Effects of cylindrical-electrode-assisted solution blowing spinning process parameters on polymer nanofiber morphology and microstructure

2.3.1 Single factor design

In this study, the processing parameters of CSBS including length of cylinder (LC), needle to cylinder distance (NCD), and left face of cylinder to collector distance (CCD) were chosen to study the impact of these parameters on the diameter of PEO nanofiber. Each of the aforementioned three factors was changed at five levels while all other factors unchangeable. The spinning conditions for each single-factor experiment were listed in Tables 1-3.

2.3.2 Orthogonal design

In this paper, the effects of seven factors on the CSBS nanofibers diameter and the porosity of various layers were studied by orthogonal analysis. These seven factors were: air pressure, voltage, length of cylinder (LC), diameter of cylinder (DC), injection rate, needle to cylinder distance (NCD), and left face of cylinder to collector distance (CCD). Two levels of every factor were tested as listed in Table 4.

3.1.1 Effect of the needle to cylinder distance (NCD) on the fiber diameter

The needle to cylinder distance (NCD) was an important parameter in the CSBS process. Figures 2a-e illustrate SEM photos of PEO nanofibers fabricated at NCD of 5, 10, 15, 20, 25 cm, respectively. The dependence of the fiber diameter on NCD was shown in Figure 2f. As shown in Figure 2f, when the NCD was increased from 5 to 25 cm, the average diameter increased from 570.976 to 787.073 nm. In CSBS process, when the flying spinning jet which sprayed from the needle approaches the cylinder, the jet will be charged due to electrostatic induction effect. The evaporation of solvent and air stretching reduces the thickness of the jet and increases the surface charge density of the jet. The increased density of surface charge splits the solution jet to thinner jets due to mutual charge repulsion. This procedure may repeat many times to decrease the fiber diameter (17,22). It is known from electrical knowledge that the electric field strength is greater as it is closer to the charged body. Conversely, the farther away from the charged body, the electric field strength is small. A smaller NCD means that the distance between the needle tip and the electrode is smaller. Therefore, the electric field strength at the tip of the needle is large when the NCD is small. Since the jet is ejected from the tip of the needle, the electric field intensity around the jet is larger as the NCD is smaller. At this time, the electrostatic induction effect of the jet is more obvious. This results in a greater induced charge density on the jet. An increase in the jet charge density results in an increase in the splitting frequency of the spinning jet, and finally, decreases the nanofiber diameter (17,22). In addition, reducing the NCD causes the jet to enter the electric field in the cavity of the electrode earlier, thereby charging earlier in the electric field in the electrode’s cavity. This is also one of the reasons for the fiber to become finer.

Figure 2

SEM photos of nanofibers prepared by CSBS system at the NCD: (a) 5 cm, (b) 10 cm, (c) 15 cm, (d) 20 cm, (e) 25 cm, (f) relationship between NCD and the fiber diameter.

3.1.2 Effect of length of cylinder (LC) on the fiber diameter

One of the major originalities of CSBS has lain in adding the charged cylinder. So it’s necessary to investigate the impact of cylinder length on the fiber diameter. Figures 3a-e show SEM images of spun nanofiber at LC of 6, 8, 10, 12, 14 cm, respectively. The relationship between LC and the fiber diameter is illustrated in Figure 3f. As depicted in Figure 3f, when the LC was increased from 6 to 14 cm, the average diameter decreased from 1354.209 to 1184.889 nm. The following is the possible explanation for the above result: an increase LC can increase the time of solution jet pass through the cylinder. The increase in the time it takes for the jet to pass through the cylinder extends the duration of the electrostatic induction effect, causing the jet to charge more fully due to electrostatic induction effects. For the above reasons, the splitting frequency of the jet increases, eventually making the fiber diameter smaller. During the course of the experiments, we found that too large LC is not conducive to the jet flow through the cylindrical electrode.

Figure 3

SEM images of nanofiber nonwoven produced by CSBS at the LC: (a) 6 cm, (b) 8 cm, (c) 10 cm, (d) 12 cm, (e) 14 cm, (f) relationship between LC and fiber diameter.

3.1.3 Effect of left face of cylinder to collector distance (CCD) on the fiber diameter

Figures 4a-e illustrate SEM images of nanofiber mats prepared at CCD of 75, 80, 85, 90, 95 cm, respectively. Figure 4f shows the relationship between CCD and the fiber diameter. As shown in Figure 4f, when the CCD was increased from 75 to 95 cm, the average diameter of PEO fibers could be varied in a very narrow range of 572.8425 to 613.4979 nm. That is to say, the influence of CCD on the fiber diameter was not obvious. In the CSBS process, the solution jets were stretched by electrostatic force plus air shearing force (17). Intuitively, increasing the CCD means increasing the distance the jet flies, which in turn increases the jet flight time. In longer flight times, the jet will be stretched more fully by the airflow pull and electrostatic forces, which will result in a smaller fiber diameter.

Figure 4

SEM photos of PEO nanofiber mats obtained by CSBS at the CCD: (a) 75 cm, (b) 80 cm, (c) 85 cm, (d) 90 cm, (e) 95 cm (f) relationship between CCD and fiber diameter.

However, the effective air stretching distance is very short (≈100μm) from the injection nozzle (23,24). Increase CCD couldn’t improve the effect of air-stretch. Medeiros et al also found the working distance did not have significant effect on the fiber diameter when nanofibers obtained by solution blowing spinning system (1). During the CSBS process, the solution jet will solidify into fiber due to solvent evaporation. The solidification of the jet increase the viscosity of the jet, which in turn increase the surface tension of the jet. When the surface tension of the jet exceeds the repulsive electric force of spinning jet, the splitting of jets ended. When the above situation occurs, only increasing the CCD without changing the other parameters will not affect the frequency of the splitting. That is to say, in a certain range, enhancement of CCD couldn’t increase the splitting frequency. Chang Seok Ki et al also found the effect of spinning distance (SD) on the size of nanofibers was not significant when nanofibers were prepared by electrospinning system (25). In the course of the experiments, we also found that an excessively large CCD is not conducive to the accumulation of nanofibers on the receiver. This is because the increase of the CCD leads to an increase in the distance of the jet flight, which in turn causes an increase in the divergence of the airflow, which is not conducive to the reception of the nanofiber. However, too short CCD may produce bead defects, and cause fibers adhere to each other due to not enough distance was provided for evaporation of solvent.

3.2.1 Morphology of CSBS nanofibers

The process variables of SBS are few (26). However, the process parameters of CSBS are more than SBS. Thus, it is important to research the effects of these process parameters on the fiber diameter and nanofiber webporosity in the CSBS process. The orthogonal array design plan, the average fiber diameter, the standard deviation of the fiber diameter, and the different layers’ porosity (P1, P2, P3) for each experiment were shown in Table 5. SEM photos of CSBS PEO nanofibers correspond with experiments list in Table 5 were shown in Figure 5.

Figure 5

SEM photos illustrating the morphology of the PEO nanofiber mats: (a-h) correspond to Experiments 1-8.

As showed in Table 5, where M_ij indicate the sum of each fiber morphology or microstructure properties indexes values (including the average fiber diameter, fiber diameter standard deviation, P1, P2, and P3) of row ”j” if the level is i. R_j represents the difference between the M_1j value and the M_2j value in the “j” column. A large R_j value indicated the effect of the factor j on this corresponding properties index was significant. From Table 5, the R_j values of the fiber average diameter corresponding to each factor range from 484.025>483.477>439.519>324.417>248.885>53.0 91>37.583. Thus, the importance order of the seven factors on the nanofiber average diameter can be listed as follows: injection rate>air pressure>NCD>DC>LC>voltage>CCD.

Similarly, the importance order of each factor on the fiber diameter standard deviation was as follows: NCD>DC>voltage>CCD>LC>air pressure>injection rate. The R_j value of the fiber average diameter corresponding to CCD was 37.583, which was the smallest of all seven process parameters. This means that the CCD has no significant effect on the average fiber diameter. This conclusion is consistent with the conclusions in Section 3.1.3of this paper. Table 5 also indicated that the average diameter and standard deviation of the diameter varied from 539.121 to 904.149 nm and 127.903 to 303.253 nm, respectively.

As can be seen from Figure 5 and Table 5, the nanofibers prepared in Experiment 7 had the smallest fiber diameter and the best diameter uniformity. According to Table 5, injection rate, air pressure and NCD were the three most important factors affecting the average diameter of fibers, the R_j values of these three process parameters were much larger than other parameters. In all of the eightexperiments, Experiment 7 had the highest air pressure, the lowest injection rate, and the smallest NCD. According to the conclusions in Section 3.1.1 and references (1,27), reducing the injection rate or increasing the gas pressure or reducing NCD leads to a decrease in nanofiber diameter. Therefore, the fiber diameter of Experiment 7 was the smallest.

It is known that the fiber diameter could affect the mechanical properties of nanofiber (28), nanofiber web pore size and other properties. Therefore, the conclusions of this paper can be applied to many applications of CSBS nanofibers.

3.2.2 Porosity of various layers of CSBS nanofiber web

After converting the SEM images to binary form by using ImageJ, different thresholds can be used to observe the various layers of nanofiber web. The

magnifications of Experiments 3, 6, and 7 were 5000×, the magnifications of Experiments 5 and 8 were 4000×, and the magnifications of Experiments 1, 2, and 4 were 3000×. Figure 6 shows the histogram of different images with different magnification. As it can be seen in Figure 6, SEM images of different magnifications had different histograms. However, as described by Ghasemi-Mobarakeh et al. (21), the results of the method used herein did not depend upon the histogram or magnification of the images. This method can be used to calculate different layers’ porosity from most of images histogram and magnifications.

Figure 6

Histogram of different SEM photos. Magnification of (a) 3000, (b) 4000 and (c) 5000.

Figure 7 shows the different binary images of the no. 7 experiment under various thresholds. The morphologies of the various layers of the nanofiber web were shown by the selection of different thresholds. When threshold 1 was applied, only the surface layer appeared in the binary image. By applying threshold 2, the surface and the middle layer were displayed. When the threshold was 3, all layers of the nanofiber web can be seen.

Figure 7

Binary image with different thresholds for Experiment 7 sample: (a) original image, (b) binary image with threshold 1, (c) binary image with threshold 2 and (d) binary image with threshold 3.

After converting the each original image into different binary images, the each binary image’s porosity was calculated using Eq. 1:

where N is the total number of pixels in the binary image, n is the number of white pixels, and P is the porosity of each binary image. The porosity of binary images with thresholds of 1, 2 and 3 were presented as P1, P2 and P3, respectively. The calculation results of the porosity of different layers of each sample in the orthogonal experiment were shown in Table 5. Four SEM photographs were taken for each sample, and each of the porosity values in Table 5 was the average porosity values of the four SEM photographs.

The results in Table 5 show that the R_j values of the P1 corresponding to each factor range from 23.0035>14.8 69>13.1915>10.642>9.1275>8.125>7.71. Thus, the importance order of the seven process parameters on P1 can be listed as follows: DC>voltage>air pressure>CCD>injection rate>NCD>LC. Similarly, the importance order of each factor on P2 was: voltage>DC>injection rate>air pressure>NCD>LC>CCD. The importance order of seven factors on P3 was as follows: voltage>NCD>injection rate>air pressure>DC>LC>CCD.

As shown in Table 5, the varied range of P1, P2, and P3 was 71.464-85.1415%, 60.32725-75.46625%, 48.23925-70.08875%, respectively. The P1, P2, and P3 values of Experiment 5 were the largest, which indicates that the porosity of the different layers of Experiment 5 was the

largest. Both P1 and P2 values of Experiment 8 were the smallest. This means that the porosity of the first and second layers of Experiment 8 was the smallest. The P3 values of Experiment 3 and Experiment 7 were approximately equal and smaller than the P3 values of the other experiments. The reason was followed: it can be seen from Table 5 that voltage and NCD were the parameters that had the greatest influence on P3, and their R_j values were much larger than other parameters. The voltage and NCD of Experiment 3 and Experiment 7 had the same value, so their P3 values were approximately equal. Although the P3 values of Experiment 3 and Experiment 7 were not significantly different, it can be seen from Figure 5 that the morphologies of the pores of these two experiments were different, the pores of Experiment 3 were larger than that of Experiment 7. The reason is that the larger diameter fibers formed larger pores than those formed by smaller diameter fibers (29). It also can be seen from Figure 5 that the number of pores in Experiment 3 and Experiment 7 were also different.

As it can be seen from Table 5, although the effect of the voltage on the average fiber diameter was not significant. However, the voltage had a great influence on P1, P2, and P3. Therefore, if we want to achieve the effect of changing the porosity of each layer without significantly affecting the fiber diameter, adjusting the voltage was a good method.

It is well known that the porosity of scaffolds is an important parameter in many tissue engineering applications. Appropriate porosity and pore morphology provide space for growth of blood vessels and migration of cells and ensure efficient exchange of waste products and nutrients between cells and their microenvironment. The following is the Jain et al.’s description of the different functions of different layers’ porosity: The decreasing porosity from the top surface toward the substrate suggests that while top layer would provide support for cell proliferation and growth, the subsequent layers with sufficient but lower porosity will allow transport of nutrient, blood, and waste (30). It can be seen that proper porosity of various layer is very important for cell growth. Therefore, using the conclusions of this paper, the porosity of each layer can be adjusted to prepare the CSBS nanofiber web suitable for cell growth and proliferation. For the first time, this paper used orthogonal experiments to study the effects of CSBS process parameters on the porosity of different layers of nanofiber web. These conclusions can promote the applications of the CSBS method in tissue engineering, bioengineering, filtration engineering and so on.

The porosity of nanofiber webs produced by electrospinning (ES) and CSBS technology was also compared. The process parameters and porosity of the electrospinning and CSBS were shown in Table 6. In each experiment, inner diameter of the needle (0.42 mm) and solution concentration (7%) were kept constant. The injection rates were kept constant in Experiments 1-4 (0.2 ml/h), and the injection rates of Experiments 5 and 6 were the same (0.5 ml/h). It can be seen from Table 6 that the nanofiber membranes obtained by the two spinning methods have similar porosity.