Rapidly cured out-of-autoclave laminates: Understanding and controlling the effect of voids on laminate fracture toughness

https://doi.org/10.1016/j.compositesa.2015.02.023Get rights and content

Abstract

Voids are one of the most significant defects found within composites and have been demonstrated to reduce the performance of composite structures. The understanding of the impact of the size and distribution of voids on laminate properties is still limited because voids have proven difficult to deliberately control. This study aims to understand the mechanisms by which voids are generated within out-of-autoclave cured laminates. In this study, a process of prepreg conditioning was developed to control the level of voids within test laminates. Non-conditioned laminates highlighted signs of void growth (1.5%), while conditioned laminates showed consistently low levels of voids (<0.3%). Mass spectrometry indicated higher levels of aqueous and solvent volatiles within the non-conditioned prepreg. Finally, Mode II fracture testing revealed a 21% improvement in toughness for the non-voided laminates. A model on the effect of voids within the Mode II stress state has also been proposed.

Introduction

Voids are considered one of the most critical defects found within composite structures, and as a result, industry such as aerospace undergo vigorous Non Destructive Evaluation (NDE) and set tight limits on overall content within their primary structures (typically < 1%). The reason is that voids result in reductions to composite strength, particularly for matrix dominated properties. Literature has highlighted reductions to interlaminar and in-plane shear strengths, compression strength and fracture toughness [1], [2], [3], [4], [5]. In addition, voids pose other issues such as accelerated moisture uptake due to an increase in free volume [6] as well as acting as initiation sites for the onset of failure within composite structures subjected to fatigue [7].

It is well documented that voids can be suppressed within prepreg laminates by the use of high external pressures, which is why autoclaves have been widely used [5], [8]. Though, due to their high capital cost, high energy consumption and slow heating rates, the composites industry is seeking new alternatives. As a result, it is imperative to advance Out-of-Autoclave (OOA) technologies as composites are becoming widely used across many industries including the automotive sector, where production rates are extremely high. Unlike the autoclave, OOA prepreg processes mostly rely on vacuum bag consolidation pressure, which can only achieve external pressures of approximately 1 atmosphere. According to vapour pressure maps, this pressure is not high enough to prevent moisture and solvent volatiles from erupting over typical curing temperatures, particularly during the second ramp within two step cure cycles [9], [10].

Trace solvent and aqueous volatiles can result from the resin synthesis and prepreg manufacture [9] or through incorrect handling practices, as many prepreg resins are very hygroscopic in nature. The second point is an increasing issue particularly as composite parts are now being manufactured in parts of the world such as South East Asia, where the humidity levels can exceed 90%. Industry sees OOA manufacture as an attractive alternative as there is a race for shorter cure cycle times without compromising properties. Controlling and reducing void content in OOA cured composites has been the subject of a number of studies in the literature.

The Double Vacuum Bag (DVB) technique is one example used to reduce the void content of laminates. Unlike the Single Vacuum Bag (SVB) technique which applies full vacuum and bag compaction to the laminate throughout the entire cure cycle, the pressure can be varied during cure using the DVB technique. The tooling has been designed to apply partial vacuum without compaction for the volatile management stage of the cure cycle; and full vacuum and compaction for the curing stage of the cure cycle. The former part of the process allows for volatiles to be removed through a less restrictive medium. Investigations using this approach [11], [12], highlighted reductions in void content, however, the complex nature of this technique may limit its adoption by industry.

Based on a similar principle, prepreg composites with low void contents using OOA processes have been achieved through the development of specially designed prepregs. OOA prepregs are partially impregnated (resin rich on one side) which allows volatised gases to escape during initial cure when the resin viscosity is high. Literature has shown that if these prepregs are cured using the correct conditions, void contents of less than 1% can be achieved [13]. One vital curing condition is that OOA prepregs require slow heating rates 1–3 °C/min, as well as long intermediate dwells in order to keep the viscosity high for as long as possible in order to keep the air paths open. As soon as the viscosity becomes low enough to permeate the ply, the air paths close and any entrapped volatiles become trapped within the composite. Similarly to autoclave cure, the slow heating rates used with OOA prepregs result in cures just as long. In addition, OOA systems like any other prepreg require stringent storage measures to keep moisture uptake low. Grunenfelder et al. [13] studied the effect of dissolved moisture within prepreg composites and found that regardless of the opened air paths within the lay-up, if the prepreg is handled incorrectly (storage environments >70% RH) resulted in composites with unacceptable void contents regardless of the curing conditions.

Novel OOA curing techniques have also been shown to minimise void content. Davies et al. [14] used a number of curing techniques including the rapid technology Quickstep, and characterised their effect on void content within carbon/epoxy prepregs. In this study a novel cure cycle referred to as a “spike cure” was explored and compared to both conventional two step and isothermal cure cycles. The spike cure was developed to maintain a wider processing window, while maintaining a relatively low viscosity to facilitate the removal of volatiles. It was reported that the spike cure produced laminates with considerably lower volatiles (1.8%) when compared to other vacuum assisted technologies (∼10%). However, panels cured using the spike cure were not able to match the void content of autoclave panels (0.6%), which demonstrated the importance of pressure to achieve aerospace quality composites. Hernández et al. [2], [3] conducted extensive studies on investigating voids within OOA composites where cure cycles were optimised for low pressure (0.2 MPa) compression moulded prepreg composites. Similarly to Davies et al. [14], it was shown that the cure cycles designed with wide processing windows were able to manufacture composites with low void contents less than 0.5%. Muric-Nesic [15], [16] optimised cure cycles in terms of void content reduction by incorporating vibrations throughout the volatile management stage of the cure cycle. A speaker was used to excite the composite during cure where a range of frequencies were studied (0–8 kHz). The work found that void contents could be reduced by up to 60% when compared to composites without vibration assistance.

The current study explores a route to reducing cure cycle time via void content control. It will achieve this using a number techniques designed to control and minimise the effect of voids within OOA composites. The first technique, a conditioning procedure, aims to remove entrapped volatiles within carbon fibre/epoxy prepreg prior to lay-up and cure rather than during cure which is commonly seen within the literature. The parameters required for successful removal of the trace volatiles within the prepreg are validated using simple mass measurements, mass spectrometry, Thermal Gravimetric Analysis (TGA), Differential Scanning Calorimetry (DSC) and prepreg tack testing. A second technique, a cure monitoring procedure was then followed to assess the development of voids at critical points within both non-conditioned and conditioned laminates cured using both slow and rapid heating technologies (1.5 and 10 °C/min). Rapid heating was studied as a recent investigation has shown that faster heating rates (10 °C/min) resulted in a reduced resin viscosity throughout cure [17], which has the potential to influence properties such as fracture toughness and interlaminar shear strength by improving the adhesion between resin and fibre as publicised in previous studies [18], [19], [20]. However, investigating such an idea has proven difficult as literature has shown that these properties are highly sensitive to voids [1], [3], [5], [21]. Finally, Mode II fracture toughness testing was conducted to understand the effect of the conditioning technique on fracture toughness.

Section snippets

Material

The laminates in this study were manufactured using Hexply 8552 prepreg intended for autoclave cure. Hexply 8552 was specially chosen for this study as it is a prepreg material currently in high demand within the aerospace industry. Secondly the resin formulation contains a polyethersulfone toughening agent, which increases the hydroscopic nature of the material posing problems in terms of voids. In addition the prepreg system has been manufactured via the hot melt production process [22] where

Thermal Gravimetric Analysis–Mass spectrometry

In previous work [17], [27], the authors identified that volatiles within the prepreg were detrimental to the overall performance of the composite material. Here we use TGA–Mass spectrometry to identify and quantify these. Previous work [17] also indicated that residual moisture and acetone are commonly present in prepregs. Residual moisture is the result from both the manufacture of the prepreg during the hot melt process as well as from the uptake of moisture during handling. Acetone on the

Conclusion

The effect of conditioning prepreg in order to understand, control and minimise the effect of voids within OOA laminates has been successfully investigated. Mass and TGA measurements revealed that exposing the prepreg to 40 °C and −97 kPa absolute pressure for 120 min resulted in a weight loss between 0.54% and 0.62%. The effect of the conditioning conditions on prepreg age was verified using DSC, which determined minor shifts in glass transition temperature. Further evidence of successfully

Acknowledgements

The author would like to acknowledge the financial support of the Victorian Centre for Advanced Materials Manufacturing (VCAMM), John Ellis at Hexcel composites for supplying the 8552 prepreg material and Commonwealth Scientific and Industrial Research Organisation – Australia (CSIRO) for mass spectrometry support.

References (33)

Cited by (0)

View full text