Effect of carbon fiber oxidization parameters and sizing deposition levels on the fiber-matrix interfacial shear strength

Abstract

This paper investigates fifteen fiber types against two epoxy resin systems and the effects of altering electrochemical oxidization conditions and sizing deposition ratio on interfacial shear strength (IFSS). Oxidization current was altered between 0, 2, and 3.4 A while sizing deposition ratio was altered between unsized, 1:10, 1:15 and 1:20 parts sizing to water. Desized fibers were also compared against pristine unsized fibers. Results show, a correlation between increasing current and IFSS, however sizing has an optimal ratio for best performance. Improvements through oxidization are attributed to the introduction of oxygenated functional groups on the fiber surface while improvements due to sizing are attributed to the promotion of a chemically active intermediate layer between the fiber and resin. Fiber roughness was seen to play no effect on IFSS. Desized fibers and unsized fibers had similar IFSS results however characterisation shows chemical composition of the fiber surfaces to be very different.

Introduction

With their high strength to weight ratios contributing to reduction of component structural mass, carbon fiber reinforced polymers have already made a notable mark on the engineering industry [1]. While carbon fiber composites are synonymous with high performance materials, micro-cracking, a known failure mechanism of these materials commonly arises as a result of a weak interface between fibers and resin. One means to obviate this limitation is to chemically modify the surface of carbon fibers to promote fiber/resin compatibility, which typically results in a simultaneous increases in fiber/matrix adhesion [2], [3], [4]. For this reason surface treatments have become of intense interest, with a number of reviews published on the topic recently [5], [6], [7]. Over the past decade many oxidative and non-oxidative techniques have been employed to etch [8], [9], [10], roughen [11], [12], and alter surface chemistry [10], [13], [14] with varying degrees of success. Whilst much of this research demonstrates enormous promise, uncertainties still remain in interfacial science.

Academic consensus between the relative contribution of different bonding mechanisms, namely chemical adhesion, mechanical interlocking or fiber wetting, that constitutes an improved interfacial bond remains contested. The influence of several manufacturing variables during carbon fiber fabrication and final interfacial performance is also unclear, as is the relationship among these variables. Likewise performance comparison among sized, desized and pristine virgin fibers remains ambiguous requiring further study.

In this work we have used a pilot scale carbon fiber line (www.carbonnexus.com.au) for the manufacture of virgin carbon fibers to enable the examination of bonding interactions in an industrially relevant manner. The goal is to clarify the effects of two processing parameters; electrolytic bath amperage and sizing deposition ratio, on interfacial shear strength (IFSS). It is generally agreed that the surface treatment is critical to ensure improved IFSS [13], [14], [15]. Studies have attributed the importance of the electrochemical oxidization to be the removal of weakly bound graphite from the surface of the fiber and the introduction of polar functional groups (e.g. carboxylic acids, phenols, ketones, etc.) [7], [11]. However no systematic study on varying electrochemical amperage and subsequent effects on IFSS has been conducted. Conversely, the influence of sizing compatibility and whether its effects are beneficial [8], [16], [17], [18] or detrimental [12], [13], [19] to IFSS performance have not reached a unified scientific consensus. This convolution alone constitutes the requirement for further research but by adding manufacturing variables such as the effects of oxidization, the requirement for this research becomes crucial (see Table 1).

Conflicting literature is also observed within studies claiming mechanical interlocking to improve IFSS [12], [19] while others show no evidence of this [20]. Similarly contentions supporting polar functional groups [10], [8], [19], [20] and dispersive functional groups respectively [13], [18] as the key contributor of improved IFSS highlight gaps in scientific understanding that require further exploration [7], [21]. The truth may be more complex and rather than a uniform answer, may be a matter of fiber-resin compatibility to be considered on a case by case basis [22]. In either case further fundamental research is required.

Reasons for these aforementioned discrepancies are likely due to numerous factors. Firstly, the constituents of commercial sizings are typically not publicly known and thus a comparison of two sizings may not be strictly scientific as more than one variable may be changed without control. Secondly the ability to examine both surface treatments and sizings separately and simultaneously to observe any masking or interaction effects has not been afforded to researchers without access to a commercial manufacturing line. This has fostered a third issue, with fibers in the majority of previous studies that first desized and then subsequently resized or surface treated as sizing is routinely applied in the manufacture of commercial fiber [13], [15]. Undoubtedly this process introduces further variability in testing control which may not necessarily give an accurate representation of virgin carbon fibers.

In this work, we report our findings focussing on the effects of surface treatment and sizing processes on IFSS. These manufacturing effects are observed individually (i.e. impact of only sizing or only oxidization), and simultaneously (impact of both sizing and oxidization together). Results of treated pristine fibers are also compared against fibers that have been desized and then treated to understand if the desizing process alters the subsequent IFSS. Extensive physical and chemical characterisation was done on all fibers to compile an accurate understanding of fiber properties.

It was concluded that each of these processes (surface treatment and sizing) improved IFSS independently with no masking effects influencing results. Improved IFSS was observed to be a factor of both chemical modification introducing active groups to the surface and sizing dispersion encouraging a greater fiber/matrix compatibility at the interface. Mechanical interlocking (improvement in interfacial adhesion attributed to fiber surface roughness) was not seen to play any statistically significant role on IFSS. Desized fibers were also observed to provide mixed results suggesting that they may not necessarily be representative of pristine fibers.