368

electrospinnable, instead, micro to nano-scale particles were obtained. Further 369

optimizing electrospinning parameters (distance, voltage, and flow rate) for the 30% 370

(w/v) PGU dispersion could not produce bead-free and bead-free nanofibers, and thus 371

PUL was added to aid in molecular entanglement. Additions of a relatively small 372

amount of PUL into PGU dispersions were able to create continuous, smooth, and bead-373

free nanofibers. Electrospinnable PGU-PUL dispersions were demonstrated on an 374

electrospinnability map, and their shear viscosity at 100 s-1 fell into a range from 1.52 375

to 7.50 Pa·s. The mean diameter of PGU-PUL nanofibers obtained ranged from 147 to 376

250 nm in this study and was affected by total polymer concentration. This study 377

successfully fabricated nanofibers from food-grade starch and PUL with a greener 378

electrospinning process than using DMSO solvent or formic acid solution. Further 379

studies are suggested to focus on the physicochemical properties of the starch-based 380

26

nanofibers and the utilization of the nanofibers. The nanofibers obtained showed bead-381

free morphology and may find potential uses in food, pharmaceutical, cosmetic, and 382

biomedical applications. 383

Acknowledgments 384

This work was supported by the USDA National Institute of Food and Agriculture 385

Federal Appropriations under Project PEN04708. China Scholarship Council is 386

acknowledged for sponsoring Songnan Li as a visiting student (No. 201906150085) at 387

the Pennsylvania State University and the University of Alabama. 388

Declarations of interest 389

None 390

Supplementary material 391

Figure S1. Flow curves of OS starches (HC100, PG2000, and PGU) as a function of 392

starch concentration (0.1-30%, w/v). 393

Figure S2. SEM images of electrospun OS starch nanofibers produced from 15-30% 394

(w/v) of PGU aqueous dispersions with fixed spinning distance of 5~10 cm, feed rate 395

of 0.2~0.6 mL/h, and voltage of 15~25 kV. 396

Figure S3. SEM images of electrospun OS starch particles produced from 30% (w/v) 397

of HC100 and PG2000 aqueous dispersions with fixed spinning distance of 5~10 cm, 398

feed rate of 0.2~0.6 mL/h, and voltage of 15~25 kV. 399

Figure S4. SEM images of electrospun OS starch nanofibers produced from 30% (w/v) 400

of PGU aqueous dispersion with various electrospinning parameters (A1-4: distance of 401

5, 7, 9, and 11 cm with fixed feed rate of 0.2 mL/h and voltage of 21 kV; B1-4: voltage 402

of 21, 22, 24, and 26 kV with fixed feed rate of 0.2 mL/h and distance of 9 cm; C1-4: 403

feed rate of 0.2, 0.4, 0.75, and 1 mL/h with fixed distance of 24 kV). 404

Figure S5. Flow curve and plot of specific viscosity of PUL aqueous dispersions as a 405

function of concentration (0.1-30%, w/v). 406

Figure S6. SEM images of electrospun nanofibers produced from 10-15% (w/v) of PUL 407

aqueous dispersions with fixed spinning distance of 5~10 cm, feed rate of 0.2~0.6 mL/h, 408

and voltage of 5~15 kV. 409

27

Reference 410

AACC International (1999). Method 44-15.02: moisture-air-oven methods. Approved 411

methods of analysis. 412

Aceituno-Medina, M., Mendoza, S., Lagaron, J. M., & López-Rubio, A. (2013). 413

Development and characterization of food-grade electrospun fibers from 414

amaranth protein and pullulan blends. Food research international, 54(1), 667-415

674. 416

Ameer, J. M., Pr, A. K., & Kasoju, N. (2019). Strategies to Tune Electrospun Scaffold 417

Porosity for Effective Cell Response in Tissue Engineering. Journal of 418

Functional Biomaterials, 10(3). 419

Angel, N., Guo, L., Yan, F., Wang, H., & Kong, L. (2020). Effect of processing 420

parameters on the electrospinning of cellulose acetate studied by response 421

surface methodology. Journal of Agriculture and Food Research, 2, 100015. 422

Blanco-Padilla, A., López-Rubio, A., Loarca-Pina, G., Gómez-Mascaraque, L. G., & 423

Mendoza, S. (2015). Characterization, release and antioxidant activity of 424

curcumin-loaded amaranth-pullulan electrospun fibers. LWT-Food Science and 425

Technology, 63(2), 1137-1144. 426

Chiu, N., Tarrega, A., Parmenter, C., Hewson, L., Wolf, B., & Fisk, I. D. (2017). 427

Optimisation of octinyl succinic anhydride starch stablised w1/o/w2 emulsions 428

for oral destablisation of encapsulated salt and enhanced saltiness. Food 429

Hydrocolloids, 69, 450-458. 430

Colby, R. H. (2009). Structure and linear viscoelasticity of flexible polymer solutions: 431

comparison of polyelectrolyte and neutral polymer solutions. Rheologica Acta, 432

49(5), 425-442. 433

Colby, R. H., Fetters, L. J., & Graessley, W. W. (1987). The melt viscosity-molecular 434

weight relationship for linear polymers. Macromolecules, 20(9), 2226-2237. 435

de Gennes, P. (1979). Scaling Concepts in Polymer Physics. Ithaca, NY: Cornell 436

University Press. 437

Dokić, P., Dokić, L., Dapčević, T., & Krstonošić, V. (2008). Colloid characteristics and 438

emulsifying properties of OSA starches. In Colloids for nano-and biotechnology 439

(pp. 48-56). Springer, Berlin, Heidelberg. 440

Drosou, C., Krokida, M., & Biliaderis, C. G. (2018). Composite pullulan-whey protein 441

nanofibers made by electrospinning: Impact of process parameters on fiber 442

morphology and physical properties. Food Hydrocolloids, 77, 726-735. 443

Fong, H., Chun, I., & Reneker, D. H. (1999). Beaded nanofibers formed during 444

electrospinning. Polymer, 40(16), 4585-4592. 445

Greiner, A., & Wendorff, J. H. (2008). Functional self-assembled nanofibers by 446

electrospinning. In Self-Assembled Nanomaterials I . Springer, Berlin, 447

Heidelberg, pp 107−171. 448

Gupta, P., Elkins, C., Long, T. E., & Wilkes, G. L. (2005). Electrospinning of linear 449

homopolymers of poly(methyl methacrylate): exploring relationships between 450

fiber formation, viscosity, molecular weight and concentration in a good solvent. 451

Polymer, 46(13), 4799-4810. 452

Jafari, S. M. (2017). An overview of nanoencapsulation techniques and their 453

classification. In Nanoencapsulation technologies for the food and 454

nutraceutical industries (pp. 1-34). Academic Press. 455

28

Karim, M. R., Lee, H. W., Kim, R., Ji, B. C., Cho, J. W., Son, T. W., . . . Yeum, J. H. 456

(2009). Preparation and characterization of electrospun 457

pullulan/montmorillonite nanofiber mats in aqueous solution. Carbohydrate 458

Polymers, 78(2), 336-342. 459

Kerr, R. W., Cleveland, F. C., & Katzbeck, W. J. (1951). The molecular magnitude of 460

amylopectin. Journal of the American Chemical Society, 73(1), 111-117. 461

Klossner, R. R., Queen, H. A., Coughlin, A. J., & Krause, W. E. (2008). Correlation of 462

chitosan’s rheological properties and its ability to electrospin. 463

Biomacromolecules, 9(10), 2947-2953. 464

Kong, L., & Ziegler, G. R. (2012). Role of molecular entanglements in starch fiber 465

formation by electrospinning. Biomacromolecules, 13(8), 2247-2253. 466

Kong, L., & Ziegler, G. R. (2013). Quantitative relationship between electrospinning 467

parameters and starch fiber diameter. Carbohydrate Polymmers, 92(2), 1416-468

1422. 469

Kong, L., & Ziegler, G. R. (2014a). Fabrication of pure starch fibers by electrospinning. 470

Food Hydrocolloids, 36, 20-25. 471

Kong, L., & Ziegler, G. R. (2014b). Rheological aspects in fabricating pullulan fibers 472

by electro-wet-spinning. Food Hydrocolloids, 38, 220-226. 473

Lancuski, A., Vasilyev, G., Putaux, J. L., & Zussman, E. (2015). Rheological Properties 474

and Electrospinnability of High-Amylose Starch in Formic Acid. 475

Biomacromolecules, 16(8), 2529-2536. 476

Li, R., Tomasula, P., De Sousa, A. M. M., Liu, S. C., Tunick, M., Liu, K., & Liu, L. 477

(2017). Electrospinning Pullulan Fibers from Salt Solutions. Polymers, 9(1). 478

Masina, N., Choonara, Y. E., Kumar, P., du Toit, L. C., Govender, M., Indermun, S., & 479

Pillay, V. (2017). A review of the chemical modification techniques of starch. 480

Carbohydrate Polymers, 157, 1226-1236. 481

McKee, M. G., Wilkes, G. L., Colby, R. H., & Long, T. E. (2004). Correlations of 482

solution rheology with electrospun fiber formation of linear and branched 483

polyesters. Macromolecules, 37(5), 1760-1767. 484

Mendes, A. C., Stephansen, K., & Chronakis, I. S. (2017). Electrospinning of food 485

proteins and polysaccharides. Food Hydrocolloids, 68, 53-68. 486

Morris, E. R., Cutler, A., Ross-Murphy, S., Rees, D., & Price, J. (1981). Concentration 487

and shear rate dependence of viscosity in random coil polysaccharide solutions. 488

Carbohydrate Polymers, 1(1), 5-21. 489

Niu, B., Shao, P., Luo, Y., & Sun, P. (2020). Recent advances of electrosprayed particles 490

as encapsulation systems of bioactives for food application. Food Hydrocolloids, 491

99, 105376. 492

Prajapati, V. D., Jani, G. K., & Khanda, S. M. (2013). Pullulan: an exopolysaccharide 493

and its various applications. Carbohydrate Polymers, 95(1), 540-549. 494

Sun, J., Perry, S. L., & Schiffman, J. D. (2019). Electrospinning Nanofibers from 495

Chitosan/Hyaluronic Acid Complex Coacervates. Biomacromolecules, 20(11), 496

4191-4198. 497

Sun, X. B., Jia, D., Kang, W. M., Cheng, B. W., & Li, Y. B. (2013). Research on 498

electrospinning process of pullulan nanofibers. Applied Mechanics and 499

Materials, 268, 198-201. 500

Sweedman, M. C., Tizzotti, M. J., Schafer, C., & Gilbert, R. G. (2013). Structure and 501

physicochemical properties of octenyl succinic anhydride modified starches: a 502

review. Carbohydrate Polymers, 92(1), 905-920. 503

29

Tao, J., & Shivkumar, S. (2007). Molecular weight dependent structural regimes during 504

the electrospinning of PVA. Materials Letters, 61(11-12), 2325-2328. 505

Vasilyev, G., Vilensky, R., & Zussman, E. (2019). The ternary system amylose-506

amylopectin-formic acid as precursor for electrospun fibers with tunable 507

mechanical properties. Carbohydrate Polymers, 214, 186-194. 508

Wang, H., Kong, L., & Ziegler, G. R. (2018). Plasticization and conglutination improve 509

the tensile strength of electrospun starch fiber mats. Food Hydrocolloids, 83, 510

393-396. 511

Wang, H., Kong, L., & Ziegler, G. R. (2019a). Aligned wet-electrospun starch fiber 512

mats. Food Hydrocolloids, 90, 113-117. 513

Wang, H., Kong, L., & Ziegler, G. R. (2019b). Fabrication of starch - Nanocellulose 514

composite fibers by electrospinning. Food Hydrocolloids, 90, 90-98. 515

Wang, H., & Ziegler, G. R. (2019). Electrospun nanofiber mats from aqueous starch-516

pullulan dispersions: Optimizing dispersion properties for electrospinning. 517

International Journal of Biological Macromolecules, 133, 1168-1174. 518

Xue, J., Wu, T., Dai, Y., & Xia, Y. (2019). Electrospinning and Electrospun Nanofibers: 519

Methods, Materials, and Applications. Chemical Reviews, 119(8), 5298-5415. 520

Yang, Y., Xie, B., Liu, Q., Kong, B., & Wang, H. (2020). Fabrication and 521

characterization of a novel polysaccharide based composite nanofiber films with 522

tunable physical properties. Carbohydrate Polymers, 236, 116054. 523

Zhang, B., Xie, F., Shamshina, J. L., Rogers, R. D., McNally, T., Halley, P. J., . . . Zhao, 524

S. (2017). Dissolution of Starch with Aqueous Ionic Liquid under Ambient 525

Conditions. ACS Sustainable Chemistry & Engineering, 5(5), 3737-3741. 526

527

528