Co-authors​ :

 Panayiotis TSOKANAS (BELGIUM), Yentl SWOLFS (BELGIUM), Thanasis CHATZIATHANASIOU (BELGIUM) 

Among the largest drawbacks of fibre-reinforced composites is their poor delamination resistance, which hinders their industrial exploitation. Delamination, or interlaminar fracture, is among the most prevalent life-limiting failure modes in laminated composites [1] since it significantly reduces stiffness and strength [2] and the compression-after-impact resistance [1]. Shear-driven (or mode II) delaminations are usually triggered by off-axis cracks in tensile loading or through interlaminar shear stresses in flexural or impact loading. They are more challenging to hinder than opening (or mode I) delaminations, partially because through-the-thickness reinforcements, such as Z-pins, are much less effective in mode II [3]. At the same time, the mode II interlaminar fracture toughness is crucial in the compression-after-impact resistance that is, in turn, vital for aerospace and other applications. A novel experimental campaign is conducted in this work to improve our understanding of the delamination initiation and propagation at the micro- and mesoscale. Figure 1 schematically depicts the experimental setup. Three-dimensional, in-situ, X-ray computed tomography (CT) experiments are employed on our in-house, lab-scale X-ray CT scanner to identify microstructural features that delay delamination. To the best of our knowledge, such observations of delaminations at the microscale have only been performed in mode I [4], and this work is the first to deal with mode II. We employ for the first time the four-point end-notched flexure configuration that offers more stable crack growth and less X-ray beam interference with the fixture than the standardised three-point end-notched flexure configuration. A recently proposed [5] load rig is used to introduce mode II loading conditions at the crack front. The load rig is mounted on the rotating stage inside the CT scanner. The voxel size is limited to around 2–3 µm, still precise enough to track the overall crack development and confirm that the delamination follows the desired interface throughout the test. We use unidirectional, thin-ply (50–70 µm) prepregs. Using thin plies is crucial to have enough plies without an excessive thickness and ensure we can downscale the specimens to the required dimensions for the CT scanning without violating width-to-thickness ratio specifications that can secure pure mode II conditions. Our overall aim is to identify the precise crack path and determine the involved micromechanisms. The obtained information will hopefully contribute to answering the critical question of how the local microstructure affects delamination initiation and growth.