Experimental investigation of the effect of the mould thermal expansion on the development of internal stresses during carbon fibre composite processing

https://doi.org/10.1016/j.compositesa.2008.04.011Get rights and content

Abstract

The influence of the mould material on the build-up of internal stresses during autoclave cure of carbon fibre composites was investigated using embedded optical fibre Bragg gratings. Unidirectional and cross-ply laminates were cured on four flat moulds: aluminium, steel, carbon composite and carbon foam. From the beginning of cure, the fibres were strained by contact with the expanding mould. The strain increased with the mould thermal expansion and autoclave pressure. As a result, large residual strains were found in the longitudinal direction for the samples cured on aluminium or steel moulds, whereas the samples cured on carbon-based moulds exhibited almost no strain after cooling-down.

Introduction

The quality and mechanical properties of a thermoset composite material are dictated by the material processing steps and parameters including temperature, time and pressure [1], [2]. In particular, differences in thermal properties between the resin and the fibres or between plies of different orientations lead to the build-up of large stresses in the parts during processing. These stresses may exert a strong influence on the short and long term mechanical properties of a composite structure, anticipating the matrix crack initiation or being at the origin of delamination. Additionally, they can lead to warpage, and lack of dimensional stability, in turn complicating the assembly of structures. A large number of experimental and numerical investigations have thus been conducted to analyse and control these effects [3], [4], [5], [6], [7].

In situ cure monitoring procedures have been developed to optimise the processing step. Among these methods, optical fibre monitoring has demonstrated several advantages for this application as the sensors can easily be integrated in the composite material during manufacturing. Different types of optical fibre sensors have been used, including refractive index sensors [8], [9], [10], extrinsic Fabry-Pérot interferometric (EFPI) sensors [11], [12], fibre Bragg grating (FBG) sensors [9], [13] and FBG/EFPI hybrid sensors [14]. They have mainly been used to measure internal strains and temperature during composite curing [9], [10], [11], [12], [13], [14], [15], [16]. Moreover, they have been used to determine the resin gel point [8], [9], the degree of resin conversion [10], the glass transition temperature [9], [13] and to evaluate the thermal expansion coefficients of moulds and fully cured resins [13].

The influence of the mould material was generally not taken into account in the majority of these studies, which concerned the autoclave cure of laminates. However, it was shown that the mould material can exert a measurable influence. Twigg et al. [17], [18] presented an experimental study of the effect of mould-part interaction on the induced warpage during processing of unidirectional carbon-fibre reinforced plastic (CFRP) samples fabricated in autoclave using a flat aluminium tool. Electrical strain gages were used for the characterization of the interfacial shear stresses between the tool and the composite part during processing. The strain inside the composite part was not measured as the electrical strain gages could not be embedded into the laminate. When relating the tool–part interface conditions to the measured mould in-plane strains, they observed high interfacial shear stresses that resulted from the mismatch between the mould and part coefficients of thermal expansion (CTE). They also noticed that the fibres in the prepreg were stressed even prior to resin gelation. In addition, it was observed that the part aspect ratio had a much greater influence on warpage than the autoclave pressure, while on the other hand the tool surface condition had no significant effect. An empirical relationship between part aspect ratio and warpage could be predicted from an analytical model [18]. A numerical model [19] that takes into account the tool–part interaction on the final part shape was also developed for calculating the deformation induced during manufacturing. More recently, Potter et al. [3] observed that the tool exerts a strong influence on the final stress state in the composite part. Moreover, they showed that stresses arise even before gelation most probably through friction between the prepreg and the tool surface. However, all these studies concentrated on one type of mould material which had a rather large coefficient of thermal expansion.

Finally, Antonucci et al. [15] used fibre Bragg gratings to investigate the effect of the mould material on the process induced strain field in a thermoset resin both with an aluminium and a steel mould. They observed that the strain evolution strongly depended on the geometrical constraints imposed by the mould and was affected by the mismatch between the different properties of the metallic mould and the epoxy resin.

In the present work, the capability of optical fibre sensors to be embedded in composite materials during manufacturing was used to investigate the build-up of internal stresses in a composite laminate as a function of the mould material and its thermal expansion properties. We considered the case of unidirectional and cross-ply carbon–epoxy laminate composites from prepreg layers cured in an autoclave. Strain in the material was measured using fibre Bragg grating sensors embedded along the reinforcement orientation. Four different moulds with different thermal expansion coefficients (i.e. aluminium, steel, carbon foam and carbon–epoxy composite-based) were used. In the case of the unidirectional laminates, the influence of curing pressure (i.e. vacuum bag only and with an additional pressure of 5 bars) was also evaluated.

Section snippets

Fibre Bragg grating sensors

A fibre Bragg grating consists of a periodic variation of the core refractive index along the fibre, with a typical period around 500 nm. Bragg gratings are fabricated by exposing a section of the optical fibre to an interference pattern of ultra-violet (UV) light [20]. The photosensitivity of the germano-silicate glass allows the index of refraction in the core to be changed by the UV laser radiation. When a broadband light source is coupled to the optical fibre containing a FBG, only a narrow

Composite materials

Four different flat moulds made of materials with different thermal expansions were used for the manufacturing of carbon fibre-epoxy composite laminates: a steel plate (α = 12 × 10−6 K−1), an aluminium plate (α = 23 × 10−6 K−1), a carbon fibre–epoxy quasi-isotropic composite rectangular plate (αL = 1.9 × 10−6 K−1 in the longitudinal direction and αT = 0.9 × 10−6 K−1 in the transversal direction) and a Grafoam™ FPA35 carbon foam mould from GrafTech (α = 2.3 × 10−6 K−1).

Carbon–epoxy [04] unidirectional samples of

Resin chemorheology

Fig. 2 presents the evolution of G′, G″ and tan δ = G″/G′ during cure of the prepreg for the temperature cycle as presented in Fig. 1. The viscosity initially decreased following G″ (i.e. shear loss modulus) as the temperature was increased, until the chemical cross-linking reaction took over resulting in a viscosity increase. The resin gel point was taken when the curve of G″ stopped its sharp rise at around 140 min. This means that the gel point was reached at the beginning of the second

Conclusions

This study showed that at the end of cure, an optical fibre embedded in a unidirectional and a cross-ply carbon–epoxy composite along the longitudinal direction is under tension when using a metallic mould. Assuming that there is a stress transfer from the host material to the FBG sensor it can be concluded that, during curing of a carbon-epoxy composite sample on a metallic mould, internal stresses are induced in the material and that these are related to the mould expansion. The choice of the

Acknowledgements

R. de Oliveira would like to thank the Portuguese Foundation for Science and Technology (FCT) for financial support under the grant SFRH/BPD/41347/2007. Part of this work was funded by the Swiss Commission for Technology and Innovation no. 6972.2 in collaboration with Alinghi SA. Graftech International Ltd is acknowledged for donating carbon foam mould material, and F. Rakusa for contributing to this work during her studies.

References (28)

Cited by (80)

View all citing articles on Scopus
View full text