University of Limerick
Browse

Computational modelling of energy absorption characteristics of epoxy/CNT nanocomposites

Download (4.7 MB)
thesis
posted on 2022-12-20, 14:59 authored by David Weidt
Epoxy-based composite laminates frequently behave as brittle materials under various types of loading, from quasi-static to impact, as they cannot absorb/dissipate enough energy. Matrix cracks develop due to the brittle nature of epoxy, which after further crack growth can cause delaminations and catastrophic failure. Carbon nanotubes (CNTs) offer promising means for enhancing energy absorption characteristics of epoxy matrices, through CNT crack-bridging, and enhanced nonlinear deformation of the epoxy. However, better understanding of the relationship between process-induced nanocomposite morphologies (e.g. CNT distribution, CNT functionalisation and CNT curvature) and nanocomposite properties is crucial for a successful exploitation of these materials for energy absorbing applications. This work addressed this issue from a computational perspective by focussing on two cases, (1) crack resistance characteristics of epoxy/CNT nanocomposites in tension, and its (2) rate-dependent nonlinear compressive behaviour across different strain rates. It was found in case (1) that CNTs significantly reduce the crack driving force in epoxy and increase strains to failure as a result of the damage propagation at the epoxy-CNT interface. Particularly, enhancements of shear stiffness, shear strength and mode II fracture energy of epoxy-CNT interfaces via CNT functionalisation and minor increases of low sp3-bond densities in the interwall phase of DWCNTs were shown to increase the crack resistance of the nanocomposite. However, major focus was on case (2), and related development of a holistic multiscale modelling approach that links various nanocomposite length scales in a sequential manner. In particular, CNT and interface properties were predicted from molecular mechanics/dynamics and used in a mesoscale model that was formulated within the representative volume element (RVE) concept, nonlinear finite element (FE) framework and employed first-order nonlinear homogenisation. It was found that the nanocomposite nonlinear compressive stress-strain response cannot be accurately captured by 2D RVEs (assuming the plane strain condition), when compared to 3D RVEs, primarily because of the stress transfer effect and the particle interaction accurately captured only in 3D. In general, the multiscale models predicted that the increasing CNT aspect ratio, CNT volume fraction and CNT alignment enhance the nonlinear finite strain compressive response by increasing the yield peak true stress and changing the post-yield deformation behaviour from softening to hardening. Also, the CNT alignment was identified as the major factor for enhancing the nonlinear stress-strain response at both quasi-static and impact rates of strain. However, weak van der Waals (vdWs) bonding at the epoxy/CNT interface as well as CNT curvature significantly limit their reinforcement capabilities as predicted in terms of the nanocomposite Young’s modulus and yield peak stress. The model validation involved the preparation of epoxy/CNT nancomposites composed of randomly distributed and oriented CNTs of low mass fraction, mechanical testing under quasi-static compressive loading and the study of fracture surfaces and obtained CNT morphologies using scanning electron microscopy (SEM). 3D models agreed relatively well with the results of experimental programme, when CNT waviness and imperfect bonding at the epoxy-CNT interface were taken into account.

Funding

Using the Cloud to Streamline the Development of Mobile Phone Apps

Innovate UK

Find out more...

History

Faculty

  • Faculty of Science and Engineering

Degree

  • Doctoral

First supervisor

Lukasz, Figiel

Second supervisor

McCarthy, Michael A.

Note

peer-reviewed

Other Funding information

IRC

Language

English

Department or School

  • School of Engineering

Usage metrics

    University of Limerick Theses

    Categories

    No categories selected

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC