Fluorinated vitrimer elastomers with a dual temperature
response
Marc Guerre1*, Christian Taplan1, Renaud Nicolaÿ,2 Johan M.
Winne1 and Filip E. Du Prez1*
1Polymer Chemistry Research Group and Laboratory for Organic
Synthesis, Department of Organic and Macromolecular Chemistry,
Ghent University Krijgslaan 281 S4-bis, B-9000 Ghent, Belgium.
E-mail: fi[email protected], [email protected].
2Matière Molle et Chimie, UMR 7167 CNRS-ESPCI Paris, Paris
Sciences et Lettres (PSL) Research University, 10 rue Vauquelin,
75005 Paris, France.
Abstract
Vitrimers are an emerging new class of permanently cross-linked
polymeric materials that show a liquid behavior upon heating
wherein the macroscopic deformation is controlled by the rate of
internal chemical bond exchange reactions. Thus, quite uniquely
among polymeric materials, flow rates and material viscosities can
be enhanced or controlled by the addition of catalysts and
additives. We now report a catalyst-free vitrimer system, prepared
from mixing two simple components, wherein two competing bond
exchange mechanisms co-exist, each showing a strikingly different
temperature dependence, related to the large difference in
activation energy for the different exchange pathways (60 vs
130-170 kJ/mol). The low barrier process is predominant at lower
temperatures, but is outcompeted by the high barrier process that
becomes dominant at higher temperatures because of its much more
pronounced temperature dependence. The result is an interesting and
highly unusual dual viscosity profile for this new class of
vitrimer materials: a very gradual decrease in viscosity at lower
temperatures, intercepted by a much sharper drop in viscosity at
higher temperatures. The highly counterintuitive effect where a
higher barrier pathway is dominant over a much lower barrier
process, can be rationalized by the exchange mechanisms that
involve different reactive species, but lead to the overall same
exchange. We observed this unusual but highly promising behavior
first for fluorinated vitrimer elastomers, aimed at high
performance materials, but the effect was also shown to hold in
related non-fluorinated elastomers. A new way to control and design
the rheological behavior of vitrimers towards finely tuned and
precisely controlled processing applications has thus been
provided.
Introduction
Synthetic materials that are built up from high molecular weight
polymer chains derive their material properties from intermolecular
forces, chain mobility and physical chain entanglements. The
resultant thermoplastic polymers are easily processed, but show a
limited durability, and poor resistance to deformation (creep),
weathering or dissolution. For applications where aspects such as
high dimensional stability and integrity are important, covalent
cross-linking of polymer chains is the desired strategy.
Classically, this results in thermosetting polymers that need to be
synthesized or at least fully hardened (or cured) in their final
shape, prohibiting most of the standard ‘plastic’ processing
possibilities. However, a clear paradigm shift has emerged around
the turn of the century: reversible cross-linking of polymer
chains, via strong supramolecular interactions1-3 or reversible
covalent bonds,4-7 can combine the easy processing and high
durability of thermoplastic and thermosetting polymers. The
viscoelastic properties of such materials can be changed or
introduced by exposure to a specific stimulus, such as light8 or
heat,9 giving a temporarily de-cross-linked material that can then
be reshaped, reprocessed, repaired or recycled. Given the growing
global demands and societal incentives towards more sustainable
materials, interest in this type of polymers has grown
exponentially.
Covalent cross-links that can be reversed to an open form, can
self-evidently transform a thermosetting topology into a
thermoplastic (processable) one, via a net-depolymerisation. In
2011,10 Leibler and coworkers introduced another paradigm shift
within the field of reversibly cross-linked polymer networks.
Somewhat counterintuitively, for macroscopic flow to occur in a
polymer network, depolymerisation is not a condition. Polymer
materials can indeed be designed wherein the network integrity is
preserved under all conditions, but where macroscopic flow is
enabled by covalent exchange reactions that rearrange the network
connectivity. For this, network defects or reactive side chain
groups must be able to associate themselves with existing network
points, and dislodge another polymer chain from this bonding site.
Interestingly, in such materials, the viscoelastic properties or
the rate of macroscopic flow are not primarily governed by
intrinsic chain mobility, or by the properties of the
non-cross-linked material, but is controlled by the rate of the
chemical exchange reaction, as the rate-limiting step in the
physical process of macroscopic flow.11 The result is a polymer
material that is completely insoluble at all temperatures, but that
can be processed and shows stress relaxation with a temperature
dependence of the viscosity, which aligns almost perfectly with the
temperature dependence of the chemical reaction. In fact, it has
been found consistently that measuring the stress relaxation times
of these materials at different temperatures, readily allows the
retrieval of the activation energy of the exchange reaction.12, 13
In contrast to all previously studied organic polymer liquids,
these non-depolymerising but deformable networks show a very
gradual viscosity decrease upon heating, a thermal behavior found
previously only in certain inorganic materials (strong liquids)
such as vitreous silica. Hence, upon the discovery of the first
such material by Leibler and coworkers, i.e. a catalyst-doped
polyester epoxy network, they introduced the term ‘vitrimers’ for
this new class of polymer materials. 14-25
Fluorinated polymers show remarkable differences in properties
compared to classical polymer materials, and have quickly found
commercially relevant applications because of their high inertness
and large impact on surface properties, most notably in inert
rubbers,26 coatings27 and membranes.28 Among fluoropolymers,
industrially produced perfluoropolyethers (PFPEs)29 are a class of
growing interest: they are liquids at room temperature, exhibit low
surface energy, good oxidative and thermal stability, are
characterized by excellent lubricating properties, biological
inertness, radiation resistance, pressure stability and are
non-flammable. Today, PFPE building blocks are widely used as
reactive intermediates in the synthesis of a large range of
materials for the creation of robust fluorinated networks (e.g.
polyurethanes, epoxy or polyester-based networks)30-33 for
aerospace materials, microfluidic devices, protective coatings,
optical applications,34-40 as well as electrolytes for lithium
batteries.41, 42
While the introduction of reversible covalent bonds inside a
fluoropolymer matrix could increase the mechanical properties and
temperature resistance, and would in principle give easily
(re)processable and repairable high performance fluorinated resins
or elastomers, the combination of Covalent Adaptable Networks
(CANs) with fluorinated materials has, to the best of our
knowledge, never been explored so far. We found one report in which
Li et al.43 reported a PFPE-based elastomer cross-linked via
supramolecular interactions, showing self-healing behavior in
combination with moderate mechanical properties, as expected for a
weakly cross-linked system.
Because of our ongoing research in the area of vitrimer-type
chemistries, leading to several interesting chemistry platforms for
material design, we were highly interested in implementing these
chemistries to PFPE with the aim to create fluorinated vitrimers.
We report here the full details of this study, in particular aimed
at introducing the catalyst-free vinylogous urethane vitrimer
chemistry44-47 into cross-linked PFPE elastomers. This study not
only reveals novel promising fluorinated materials with excellent
properties but also the full characterization of their
visco-elastic properties, which yields a unique insight into the
factors that govern their macroscopic flow.
Our study is the first and very clear example of a dual vitrimer
relaxation behavior wherein the covalent exchange reaction between
network points follows two different chemical pathways with a high
and low activation energy respectively. Counterintuitively, but
actually within the expectations of the underlying chemistry, the
high activation energy exchange can become the dominant process at
higher temperatures (130-160°C), leading to a highly unusual and so
far unprecedented thermal response of the viscosity of these
materials: a gradual and moderate decrease of the viscosity in the
lower temperature ranges, followed by a much sharper and pronounced
temperature response at higher temperatures. Our in-depth study of
the PFPE elastomers focused in first instance on stress-relaxation
experiments, but also included creep experiments, which fully
complement our findings. Moreover, we could also demonstrate that
the conceptually novel dualistic temperature response effect is not
related to the fluoropolymer backbone or the presence of additives
or impurities, but is rather related to the network architecture,
combined with the intrinsic properties of the remarkable exchange
reaction. Thus, this new and peculiar behavior should hold
relevance for many vitrimer-type materials.
Experimental part
Material
Benzylamine (>99%), hexylamine (>99 %), Polypropylene
glycol (PPG, Mn = 400 and 2000 g.mol-1), tert-butylacetoacetate
(TBAA, >98 %) and tris(2-aminoethyl)amine (TREN, 96%) were
purchased from Sigma Aldrich and used without further purification.
Fluorolink® E10H (PFPE-OH, Mn ≈ 1800 g·mol-1) was purchased from
Acota (UK). 1,1,1,3,3-Pentafluorobutane (Solkane® 365-mfc) was
purchased from Alpha Aesar. Tetrahydrofuran (THF),
dimethylacetamide (DMAc), dimethylformamide (DMF) and methanol
(MeOH) were purchased from Acros Organics. Sodium sulfate (Na2SO4)
and sodium hydroxide (NaOH) were purchased from Carl Roth.
Hydrochloric acid (HCl, 36 %) was purchased from Chem-lab.
Instrumentation
Nuclear magnetic resonance (NMR) spectra were recorded on a
Bruker Avance I Ultrashield 300 MHz spectrometer. Deuterated
acetone was used as the solvent in each sample. Coupling constants
and chemical shifts are given in hertz (Hz) and parts per million
(ppm), respectively. The experimental conditions for recording NMR
spectra were as follows: 1D 1H: flip angle, 30°; acquisition time,
4.1 s; pulse delay, 7.75 µs; number of scans, 16; spectral window,
8 kHz. 1D 13C: jmod pulse program was used with acquisition time of
1.36 s and 3072 number of scans. 2D COSY 1H-1H: The cosygpppqf
pulse sequence from Bruker catalogue was used without modification;
acquisition time, F1 0.01 s and F2 0.17 s; spectra window, 6 kHz
for both F1 and F2 dimensions; number of scans, 8. 2D HSQC 1H-13C:
The hsqcedetgp pulse sequence from Bruker catalogue was used
without modification; acquisition time, F1 7.68 ms and F2 0.16
s; spectra window, 6.4 kHz for F2 and 16.6 kHz for F1; number of
scans, 8. Diffusion ordered NMR spectroscopy (DOSY) experiments were
performed at 25 °C with 16 number of scans and 64 planes in F1
dimension.
Infrared (ATR-FTIR) spectra were collected using a Perkin–Elmer
Spectrum1000 FTIR infrared spectrometer with a diamond ATR
probe.
Size-Exclusion Chromatography (SEC) was recorded using a
triple-detection SEC from Agilent Technologies with its
corresponding Agilent software, dedicated to multidetector SEC
calculation. The system used two PL1113-6300 ResiPore 300 × 7.5 mm
columns with DMF (containing 0.1 wt. % of LiCl) as the
eluent with a flow rate of 0.8 mL·min-1 and toluene as the flow rate
marker. The detector suite was composed of a PL0390-06034 capillary
viscometer detector. The entire SEC-HPLC system was thermostated at
70 °C.
Thermogravimetric analyses (TGA) were performed with a Mettler
Toledo TGA/ SDTA851e instrument under air or nitrogen atmosphere
respectively at a heating rate of 10 K·min-1 from 25 °C to
600 °C for dynamic mode or at 100 °C and 150 °C
during 120 min for isothermal mode.
Differential scanning calorimetry (DSC) analyses were performed
with a Mettler Toledo instrument 1/700 under nitrogen atmosphere at
a heating rate of 10 K ·min-1.
Dynamic mechanical analyses (DMA) was performed on a SDTA861e
DMA from Mettler Toledo utilizing rectangular sample (ca. 2.1 mm
(T), 2.0 mm (W) and 8.0 mm (L)) and using tension clamp. The axial
force was adjusted to 0.1 N and a temperature ramp was then
performed from -140 °C to 100 °C at a rate of 3 K·min-1 with an
oscillating force of 0.1 N and a regular frequency of
1 Hz.
Rheology experiments were performed on an Anton Paar MCR 302.
The experiments were performed in parallel plate geometry using 8
mm sample disks. Unless specified, the experiments were performed
using a normal force of 1 N, an oscillating frequency of
1 Hz and a strain of 1 %. For all rheology experiments,
the applied stress was comprised in the linear viscoelastic region
at the measured temperatures. For amplitude sweep experiments, the
strain was varied from 0.01 to 100 %. For time sweep
experiments, a frequency of 1 Hz and a strain of 1 % were
applied and the storage modulus (G’) was followed over time at a
constant temperature. For stress relaxation experiments, a strain
of 1 % was applied to the material and the relaxation modulus
(G(t)) was followed over time at a constant temperature. The series
of stress relaxation experiments at different temperatures were
performed successively on the same sample. Creep experiments were
performed using a stress of 2800 Pa and the resulting strain
was followed over time at constant temperature. Creep recovery
experiments were performed using a stress alternating 0 Pa and
2800 Pa. In a typical test, a stress of 2800 Pa was
applied for 300 s, before a 100 s-recovery period. This
cycle was repeated multiple times and the corresponding strain
profile was recorded over time.
Uniaxial Tensile testing was performed on a Tinius-Olsen H10KT
tensile tester equipped with a 100 N load cell using ASTM standard
type IV dog bones (ISO 527-2-2B). The dog bone shaped samples had
an effective gauge length of 12 mm, a width of 2 mm and a
thickness of ± 2 mm and they were cut using a Ray-Ran
hand operated cutting press. The tensile measurements were
performed using a preload of 0.05 N and a pulling speed of
10 mm·min-1 until sample failure. The stress was recorded as a
function of strain . Reported values (elongation (%), stress
at break (Pa) and young modulus (Pa) are the average and standard
deviations of at least seven samples. Exhibited tensile curves were
selected to be representative of the average values.
Reprocessability. To reprocess the network, the polymer was
broken into small pieces (< 2mm) and placed into a rectangular
mold (70 mm x 40 mm x 2 mm) for compression molding.
This assembly was placed in a 150 °C preheated compression
press for 2 min under 1 metric ton of pressure. Then the
pressure was increased to 3 tons and kept constant for an
additional 3 min. After 5 min of pressing in total, the mold was
cooled down for 30 min and the sample was removed from the
metal sheet with a razor blade.
Solubility and hydrolytic stability experiments. Solubility
tests were carried out with samples of 4 mm diameter and
2 mm of thickness with a weight of around 60 mg and
5 mL of diverse solvents such as: DMF, MeOH, THF, and Solkane.
Those tests were performed for 24 h at 25 °C in DMF,
MeOH, THF, and Solkane, and also at 50 °C in THF and
80 °C in DMF. Then, the solvent was removed and the samples
were dried under vacuum, overnight at 100 °C. The soluble
fraction was calculated using equation (1), while the swelling
ratio was calculated using equation (2).
with mi, ms and md standing for initial, swollen and dry mass,
respectively.
As a control experiment, Soxhlet extraction was also carried out
in refluxing THF for 24 h in THF. Typically, 1 g of
sample was placed in an extraction unit of known weight and
subsequently immersed in dried THF as solvent. Next, the sample was
dried at 100 °C under vacuum overnight. The remaining dried
sample was weighted and the soluble fraction was calculated using
equation (1).
Hydrolytic stability experiments were carried out using a
similar sample size. Three different conditions were tested:
neutral (deionized water), basic (NaOH(aq) 1M) and acidic (HCl(aq)
1M). Tests were performed during 24 h at both 25 °C and
80 °C. Soluble fractions and swelling ratios were calculated
following equation (1) and (2), respectively.
Synthetic Procedures
Bis-perfluoropolyether acetoacetate (PFPE-AA): Fluorolink® E10H
(PFPE, Mn ≈ 1800 g.mol-1, 100 g, 5.56x10-2 mol) and
tert-butyl acetoacetate (19.3 g, 1.22x10-1 mol) were added in a
250 mL bottom flask. The viscous mixture was heated for
60 min at 130 °C, and tert-butanol product was
continuously removed by distillation throughout the reaction. When
the temperature of the still head dropped to 40 °C and
tert-butanol condensation stopped, the unreacted tert-butyl
acetoacetate was removed by distillation at 130 °C under
1 mbar vacuum. Yield = 99 %, functionality =
98 %, 1H NMR (300 MHz, (CD3)2CO, δ ppm): 1.92-1.97
(-CH=C(OH)-CH3, enol form (8.5 %)), 2.19-2.26
(-CH2-(C=O)-CH3), 3.49-3.75 (OCF2CH2-OCH2CH2-OCH2CH2- and
-(C=O)-CH2-(C=O)-CH3), 3.76-3.96 (-OCH2CH2-O-(C=O)-CH2-), 3.98-4.17
(-OCF2CH2-), 4.18-4.35 (-OCH2CH2-O-(C=O)-CH2-), 5.03-5.10
(-CH=C(OH)-CH3, enol form (8.5 %)).
Bis-perfluoropolyether hexyl vinylogous urethane (PFPE-HVU):
PFPE-AA (Mn ≈ 2000 g.mol-1, 3.0 g, 1.50x10-3 mol) and hexylamine
(1.52 g, 1.50x10-2 mol) were mixed in a 25 mL bottom flask and
stirred overnight at 60 °C. Then, water was added and the
mixture was extracted with Solkane twice. The combined fluorinated
phases were dried over Na2SO4 and the fluorinated solvent was
evaporated in vacuo yielding the desired compound.
Yield = 76 %, functionality = 98 %,
1H NMR (300 MHz, (CD3)2CO, δ ppm): 0.82-0.98 (-N(H)CH2(CH2)5CH3),
1.23-1.65 (-N(H)CH2(CH2)4CH3), 1.90-1.96 (-CH=C(NH-)-CH3),
3.20-3.34 (-N(H)CH2(CH2)5CH3), 3.44-3.92
(OCF2CH2-OCH2CH2-OCH2CH2-O-(C=O)-), 3.95-4.22 (-OCF2CH2- and
-OCH2CH2-O-(C=O)-), 4.34-4.44 (-CH=C(NH-)-CH3), 8.38-8.83
(-N(H)CH2(CH2)5CH3).
PPG-AA building blocks and PPG-VU networks were prepared
following the same conditions reported above for the PFPE-AA and
PFPE-VU network.
Model studies: Benzylamine (0.194 g, 1.82x10-3 mol) was
added to a solution of PFPE-HVU model compound (0.400 g,
1.83x10-4) in DMAc (1.0 mL). Five equivalents of benzylamine
were used to obtain a pseudo-first order reaction at low
conversions. The mixture was heated (100 °C, 120 °C, and
140 °C) in a Schlenk tube equipped with a septum. Samples were
taken at different time intervals, quenched in liquid nitrogen and
analyzed by 1H NMR in (CD3)2CO. The reaction was followed by
integration of the two distinct signals at 8.61 ppm and
8.96 ppm for the NH of hexyl and NH of benzyl model compounds,
respectively.
Network synthesis: PFPE-AA (fct. = 98 %,
20 g, 1.0x10-2 mol) and tris(2-aminoethyl)amine
(1.00 g, 6.86x10-3 mol) were mixed in a 60 mL
poly(propylene) vial using a DAC 150.1 FVZ speed mixer (typical
conditions of mixing: 5 min with a speed of 2500 rpm).
Then, after gelation, the polypropylene vial was broken, and the
network was cured overnight at 80 °C under vacuum. Finally,
the sample was cut into two pieces of 10 g, and pressed
separately in steel molds, following the aforementioned
reprocessability procedures.
Results and Discussion
PFPE functionalization and model reaction
Bis-perfluoropolyether acetoacetate (PFPE-AA) was synthesized
from Fluorolink® E10H (PFPE-OH), following a straightforward, high
yielding acetoacetylation reaction as depicted in Scheme 1
(top).
Scheme 1. a) Functionalization of Fluorolink E10H (PFPE-OH) into
PFPE-AA, b) network synthesis via polycondensation reaction of
acetoacetate and amine leading to vinylogous urethane bonds and c)
general scheme for transamination of vinylogous urethanes.
The acetoacetylation reaction between the bulk chemical TBAA and
polyol precursors allows for the simple preparation of
multifunctional acetoacetate building blocks in large scale. This
strategy has the advantage of releasing a low boiling point
compound, tert-butanol, which can easily be removed from the
reaction by distillation. Furthermore, the mixture does not require
supplementary purification steps since the remaining excess of TBAA
can be conveniently distilled under reduced pressure. In addition
to this facile purification protocol, the reaction can be carried
out in bulk, thus not requiring any organic or fluorinated solvent.
Indeed, the viscous liquid Fluorolink E10H is miscible with
tert-butylacetoacetate at 130 °C. This strategy differs from
prior PFPE functionalization procedures, which report the use of
fluorinated solvents, commonly classified as highly toxic and
bio-accumulative. 31, 32, 48
The obtained PFPE-AA structure was confirmed by 1D and 2D NMR
experiments (1H and 13C). Figure 1 displays the 1H NMR spectra of
TBAA (1), Fluorolink E10H (2), and the prepared PFPE-AA (3).
Figure 1. 1H NMR spectra of: 1) tert-butylacetoacetate, 2)
Fluorolink E10H and 3) PFPE-AA in (CD3)2CO at 25 °C.
Upon functionalization, a significant downfield shift was
observed for the ethylene glycol signals linked to the acetoacetate
group. Indeed, signals assigned to the -CH2CH2-OH group observed in
the area 3.50-3.90 ppm for Fluorolink E10H, were shifted to
3.76-3.96 ppm and 3.98-4.17 ppm for -CH2CH2-O-AA and
-CH2CH2-O-AA, respectively. The characteristic acetoacetate signals
were also shifted after the functionalization. The corresponding
methyl signal can clearly be identified as a sharp singlet at
2.22 ppm, which has slightly shifted in comparison to the one
of the starting TBAA. However, the signal assigned to the CH2 of
the acetoacetate function is not clearly visible and possibly
overlaps with signals of the ethylene glycol unit. The exact
position of the CH2 signal (at 3.53 ppm) in the 3.49-3.75 ppm
region was confirmed by a 2D correlation via COSY 1H-1H experiment
(Figure S1). The chemical structure was also confirmed by 13C, 1H
DOSY NMR and ATR-FTIR (Figure S2-S5). The DOSY 1H NMR spectrum
(Figure S4) shows the presence of a single diffusion coefficient
for the PFPE and acetoacetate signals, which suggests a successful
and quantitative functionalization of the alcohol moieties. The
FTIR spectrum (Figure S5) also confirms the qualitative character
of the acetoacetalization from the disappearance of the broad OH
vibrational band of Fluorolink E10H above 3400 cm-1 and the
appearance of new ketone vibrational bands at 1719 and
1740 cm-1.
Considering the importance of chain-end functionality in
stoichiometry calculations, the acetoacetate chain-end
functionality was carefully quantified by 1H NMR. Special attention
was directed to the reference peak (-OCF2CH2- signal,
4.18-4.35 ppm), since a small degree of coupling was reported
for the industrial synthesis of Fluorolink E10H.42 Indeed, the SEC
trace of Fluorolink E10H (Figure S6) clearly shows the presence of
mono-, di-, tri- and tetra- adducts at variant elution times. It is
worth noting that the result can be severely affected if these
coupling reactions are not adequately considered in functionality
calculations. Thus, by simply considering the -OCF2CH2- signal as
reference (four protons), and by assuming that this -OCF2CH2- group
is always alcohol end-terminated, the amount of acetoacetate
functions would be underestimated because the calculation does not
take into consideration the proportion of -OCF2CH2- per polymer
chain arising from coupling reactions. Nonetheless, this fraction
can be estimated by deconvolution of the PFPE SEC trace. Thereby,
the percentage of PFPE-AA acetoacetate functionality was finally
evaluated to be more than 98 % (See Figure S7 and supporting
information for calculation details). To further confirm the
functionality of the PFPE precursor, PFPE-AA was reacted with a
precise amount of mono-functional benzyl amine. After reaction of
0.90 and 1.00 eq of benzyl amine with respect to AA
functionalities, a functionalization of 89 and 98 % was achieved,
respectively (Figure S8). These results are in a good agreement
with the calculation achieved by 1H NMR.
The profound inductive effect of fluorine on the chemical
reactivity of organic compounds is well-described in organic and
polymer chemistry.49 Therefore, to dismiss a potential remote
withdrawing effect of the fluorinated chain on the transamination
exchange kinetics of vinylogous urethane, a model study with low MW
compounds was conducted. In order to compare the results and the
respective activation energy to a previously reported study by
Denissen et al.,46 the kinetic study strictly followed the therein
described conditions. Firstly, a PFPE-hexyl vinylogous urethane
(PFPE-HVU) was synthesized by condensation reaction of PFPE-AA with
hexylamine. Then, PFPE-HVU was reacted with five equivalents of
benzylamine at different temperatures (100 °C, 120 °C and
140 °C), and samples were taken at different time intervals
and subsequently analyzed by 1H NMR in deuterated acetone (Figure
2). The exchange kinetics can easily be followed by integration of
the two distinct signals at 8.61 ppm (NH of hexyl-VU) and
8.96 ppm (NH of benzyl-VU), respectively (Figure S9).
Figure 2. Progression of the exchange reaction of PFPE-HVU and
benzylamine at 100 °C, 120 °C and 140 °C.
Kinetics displayed in Figure 2 evolve as a first-order linear
decay at low conversion and tend to a chemical equilibrium.
Furthermore, the kinetic curves clearly show a temperature
dependence of the exchange reaction. From these kinetic studies,
resulting rate constants and Arrhenius law, an activation energy of
69 ± 2 kJ.mol-1 was calculated (Figure S10), which
is in good agreement with those earlier reported
(59 ± 6 kJ.mol-1 and
73 ± 12 kJ.mol-1).45, 46 In other words, the
withdrawing effect of perfluorinated chains does not significantly
influence the transamination exchange of vinylogous urethane groups
and hence, can be neglected for the following discussion.
Vitrimer synthesis and molecular characterization
PFPE vinylogous urethane (PFPE-VU) vitrimers were prepared
through the condensation reaction of PFPE-AA and cross-linker
tris(2-aminoethyl)amine (TREN) (Scheme 1, b) in bulk at
room temperature. Due to the high viscosity of PFPE-AA, which
further increased upon the reaction, a speed mixer was used to
enhance the homogeneity of the mixture. Since the polycondensation
reaction between acetoacetate and primary amine readily occurs at
room temperature, the gel point is reached after just a few
minutes. Then, the vitrimer was isolated as an orange rubber
(circular cube shape, Figure S11) before an additional post-curing
step of 16 h at 80 °C under vacuum was performed. The
curing step not only allows quantitative transformation of the
acetoacetate groups into vinylogous urethane, but also removes
water arising from the polycondensation. The obtained fluorinated
vitrimer was subsequently pressed at 150 °C in its final shape
(Figure S11).
In the transamination reaction of vinylogous urethane moieties,
free primary amines play a determinant role in the exchange
mechanism (Scheme 1, c). Indeed, they not only promote exchange
reactions without loss of integrity, but also allow a control over
the exchange kinetics. In other words, an excess of free primary
amines in vinylogous urethane vitrimers could significantly
accelerate the exchange reactions, while an amine shortfall could
drastically decrease or even stop them. Thus, in order to evaluate
the effect of free amines with respect to exchange kinetics and to
the resulting material properties, three PFPE-VU vitrimers with
different compositions were made: PFPE-VU1.00, PFPE-VU1.05 and
PFPE-VU1.10 with the subscript referring to 0, 5 and
10 mol% excess primary amine groups, respectively.
FTIR-spectra of the obtained PFPE-VU vitrimers are displayed in
Figure S12. The progress of the reaction was qualitatively
confirmed by the conversion of the acetoacetate group into
vinylogous urethanes as shown by the complete disappearance of
ketone vibration bands at 1719 and 1740 cm-1 and by the
appearance of new vibration bands at 1601 and 1650 cm-1,
assigned to C=C and C=O of vinylogous urethane, respectively. No
significant differences were noticed between the three PFPE-VU
vitrimers and no vibration band assigned to residual water was
observed around 3000 cm-1. The quantitative elimination of
water was further confirmed by isothermal TGA thermograms, in which
no weight loss was observed after two hours at 100 °C (Figure
S13).
Material characterization of vitrimers
Notably, the displayed DSC thermograms (Figure S14) reveal two
distinct glass transition temperatures: one transition at
-41 °C assigned to the amine core and vinylogous urethane
structures, and a second Tg around -100 °C assigned to the
fluorinated PFPE domains (similar to the Tg of Fluorolink E10H,
Figure S14 black curve). This observation was expected since
fluorinated polymers are prone to phase separation in the presence
of hydro-carbonated molecules and polymers.34, 35 DMA-experiments
performed for PFPE-VU vitrimers also confirm the presence of two Tg
values (Figure S15), i.e. a pronounced and broad glass
transition around 40 °C and a second transition at 90 °C.
Those values are in good agreement with the Tg of PFPE-methacrylate
thermosets reported by Hu et al.35 via DMA (Tg1 = 55 °C
and Tg2 = 80 °C). The appearance of a constant
plateau modulus in the range of 1.6 MPa further confirms the
cross-linked structure. Similar moduli were reported for
permanently and non-permanently cross-linked elastomers. For
instance, Ishibashi et al.50 found a modulus of 0.4 MPa for
PDMS-based vitrimers, while Vitale et al.51 reported a modulus of
6.3 MPa for highly cross-linked PFPE networks.
As reported by Desimone et al. for PFPE/PEG blends, the bulk
morphologies of segregated networks are known to be heterogeneous
and opaque.37 Indeed, although the PFPE-VU vitrimers appear
transparent after their preparation, those with 5 mol.% and
10 mol.% of free amines turned turbid within three days at
room temperature. This suggests a dynamic phase separation within
the fluorinated matrix at room temperature from nano-sized domains,
which do no scatter light, to micro-sized domains. Indeed, due to
the low Tgs, the overall mobility of both hydrocarbonated and
fluorinated domains is high at room temperature, which also implies
that exchange reactions are taking place at this temperature on
this time scale. Furthermore, after exposing the PFPE-VU vitrimers
for two minutes at 150 °C, the turbidity vanishes and the
materials become transparent again (Figure S16). This reversible
transition is likely the result of hydrocarbonated domains
rearrangement from microscopic into nanodomains because of dynamic
exchanges. Finally, the contrasting evolution of turbidity for the
vitrimers with different compositions can be rationalized by taking
into account the different excess of free amine groups. Indeed, as
will be reported in the creep analysis section (vide infra), an
excess of primary amine groups promotes a faster deformation upon
stress while the PFPE-VU1.00 vitrimer exhibits a much less
pronounced creep. Thus, we assume that the amine groups enable a
better migration and reorganization of the hydrogenated domains
within the fluorinated matrix.
The thermostability of all vitrimers was analyzed by TGA under
both nitrogen and air atmosphere (Figure S17). With a heating rate
of 10 K.min-1, the onset of degradation (5% weight loss) is
observed around 280-290°C in nitrogen and 265 °C in air for
the three samples. These results are consistent with the values
obtained for robust48 permanently cross-link PFPE-based networks
reported in the literature. Thus, the reversible vinylogous
urethane chemistry does not have a significant impact on the
thermal stability of PFPE-based materials. To estimate the
effective thermal stability of PFPE vitrimers, isothermal
experiments were also conducted. A weight loss of 1 % was
observed after two hours in air at 150 °C (Figure S18) while
the thermal degradation temperature of PFPE vitrimers in air was
about 265 °C. In other words, such PFPE vitrimers cannot be
continuously exposed over a longer period to temperatures exceeding
150 °C without taking into account a potential damage of the
material. However, since the time required for reprocessing is in
the order of minutes at 150 °C (vide infra), no significant
influence is expected from network degradation.
In the next step, viscoelastic properties of PFPE-VU vitrimers
were in first instance studied by stress relaxation experiments
(Figure 3, S19 and S20). For the rheological measurements, a
constant strain of 1 % was applied, which is comprised in the
linear viscoelastic region at the measured temperatures, and the
relaxation modulus was followed as a function of time.
Figure 3. Representative stress relaxation experiments and
Arrhenius plots of PFPE-VU vitrimers (a) Stress relaxations of
PFPE-VU1.00 (black), PFPE-VU1.05 (red), and PFPE-VU1.10 (blue)
vitrimers at 150 °C (b) Stress relaxation curves of
PFPE-VU1.05 for temperature interval between 90 °C and
160 °C. The dotted line represents E/E0 = 1/e, which defines
the characteristic relaxation time ԏ* (c) Arrhenius plots of
PFPE-vitrimers with corresponding exchange mechanisms involved in
the dual relaxation behavior.
The stress relaxation curves shown in Figure 3a and b, show a
full stress relaxation as a function of time. This relaxation
behavior, which follows the Maxwell model, is the result of a
vitrimer network that is entirely composed of reversible bonds
without irreversible cross-links. As expected, relaxation time
values (ԏ*) at 150 °C decreased with an increasing content of
free primary amines in the network (Figure 3a), from approximately
390 s (PFPE-VU1.00), over 160 s (PFPE-VU1.05) to
57 s (PFPE-VU1.10). Also, when the temperature is increased
from 90 °C to 160 °C for the PFPE-VU1.05 sample, ԏ* similarly
decreased much from 4390 s to 33 s, which is consistent with
regard to other reported vitrimer systems. From those relaxation
times, the relationship between Ln(ԏ*) and 1000/T was fitted to the
Arrhenius equation (Figure 3c) and the corresponding activation
energy was calculated from the slope.
In the initial study of vinylogous urethane vitrimers, Denissen
et al.46 reported a good agreement between the Ea determined via
chemical kinetic studies with low molar mass model compounds on the
one hand, and the value determined via vitrimer stress relaxation
measurements on the other hand. In contrast, within these systems
containing much longer chain fragments, we observed a significant
shift in Ea, which not only depends considerably on the network
stoichiometry, but remarkably also on temperature: Ea of 58 ±
4 kJ.mol-1 for PFPE-VU1.00, 60 ± 2 and 169 ± 17 kJ.mol-1 for
PFPE-VU1.05 as well as 58 ± 3 and 129 ± 3 kJ.mol-1 for
PFPE-VU1.10 (from two different slopes). Previously, such shifts in
activation energy had only been observed in catalyst- or
additive-doped vinylogous urethane vitrimers,45 and they have never
been observed within one and the same material. As no significant
difference of shear modulus was noticed between all vitrimers (see
time sweep experiments in Figure S21), this variation in
activation energies can neither be related to a difference of
cross-linking density nor to a change in chain mobility.
Given what is known about the molecular basis of the rheology of
vitrimers, we hypothesize that the observed peculiar shift in
activation energy, with a stronger temperature dependence at higher
temperatures, is related to the dynamic exchange mechanism. Indeed,
the Arrhenius plots of Figure 3c can be explained by the presence
of two distinct mechanisms (or pathways) for the same dynamic bond
exchange process: one mechanism or pathway corresponding to an
activation energy barrier of ca. 60 kJ.mol-1, and a second one
corresponding to an enthalpic barrier of ca. 130-170 kJ.mol1.
Interestingly, PFPE-VU1.05 and PFPE-VU1.10 show this dual exchange
mechanism on the same Arrhenius plot, while the transition point
appears to depend on the network composition. Based on the
Arrhenius plot, this transition was estimated to be ca. 145 °C
and ca. 130 °C for PFPE-VU1.05 and PFPE-VU1.10, respectively.
To verify the reproducibility and the significance of this dual
temperature response, stress relaxation experiments were repeated 3
times on the same PFPE-VU1.05 vitrimer (see error bars in Figure
3). Arrhenius plots of those experiments (Figure S22) clearly show
that the dual-viscosity behavior remains distinctively visible
The herein described presence of a dual exchange mechanism
within one material has never been reported so far for any vitrimer
material, and thus deserves closer scrutiny. In previous studies,
we have found45 that under neutral or slightly acidic conditions,
the amine exchange invariably proceeds via a ‘protic pathway’
involving an activated iminium intermediate as electrophilic
species, corresponding to an enthalpic barrier (Ea) of about 73-81
kJ.mol-1 (Figure 3c, pathway (i)). We have also found that adding a
strong base blocks this pathway and rather promotes a much slower
direct addition pathway between a neutral free amine and a neutral
VU-moiety as a Michael-type acceptor (Figure 3c, pathway (ii)).
This ‘background’ process was previously only revealed in the
presence of a strong base, and was found to correspond to a much
higher activation barrier (Ea) ranging from 102 to 122 kJ.mol-1. By
proposing these known distinct transamination mechanisms as
competing relaxation pathways within PFPE-VU1.00, PFPE-VU1.05 and
PFPE-VU1.10, the observed rheological behavior can thus be
satisfactorily understood at a molecular level, i.e. at
sufficiently low temperatures, exchange will proceed via an iminium
exchange, while at higher temperatures, the higher barrier process
will become the dominant exchange pathway, as this reaction has a
much more pronounced temperature dependence.
Another clear indication of the two distinct pathways can be
found in the observation that the high barrier (Michael-type)
addition pathway is more strongly dependent on the free amine
content of the material, as can be seen in Figure 3c: the steeper
Arrhenius fitted lines, corresponding to the higher barrier
exchange, drop more significantly with the increase of the free
amine content. In the low barrier iminium pathway (i), attack of an
amine species to the very electrophilic (and high energy)
protonated intermediate will be rapid and is likely non-rate
determining. In contrast, the bimolecular step between an
unactivated, neutral VU and an amine will be rate determining for
pathway (ii), as the only other reaction step involved is an
internal proton transfer between the exchanging amine groups, which
should be very fast.
Remarkably, the high barrier mechanism is absent or kinetically
insignificant in all other investigated vinylogous urethane based
materials, while it is quite pronounced in the fluorinated
vitrimers studied here. Because this unprecedented and quite
unexpected double relaxation behavior can have serious implications
for vitrimer processing applications, and could allow for the
rational design of new materials with highly ‘tailored’ rheology
profiles, we examined if this novel dual temperature response
effect is related to the fluoropolymer backbone - or the presence
of additives or impurities therein - or if it is rather related to
the specific network architecture that includes long chain
fragments between the exchangeable VU cross-links. All previous VU
vitrimers were built up from small molecule monomers only, and did
not contain larger non-dynamic fragments or chains. In fact, we
found a very similar dual relaxation profile based on a VU vitrimer
derived from a non-fluorinated poly(propylene glycol) backbone of
comparable MW (PPG-VU1.05, Figure S23), implicating the molecular
architecture as the main facilitator of the high energy pathway,
and also pointing towards the general applicability of this dual
rheology as a novel design concept in vitrimer chemistry.
With respect to the established operative exchange mechanisms in
these VU materials, it is worth noting that PFPE-VU1.00, which
theoretically should not have free amine groups, also exhibits a
full stress relaxation. It is hypothesized that this behavior can
be related to minor stoichiometry discrepancies and network defects
as ‘exchange hotspots’ in the vitrimer matrix. This further
confirms the view that VU-protonation to the reactive iminium form
is the most relevant step in the dynamic exchange mechanism, as a
depletion of free amines does not shut down the exchange pathway.
As expected, based on the bimolecular rate determining step, the
second pathway is absent in stoichiometrical balanced VU networks.
To verify if relaxation could also occur within a network without
any free amines, the synthesis of a control sample was attempted
with 5 % excess of acetoacetate moieties. However, even after
24 hours at 80 °C, the gel point for the reaction between
PFPE-AA and TREN was not reached and the mixture remained
liquid.
In 2011,10 with the first vitrimer report, the correlation
between the chemical relaxation and the viscoelastic properties
resulted in the introduction of a specific temperature associated
to vitrimer systems, i.e. the topology freezing transition
temperature, labeled Tv. By definition, Tv is defined as the
temperature above which a reversible network topology can be
achieved through dynamic exchange. More precisely, the Tv has been
defined as the temperature at which solid to liquid transition
occurs with a defined viscosity of η = 1012 Pa·s.11 To estimate the
related Tv values, the Maxwell equation is typically used to
correlate the relaxation time to Tv (see SI for calculation
details). A hypothetical Tv of around 10 °C was calculated for the
fluorinated vitrimers (vide infra). Since for those elastomers, the
Tv is much higher than the detected low Tg, the material
viscosities will be mainly governed by the Tv,12 meaning that
exchange reactions are expected to already occur at room
temperature, which is perfectly consistent with the observed
turbidity appearance after 3 days at ambient conditions.
To evaluate the effect of such low Tv-value on the elastomeric
material properties, cyclic creep recovery experiments, i.e. with
the stress varying between 0 and 2800 Pa, have been conducted for
PFPE-VU1.00, PFPE1.05 and PFPE-VU1.10 at 25 °C (Figure S24). The
PFPE-VU vitrimers behave like a standard rubber at room
temperature52 on these time scales, with a nearly full recovery of
their initial shape after successive shearing. For the evaluation
of creep at higher temperature, a similar experiment was conducted
with temperature variations between 25 and 80 °C (Figure 4).
Figure 4. Cyclic creep recovery profiles of PFPE-VU1.05 vitrimer
at 25 °C and 80 °C. The lower dashed black line
corresponds to the applied force (2800 Pa) while the upper red line
shows the strain response.
As expected from the acceleration of the dynamic exchange at
high temperature, the materials’ response at 80 °C is
completely different. Upon a first administration of 2800 Pa
at 80 °C, the PFPE-VU1.05 elastomer instantaneously deforms
about 1.2 % and does not return to its initial shape after
removing the applied stress. After two cycles at 80 °C, a
deformation of 1.5 % is observed, while a permanent
deformation of 0.9 % was obtained after stress release.
Nonetheless, upon cooling to 25 °C, the PFPE-VU1.05 elastomer
regains its initial properties and behaves as a typical rubber with
a full recovery of its shape after three cycles. To further explore
the differences of creep in relation to the amine content for
PFPEVU1.00, PFPEVU1.05 and PFPEVU1.10 vitrimers, creep recovery
experiments at different temperatures were achieved. (Figure 5,
Figure S25).
Figure 5. (a,b) Creep recovery profiles of PFPE-VU1.00 (a) and
PFPE-VU1.05 (b) at 20, 40, 60, 80 and 100 °C with an applied
stress of 2800 Pa for 20 min and a recovery of 20 min (c)
Creep rates of PFPE-VU1.00, PFPE-VU1.05 and PFPE-VU1.10 from
20 °C to 100 °C (d) Arrhenius plots of PFPE-VU1.00,
PFPE-VU1.05 and PFPE-VU1.10 extracted from creep recovery
experiments.
Figure 5a,b shows that PFPE-VU1.05 and PFPE-VU1.00 can withstand
temperatures above 40 °C without significant deformation,
while its response at high temperature is much more significant. A
similar temperature dependence was observed for PFPE-VU1.10
vitrimers (Figure S25). From those creep recovery experiments, the
rate of deformation, also called creep rate (),53, 54 was plotted
as a function of the temperature (Figure 5c). A similar creep onset
for the different PFPE-VU vitrimers can be observed, i.e. no
significant creep rate is observed until 40 °C but it
increases suddenly with the temperature and at different rates
according to the vitrimer composition. This difference in creep
behavior, especially at high temperature, can again be related to
the amount of excess amine groups: an excess of amine promotes fast
deformation upon stress as a result of better ability of the
network to rearrange its structure through the underlying
associative mechanism. However, although an increase of free
primary amines allows a fast relaxation of the stress and thus fast
reshaping/recycling of the network, it can also significantly
affect the mechanical properties and oxidative sensitivity of the
resulting materials. For that reason, the amount of free amines
must be judiciously selected as a function of the desired
applications and specifications.
It is important to mention that the determination of Tv from the
average slope of the Arrhenius data points, without taking the dual
exchange mechanism into account, results in Tv values of 9 °C,
21 °C and 50 °C for PFPE-VU1.00, PFPE-VU1.05 and
PFPE-VU1.10, respectively. However, such Tv values are in
disagreement with the results of creep experiments, where the creep
onset temperatures are the same for all PFPE-VU vitrimers.
Moreover, from the creep recovery experiments, the relationship
between ln () and temperature (1000/T) was fitted to the
Arrhenius equation55 (Figure 5d) and the corresponding activation
energy was assessed from the slope. Activation energies of ca.
60 kJ.mol-1 were obtained in relatively good agreement with
activation energies determined from stress relaxation experiments.
This match in temperature dependence confirms the fact that at low
temperature, exchange reactions continue to be governed by the low
barrier process. Thus, even though the estimated Tv, obtained by
extrapolation from a high temperature Arrhenius-fitted line, should
give an approximate idea of the temperature where the dynamic
exchange becomes relevant, this can be an ambiguous and flawed way
to characterise the viscoelastic behavior of vitrimer materials.
Indeed, it should always be taken into account that a lower barrier
background exchange reaction can take over at lower temperatures,
leading to a lower Tv than anticipated based on the measurement of
the high temperature rheology. In this context, for low Tg
materials, creep experiments should be regarded as a valuable and
complementary analytical technique for a more rigorous
characterisation of vitrimers in addition to the high temperature
stress relaxation experiments.
Reprocessability, soluble fraction/swelling ratio and hydrolytic
stability
To evaluate the mechanical properties of the PFPE-VU vitrimers
after successive recycling steps, the vitrimers were reprocessed
five times using each time the same conditions, i.e. thermal
compression moulding at 150 °C for a duration of
5 minutes after a grinding process. Tensile strength
experiments were achieved at 25 °C before each recycling step
(Figure S26-S29). The stress at break, Young modulus and elongation
at break of these samples are given in Figure 6. Initially,
PFPE-VU1.00, PFPE-VU1.05 and PFPE-VU1.10 exhibited similar
properties. PFPE-VU1.05 and PFPE-VU1.10 recovered 95 % of
their original properties and maintained this level, even after
multiple recycling steps. Conversely, a noticeable decrease in
stress at break from 0.93 MPa to 0.59 MPa upon the first
reprocessing step of PFPE-VU1.00 was observed. After additional
recycling steps, the stress at break progressively increased to a
plateau of 0.73 MPa, but yet remained 22 % lower than the
initial values. This drop in mechanical properties is probably the
result of an insufficient healing step due to the lack of free
amines, i.e. after grinding the network into small pieces, the low
amount of free amines is not sufficient to efficiently fuse all
pieces together and restore the original properties.
The recycled PFPE-VU vitrimers were further analyzed by
ATR-FTIR, DSC, TGA and rheology. On FTIR spectra, the recycled
samples exhibit identical band vibrations without additional
signals that could indicate the alteration of the network structure
through the occurrence of side reactions (Figure S29). Accordingly,
the same Tgs were observed on DSC thermograms (Figure S30). The
detected thermal stability, however, showed a slight decreased from
286 °C to 276 °C compared to the non-reprocessed
vitrimers (Figure S31). Stress relaxations experiments and the
resulting Arrhenius plots were also reproduced for a recycled
sample, again exhibiting the remarkable dual relaxation profile
(Figure S32). Finally, time sweep experiments (Figure S33) showed
identical shear modulus for the original and recycled samples,
suggesting that the cross-linking density remains constant over
successive recycling steps.
Figure 6. Elongation, Young modulus and stress at break values
of PFPE-VU1.00, PFPE-VU1.05 and PFPE-VU1.10 after different
recycling steps (up to five). Note: Full lines are a guide for the
eyes.
The soluble fraction and swelling ratio of PFPE vitrimers are
displayed in Table 1. The soluble fractions from swelling
experiments performed in different solvents (THF, DMF, MeOH and
Solkane) remained below 8 %, even at high temperature, which
confirms the high and permanent cross-link density of the
materials. At 25 °C, the soluble fractions of the original
vitrimers were even lower than 1 %. The increase in free amine
content in off-stoichiometric networks do not significantly affect
the swelling of the network. As expected, the highest swelling
ratio is obtained in the fluorinated solvent (i.e. Solkane, 100120
%). However, the soluble fraction in such a solvent is similar to
DMF and THF at high temperature, meaning that PFPE has also a good
solubility in those solvents. Importantly, the soluble fraction and
swelling ratio of PFPE vitrimers after five recycling steps are
comparable to the original vitrimers, i.e. an increase of maximum
2 % soluble fraction and 60 % swelling ratio is observed.
Additionally, the more significant increases are observed with to
PFPE-VU1.00, which is consistent with a decrease of tensile
strength and modulus in time sweep experiments. This again
indicates the slight decrease of the cross-linking density, due to
the incomplete healing of the material.
Table 1. Soluble fraction and swelling ratio of PFPE-VU1.00,
PFPE-VU1.05 and PFPE-VU1.10 vitrimers before and after five
recycling steps.
Soluble fraction Swelling ratio
(%)
1.00
1.05
1.10
1.00 (RC x5)
1.05 (RC x5)
1.10 (RC x5)
THF 25 °C
0.5
31
0.8
35
0.9
36
1.5
33
1.3
36
1.2
35
DMF 25 °C
0.8
18
1.0
19
1.0
19
0.9
18
1.0
20
1.1
19
MeOH 25 °C
0.6
6
1.0
8
1.0
8
1.9
6
1.1
7
1.1
7
Solkane 25 °C
3.8
101
5.0
120
5.2
119
7.0
160
7.0
127
6.7
124
DMF 80 °C
5.0
19
6.2
20
6.0
20
7.3
19
7.5
22
6.2
21
THF 50 °C
5.3
30
6.6
35
5.4
34
7.1
31
7.9
35
6.5
30
Note: A control experiment using Soxhlet extraction for
PFPE-VU1.05 vitrimer in refluxing THF for 24 h resulted in a
soluble fraction of 6.0 %.
As expected for VU vitrimer materials, a treatment with an
excess of primary amine in a good solvent at high temperature
resulted in a complete dissolution of the vitrimers,46, 56 caused
by dynamic amine exchanges leading to a net depolymerisation
(Figure S34).
Finally, in view of potential applications for fluorinated
vitrimers, the first examples of which we report herein, the
hydrolytic stability of our materials was tested by subjecting
PFPE-VU1.05 samples to three different hydrolytic conditions: 1M
NaOH, neutral water and 1M HCl solutions (Table S2). After
24 h at 25 °C in those conditions, the general appearance
of the samples remained the same (Figure S35). Generally, the
PFPE-VU vitrimers were remarkably stable under both neutral and
basic conditions ranging from 25 °C until 80 °C (See
Table S1, soluble fraction and swelling ratio). However, the
stability in acidic conditions is less pronounced. At 25 °C in
1M HCl solution, the network keeps its integrity but it does
markedly deteriorate when heated to 80 °C in this strong
acidic environment, as the sample becomes brittle and degrades into
small fragments. Thus, the material’s stability is preserved under
basic and neutral conditions, even at high temperature, but suffers
from hydrolysis under harsh acidic conditions.
Conclusion
In summary, fluorinated vitrimers were prepared by implementing
the catalyst-free transamination of vinylogous urethanes into a
perfluoropolyether (PFPE) network matrix. These fluoroelastomers
were synthesized in two simple steps from a commercially available
ethoxylated prepolymer, using first a bulk end group
functionalization with an acetoacetylating agent. This
straightforward setup allows the preparation of PFPE precursors in
very high yield through a cost-effective and solvent-free protocol.
The final elastomers were prepared by simply combining the
end-functionalized PFPEs with a commercial trifunctional amine,
giving fluoroelastomers that, over the conducted experiments
timescale, behave like a normal rubber with a full elastic response
at room temperature. At higher temperatures, however, the herein
presented fluoroelastomers exhibit a fast stress relaxation without
the use of additives or catalysts and can be recycled and reshaped
multiple times thanks to the dynamic exchange reactions. In this
context, the reprocessability of PFPE vitrimers was also verified.
Thermal, mechanical and rheological properties of the materials
were investigated and compared before and after five recycling
steps. It could be shown that the created vitrimers recovered 95 %
of their initial mechanical properties. Besides, similar analyses
such as thermal stability, stress relaxation and infrared
spectroscopy demonstrated that the PFPE vitrimers can be recycled
multiple times without noticeable degradation. In addition to that,
very low soluble fractions and swelling ratios were obtained in
common organic solvents. A first general conclusion is that this
paper describes a promising substitute for currently available
permanently cross-linked fluorinated elastomers.
With respect to the observed relaxation times, highly remarkable
and unprecedented effects were noted and verified in several ways.
The rheological behavior of the materials can be readily controlled
by adjusting the amount of free primary amine moieties, as observed
in previous systems. However, these catalyst-free elastomer
networks obtained from two simple components, showed an unexpected
and quite counterintuitive double relaxation behavior, implicating
different chemical exchange reactions that are operating within the
same simple polymer matrix. At low temperature, the exchange
reactions are governed by the normal, previously established
mechanism for amine exchange with an activation energy of 60
kJ.mol-1. On the other hand, at higher temperature, a sharp and
clear change in temperature dependence was noted, with a fitted
activation energy that is more than double that of the low
temperature exchange (130 kJ.mol-1).
In light of our in-depth investigation of the unusual dual
rheological behavior of these materials, we found that the
activation energies at low temperatures could also be readily
determined from creep recovery experiments. This suggests that
creep recovery experiments for vitrimer elastomers can be
considered as a rapid alternative to get a better view of the
(complete) rheological profile, which we now speculate may often be
more complicated than anticipated based solely on stress-relaxation
experiments, which are often limited to high, relatively small
temperature intervals. Our work also points out an obvious danger
in calculating the topology freezing transition temperature, Tv,
since extrapolation from a high temperature zone to a lower
temperature may miss the fact that the higher temperature process
will no longer be the dominant process at the lower temperature. In
short, we can conclude that, as vitrimer rheology is a physical
process essentially governed by chemical reactivity, polymer
chemists should be prepared and ready to find and take into account
all the peculiarities and extreme context-dependence that typify
the study of chemical reactivity, even within a very clear-cut and
chemically uncomplicated situation.
Acknowledgements
The authors thank Prof. Dr. Ludwik Leibler for his valuable
comments. M. G. and C. T. acknowledge the Research
Foundation-Flanders (FWO) for the Postdoctoral and Ph. D.
fellowships. The authors would like to thank Bernhard de Meyer and
Bastiaan Dhanis for the technical support. F.D.P thanks BOF-UGent
(GOA-funding).
Author contributions
The manuscript was written through contributions of all authors.
All authors have given approval to the final version of the
manuscript.
Supporting Information
The supporting information is available free of charge on the
ACS Publications website at DOI: 10.1021/XXXXX.
Detailed information about material characterization, NMR
Spectroscopy, end group quantification, model study, FTIR, thermal
analysis (DSC, TGA and DMA), stress relaxations, Tv determination,
creep recovery experiments, tensile strength, hydrolytic
stability.
References
1.Appel, E. A.; Biedermann, F.; Rauwald, U.; Jones, S. T.;
Zayed, J. M.; Scherman, O. A., Supramolecular Cross-Linked Networks
via Host−Guest Complexation with Cucurbit[8]uril. J. Am. Chem. Soc.
2010, 132 (40), 14251-14260.
2.Yan, X.; Wang, F.; Zheng, B.; Huang, F., Stimuli-responsive
supramolecular polymeric materials. Chem. Soc. Rev. 2012, 41 (18),
6042-6065.
3.Yang, L.; Tan, X.; Wang, Z.; Zhang, X., Supramolecular
Polymers: Historical Development, Preparation, Characterization,
and Functions. Chem. Rev. 2015, 115 (15), 7196-7239.
4.Chen, X.; Dam, M. A.; Ono, K.; Mal, A.; Shen, H.; Nutt, S. R.;
Sheran, K.; Wudl, F., A Thermally Re-mendable Cross-Linked
Polymeric Material. Science 2002, 295 (5560), 1698-1702.
5.Scott, T. F.; Schneider, A. D.; Cook, W. D.; Bowman, C. N.,
Photoinduced Plasticity in Cross-Linked Polymers. Science 2005, 308
(5728), 1615-1617.
6.Bowman, C. N.; Kloxin, C. J., Covalent Adaptable Networks:
Reversible Bond Structures Incorporated in Polymer Networks. Angew.
Chem. Int. Ed. 2012, 51 (18), 4272-4274.
7.Zou, W.; Dong, J.; Luo, Y.; Zhao, Q.; Xie, T., Dynamic
Covalent Polymer Networks: from Old Chemistry to Modern Day
Innovations. Adv. Mater. 2017, 29 (14), 1606100.
8.Amamoto, Y.; Kamada, J.; Otsuka, H.; Takahara, A.;
Matyjaszewski, K., Repeatable Photoinduced Self‐Healing of
Covalently Cross‐Linked Polymers through Reshuffling of
Trithiocarbonate Units. Angew. Chem. Int. Ed. 2011, 50 (7),
1660-1663.
9.Gheneim, R.; Perez-Berumen, C.; Gandini, A., Diels−Alder
Reactions with Novel Polymeric Dienes and Dienophiles: Synthesis
of Reversibly Cross-Linked Elastomers. Macromolecules 2002, 35
(19), 7246-7253.
10.Montarnal, D.; Capelot, M.; Tournilhac, F.; Leibler, L.,
Silica-Like Malleable Materials from Permanent Organic Networks.
Science 2011, 334 (6058), 965-968.
11.Denissen, W.; Winne, J. M.; Du Prez, F. E., Vitrimers:
permanent organic networks with glass-like fluidity. Chem. Sci.
2016, 7 (1), 30-38.
12.Capelot, M.; Unterlass, M. M.; Tournilhac, F.; Leibler, L.,
Catalytic Control of the Vitrimer Glass Transition. ACS Macro Lett.
2012, 1 (7), 789-792.
13.Capelot, M.; Montarnal, D.; Tournilhac, F.; Leibler, L.,
Metal-Catalyzed Transesterification for Healing and Assembling of
Thermosets. J. Am. Chem. Soc. 2012, 134 (18), 7664-7667.
14.Snyder, R. L.; Fortman, D. J.; De Hoe, G. X.; Hillmyer, M.
A.; Dichtel, W. R., Reprocessable Acid-Degradable Polycarbonate
Vitrimers. Macromolecules 2018, 51 (2), 389-397.
15.Fortman, D. J.; Brutman, J. P.; Cramer, C. J.; Hillmyer, M.
A.; Dichtel, W. R., Mechanically Activated, Catalyst-Free
Polyhydroxyurethane Vitrimers. J. Am. Chem. Soc. 2015, 137 (44),
14019-14022.
16.Obadia, M. M.; Mudraboyina, B. P.; Serghei, A.; Montarnal,
D.; Drockenmuller, E., Reprocessing and Recycling of Highly
Cross-Linked Ion-Conducting Networks through Transalkylation
Exchanges of C–N Bonds. J. Am. Chem. Soc. 2015, 137 (18),
6078-6083.
17.Nishimura, Y.; Chung, J.; Muradyan, H.; Guan, Z., Silyl Ether
as a Robust and Thermally Stable Dynamic Covalent Motif for
Malleable Polymer Design. J. Am. Chem. Soc. 2017, 139 (42),
14881-14884.
18.Liu, W.-X.; Zhang, C.; Zhang, H.; Zhao, N.; Yu, Z.-X.; Xu,
J., Oxime-Based and Catalyst-Free Dynamic Covalent Polyurethanes.
J. Am. Chem. Soc. 2017, 139 (25), 8678-8684.
19.Lu, Y.-X.; Tournilhac, F.; Leibler, L.; Guan, Z., Making
Insoluble Polymer Networks Malleable via Olefin Metathesis. J. Am.
Chem. Soc. 2012, 134 (20), 8424-8427.
20.Lu, Y.-X.; Guan, Z., Olefin Metathesis for Effective Polymer
Healing via Dynamic Exchange of Strong Carbon–Carbon Double Bonds.
J. Am. Chem. Soc. 2012, 134 (34), 14226-14231.
21.Taynton, P.; Yu, K.; Shoemaker, R. K.; Jin, Y.; Qi, H. J.;
Zhang, W., Heat‐ or Water‐Driven Malleability in a Highly
Recyclable Covalent Network Polymer. Adv. Mater. 2014, 26 (23),
3938-3942.
22.Ogden, W. A.; Guan, Z., Recyclable, Strong, and Highly
Malleable Thermosets Based on Boroxine Networks. J. Am. Chem. Soc.
2018, 140 (20), 6217-6220.
23.Cromwell, O. R.; Chung, J.; Guan, Z., Malleable and
Self-Healing Covalent Polymer Networks through Tunable Dynamic
Boronic Ester Bonds. J. Am. Chem. Soc. 2015, 137 (20),
6492-6495.
24.Rekondo, A.; Martin, R.; Ruiz de Luzuriaga, A.; Cabanero, G.;
Grande, H. J.; Odriozola, I., Catalyst-free room-temperature
self-healing elastomers based on aromatic disulfide metathesis.
Mater. Horiz. 2014, 1 (2), 237-240.
25.An, X.; Aguirresarobe, R. H.; Irusta, L.; Ruiperez, F.;
Matxain, J. M.; Pan, X.; Aramburu, N.; Mecerreyes, D.; Sardon, H.;
Zhu, J., Aromatic diselenide crosslinkers to enhance the
reprocessability and self-healing of polyurethane thermosets.
Polym. Chem. 2017, 8 (23), 3641-3646.
26.Améduri, B.; Boutevin, B.; Kostov, G., Fluoroelastomers:
synthesis, properties and applications. Prog. Polym. Sci. 2001, 26
(1), 105-187.
27.Vitale, A.; Bongiovanni, R.; Ameduri, B., Fluorinated
Oligomers and Polymers in Photopolymerization. Chem. Rev. 2015, 115
(16), 8835-8866.
28.Cui, Z.; Drioli, E.; Lee, Y. M., Recent progress in
fluoropolymers for membranes. Prog. Polym. Sci. 2014, 39 (1),
164-198.
29.Friesen, C. M.; Améduri, B., Outstanding telechelic
perfluoropolyalkylethers and applications therefrom. Prog. Polym.
Sci. 2018, 81, 238-280.
30.Yao, X.; Dunn, S. S.; Kim, P.; Duffy, M.; Alvarenga, J.;
Aizenberg, J., Fluorogel Elastomers with Tunable Transparency,
Elasticity, Shape‐Memory, and Antifouling Properties. Angew. Chem.
Int. Ed. 2014, 53 (17), 4418-4422.
31.Yang, Y.-W.; Hentschel, J.; Chen, Y.-C.; Lazari, M.; Zeng,
H.; Michael van Dam, R.; Guan, Z., "Clicked" fluoropolymer
elastomers as robust materials for potential microfluidic device
applications. J. Mater. Chem. 2012, 22 (3), 1100-1106.
32.Hu, Z.; Finlay, J. A.; Chen, L.; Betts, D. E.; Hillmyer, M.
A.; Callow, M. E.; Callow, J. A.; DeSimone, J. M., Photochemically
Cross-Linked Perfluoropolyether-Based Elastomers: Synthesis,
Physical Characterization, and Biofouling Evaluation.
Macromolecules 2009, 42 (18), 6999-7007.
33.Turri, S.; Radice, S.; Canteri, R.; Speranza, G.; Anderle,
M., Surface study of perfluoropolyether–urethane cross‐linked
polymers. Surf. Interface Anal. 2000, 29 (12), 873-886.
34.Wang, Y.; Betts, D. E.; Finlay, J. A.; Brewer, L.; Callow, M.
E.; Callow, J. A.; Wendt, D. E.; DeSimone, J. M., Photocurable
Amphiphilic Perfluoropolyether/Poly(ethylene glycol) Networks for
Fouling-Release Coatings. Macromolecules 2011, 44 (4), 878-885.
35.Hu, Z.; Chen, L.; Betts, D. E.; Pandya, A.; Hillmyer, M. A.;
DeSimone, J. M., Optically Transparent, Amphiphilic Networks Based
on Blends of Perfluoropolyethers and Poly(ethylene glycol). J. Am.
Chem. Soc. 2008, 130 (43), 14244-14252.
36.Weiss, M.; Frohnmayer, J. P.; Benk, L. T.; Haller, B.;
Janiesch, J.-W.; Heitkamp, T.; Börsch, M.; Lira, R. B.; Dimova, R.;
Lipowsky, R.; Bodenschatz, E.; Baret, J.-C.; Vidakovic-Koch, T.;
Sundmacher, K.; Platzman, I.; Spatz, J. P., Sequential bottom-up
assembly of mechanically stabilized synthetic cells by
microfluidics. Nat. Mater. 2017, 17, 89.
37.Wong, D. H. C.; Vitale, A.; Devaux, D.; Taylor, A.; Pandya,
A. A.; Hallinan, D. T.; Thelen, J. L.; Mecham, S. J.; Lux, S. F.;
Lapides, A. M.; Resnick, P. R.; Meyer, T. J.; Kostecki, R. M.;
Balsara, N. P.; DeSimone, J. M., Phase Behavior and Electrochemical
Characterization of Blends of Perfluoropolyether, Poly(ethylene
glycol), and a Lithium Salt. Chem. Mater. 2015, 27 (2),
597-603.
38.Williams, S. S.; Retterer, S.; Lopez, R.; Ruiz, R.; Samulski,
E. T.; DeSimone, J. M., High-Resolution PFPE-based Molding
Techniques for Nanofabrication of High-Pattern Density, Sub-20 nm
Features: A Fundamental Materials Approach. Nano Lett. 2010, 10
(4), 1421-1428.
39.Zhou, Z.; Dominey, R. N.; Rolland, J. P.; Maynor, B. W.;
Pandya, A. A.; DeSimone, J. M., Molded, High Surface Area Polymer
Electrolyte Membranes from Cured Liquid Precursors. J. Am. Chem.
Soc. 2006, 128 (39), 12963-12972.
40.Rolland, J. P.; Van Dam, R. M.; Schorzman, D. A.; Quake, S.
R.; DeSimone, J. M., Solvent-Resistant Photocurable “Liquid Teflon”
for Microfluidic Device Fabrication. J. Am. Chem. Soc. 2004, 126
(8), 2322-2323.
41.Devaux, D.; Villaluenga, I.; Bhatt, M.; Shah, D.; Chen, X.
C.; Thelen, J. L.; DeSimone, J. M.; Balsara, N. P., Crosslinked
perfluoropolyether solid electrolytes for lithium ion transport.
Solid State Ionics 2017, 310, 71-80.
42.Olson, K. R.; Wong, D. H. C.; Chintapalli, M.; Timachova, K.;
Janusziewicz, R.; Daniel, W. F. M.; Mecham, S.; Sheiko, S.;
Balsara, N. P.; DeSimone, J. M., Liquid perfluoropolyether
electrolytes with enhanced ionic conductivity for lithium battery
applications. Polymer 2016, 100, 126-133.
43.Li, G.; Wie, J. J.; Nguyen, N. A.; Chung, W. J.; Kim Eui, T.;
Char, K.; Mackay, M. E.; Pyun, J., Synthesis, self‐assembly and
reversible healing of supramolecular perfluoropolyethers. J. Polym.
Sci., Part A: Polym. Chem. 2013, 51 (17), 3598-3606.
44.Denissen, W.; De Baere, I.; Van Paepegem, W.; Leibler, L.;
Winne, J.; Du Prez, F. E., Vinylogous Urea Vitrimers and Their
Application in Fiber Reinforced Composites. Macromolecules 2018, 51
(5), 2054-2064.
45.Denissen, W.; Droesbeke, M.; Nicolay, R.; Leibler, L.; Winne,
J. M.; Du Prez, F. E., Chemical control of the viscoelastic
properties of vinylogous urethane vitrimers. Nat. Commun. 2017, 8,
14857.
46.Denissen, W.; Rivero, G.; Nicolaÿ, R.; Leibler, L.; Winne, J.
M.; Du Prez, F. E., Vinylogous Urethane Vitrimers. Adv. Funct.
Mater. 2015, 25 (16), 2451-2457.
47.Stukenbroeker, T.; Wang, W.; Winne, J. M.; Du Prez, F. E.;
Nicolay, R.; Leibler, L., Polydimethylsiloxane quenchable
vitrimers. Polym. Chem. 2017, 8 (43), 6590-6593.
48.Lopez, G.; Améduri, B.; Habas, J.-P., A
perfluoropolyether-based elastomers library with on-demand
thermorheological features. Eur. Polym. J. 2017, 95, 207-215.
49.Ni, C.; Hu, J., The unique fluorine effects in organic
reactions: recent facts and insights into fluoroalkylations. Chem.
Soc. Rev. 2016, 45 (20), 5441-5454.
50.Ishibashi, J. S. A.; Kalow, J. A., Vitrimeric Silicone
Elastomers Enabled by Dynamic Meldrum’s Acid-Derived Cross-Links.
ACS Macro Lett. 2018, 7 (4), 482-486.
51.Vitale, A.; Quaglio, M.; Cocuzza, M.; Pirri, C. F.;
Bongiovanni, R., Photopolymerization of a perfluoropolyether
oligomer and photolithographic processes for the fabrication of
microfluidic devices. Eur. Polym. J. 2012, 48 (6), 1118-1126.
52.Imbernon, L.; Oikonomou, E. K.; Norvez, S.; Leibler, L.,
Chemically crosslinked yet reprocessable epoxidized natural rubber
via thermo-activated disulfide rearrangements. Polym. Chem. 2015, 6
(23), 4271-4278.
53.Fielding, S. M.; Sollich, P.; Cates, M. E., Aging and
rheology in soft materials. J. Rheol. 2000, 44 (2), 323-369.
54.Kumpfer, J. R.; Rowan, S. J., Thermo-, Photo-, and
Chemo-Responsive Shape-Memory Properties from Photo-Cross-Linked
Metallo-Supramolecular Polymers. J. Am. Chem. Soc. 2011, 133 (32),
12866-12874.
55.M., L. H.; G., E., Methodologies for lifetime predictions of
rubber using Arrhenius and WLF models. Angew. Makromol. Chem. 1998,
261-262 (1), 135-142.
56.Azcune, I.; Odriozola, I., Aromatic disulfide crosslinks in
polymer systems: Self-healing, reprocessability, recyclability and
more. Eur. Polym. J. 2016, 84, 147-160.
For table of content only
27
1.01.52.02.53.03.54.04.55.0
(ppm)
OO
F
F
O
F
F
F
F
O
F
F
OH
n
m
q
OO
F
F
O
F
F
F
F
O
F
F
O
n-1
m
q
O
OO
O
OO
aabbcc
XX
(CD
3
)
2
COH
2
O
X
cbaab+ccab
X
b+cdefggfdea
1
H
O
OOH
hihi
1)2)3)
150 °C
(a)(b)(c)
R
1
NH
2
+
O
OHN
R
2
O
O
NH
2
R
1
HN
R
2
ii) Aprotic Michael-type pathway:
i) Protic iminium-pathway:
R
1
NH
3
O
OHN
R
2
O
ON
R
2
+
H
H
H
2
NR
1
O
O
NH
2
R
1
HN
R
2
E
a
ref
= 73-81 kJ.mol
-1
E
a
ref
= 102-122 kJ.mol
-1
O
O
NHR
1
H
2
N
R
2
slow
fast
slow
fast
O
O
NHR
1
H
2
N
R
2
fast
fast
O
ON
R
1
H
H
2
NR
2
+
O
OHN
R
1
R
2
NH
3
(a)(b)(c)(d)
PFPE-VU
1.00
PFPE-VU
1.05
ii) Michael-type pathway
R
1
NH
2
+
O
OHN
R
2
i) Iminium pathway
O
ON
R
2
H
H
R
1
NH
2
NH
2
(0 mol % excess)
NH
2
(5 mol % excess)
NH
2
(10 mol % excess)
Bulk, 120 °C, 3h
- tBuOH
HO
O
O
FF
O
F
F
O
F
F
F
F
O
F
F
OH
1.75
1.75
6.1
8.8
OO
O
O
O
O
F
F
O
F
F
O
F
F
F
F
O
F
F
O
1.75
1.75
6.1
8.8
OO
OO
2.2 eq
Bulk, 25 °C, 5 min
80 °C, 16 h
- H
2
O
N
H
2
NNH
2
NH
2
O
O
O
F
F
O
F
F
O
F
F
F
F
O
F
F
O
1.75
1.75
6.1
8.8
OHN
ONH
a) Functionalization:
Acetoacetylation
b) Network synthesis
Condensation
NH
2
R
1
+
NH
2
R
2
+
c) Transamination of vinylogous
urethanes
O
OHN
R
2
O
OHN
R
1
PFPE-OH
PFPE-AA
TREN
TBAA
PFPE-VU
