Diffusion- and mobility-controlled self-healing polymer networks with dynamic covalent bonding

Research output: ThesisPhD Thesis

Abstract

Thermoreversible Diels-Alder (DA) networks are known to be valuable candidates for intrinsically self-
healing materials. DA bonds preferentially break in case of damage and reversibly reform, which leads
to repeatable healing and an increased lifetime in comparison to classical irreversible networks, making
them more sustainable alternatives for numerous applications. The self-healing performance of DA
networks is directly linked to the DA reaction kinetics, which were however up till now never
investigated as thoroughly as their irreversible counterparts. Moreover, most studies regarding this type
of material focus on the elastomeric state, i.e., at temperatures where sufficient cooperative segmental
mobility is available for the sealing and healing of damage, while thermosets to the contrary possess
only limited mobility, resulting in largely decreased reaction rates. Therefore, the evaluation of the
influence of diffusion- and mobility-controlled conditions on the DA reaction is crucial in view of an
application as self-healing materials.
In this work, self-healing reversible networks based on furan-maleimide DA bonds were investigated
in (partially) vitrified conditions, i.e., at temperatures where mobility is restricted by diffusion. In a first
step, classical cure monitoring methodologies that were typically designed for irreversible network cure,
were successfully applied to the cure of a large range of reversible networks in kinetically-controlled
conditions. Modulated Temperature Differential Scanning Calorimetry (MTDSC), with measurement
of heat capacities and heat flows, microcalorimetry, for more reliable heat flow information, as well as
Dielectric Relaxation Spectroscopy (DRS), to obtain evolutions of dielectric permittivity and loss, were
proven to be especially suitable for the study of reversible DA networks in isothermal and non-
isothermal conditions. From these experimental data in kinetically-controlled conditions, the kinetic
and thermodynamic parameters of a mechanistic model describing the DA reaction, taking into account
the formation of endo and exo cycloadducts, were optimized to have a global description of all the
studied reversible networks. A new cure-monitoring technique was developed by the modelling of the
-relaxation times retrieved from DRS data, characteristic of the glass transition temperature (Tg),
which gave a new tool for the (quasi-)continuous progress of cure evaluation for any isothermal or non-
isothermal temperature profile.
In a second step, the diffusion-controlled cure was studied using similar techniques to evaluate the effect
of vitrification. Heat capacity information and relaxation processes observed in DRS gave clear
indications of the appearance of the frequency-dependant vitrification, while the heat flow data proved
to be more challenging to evaluate without a kinetic model that could perfectly fit the data before the
appearance of vitrification effects. From calorimetric data, a model taking into account diffusion was
developed, based on the hypothesis that an encounter pair should be formed between furan and
maleimide groups before their DA reaction, and that this formation is influenced by diffusion
limitations. The aim of this model was to properly describe vitrifying systems that still react, even with
the slowing down of the reactions, which was experimentally confirmed by isothermal Tg evolutions
for extended cure times (up to 64 weeks), but could not be simulated by the classical kinetic model.
Dynamic rheometry was then applied to directly study the mechanical behaviour and determine the
gelation during isothermal and non-isothermal cure of elastomeric and thermosetting DA reversible
networks. The eventual appearance of vitrification effects, also visible with this technique, made it
Abstract
ii
possible to distinguish between the formation of a gelled glass with a network architecture, or an
ungelled thermoplastic glass, depending on the system and cure conditions. Using the kinetic models
previously derived, Time-Temperature-Transformation (TTT), Temperature-conversion-
Transformation (TxT) and Continuous-Heating-Transformation (CHT) diagrams were constructed for
both elastomeric and thermosetting reversible networks. The main transformation lines, i.e., gelation
lines and vitrification lines, were confirmed experimentally by dynamic rheometry and MTDSC
measurements, respectively, showing the suitability of the optimized kinetic models with or without
taking into account diffusion. In particular, the TTT and CHT diagrams showed unusual shapes for their
simulated iso-conversion lines, where double horizontal asymptotes at the high temperature side are
present. This was confirmed experimentally by two subsequent gelation/de-gelation events occurring
in the same heating segment. This was proven to be due to the reversibility of the system and slight
differences in kinetics and equilibrium constants between the formation of the endo and exo
cycloadducts. At every step of this investigation, the reversible networks were compared to a classical
irreversible epoxy-amine system, to highlight similarities and differences due to the (ir-)reversibility of
the networks. This comprised a similar calorimetric analysis for both kinetically and diffusion-
controlled conditions as well as the construction of a TTT diagram.
Moreover, the ability to self-heal these reversible networks in (partially) vitrified conditions was proven
for the first time by a newly designed testing set-up, which mimics the repair of micro-cracks using
dynamic mechanical analysis. This promising result, however, highlighted the practically low rates at
which the self-healing takes place for the furan-maleimide based systems (4 days in this work). Density-
functional theory (DFT) calculation allowed to identify alternative DA couples possessing faster
dynamics. Based on these results, alternative dienes and dienophiles, i.e., multi-functional fulvenes and
a bi-functional alkene, were then synthesized and characterized by Nuclear Magnetic Resonance
(NMR), MTDSC and dynamic rheometry. The fulvenes showed especially promising results, even
though a noticeable lack of reversibility when combined with a maleimide-based compound, which was
predicted by DFT calculations, makes this combination unsuitable. The bisalkene, for its part, showed
poorer results, as the compound appeared crystalline with an uncertain bi-functionality, making its study
more complicated. Nevertheless, these preliminary results indicate that it is possible to move beyond
furan-maleimide to newer DA reactions, which can present faster dynamics. It is clear that the current
study has developed important experimental and simulation methods which can be used for the future
study of reversible (thermoset) networks.
Original languageEnglish
Awarding Institution
  • Vrije Universiteit Brussel
Supervisors/Advisors
  • Van den Brande, Niko, Supervisor
Award date22 Jun 2023
Publication statusPublished - 2023

Fingerprint

Dive into the research topics of 'Diffusion- and mobility-controlled self-healing polymer networks with dynamic covalent bonding'. Together they form a unique fingerprint.

Cite this