James A. Newell

Kinetics of carbonization and graphitization of PBO fiber

PBO [poly(p-phenylenebenzobisoxazole)] fiber has been shown to convert to an ordered carbon fiber without the need for stabilization. This article presents the first detailed analysis of the carbonization and graphitization behavior of this unique material. The carbonization process was modeled as a series of free-radical reactions, and thermogravimetric analysis was used to determine an activation energy of 76 ± 6 kcal/mol for the thermal initiation of free radicals. The initiation reaction data then were applied to determine the temperature dependence of the termination reaction. Additionally, the development of long-range order in the graphitizing fiber was examined. The spacing between graphene planes was shown to decrease with increasing treatment temperature and soak duration. Carbonized PBO fibers developed more long-range order than carbon fibers produced from other polymers, which may partially explain why these PBO-based fibers display excellent lattice-dependent properties. Finally, an Arrhenius analysis found the activation energy for graphitization to be 120 ± 17 kcal/mol.

Factors limiting the tensile strength of PBO-based carbon fibers

Abstract Poly p-phenylene benzobisoxazole (PBO) converts directly to carbon fiber without stabilization. However, the applicability of PBO-based carbon fibers is limited by their low tensile strength. This paper represents the first study of the high-temperature conversion of PBO to carbon fiber and examines some factors leading to this low tensile strength. A mixed-mode Weibull analysis implies that the flaws present in the precursor fiber persist throughout carbonization and cause the tensile failure of the carbonized fiber. Improvements in the PBO spinning process resulted in increased tensile strengths for the carbonized fibers. Further, the tensile strengths of the carbonized fibers appear to be related to the release of nitrogen during the onset of crystallite growth. The rapid heating rates associated with continuous carbonization were shown to minimize the negative effects of this nitrogen release. Fibers heated to 1600 °C in a continuous operation displayed nearly double the tensile strength of those produced in a batch operation. Ultimately, this study showed that, while the tensile strengths of PBO-based carbon fibers can be increased by using a continuous process, enhancements in the solution spinning process of the polymer will be required for PBO to become a commercially viable carbon fiber precursor.

Direct carbonization of PBO fiber

Abstract Poly p-phenylene benzobisoxazole (PBO) fiber appears to display unusual characteristics during its conversion to carbon fiber. This introductory study focused on specific aspects of the carbonization behavior that differentiate PBO from other materials. Results showed that oxidative stabilization did not affect the tensile properties of carbonized fiber produced from PBO precursor fiber. Unlike other polymeric precursors, unstabilized PBO fibers could be directly converted to carbon fibers with promising mechanical and thermal properties. The carbonization characteristics of unstabilized PBO were found to be similar to those of stabilized PAN-based fibers. However, the PBO-based carbon fibers exhibited a radial texture similar to pitch-based fibers. Surprisingly, the PBO-based carbon fibers exhibited electrical resistivities in the range of commercial pitch-based fibers, implying that their thermal conductivities also will be similar. Low-temperature carbonization studies showed that PBO fiber can be carbonized at rapid rates without adversely affecting the tensile properties of the carbonized material. Its ability to be directly carbonized, combined with the unusual thermal properties of the final carbon fiber, may make PBO an attractive precursor for some carbon fiber applications.

Experimental Verification of the End-Effect Weibull Model as a Predictor of the Tensile Strength of Kevlar-29 (Poly p-Phenyleneterephthalamide) Fibres at Different Gauge Lengths

This paper describes the first application of the four-parameter end-effect Weibull model as a predictor of tensile failure frequency at a gauge length different from that used to generate the model parameters. As-received Kevlar-29 fibres were tensile tested at gauge lengths of 10, 25 and 40 mm. The resulting failure data were used to determine four empirical constants using a maximum-likelihood regression. The model and parameters were used to predict failure frequency at a gauge length outside the initial range (5 mm). The results show that the end-effect Weibull model accurately represents the data from which its parameters are evaluated and that the model may be applied as an effective predictor for gauge lengths beyond the original testing range.

Development of a Pneumatic Spreading System for Kevlar-Based Sic-Precursor Carbon Fibre Tows:

A novel spreading system based upon the Venturi principle was designed and constructed. The system was used to spread Kevlar-based carbon fibres and to maintain that spread for sufficient time to facilitate chemical vapour deposition of an SiC coating. The Kevlar-based carbon fibres were shown to possess excellent thermal stability and a combination of properties that make them an ideal silicon carbide precursor fibre. Through a combination of pinch-rollers and two distinct axial Venturi spreaders, a highly chaotic and effective spread was achieved and maintained. This work provides a novel process for producing potentially excellent silicon carbide precursor fibres as well as providing a novel use for a high-performance polymer.

Influence of Thermal Treatment Conditions on the Recoil Compressive Strength of Kevlar-29 Fibers:

In this paper, thermal treatments of Kevlar-29 fiber resulted in a significant enhancement of compressive strength and corresponding decrease in tensile strength. These changes result from a decreased linearity at the macroscopic level in the fiber. Statistically designed experiments indicate that treatment temperature is the most significant variable in the thermal processing, while treatment soak time also played a role in the shift of properties.

An Improved Interpretation of Recoil Compressive Failure Data for High-Performance Polymers

This paper describes an improved methodology for the accurate interpretation of recoil compressive failure data. This new procedure uses a wider portion of the data set in its interpretation than does the method of Allen and is far simpler and less data intensive than the Weibull model proposed by Hayes. Computer simulations using various statistical distributions for the intrinsic recoil compressive strength of a batch of filaments show that this method quickly and accurately converges to an accurate approximation of the mean for the sample set.

Analysis of Recoil Compressive Failure in High Performance Polymers Using Two and Four Parameter Weibull Models

In this study, large amounts of recoil compressive failure data were gathered for Kevlar-29 fibers. Once the dependence of failure frequency on stress level had been determined experimentally, two-parameter and four-parameter Weibull models were used to represent the data. An analysis of these results shows that each model represents the failure well, but that the four-parameter model offers little improvement over the two-parameter Weibull model. Deconstruction of the four-parameter model indicated that a single failure mechanism dominated and that this mechanism could be well represented by the two-parameter Weibull model. These results imply that while recoil compressive failure is more complicated than tensile failure, recoil failure is dominated by a single flaw distribution and can be represented by a more simple model than tensile failure.

Enhancement of the Compressive Strength of Kevlar-29/Epoxy Resin Unidirectional Composites

This paper examines the compressive and tensile strength of Kevlar-29/Derakane composites using as-received Kevlar-29 and a thermally enhanced version of the fiber. Previous studies have shown that thermal enhancement results in a form of Kevlar-29 with significantly better compressive strength at a cost of 10—20% of its tensile strength. This study compared the mechanical properties of unidirectional composites made using as-received and thermally enhanced forms of Kevlar-29 fibers. The composites made with the thermally treated fiber showed a decrease in tensile strength roughly proportional to the difference in tensile strength of the fibers. Although the improvement in the compressive strength of the composites made with the thermally treated fibers was statistically significant, it was less than anticipated. A change in predominant failure mechanism also occurred with most as-received fiber-based composites failing from a debonding/pull-out mechanism, while composites made with the thermally enhanced fiber failed primarily from microbuckling with attendant splitting.

A Comparative Analysis of Techniques Used to Estimate the Mean Recoil Compressive Strength of High Performance Polymers

This study provides the first experimental analysis comparing multiple methods of interpreting recoil compressive failure data. A two-parameter Weibull model was used to provide an accurate mean recoil compressive strength for a batch of Kevlar-29 fibers. The Weibull model provides highly accurate results, but requires the testing of more than 1000 filaments. The results from the Weibull model were used as a basis to compare two simplified and less data-intensive models: The method of Allen and the moving average method. Each method was applied to 10 sets of 80 filaments each and the results of these tests were compared with the two-parameter Weibull model results from the same fiber batch. The results of these tests have shown that both the method of Allen and moving average method tend to slightly over-estimate the compressive strength calculated by the two-parameter Weibull model, yet both estimation methods fall in the range of published values (200-400 MPa) for the recoil compressive strength of Kevlar-29. This inaccuracy probably results from the assumption of a symmetric failure distribution that is made in both simplified models, but not in the Weibull model. A statistical analysis of the results was completed and shows with a 95% confidence level that there is no statistical significant difference between the two methods.