Low-Cost Carbon Fibers

Andrea Forrest, Louisiana Tech University
Jon Pierce, Alfred University
Willie Jones, Clemson University
Esther Hwu, Emory University

Introduction

Environmental regulations on automobile emissions increase the need for lighter automobiles. One way to reduce vehicle weight without losing any structural and mechanical strength is to use carbon fiber reinforced composites. Unfortunately, high production costs have limited the widespread use of carbon fibers. The current cost of carbon fibers is approximately $15/pound. In order for carbon fibers to be used widely, the costs need to be lowered to $5/pound or less. To reach this goal, alternate precursors need to be explored and the heat treatments involved in stabilizing and carbonizing the fibers need to be optimized. Therefore, the goal of this research was to explore alternate, potentially low-cost routes for carbon fiber production.

Currently 90% of all commercial carbon fibers are produced by the thermal conversion of a polyacrylonitrile (PAN) precursor fiber (1, 2). In these processes a solution of 10 to 30% by weight of a PAN copolymer in a highly polar solvent such as sodium thiocyanate or dimethylacetamide is extruded through a multihole spinneret into a coagulation bath. As the solution exits the multihole spinneret and enters the bath it precipitates forming a multifilament bundle. The bundle is washed to remove excess solvent, stretched to increase the molecular orientation within the filaments, and finally wound onto a takeup bobbin. Next the PAN precursor fiber is heated in air to temperatures ranging from 230 to 280°C for times ranging from 30 minutes to two hours. During this step oxygen diffuses into the fibers and crosslinking occurs. Finally, the crosslinked fibers are heated in an inert environment to approximately 1300°C. This drives off the non-carbon elements, yielding a carbon fiber.

One obvious method for reducing the price is to utilize melt-spinning rather than solution spinning to produce the precursor fiber. Melt-spinning is inherently less expensive than solution-spinning because it eliminates the extra cost of solvent handling and it also operates at much higher throughputs. Another possible approach is to employ a precursor that contains more carbon. PAN precursor fiber actually contains only 68% carbon. Therefore, two pounds of PAN precursor fiber normally yield about a pound of carbon fiber. By comparison, potential carbon fiber precursors such as pitch and high carbon content polymers contain between 80 to 90% carbon. If one of these materials could be melt spun and successfully converted into carbon fibers, the overall process conversion could be substantially increased. In this preliminary study, both approaches were explored and both appear to offer promise for reducing process costs.


Materials and Methods

The materials investigated in our experiments were a Mitsubishi® PAN (95% PAN, 5% methyl acrylate) pre-spun fiber, and isotropic pitch, and Amlon® (80% PAN, 20% methylacrylate) precursors. The Mitsubishi® PAN is a commercial precursor fiber that is solution-spun. This fiber served as a control for comparing the final carbonized properties of the melt-spun fibers produced in our study. The Amlon® PAN copolymer was supplied by BP-Amoco, and the isotropic pitch was supplied by Conoco, Inc. These two precursors melt below their degradation temperatures (the prerequisite for melt-spinning). Also, the carbon content of the isotropic pitch was greater than 85%. Thus, the Amlon® precursor allowed us to explore the advantages, as well as evaluate the problems, associated with producing carbon fiber using a melt-spinnable PAN precursor. By contrast, the isotropic pitch allowed us to evaluate the benefits as well as penalties of producing carbon fiber with a high-carbon content precursor.

A Haake rheometer was used to determine the optimum spinning temperature for the Amlon® and the isotropic pitch precursors. Then, a batch melt spinning apparatus was used to melt-spin fibers from the two precursors (Figure 1). This involved first attaching a multihole spinneret to the bottom of the melting chamber of the apparatus. Then, solid pellets of each precursor were loaded into the chamber. Next, the extrusion piston was inserted and the chamber was purged with nitrogen. An electrical heating system was then attached, and the chamber heated to the optimum spinning temperature. After the precursor temperature reached equilibrium, the piston drive motor was turned on and set to a predetermined speed. This lowered the piston at a fixed speed, forcing the precursor through the multihole spinneret at a fixed rate. The surrounding air-cooled the molten precursor as it emerged from the spinneret capillaries, forming fibers. These fibers were collected on a takeup roll that was set to a predetermined rotation rate.

Figure 1. Batch melt-spinning apparatus.

A two-step process was used in our initial attempts to convert these melt-spun precursor fibers into carbon fibers. As a control, the Mitsubishi® PAN (95% PAN, 5% methyl acrylate) pre-spun fiber was also converted to carbon fibers using the same two-step process. This allowed us to compare the final properties of carbon fibers produced from the two melt-spun precursors with those of a commercial solution-spun precursor fiber that had been subjected to a similar conversion process.

In the first step of this thermal treatment process, the precursor fibers were inserted in an oven with air circulation, heated to just below their softening point, and held at this temperature for a set amount of time. A flame test was used as a measure of complete stabilization. This involved placing the stabilized fibers in a flame. If the fibers glowed red and did not melt, stabilization was considered to be successful. In the second and final step of the thermal treatment process, the stabilized fibers were placed in an Astro furnace, purged with helium and heated to either 1200°C or 1500°C for 30 minutes.

Once the fibers were carbonized, they were mounted individually on tabs and tested using the MTI for tensile properties. Fibers were mounted on glass slides for Raman spectroscopy analysis. Also, fibers were mounted on cartridges for the S-4700 scanning electron microscope.

Results

The Mitsubishi® PAN-based pre-spun fibers were used as a control in our experiments. The fibers were stabilized for two hours at 230°C, 240°C, or 250°C. Then they were carbonized at 1200°C and 1500°C for 30 minutes. The mechanical testing results for the Mitsubishi® PAN fibers are shown below in Figures 2 and 3.

Figure 2: Strength of Mitsubishi® Pan		Figure 3: Modulus of Mitsubishi® Pan

Isotropic pitch-based fibers were spun at 250°C, and then were chemically and thermally stabilized. The fibers were chemically stabilized by soaking them in a 15wt% to 70wt% nitric acid solution for various lengths of time. These fibers were then carbonized but could not be tested because they were stuck together. Thermal stabilization followed a four-day heating schedule: 180°C for 24hrs, 200°C for 24hrs, 220°C for 24hrs, and 240°C for 24hrs. Some fibers were heated at either 260°C or 270°C for two additional hours. These fibers were then carbonized at 1200°C and 1500°C for 30 minutes. Isotropic pitch fibers that were stabilized at 260°C and 1200°C had a tensile strength and modulus close to those of commercially produced pitch fibers. The mechanical testing results for isotropic pitch are shown in Figures 4 and 5.

Figure 4		Figure 5

Amlon® fibers were spun at 230-235°C, and then stabilized using the following schedule: 180°C for 24hrs, then 230°C for 2hrs, and then some of the fibers were ramped to 250°C for an additional 2hrs. After being stabilized at these conditions, the Amlon® fibers were carbonized. After carbonization a visual inspection showed that some sticking had occurred between the fibers. This indicated that these stabilization conditions were not ideal.

Conclusions

Amlon® PAN and the isotropic pitch were both successfully melt spun into precursor fibers. Melt spinning carbon fibers from an Amlon® PAN precursor could lower the production costs because it eliminates the solvents used in wet spinning. However, heat treatment trials showed that standard oxidative stabilization of a melt-spun PAN is likely to require process times on the order of days. Obviously, alternate stabilization methods will need to be developed if this melt-spinnable PAN copolymer is to be commercially practical. The isotropic pitch precursor offers the additional potential advantage of high process conversions. This precursor also required long stabilization times. Nevertheless, stabilization was successful and fibers could be carbonized and mechanically tested. Although isotropic pitch-based carbon fibers display strengths approximately one third that of PAN-based fibers, the precursors cost much less than PAN precursors. Slight modifications of the chemical compositions of isotropic pitch and Amlon® PAN are likely to make them practical for industrial applications. Also, alternate methods of stabilization should be evaluated in future research.

Acknowledgements

The authors wish to thank Jane Jacobi, director of the Center for Advanced Engineering Fibers and Films National Science Foundation Research Experiences for Undergraduates program. We also would like to thank our faculty advisor, Dr. Dan Edie, of Clemson University as well as our industrial advisors, Mark Southard and Ernie Romine, of Conoco, Inc. This work was supported primarily by the ERC Program of the National Science Foundation under Award Number EEC-9731680.

References

1. Edie, D. D., "Carbon Fiber Processing and Structure/Property Relations," in Design and Control of Structure of Advanced Carbon Materials for Enhanced Performance, B. Rand et al. (eds.), Kluwer Academic Publishers, Netherlands (2001) pp. 163-181.

2. Edie, D. D. "The Effect of Processing on the Structure and Properties of Carbon Fibers," Carbon, 36(11), pp. 345-362 (1998).