Polycyanurate networks from dehydroanethole cyclotrimers: Synthesis and characterization

Article abstract of DOI:10.1016/j.polymer.2013.10.050

Novel biobased trisphenols were obtained by palladium- and cobalt-catalyzed [2 + 2 + 2] cycloaddition of dehydroanethole isomers which were readily prepared from the natural product trans-anethole. The trisphenols were transformed into their corresponding

Full text of DOI:10.1016/j.polymer.2013.10.050

Accepted Manuscript
Polycyanurate Networks from Dehydroanethole Cyclotrimers: Synthesis and
Characterization
Matthew C. Davis, Andrew J. Guenthner, Christopher M. Sahagun, Kevin R. Lamison,
Josiah T. Reams, Joseph M. Mabry
PII:
S0032-3861(13)01019-7
10.1016/j.polymer.2013.10.050
JPOL 16581
DOI:
Reference:
To appear in: Polymer
Received Date: 5 September 2013
Revised Date: 24 October 2013
Accepted Date: 27 October 2013
Please cite this article as: Davis MC, Guenthner AJ, Sahagun CM, Lamison KR, Reams JT, Mabry JM,
Polycyanurate Networks from Dehydroanethole Cyclotrimers: Synthesis and Characterization, Polymer
(2013), doi: 10.1016/j.polymer.2013.10.050.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Biobased
Polycyanurate
networks
Star anise
Trianethole
tricyanate

ACCEPTED MANUSCRIPT
Polycyanurate Networks from Dehydroanethole Cyclotrimers:
Synthesis and Characterization
a,*
b,*
c
d
Matthew C. Davis, Andrew J. Guenthner, Christopher M. Sahagun, Kevin R. Lamison,
d
b
Josiah T. Reams, and Joseph M. Mabry
a
Chemistry & Materials Division, Michelson Laboratory, Naval Air Warfare Center, China
Lake, California, 93555, USA
b
Air Force Research Laboratory, Aerospace Systems Directorate Edwards AFB, California,
9
3524, USA
c
National Research Council / Air Force Research Laboratory, Aerospace Systems Directorate,
Edwards AFB, California, 93524, USA
d
ERC Incorporated, Edwards AFB, California, 93524,USA
Abstract
Novel biobased trisphenols were obtained by palladium- and cobalt-catalyzed [2+2+2]
cycloaddition of dehydroanethole isomers which were readily prepared from the natural product
trans-anethole. The trisphenols were transformed into their corresponding tricyanate esters and
thermally cured to give polycyanurate networks. Comparison of the thermal properties, from
differential scanning calorimetry, thermogravimetric and oscillatory thermomechanical analyses,
of the new tricyanate esters with similar commercial products is presented. A new tricyanate
ester with an acceptable processing temperature that yields a polycyanurate with very high glass
transition temperature and low water absorption was found. Interestingly, the low moisture
absorption was found to help limit the degradation of residual cyanate ester groups during
*
Corresponding authors. Tel.: 01 7609390196; E-mail address: matthew.davis@navy.mil (M.C. Davis);
Tel.: 01 6612758020; E-mail address: andrew.guenthner@us.af.mil (A.J. Guenthner).

ACCEPTED MANUSCRIPT
exposure to hot water, allowing for a substantial recovery in the glass transition temperature
upon subsequent heating, an unusual phenomenon for polycyanurate networks.
Keywords: high temperature materials, networks, cyanate ester, polycyanurate, renewable
resources, thermosets, thermal properties, cycloaddition, alkynes
1
. Introduction
The enormous variety and quantity of plastics that make up the consumer goods on the
market today are almost entirely composed of petrochemicals or chemicals derived from fossil
based sources. Predictions are continually changing about when this invaluable natural resource,
crude oil, will expire [1-4]. Of course, there is already inherent competition for the use of crude
oil as a chemical feedstock [5], because of the need for liquid transportation fuels. In addition, it
is clear that there will be increasing demands on this resource as economic growth and
industrialization in the developing regions of the world continues at a dramatic pace [6,7]. The
theoretically best solution to achieve a secure supply of chemical feedstock without the need for
crude oil would be to obtain chemical building blocks from biomass [8-13]. Although this
concept can present new complications and concerns, both economic and environmental in
nature (e. g., arable land, water, fertilizer, climate, etc.), these concerns are far beyond the scope
of the present theme. Rather, we focus on the fact that, as Dijkstra and Langstein [14] recently
pointed out, for engineering plastics, namely those constructed from substituted benzene (e.g.
polyethylene terephthalate), benzene-toluene-xylene (BTX) chemicals are vital [15]. In a cursory
glance at the top chemicals derived from biomass, however, one sees a void in this category [16-
1
9]. A renewable and direct source of aromatic-type chemical feedstock will address this
important niche for the polymer community

ACCEPTED MANUSCRIPT
Among possible biomass sources for BTX chemicals, a plant that appears interesting
from a chemical standpoint is star anise (Illicium verum), whose essential oil is collected
annually on a scale of 2,200 metric tons [20]. Found primarily in Guangxi province China
(hardiness zone 10) [21], the plant is not a food source and only the leaves, fruits and flowers are
harvested for the distillation process. The essential oil is routinely assayed at 90% or better
trans-anethole (1-(4-methoxyphenyl)propene) and [22], because of its bifunctional nature, is a
suitable ‘phytofeedstock’ for materials research in the engineering plastics realm. In searching
for high-performance cyanate esters (aryl–O–C≡N) starting from renewable, chemical building
blocks, we recently described results from diphenols obtained through dimerization of trans-
anethole [23]. These dicyanates thermally polymerized to polycyanurate networks (Fig. 1) with
high glass transition temperatures, excellent water resistance, and simple processing.
Drawing inspiration from the work of Lligadas et al. [24] with vegetable oils, this report
describes the derivitization of trans-anethole to its dehydro or carbon–carbon triple bond analog
(dehydroanethole), which by way of intermolecular [2+2+2] cycloaddition gave triphenols. The
trisphenols were converted into tricyanate esters and readily cured into polycyanurate networks
with low moisture uptake and glass transition temperatures of up to 362 °C, thus demonstrating
that the improved thermo-mechanical performance of networks based on tricyanate esters (in
comparison to those based on dicyanate esters) is available while preserving the low moisture
uptake associated with anethole-based monomers. Also demonstrated is the somewhat surprising
but very important discovery that in these uncatalyzed cyanurate networks with a high glass
transition temperature and low moisture uptake, the degradation in performance due to exposure
to moisture is significantly reversible.

ACCEPTED MANUSCRIPT
Figure 1. General reaction for polycyanurate network formation by heating dicyanate ester.
2
. Experimental
2
.1. Materials
The CoI (95%), ZnBr (98%), Zn dust (98%) cyanogen bromide (BrCN), pyridine, 5 %
2
2
palladium on carbon (Pd/C), and phosphorus oxychloride (POCl ) were purchased from Sigma-
3
Aldrich (Milwaukee). Trans-anethole (> 98 %) was purchased from TCI America (Portland).
All other reagents and solvents were obtained commercially and used as received.
2
.2. Synthesis of monomers 5 and 9
For complete experimental details of the synthesis of tricyanate monomers 5 and 9 refer
to the Supplementary Data.
2
.3. Thermal curing of monomers
Prior to cyclotrimerization, monomers were partially de-gassed along with a silicone
mold for 30 minutes under reduced pressure (300 mm Hg) at a temperature of 10 – 20 °C above
the melting point. If the melting point exceeded 150 °C, then only the silicone mold was de-
gassed under 300 mm Hg at 150 °C for 30 minutes, and the monomer was melted as quickly as
possible at 10 – 20 °C above the melting point. Each monomer was then poured into the mold
and cured under flowing nitrogen for 1 hour at 150 °C followed by 24 hours at 210 °C to produce

ACCEPTED MANUSCRIPT
void-free discs measuring approximately 13.7 mm in diameter and 1.6 mm thick and weighing
2
00-300 mg. The temperature ramp rate during cure was 5 °C/min. For monomers with melting
points above 150 °C, the initial 1 hour hold at 150 °C was skipped. Note that no catalyst was
added to any of the monomers. Cured discs were cut with a Struers Somotom-10 saw using a
diamond-coated non-ferrous disc to produce surfaces with appropriate, reproducible roughness
characteristics.
2
.4. Conditioning of polycyanurate networks
For water uptake tests, some cured sample discs were placed in approximately 300 mL of
deionized water at 85 °C for 96 hours. For weighing, samples were removed and their surfaces
dried. Mass was measured before and after exposure.
2
2
.5. Measurement and characterization
.5.1. General methods
The melting points were collected on a Mel-Temp II from Laboratory Devices (Holliston,
MA) and are not corrected. All NMR data were collected on a Bruker Avance II 300 MHz
spectrometer. Nuclear magnetic resonance data (free-induction decay’s) were processed using
NUTS software from Acorn NMR (Livermore, CA). All spectra are referenced to solvent or
tetramethylsilane. Elemental analyses were performed by Atlantic Microlab, Inc. (Norcross,
GA).
2
.5.2. Fourier transform infrared analyses
Fourier transform infrared spectroscopy (FT-IR) was carried out using a Thermo
Corporation Nicolet 6700 spectrometer in attenuated total reflectance mode with the ZnSe crystal

ACCEPTED MANUSCRIPT
-1
attachment. A total of 512 scans were completed on these surfaces with a resolution of 4 cm to
-1
-1
obtain spectra. The integrated area between 1480 cm and 1510 cm , containing the aromatic
-
1
ring peak near 1500 cm , was used for normalization, with a quadratic baseline determined by
-1
-1
-1
using 1810 cm , 2100 cm , and 2500 cm as zero points.
2
.5.3. Density measurement
The density of the cured resins was determined by placing cured discs in solutions of
CaCl (as the dihydrate) and deionized water and varying the CaCl concentration until neutral
2
2
buoyancy was observed on bubble-free samples over a period of several minutes. The density of
the neutrally buoyant solution was determined by placing 10.00 mL in a volumetric flask
(calibrated with deionized water at 20 °C) and weighing, and checked against the predicted
density of the solution at ambient temperature based on the known concentration of CaCl2.
2
2
.5.4. Thermal analysis
.5.4.1. Differential scanning calorimetry
Differential scanning calorimetry (DSC) was performed on ~10 mg of monomer reserved
after de-gassing using a TA Instruments Q2000 calorimeter under 50 mL/min. of flowing
nitrogen. Samples were heated to 350 °C, then cooled to 100 °C and re-heated to 350 °C, all at
1
0 °C/min. The same protocol was also carried out on ~10 mg fragments of cured discs.
.5.4.2. Oscillatory thermomechanical analyses
2
Cured discs were also tested via oscillatory thermomechanical analysis (OTMA) with a
TA Instruments Q400 series analyzer under 50 mL/min of nitrogen flow. The discs were
initially held in place with a compressive force of 0.2 N using the standard ~5 mm diameter flat

ACCEPTED MANUSCRIPT
cylindrical probe. The force was then modulated at 0.05 Hz over an amplitude of 0.1 N (with a
mean force of 0.1 N) and the temperature was ramped twice (heating and cooling) between 0 °C
and 200 °C (to aid in determination of thermal lag) followed by heating to 350 °C, cooling to 100
°
C, and re-heating to 350 °C, all at 50 °C/min. For samples previously exposed to hot water, the
heating rate was decreased to 20 °C/min and the order of segments were: heating to 350 °C,
cooling to 100 °C, two cycles between 100 °C and 200 °C for thermal lag determination, and
finally heating to 350 °C. The details of the thermal lag determination procedure have been
reported elsewhere [25].
2
.5.4.3. Thermogravimetric analyses
Thermogravimetric analysis (under both nitrogen and air) was performed on ~10 mg
samples of cured discs, using a TA Instruments Q5000 analyzer, at a gas flow rate of 10 mL/min
balance) and 25 mL/min (purge), heating the samples at 10 °C/min to 600 °C.
(
3
. Results and discussion
3
.1. Synthesis of monomers
For expediency, the two-step process of bromination/dehydrobromination used by Gramatica et
al. [26] was chosen as the method to effect oxidation of trans-anethole (1) to dehydroanethole (2)
(Scheme 1) [27,28]. Typical conditions for bromination of 1 use an unreactive solvent such as
chlorinated hydrocarbons that give a diastereomeric mixture of vicinal dibromides [29-32].
Tetrahydrofuran (THF) has rarely been employed as a solvent for bromination of alkenes [33],
most likely due to the possibility of forming a charge transfer complex like that of the stable
solid Br ·1,4-dioxane [34]. The bromination of 1 was found to work well using THF as solvent
2
1
at 0 °C requiring only a slight excess of bromine to achieve completion based on H NMR

ACCEPTED MANUSCRIPT
analysis of the reaction mixture. Therefore, rather than isolate the mixture of vicinal dibromide
isomers, the reaction mixture was added to a solution of potassium tert-butoxide in THF which
smoothly brought about double dehydrobromination to give the alkyne 2 [35-38]. This
combination of reactions saved time yet did not effect much loss in overall yield (92%).
Scheme 1. Synthetic route for tricyanate esters 5 and 9 from trans-anethole (1).
Since the discovery by Reppe [39,40] that soluble transition metal complexes can catalyze
the intermolecular [2+2+2] cycloaddition of alkynes, there has been a steady increase in this
transformation over the past 50 years [41-47]. With dehydroanethole (2) in hand, the conditions

ACCEPTED MANUSCRIPT
for cyclotrimerization of Jianghan and Maier (Pd/C, trimethylsilyl chloride) [48] were carried out
though using 1,4-dioxane at reflux to speed up the reaction. The product was a mixture of the
1
,2,4-regioisomer 3 and its symmetrical isomer in a 3:1 ratio, respectively [49]. Fortunately, 3
could be separated by fraction crystallization. Demethylation of 3 was accomplished by
refluxing in a previously prepared mixture of pyridine, water, and phosphorus oxychloride (3:3:1
ratio respectively). The resulting triphenol 4 was reacted under typical cyanation conditions
(cyanogen bromide, triethylamine, acetone, - 20 °C) [50] to give the tricyanate ester 5 in an
overall yield of 14%. Tricyanate 5 had a rather high melting point (T ) of 165 °C (329 °F). The
m
symmetrical isomer from this cycloaddition reaction was anticipated to make a tricyanate ester
with an even higher melting point making curing and processing difficult (vide infra) and,
therefore, was not pursued [51].
Ideally a liquid tricyanate ester was desired but one with a melting point less than 100 °C
would be satisfactory and so further chemistry was explored. It is well known that 1-
arylpropynes can be isomerized by strong base resulting in 3-arylpropynes, an interesting
transformation proceeding through an allene intermediate [52-57]. Although the reaction was
slow, compound 2 was efficiently isomerized to 6 following the conditions of Mulvaney et
al.[58] though it was found critical to keep the reaction mixture cold during the water work-up
for best yield (Scheme 1). The [2+2+2] cycloaddition of the parent 3-phenylpropyne had been
studied previously by Choi et al. [59] and Rüba et al. [60] using metal catalyst ZrCl /Li and
4
[
RuCp(CH CN) ]PF , respectively, each leading to ~3:1 mixtures favoring the unsymmetrical
3 3 6
tribenzylbenzene isomer. So it was anticipated that cycloaddition would be successful with 6, at
least with those catalysts, but it was decided to explore the convenient, “instant” cobalt catalyst
system developed by Hilt’s group [61-65]. Reaction of 6 with CoI /Zn/ZnBr in acetonitrile,
2
2

ACCEPTED MANUSCRIPT
without any triarylphosphine or 1,2-diimine ligand, gave a crude product mixture of cyclotrimer
7
and symmetrical isomer in a 19:1 ratio, respectively, based on integration of the CH peaks in
2
1
3
C NMR. Owing to this remarkably selective reaction, product 7 was purified by
recrystallization in a yield of 55%. Indeed, the reaction is also interesting for the fact that the
solvent, with an sp-hybridized bond, did not participate in the cycloaddition reaction which has
been demonstrated with cobalt and other catalysts [66-68]. Tricyanate ester 9 was then made
from 7 in 15% overall yield by a similar three-step sequence of reactions mentioned above in the
synthesis of 5, in this case using commercial pyridine hydrochloride. Tricyanate 9 had a T of
m
8
5 °C (185 °F) and so too were intermediates 7 and 8 solids, but despite significant effort,
appropriate solvent/conditions for X-ray quality crystals of these compounds, to provide further
1
13
support for these structures, was not found. The H and C NMR spectra of 9 clearly show the
unsymmetrical nature of the molecule, for example the individual methylene peaks and complex
aromatic region (Fig. S1, Supplementary Data). Unfortunately, without a more powerful NMR
1
3
instrument, the three individual cyanate resonances of 9 overlap in the C spectrum.
3
.2. Synthesis of polycyanurates and characterization
Analysis of monomer 5 by differential scanning calorimetry (DSC) (Fig. 2) showed the
main polymerization exotherm began shortly after melting.
Unfortunately, potential
thermosetting monomers that do not provide an adequate temperature range between the melting
point and the onset of rapid gelation, the so-called processing window (vide infra), are not useful
for molding operations [69,70]. Therefore, monomer 9 with its greater degree of conformational
freedom and lower melting point, became the focus of our attention. Analysis of monomer 9 by
DSC (Fig. 3) shows good baseline separation between melting and the main polymerization (or
cure) exotherm peak commencing at ~ 200 °C, which is typical of cyanate esters. The total heat

ACCEPTED MANUSCRIPT
output of the exothermic event was 620 J/g alternately expressed as 97 kJ/cyanate ester, a more
convenient value with which to compare different cyanate esters (Table 1). The 97 kJ/cyanate
ester is a typical value seen in other cyanate esters and represents near complete cure.
Figure 2. DSC trace of monomer 5.

ACCEPTED MANUSCRIPT
3
2
1
0
1
st heating
2nd Heating
-1
-2
-3
-4
-5
50
100
150
200
250
300
350
Temperature (°C)
Figure 3. DSC trace of monomer 9.
A mild cure schedule of 150 °C for 1 hour followed by 210 °C for 24 hours was used to
prepare sample discs for further testing. By this heating protocol, curing of 9 produced a high
conversion, as evidenced by Fourier transform infrared (FTIR) analysis which showed some
-1
-1
residual cyanate stretch at 2250 cm and strong peaks at 1360 and 1570 cm corresponding to
triazine (Fig. S2, Supplementary Data) [71]. A DSC scan of a small chip from the cured disc
(Fig. 4) indicates a residual cure amounting to 19 ± 2 kJ/eq, with the onset of residual cure
preceded by what appears to be a glass transition temperature (T ) of around 235 °C. Since the
g
Tg signals for polycyanurate networks can sometimes be very broad [72] while the signal in Fig.
4
appears narrow and truncated, a confirmatory indication of the “as cured” T was sought. An
g
oscillatory thermomechanical analysis (TMA) trace (Fig. S3, Supplementary Data) showed that
under this cure schedule polycyanurate 9 exhibited a T of 277 °C (tan δ peak). Despite the use
g
of a very rapid heating rate (50 °C / min.), this higher T likely reflects the effects of some in-situ
g
cure during the analysis experiment. After continued heating to 380 °C to achieve full cure, the

ACCEPTED MANUSCRIPT
oscillatory TMA experiments clearly show T of 362 °C (DSC scans at these temperatures are
g
significantly affected by slight decomposition of the network, whereas signals from mechanical
methods are minimally affected).
Figure 4. DSC of 9, after curing at 210 °C for 24 hours.
The determination of the T at full cure provides the key parameter needed for use of the
g
diBenedetto equation [73,74] to relate the T to the extent of cure in polycyanurate networks. As
g
described in the Supplementary Data, data for the diBenedetto equation was generated using
partially cured samples prepared in the DSC, and a good fit with a typical value for the parameter
λ of 0.44 ± 0.06 was obtained. The diBenedetto equation predicts that a polycyanurate network
cured from 9 with a residual exotherm of 19 kJ/eq. will exhibit a T of 240 ± 7 °C, corresponding
g
to the location of the probable T seen in the Fig. 4.
g
Reduction in mechanical and physical properties (e.g. reduction of T , swelling) of
g
polycyanurate networks can be linked to absorption of moisture and plasticization and/or
chemical alteration of the polymer matrix [75,76]. To examine this effect, a sample of 9 cured
for 24 hours at 210 °C was subjected to hygrothermal ageing by exposure to hot water (85 °C).

ACCEPTED MANUSCRIPT
After 96 hours the sample had water uptake of only 1.4 wt%. This water challenged sample was
analyzed again by oscillatory TMA (Fig. 5) which showed two peaks at 187 °C and 303 °C in the
loss modulus curve.
60000
50000
40000
30000
20000
10000
0
0.5
Storage
Loss
tan delta
0
0
.45
.4
0.35
.3
0.25
0
0
0
0
0
.2
.15
.1
.05
0
100
200
300
Temperature (°C)
Figure 5. TMA of cured 9 after exposure to 85 °C water for 96 hours.
In order to better understand the behavior observed in Fig. 5, a similarly-cured sample of
9
(conversion 87%, dry “as cured” T 270 °C) was tracked via FT-IR, DSC, TMA, and weight
g
gain measurements during the course of exposure to 85 °C water for 96 hours followed by
heating to 350 °C in the TMA. The heating run was interrupted when the temperature set point
reached 200 °C, the sample was weighed, and a portion of the sample was removed for DSC and
FT-IR testing; the remainder was then returned to the TMA and the heating run continued. In
this manner, the effects of water exposure, heating below the first “wet” T , and subsequent
g
heating could be determined separately.
Figure 6 shows the FT-IR traces of the sample “as cured”, exposed to water (but not
subsequently heated), and exposed with re-heating to 200 °C and 350 °C. Additional views
-1
-1
showing the key peaks for cyanurate (1360 and 1555 cm ), cyanate ester (2200-2300 cm ),
-1
-1
carbamate (1720 cm ), and phenol (1600 cm ) are available in Supplementary Data (Figs. S6

ACCEPTED MANUSCRIPT
and S7). Integration of the relevant peaks revealed that around 35% of the cyanate ester groups
(or 5 mol% of the cyanate ester groups originally present in the monomer) disappeared on
exposure to water, with the appearance of a significant carbamate peak and little change in the
cyanurate peaks. The simplest explanation is that residual cyanate ester groups are converted to
carbamate on exposure to water, that heating to 200 °C does not produce further carbamate
groups due to insufficient available time, and that carbamate groups decompose after heating to
2
00 °C. This explanation matches the generally accepted picture of the formation and
destruction of carbamate groups in cyanurate networks [77-81].
Figure 6. FT-IR scans of cured 9 at various points before, during, and after exposure to 85 °C
water and subsequent heating.
Moreover, the weight change data indicates a gain of 2.4% on exposure to water. (This
value is higher than the sample with 80% conversion due to the strong increase in water uptake
with increasing conversion in cyanurate networks, combined with the effect of curing well below
T (increases water uptake) as opposed to near T .) The weight then decreases by 0.4% (original
g
g
weight basis) on heating to 200 °C, presumably due to some loss of water of plasticization,
followed by a further decrease of 2.4% (net mass loss of 0.4% compared to original sample
mass) on heating to 350 °C. These data are consistent with reversible water uptake, formation of

ACCEPTED MANUSCRIPT
roughly 5 mol carbamate per 100 mol of uncured cyanate ester on exposure to water (as
indicated by FT-IR), subsequent decomposition of carbamate with release of CO when heated
2
above 200 °C, and little or no permanent weight gain due to hydrolysis of cyanurate ring
structures.
Additional details may be inferred by examining the corresponding DSC traces (Fig. 7).
As in Fig. 4, both the extent and onset temperature of residual cure in the “as cured” sample
match the expected behavior for a network with a T of 270 °C. After exposure to water and
g
heating to 200 °C, the exotherm shifts to lower temperatures (an indication of a significant drop
in T ) and, despite the loss of roughly 35% of the uncured cyanate ester groups, the residual
g
enthalpy increases, at 82 ± 6 J/g before exposure to 117 ± 23 J/g after exposure and heating to
2
00 °C. A DSC scan of the sample after exposure to water and heating to 350 °C in a TMA
experiment shows only a broad residual enthalpy (using the subsequent re-scan as a baseline).
These results clearly indicate that exposure to water induces additional chemistry on subsequent
heating to 350 °C not seen in dry samples, with carbamate decomposition being the most likely
candidate.

ACCEPTED MANUSCRIPT
st
Figure 7. DSC scans (1 heating) of cured 9 at various points before, during, and after exposure
nd
to 85 °C water and subsequent heating, with the baseline (based on a 2 heating scan)
subtracted.
A further bit of information may be obtained by comparing the second scans of the
aforementioned DSC traces. These second scans are shown in Fig. 8. Each of these samples has
been previously heated to 350 °C, but via different paths. The “as cured” sample has not been
exposed to water, and the “heated to 200 °C” and “heated to 350 °C” samples have been heated
as indicated in the TMA as well as being re-heated to 350 °C in the DSC. Exposure to 350 °C is
expected to remove any uncured cyanate ester groups either by conversion to cyanurate or
thermal decomposition, and to decompose any carbamate groups.
nd
Figure 8. DSC scans (2 heating) of cured 9 at various points before, during, and after exposure
to 85 °C water and subsequent heating. The curves have been offset to coincide at low
temperatures, in order to make the broad glass transitions more apparent.
Despite having all been heated to 350 °C at least once, the samples show different Tg
characteristics. The “as cured” sample shows no T , presumably because a fully-cured network
g

ACCEPTED MANUSCRIPT
has formed with an expected T of 362 °C, which is too high to appear in the scan. The sample
g
exposed to water but heated only to 200 °C in the TMA shows a clearly defined T around 300
g
°
C, indicating a lower density of cross-links. Because roughly 5 mol% of the original cyanate
ester groups have been converted to carbamates and decomposed in this sample, these groups
may be unavailable to form cross-links and instead form dangling chain ends, depressing the Tg
even after all remaining cyanate ester groups have reacted to form cross-links. The sample
heated to 350 °C in the TMA shows a broad T of 250 – 300 °C, in agreement with the T value
g
g
of 260 °C seen in the TMA scan itself (all data from the TMA scans is shown in Supplementary
Data, Figures S8-S10). The sample heated to 350 °C in the TMA thus shows a lower and
broader T than the sample heated to only 200 °C in the TMA and 350 °C in the DSC. The DSC
g
sample is comprised of a ~5 mg chip while the TMA sample consists of a ~300 mg disc. Both
are exposed to a similar 50 mL/min. flow of dry nitrogen, but due to the smaller sample size,
drying will be much faster in the DSC sample. Retained moisture in the TMA sample therefore
likely causes increased conversion of residual cyanate ester groups to carbamates followed by
decarboxylation above 200 °C [82]. This effect also helps explain the differences in the “wet”
TMA scans in the two samples of cured 9 studied. For the sample with 81% conversion and a
low moisture uptake of just 1.4%, fewer residual cyanate ester groups will be lost, allowing for
greater recovery of T on heating due to in-situ cure (the dual peaks in the TMA scan are
g
characteristic of such in-situ cure) [83].
As previously mentioned, exposure of cyanurate networks to hot water has traditionally
been expected to result in an irreversible decline in T , though at least one exception has been
g
reported [84]. Previous studies of carbamate formation in cyanate esters [79] have shown that
the rate of carbamate formation may be changed by at least an order of magnitude depending on

ACCEPTED MANUSCRIPT
the type and level of catalyst used. Similarly dramatic effects on the rate of T decline as a
g
function of catalyst level and type have been reported for cyanate esters based on polyphenolic
resins [85]. Although concrete experimental data is lacking, it is also reasonable to assume that
the rate of degradation will be proportional (though perhaps not linearly) to the concentration of
absorbed water. Because the performance of cyanurate networks is more commonly studied
using catalyzed systems, because most cyanurate networks absorb more water than 9, and
because most cyanurate networks achieve a higher degree of conversion than 9 during typical
cure (due to a lower T ), the conditions employed in the analysis of cured 9 are unusually
g
favorable for the recovery of T . The fact that recovery has been demonstrated, however, and the
g
elucidation of the conditions under which it is likely to occur, represent an important advance in
the study of cyanurate networks. These results also indicate that, if a sufficiently active cure
catalyst that does not simultaneously promote hydrolysis can be identified and utilized, then very
significant improvements in the performance of cyanurate networks in terrestrial applications can
be achieved.
Contextualization of the thermal analysis data of the bio-derived tricyanate ester 9 was
made with that from two cyanate esters derived wholly from petroleum sources, the tricyanate
ester 10 (Fig. 9) and the commercial dicyanate of bisphenol A (BADCy) [86] (Tables 1 and 2).
The data for the latter compounds was taken largely from our previous reports [23,87].
Interestingly, tricyanate 10, though prepared by a completely different synthetic route, is the
symmetrical isomer of tricyanate 9. The DSC data for tricyanates 9 and 10 are almost identical
(cure onset T and ∆H) which is understandable considering the slight difference in chemical
structure (Table 1). Owing to its asymmetric shape [51], tricyanate 9 has a melting point 12 °C

ACCEPTED MANUSCRIPT
lower than its symmetrical counterpart, and almost the same as BADCy which is significant for
molding operations.
Figure 9. Isomeric tricyanate ester 10 previously
prepared from petrochemical starting materials.
Table 1
Comparison of DSC data for monomers 5, 9, 10 and BADCy
∆
Hcure
Cure Onset
(°C)
Exotherm
Max (°C)
215
(kJ/cyanate
ester)
54
Compound
Tm (°C)
165
85
∆Hcure (J/g)
600
5
9
170
225
280
620
97
1
0
97
230
280
780
123
BADCy
82
270
330
810
108
According to Table 2, after full cure, the T of polycyanurate 9 is ~ 40 °C higher than the
g
polycyanurates of BADCy and 20 °C higher than 10. Both water uptake and density of cured
BADCy, 9, and 10 are strongly correlated to cyanate ester conversion. The higher density and
lower water uptake of 9 compared to 10 are the result of both lower conversion in 9 and the
relative similarity of the cure temperature (210 °C) and the “as cured” T . When cured near T ,
g
g
the free volume (or unoccupied space between molecules created by the decrease in Van der
Waals volume as conversion increases) is able to disappear via physical relaxation, leading to

ACCEPTED MANUSCRIPT
higher density and lower uptake, whereas when the cure temperature is below T , the free
g
volume remains frozen in place. These phenomena have been confirmed by positron
annihilation studies in BADCy [88].
Table 2
Comparison of key thermomechanical analysis data of cured 9, 10 and BADCy
Cyanurate
Density at
Full Cure
(mmol/mL) Dry T (°C)
Tg 350 °C
Post-Cure
(°C)
Water
Uptake
(wt%)
1.35
2.2
1.34
Density
(g/mL)
1.23
1.17
1.21
As-Cured
“Wet” Tg
(°C)
Compound
g
9
2.61
2.48
2.89
277
280
275
362
338
323
303
238
240
1
0
BADCy
100%
80%
60%
40%
20%
N2
Air
0%
200 250 300 350 400 450 500 550 600
Temperature (°C)
Figure 10. TGA traces of polycyanurate 9.
The thermogravimetric analysis (TGA) curve up to 600 °C for 9 in both nitrogen and air
atmospheres is shown in Fig. 10 and key data is compared with 10 and BADCy in Table 3. The
thermal stability of 9 is only slightly less than 10 (Fig. S4, Supplementary Data), but 9 loses
more mass in the initial degradation step. Polycyanurate 9 is more stable that BADCy, however,
with char yields in nitrogen and air that are 20% and 50% higher, respectively. This is most

ACCEPTED MANUSCRIPT
likely due to the thermal lability of quaternary aliphatic carbon content which is absent in 9 and
1
0 [89].
Table 3
Comparison of TGA data for cured 9, 10 and BADCy
%
weight loss @ T (°C)
T @ 5%
weight loss
Char Yield
(°C)
600 °C (%)
N2 Air
Compound
N2
Air
415
443
400
N2
69
77
47
Air
70
75
25
450
13
5
550
15
12
600
3
6
450
12
7
550
11
13
600
7
5
9
419
449
402
1
0
BADCy
35
15
3
37
13
25
4
. Conclusions
Novel bio-based tricyanate esters 5 and 9 were prepared from the natural product trans-
anethole by way of [2+2+2]-cycloaddition. Tricyanate 9, with more conformational flexibility
and lower melting temperature gave a wide processing window of at least 100 °C and could be
cured to polycyanurate with thermal properties in the realm of high-performance polymers. The
new bio-based tricyanate 9 had T (~ 362 °C at full cure), water uptake (1.4-2.4 %) and T (85
g
m
°
C) equal or better in all respects to the petrochemical-based BADCy and could make a ‘drop in’
replacement for the latter. When cure far below T is avoided, so as to keep moisture uptake
g
low, the uncatalyzed cured 9 demonstrated the unexpected ability to significantly recover
thermo-mechanical performance after hygrothermal exposure, a discovery with significant
potential technological importance for cyanurate networks in general. Tricyanate 9 and its
isomer 10 were very close in their thermal parameters and both had better thermal stability and
higher char yields than BADCy. Trans-anethole appears to be versatile molecule with which to

ACCEPTED MANUSCRIPT
generate polyphenols and we hope to synthesize and characterize a variety of other polymer
systems (e.g., epoxy, polycarbonate, polyester, etc.) from them in the future.
Acknowledgements
Financial support from the Office of Naval Research, the Air Force Office of Scientific
Research, and the Strategic Environmental Research and Development Program are gratefully
acknowledged. Support from the National Research Council’s Research Associateship Program
(C.M.S.) is also acknowledged. Thanks to Ann M. Moorehead, Mary E. Ray and Cynthia M.
Kitchens of the NAWC Technical Library (China Lake) for obtaining several of the references.
Appendix. Supplementary data
Supplementary data associated with this article can be obtained free of charge, in the
online version, at XXXXX
References
[
1] Hubbert MK, Nuclear Energy and the Fossil Fuels. Publication No. 95, Shell
Development Company, Exploration and Production Research Division, Houston, Texas,
1
956.
[
2] Meadows DH, Meadows DL, Randers J, Behrens III WW. Limits to Growth: a Report for
the Club of Rome’s Project on the Predicament of Mankind. New York: Universe Books;
1
972.

ACCEPTED MANUSCRIPT
[
3] Heinberg R. The Party’s Over: Oil, War, and the Fate of Industrial Societies. British
Columbia: New Society Publishers; 2003.
[
[
[
[
[
4] Campbell CJ, Laherrère JH. Sci Am 1998;278:78-83.
5] Sherwood M. New Sci 1973;60:777-8.
6] Komiyama R, Li Z, Ito K. Int J of Global Energy Issues 2005;24:183-210.
7] United Nations. World Population Prospects: The 2002 Revision. New York; 2003.
8] Plant/crop-based renewable resources 2020: A vision to enhance U.S. economic security
through renewable plant/crop-based resource use. US Department of Agriculture;
1
998 January Report No. DOE/GO-10097-385.
[
9] Armstrong RE. From Petro to Agro: Seeds of a New Economy. Center for Technology
and National Security Policy, National Defense University. Washington (DC):
Government Printing Office; 2002.
[
[
10]
11]
Wang SC, Huffman JB. Econ Bot 1981;35:369-82.
Lipinsky ES. Science 1981;212:1465-71.

ACCEPTED MANUSCRIPT
[
[
[
[
12]
13]
14]
15]
Karger-Kocsis J. Express Polym Lett 2009;3:676.
Calvin M. Bioscience 1979;29:533-8.
Dijkstra DJ, Langstein G. Polym Int 2012;61:6-8.
Szmant HH. Organic building blocks of the chemical industry. New York: John
Wiley & Sons; 1989.
[
[
16]
Corma A, Iborra S, Velty A. Chem Rev 2007;107:2411-502.
17]
Werpy T, Peterson G, editors. Top value-added chemicals from biomass, volume
I: results from screening for potential candidates from sugars and synthesis gas. US
Department of Energy. Washington (DC): Government Printing Office; 2004.
[
18]
Holladay JE, Bozell JJ, White JF, Johnson D. Top value-added chemicals from
biomass, Volume II: results of screening for potential candidates from biorefinery lignin.
Pacific Northwest National Laboratory Report No. 16983. Washington (DC):
Government Printing Office; 2007.
[
19]
Christensen CH, Rass-Hansen J, Marsden CC, Taarning E, Egeblad K.
ChemSusChem 2008;1:283-9.

ACCEPTED MANUSCRIPT
[
[
[
[
20]
Consultative Group on International Agricultural Research. China: voices for
sustainable agriculture. Washington (DC); 2007. p 14-7.
21]
Hardiness zones in China. [climate map]. Ames: US Department of Agriculture-
Agricultural Research Service, Iowa State University; 1997.
22]
Environmental Protection Agency. Test plan for anethole (isomer unspecified)
and trans-anethole. Washington (DC); 2002.
23]
Davis MC, Guenthner AJ, Groshens TJ, Reams JT, Mabry JM. J Polym Sci Part
A: Polym Chem 2012;50:4127-36.
[
[
24]
Lligadas G, Ronda JC, Galia M, Cádiz V. Biomacromolecules 2007;8:1858-64.
Guenthner AJ, Yandek GR, Mabry JM, Lamison KR, Vij V, Davis MC, Cambrea
25]
LR. Insights into moisture uptake and processability from new cyanate ester monomer
and blend studies. SAMPE International Technical Conference, vol. 55. Salt Lake City,
UT: SAMPE International Business Office; 2010. p. 42ISTC-119.
[
[
26]
Gramatica P, Monti D, Manitto P. Gazz Chim Ital 1974;104:629-32.
Orita A, Otera J. Chem Rev 2006;106:5387-412.
27]

ACCEPTED MANUSCRIPT
[
[
[
[
[
[
28]
29]
30]
31]
32]
33]
Naves Y–R. Helv Chim Acta 1960;43:230-2.
McGahey L. J Chem Educ 1990;67:554-5.
Arrieta A, Ganboa I, Palomo C. Synth Commun 1984;14:939-45.
Fahey RC, Schneider H–J. J Am Chem Soc 1968;90:4429-34.
Hell C, von Günther O. J Prakt Chem 1895;52:193-210.
Kunitake T, Aso C, Ito K. Makromol Chem 1966;97:40-8.
[
[
[
34]
35]
36]
Metsger L, Bittner S. Tetrahedron 2000;56:1905-10.
Grob CA, Nussbaumer R. Helv Chim Acta 1971;54:2528-35.
Taherirastgar F, Brandsma L. Synth Commun 1997;27:4035-40.
[
37]
Krause M, Ligneau X, Stark H, Garbarg M, Schwartz J–C, Schunack W. J Med
Chem 1998;41: 4171-6.
[
[
38]
39]
Schulze K, Mühlstädt M. J Prakt Chem 1966;33:84-95.
Reppe W, Schwekendiek WJ. Justus Liebigs Ann Chem 1948;560:104-16.

ACCEPTED MANUSCRIPT
[
[
[
[
[
[
[
40]
41]
42]
43]
44]
45]
46]
Rose JD, Statham FS. J Chem Soc 1950:69-70.
Kotha S, Brahmachary E, Lahiri K. Eur J Org Chem 2005:4741-67.
Frühauf H–W. Chem Rev 1997;97:523-96.
Chopade P R, Louie J. Adv Synth Catal 2006;348:2307-27.
Domínguez G, Pérez-Castells J. Chem Soc Rev 2011;40:3430-44.
Broere DLJ, Ruijter E. Synthesis 2012;2639-72.
Galan BR, Rovis T. Angew Chem Int Ed Eng 2009;49:2830-34.
[
[
47]
48]
Schore NE. Chem Rev 1988;88:1081-119.
Jhingan AK, Maier WF. J Org Chem 1987;52:1161-5.
[
[
49]
Rodríguez JG, Lafuente A, Martín-Villamil R. J Polym Sci Part A: Polym Chem
2
005;43:5987-97.
50]
Grigat E, Pütter R. Chem Ber 1964;97:3012-7.