Multiscale characterization of the fracture mechanics of additively manufactured short fiber-reinforced composites

We present a multiscale characterization approach to experimentally investigate the influence of architectural features, namely pores, weakly bonded filaments’ interfaces, and layer stacking, on the failure of composites manufactured by fused filament fabrication. Combining standard approaches, such as tensile and flexural tests, with contactless techniques, we identified the local phenomena driving the damage of 3D-printed short carbon fiber-reinforced polyetheretherketone specimens. The elastic and fracture tests highlighted the process-induced anisotropy and the incomplete interface adhesion resulting in the transverse tensile modulus drop to around 20% of the longitudinal value. Equivalent fracture toughness was measured for a crack propagation along filaments’ and layers’ interfaces, while the stress intensity factors doubled for a crack propagation involving the filaments’ breakage, when compared to the interface failure. The displacement and strain contours obtained by digital image correlation emphasize the influence of the stacking (i.e., 0ᵒ-90ᵒ, ±45⁰) on the preferential crack propagation at the layers’ and filaments’ interfaces. The specimen inspection by scanning electron microscopy and by X-ray tomography further highlighted the influence of the printed composites’ meso and microscale architecture on the fracture mechanisms, such as the simultaneous damage of parallel ±45⁰ oriented layers’ interfaces or the zig-zag crack propagation for specimens with ±45⁰ stacking undergoing intralayer delamination. The elastic and fracture properties, together with the full-field measurements, provide the tools to guide the design of complex and reliable components for high-performance applications (e.g., aerospace, automotive) and benchmark for damage prediction models.

Mots clés

fused filament fabrication; short fiber-reinforced composites; fracture mechanics; digital image correlation; computed tomography