Four specimens were prepared from one continuous Carbon Fiber Reinforced Thermoplastic Polymer (CFRP) tape and nondestructively tested using 2D X-ray micrographs and 3D X-ray Computed Tomography (CT). They were each polished on one front side and imaged by optical microscopy using a Keyence VK-X1000 confocal scanning microscope. These two-dimensional micrographs provided high-resolution reference data of the polished tape surfaces. CT was performed on the same specimens with a Zeiss Xradia 520 Versa at voxel sizes of 0.8, 2.0, and 3.5 µm each. The field of view was adjusted to include the polished front side, and the rotation axis was kept constant in between scans of one specimen. This configuration enabled the CT datasets to be registered into a common coordinate system. The registered stacks were subsequently cropped to the tape volume to optimize memory usage. The 3D CT datasets were segmented using structure tensor analysis and Trainable Weka Segmentation to extract fiber, matrix and pore regions in the CFRP tapes’ microstructure. The 2D microscopy images were used as complementary benchmarks to evaluate the required spatial resolution. The overall aim was to determine whether reliable microstructural characterization demands full fiber-level resolution, or whether coarser CT scans provide sufficient information.

Understanding the microstructural variability in unidirectional composite prepreg tapes is relevant to investigating mechanisms of tape microstructure formation, their impact on its processability and the mechanical performance of the final composite part. It has been shown that three-dimensional microstructural variability at the single-fibre level can be resolved by X-ray microcomputed tomography (XCT). However, to define a representative microstructural fingerprint of a given tape, investigations at the required small voxel size lead to limited volumes of observation, which might not be representative. This research aims to extend these findings via a multiscale approach, considering scales of observations, from microscopic (single fibre) up to mesoscopic (dimension of tape) length scale, to generate further insight into the microstructural organisation of thermoplastic prepreg tapes. By exploring the ability of XCT imaging for carbon fibre-reinforced thermoplastic composites at different voxel sizes, the work aims to identify the limitations of the use of different scales of observations to capture features of microstructures and their propagation from micro- to mesoscale level. While structure tensor analysis appeared to correctly capture misaligned regions in XCT images with small voxel size (1/10 of the fibre diameter), the method proved ineffective for larger voxel size images (1/2 of the fibre diameter).

The main bottleneck of using composites for cryogenic storage of clean hydrogen fuel is the permeation of gas molecules. In this work, the permeation of hydrogen gas through thermally cycled thermoplastic composite laminates with two different stacking sequence is investigated. The experimental study is based on a methodology of cryogenically cycling the composite specimen and measuring the permeability in a dedicated hydrogen permeation setup. An optical microscope and X-ray computed tomography scanner are employed to investigate the existence of cracks. The results reveal that thermal cycling does not have a profound influence on permeability, while the stacking sequence has a considerable effect. Laminates with dispersed 0° layers resulted in lower permeation values compared to the laminate with grouped 0° layers at the laminate’s core. The imaging techniques did not reveal any observable crack which supports the hypothesis that permeation is mostly driven by bulk diffusion in the polymer.

Thin ply composites open up new opportunities to exceed the limits of conventional composite materials by improving first-ply/ first-damage criteria, fatigue life and ultimate strength [1]. By definition, individual plies with a ply thickness of less than 0.100 mm can be called a thin ply [2,3]. The size effects [1] and design freedom as it allows smaller pitch angles for a specific thickness [4] make thin ply composites superior to conventional ply composites. Obtaining thin plies is possible spreading conventional tows via techniques based on airflow, ultrasonic vibration, and mechanical means as the common alternatives [5]. Among these, mechanical tow spreading is based on pulling dry tows through bars/ pins with a certain tension. An example of a lab-scale tow spreading line, developed at TU Delft, including static spreader bars and tension sensors can be seen in Fig. 1 (a). Wrap angle, the pre-tension of the tow, relative friction between the spreading bar and tow, temperature and pulling speed are some of the parameters influential on spreading. Therefore, we propose an experimental framework to assess and quantify the effect of individual mechanisms on tow spreading. In the present study, we first investigate the effect of tension and wrap angle on tow spreading under the static condition schematically shown in Fig. 1 (b). Then, the effect of an actively controlled bar on tow spreading is tested under static conditions. Spreading bar rotation is controlled with a motor that reveals the influence of relative motion between the bar and the tow and thus the induced friction. Tow spreading is a continuous process, where filaments’ reorganization is time dependent. That means the time spent on a spreading bar, defined by the pulling speed, is influential on spreading, noting that the pulling speed is also intrinsically related to friction. Thus the third experimental method is based on observing tow spreading while the tow is pulled with various pulling rates, as schematically shown in Fig. 1 (d). During the talk, we will report our findings regarding the extent of spreading of different fiber types under different process conditions obtained by controlling the wrap angle, relative speed between the tow and bar(s) as well as by using bars that either are heated and/or exhibits different surface roughness. The analyses will be further enriched by microstructural analyses of specimens.

The effect of thermal contact resistance (TCR) correlated to the degree of intimate contact (DIC) between the incoming tape and the substrate on the temperature history during laser-assisted fiber placement (LAFP) was investigated. A novel experimental methodology was designed to understand the effect with a non-contact method which did not influence the local consolidation quality. To assess the influence of TCR numerically, a three-dimensional optical-thermal model was developed. Experimental results indicated that, for the same tape temperature near the nip point, an increase in the compaction force resulted in a decrease in the temperature at the roller exit and the following cooling phase, in correlation with an increase in the final DIC. Also, the effect of the laser power on the final DIC was less pronounced than the compaction force. In the thermal model, when TCR at the tape-substrate interface was not considered, the temperature predictions underestimated the experimental measurements.