Aromatic dialdehyde-based bisbenzoxazines: The influence of relative position of oxazine rings

Article abstract of DOI:10.1016/j.polymer.2020.123270

Polybenzoxazines are known to have superior thermal stability. Especially, the use of aromatic aldehydes instead of formaldehyde for the synthesis of benzoxazines monomers could improve this stability. However, the increase in aromatic content of the mono

Full text of DOI:10.1016/j.polymer.2020.123270

Polymer xxx (xxxx) xxx
Contents lists available at ScienceDirect
Polymer
journal homepage: http://www.elsevier.com/locate/polymer
Aromatic dialdehyde-based bisbenzoxazines: The influence of relative
position of oxazine rings
a
a
b
a
a,*
´
Romain Tavernier , Lerys Granado , Gabriel Foyer , Ghislain David , Sylvain Caillol
a ICGM, Univ Montpellier, CNRS, ENSCM, Montpellier, France
b ArianeGroup, Rue de Touban, 33185, Le Haillan, France
A R T I C L E I N F O
A B S T R A C T
Keywords:
Polybenzoxazines are known to have superior thermal stability. Especially, the use of aromatic aldehydes instead
of formaldehyde for the synthesis of benzoxazines monomers could improve this stability. However, the increase
in aromatic content of the monomer can prevent the processability. The objective of this article is to understand
the relationships between the conformation of an aromatic dialdehyde and its thermal properties. Two ben-
zoxazine monomers were synthesized, based on salicylaldehyde, furfurylamine and respectively tereph-
thalaldehyde (TPA, para configuration) and isophthalaldehyde (IPA, meta). Structural characterizations showed
differences in intramolecular interactions between the two isomers. The melting transition of “meta” isomer was
50 ◦C lower than for “para” monomer, whereas polymerization temperatures and enthalpies were nearly the
same. Volatile compounds released during the polymerization were the same in both cases, as investigated by
mass spectrometry (Py-GC/MS). Curing kinetics using model-free kinetic methods revealed that the E-de-
pendency follows the same trend for both monomers. Finally, thermal degradation monitored by thermogravi-
metric analysis under inert atmosphere was similar, with high degradation temperatures and high char yields
(64%).
Polybenzoxazines
Isoconversional analysis
High char yield
High performance thermoset
1. Introduction
prime importance regarding the polymerization behavior. The poly-
merization temperature tends to decrease with stronger
Benzoxazines (BZX) are an interesting class of phenolic thermoset,
which has gained increased attention during the last thirty years [1].
Their outstanding properties such as high thermal stability, initiator-free
thermally activated polymerization and curing, and their relatively easy
syntheses are some of the reasons that make BZX interesting from both
academic and industrial communities. Among these properties of in-
terest, the char formation at high temperature is desirable in the
manufacturing of high performance thermosets, used in aerospace in-
dustry [2]. The particular structure of polymerizable benzoxazines is the
arrangement of the oxazine ring relatively to the phenol, which confers
the ability to polymerize. Only 1,3-benzoxazines undergo a specific
ring-opening polymerization (ROP), due to the particular conforma-
tional structure of the heterocycle. Many studies have been conducted in
order to understand the relationships between monomer design and
subsequent properties of thermosets [3–6].
electron-withdrawing groups (EWG) in para position relatively to
phenol, whereas opposite trend is obtained with EWG groups in para
position to the amine. In contrast, electron-donating groups showed only
very slight effects. This study showed the importance of electronic
environment regarding the mechanistic properties of benzoxazine
polymerization.
More recently, Kolanadiyil et al. investigated the role of the number
of oxazine moieties per monomer in the ring-opening polymerization
behavior [7]. Hence, higher functionality was shown to lower poly-
merization temperatures and to enhance the thermal stability. It also
reduced the volatilization of benzoxazine fragments during the poly-
merization. However, the increase of the number of oxazine rings also
required longer curing times (or higher temperature). Thus, it has been
suggested that the functionality dictates the curing behavior and final
properties. Usually, bisbenzoxazines – comprising two oxazine rings per
monomer, are used to ensure complete crosslinking of polymers, since
monobenzoxazines are known to yield only low molar mass polymers.
This effect has been attributed to a terminating effect of hydrogen
For instance, Andreu et al. studied the effect of different substituents
on the ring-opening polymerization of monobenzoxazines [3]. They
showed that electronic effect of substituents on phenol and amine are of
* Corresponding author.
E-mail addresses: sylvain.caillol@enscm.fr, Sylvain.caillol@enscm.fr (S. Caillol).
https://doi.org/10.1016/j.polymer.2020.123270
Received 29 July 2020; Received in revised form 12 November 2020; Accepted 27 November 2020
Available online 30 November 2020
0032-3861/© 2020 Elsevier Ltd. All rights reserved.
Please cite this article as: Romain Tavernier, Polymer, https://doi.org/10.1016/j.polymer.2020.123270

R. Tavernier et al.
Polymer xxx (xxxx) xxx
bonding [8].
Deuterated chloroform was purchased from Eurisotop. All solvents and
reagents were used without further purification.
In another research article, Kolanadiyil et al. studied the influence of
relative position of the two oxazine rings on polymerization of bisben-
zoxazines [9]. They synthesized three bisbenzoxazines with different
oxazine-ring relative positions, i.e. one bisbenzoxazine having another
oxazine ring attached in ortho position to the phenol moiety of the first
benzoxazine, and two others with para and meta-substitution. They
showed that ROP temperature was influenced by these relative posi-
tions. Firstly, intermolecular interactions were related to configuration,
influencing hydrogen bonds. Secondly, intramolecular interactions were
also affected, with the electronic effects of the attached rings impacting
the opening of the other rings. The meta position seemed to be the most
affected by the relative position of the rings, because it was supposed to
have a catalytic effect both by the intramolecular proximity of the pol-
ymerizable moieties and by intermolecular assistance of hydrogen
bonding which has been shown for some other benzoxazines. This strong
reactivity of meta structural configuration has also been confirmed with
resorcinol-based and meta-phenylenediamine-based benzoxazines
[10–12].
2.2. Synthesis of furfurylaminomethylphenol (Scheme 1, FA)
In a round-bottom flask, salicylaldehyde (20.0 g, 164 mmol) was
dissolved in 100 mL of methanol and then furfurylamine (15.9 g, 164
mmol) was added to the solution. The mixture was heated in a ther-
mostated oil bath to reflux for 1.5 h. After cooling to room temperature,
solvent was removed under reduced pressure in order to isolate furfur-
yliminomethylphenol (FI) and for characterization purpose. Then, FI
was dissolved in 75 mL of methanol and 6.20 g (164 mmol) of NaBH4
were added in small portions, starting at 0 ◦C. After complete addition of
NaBH4, the mixture has been heated to reflux for 2 h. In order to quench
NaBH4 residues, the mixture has been precipitated against 600 mL of
distilled water. A viscous oil was formed in suspension in water, and it
was extracted twice with 250 mL of ethyl acetate, washed with distilled
water, and the organic phase was dried on magnesium sulfate and
concentrated under vacuum. The resulting brown oil was then dried
under vacuum at 40 ◦C overnight. 29.6 g of the desired product has been
recovered (88% yield).
Bisbenzoxazines based on isomers of bisphenol F have been studied
by Liu and Ishida, and an unexpected effect on glass transition tem-
peratures, Tg, has been observed [13]. The three isomers of this
bisphenol, i.e. ortho-ortho, para-para and ortho-para led to differences
in Tg that appeared unusual. Indeed, usually, “para” configured poly-
mers display higher Tg than “ortho” configured chains. In their work,
ortho-ortho configuration of bisphenol F led to polybenzoxazines with
higher Tg, higher crosslinking density and higher degradation temper-
ature. These findings underlined the absolute necessity of controlling the
molecular design to tune the final thermoset properties.
FI (imine): 1H NMR (CDCl3, 7.26 ppm) δ 8.38 (t, 1H, 3J = 1.1 Hz),
7.40 (dd, 1H, 3J = 1.85 Hz, 4J = 0.9 Hz), 7.35–7.30 (ddd, 1H, 3J = 9.3
Hz, 4J = 7.3 Hz, 5J = 9.3 Hz), 7.25 (1H, dd, 3J = 7.7 Hz, 4J = 1.9 Hz),
6.99–6.97 (ddd, 1H, 3J = 8.4 Hz, 4J = 0.9 Hz, 5J = 0.4 Hz), 6.91–6.87
(m, 1H), 6.36 (dd, 1H, 3J = 3.4 Hz, 4J = 1.9 Hz), 6.28 (m, 1H), 4.76 (s,
2H).
FA (amine): 1H NMR (CDCl3, 7.26 ppm) δ 7.40 (dd, 1H, 3J = 1.9
Hz, 4J = 0.9 Hz), 7.21 (m, 1H), 6.99–6.96 (m, 1H), 6.86 (dd, 3J = 8 Hz,
4J = 1.1 Hz), 6.81–6.77 (td, 1H, 3J = 7.5Hz, 4J = 1,3 Hz), 6.34 (dd, 1H,
3J = 3.2 Hz, 4J = 1.9 Hz), 6.21 (m, 1H), 3.96 (s, 2H), 3.82 (s, 2H).
HRMS (m/z, positive mode, [M + H]+): C12H14NO2; calculated
204.1019, found 204.1027.
Recently, the trend toward the use of safer chemicals led to the
development of new chemistries for thermosets. For example, cyclic
carbonates are replacing isocyanates for polyurethanes [14], biobased
compounds are looked forward to replace bisphenol A and aromatic
amines in polyepoxides [15,16], and new phenols and aldehydes have
shown promising properties as an alternative to formophenolics [17]. In
this context, the use of aromatic aldehydes in substitution of formalde-
hyde provides new possibilities for the synthesis of polybenzoxazines.
Ohashi et al. reported the first formaldehyde-free synthesis of 2-subti-
tuted benzoxazine monomers [18] and in a previous article, we pro-
vided deeper insights into ROP mechanisms of these new benzoxazines
[19]. Our study suggested that 2-substituted benzoxazines exhibit
similar polymerization behavior and analogous network structures than
1,3-2H-benzoxazines. The reactivity of these benzoxazines with higher
substitution on the oxazine ring provided a new way to obtain bifunc-
tional monomers, by the use of dialdehydes instead of diphenols or
diamines.
2.3. Synthesis of benzoxazine monomers
Ph-fa[2,2’]tpa: 4.00
g
(19.7 mmol,
2
eq) of furfur-
ylaminomethylphenol (FA) and 1.31 g (9.80 mmol, 1 eq) of tereph-
thalaldehyde were dissolved in 3 mL of acetone in a glass vial. The
mixture has been stirred with a magnetic stirrer for 1 h. After few mi-
nutes, a white solid precipitated, but the mixture has been kept stirring
at ambient temperature in order to ensure maximum precipitation of the
product. The solid was recovered by filtration and washed with acetone,
affording 4.02 g of the desired product (after overnight vacuum drying
at 40 ◦C). Yield = 82%.
1H NMR (CDCl3, 7.26 ppm) δ 7.64 (d, 4H), 7.40 (s, 2H), 7.18 (td,
2H), 6.98 (dt, 2H), 6.93 (d, 2H), 6.88 (tt, 2H), 6.32 (q, 2H), 6.21 (m, 2H),
5.98 (2 s, 2H), 3.95 (m, 4H), 3.82 (s, 4H).
In the present article, we aim to study the effect of relative position of
the oxazine ring by the use of different phthalic aldehydes that are
drastically less toxic than formaldehyde, i.e. terephthalaldehyde (TPA)
with two aldehydes in para position and isophthalaldehyde (IPA) which
has a meta configuration. We synthesized two bisbenzoxazines using the
same synthetic pathway and studied their polymerization behavior,
through a thermo-kinetic study carried out using infrared spectroscopy,
differential scanning calorimetry and isoconversional computations.
The final thermal performances of the two BZX were compared. Overall,
we provide herein some new insights into the structure-properties re-
lationships of these difunctional benzoxazines.
13C NMR (CDCl3, 77.16 ppm) δ 153.56, 153.53, 152.31, 142.45,
142.44, 138.61, 138.58, 128.00, 127.78, 126.90, 120.88, 119.55,
119.54, 116.66, 110.26, 108.67, 108.65, 90.01, 89.95, 47.36, 47.35,
46.41, 46.35.
HRMS (m/z, positive mode, [M + H]+): C32H29N2O4; calculated
505.2127, found 505.2126.
Ph-fa[2,2’]ipa: 4.00
g
(19.7 mmol,
2
eq) of furfur-
ylaminomethylphenol (FA) and 1.30 g (9.70 mmol, 1 eq) of iso-
phthalaldehyde were dissolved in 3 mL of ethanol in a glass vial. The
mixture has been stirred with a magnetic stirrer for 1h. After few mi-
nutes, a brown precipitate appears which then redissolved. Another
precipitation occurred during the stirring and the solid did not redis-
solved. After 1 h h stirring, the solid was recovered by filtration and
washed with ethanol. The recovered solid was then recrystallized in
ethanol, and the resulting compound was then dried under vacuum
overnight at 40 ◦C. 3.39 g of product was afforded as a white solid. Yield
= 69%.
2. Experimental
2.1. Materials
Salicylaldehyde, terephthalaldehyde, isophthalaldehyde and furfur-
ylamine were purchased from TCI. Sodium borohydride was purchased
from Sigma-Aldrich. Anhydrous magnesium sulfate, acetone, methanol,
ethyl acetate and absolute ethanol were purchased from VWR.
1H NMR (CDCl3, 7.26 ppm) δ 7.91 (s, 1H), 7.59 (d, 2H), 7.40 (d,
2

R. Tavernier et al.
Polymer xxx (xxxx) xxx
1H), 7.38 (dd, 1H), 7.36 (dd, 1H), 7.17–7.22 (m, 2H), 6.97–7.00 (dd,
2H), 6.87–6.95 (m, 4H), 6.30–6.32 (m, 2H), 6.22 (dd, 1H), 6.18 (dd,
1H), 6.00 (s, 1H), 5.98 (s, 1H), 6.79–4.01 (m, 4H), 3.80 (s, 4H).
13C NMR (CDCl3, 77.16 ppm) δ 153.65, 153.60, 152.45, 152.41,
142.34, 138.94, 138.89, 128.65, 128.63, 127.96, 127.76, 126.65,
125.40, 125.38, 120.87, 120.84, 119.61, 119.60, 116.74, 116.72,
110.28, 110.26, 108.56, 108.53, 90.19, 90.17, 47.56, 47.51, 46.29,
46.12.
Differential Friedman and integral Vyazovkin approaches were used.
The theoretical considerations and computational detail can be found in
our previous articles [17,20,21].
3. Results and discussion
3.1. Monomer syntheses and characterizations
HRMS (m/z, positive mode, [M + H]+): C32H29N2O4; calculated
505.2127, found 505.2123.
Bisbenzoxazines based on aromatic dialdehydes have been synthe-
sized in three steps, according to our previous work. Furfurylamine was
reacted with salicylaldehyde, in order to obtain an imine, which was
then reduced by sodium borohydride. The obtained secondary amine,
furfurylaminomethylphenol (FA) was then reacted with corresponding
aldehyde, as shown in Scheme 2. Respectively 1 equivalent of tereph-
thalaldehyde (TPA, para) or isophthalaldehyde (IPA, meta) was reacted
with 2 equivalents of furfurylaminomethylphenol, to get corresponding
bisbenzoxazines. According to the abbreviations introduced by Ohashi
et al. and our previous contribution, para oriented benzoxazine is Ph-fa
[2,2’]tpa and meta is Ph-fa[2,2’]ipa [18].
2.4. Measurements
NMR spectroscopy has been performed using a Bruker AC 400 NMR
spectrometer. Chemical shift reference was determined using residual
non deuterated solvent.
Fourier transform infrared absorption spectroscopy (FTIR) mea-
surements were performed on a Nicolet 6700 spectrometer (Thermo-
Scientific), with a mercuryꢀ cadmiumꢀ tellurium detector, resolution
was 4 cmꢀ 1 and 32 scans were added for each spectrum.
Reaction occurred for all monomers at ambient temperature, within
a few minutes. Product recovery was easier by choosing the appropriate
solvent. Indeed, we performed the syntheses in a non-solvent of the final
product. Before optimization, reactions had been performed first in
refluxing toluene, and precipitation was observed for Ph-fa[2,2’]tpa
but not for Ph-fa[2,2’]ipa (results not shown). Reaction conditions
have then been adapted in order to get the corresponding monomers in
milder and more convenient conditions. Monomer solubilities were
different, Ph-fa[2,2’]tpa being insoluble in acetone, whereas Ph-fa
[2,2’]ipa was soluble in acetone, toluene and insoluble in ethanol. We
thus selected respectively acetone and ethanol as reaction medium,
since desired products are insoluble in these solvents. Hence, they
precipitated when formed and could be recovered by simple filtration.
Structural characterizations of the monomers have been performed
by 1H and 13C NMR spectroscopies. 1H NMR spectroscopy, displayed in
Fig. 1-a, revealed the presence of diastereomers. Indeed, the signal OCH
(Ar)N (a and a’) is observed as two singlets for both monomers, at 5.98
ppm for Ph-fa[2,2’]tpa and 6.00 ppm for Ph-fa[2,2’]ipa. Aromatic
protons from aryl attached to the two oxazine rings in Ph-fa[2,2’]tpa is
also observed as two singlets, at 7.64 ppm whereas in Ph-fa[2,2’]ipa,
these protons are not equivalent. The signal from aromatic proton
located between the two bonds attaching the oxazine rings (i) is
observed as a singlet at 7.94 ppm and other protons from the aromatic
ring (g and h) are observed respectively at 7.41 (1H, d) and 7.59 (2H, d).
Protons from furan rings are also influenced by the conformation of the
monomer. When oxazine rings are attached in para position, furan ring
protons are equivalent in both diastereomers (d’, e’ and f’), and they are
observed respectively at 6.21, 6.32 and 7.40 ppm. However, when the
rings are bonded in meta position, some furan rings protons have
different chemical shifts. We can explain this observation by the fact that
dangling furan rings may interact with aryl protons, but only in Ph-fa
[2,2’]ipa. Indeed, the closest furan proton to the IPA residue (proton d)
is observed at 6.18 and 6.22 ppm. Its neighboring proton (e) is observed
at 6.30 ppm as a unique (but complex) signal. The furthest furan proton
is also affected and observed as two signals, at 7.37 and 7.38 ppm. The
free rotation of the furan ring due to the methylene bond, can explain
these strong interactions on protons d and f, and weaker one on proton e.
Infrared spectroscopy has also been performed on monomers and
results are shown in Fig. 1-b. According to the literature, Ar–O phenolic
bands are observed at 1217 and 1035 cmꢀ 1 for Ph-fa[2,2’]tpa and at
1230 and 1035 cmꢀ 1 for Ph-fa[2,2’]ipa. Moreover, the specific ben-
zoxazine band is observed as a broad band in Ph-fa[2,2’]tpa at 938
cmꢀ 1, corresponding to the O–C–N bonds. This broad band may be
explained by the existence of diastereomers, as we reported before and
according to the NMR spectroscopy results. For Ph-fa[2,2’]ipa, several
bands are observed in the region of the O–C–N bond vibration. Three
main bands are observed at 951, 926 and 909 cmꢀ 1. This more complex
High Resolution Mass Spectrometry measurements (HRMS) for syn-
thesized compounds were done on a Waters Synapt G2-S High Resolu-
tion Mass Spectrometer. An electrospray ionization (ESI) method was
used.
Differential scanning calorimetry (DSC) were performed on a DSC-3
F200 Maia from Netzsch GmbH. The machine is equipped with an
intracooler module. Nitrogen flow used was 70 mL minꢀ 1. The tem-
perature sensor was calibrated with biphenyl, indium, bismuth, and CsCl
high purity standards at 10 ◦C⋅minꢀ 1. In order to prevent volatile
evaporation, high-pressure stainless-steel pans and lids sealed at 3 N cm
were used.
Dry content was measured after curing at 200 ◦C. Between 90 and
125 mg of sample were cured in aluminum pans and weighed after
curing. The dry content is the weight loss that happened during curing. It
is presented as an average of 4 samples for each monomer.
Thermogravimetric analyses (TGA) were performed with a TGA-3
Libra (Netzsch).Alumina crucibles were used. Samples used were
weighed between 10 and 12 mg, and introduced as monoliths.The
heating rates were 10 ◦C⋅minꢀ 1 from RT to 900 ◦C. All TGA analyses
were performed under 40 mL minꢀ 1 nitrogen flow.
Py-GC/MS setup: Compounds were pyrolyzed with a Pyroprobe 5000
pyrolyzer (CDS Analytical) under a helium flow. The pyrolyzer sample
holder consists in a quartz tube equipped with an electrically heating
platinum filament. Less than 1 mg of sample is placed into the quartz
tube, between two pieces of quartz wool. Pyrolyzis was enabled by a coil
probe. Successive pyrolysis step were done at 200, 300, 400, 500, 600
and 900 ◦C. Each temperature was held for 15 s before gases were drawn
to the GC for 5 min. The pyrolyzer was coupled to a 450-GC gas chro-
matograph (Varian) by means of a transfer line heated at 270 ◦C. The GC
oven initial temperature of 70 ◦C was held for 0.2 min, and then raised to
310 ◦C at 10 ◦C⋅minꢀ 1. GC column was a Varian Vf-5 ms capillary col-
umn (30 m × 0.25 mm) and the carrier gas was helium (1 mL minꢀ 1); a
split ratio was set to 1:50. The MS analyzer used was an ion trap analyzer
(240-MS from Varian) directly coupled to the capillary column. Identi-
fication of the products was achieved using the American National
Institute of Standards and Technology mass spectral library.
2.5. Thermo-kinetics analysis
The isoconversional computations were performed from non-
isothermal DSC data, at different heating rates: β = 5.0, 7.5, 10.0,
12.5 and 15.0 ◦C⋅minꢀ 1. The degree of curing,
α, was calculated from the
time-integration of exothermic peak with respect to the overall curing
enthalpy (
taken to draw the baseline for integration. The activation energy values
were calculated as a function of , using custom-made programs.
α = ΔHt/ΔHTOTAL). A simple straight-line approximation was
α
3

R. Tavernier et al.
Polymer xxx (xxxx) xxx
Scheme 1. Reaction scheme for the synthesis of Furfurylaminomethylphenol.
Scheme 2. Reaction scheme for the synthesis of dialdehyde-based benzoxazines.
behavior is in line with the NMR spectroscopy results, confirming that
the presence of diastereomer has more influence in Ph-fa[2,2’]ipa than
in Ph-fa[2,2’]tpa due to the non-symmetrical character of the mole-
cule. Finally, the sharp peaks at 1584 cmꢀ 1 for Ph-fa[2,2’]tpa and
1582 cmꢀ 1 for Ph-fa[2,2’]ipa are characteristic of the furan ring
[22–25].
symmetrically. The work of Kolanadiyil et al. was based on isomeric
benzoxazines which displayed very different electronic configuration for
each ring in all their 3 monomers, and they consequently observed
different polymerization temperatures [9]. Polymerization enthalpies
are also found rather similar, namely 257 and 278 J gꢀ 1 for Ph-fa[2,2’]
tpa and Ph-fa[2,2’]ipa, respectively. However, a slight difference is
observed in the shape of the polymerization peaks. Ph-fa[2,2’]tpa
displays a large peak, with a small shoulder after the peak maximum,
whereas Ph-fa[2,2’]ipa displays a clear bimodular peak. In addition for
Ph-fa[2,2’]ipa, the maximum heat flux (peak maximum at which the
reaction rate is the fastest) occurs during the second phenomenon, with
a difference of 14 ◦C observed between the two monomers peak maxima.
This bimodular behavior can find an explanation in the fact that each
monomer is a mixture of two diastereomers, which may display slightly
different polymerization behaviors. It is unlikely that each of the two
rings will ring-open owing to different ways. Indeed, in that case, such
this behavior should have been observed by Kolanadiyil et al. with
monomers with higher functionalities [7]. In our last article, we re-
ported polymerization behaviors of several 2-substituted benzoxazines
[19]. One of them displayed two partially separated polymerization
enthalpies, for a symmetric monomer, that was also a mixture of di-
astereomers. Thus, the influence of the presence of diastereomers is
likely to explain our observations. However, the study of diaster-
eomerically pure samples would bring more insights into this phenom-
enon. Overall, DSC study of the monomers led to the conclusion that
3.2. Curing kinetics
Fig. 2 depicts the DSC thermograms of Ph-fa[2,2’]tpa and Ph-fa
[2,2’]ipa, recorded at 10 ◦C⋅minꢀ 1. A sharp endothermic peak is
observed at 164 ◦C for Ph-fa[2,2’]tpa, corresponding to the melting
transition of monomers. Ph-fa[2,2’]ipa behaved differently, with a
melting temperature at 112 ◦C, which can be explained by the effect of
the isomerism. Even after recrystallization, the melting endotherm of
Ph-fa[2,2’]ipa is broad, indicating that the crystalline network is less
ordered than in Ph-fa[2,2’]tpa. The arrangement of different rings in
the monomer could partially prevent the crystallization, thus requiring
less energy to melt the compound. Polymerization occurred for both
monomers, resulting in exothermic peaks in the thermograms, with an
onset at 216 ◦C and maximum at 235 ◦C for Ph-fa[2,2’]tpa and an onset
at 212 ◦C and maximum at 249 ◦C for Ph-fa[2,2’]ipa. These very
similar values were expected, since the configuration of both oxazine
rings are the same within each monomer structure. Indeed, oxygen and
nitrogen atom of each ring are attached to the same aromatic ring,
4

R. Tavernier et al.
Polymer xxx (xxxx) xxx
Fig. 1. - a) NMR 1H spectra and; b) FT-IR spetra of both monomers Ph-fa[2,2’]ipa and Ph-fa[2,2’]tpa.
processing window of only 50 ◦C.
In order to get more insights about the crosslinking mechanisms, a
kinetic analysis was performed, using a model-free kinetic method. We
used non-isothermal differential scanning calorimetry integral and dif-
ferential data, with multiple heating rates. The advanced isoconver-
sional method of Vyazovkin [26] has been used and compared with
results obtained with Friedman’s method [27]. This kinetic analysis is
based on the elucidation of activation energy, as described by Arrhenius
equations. Isoconversional kinetic analysis has proven to be very useful
in the elucidation of curing mechanisms [28]. Fig. 3-a depicts the degree
of curing as a function of temperature, at different heating rates for both
Ph-fa[2,2’]tpa and Ph-fa[2,2’]ipa. We can observe that for all heating
rates, kinetic profiles are parallel. This confirms that generated data can
be used for isoconversional analysis.
Reaction rate as a function of crosslinking degree is shown in Fig. 3-b
for both monomers, at different heating rates. We can observe two shifts
of the reaction rates, occurring for Ph-fa[2,2’]tpa and Ph-fa[2,2’]ipa
at the same conversion, for all heating rates. Three regimes can thus be
ascribed for these polymerizations. The first regime occurred for
α
<
0.25, which is thought to correspond to the initiation of the ROP and
starting of the propagation of reactive species in the network. For 0.25 <
α
< 0.55, we observe a transient regime, with a slower rate, corre-
sponding to the propagation of polymerization. For
α > 0.55, reaction
rate drastically reduces, and this corresponds to a limitation of the
diffusion of reactive species, due to an increase of the viscosity. These
different regimes happen at the same conversion for both monomer.
However, at low conversion, the increase of reaction rate is more
important for Ph-fa[2,2’]tpa, than for Ph-fa[2,2’]ipa. This can be
explained by the less hindered para structure, which facilitates the
efficient collisions of molecule, and thus the diffusion.
Fig. 2. – DSC of Ph-fa[2,2’]ipa and Ph-fa[2,2’]tpa at 10 ◦C⋅minꢀ 1
.
Activation energy, Ea, as a function of crosslinking degree is shown in
Fig. 4. For both monomers, activation energy ranges between 55 and
both monomers had nearly the same polymerization behavior. However,
the advantage of Ph-fa[2,2’]ipa is the larger processing window, with a
favorable difference of nearly 100 ◦C between melting point and poly-
merization onset, compared to Ph-fa[2,2’]tpa that has a theoretical
103 kJ molꢀ 1. At low conversion (
α < 0.25), Ea decreases to reach a
plateau for 0.25 <
α < 0.55, and then it increases toward the end of
5

R. Tavernier et al.
Polymer xxx (xxxx) xxx
Fig. 3. a) Degree of curing as a function of temperature and - b) Reaction rate as a function of degree of curing for Ph-fa[2,2’]ipa and Ph-fa[2,2’]tpa.
reaction. At low conversion, a decrease in the activation energy is
observed with increasing conversion. This could be explained by the fact
that the ring-opening is a high energy barrier to overcome. As the con-
version from unreacted to ring-opened monomer increases, the energy
barrier decreases, represented by the decrease in activation energy. The
plateau for 0.25 <
α < 0.55 with the lowest activation energy for both
monomers, i.e. 55 kJ molꢀ 1 for Ph-fa[2,2’]tpa and 65 kJ molꢀ 1 for Ph-
fa[2,2’]ipa could represent a transient regime. During this regime,
polymerization seems limited neither by diffusion of reactive species nor
gelation. It can be mentioned that for
α < 0,8, the calculated Ea for the
para monomer is lower than for the meta monomer, and it could be due
to a lower steric hindrance for the para position. At the end of the curing
reaction, Ea increases, as generally observed because the reaction is
diffusion-controlled. This is also illustrated by the lower confidence in
calculated values (Fig. S13), especially for Ph-fa[2,2’]ipa that shows an
important lowering of the correlation coefficient. This is however the
only significant difference of reactivity observed between the two con-
figurations of the monomers. The steric hindrance of the meta position
has probably an important impact at high conversion degree, and during
this step, the heating rate may have an influence on the curing mecha-
nism. In addition, the increase in Ea value for the para position could also
be due to a higher crosslink density, as it can be postulated from the
higher degradation temperature seen in TGA and discussed later. In the
literature, Pereira et al. performed a kinetic analysis of 2-substituted
benzoxazines, and they found activation energies between 102 and
109 kJ molꢀ 1 for 2-substituted monomers obtained from benzaldehyde
and valeraldehyde. Our results are thus in accordance with those values
Fig. 4. Activation energy of Ph-fa[2,2’]ipa and Ph-fa[2,2’]tpa as a function of
degree of curing, with both methods.
6

R. Tavernier et al.
Polymer xxx (xxxx) xxx
[29]. Interestingly, because we synthesized bifunctional monomers, we
obtained a mixture of diastereomers, as described in monomer charac-
terizations. Wang et al. showed a different behavior between polymer-
ization of a unique diastereomer of benzoxazine, and a racemic mixture
of both diastereomers [30]. In particular, the diasteromer mixture dis-
played a bimodal polymerization enthalpy, and a decrease in Ea for 0.5
released during polymerization, whose structural analysis is important
for the determination of curing mechanism. Dry contents of Ph-fa[2,2’]
tpa and Ph-fa[2,2’]ipa are respectively 92.4 ± 0.2% and 89.7 ± 0.3%,
which means that smaller degassing occurs compared to formaldehyde-
free benzoxazines in our previous work [19]. Our results are in accor-
dance with Sudo et al. who have shown that the quantity of volatile
release is higher with bulkier amine residues [32].
<
α < 0.7. Activation energy and kinetic parameters of diastereomer
could be different, as illustrated by the different behavior of a single
enantiomer. These results could explain the behavior observed for our
2-substituted benzoxazines; however, we could not be able to find
another occurrence of the comparison of curing behavior of different
diastereomers and enantiomers.
Pyrolysis coupled with gas chromatography and mass spectrometry
has been used in order to monitor volatiles during polymerization. Gas
chromatograms for both monomers are depicted in Fig. 5-a. Identifica-
tion of the volatiles released during curing can be helpful for the un-
derstanding of reaction mechanisms. For both monomers, we observe
one main peak with a retention time of 13.93 min (corresponding MS
spectrum is shown in Fig. 5-b). The resulting mass spectra reveals that
the compound structure is an imine containing two furan groups. This
result strongly suggests that after ring opening of the oxazine ring, as
described in the literature, the dangling furan group from furfurylamine
attacks the electrophilic iminium [23,25,33–35]. The volatilization of
the compound probably occurs from scission by ring opening of the
second monomer and scission from the methylene attached to the
phenol. In our latest work using Py-GC/MS analysis, we were actually
not able to determine if imines were released from a single monomer or
through intermolecular reaction, since our monomers were designed
only from benzylic rings [19]. In this work however, the presence of an
imine bearing two furan groups can only be the product of intermolec-
ular reaction. The major product released during polymerization, as a
result of furan ring attack on the iminium, reveals that this could be the
most favorable reaction (Reaction scheme proposed in Fig. 5-c). The
furan ring is actually a better nucleophile than the phenolic ring, as it
was measured by Gotta and Mayr for Friedel–Crafts alkylations [36].
They calculated that methylfuran is a stronger nucleophile than
O-methylated phenol (anisole). By analogy, before ring opening, one
may consider that the most nucleophilic site is located on the furan. In
addition, prediction of aromatic nucleophilic sites using a simulation
tool (RegioSQM method) [37] for both monomer and iminium in-
termediates, allowed to determine that the most nucleophilic site is on
the furan ring (Figs. S14–S15). These comparisons would be
Our results shows that model-free kinetics are suitable to analyse
formaldehyde-free benzoxazines ROP. Model-free kinetics provide a
better understanding of the polymerization, without requiring any hy-
pothesis of the chemical mechanism. The apparent complexity of curing
processes, revealed by the presence of bimodal signals on DSC thermo-
grams for both monomers does not prevent the two models to give
consistent results [31]. These results are in line with our work on
formaldehyde-free resoles, whose crosslinking mechanisms involved
several chemical reactions [17,20,21].
3.3. ROP mechanism
FT-IR has been performed on cured samples and shows typical
structure of polymerized benzoxazines, as displayed in Fig. S16. Specific
benzoxazine signals disappeared around 900-950 cmꢀ 1, evidencing the
ring-opening polymerization of both monomers. In poly(Ph-fa[2,2’]
tpa) the band at 1035 cmꢀ 1 disappeared while a weak band remained
around 1230 cmꢀ 1, and in poly(Ph-fa[2,2’]ipa), the 1034 cmꢀ 1 band
disappeared and a weak band at 1226 cmꢀ 1 is observed revealing a
typical Mannich-type linkage. This is also confirmed by the presence of a
broad band around 3200 cmꢀ 1, corresponding to free OH in the cured
network. The disappearance of the 1585 cmꢀ 1 band in poly(Ph-fa
[2,2’]tpa) and the 1582 cmꢀ 1 signal in poly(Ph-fa[2,2’]ipa) is
attributed to the crosslinking on the furan rings [22,23,25].
Dry content measurements allow determining if volatiles are
Fig. 5. - a) GC chromatogram of both monomers pyrolyzed at 200 ◦C b) MS spectrum at retention time = 13.93 min and c) Proposed crosslinking mechanism via furan
cycle reaction.
7

R. Tavernier et al.
Polymer xxx (xxxx) xxx
strengthened with a DFT study of furan-containing benzoxazines, in
order to better design monomers for targeted applications and better
understand ring-opening polymerization of benzoxazines, as reported
before [38].
Other compounds were detected which may originate from degra-
dation of part of the network, such as furfurylamine (retention time 5.42
min, 96 Da), methylphenols (8.56 min, 108 Da), diphenol (15.64 min,
186 Da), and an amine with furan and benzyl group (18.26 min, 201
Da). Additionally, some imines are also proposed to correspond to some
detected ions, according to our previous work (8.75 min, 121 Da and
13.31 min, 109 Da). All those compounds (illustrated in Fig. 5-a) are in
accordance with recent work on the volatile release during benzoxazine
polymerization [39]. In overall, these structures represent a small
fraction of volatiles compared to the bisfuranic imine.
3.4. Thermal degradation
Thermal stability of the polybenzoxazines has been investigated by
TGA under nitrogen, shown in Fig. 6. The degradation study performed
under nitrogen allowed the determination of char yield, that is of prime
interest in the field of high performance composites [40]. Both poly
(Ph-fa[2,2’]tpa) and poly(Ph-fa[2,2’]ipa) displayed the same
degradation behavior, with Td10% of 412 ◦C and 401 ◦C respectively, and
char yields at 900 ◦C of 64% for both cured polybenzoxazines. These
very interesting results evidence that both cured networks have prob-
ably the same structure, and thus the same stability. In our previous
work, we reported the first synthesis of aromatic dialdehyde-based
benzoxazine, which displayed 59% of char yield [18]. Compared to
poly(Ph-fa[2,2’]tpa) which differs only by the use of a furan ring on
the amine, instead of a benzyl, it should be noted that the presence of
furan ring enhances the char formation, as reported previously [24,39].
Derivative TGA (DTG), revealed the same degradation pathway, with a
maximum weight loss respectively at 436 ◦C for poly(Ph-fa[2,2’]ipa)
and 447 ◦C for poly(Ph-fa[2,2’]tpa). A slight shoulder was also
observed at 720 ◦C for both polymers, corresponding to a release of the
lighter compounds during the char formation, as suggested by the py-
rolysis results discussed in the last part of this article.
Fig. 6. TGA of cured materials under nitrogen atmosphere.
between the two monomers, and kinetic analyses revealed no significant
difference in activation energies involved (between 55 and 103 kJ
molꢀ 1). The different kinetic regime shifts happened at the same con-
version for both monomers. Finally, the degradation pathway was the
same for both thermosets, with high degradation temperatures (Td10% of
412 ◦C for para monomer and 401 ◦C for meta). The char formation,
evidenced by Py-GC/MS, was evaluated at 64% of the initial weight,
under nitrogen. Overall, these results demonstrated that the selection of
substituent is of prime importance for the control of polymerization
parameters and thermal properties of polybenzoxazines. In addition, the
structural conformation can help tuning the physical properties of the
monomers, allowing a better control of the processability. The elabo-
ration of formaldehyde-free polybenzoxazines with high char yields and
an improved processability could be of great interest in the
manufacturing of ablation composites used in aerospace industry [40].
CRediT authorship contribution statement
The investigation of the degradation by Py-GC/MS, shown in
Figs. S17–S18, revealed the same degradation products from the two
polybenzoxazines. Compounds such as phenol (94 Da), cresols (108 Da),
xylenes (106 Da), toluene (91 Da), and benzene (78 Da) were detected.
Contrary to our previous results, the release of amine or imine com-
pounds after the pyrolysis step at 200 ◦C was low. This explains the
higher degradation temperatures observed on DTG. The late degrada-
tion of the network is due to the higher crosslinking density, thanks to
furan reaction. Aromatic compounds were released from 400 ◦C to
900 ◦C, with lighter aromatics at higher temperature, as we reported
previously [19]. Some furanic compounds were also released, as ex-
pected. The most important degradation occurred at 500 and 600 ◦C for
both materials, in correlation with the TGA traces. These results are in
line with our previous reported work, and correspond to a typical
charring material, and especially polybenzoxazines.
Romain Tavernier: Conceptualization, Investigation, Writing -
´
original draft, Writing - review & editing, Visualization. Lerys Granado:
Software, Formal analysis, Writing - review & editing, Visualization.
Gabriel Foyer: Writing - review & editing, Supervision. Ghislain
David: Writing - review & editing, Supervision. Sylvain Caillol: Writing
- review & editing, Supervision.
Declaration of competing interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
Acknowledgements
´ ´
Authors thank Direction Generale de l’Armement (DGA) for the
4. Conclusion
funding of the Ph.D. and the “Laboratoire de Mesures Physiques” of
University of Montpellier for the ESI-MS measurements.
We demonstrated the relationships between isomerism and thermal
properties of two aromatic dialdehyde-based bisbenzoxazines. NMR and
infrared spectroscopies revealed the structural differences between
monomers, especially some intramolecular interactions. Physical prop-
erties such as melting transition was greatly influenced by the topology
of the monomer. By changing the orientation of benzoxazine rings from
para to meta, the melting transition was lowered by 50 ◦C. However,
polymerization temperatures and enthalpies remained unchanged, i.e.
with polymerization onsets of 216 ◦C for para monomer and 212 ◦C for
meta, and respectively 257 J gꢀ 1 (para) and 278 J gꢀ 1 (meta). The re-
action mechanisms investigated by Py-GC/MS were not different
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.polymer.2020.123270.
8

R. Tavernier et al.
Polymer xxx (xxxx) xxx
[20] L. Granado, R. Tavernier, G. Foyer, G. David, S. Caillol, Comparative curing
References
kinetics study of high char yield formaldehyde- and terephthalaldehyde-phenolic
thermosets, Thermochim. Acta 667 (2018) 42–49, https://doi.org/10.1016/j.
tca.2018.06.013.
[1] B. Kiskan, Y. Yagci, The journey of phenolics from the first spark to advanced
materials, Isr. J. Chem. 60 (2020) 20–32, https://doi.org/10.1002/
ijch.201900086.
[21] L. Granado, R. Tavernier, G. Foyer, G. David, S. Caillol, Catalysis for highly
thermostable phenol-terephthalaldehyde polymer networks, Chem. Eng. J. 379
(2020), 122237, https://doi.org/10.1016/j.cej.2019.122237.
[2] L. Dumas, L. Bonnaud, M. Olivier, M. Poorteman, P. Dubois, Multiscale
benzoxazine composites: the role of pristine CNTs as efficient reinforcing agents for
high-performance applications, Compos. B Eng. 112 (2017) 57–65, https://doi.
org/10.1016/j.compositesb.2016.12.039.
[22] C. Wang, J. Sun, X. Liu, A. Sudo, T. Endo, Synthesis and copolymerization of fully
bio-based benzoxazines from guaiacol, furfurylamine and stearylamine, Green
Chem. 14 (2012) 2799, https://doi.org/10.1039/c2gc35796h.
[3] R. Andreu, J.A. Reina, J.C. Ronda, Studies on the thermal polymerization of
substituted benzoxazine monomers: electronic effects, J. Polym. Sci. Part A Polym.
Chem. 46 (2008) 3353–3366, https://doi.org/10.1002/pola.22677.
[4] X. Ning, H. Ishida, Phenolic materials via ring-opening polymerization: synthesis
and characterization of bisphenol-A based benzoxazines and their polymers,
J. Polym. Sci. Part A Polym. Chem. 32 (1994) 1121–1129, https://doi.org/
10.1002/pola.1994.080320614.
[23] P. Thirukumaran, A. Shakila Parveen, M. Sarojadevi, Synthesis and
copolymerization of fully biobased benzoxazines from renewable resources, ACS
Sustain. Chem. Eng. 2 (2014) 2790–2801, https://doi.org/10.1021/sc500548c.
[24] Y. Lu, M. Li, Y. Zhang, D. Hu, L. Ke, W. Xu, Synthesis and curing kinetics of
benzoxazine containing fluorene and furan groups, Thermochim. Acta 515 (2011)
32–37, https://doi.org/10.1016/j.tca.2010.12.014.
[25] Y.-L. Liu, C.-I. Chou, High performance benzoxazine monomers and polymers
containing furan groups, J. Polym. Sci. Part A Polym. Chem. 43 (2005) 5267–5282,
https://doi.org/10.1002/pola.21023.
[5] N.N. Ghosh, B. Kiskan, Y. Yagci, Polybenzoxazines—new high performance
thermosetting resins: synthesis and properties, Prog. Polym. Sci. 32 (2007)
1344–1391, https://doi.org/10.1016/j.progpolymsci.2007.07.002.
[6] Y. Yagci, B. Kiskan, N.N. Ghosh, Recent advancement on polybenzoxazine-A newly
developed high performance thermoset, J. Polym. Sci. Part A Polym. Chem. 47
(2009) 5565–5576, https://doi.org/10.1002/pola.23597.
[26] S. Vyazovkin, D. Dollimore, Linear and nonlinear procedures in isoconversional
computations of the activation energy of nonisothermal reactions in solids,
J. Chem. Inf. Comput. Sci. 36 (1996) 42–45, https://doi.org/10.1021/ci950062m.
[27] H.L. Friedman, Kinetics of thermal degradation of char-forming plastics from
thermogravimetry. Application to a phenolic plastic, J. Polym. Sci. Part C Polym.
Symp. 6 (1964) 183–195, https://doi.org/10.1002/polc.5070060121.
[28] N. Sbirrazzuoli, Interpretation and physical meaning of kinetic parameters
obtained from isoconversional kinetic analysis of polymers, Polymers 12 (2020)
1280, https://doi.org/10.3390/polym12061280.
[7] S. Nalakathu Kolanadiyil, T. Endo, Toward elucidating the role of number of
oxazine rings and intermediates in the benzoxazine backbone on their thermal
characteristics, Macromolecules 49 (2016) 8466–8478, https://doi.org/10.1021/
acs.macromol.6b01965.
[8] S. Chirachanchai, A. Laobuthee, S. Phongtamrug, Self termination of ring opening
reaction of p -substituted phenol-based benzoxazines: an obstructive effect via
intramolecular hydrogen bond, J. Heterocycl. Chem. 46 (2009) 714–721, https://
doi.org/10.1002/jhet.130.
[29] R.C.S. Pereira, L.R.V. Kotzebue, D. Zampieri, G. Mele, S.E. Mazzetto, D. Lomonaco,
Influence of natural substituents in the polymerization behavior of novel bio-based
benzoxazines, Mater. Today Commun. 21 (2019) 100629, https://doi.org/
10.1016/j.mtcomm.2019.100629.
[9] S.N. Kolanadiyil, M. Minami, T. Endo, Synthesis and thermal properties of
difunctional benzoxazines with attached oxazine ring at the para -, meta -, and ortho
-position, Macromolecules (2017), https://doi.org/10.1021/acs.
macromol.7b00487 acs.macromol.7b00487.
[30] J. Wang, H. Wang, J. Liu, W. Liu, X. Shen, Synthesis, curing kinetics and thermal
properties of novel difunctional chiral and achiral benzoxazines with double chiral
centers, J. Therm. Anal. Calorim. 114 (2013) 1255–1264, https://doi.org/
10.1007/s10973-013-3081-8.
[10] S. Ren, X. Yang, X. Zhao, Y. Zhang, W. Huang, An m -phenylenediamine-based
benzoxazine with favorable processability and its high-performance thermoset,
J. Appl. Polym. Sci. 133 (2016) 43368, https://doi.org/10.1002/app.43368.
[11] A. Arnebold, O. Schorsch, J. Stelten, A. Hartwig, Resorcinol-based benzoxazine
with low polymerization temperature, J. Polym. Sci. Part A Polym. Chem. 52
(2014) 1693–1699, https://doi.org/10.1002/pola.27169.
[31] N. Sbirrazzuoli, Advanced isoconversional kinetic analysis for the elucidation of
complex reaction mechanisms: a new method for the identification of rate-limiting
steps, Molecules 24 (2019) 1683, https://doi.org/10.3390/molecules24091683.
[32] A. Sudo, L.-C. Du, S. Hirayama, T. Endo, Substituent effects of N-alkyl groups on
thermally induced polymerization behavior of 1,3-benzoxazines, J. Polym. Sci. Part
A Polym. Chem. 48 (2010) 2777–2782, https://doi.org/10.1002/pola.24026.
[33] N.K. Sini, J. Bijwe, I.K. Varma, Renewable benzoxazine monomer from Vanillin:
synthesis, characterization, and studies on curing behavior, J. Polym. Sci. Part A
Polym. Chem. 52 (2014) 7–11, https://doi.org/10.1002/pola.26981.
[34] A. Trejo-Machin, A. Adjaoud, L. Puchot, R. Dieden, P. Verge, Elucidating the
thermal and polymerization behaviours of benzoxazines from lignin derivatives,
Eur. Polym. J. 124 (2020), 109468, https://doi.org/10.1016/j.
¨
´
[12] H. Schafer, A. Arnebold, J. Stelten, J. Marquet, R. María Sebastian, A. Hartwig,
K. Koschek, Bifunctional benzoxazines: synthesis and polymerization of resorcinol
based single isomers, J. Polym. Sci. Part A Polym. Chem. 54 (2016) 1243–1251,
https://doi.org/10.1002/pola.27966.
[13] J. Liu, H. Ishida, Anomalous isomeric effect on the properties of bisphenol F-based
benzoxazines: toward the molecular design for higher performance,
Macromolecules 47 (2014) 5682–5690, https://doi.org/10.1021/ma501294y.
´
´
[14] C. Carre, Y. Ecochard, S. Caillol, L. Averous, From the synthesis of biobased cyclic
carbonate to polyhydroxyurethanes: a promising route towards renewable non-
isocyanate polyurethanes, ChemSusChem 12 (2019) 3410–3430, https://doi.org/
10.1002/cssc.201900737.
eurpolymj.2019.109468.
[35] L. Dumas, L. Bonnaud, M. Olivier, M. Poorteman, P. Dubois, High performance bio-
based benzoxazine networks from resorcinol and hydroquinone, Eur. Polym. J. 75
(2016) 486–494, https://doi.org/10.1016/j.eurpolymj.2016.01.021.
[36] M.F. Gotta, H. Mayr, Kinetics of the Friedel-Crafts alkylations of heterocyclic
[15] A.W. Bassett, J.D. Cosgrove, K.M. Schmalbach, O.M. Stecca, C.M. Paquette, V.
H. Adams, W.S. Eck, J.M. Sadler, J.J. La Scala, J.F. Stanzione, Alternative
monomers for 4,4′-methylenedianiline in thermosetting epoxy resins, J. Appl.
Polym. Sci. 137 (2020) 48707, https://doi.org/10.1002/app.48707.
[16] C. Mantzaridis, A.-L. Brocas, A. Llevot, G. Cendejas, R. Auvergne, S. Caillol,
S. Carlotti, H. Cramail, Rosin acid oligomers as precursors of DGEBA-free epoxy
resins, Green Chem. 15 (2013) 3091, https://doi.org/10.1039/c3gc41004h.
[17] L. Granado, R. Tavernier, S. Henry, R.O. Auke, G. Foyer, G. David, S. Caillol,
Toward sustainable phenolic thermosets with high thermal performances, ACS
Sustain. Chem. Eng. 7 (2019) 7209–7217, https://doi.org/10.1021/
acssuschemeng.8b06286.
arenes: comparison of the nucleophilic reactivities of aromatic and nonaromatic π-
systems, J. Org. Chem. 63 (1998) 9769–9775, https://doi.org/10.1021/jo981373.
[37] J.C. Kromann, J.H. Jensen, M. Kruszyk, M. Jessing, M. Jørgensen, Fast and accurate
prediction of the regioselectivity of electrophilic aromatic substitution reactions,
Chem. Sci. 9 (2018) 660–665, https://doi.org/10.1039/C7SC04156J.
¨
˘
[38] T. Furuncuoglu Ozaltın, S. Catak, B. Kiskan, Y. Yagci, V. Aviyente, Rationalizing
the regioselectivity of cationic ring-opening polymerization of benzoxazines, Eur.
Polym. J. 105 (2018) 61–67, https://doi.org/10.1016/j.eurpolymj.2018.05.024.
[39] P. Wang, M. Liu, Q. Ran, The study on curing and weight-loss mechanisms of
benzoxazine during thermal curing process, Polym. Degrad. Stabil. 179 (2020),
109279, https://doi.org/10.1016/j.polymdegradstab.2020.109279.
[18] S. Ohashi, F. Cassidy, S. Huang, K. Chiou, H. Ishida, Synthesis and ring-opening
polymerization of 2-substituted 1,3-benzoxazine: the first observation of the
polymerization of oxazine ring-substituted benzoxazines, Polym. Chem. 7 (2016)
7177–7184, https://doi.org/10.1039/C6PY01686C.
[40] M. Natali, J.M. Kenny, L. Torre, Science and technology of polymeric ablative
materials for thermal protection systems and propulsion devices: a review, Prog.
Mater. Sci. 84 (2016) 192–275, https://doi.org/10.1016/j.pmatsci.2016.08.003.
[19] R. Tavernier, L. Granado, G. Foyer, G. David, S. Caillol, Formaldehyde-free
polybenzoxazines for high performance thermosets, Macromolecules 53 (2020)
2557–2567, https://doi.org/10.1021/acs.macromol.0c00192.
9