Sage Journals: Discover world-class research

Abstract

In this study, an improved free surface electrospinning was applied to large-scale production of polyvinyl alcohol nanofibers by utilizing a stepped pyramid stage. Multiple polymer jets were observed to form on the edges of the stepped pyramid stage. The influences of operating parameters (e.g. polymer solution concentration, applied voltage and collecting distance) on fiber diameter as well as productivity were experimentally investigated. Response surface methodology was utilized to obtain a quantitative relationship between selected electrospinning parameters and the average fiber diameters as well as the productivity, and the analysis of variance has been used to the statistical validation of regression models. Adjusted R-squared was found to be 98.79% and 98.55% for the fiber diameter and productivity, respectively. The results indicated that the solution concentration had a statistically significant effect on the fiber diameter and the applied potential influence, largely on the productivity.

Keywords

Free surface electrospinning response surface methodology nanofiber

Introduction

Electrospinning is a simple and versatile technique to produce continuous nanofibers with many outstanding characteristics, such as high porosity, high ratio of surface area to mass and superior mechanical properties [1]. Owing to these unique properties, electrospun nanofibers have various applications in diverse areas such as filtration, tissue scaffolds, drug-delivery devices, protective clothing, sensors and energy storage [2, 3].

However, the conventional single-needle electrospinning systems have typically low productivity, usually at the level of 0.25 g/h [4], which inhibits the application of nanofibers to commercial applications. Hence, more and more researchers try every means to enhance the production rate of electrospinning, such as increasing the number of needle nozzles (also called multi-needle setup) and electrospinning from a free liquid surface (also called free surface electrospinning). The multi-needle setup usually requires a large operating space and careful design of the relative locations of needles to minimize the strong charge repulsion. Besides, multi-needle setup is prone to clog during electrospinning. Nevertheless, these problems can be avoided by free surface electrospinning. Recently, free surface electrospinning has become a promising strategy to satisfy the increasing needs of nanofiber for commercial applications. The pioneering work was reported by Yarin et al. [5], who used a magnetic fluid to perturb the uppermost polymer solution to initiate multiple jets from the free surface of polymer layer. Later, Jirsak et al. [6] described a free surface electrospinning setup using a slowly rotating horizontal roller as the nanofiber generator, which was subsequently commercialized by the Elmarco Company under the brand name of Nanospider. Liu et al. [7] employed a gas pump to generate bubbles on a liquid surface to initiate the electrospinning process. Lin et al. [8] patented their free surface electrospinning system utilizing a spiral coil wire as the fiber generator. Lu et al. [9] demonstrated a super high-throughput free surface electrospinning technique consisting of an electriferous rotary cone. Thoppey et al. [10] reported a simple geometry for high-throughput electrospinning from a bowl edge that utilized a vessel filled with a polymer solution and a concentric cylindrical collector. More recently, Keith et al. [11] reported the formation of multiple jets from a rotating wire electrode spindle. However, such studies were carried out for the formation of multiple jet in different setups and never performed on the effect of processing parameters on jet formation in free surface electrospinning. In this paper, we studied the influence of parameter on free surface electrospinning using a response surface methodology.

The electrospinning process is influenced by many parameters including the polymer solution parameters, the processing conditions and the ambient parameters [2]. In order to control the fiber diameter and productivity, the effects of electrospinning variables on electrospinning process were required to be systematically investigated. Recently, response surface methodology (RSM) has been proven to be an effective tool for the optimization of electrospinning parameters for nanofibers. RSM is a set of mathematical and statistical techniques that can be used to optimize processes in which a certain response is influenced by many variables synchronously. Sukigara et al. [12] employed RSM to optimize the electrospinning of Bombyx mori silk. Gu et al. [13, 14] exploited the RSM for the quantitative study of polyacrylonitrile (PAN) and poly(d-lactide) PDLA, respectively. Yördem et al. [15] utilized RSM to investigate the interactive effects of the parameters on the resultant fiber and established a prediction scheme for targeted PAN fiber diameter. Konwarh et al. [16] applied RSM using Box Behnken design (BBD) technique as an optimization tool to minimize the diameter of electrospun CA nanofibers. During this study, we carried out the RSM using BBD technique to investigate the influence of operating parameters including polymer solution concentration, applied voltage, and collecting distance on fiber diameter and free surface electrospinning productivity.

We have recently demonstrated a simple and efficient free surface electrospinning setup using one-stepped pyramid stage [17]. As discussed previously, the experimental parameters have important influence on fiber diameter and productivity. In order to study and understand the electrospinning process in this novel free surface system, especially the combined effects of the processing parameters on the fiber diameter and productivity, RSM using BBD technique was performed to identify and quantify a quantitative relationship between process parameters and average fiber diameters as well as productivity.

Materials and methods

Materials

Polyvinyl alcohol (PVA, with an average molecular weight of 95,000 g/mol and degree of hydrolysis of >98%) was purchased from Acros and used without further purification. Homogeneous PVA solution with desired solution concentrations was prepared by dissolving PVA powder in deionized water under the aid of mechanical stirring for 8 h at 80℃.

Electrospinning setup

The experiments were carried out on a novel electrospinning setup. The schematic of the novel free surface electrospinning system is shown in Figure 1(a), which contained five major components: a high-voltage direct-current power supply, a stepped pyramid stage, a Teflon solution reservoir, a peristaltic pump, and a grounded collector. A stepped pyramid stage was utilized as the electrospinning generator. During the novel free surface electrospinning, the solution was continuously pumped into the spinneret so that every edge of the spinneret was covered with the polymer solution; meanwhile, the excessive solution was made to flow slowly to the reservoir. The applied voltage was increased until a number of jets were observed simultaneously from every edge (Figure (1b)). The PVA solutions were electrospun at different solution concentration, applied potentials, and the working distance (the distance between the top lip of the spinneret and the grounded collector), respectively. The experiments were performed at room temperature in the air.

Figure 1.

(a) Scheme of the electrospinning apparatus using a stepped pyramid stage and (b) Photograph of multi-jets in electrospinning process. Inset: a magnified image of the stepped pyramid stage and jets generated sites, respectively [18].

Characterization of nanofibers

The morphologies of electrospun fibers were characterized by scanning electron microscope (SEM) (JSM-5600LV, Japan). All samples were sputter-coated with gold for 60 s prior to imaging. The average fiber diameter of the nanofiber was calculated from the SEM images using Image J software (NIH, USA). Productivity of this novel electrospinning setup was calculated by electrospinning for 30 min, measuring the resultant mat mass by weighting the collected mat, and then extrapolating to obtain a rate in grams per hour (g/h).

Design of experiments

In order to assess the effect of process parameters on the fiber diameter and productivity, the BBD technique was adopted in this study. The three factors with three levels BBD design were used to investigate the significance of three selected processing parameters (i.e. solution concentration, applied potential, and working distance) on average fiber diameter and productivity. Table 1 represents the BBD design matrix for experimental factors and responses. Based on this design, the following second-order polynomial model was used to fit the experimental data presented in Table 1

Y = β_{0} + β_{1} X_{1} + β_{2} X_{2} + β_{3} X_{3} + β_{11} X_{1}^{2} + β_{22} X_{2}^{2} + β_{33} X_{3}^{2} + β_{12} X_{1} X_{2} + β_{23} X_{2} X_{3} + β_{13} X_{1} X_{3}

(1)

where Y is response (average fiber diameter or productivity in this study), X₁, X₂, and X₃ are the variables (solution concentration, applied potential, and working distance, respectively), and β₀, β_i, β_ii, and β_ij are the regression coefficients.

Table 1.

Box–Behnken design matrix containing 17 experimental runs.

Run order	Solution concentration (%)	Applied potential (kV)	Working distance (cm)	Responses
Run order	Solution concentration (%)	Applied potential (kV)	Working distance (cm)	Y₁ fiber diameter (nm)	Y₂ productivity (g/h)
1	9.00	60.00	15.00	183 ± 52	3.93
2	8.00	65.00	15.00	157 ± 45	4.96
3	9.00	60.00	15.00	179 ± 53	4.12
4	8.00	55.00	15.00	182 ± 68	2.95
5	10.00	60.00	20.00	219 ± 82	2.46
6	10.00	65.00	15.00	196 ± 62	4.72
7	9.00	55.00	20.00	213 ± 56	2.62
8	9.00	65.00	20.00	189 ± 57	4.03
9	9.00	65.00	10.00	199 ± 75	5.84
10	10.00	60.00	10.00	237 ± 78	3.91
11	10.00	55.00	15.00	228 ± 69	2.83
12	9.00	60.00	15.00	181 ± 71	4.01
13	9.00	55.00	10.00	226 ± 84	3.51
14	9.00	60.00	15.00	185 ± 59	3.87
15	8.00	60.00	20.00	175 ± 53	3.04
16	8.00	60.00	10.00	190 ± 61	4.15
17	9.00	60.00	15.00	183 ± 43	3.98

The regression and graphical analysis of the obtained data was performed using the software Design-Expert 8.0.5. The significance of each coefficient was determined from the t-values and p-values. Co-efficients in the equation with t-values greater than t-values at 95% level of confidence or p > 0.05 were considered statistically significant. The final response surface model was further refined by deleting the terms found to be associated with a level of significance greater than 5% (p > 0.05) [18].

Results and discussion

Response function

By regression analysis, the unknown coefficients were obtained. The analysis of variance (ANOVA) was performed to evaluate the quadratic response surface model. p-Value is a measure of statistical significance, and R_adj² represents the proportion of the total variability that has been explained by regression model. Values of “Prob >F” less than 0.0500 indicate model terms that are significant [19]. In this case, the ANOVA results (Tables 2 and 3) showed that X₁, X₂, X₃, X₁², X₂², X₃², and X₁, X₂, X₃, X₂X₃, X₁², X₂², and X₃² were found to be significant model terms, respectively. The model F-values of 218.51 and 156.70 implied the models to be statistically significant, respectively. There was only a 0.01% chance that a “model F-value” this large could occur due to noise. Values of R²_adj are 0.9879 for average fiber diameter and 0.9855 for productivity, which illustrate that the models are able to explain variability of 98.79% for average fiber diameter and 98.55% for productivity. The model-predicted fiber diameter (productivity) versus the experimental diameter (productivity) is shown in Figure 2. By eliminating the insignificant terms (p > 0.05) from the full quadratic model, the application of RSM yielded equations (2) and (3), which are empirical relationship between the fiber diameter (productivity) and the test variables in actual unit

Y_{1} = 1358.31 - 37.97 X_{1} - 27.66 X_{2} - 24.34 X_{3} + 3.33 X_{1}^{2} + 0.21 X_{2}^{2} + 0.77 X_{3}^{2}

(2)

Y_{2} = - 8.46 + 6.4 X_{1} - 0.85 X_{2} + 0.69 X_{3} - 0.0092 X_{2} X_{3} - 0.36 X_{1}^{2} + 0.0099 X_{2}^{2} - 0.0091 X_{3}^{2}

(3)

where Y₁ and Y₂ are average fiber diameter and productivity, respectively, and X₁, X₂, and X₃ are the uncoded values of solution concentration, applied potential, and working distance, respectively.

Figure 2.

Plot of model predicted versus experimental values: (a) fiber diameter and (b) productivity.

Table 2.

ANOVA for response surface reduced quadratic model for average fiber diameter.

Source	SS	DF	F value	p-value prob >F
Model (significant)	7519.61	6	218.51	<0.0001
X ₁	3873.32	1	675.31	<0.0001
X ₂	1466.65	1	255.71	<0.0001
X ₃	382.95	1	66.77	<0.0001
X ₁ ²	46.74	1	8.15	0.0171
X ₂ ²	113.82	1	19.84	0.0012
X ₃ ²	1541.15	1	268.70	<0.0001

DF: degrees of freedom; SS: sum of squares.

Table 3.

ANOVA for response surface reduced quadratic model for productivity.

Source	SS	DF	F value	p-value prob >F
Model (significant)	12.16	7	156.70	<0.0001
X ₁	0.17	1	15.71	0.0033
X ₂	7.30	1	658.77	<0.0001
X ₃	3.46	1	312.26	<0.0001
X ₂ X ₃	0.21	1	19.11	0.0018
X ₁ ²	0.56	1	50.23	<0.0001
X ₂ ²	0.26	1	23.10	0.0010
X ₃ ²	0.22	1	19.85	0.0016

DF: degrees of freedom; SS: sum of squares.

Results from response function for fiber diameter

Figure 3 shows the SEM micrographs of fiber samples from each experimental run from 1 to 17. Figure 4 shows the diameter distribution of fiber samples from each experimental run from 1 to 17. The 3D surface plots of the response variable versus two variables at one time (keeping the third variable constant value) are depicted in Figure 5. The effect of the solution concentration on fiber diameter versus applied potential and working distance is depicted in Figure 5(a1) and (a2), respectively. Regardless of the applied potential and working distance, a monotonic increase in fiber diameter with the solution concentration was observed. It was obvious that the effect of solution concentration on fiber diameter was independent from the applied potential and working distance. With the increase in solution concentration, the fiber diameter increased greatly. Previous work similarly reported that the solution concentration is a very important parameter that affects the nanofiber diameter [20,21]. This could be attributed to high-solution concentration, which led to higher viscoelastic forces, resulting in less stretch during the whipping motion of the jets.

Figure 3.

SEM micrographs of fiber samples from each experimental run from 1 to 17.

Figure 4.

Fiber diameter distribution of fiber samples from each experimental run from 1 to 17.

Figure 5.

3D response surface plots of fiber diameter versus (a₁) solution concentration and applied potential, (a₂) solution concentration and working distance, and (a₃) applied potential and working distance; 3D response surface plots of productivity versus (b₁) solution concentration and applied potential, (b₂) solution concentration and working distance, and (b₃) applied potential and working distance.

The effect of the applied potential on fiber diameter was illustrated in Figure 5(a1) and (a3). Since the electric field is the main driving force to generate jets [22], the applied potential is a very important parameter affecting both the electrospinning process and fiber diameter. As shown in Figure 5(a1), the fiber diameter does not dramatically change with the variation of applied potential in the range of solution concentrations in the experimental region. Our results show a slight but insignificant decrease in fiber diameter with the increasing of the applied potential. Probably, it is due to the increase in applied potential, which has two major different effects on fiber diameter. First, increasing applied potential will increase the electric field strength causing more extension of the jets, thereby favoring thinner fiber formation. Second, increasing the applied potential also draws more solution from the free surface causing thick fiber formation.

The effect of working distance on fiber diameter is shown in Figure 5(a2) and (a3). As illustrated in these figures, the effect of working distance is not always the same. Since the varying working distance will have a direct influence in both the jet flight time and the electric field strength [2], it is worth quoting that the working distance has a twofold effect on fiber diameter. Longer distance will provide more time to stretch the jet in the electric field as well as evaporate the solvent. Hence, the fiber diameter will be prone to decrease. On the other hand, longer distance will decrease the electric field resulting in less stretching of the jet, which leads to thick fiber formation. While at short distances, the working distance is a main factor. Therefore, increasing working distance spending more time to both stretch the jet and evaporate the solvent results in a decrease in fiber diameter. However, at long distances, electric field would be a determining factor. As the working distance increases, the electric field decreases, which led to a slight increase in fiber diameter.

Results from response function for productivity

Our novel free surface electrospinning setup is different from the traditional needle electrospinning. Multiple jets could be initiated on the edges of the stepped pyramid stage simultaneously. There are two main factors that determine the productivity: the number of jets and the fiber diameter.

Figure 5(b1) and (b2) shows the influence of the solution concentration on productivity. Increasing the solution concentration resulted in an increasing productivity followed by a decrease in productivity. As mentioned above, increasing the solution concentration led to enhancement in fiber diameter, which could increase the productivity. But at higher solution concentration, since the increased viscosity requires a larger electric force to initiate jets, the number of jets reduced, and a lower productivity occurred.

Figure 5(b1) and 5(b3) illustrates the effect of applied potential on productivity. The effect of applied potential on productivity was suggested to be independent from the solution concentration (Figure 5(b1)), and as for each solution concentration, increasement of applied potential had a trend of productivity increased. This could be attributed to increasing the applied potential enhanced the electric force and creates more jets on the edges of the stepped pyramid stage. The increasing of productivity with increasing applied potential was more apparent especially at short working distance as illustrated in Figure 5(b3). This could be due to the fact that decreasing the working distance has a similar effect to increasing the applied potential, which induces higher electric field strength. It was obvious that there was a strong interaction effect between applied potential and working distance on productivity as demonstrated in the refined surface model. All these results suggest that the applied potential plays an important role in improving the productivity in this novel free surface electrospinning.

The effect of working distance on productivity is shown in Figure 5(b2) and (b3). The productivity increased when a shorter working distance was applied, because shorter working distance led to enhanced electric field strength. This trend was exactly the same with increasing applied potential because the increased applied potential enhanced the electric field strength. When the working distance is too long, the electric field strength becomes weak resulting in less number of jets initiated on the free surface. On the other hand, shorter working distance causes less stretch of fiber and evaporation of solvent, resulting in greater fiber diameter.

Conclusion

In the present work, a novel free surface electrospinning setup for high-throughput nanofibers by utilizing one-stepped pyramid stage has been studied. A BBD technique-based response surface model was employed to investigate the effect of processing parameters on fiber diameter and productivity. Among the three independent parameters, solution concentration demonstrated a direct effect on the fiber diameter regardless of the applied potential and working distance. The response surface analysis indicated that applied potential had a significant effect on productivity. The interactive effect between applied potential and working distance on productivity was also observed. The decrease of fiber diameter with simultaneous increase of productivity has been achieved by using this novel free surface electrospinning system.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was partly supported by grants (51373033 and 11172064) from the National Natural Science Foundation of China and from the Foundation for the Author of National Excellent Doctoral Dissertation of PR China (200961), as well as sponsored by Shanghai Rising-Star Program in China (10QA1400100) and Fok Ying Tong Education Foundation (121071) to Prof Xiaohong Qin. Furthermore, it was also supported by Program for New Century Excellent Talents in University (NCET-10-0322) and the Fundamental Research Funds for the Central Universities as well as “Shu Guang” (11SG33) project supported by Shanghai Municipal Education Commission and Shanghai Education Development Foundation to her. This work was also supported by Chinese Universities Scientific Fund (CUSF-DH-D-2013021) to Mr Guojun Jiang. This work was also supported by the Keygrant Project of Chinese Ministry of Education (No 113027A).

References

Huang

Zhang

Kotaki

. A review on polymer nanofibers by electrospinning and their applications in nanocomposites. Compos Sci Technol 2003; 63: 2223–2253.

Ramakrishna

. An introduction to electrospinning and nanofibers, Singapore: World Scientific, 2005, pp. 277–277.

Huang

Patel

Thakar

. Myotube assembly on nanofibrous and micropatterned polymers. Nano lett 2006; 6: 537–542.

Zheng

Liu

Zeng

. Electrospun nanofiber from a multihole spinneret with uniform electric field. J Appl Polym Sci 2013; 130: 3221–3228.

Yarin

Zussman

. Upward needleless electrospinning of multiple nanofibers. Polymer 2004; 45: 2977–2980.

Jirsak O. A method of nanofibers production from polymer solution using electrostatic spinning and a device for carrying out the method. Patent 294274-ZH, CZ, 2005.

Liu

. Bubble electrospinning for mass production of nanofibers. Int J Nonlinear Sci Num 2007; 8: 393–396.

Lin T. Electrostatic spinning assembly. Patent 0311671-ZH, USA, 2011.

Wang

Liu

. Super high throughput needleless electrospinning using a rotary cone as spinneret. Small 2010; 6: 1612–1616.

10.

Thoppey

Bochinski

Clarke

. Edge electrospinning for high throughput production of quality nanofibers. Nanotechnology 2011; 26: 22–34.

11.

Forward

Rutledge

. Free surface electrospinning from a wire electrode. Chem Eng J 2012; 183: 492–503.

12.

Sukigara

Gandhi

Ayutsede

. Regeneration of Bombyx mori silk by electrospinning, Part 2: Process optimization and empirical modeling using response surface methodology. Polymer 2004; 45: 3701–3708.

13.

Ren

Vancso

. Process optimization and empirical modeling for electrospun polyacrylonitrile (PAN) nanofiber precursor of carbon nanofibers. Eur Polym J 2005; 41: 2559–2568.

14.

Ren

. Process optimization and empirical modeling for electrospun poly(d,l-lactide) fibers using response surface methodology. Macromol Mater Eng 2005; 290: 1097–1105.

15.

Yordem

Papila

Menceloglu

. Effects of electrospinning parameters on polyacrylonitrile nanofiber diameter: An investigation by response surface methodology. Mater Des 2008; 29: 34–44.

16.

Konwarh

Misra

Mohanty

. Diameter-tuning of electrospun cellulose acetate fibers: A Box–Behnken design (BBD) study. Carbohydr Polym 2013; 92: 1100–1106.

17.

Jiang

Zhang

Qin

. High throughput of quality nanofibers via one stepped pyramid-shaped spinneret. Mater Lett 2013; 106: 56–58.

18.

Box

GEP

Draper

. Empirical model building and response surfaces, New York: Wiley, 1987.

19.

Schmidt

Launsby

. Understanding industrial designed experiments, Colorado: Air Academy Press, 1992.

20.

Jia

Yang

. Preparation and properties of poly(vinyl alcohol) nanofibers by electrospinning. Int Conf Solid Dielectr 2007, pp. 215–218.

21.

Rabbi

Nasouri

Bahrambeygi

. RSM and ANN approaches for modeling and optimizing of electrospun polyurethane nanofibers morphology. Fiber Polym 2012; 13: 1007–1014.

22.

Deitzel

Kleinmeyer

Harris

. The effect of processing variables on the morphology of electrospun nanofibers and textiles. Polymer 2001; 42: 261–272.

Effect of processing parameters on free surface electrospinning from a stepped pyramid stage