Influence of Humidity, Temperature, and Annealing on Microstructure and Tensile Properties of Electrospun Polyacrylonitrile Nanofibers.
Due to simplicity and compatibility with a wide range of polymer, electrospinning process has potential for mass production of nanofibrous yarns, random mats, tissue scaffolding, filtration, drug delivery, and protective clothing [1-5]. In a typical electrospinning system, a high-voltage potential is applied to polymer solution at the tip of an electrode and at sufficient voltage the electrostatic repulsion force prevails over the surface tension of the polymer solution and triggers the formation of jet in the form of a Taylor cone at the end of the solution droplet [6, 7]. The charged polymer jet usually travels a few centimeters along a straight path due to longitudinal stress caused by the external electric field before lateral perturbation kicks in, which causes whipping instability on the polymer jet. Thus, the resulting polymer jet experiences an enormous amount of stretching before landing on to the collector in the form of nanofibers.
Many factors affect the microstructure and properties of the electrospun nanofiber. In general, these factors can be divided into three groups: (a) intrinsic properties of the polymer solution such as viscosity [8, 9], molecular weight [10, 11], conductivity , and surface tension ; (b) process parameters such as voltage [13, 14], feeding rate [13, 14], distance between the electrode and the collector [15, 16], and collector geometry ; and (c) environment conditions such as temperature [18, 19] and relative humidity (RH) . Study of these parameters is important for understanding the microstructural evolution of electrospun nanofibers and thus enabling tailoring of their properties. In our recent paper, we have reported that both applied voltage and feeding rate can lead to stable Taylor cone formation which can result in smaller and continuous nanofiber with higher mechanical properties . Most of the research work published in the literature focused on the first two groups of parameters and only a few studies have reported on the influence of RH and temperature on microstructure and mechanical properties of nanofibers. In a recent study, it has been reported that the RH can significantly affect the fiber diameter depending on the solvent present. For acidic solvent, the diameter of electrospun fibers was found to decrease with the increase of RH, while opposite trend was observed for organic solvent. For instance, Tripatanasuwan et al.  and Cai et al.  observed that the electrospun nanofibers from polyethylene oxide (PEO) aqueous solution showed decreasing fiber diameter with increasing RH. They suggested that the evaporation rate of solvent is much lower at higher RH which provides a longer stretching time for the electrospinning jet and consequently, smaller diameter fibers are formed. Similar trend was observed by Pelipenko et al.  for a variety of polymer/aqueous or acidic solvent systems such as polyvinyl alcohol (PVA)/water, PVA/hyaluronic acid (HA)/water, PEO/acetic acid, and PEO/chitosan (CS)/water. In contrast, Vrieze et al.  observed that the diameter of the electrospun nanofiber from cellulose acetate (CA)/Dimethylacetamide ([D.sub.M]Ac) solution increased as RH increased. For further understanding they added water into CA/[D.sub.M]Ac solution and found that precipitation of polymer occurs at higher RH as more water is absorbed into the polymer jet which prohibits further stretching of the jet, and consequently, higher diameter nanofibers are produced. Similar observation can also be found in the work of Fashandi et al.  for electrospun nanofiber from polyetherimide (PEI) solutions of three nonvolatile solvents: DMF, [D.sub.M]Ac, and N-Methyl-2-pyrrolidone (NMP).
RH can also affect the surface morphology and internal structure of the electrospun fibers. Lu et al.  investigated the influence of RH on surface morphology and internal microstructure of polystyrene (PS) nanofibers containing high volatile solvent such as tetrahydrofuran (THF) and low volatile solvent such as DMF. They observed that the fibers exhibited smooth surface and solid core at lower (2%) RH, while pores were formed at higher (22% and more) RH. However, they found that the pores were formed only on the surface of the fibers for PS/ THF solution, while the pores were found in the core of the fibers for PS/DMF solution. A few possible mechanisms of pore formation include breath figure formation (BFF) and vapor induced phase separation (VIPS) . BFF occurs in high volatile solvents where it is expected that rapid solvent evaporation significantly cools down the surface of the electrospun jet during its travel time. As the surface cools down, water vapors on the fiber surface condense in the form of droplets. The water droplet leaves an impression on the fiber surface which appears as pore after drying. VIPS occurs in high humid conditions where the polymer jet is exposed to non-solvent such as water vapor which diffuses into the polymer solution, and as a result thermodynamic instability occurs [24, 27-29], which leads to the separation of two phases such as polymer-rich phase and polymer-lean phase. Polymer-rich region favors a solid fiber matrix, whereas polymer-lean region creates porous structure. Dayal et al.  theoretically investigated the evolution of fiber morphology during electrospinning process in presence of nonsolvent. They used non-solvent/solvent/polymer ternary phase diagram to predict different morphological trends of electrospun fiber. In another study, Pai et al.  demonstrated that the morphology of PS/DMF electrospun fiber is a direct consequence of VIPS and two other processes such as solvent drying and buckling instability.
Vrieze et al.  also investigated the effect of temperature on the diameter of nanofiber produced from cellulose acetate (CA)/[D.sub.M]Ac and PVP/ethanol solutions. In both cases, they observed an initial increase in fiber diameter when temperature was increased from 10[degrees]C to 20[degrees]C. Further increase in temperature to 30[degrees]C reduced the diameter of the fiber. They believed that solvent evaporation rate and solution viscosity (temperature dependent) were the deciding factors influencing the diameter of the nanofiber.
Polyacrylonitrile (PAN) polymer has been widely used to produce carbon fibers due to their excellent characteristics such as spinnability, environmentally benign, thermal stability, and improved mechanical properties [30, 31]. Electrospinning provides the flexibility of tailoring the microstructure of carbon nanofibers. For example, porous carbon fibers are preferred for some applications such as energy storage and filtration due to larger surface area. Subsequent heat treatment process such as annealing can reduce the porosity in the fiber and increase the crystallinity in the polymer . Increasing crystallinity often increases the strength of polymer, but reduces the strain at break. On the other hand, reduction of internal porosity may increase both strength and strain at break. Thus, it is paramount important to investigate the effect of RH and temperature on fiber diameter, surface morphology and fiber microstructure, and mechanical properties of the electrospun PAN nanofibers and develop a better understanding of the mechanisms of the formation of various microstructures. Huang et al.  investigated the effect of RH on the surface morphology and tensile properties of PAN nanofibers. They observed increase in nanofiber diameter and surface area with the increase of RH. However, they did not look at the internal structure of the nanofiber, and the mechanism of porous structure formation is not discussed in detail in their work. Moreover, they found the highest tensile strength about 6.5 MPa for nanofiber mat produced at 20% RH, and an increase or decrease in RH lowered the strength of the nanofiber mat. According to them, nanofibers electrospun at higher RH underwent partial phase separation, resulting in a skin layer which hindered fiber-fiber bonding in the mat, however possible reasons for lower mechanical properties at lower RH was not discussed. Yu et al.  produced porous PAN fibers by electrospinning a water/ DMF/PAN ternary system. At low water content smooth fibers with beads were observed, while bead less porous fibers were found at higher water content. Thus, the influence of RH and temperature on the microstructure evolution of electrospun PAN nanofibers and their resultant mechanical properties can be better explained by using water/DMF/PAN ternary phase diagram.
In this study, we investigate the influence of RH, temperature, and annealing on the surface morphology, internal microstructure, and mechanical properties of electrospun PAN nanofibers using dilute PAN/DMF solution (10% PAN by weight). The evolution of fiber morphologies is examined by the means of ternary phase diagram of water/DMF/PAN. Tensile properties of both as-spun and annealed nanofibrous yarns are also measured and correlated with the morphology, internal structure, and crystallinity.
PAN powder (Sigma-Aldrich Inc., Cat. no.-181315) with an average molecular weight of about 150,000 g/mol was dried at 100[degrees]C under vacuum to remove any residual moisture. N,N- dimethylformamide (DMF) (Sigma-Aldrich Inc., Cat. no.-227056) was used as solvent to make PAN polymer solution. Dried PAN polymer (10% by weight) was dissolved in DMF at 80[degrees]C using a hot plate with magnetic stirrer to obtain the spinning solution. Any air bubbles entrapped into the solution were carefully purged prior to electrospinning.
The PAN/DMF solution was fed using a positive displacement syringe pump (KD Scientific Inc., Model-KDS 200) with a stainless steel needle having inner diameter of 0.41 mm. The needle was electrically connected to the positive terminal of a high voltage DC power supply (Spellman High Voltage Electronics Corp.). A copper foil of about 12.7 cm width was glued on the surface of a 25-cm diameter plastic disc collector which was electrically connected to the ground. The distance between the needle and the disc surface was maintained constant at 18 cm. About 200 pL solution was electrospun to deposit aligned nanofibers on the disc collector rotating at 600 rpm. The electrospinning voltage was kept constant at 15 kV while the flow rate was varied from 13 [micro]l/min to 26 [micro]l/min to achieve a stable polymer-solution jet during electrospinning [20, 34]. Once deposition is complete, the electrospun nanofiber bundles were peeled off the copper foil in the form of yarn and then mounted on a drying rack to keep them in tension while drying. A schematic of the custom built electrospinning setup is shown in Fig. 1. All yarns were dried at 60[degrees]C for 12 h under vacuum. During annealing, the yarns were further heat treated at 95[degrees]C for 4 h while they are still mounted on the drying rack.
The relative humidity (RH) and temperature inside the electrospinning chamber were regulated using a combination of dried silica, water vapor, and heater. A low-speed fan was used inside the chamber to attain uniform RH and temperature distribution. However, the fan was turned off during electrospinning process to avoid any disturbance. Humidity sensor and thermocouple were used to monitor relative humidity and temperature inside the chamber (Fig. 1). Once the RH and temperature reached to the desired value, electrospinning was performed. The RH and temperature were closely monitored during the entire electrospinning process. Nanofibrous yarns were produced for six humidity conditions such as [14.sup.[+ or -]2]%, [22.sup.[+ or -]1%], [30.sup.[+ or -]1]%, [40.sup.[+ or -]2]%, [50.sup.[+ or -]2]%, and [60.sup.[+ or -]2]% RH at [20.sup.[+ or -]20]% and for two humidity conditions such as [30.sup.[+ or -]2]% and [40.sup.[+ or -]3]% RH at [40.sup.[+ or -]2][degrees]C. If the variations in RH and temperature were higher than the tolerance limit mentioned above, the yarn was discarded and new yarns were made.
Surface morphologies and microstructures of nanofibers were characterized using a high resolution scanning electron microscope (SEM, Model-Neon 40 EsB from Carl Zeiss AG, Germany). Liquid nitrogen was used to freeze fracture the nanofibers and samples were coated with about 4 nm of Iridium. SEM images were also used to determine the distribution of nanofiber diameter using the ImageJ software. SEM images of the nanofiber were processed using ImageJ software. For each image, approximately 70 to 100 measurements were made and data were averaged to determine the average diameter of the nanofibers.
Dynamic mechanical analyzer ([D.sub.M]A Q800 from TA Instruments) was utilized to perform tensile experiments on nanofibrous yarns. Tensile tests of the yarns were performed at room temperature. Yarns samples with a gauge length of 9-10 mm were tested in tensile at a constant strain rate of 0.001 [s.sup.-1]. At least five samples were tested for each case and data were averaged. The cross-section area of the yarn sample was calculated by dividing the mass of the yarn by its length and the density of PAN (1.18 g/cc). Wide angle X-ray diffraction (WAXD) experiment was performed using D8 Discover system from Bruker Corp., Germany. In order to determine the crystallinity, the WAXD spectra has been modified to remove any background noise followed by fitting of the crystalline and amorphous halo using a special program written in the Microsoft Excel VBA . The percent crystallinity is then calculated by dividing the area under the crystalline peaks by the total area under the curve.
TERNARY PHASE DIAGRAM
Ternary phase diagram for water/DMF/PAN system was constructed using binodal and spinodal curves proposed by Tompa  based on the Flory-Huggins (FH) theory of polymer solutions . The Gibbs free energy of a mixer ([DELTA][G.sup.M]) for a ternary system is described as follows:
[mathematical expression not reproducible] (1)
where [n.sub.i], and [[empty].sub.i], are the number of mole and volume fraction of component i, respectively, R is the gas constant, T is the absolute temperature, and [g.sub.ij] is the interaction parameter between component i and j. The subscripts 7, 2, and 3 refer to nonsolvent, solvent, and polymer, respectively. Considering the binary interactions the chemical potentials of each component in a ternary solution are calculated as follows :
[mathematical expression not reproducible] (2)
[mathematical expression not reproducible] (3)
[mathematical expression not reproducible] (4)
where [DELTA][[mu].sub.i], denotes the chemical potential difference of component i in the mixture and the pure state, [v.sub.i] is the molar volume of component i, [[mu].sub.1] = [[empty set].sub.1]/([[empty set].sub.1] + [[empty set].sub.2]), and [u.sub.2] = [[empty set].sub.2]/([[empty set].sub.1] + [[empty set].sub.2]) are pseudo binary mixture. Based on the definition of the binodal curve, the chemical potentials of the polymer-rich phase and the polymer-lean phase at equilibrium are,
[DELTA][[mu].sub.i,A] = [DELTA] [[mu].sub.i,B]; i = 1,2,3 (5)
where the subscript A and B represent polymer-rich and polymer-lean phases, respectively. In addition, each component obeys the material conservation law as below:
[summation][[empty set].sub.i,A] = [summation][[empty set].sub.i,B] = 1; i = 1,2,3 (6)
Equations 5 and 6 include five coupled nonlinear equations with six unknowns such as [[empty set].sub.1,A], [[empty set].sub.2,A], [[empty set].sub.3,A], [[empty set].sub.1,B], [[empty set].sub.2,B], and [[empty set].sub.3,B]. By choosing one variable as independent (e.g., [[empty set].sub.3,A]), others can be determined using a set of interaction parameters.
The spinodal curve of a ternary system satisfies the following equation .
[G.sub.22] x [G.sub.33] = [([G.sub.23]).sup.2] (7)
where [mathematical expression not reproducible]
It follows, using Eq. 1
[mathematical expression not reproducible] (8)
[mathematical expression not reproducible] (9)
[mathematical expression not reproducible] (10)
and the material conservation law follows:
[summation][[empty set].sub.i] = 1; i = 1,2,3 (11)
Equations 7 and 11 include two coupled nonlinear equations with three unknowns such as [[empty set].sub.1], [[empty set].sub.2], and [[empty set].sub.3]. Again, by choosing one variable as independent (e.g., [[empty set].sub.1]), the spinodal curve can be obtained.
To construct a phase diagram numerically, binary interaction parameters of non-solvent/solvent, solvent/polymer, and non-solvent/polymer are necessary. The term [g.sub.12] in Eq. 7 is a generalized non-solvent/solvent interaction function and depends on the volume fraction [u.sub.2] of a pseudo binary mixture [38, 39]. The binary interaction parameters of water/DMF at 20[degrees]C and 40[degrees]C were collected from the literatures  and presented based on Koningsveld and Kleintjens model  (Eq. 12).
[mathematical expression not reproducible] (12)
where [alpha], [beta], and y are temperature-dependent constants. Non- solvent/polymer and solvent/polymer interaction parameters such as [g.sub.13] and [g.sub.23] often assumed to be temperature dependent but concentration independent [42-44]. The water/PAN interaction parameter (#13) can be evaluated by equilibrium swelling method [43, 44], Dong et al.  determined water/PAN interaction parameter ([g.sub.13]) as a function of temperature. In this study we have used [g.sub.13] parameter at 20[degrees]C and 40[degrees]C from Dong et al. . There are many approaches to determine the solvent/polymer (DMF/PAN) interaction parameters ([g.sub.23]), such as light scattering, gas-liquid equilibrium, and osmotic pressure [45, 46]. The data set of binary interaction parameters of DMF/PAN system were collected from literature . Knowing the interaction parameters for each component at 20[degrees]C and 40[degrees]C, the binodal and spinodal curves were calculated with the help of Lsqnonlin function in MATLAB program. More details of the mathematical/numerical treatment can be found elsewhere .
RESULTS AND DISCUSSION
Construction of Ternary Phase Diagram of Water/PANIDMF System
During electrospinning process, the liquid jet experiences solvent drying as well as water ingress. Both solvent evaporation and water ingress affect the free energy of the polymer solution. If the net energy change causes the ternary mixture thermodynamically unstable, phase separation occurs in the form of solvent-rich phase and polymer-rich phase [24, 48]. Two possible mechanisms can be attributed to phase separation : (i) nucleation and growth (NG) which leads to isolated pores and (ii) spinodal decomposition (SD) which leads to interconnected network of pores. The phase behavior of an initially homogenous polymer-solvent mixture such as PAN/DMF in a non-solvent vapor such as water can be described in terms of thermodynamic interactions of the individual components which can be visualized by ternary phase diagram [24, 49] as discussed below.
The ternary phase diagram of water/PAN/DMF system is shown in Fig. 2. The phase diagram is constructed using the interaction parameters at 20[degrees]C and 40[degrees]C as shown in Table 1. Interaction parameter for DMF/PAN at 40[degrees]C was interpolated between the reported values at 35[degrees]C and 50[degrees]C . As illustrated in Fig. 2, the binodal and spinodal curves divide the ternary phase diagram into three regions: miscible region (I), metastable region (II), and unstable region (III). The mechanism of phase separation process is different for different regions. In metastable region (region II), liquid-liquid demixing is governed by NG while SD is dominant in unstable region (region III). Careful observation reveals that at higher temperature the miscibility area (region I) becomes wider due to the differences in the binary interaction parameters of the components in the mixture.
Morphology of Electrospun Nanofibers
Effect of Humidity. Figure 3 shows the SEM images of the surface morphology of the nanofibers produced at 20[degrees]C under different RH conditions. At low RH (14% and 22%) the electrospun nanofiber shows almost texture free smooth surface. Although no pores were observed on fiber surface at 14% RH, a few isolated pores were observed at 22% RH. With increasing RH, the fiber surface becomes more irregular and textured due to possible phase separation as discussed below. When electrospinning is performed at 14% RH, the mass transfer path does not cross the binodal curve (Fig. 2a). At this condition, fiber surface morphology is dominated by the solvent drying phenomena due to longer phase separation time which resulted in smooth and texture free surface. As the humidity of the chamber increases, accumulation of water molecules in the polymer solution jet reduces the phase separation time, and as a result, cavities or pores are formed on the fiber surface due to the mass transfer path crossing the binodal curve. At 22% RH, the mass transfer path barely crosses the binodal curve, and hence, occasional small pores are observed on the fiber surface due to phase separation. At 30% or higher RH, the irregular structure on the surface of fibers resembles bicontinuous structure which is the characteristic feature of SD mechanisms. Considering the nonvolatile behavior of DMF and the initial composition of the polymer solution (10% PAN/DMF) being so close to the solvent-water axis (Fig. 2a), the metastable region at 20[degrees]C is very small to believe that the composition of the solution changes from metastable region to unstable region of the phase diagram. As a result the isolated pores are replaced by the interconnected network of pores on the surface with increasing RH in the chamber.
The dominating factor for bead formation during electrospinning is believed to be due to capillary instability [50, 51] as this event emerges from the surface tension of polymer solution and must be prevented to stabilize the fiber formation. It has been demonstrated that this type of instability can be effectively suppressed by viscoelastic stress . During electrospinning at higher humidity, ingress of water into the polymer solution jet raises the thermodynamic instability and evaporation of the solvent leading to increasing concentration of the polymer jet. Both events lead to increase in viscoelastic forces and eventually prevent the capillary instability to occur and as a result no bead formation occurs at higher humidity conditions. On the other hand, at low humidity (14% RH for example), the amount of water ingress into the polymer jet is very small, the viscoelastic stress is not enough to overcome the capillary instability which results in formation of beads.
SEM images of electrospun nanofibers produced at 20[degrees]C at lower magnification are shown in Fig. 4 (a-c) to discuss bead formation and orientation of the fibers. Bead formation is only seen at low humidity (14% RH) condition (Fig. 4a). It is also observed that nanofibers at lower humidity and temperature conditions the nanofibers are more random. With increasing RH the bead formation is not observed and also the fibers are found to be more aligned.
Close observation of the bead microstructure shown in Fig. 5 a reveals that a few small pores exists on the surface of the large elliptical bead which indicates that the composition path (arrow 1) has been changed to a new direction (arrow 2) crossing the binodal curve (dotted line in Fig. 5b) and finally ends up into the metastable region (region II in Fig. 5b). Once bead is formed a large amount of solvent (DMF in this case) is trapped inside which makes it longer for the bead to completely dry. This allows more water molecules to precipitate on the surface which may lead the composition path into the metastable state, and consequently, creates pores on the surface of the bead.
Effect of Temperature. Temperature is believed to influence the morphology of electrospun fibers by affecting the electrospinning environment, expansion of the miscibility area (Fig. 2b), and rate of mass transfer across the interface between the solution jet and the water vapor. SEM images of electrospun nanofibers were compared at two different temperatures (20[degrees]C and 40[degrees]C) under two humidity conditions such as 30% and 40% RH, as shown in Fig. 6.
As seen from the SEM images, significant surface irregularities were observed at 20[degrees]C due to phase separation for both humidity conditions. On the other hand, at 40[degrees]C, the fiber surface does not show any pores at 30% RH except a few uneven wrinkles on the surface as a result of impression left by water after drying. No phase separation occurred during electrospinning at 40[degrees]C-30% RH, whereas phase separation occurred at 40[degrees]C when humidity was increased to 40% RH. Thus it can be said that, the morphological transition from smooth to rough surface occurs at certain humidity conditions depending on the process temperature. Increasing the process temperature increases the solubility of water into the polymer solution without phase separation, meaning expansion of the miscibility area in the phase diagram (Fig. 2b). Thus, a delayed phase separation can be adopted at higher temperature. At 30% RH, water absorbed by the electrospinning jet was not high enough to induce phase separation at 40[degrees]C, while at 20[degrees]C the presence of water can trigger phase separation resulting in continuous interconnected network of pores on the surface of the fibers, by bringing the mass transfer path passed the binodal and spinodal curves of the ternary phase diagram. Careful observation of the fiber surface at 40[degrees]C-40% RH shows much rougher surface than those at 20[degrees]C-40% RH. This could be due to longer wait time before the mass transfer path crosses the spinodal curve, owing to the larger miscibility area in the ternary phase diagram at 40[degrees]C, which allows the fiber to absorb more water molecules and consequently brings more roughness.
Effect of Annealing. Figure 7 shows the effect of annealing on the surface morphology of PAN nanofibers electrospun under low and high humidity conditions. As seen there is not a very significant surface morphological change due to annealing at low humidity conditions, however at high humidity conditions annealing shows diminishing effect of surface porosity and alleviation of the surface roughness. It is to be mentioned that all as-spun fibers were dried at 60[degrees]C for 12 h and all annealed fibers were further heat treated in a vacuum oven at 95[degrees]C for 4 h. It is reported in the literature that the glass transition temperature is in the temperature range 90-120[degrees]C  and the annealing temperature is within the glass transition temperature range. It is expected that annealing near the glass transition temperature for several hours helps reduce the surface energy by smoothing out the surface roughness as well as better alignment of the internal microstructure.
Interior Structures of Electrospun Nanofibers
Figure 8a-d shows a few micrographs of the cross-section of the PAN nanofibers electrospun at 20[degrees]C under various RH conditions. At 60% RH, the highest RH in this study, nanofibers exhibit porous structure. However, the porosity decreased with decrease in RH and eventually, fibers with dense polymer structure was developed at 22% RH.
Once the electrospinning jet is exposed to the humid air, interactions among polymer, solvent, and non-solvent (water) molecules begin at the interface. With increasing humidity, the concentration of water in the polymer solution increases and as a result the mass transfer path moves the polymer-solvent axis towards the solvent-nonsolvent axis. The trajectory of the mass transfer path depends on the non-solvent concentration in the resulting polymer solution. If the mass transfer path does not cross the binodal curve, the polymer solution remains within the miscibility region and as a result the surface as well as the interior structure does not show any pores. On the other hand, if the mass transfer path crosses the binodal curve, the polymer solution becomes thermodynamically unstable, bringing the composition of the electrospinning jet into the metastable region where polymer-rich and polymer-lean domains are emerged. The former makes solid polymer phase while the later develops porous structure. However, the extent of the porous structure depends on whether the composition path passed the spinodal curve or not. For example, if the composition path ends into the metastable region and does not cross the spinodal curve, the porous structure can only be seen on the surface and does not extend to the interior of the fiber. On the other hand, if the composition path crosses the spinodal curve, phase inversion occurs by the means of SD mechanism which permits porous structure to extend to the interior of the fiber. Thus, electrospinning at 20[degrees]C under RH of 30% or higher, the mass transfer path must have crossed the spinodal curve of the ternary phase diagram which was also confirmed from surface morphology. At 30% RH, electrospinning at higher temperature such as 40[degrees]C produced dense internal structure, as shown in Fig. 8e compared to the microstructure at 20[degrees]C (Fig. 8b), as no phase demixing is expected for these fibers at higher temperature. Annealing near the glass transition temperature for several hours was also found to have significant effect on microstructure of the fiber as shown in Fig. 8f compared to as-pun fiber as shown in Fig. 8d. Annealing seems to reduce the internal porosity of the as-spun fibers significantly.
Effect of Humidity, Temperature, and Annealing on Diameter
Variations in nanofiber diameter as a function of RH at different process temperature and annealing are shown in Fig. 9. It can be seen that the diameter of nanofibers increases with increasing RH for both as-spun and annealed conditions. Increasing the nanofiber diameter with increasing RH can be explained by considering the rate of phase separation. At higher RH, more water can be diffused into the polymer solution and as a result phase separation occurs at a faster rate. This event accelerates the solidification process and as a result further stretching of the jet is hindered, and hence thicker fibers are produced. Effect of temperature on fiber diameter strongly depends on the state of phase separation. For example, at 30% RH, no phase separation was observed at 40[degrees]C and consequently the nanofiber resulted in lower diameter compared to at 20[degrees]C where phase separation was observed. Whereas at 40% RH the nanofiber experiences significant phase separation at both temperatures and as a result similar diameter was observed.
Fridrikh et al.  developed a simple model to predict the fiber diameter by analyzing the dynamic equations, describing the motion of whipping jet, of Hohman [54, 55]. Evaluating the asymptotic balance between normal stresses due to surface tension and surface charge repulsion, they obtained:
[mathematical expression not reproducible] (13)
where c is the polymer concentration; y is the surface tension (N/m); [bar.[epsilon]] is the dielectric constant of ambient air; Q is the flow rate ([m.sup.3]/s); I is the measured fiber current (A); [chi] is the dimensionless wavelength of the instability response for the normal displacements. Surface tension is a function of temperature and dielectric constant of air varies very slowly (2~4 X [10.sup.-4]%) with RH. Thus, at a constant temperature Eq. 13 can be reduced to,
[d.sub.fiber] = constant * [(I/Q).sup.-2/3] (14)
In Eq. 14, [d.sub.fiber] is a limiting (minimum) diameter which arises from the balance between normal stresses due to surface tension and surface charge repulsion; diffusion of nonsolvent (in our case water) molecule from the electrospinning environment into the solution jet may cause phase separation leading to arrest the jet prematurely with larger diameter. As fibers produced at 22% RH barely experienced phase separation; it can be assumed that 22% RH obey Eq. 14, and based on this the effect of volumetric charge density on nanofibers diameter alone can be predicted by considering no solidification takes place due to phase demixing. Table 2 reports the predicted diameter, actual diameter, and difference between the diameters for electrospun nanofiber at 20[degrees]C. The difference between actual diameter and predicted diameter based on volumetric charge density increases with increasing RH. This further confirms higher rate of phase demixing at higher RH which brings the electrospun jet to solidification stage much earlier and consequently thicker nanofibers are produced.
Mechanical Properties of Electrospun Nanofibrous Yarns
Effect of Humidity and Temperature. Representative stressstrain behavior of as-spun nanofibrous yarns prepared at different process temperatures and RH are shown in Fig. 10. Tensile strength was found to be the lowest for the yarns produced at highest RH. As humidity decreases tensile strength increases. The increase in tensile strength can be attributed to the reduction in average nanofiber diameter with decreasing humidity. However, yarns produced at 20[degrees]C and 14% RH did not exhibit further increase in strength despite having the lowest average nanofiber diameter. This could be due to the presence of beads in the yarns which could act as a defect in the yarn. Moreover, beads could contribute to the error in the cross-sectional area of the yarns as it was estimated based on the mass and length of the yarn.
At 22% RH the nanofibrous yarns not only show highest tensile strength but also exhibit highest ductility among all fibers investigated. Such a pronounced elasto-plastic behavior with large deformation to failure strain can have significant importance in many applications. A significant reduction in failure strain (47% to 11%) of the yarns was observed when RH was increased from 22% to 30%. These could be due to either huge increase in crystallinity or significant defects presence in the structure produced at 30% RH. Wide angle X-ray diffraction (WAXD) patterns along with calculated fractional crystallinity (using Eq. 13) are shown in Fig. 11. It is observed that there is no significant difference (~1%) in the crystallinity between the nanofibers produced at 22% RH and 30% RH. Thus huge reduction in failure strain and/or toughness (56.2 MJ/m3 to 11.5 MJ/ m3) could be due to presence of defects in the nanofiber. Nanofibers produced at higher RH undergo phase demixing which creates pores in the internal structures of the fibers. These pores act as stress concentration sites causing the fibers to fail prematurely. Similarly, at 40[degrees]C, the nanofibers produced at 30% exhibit higher tensile strength owing to their solid and/or less porous structure; whereas at 40% RH the nanofibers exhibit much lower tensile strength due to presence of porous structure as a result of phase demixing.
Effect of Annealing. The average tensile strength and strain at break of as-spun and annealed nanofibrous yarns at 20[degrees]C under various RH conditions are shown in Fig. 12. The results show significant improvement in the tensile strength of annealed nanofibers as compared to as-spun nanofibers. Annealing is often used to increase the degree of crystallinity of polymers which improves the mechanical strength . The percentage crystallinity extracted from the WAXD data indicates that the annealed fibers show about 10-12% increase in the crystallinity (Figs. 11 and 13).
As shown in Fig. 12, annealing also affects the strain at break of the nanofibrous yarns. At 22% RH, the failure strain decreased by 44% due to annealing. However, with the increase of RH the amount reduction in failure strain decreases and eventually increases by 16% at 60% RH. This trend can be explained by two contrasting influence of annealing on the nanofibers at higher RH. Annealing reduces the toughness of fibers by increasing the crystallinity while it reduces the structural defects of the fiber by alleviating the surface roughness and reducing the porosity in the structure, as discussed before. At 22% RH, as-spun nanofibrous yarns are mostly defect free and as a result exhibit large deformation to failure owing to their low crystallinity. Due to annealing crystallinity of those fibers increased and as a result the fibers failed solely due to higher crystallinity. In contrast, the porosity in the internal structure of as-spun nanofibers produced at high RH act as stress concentration sites and hence fibers exhibit low failure strain. Annealing reduces the porosity and consequently the difference between failure strain of as-spun and annealed yarns decreased and finally at 60% RH an improvement in failure strain was observed for annealed yarns.
This study demonstrated that both RH and temperature have significant effect on the surface morphology, microstructure, diameter, and tensile properties of electrospun PAN nanofibers. The interplay between solvent drying and water vapor ingress was found to be crucial factor for final morphology evolution of the fibers during electrospinning. At low RH, the solvent drying prevails and as a result solid and smooth nanofibers are produced, while VIPS plays a vital role at high RH and as a result porous nanofibers with large diameter and rough surface roughness are observed. Ternary phase behavior of water/DMF/PAN system can be used as a powerful and effective tool to explain the evolution of morphology of the electrospun fibers. The size of the miscible area in the ternary phase diagram changes with process temperature which affect the RH condition at which the transition from solid structure to porous structure occurs. At 20[degrees]C process temperature the transition was observed in between 22-30% RH, while 30-40% RH was found for 40[degrees]C due to an increase in the size of miscibility area. At very low RH, a few large size beads were found as a consequence of capillary instability which emerges from surface tension of polymer solution and can be prevented by increasing RH. Thus, the morphology of electrospun nanofiber can be maneuvered by changing RH and temperature during electrospinning.
Tensile properties of electrospun nanofibrous yarns were found to be affected by the diameters as well as surface morphology and formation of beads. At 20[degrees]C, the highest tensile strength and strain at break were observed for yarns produced at 22% RH due to their solid structure, smaller diameter, and bead free fibers. With increasing RH to 30%, both tensile strength and stain at break were found to decrease due to the porous structure that formed as a result of phase separation. Annealing was found to improve the strength and reduce the strain at break for solid nanofibrous yarns. However, annealing improved the overall performance of porous nanofibrous yarns by alleviating surface roughness and reducing internal porosity.
We thank Dr. Altan for the high voltage power system. We also thank Dr. Grady for his invaluable help with the WXRD experiments.
[1.] Z.M. Huang, Y.Z. Zhang, M. Kotaki, and S. Ramakrishna, Compos. Sei. Technol., 63, 2223 (2003).
[2.] T.J. Sill, and H.A. von Recum, Biomaterials, 29, 1989 (2008).
[3.] S. Ramakrishna, K. Fujihara, T. Wee-Eong, T. Yong, Z. Ma, and R. Ramaseshan, Mater. Today, 9, 40 (2006).
[4.] N. Bhardwaj, and S.C. Kundu, Biotechnol. Adv., 28, 325 (2010).
[5.] J. Doshi, and D.H. Reneker, J. Electrostat., 35, 151 (1995).
[6.] D.H. Reneker, and I. Chun, Nanotechnology, 7, 216 (1996).
[7.] J.J. Feng, Phys. Fluids, 14, 3912 (2002).
[8.] RX. Ma, and R.J. Zhang, J. Biomed. Mat. Res., 46, 60 (1999).
[9.] G.M. Whitesides, and B. Grzybowski, Science, 295, 2418 (2002).
[10.] P. Gupta, C. Elkins, T.E. Long, and G.L. Wilkes, Polymer, 46, 4799 (2005).
[11.] X. Geng, K. Oh-Hyeong, and J. Jang, Biomaterials, 26, 5427 (2005).
[12.] H. Fong, I. Chun, and D.H. Reneker, Polymer, 40, 4585 (1999).
[13.] C. Zhang, X. Yuan, L. Wu, Y. Han, and J. Sheng, Eur. Polym. J., 41, 423 (2005).
[14.] J. Du, S. Shintay, and X. Zhang, J. Polym. Sei., Part B: Polym. Phys., 46, 1611 (2008).
[15.] A.A. Ashraf, and M.A. El-Hamid, Composites Part A, 37, 1681 (2006).
[16.] P. Heikkila, and A. Harlin, Eur. Polym. J., 44, 3067 (2008).
[17.] W.E. Teo, and S. Ramakrishna, Nanotechnology, 17, R89 (2006).
[18.] M. Pakravan, M.C. Heuzey, and A. Ajji, Polymer, 52, 4813 (2011).
[19.] S.D. Vrieze, T.V. Camp, A. Nelvig, B. Hagstrom, P. Westbroek, and K.D. Clerck, J. Mater. Sei., 44, 1357 (2009).
[20.] B. Barua, and M.C. Saha, J. Appl. Polym. Sei., 132, 41918 (2015).
[21.] S. Tripatanasuwan, Z. Zhong, and D.H. Reneker, Polymer, 48, 5742 (2007).
[22.] Y. Cai, and M. Gevelber, J. Mater. Sei., 48, 7812 (2013).
[23.] J. Pelipenko, J. Kristl, B. Jankovic, S. Baumgartner, and P. Kocbek, Int. J. Pharm., 456, 125 (2013).
[24.] H. Fashandi, and M. Karimi, Ind. Eng. Chem. Res., 53, 235 (2013).
[25.] P. Lu, and Y. Xia, Langmuir, 29, 7070 (2013).
[26.] M. Srinivasarao, D. Collings, A. Philips, and S. Patel, Science, 292, 79 (2001).
[27.] C.L. Pai, M.C. Boyce, and G.C. Rutledge, Macromolecules, 42, 2102 (2012).
[28.] P. Dayal, J. Liu, S. Kumar, and T. Kyu, Macromolecules, 40, 7689 (2007).
[29.] H. Fashandi, and M. Karimi, Thermochim Acta, 547, 38 (2012).
[30.] S.K. Nataraj, K.S. Yang, and T.M. Aminabhavi, Prog. Polym. Sei., 37, 487 (2012).
[31.] O.P. Bahl, Z. Shen, J.G. Lavin, and R.A. Ross, in Carbon Fibers, J.B. Donnet, T.K. Wang, S. Rebouillat, and C.M. Peng, Eds., Marcel Dekker, New York (1998).
[32.] L. Huang, N. Bui, S.S. Manickam, and J.R. McCutcheon, J. Polym. Sei. part B: Polym. Phys., 49, 1734 (2011).
[33.] X. Yu, H. Xiang, Y. Long, N. Zhao, X. Z, and J. Xu, Mater. Lett., 64, 2407 (2010).
[34.] B. Barua, Investigation of Electrospinning Process Parameters and Studies of Stabilization Kinetics of Polyacrylonitrile-based Electrospun Carbon Nanofibers. PhD diss., Unversity of Oklahoma, Norman (2015).
[35.] http://coecs.ou.edU/Brian.P.Grady/saxssoftware.html (Last accessed: July 10, 2016).
[36.] H. Tompa, Polymer Solutions, Butterworths, London (1956).
[37.] P.J. Flory, Principles of Polymer Chemistry, Cornel University Press, New York, (1953).
[38.] L. Yilmaz, and A.J. McHugh, J. Appl. Polym. Sei., 31, 997 (1986).
[39.] F.W. Altena, and C.A. Smolders, Macromolecules, 15, 1491 (1982).
[40.] H. Fashandi, and M. Karimi, Polymer, 53, 5832 (2012).
[41.] R. Koningsveld, and L.A. Kleintjens, Macromolecules, 4, 637 (1971).
[42.] J. Zhang, Y. Zhang, and J. Zhao, Polym. Bull., 67, 1073 (2011).
[43.] L. Tan, D. Pan, and N. Pan, J. Appl. Polym. Sei., 110, 3439 (2008).
[44.] R. Dong, J. Zhao, Y. Zhang, and D. Pan, J. Polym. Sei., 47, 261261 (2009).
[45.] N. Schuld, and B.A. Wolf, In Polymer Handbook, J. Brandrup, E.H. Immergut, and E.A. Grulke, Eds., Wiley, New York (1999).
[46.] M. Karimi, W. Albrecht, M. Heuchel, T. Weigel, and A. Lendlein, Polymer, 49, 2587 (2008).
[47.] R.B. Beevers, J. Polym. Sei.: Macromol. Rev., 3, 113 (1968).
[48.] S.P. Nunes, and T. Inoue, J. Membrane. Sei., Ill, 93 (1996).
[49.] H.C. Park, Y.P. Kim, H.Y. Kim, and Y.S. Kang, J. Membrane. Sei., 156, 169 (1999).
[50.] Y.M. Shin, M.M. Hohman, M.P. Brenner, and G.C. Rutledge, Polymer, 42, 09955 (2001).
[51.] K.H. Lee, H.Y. Kim, H.J. Bang, Y.H. Jung, and S.G. Lee, Polymer, 44, 4029 (2003).
[52.] L. Vaisman, E. Wachtel, H.D. Wagner, and G. Marom, Polymer, 48, 6843 (2007).
[53.] S.V. Fridrikh, H.Y. Jian, M.P. Brenner, and G.C. Rutledge, Phys. Rev. Lett., 90, 144502 (2008).
[54.] M.M. Hohman, M. Shin, G. Rutledge, and M.P. Brenner, Phys. Fluids, 13, 2201 (2001).
[55.] M.M. Hohman, M. Shin, G. Rutledge, and M.P. Brenner, Phys. Fluids, 13, 2221 (2001).
[56.] E.P.S. Tan, and C.T. Lim, Nanotechnology, 17, 2649 (2006).
Bipul Barua, (1) Mrinal C Saha (DI) (2)
(1) Division of Nuclear Engineering, Argonne National Laboratory, Lemont, Illinois 60439
(2) School of Aerospace and Mechanical Engineering, University of Oklahoma, Norman, Oklahoma 73019
Correspondence to: Mrinal C. Saha; e-mail: email@example.com
The content of this paper has been taken from Ph.D. dissertation of coauthor, Bipul Barua, who holds the copyright of the dissertation. The online version of the dissertation can be found at https://shareok.org/handle/11244/ 15490 (Last accessed: July 10, 2016).
Part of this work has been presented at AIAA/ASME Oklahoma Symposium XXXIV on March 29, 2014.
Caption: FIG. 1. Schematic representation of electrospinning process.
Caption: FIG. 2. Ternary phase diagram of water/DMF/PAN system at (a) 20[degrees]C and (b) 40[degrees]C. Three regions in the diagrams are miscible region (I), metastable region (II), and unstable region (III).
Caption: FIG. 3. SEM images showing different surface morphologies of electrospun nanofibers produced at 20[degrees]C under various RH conditions.
Caption: FIG. 4. SEM images showing formation of beads (indicated by circles) and orientation of nanofibers (at lower magnification) electrospun at 20[degrees]C.
Caption: FIG. 5. (a) Photomicrograph shows large elliptical bead during electrospinning of nanofibers at 20[degrees]C and 14% RH. (b) Ternary phase diagram showing possible composition paths for nanofibers without bead (arrow 1) and with bead (arrow 2). Dashed line: binodal curve; solid line: spinodal curve; I, II, and III are miscible, metastable, and unstable regions, respectively.
Caption: FIG. 6. Surface morphology of as-spun nanofibers electrospun at 20[degrees]C (top) and 40[degrees]C (bottom) at two different humidity conditions (30% and 40% RH).
Caption: FIG. 7. Photomicrographs showing surface morphology of as-spun (top) and annealed (bottom) nanofibers at two different humidity conditions (14% and 60% RH).
Caption: FIG. 8. Photomicrographs showing cross-sections of the as-spun (a-e) and annealed (f) nanofibers under various environmental conditions.
Caption: FIG. 9. Effects of RH, temperature, and annealing on the average diameter of nanofibers.
Caption: FIG. 10. Representative stress-strain curves for nanofibrous yarns electrospun at (a) 20[degrees]C and (b) 40[degrees]C under various RH conditions.
Caption: FIG. 11. Integrated WAXD plots and calculated percentage crystallinity of as-spun nanofibrous yarns electrospun at (a) 40[degrees]C-30% RH, (b) 20[degrees]C-22% RH, (c) 20[degrees]C-30% RH, and (d) 20[degrees]C-60% RH. Intensity profiles of integrated scans are shifted upward for clear comparison.
Caption: FIG. 13. Integrated WAXD plots and calculated percentage crystallinity of annealed nanofibrous yarns electrospun at 20[degrees]C and various RH conditions. Intensity profiles of integrated scans are shifted upward for clear comparison.
TABLE 1. Interaction parameters for ternary phase diagram of water/ DMF/PAN system. [g.sub.12] ([[empty set].sub.2]) = [alpha] + [[beta]-(1 - [gamma][[empty Temperature [[empty set].sub.2])] (a) 20[degrees]C 0.218+[0.276/(1 +0.622 [[empty set].sub.2])] 40[degrees]C 0.425 + [0.0619/ (1 - 0.94940 [[empty set].sub.2])] Temperature [g.sub.13] (b) [g.sub.23] (c) 20[degrees]C 1.65 0.280 40[degrees]C 1.35 0.292 (a) Data taken from ref. 40. (b) Data taken from ref. 44. (c) Data taken from ref. 47. TABLE 2. Effect of volumetric charge density on the average diameter of nanofibers electrospun at 20[degrees]C. Predicted nanofiber diameter based on volumetric charge Measured average density nanofiber diameter [d.sub.predicted] [d.sub.measured] RH (%) (nm) (nm) 22 (control) -- 282 30 310 351 40 328 442 50 365 564 60 375 664 [d.sub.measured] - [d.sub.predicted] RH (%) (nm) 22 (control) -- 30 41 40 114 50 199 60 289 FIG. 12. Effect of annealing on tensile strength and strain at break of nanofibrous yams produced at 20[degrees]C under various RH conditions. Numbers showing on histograms are percentage change in strength and strain at break due to annealing. Tensile Strength (MPa) As-spun Annealed 22 +25% 30 +14% 40 +11% 50 +11% 60 +18% Strain at Break (%) 22 -44% 30 -27% 40 -26% 50 -19% 60 -16% Note: Table made from bar graph.
|Printer friendly Cite/link Email Feedback|
|Author:||Barua, Bipul; Saha, Mrinal C.|
|Publication:||Polymer Engineering and Science|
|Date:||Jun 1, 2018|
|Previous Article:||Confocal Raman Spectroscopy Studies on the Mutual Diffusion Behavior at the Interface Between Two Different Polyesters.|
|Next Article:||A Review of Conductive Polymer Composites Filled With Low Melting Point Metal Alloys.|