Mechanical characterisation of polymer-based composites at fibre/matrix level by nano digital image correlation

Details

Supervisors

Pardoen, Thomas ; Swolfs, Yentl

Faculty

Ecole polytechnique de Louvain

Degree label

Master [120] : ingénieur civil en chimie et science des matériaux, à finalité spécialisée

Abstract

Owing to a combination of low density and high stiffness and strength, fibre-reinforced polymers (FRP) are the material of choice for high-performance, lightweight applications. Given the heterogeneous nature of FRPs, multiscale bottom-up models are developed to predict the deformation and failure of composite components. Yet, to improve the numerical predictions and aid with the design of future composite structures, these models require constituent properties as input parameters. An accurate characterisation of the response of the constituents is thus essential. Existing multiscale models are currently lacking a complete understanding and reliable experimental data of the nano- and microscale stress and strain distributions. The literature points towards the presence of an in-situ effect on the matrix, meaning that its properties at the constituent level differ from those at the macroscale. Nanoscale digital image correlation (DIC) enables characterising the deformation and failure of the polymer matrix confined in between the fibres. The goal of this master’s thesis is to determine the nanoscale mechanical response in FRPs. For this purpose, glass/carbon fibre-reinforced amorphous thermoplastic/thermoset composites are examined by nano-DIC. Unidirectional samples were subjected to transverse compression within a scanning electron microscope. Nano-DIC relied on an indium speckle pattern of 15 nm in diameter to analyse several fibre configurations, classified according to a horizontal, vertical or diagonal position of two close fibres. Their corresponding nanoscale strain maps were characterised and compared while considering the thickness of the matrix strip separating the fibres and the macroscopic strain. Nano-DIC allows identifying regions of strain concentration and shows that the strains tend to localise around the fibres. The formation of local submicron shear bands, oriented along the matrix strip, is captured in the resulting shear strain maps. The shear band orientation in vertical fibre configurations increased with the matrix strip length. The orientation obtained from nano-DIC coincided well with the literature. Furthermore, the failure locations in the regions of interest could be pre-emptively identified via concentrations of maximal shear strain values. A comparison of the local and macroscopic failure strains suggested a local increase in triaxiality. The observed patterns are compared to finite element models based on constituent properties validated at the macroscale. The model results revealed similar patterns for the strain concentrations but showed inconsistencies in the local strain values and shear band orientations. Part of these discrepancies were related to the limitations of the model. The results of this master’s thesis consolidate the use of nano-DIC for the characterisation of FRPs at the fibre/matrix level, provide a systematic semi-quantitative framework to analyse and compare nano-DIC results across different systems and create valuable data for experimental validation of micromechanical models.