Microstructure and tensile properties of carbon–epoxy laminates produced by automated fibre placement: Influence of a caul plate on the effects of gap and overlap embedded defects

Abstract

Automated fibre placement (AFP) enables the trajectory of unidirectional composite tape to be optimized, but laying down complex shapes with this technology can result in the introduction of defects. The aim of this experimental study is to investigate the influence of gaps and overlaps on the microstructure and tensile properties of carbon–epoxy laminates. First, a comparison between a hand-layup and AFP layup, draped and cured under the same conditions, shows equivalent microstructures and tensile properties. This provides the reference values for the study. Then, gap and overlap embedded defects (more or less severe) are introduced during manufacturing, on two cross-ply layups [(0°/(90°)5/0°] and [(90°/0°)2/90°]. Autoclave cure without a caul plate results in local thickness variation and microstructural changes which depend on the defect type. This has a strong influence on mechanical performance. Use of a caul plate avoids these variations and in this case embedded defects hardly affect tensile properties.

Introduction

Composite materials have been used for many years in the aeronautical industry to reduce aircraft weight and to optimize performance. The traditional manufacturing techniques are being required to produce increasingly complex shapes while reducing defect levels. In order to address this, new processes such as AFP (Automated Fibre Placement) have been developed, which allow the optimization of reinforcement lay-up and close control of process parameters [1]. An automated fibre placement machine applies tows (of 3.175–12.7 mm width), in the form of a ribbon of unidirectional prepreg with fibres in either thermosetting or thermoplastic matrix but also powder tows for RTM, onto the surface of a mould through a placement head. Thirty-two tows may be placed simultaneously side by side on the mould surface. The orientation of plies, the precision of draping, repeatability and reliability of manufactured parts is ensured while reducing mechanical tolerances. This technology has the advantage of making parts of complex curvilinear shapes, with single or double curvature, in order to optimize the composite structure [2]. It is also possible to produce laminates with large thickness variations. The combinations of controlled pressure exerted by roller compaction and the local application of temperature by a heating system (infrared lamp or hot air torches) removes air trapped between the layers, reduces porosity and allows the adhesion of different plies of the laminate to the mould or onto previously draped plies [3], [4], [5]. Following the layup, the manufactured part is placed in an autoclave to consolidate the laminate plies, polymerizing the resin contained in the prepreg tapes. Mechanical tests on specimens from laminated AFP components have highlighted an increase in buckling load [6], [7], reduced effects of stress concentrations [8], [9] and reduced notch sensitivity [10].

During the layup of composite parts by AFP, gaps and overlaps are generally avoided so as not to impact on the mechanical properties. However, aircraft manufacturers impose a gap of 0.5 mm between each series of n ribbons, with n depending on the placement head used for draping. Fig. 1 illustrates this rule for a tape laying head of n = 8 ribbons. This gap prevents the creation of an overlap which is generally forbidden by the aircraft manufacturers. The value of this gap is defined according to the machine tolerances.

During the draping of non-developable surfaces, misalignment of the tows’ edges will be induced, so management of the gap and/or overlap will be necessary. The AFP programming software allows optimization of the positioning of the tow drop offs. This option allows the location of singularities and their dimensions to be understood, but discontinuities may occur. The tows are cut perpendicular to the fibres and this may create heterogeneous areas depending on the draping strategy. In addition, gaps and tape overlaps, directly related to machine and material tolerances, can be introduced and affect the mechanical properties of the manufactured parts. Research has been conducted for many years to understand the impact of these constraints on carbon fibre reinforced thermoset materials. There are now methods which allow the tape laying trajectory to be optimized, to reduce the impact on mechanical properties [11], [12], [13].

Sawicki and Minguett [14] modelled the influence of gaps and overlaps in plies oriented at 90° to the loading direction in compression. Ply waviness in defect regions resulted in a loss of compression strength. Blom et al. [15] presented a numerical investigation of the compression behaviour that showed the effect of the tow drop areas on AFP laminate panel properties. They concluded that a significant decrease of the strength and the stiffness results when the numbers of tows per course increases. The position of the tow drop area influences the strength and the stress distribution. Turoski [16] performed experimental and numerical tests to investigate the impact of the number of gaps (one, two, three or four) and offset of gap on the tensile and compression strength for unnotched and notched quasi-isotropic laminates. He showed that in tension, the gaps have a smaller effect than when there is an open hole. For compression, unnotched and notched samples showed failure stress reductions as the number of gaps increases. In general, the compression tests revealed larger variations than the tension tests. These previous studies show that the losses in mechanical properties are mainly due to modifications to the geometry in the gap area, related to thickness variation and out-of-plane ply waviness.

Croft et al. [17] presented results from an experimental study which allowed the influence of the main defects (gap, overlap, half spacing/overlap defects and twisting) to be quantified for laminates loaded in tension, compression and shear. They showed that defects present along the specimen length caused a larger drop in mechanical properties than defects across the specimen width. Open-hole test specimens showed a larger variation in properties due to fibre waviness near defects, which initiated micro-buckling. For open hole compression [18], gaps and overlaps have a limited effect when they are placed in the loading direction. When plies are oriented at 90°, the fibre waviness causes a drop in properties. More recently, Falco et al. [19] studied the effect of fibre angle discontinuities on the un-notched and open-hole tensile strength of quasi-isotropic laminates. Observation of the geometry of embedded defects and the resulting mechanical behaviour showed that the influence of gaps was more significant when staggering rules (offsetting of the defects in successive plies) were not applied. However, the test results showed that an open hole has a stronger negative influence than that of a gap singularity. The effects of impact and compression after impact were also studied, experimentally and numerically [20], [21].

Li et al. [22] developed a 3D meshing tool to automatically generate gaps and overlaps numerically. This enabled them to investigate the thickness variations associated with singularities, as well as out-of-plane ply waviness. Models with different sizes and distributions of gaps and overlaps were set up in order to study how these singularities affect properties. Numerical studies have also been performed to examine the influence of gap and overlap defects on the in-plane stiffness and buckling of laminates laid up by AFP [23], [24], [25]. These indicate that overlaps can result in an increase in properties, in contrast to gaps which tend to weaken these materials.

The aim of the work presented here is to develop an experimental method for the analysis of the consequences of the presence of singularities, which are introduced by AFP, for the mechanical properties of carbon/epoxy laminates. Two singularity configurations, gaps and overlaps, of different geometries, have been deliberately introduced into laminates with different stacking sequences, in order to examine how they affect tensile properties. The sequences are only based on 0° and 90° layers and are loaded in tension to avoid membrane-bending coupling. First, a comparison is made between a plate produced by hand layup and one made by AFP, without embedded defects, in order to provide reference values. Then plates with two stacking sequences but the same number of 0° plies, [0°/90°₅/0°] and [(90°/0°)₂/90°], were prepared with embedded defects. These sequences were chosen to study first an extreme case, with superposed embedded defects, then a more realistic case. The embedded defects were only introduced in the 90° plies. C-scan inspections were used to check the quality of all plates. Then scanning electron microscopy (SEM) was used to examine the microstructure of the stacks at the location of embedded defects in detail. Finally, tensile loading enabled the influence of the embedded defects on tensile properties to be quantified. Plates were cured both with and without a caul plate; caul plates are often but not always used during cure in the aeronautical industry. The influence of this plate is examined as it has a strong effect on both geometry and microstructure, but few systematic studies of caul plate effects on AFP composite performance are available.