ELECTROSTATIC SPINNING OF ULTRAFINE POLYMER FIBERS B. Torres, M. J. Díaz-de León and N. J. Pinto Department of Physics and Electronics University of Puerto.

ELECTROSTATIC SPINNING OF ULTRAFINE POLYMER FIBERS 

B. Torres, M. J. Díaz-de León and N. J. Pinto 

Department of Physics and ElectronicsUniversity of Puerto Rico

Humacao, PR 00791ABSTRACT

A relatively simple non-mechanical technique of preparing ultra fine fibers of polymers (conducting and non-conducting) is presented. Although discovered in the 1930’s[1] this electrospinning technique has received little attention with very few groups around the world actively involved in research using this method of fiber preparation. In a typical experimental setup the polymer to be electrospun is dissolved in a low boiling point solvent like chloroform and inserted into a glass pipette. A copper wire (anode) inserted into the pipette making contact with the solvent is then connected to a high voltage source. The pipette is placed a few degrees below the horizontal to obtain a slow flow rate of the polymer solution. As the voltage is increased there comes a point when the electrical forces overcome the surface tension of the polymer droplet formed at the end of the pipette and a continuous fine jet is formed that travels toward a cathode which is placed a few centimeters from the anode. As the solvent evaporates ultra fine fibers of the polymer of diameters in the range 1μm down to few tens of nanometers deposit themselves on the cathode. Pure conducting polymers like polyaniline and polyaniline blended with non conducting polymers like polyethylene oxide and polystyrene have been electrospun. Single fibers have been captured on Si/SiO2 wafers by intercepting the

jet momentarily. The dc conductivity of some of these fibers has been measured and the effects of an external electric field on the observed conductivity investigated.

Figure 1: (a) Insulating emeraldine base form of PANi, (b) conducting emeraldine salt form of PANi.

INTRODUCTION

Ever since miniaturization of electronic items hit the markets in the early 1970’s, people have been fascinated in nanotechnology. For this reason, many scientists worldwide have devoted their efforts to study if fundamental processes observed in the bulk material are affected by reducing sample dimensions. Last year, then President Clinton’s nanotechnology initiative led to an increase in the amount of funding into this area of research. The National Science Foundation definition of a nano object is that at least one dimension of the sample be less than 100 nm (1 nm = 10-9 m). If nanotechnology reaches even one-tenth of its projected capabilities, the effects will be nothing short of revolutionary in a broad spectrum of areas. With this potential still untapped, ways to produce materials with dimensions < 100 nm are of great importance.

Polystyrene (PS) and polyethylene oxide (PEO) are two commercially available non conducting polymers that can be purchased having different molecular weights. Being non conducting their potential applications in electronic circuits are limited and so blends of PS and PEO with polyaniline (PANi) were prepared. PANi is an organic conducting polymer that was first reported in the literature about 100 years ago. The insulating emeraldine base form of polyaniline (PANiEB) consists of equal numbers of reduced [-(C6H4)-N(H)-(C6H4)-N(H)-] and

oxidized [-(C6H4)-N=(C6H4)=N-] repeat units as seen in Figure 1. The conducting emeraldine

salt form (PANiES) is achieved by doping with aqueous protonic or functionalized acids where protons are added to the -N= sites while maintaining the number of electrons in the polymer chain constant. Most of the research carried out in the past two decades were done on electrical characterization of bulk films with very few studies on ultrathin (diameter < 100 nm) fibers[2].

EXPERIMENTAL

1. Electrospinning Apparatus: The basic elements of the electrospinning apparatus is shown in Figure 2. A high voltage dc source is connected to a hypodermic needle that contains the polymer dissolved in a proper solvent. A metal cathode (Al foil) is placed about 15 cm from the tip of the needle at an angle of about 5o below the horizontal to permit a slow flow of the solution under gravity of approximately one drop every 30 seconds. As the voltage to the needle is increased the electric forces on the drop overcome the surface tension and fine jet is forced towards to cathode. As the solvent evaporates ultra fine fibers are collected on the cathode. We have used slight variations of this technique for the work presented here. As for example, we used a glass pipette instead of a hypodermic needle with a copper wire inserted that acted as the anode.

3. Sample preparation: (Manuel Díaz) 100 mg of PANiEB was mixed with 129 mg of camphor sulfonic acid (CSA) in a mortar and pestle and then added to 10 ml of chloroform. After about 6 hours of constant stirring the solution was filtered. 400 mg of PS was dissolved in the above solution and electrospun to get PANi/PS fibers. For PANi/PEO fibers, 100 mg of PANEB was mixed with 129 mg of CSA in a mortar and pestle and then added to 10 ml of chloroform. After about 6 hours of constant stirring the solution was filtered and 100 mg of PEO was dissolved in the above solution and electrospun. For the pure PANi fibers we used PANi doped with AMPSA dissolved in concentrated sulfuric acid and this was provided by Dr. MacDiarmid at UPENN. The pure PANi fibers were prepared by dissolving 900 mg of PANiAMPSA in 3 ml of concentrated sulfuric acid. The solution was placed in a glass pipette and the pipette was held vertically down with the outlet about 3 cm from a beaker of water that acted as the cathode. The spinning voltage used was 15 kV, this resulted in jets emerging from the end of the pipette (sometimes with sparks) in an erratic manner and depositing itself on the surface of the water in the form of fine fibers. After allowing 12 hours for the sulfuric acid to dissolve in the water the fibers were collected on a Si/SiO2 wafer. For the blends the

pipette was held at about 5o below the horizontal position and the applied electric field was about 1 kV/cm and the jet was much more consistent that in the previous sample. The fibers were collected on Si/SiO2 wafer by intercepting the jet with the wafer for a few seconds.

2. Sample preparation: (Billy Torres) Polystyrene of molecular weights 280000, 28000 and 2430 together with THF were purchased from Aldrich Chemicals. Four different solutions of PS and THF were prepared and electrospun. Sample A was prepared by dissolving PS (MW=280,000) in THF and having 15 wt.% in solution. Sample B was prepared by dissolving PS (MW=280,000) and PS (MW=28,000) in the ratio 90/10 in THF and having a 15 wt.% in solution. Sample C was prepared by dissolving PS (MW=280,000) and PS (MW=28,000) in the ratio 50/50 in THF and having a 15 wt.% in solution. Sample D was prepared by dissolving PS (MW=280,000) and PS (MW=2,430) in the ratio 90/10 in THF and having a 15 wt.% in solution. The spinning conditions were typically applying 20 kV to the anode and having the Al foil cathode placed about 20 cm from the anode. Pieces of the Al coated foils were then coated with gold and placed in a SEM for surface characterization.

OBJECTIVE

Our objective is to prepare fibers of polymers, conducting and/or non conducting that have an average diameter < 100 nm. In addition, the conducting polymer fibers were electrically characterized at room temperature and the dependence of the dc conductivity on an externally applied elelctric field was investigated. The objective of this being to fabricate nanoelectronic FET’s.

Figure 2: Basic electrospinning apparatus

RESULTS AND DISCUSSION

 Figure 3 shows the SEM photograph of sample A. This is a low magnification of the “fibers” and this topography was typical when scanned over different portions of the fiber mat.

 Figure 3: SEM photograph of sample A showing the formation of beads on the fibers.

The electric field gradient of 0.67 kV/cm was used in the electrospinning process.

This sample was electrospun using a very long syringe needle (length of the syringe was about 30 cm) to check if there would be any polymer alignment in the spun fibers due to the longer distance of travel before splaying of the jet began. Fiber diameters were in the range 500 nm to 4000nm. The presence of beads on some fibers could be a result of the high viscosity of the solution. Such beaded effects were also observed previously in solutions of polyethylene oxide and ethanol[3].

Figure 4: SEM photograph of sample B at two different applied electric fields. Left: 1 kV/cm; Right: 0.67 kV/cm.

 Figure 4 shows the SEM photograph of sample B at a low magnification typical of the sample surface under slightly different spinning voltages. Qualitatively there does not appear to be much difference in the two photographs implying that a slight change in the electric field gradient did not affect the nature of the fibers. There appears to be many hemispheres in both figures and on further magnification as seen in Figure 5, these mushroom like structures have some fine structure.

Figure 5: Close up view of Figure 4 showing “mushroom” like structures. The average pore diameters were in the range of a few 100 nm.

The porous features in Figure 5 have an average diameter of about a few 100 nm. In contrast to Figure 3, these are not beaded and could be a result of nucleation of the high molecular weight polymer around the low molecular weight polymer which act as “seeds”. The fibers connecting these mushroom like structures had diameters in the range 80 nm – 1000 nm. In order to check the concentration dependence of these mushroom like structures on the relative concentrations of the high molecular weight and low molecular weight species sample C was prepared with a 50/50 percent of the two polymers in solution with THF. Figure 6 shows a typical SEM image of the sample surface.

 Figure 6: SEM of a sample C having 50/50 mixture of molecular weights of PS dissolved in THF.

When Figure 6 is compared to Figure 4 one notices fewer mushroom like structures. The higher concentration of the low molecular weight polymer resulted on fewer fibers with diameters in the range 200 nm – 3000 nm. To check the dependence of the fibers on the presence of the low molecular weight polymer, sample D was prepared and electrospun. Figure 7 shows typical a SEM photograph of this sample surface. The thinnest fiber measured was 71 nm. The presence of fewer mushroom like structures could be as a result of the inability of the high molecular weight polymer to efficiently

Figure 7: SEM of sample D. The thinnest fiber was observed in this sample with diameter 71 nm.

 condense on the low molecular weight “seeds”. We would like to point out that all the SEM studies done on samples with mixed molecular weights did not show the beaded structure seen in Figure 3 where a single molecular weight polymer was used.

Figure 8 shows the IV curve for a pure PANiAMPSA fiber that had an average diameter of 10 m. The IV curve shows a linear relationship indicating that power dissipation was negligible within the sample and the calculated conductivity of 0.76 S/cm represents the intrinsic conductivity of the fiber. The conductivity of similar fibers was reported to be of the same order of magnitude[4]. The data in Figure 8 were taken with the gate voltage set to 0 V and 10 V. There is no apparent field effect on the conductivity suggesting that perhaps the fiber is not in the semiconducting regime but rather in the metallic regime of charge transport where field effects are negligible.

Figure 8: IV curve for a pure PANi single fiber of diameter 10m.

Figure 9 shows a photographic image of the pure PANi fiber studied above mounted on a degenerately doped Si/SiO2 wafer with three leads that have been connected with silver paste.

These include two leads for the bias voltage and a third lead for the gate voltage. The average diameter of the fiber was 10 μm.

Figure 9: PANiAMPSA single fiber cast on a Si/SiO2 wafer showing the gate and bias voltage

connections.

Figures 10(a) and (b) show AFM images of PANi/PEO fibers. As seen in these figures (which are representative of the fibers electrospun) one can obtain meter long fibers with controlled fiber collection techniques. The fibers in this sample have a minimum diameter of 500 nm and a maximum diameter of 3 μm and show some structure. These features appear as beads formed at approximately regular intervals giving the fibers a bamboo stem like feature. One possible reason for this could be that the PEO did not dissolve completely or the solution was not homogenous enough prior to spinning. It could also be due to the effects of surface tension which tries to reduce the surface area per unit mass[3]. One observes from Figure 10(a) that the fibers deposit themselves in a sinusoidal fashion with a branching effect in the upper right hand corner of Figure 10(a). This leads us to conclude that the thick fibers are composed of perhaps a bundle of smaller diameter fibers running parallel to the axis of the thick fiber. Figure 11 (a), (b) and (c) show AFM images of PANi/PS fibers. These fibers clearly have a smaller diameter than the PANi/PEO fibers with the thinnest fiber reported to be 50 nm. One possibility of obtaining thinner diameter fibers of PANi blends with PS is that the PS used was a standard polymer and so had few if any of defects. Fewer defects means fewer branches that could lead to polymer entanglement with PANi and hence thicker fibers of the belnds when electrospun as is the case of PANi/PEO. It could also mean that PS form a more uniform blend with PANi in chloroform The morphology of the PANi/PS fibers as seen in Figures 11(a) and 11(b) are also very uniform supporting this idea and suggesting better chain alignment in the blends. Once again in Figure 11(c) we notice bundles of fibers of smaller diameter bunched together forming a thicker fiber. Since the conductivity in PANi is very sensitive to the amount of defects it is reasonable to suggest that the overall conductivity of the blended fibers will be higher in the PANi/PS blends than in the PANi/PEO blends. Studies to measure these conductivities are in progress and will be reported at a later date. Although not studied in this report we plan to measure the conductivity of these PANi blends and also check to see if any field effect on the conductiivty is observed. We also would like to point out the regions Figures 11(a) and (b) there two polymers cross, this polymer-polymer junction is also of interest to us in making nanoelectronic junction devices. We are also in the process of trying out various other solvents other than chloroform, like tetrahydrofuran and acetonitrile to see which solvent yields thinner and more uniform fibers of PANi blends.

Figure 11: AFM images of PANi/PS electrospun fibers at 1kV/cm.

Figure 10: AFM images of PANi/PEO electrospun fiberselectrospun at 1kV/cm.

CONCLUSION:

While the diameter of the fibers formed did not have any correlation with the differing molecular weights of PS in THF, we did observe some interesting results. The formation of mushroom like features on the spun fibers was most unexpected. The extremely large surface to volume ratio of the polymer matrix could be exploited for many uses for example as a catalyst or in the preparation of nanopore filters. Further, some nanofibers are seen to hang in air between these mushrooms and lend themselves to being “picked” up and isolated from the rest making single fiber characterization easier.

Fibers of pure PANi were spun and their room temperature conductivity was measured to be similar to the bulk value. No field effect was observed on the conductivity of these fibers suggesting that the sample was in the metallic regime rather that in the semiconducting regime where field effects are generally seen. In the polyaniline blended with PS and PEO it was found that PS is a better polymer to use on order to get thinner (< 100 nm) diameter fibers. This could be due to the availability of a standard with fewer defects as compared to PEO.

 Acknowledgements:

The authors would like to thank Dr. Junbo Gao and Dr. Alan MacDiarmid at UPENN and Dr. Frank Ko at Drexel University for their collaboration on this project. Special thanks to Jorge Colón for AFM studies. This work was supported by NSF under grants DMR-9872689 and DMR-0098603.

References:

[1]A. Formhals, US Patent #1,975,504 (1934).

[2]I.D. Norris, M.M. Shaker, F.K. Ko and A.G. MacDiarmid, Synth. Metals 114, 109 (2000).

[3]H. Fong, I. Chun and D.H. Reneker, Polymer 40, 4585 (1999).

[4]A.G. MacDiarmid et al., Synth. Metals (2001) in press.