Article

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

2016, Vol. 45(4) 483494

! The Author(s) 2015
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/1528083714537101
jit.sagepub.com

Guojun Jiang, Sai Zhang and Xiaohong Qin

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 nanobers with many outstanding characteristics, such as high porosity, high ratio of
Key Laboratory of Textile Science & Technology, Ministry Education, Shanghai Key Laboratory of Advanced
Micro & Nano Textile Materials, College of Textiles, Donghua University, Songjiang, Shanghai, China
Corresponding author:
Xiaohong Qin, Key Laboratory of Textile Science & Technology, Ministry Education, Shanghai Key Laboratory
of Advanced Micro & Nano Textile Materials, College of Textiles, Donghua University, No. 2999 North
Renmin Road, Songjiang, Shanghai 201620, China.
Email: xhqin@dhu.edu.cn

Downloaded from jit.sagepub.com at COLORADO STATE UNIV LIBRARIES on October 31, 2016

484

Journal of Industrial Textiles 45(4)

surface area to mass and superior mechanical properties [1]. Owing to these unique
properties, electrospun nanobers have various applications in diverse areas such
as ltration, tissue scaolds, 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 nanobers 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 multineedle 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
nanober for commercial applications. The pioneering work was reported by
Yarin et al. [5], who used a magnetic uid 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 nanober 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 ber
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 lled 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 dierent setups and never performed on the eect of processing parameters on jet formation in free surface
electrospinning. In this paper, we studied the inuence of parameter on free
surface electrospinning using a response surface methodology.
The electrospinning process is inuenced by many parameters including the
polymer solution parameters, the processing conditions and the ambient parameters [2]. In order to control the ber diameter and productivity, the eects
of electrospinning variables on electrospinning process were required to be
systematically investigated. Recently, response surface methodology (RSM) has
been proven to be an eective tool for the optimization of electrospinning
parameters for nanobers. RSM is a set of mathematical and statistical techniques
that can be used to optimize processes in which a certain response is inuenced
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

Downloaded from jit.sagepub.com at COLORADO STATE UNIV LIBRARIES on October 31, 2016

Jiang et al.

485

RSM for the quantitative study of polyacrylonitrile (PAN) and poly(D-lactide)

PDLA, respectively. Yordem et al. [15] utilized RSM to investigate the interactive
eects of the parameters on the resultant ber and established a prediction scheme
for targeted PAN ber diameter. Konwarh et al. [16] applied RSM using Box
Behnken design (BBD) technique as an optimization tool to minimize the diameter
of electrospun CA nanobers. During this study, we carried out the RSM using
BBD technique to investigate the inuence of operating parameters including polymer solution concentration, applied voltage, and collecting distance on ber diameter and free surface electrospinning productivity.
We have recently demonstrated a simple and ecient free surface electrospinning setup using one-stepped pyramid stage [17]. As discussed previously, the
experimental parameters have important inuence on ber diameter and productivity. In order to study and understand the electrospinning process in this novel free
surface system, especially the combined eects of the processing parameters on the
ber diameter and productivity, RSM using BBD technique was performed to
identify and quantify a quantitative relationship between process parameters and
average ber 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
purication. 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 C.

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 ve major components: a high-voltage direct-current power supply, a
stepped pyramid stage, a Teon 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 ow
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 dierent 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.

Downloaded from jit.sagepub.com at COLORADO STATE UNIV LIBRARIES on October 31, 2016

486

Journal of Industrial Textiles 45(4)

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 bers 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 ber diameter of the nanober 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 eect of process parameters on the ber diameter and productivity, the BBD technique was adopted in this study. The three factors with three

Downloaded from jit.sagepub.com at COLORADO STATE UNIV LIBRARIES on October 31, 2016

Jiang et al.

487

Table 1. BoxBehnken design matrix containing 17 experimental runs.

Responses
Run
order

Solution
concentration
(%)

Applied
potential
(kV)

Working
distance
(cm)

Y1 fiber
diameter (nm)

Y2 productivity
(g/h)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17

9.00
8.00
9.00
8.00
10.00
10.00
9.00
9.00
9.00
10.00
10.00
9.00
9.00
9.00
8.00
8.00
9.00

60.00
65.00
60.00
55.00
60.00
65.00
55.00
65.00
65.00
60.00
55.00
60.00
55.00
60.00
60.00
60.00
60.00

15.00
15.00
15.00
15.00
20.00
15.00
20.00
20.00
10.00
10.00
15.00
15.00
10.00
15.00
20.00
10.00
15.00

183  52
157  45
179  53
182  68
219  82
196  62
213  56
189  57
199  75
237  78
228  69
181  71
226  84
185  59
175  53
190  61
183  43

3.93
4.96
4.12
2.95
2.46
4.72
2.62
4.03
5.84
3.91
2.83
4.01
3.51
3.87
3.04
4.15
3.98

levels BBD design were used to investigate the signicance of three selected processing parameters (i.e. solution concentration, applied potential, and working
distance) on average ber 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 t the experimental data
presented in Table 1
Y 0 1 X1 2 X2 3 X3 11 X21 22 X22 33 X23 12 X1 X2 23 X2 X3
13 X1 X3

where Y is response (average ber diameter or productivity in this study), X1, X2,
and X3 are the variables (solution concentration, applied potential, and working
distance, respectively), and 0, i, ii, and ij are the regression coecients.
The regression and graphical analysis of the obtained data was performed using
the software Design-Expert 8.0.5. The signicance of each coecient was determined from the t-values and p-values. Co-ecients in the equation with t-values
greater than t-values at 95% level of condence or p > 0.05 were considered

Downloaded from jit.sagepub.com at COLORADO STATE UNIV LIBRARIES on October 31, 2016

488

Journal of Industrial Textiles 45(4)

statistically signicant. The nal response surface model was further rened by
deleting the terms found to be associated with a level of signicance greater than
5% (p > 0.05) [18].

Results and discussion

Response function
By regression analysis, the unknown coecients were obtained. The analysis of
variance (ANOVA) was performed to evaluate the quadratic response surface
model. p-Value is a measure of statistical signicance, and Radj2 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 signicant [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 signicant model
terms, respectively. The model F-values of 218.51 and 156.70 implied the models to
be statistically signicant, respectively. There was only a 0.01% chance that a
model F-value this large could occur due to noise. Values of R2adj are 0.9879
for average ber diameter and 0.9855 for productivity, which illustrate that the
models are able to explain variability of 98.79% for average ber diameter and
98.55% for productivity. The model-predicted ber diameter (productivity) versus
the experimental diameter (productivity) is shown in Figure 2. By eliminating the
insignicant terms (p > 0.05) from the full quadratic model, the application of
RSM yielded equations (2) and (3), which are empirical relationship between the
ber diameter (productivity) and the test variables in actual unit
Y1 1358:31  37:97X1  27:66X2  24:34X3 3:33X21 0:21X22 0:77X23
Table 2. ANOVA for response surface reduced quadratic model for
average fiber diameter.
Source

SS

DF

F value

p-value
prob >F

Model
(significant)
X1
X2
X3
X12
X22
X32

7519.61

218.51

<0.0001

3873.32
1466.65
382.95
46.74
113.82
1541.15

1
1
1
1
1
1

675.31
255.71
66.77
8.15
19.84
268.70

<0.0001
<0.0001
<0.0001
0.0171
0.0012
<0.0001

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

Downloaded from jit.sagepub.com at COLORADO STATE UNIV LIBRARIES on October 31, 2016

Jiang et al.

489
Y2 8:46 6:4X1  0:85X2 0:69X3  0:0092X2 X3  0:36X21
0:0099X22  0:0091X23

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

Results from response function for fiber diameter

Figure 3 shows the SEM micrographs of ber samples from each experimental run
from 1 to 17. Figure 4 shows the diameter distribution of ber samples from each
experimental run from 1 to 17. The 3D surface plots of the response variable versus

Table 3. ANOVA for response surface reduced quadratic model for productivity.
Source

SS

DF

F value

p-value
prob >F

Model
(significant)
X1
X2
X3
X2X3
X12
X22
X32

12.16

156.70

<0.0001

0.17
7.30
3.46
0.21
0.56
0.26
0.22

1
1
1
1
1
1
1

15.71
658.77
312.26
19.11
50.23
23.10
19.85

0.0033
<0.0001
<0.0001
0.0018
<0.0001
0.0010
0.0016

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

Figure 2. Plot of model predicted versus experimental values: (a) fiber diameter and
(b) productivity.

Downloaded from jit.sagepub.com at COLORADO STATE UNIV LIBRARIES on October 31, 2016

490

Journal of Industrial Textiles 45(4)

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.

two variables at one time (keeping the third variable constant value) are depicted in
Figure 5. The eect of the solution concentration on ber 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
ber diameter with the solution concentration was observed. It was obvious that
the eect of solution concentration on ber diameter was independent from the
applied potential and working distance. With the increase in solution

Downloaded from jit.sagepub.com at COLORADO STATE UNIV LIBRARIES on October 31, 2016

Jiang et al.

491

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.

concentration, the ber diameter increased greatly. Previous work similarly

reported that the solution concentration is a very important parameter that aects
the nanober 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.
The eect of the applied potential on ber diameter was illustrated in Figure
5(a1) and (a3). Since the electric eld is the main driving force to generate jets [22],
the applied potential is a very important parameter aecting both the electrospinning process and ber diameter. As shown in Figure 5(a1), the ber 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
insignicant decrease in ber diameter with the increasing of the applied potential.
Probably, it is due to the increase in applied potential, which has two major different eects on ber diameter. First, increasing applied potential will increase the

Downloaded from jit.sagepub.com at COLORADO STATE UNIV LIBRARIES on October 31, 2016

492

Journal of Industrial Textiles 45(4)

electric eld strength causing more extension of the jets, thereby favoring thinner
ber formation. Second, increasing the applied potential also draws more solution
from the free surface causing thick ber formation.
The eect of working distance on ber diameter is shown in Figure 5(a2) and (a3).
As illustrated in these gures, the eect of working distance is not always the same.
Since the varying working distance will have a direct inuence in both the jet ight
time and the electric eld strength [2], it is worth quoting that the working distance
has a twofold eect on ber diameter. Longer distance will provide more time to
stretch the jet in the electric eld as well as evaporate the solvent. Hence, the ber
diameter will be prone to decrease. On the other hand, longer distance will decrease
the electric eld resulting in less stretching of the jet, which leads to thick ber
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 ber diameter. However, at
long distances, electric eld would be a determining factor. As the working distance
increases, the electric eld decreases, which led to a slight increase in ber diameter.

Results from response function for productivity

Our novel free surface electrospinning setup is dierent 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 ber diameter.
Figure 5(b1) and (b2) shows the inuence 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 ber 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 eect of applied potential on productivity.
The eect 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 eect to increasing the
applied potential, which induces higher electric eld strength. It was obvious that
there was a strong interaction eect between applied potential and working distance on productivity as demonstrated in the rened surface model. All these
results suggest that the applied potential plays an important role in improving
the productivity in this novel free surface electrospinning.

Downloaded from jit.sagepub.com at COLORADO STATE UNIV LIBRARIES on October 31, 2016

Jiang et al.

493

The eect 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 eld strength. This trend was
exactly the same with increasing applied potential because the increased applied
potential enhanced the electric eld strength. When the working distance is too
long, the electric eld 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 ber and evaporation of solvent, resulting in greater ber diameter.

Conclusion
In the present work, a novel free surface electrospinning setup for high-throughput
nanobers by utilizing one-stepped pyramid stage has been studied. A BBD technique-based response surface model was employed to investigate the eect of processing parameters on ber diameter and productivity. Among the three independent
parameters, solution concentration demonstrated a direct eect on the ber diameter
regardless of the applied potential and working distance. The response surface analysis indicated that applied potential had a signicant eect on productivity. The
interactive eect between applied potential and working distance on productivity was
also observed. The decrease of ber diameter with simultaneous increase of productivity has been achieved by using this novel free surface electrospinning system.
Declaration of Conflicting Interests
The author(s) declared no potential conicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) disclosed receipt of the following nancial 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 Scientic 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
[1] Huang ZM, Zhang YZ, Kotaki M, et al. A review on polymer nanofibers by electrospinning and their applications in nanocomposites. Compos Sci Technol 2003; 63: 22232253.

Downloaded from jit.sagepub.com at COLORADO STATE UNIV LIBRARIES on October 31, 2016

494

Journal of Industrial Textiles 45(4)

[2] Ramakrishna S. An introduction to electrospinning and nanofibers. Singapore: World

Scientific, 2005, p.277.
[3] Huang NF, Patel S, Thakar RG, et al. Myotube assembly on nanofibrous and micropatterned polymers. Nano lett 2006; 6: 537542.
[4] Zheng YS, Liu XK and Zeng YC. Electrospun nanofiber from a multihole spinneret
with uniform electric field. J Appl Polym Sci 2013; 130: 32213228.
[5] Yarin AL and Zussman E. Upward needleless electrospinning of multiple nanofibers.
Polymer 2004; 45: 29772980.
[6] 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.
[7] Liu Y and He JH. Bubble electrospinning for mass production of nanofibers. Int J
Nonlinear Sci Num 2007; 8: 393396.
[8] Lin T. Electrostatic spinning assembly. Patent 0311671-ZH, USA, 2011.
[9] Lu BA, Wang YJ, Liu YX, et al. Super high throughput needleless electrospinning
using a rotary cone as spinneret. Small 2010; 6: 16121616.
[10] Thoppey NM, Bochinski JR, Clarke LI, et al. Edge electrospinning for high throughput
production of quality nanofibers. Nanotechnology 2011; 26: 2234.
[11] Forward KM and Rutledge GC. Free surface electrospinning from a wire electrode.
Chem Eng J 2012; 183: 492503.
[12] Sukigara S, Gandhi M, Ayutsede J, et al. Regeneration of Bombyx mori silk by electrospinning, Part 2: Process optimization and empirical modeling using response surface methodology. Polymer 2004; 45: 37013708.
[13] Gu SY, Ren J and Vancso GJ. Process optimization and empirical modeling for electrospun polyacrylonitrile (PAN) nanofiber precursor of carbon nanofibers. Eur Polym
J 2005; 41: 25592568.
[14] Gu SY and Ren J. Process optimization and empirical modeling for electrospun
poly(D,L-lactide) fibers using response surface methodology. Macromol Mater Eng
2005; 290: 10971105.
[15] Yordem OS, Papila M and Menceloglu YZ. Effects of electrospinning parameters on
polyacrylonitrile nanofiber diameter: An investigation by response surface methodology. Mater Des 2008; 29: 3444.
[16] Konwarh R, Misra M, Mohanty AK, et al. Diameter-tuning of electrospun cellulose
acetate fibers: A BoxBehnken design (BBD) study. Carbohydr Polym 2013; 92:
11001106.
[17] Jiang GJ, Zhang S and Qin XH. High throughput of quality nanofibers via one stepped
pyramid-shaped spinneret. Mater Lett 2013; 106: 5658.
[18] Box GEP and Draper NR. Empirical model building and response surfaces. New York:
Wiley, 1987.
[19] Schmidt SR and Launsby RG. Understanding industrial designed experiments.
Colorado: Air Academy Press, 1992.
[20] Li Q, Jia ZD, Yang Y, et al. Preparation and properties of poly(vinyl alcohol) nanofibers by electrospinning. Int Conf Solid Dielectr 2007; 215218.
[21] Rabbi A, Nasouri K, Bahrambeygi H, et al. RSM and ANN approaches for modeling
and optimizing of electrospun polyurethane nanofibers morphology. Fiber Polym 2012;
13: 10071014.
[22] Deitzel JM, Kleinmeyer J, Harris D, et al. The effect of processing variables on the
morphology of electrospun nanofibers and textiles. Polymer 2001; 42: 261272.

Downloaded from jit.sagepub.com at COLORADO STATE UNIV LIBRARIES on October 31, 2016