Terpene-based renewable monomers and polymers via thiol-ene additions

Article abstract of DOI:10.1021/ma201544e

Solvent and radical initiator-free addition of thiols to terpenes ((R)-(+)- and (S)-(-)-limonene and (-)-β-pinene) are described as a simple approach to obtain a wide range of alcohol and/or ester functionalized renewable monomers. (R)-(+)-Limonene (1) and (S)-(-)-limonene (2), presenting different reactivity at the endocyclic and exocyclic double bonds, have yielded the monoaddition or diaddition product by simple variation of the thiol feed ratio. In the same manner, (-)-β-pinene (3) derived alcohol and ester monomers have been prepared. The monomers thus obtained have been characterized, and their behavior in polycondensation has been studied. It has been found that long chain diesters or diols, which were synthesized from a castor oil derived platform chemical, are suitable comonomers and result in polycondensates with number-average molecular weights of up to 25 kDa. Thus, terpene/fatty acid-based polyesters were prepared, and their structure-thermal property relationships were studied.

Full text of DOI:10.1021/ma201544e

ARTICLE
pubs.acs.org/Macromolecules
Terpene-Based Renewable Monomers and Polymers via ThiolÀEne
Additions
Maulidan Firdaus, Lucas Montero de Espinosa, and Michael A. R. Meier*
Institute of Organic Chemistry, Karlsruhe Institute of Technology (KIT), Fritz-Haber-Weg 6, 76131 Karlsruhe, Germany
S Supporting Information
b
ABSTRACT: Solvent and radical initiator-free addition of thiols to terpenes
((R)-(+)- and (S)-(À)-limonene and (À)-β-pinene) are described as a simple
approach to obtain a wide range of alcohol and/or ester functionalized renewable
monomers. (R)-(+)-Limonene (1) and (S)-(À)-limonene (2), presenting different
reactivity at the endocyclic and exocyclic double bonds, have yielded the mono-
addition or diaddition product by simple variation of the thiol feed ratio. In the same
manner, (À)-β-pinene (3) derived alcohol and ester monomers have been prepared.
The monomers thus obtained have been characterized, and their behavior in
polycondensation has been studied. It has been found that long chain diesters or
diols, which were synthesized from a castor oil derived platform chemical, are
suitable comonomers and result in polycondensates with number-average molecular
weights of up to 25 kDa. Thus, terpene/fatty acid-based polyesters were prepared,
and their structureÀthermal property relationships were studied.
Scheme 1. Mechanism of Radical ThiolÀEne Additions
’ INTRODUCTION
The utilization of renewable resources for the synthesis of new
platform chemicals has been accepted as a great challenge in
order to contribute to a sustainable development.1,2 Among the
wide variety of available renewable resources, terpenes are found
in many essential oils and represent a versatile chemical
feedstock.3 R-Pinene and β-pinene are the major components
of wood turpentine, which can be obtained from the resinous sap
of pine trees by steam distillation. It is also one of the main
byproducts of the Kraft process, which is used in the paper
industry to extract lignin from wood in the production of pulp.4
Limonene can be obtained as a byproduct of the citrus industry
and is a very common terpene, being produced by more than 300
the one most used as a fragrance in soaps and cosmetics due to its
typical lilac odor. The selective epoxidation of limonene leads to
limonene oxide, which is found in natural sources and used in
plants.5 The (R)-enantiomer represents 90À96% of citrus peel
oil,6 and its world production is over 70 000 tons per year.7 In
contrast to many other terpenes, pinenes and limonenes are
fragrances. It is also an active cycloaliphatic epoxide with low
abundant and inexpensive natural compounds that are real
viscosity that can be used with other epoxides in applications
building blocks for the synthesis of new important chemicals
including metal coatings, varnishes, and printing inks.
for use as fragrances, flavors, pharmaceuticals, solvents, and chiral
ThiolÀene chemistry, although already known for more than
100 years,9 has not been studied in detail using terpenes as
olefinic substrates. It proceeds via a free radical chain mechanism
and mainly yields the anti-Markovnikov products.10 First, the
initially formed thiyl radical attacks the unsaturated substrate
forming a carbon radical. The carbon radical then reacts with a
thiol molecule to give the final product and a new thiyl radical,
thus propagating the radical chain (Scheme 1). Since this
intermediates.3,8
Known chemical transformations of these renewables were
extensively reviewed by Corma et al.,3 including isomerization,
epoxidation, hydration, or dehydrogenation reactions among
others. The isomerization of terpenes provides a wide range of
products such as camphene, which is produced industrially by
isomerization of R-pinene and is used as an intermediate in the
chemical industry for production of fragrance compounds,
terpeneÀphenol resins, and other derivatives. The hydration of
R-pinene or turpentine oil with aqueous mineral acids yields cis-
terpin hydrate, which can be partially dehydrated into R-terpi-
neol, the most important monocyclic monoterpenic alcohol and
Received: July 19, 2011
Revised:
August 3, 2011
Published: August 18, 2011
r
2011 American Chemical Society
7253
dx.doi.org/10.1021/ma201544e Macromolecules 2011, 44, 7253–7262
|

Macromolecules
ARTICLE
involves the cleavage of a SÀH bond, the overall reaction rate will
be strongly influenced by the structure of the thiol and the
lifetime of the intermediate carbon radical. Moreover, the CÀS
bond formation is a reversible process which also depends on the
structure of the olefin. Thus, the addition of thiols to terminal
olefins, monosubstituted alkynes, and olefins leading to reso-
nance stabilized radical intermediates is less reversible than the
addition to internal olefins.10 Therefore, terminal double bonds
are generally more reactive than internal ones.11
Nowadays, thiolÀene additions are widely used as very
efficient transformations since in many cases they display click
reaction features12 and are thus versatile tools for the preparation
of value-added chemical intermediates.13,14 ThiolÀene additions
can be performed under mild reaction conditions by simply
mixing thiols and olefinic substrates if terminal olefins are used.
Moreover, the products are generally obtained in high yields with
low byproduct formation. The addition of thiols to less reactive
olefins can be initiated thermally with radical initiators or with
radical photoinitiators in presence of UV irradiation.
Even though thiolÀene reactions have been studied for a long
time, covering different scientific fields from biochemistry to
polymer science, only very few examples have been reported on
thiolÀene additions with terpenes as olefinic substrates. In 1957,
Marvel and Olson described the synthesis of a terpene-based
dithiol by addition of thioacetic acid to (R)-limonene.15 They
subsequently used this dithiol in the synthesis of polyalkylene
sulfides via thiolÀene polymerization with the original (R)-
limonene. The obtained poly(alkylene sulfide)s were soft and
sticky with inherent viscosities in the 0.12À0.36 range. Later on,
the addition of hydrogen sulfide to limonene was shown by
Tolstikov et al. to give menth-1-ene-8-thiol, characteristic of the
grapefruit flavor.16 The same researchers reported that this
compound is also obtained by the Lewis acid catalyzed addition
of hydrogen sulfide to pinenes.17 More recently, Janes et al.
reported the synthesis of several terpene-based thiols by reaction
of hydrogen sulfide with different monoterpenoids, observing in
some cases the formation of bridged epi-sulfides as minor
products.18 Overall, the knowledge about the reactivity of
terpenes toward thiol additions is thus limited. However, the
high efficiency of thiolÀene additions, which permits the intro-
duction of different functionalities to olefinic structures, makes it
a suitable way to obtain terpene-based monomers for the
synthesis of renewable polymers.
In this aspect, terpenes have been used since long as renewable
monomers and precursors of monomers in polymer synthesis
through many different approaches.19 Moreover, since terpenes
can be obtained from nature as pure enantiomers, the chiral
polymers derived from them have potential applications in chiral
purification,20 in asymmetric catalysis,21 in nonlinear optics,22
or as conducting materials.23 The cationic polymerization of
β-pinene has been intensively studied in presence of several
Lewis acids at temperatures between À80 and 0 °C in order to
avoid chain transfer.24 Among them, EtAlCl2 was shown to be the
most efficient one producing polymers with molecular weights
up to 40 kDa.25 The cationic polymerization of β-pinene has also
been achieved in a living manner by Lu et al., who prepared block
copolymers of β-pinene with styrene or p-methylstyrene (PDI as
low as 1.2) using isopropoxytitanium trichloride and tetra-n-
butylammonium chloride at À40 °C.26 The radical homopoly-
merization of monoterpenes is an inefficient process that leads
to oligomers with molecular weights below 1000 Da;27 however,
when they are used as comonomersinfree radicalpolymerizations,
high molecular weight polymers can be obtained. Thus, β-pinene
has been successfully copolymerized with styrene,24b methyl
methacrylate,27 acrylonitrile,28 maleimide,29 or a sucrose-based
monomer30 to name a few examples. Likewise, limonene has
been radically copolymerized by Sharma and Srivastava with
styrene,31 methyl methacrylate,32 acrylonitrile,33 N-vinylpyr-
rolidone,34 and other comonomers.19 Regarding the polymeri-
zation of terpene derivatives, Aikins and Williams reported the
radiation-induced cationic polymerization of limonene oxide, R-
pinene oxide, and β-pinene oxide, which led to low molecular
weight polyethers with high monomer conversions.35 More
recently, Coates and co-workers reported the syntheses of
alternating polycarbonate copolymers of (R)-limonene oxide
and carbon dioxide using β-diiminate zinc complexes.36 This
work was the first example of a non-petroleum route to carbon
dioxide copolymers.
Within this work, we aim to extend the current knowledge on
the reactivity of naturally occurring terpenes toward thiol addi-
tions and to establish the reaction conditions that maximize their
efficiency. Furthermore, we present thiolÀene terpene modifica-
tion as a suitable tool to obtain renewable monomers for the
synthesis of polyesters.
’ RESULTS AND DISCUSSION
Syntheses of Monomers. As discussed in the Introduction,
despite the large amount of work that has been done on the
chemical transformation of terpenes for the synthesis of indus-
trially valuable chemicals, building blocks, or monomers for
polymer synthesis, the thiolÀene reaction using limonene and
pinene as olefinic substrates has not been fully explored yet, and
only a few studies have been reported. The reaction of these
abundantly available terpenes with functional thiols is attractive
since it can provide chiral monomers in a one-step procedure
directly from a renewable feedstock. More specifically, the
extensive reaction of both double bonds of limonene with
hydroxyl or methyl ester functionalized thiols would lead to
suitable monomers for polycondensation. On the other hand, the
reaction of pinene with these functional thiols would lead to
terpene-based building blocks that could be used for the synthe-
sis of grafted polymers following a grafting-onto approach or for
the synthesis of monomers of higher complexity. Furthermore,
the different reactivity of the double bonds of limonene toward
thiol additions can be exploited to selectively functionalize the
more reactive exocyclic (terminal) double bond, leaving the
endocyclic (internal) one unreacted for further transformations.
It is worth to highlight the simplicity of these reactions, which can
usually be performed by simply mixing thiol and olefin in the
absence of solvent and radical initiator and at temperatures close
to room temperature.37
Taking this information, we started by studying the solvent-
free, radical initiator-free thiolÀene addition of 2-mercaptoetha-
nol (4), methyl thioglycolate (5), and thioglycerol (6) to (R)-(+)-
limonene (1), (S)-(À)-limonene (2), and (À)-β-pinene (3) at
room temperature (Scheme 2). During the course of our in-
vestigations, we found that 5 shows a higher reactivity than 4 and
6. For this reason, in the following discussion, the results of the
thiolÀene additions using 5 will be addressed separately since a
more thorough optimization of the reaction conditions was
needed. The high reactivity of thiols based on glycolate esters
has been observed before and was ascribed to a weakening of the
thiol bond by formation of a hydrogen bond with the carbonyl
7254
dx.doi.org/10.1021/ma201544e |Macromolecules 2011, 44, 7253–7262

Macromolecules
ARTICLE
Scheme 2. Mono- and Diadditions of Thiols 4À6 to
Terpenes 1À3
Table 2. Results of Diadditions of Thiols 4À6 to Terpenes 1
and 2 and Additions of 4 and 5 to Monoaddition Products
terpene
thiol (equiv)
product
yield (%)d
1
4 (2.5)a
5 (2.5)a
6 (2.5)a
4 (2.5)a
5 (2.5)a
6 (2.5)a
5 (2.0)b
4 (1.2)c
5 (2.0)b
4 (1.2)c
7 + 16
36 (7); 62 (16)
5 (8); 93 (17)
1
8 + 17
1
e
2
10 + 18
11 (10); 82 (18)
4 (11); 93 (19)
2
11 + 19
2
e
7
20
21
22
23
83
92
81
90
8
10
11
a At rt for 48 h. b At rt for 72 h. c At rt for 24 h. d After column
chromatography. e Only the monoaddition product was obtained.
dropped ∼10%. An increase of the temperature to 35 °C gave
only a very slight increase in conversions (below 5% increase)
and at the same time generally led to a drop of regioselectivity. All
thiolÀene additions were also performed in the presence of
AIBN at 70 °C. Adding a radical initiator resulted in faster
conversion of the terpenes; however, it also resulted in lower
regioselectivities, and the conversion rates decreased at reaction
times above 5 h, which is about the half-life time of AIBN. As a
summary, Table 1 shows the results obtained after optimization
of the reaction conditions for the thiolÀene monoadditions and
clearly reveals that the exocyclic double bond of limonene can be
addressed regioselectively in these radical additions.
Once the thiolÀterpene monoadditions were optimized, we
focused on the synthesis of the diaddition products of thiols 4, 5,
and 6 to terpenes 1 and 2 (Scheme 2), for which a thiolÀterpene
ratio of 2.5:1 was used. The addition of 6 to both 1 and 2 did not
give the diaddition products, and only the monoadditions to the
exocyclic double bonds were observed, probably due to steric
hindrance around the endocyclic double bonds. The additions of
4 and 5 led to a mixture of mono- and diaddition products with
conversions over 97% (Table 2). As already mentioned, the
monoaddition products from (R)-(+)- and (S)-(À)-limonene
can undergo a second thiol addition with a different thiol to
obtain heterodifunctional monomers. Thus, the alcohol func-
tionalized products 7 and 10 were reacted with methyl thiogly-
colate, and the ester functionalized products 8 and 11 were
reacted with 2-mercaptoethanol (Scheme 3). The reactions were
conducted at room temperature, in the absence of solvent and
radical initiator, and using 2-fold and 1.2-fold excesses of methyl
thioglycolate and 2-mercaptoethanol, respectively. In this way,
difunctional monomers containing both ester and alcohol groups
(20À23, Table 2) were obtained.
Table 1. Results of Monoaddition of Thiols 4À6 to Terpenes
1À3a
terpene
thiol
product
yield (%)b
1
1
1
2
2
2
3
3
3
4
5
6
4
5
6
4
5
6
7
85
55
80
82
52
61
81
60
76
8
9
10
11
12
13
14
15
a Reaction conditions: terpene:thiol ratio 1:1.2, rt for 24 h. b After
column chromatography.
group.13 With the exception of additions of 5, the rest of the
reactions were performed under vacuum (200 mbar) in order to
remove oxygen,37 which is an efficient radical scavenger in these
reactions. First, the thiolÀterpene ratio was optimized to max-
imize the efficiency of the monoaddition reactions. Initially, a 1:1
ratio was used for the addition of 4 and 6 to 1, 2, and 3. However,
the conversion of the double bonds was rather low (below 50%),
even after 48 h. When the amount of thiol was increased to 1.2
equiv, the double bond conversions also increased, and 60À80%
isolated yields of monoaddition products were obtained after
column chromatography (see Table 1). Interestingly, the addi-
tion of 4 and 6 to 1 and 2 took place exclusively at the exocyclic
double bond, and no diaddition products were observed. On the
other hand, the addition of 5 to 1 and 2 showed a lower
selectivity, giving additions both to the exocyclic and endocyclic
double bonds when working in a 1:1 ratio. Moreover, low double
bond conversions were obtained even after 24 h (below 75%).
Also for the addition of 5, the use of a thiol excess (1.2 mol equiv)
provided better results, leading to yields ranging from 52 to 85%
after column chromatography (Table 1), but also to a decrease in
the regioselectivity (from 74 to 55%). A thiolÀterpene ratio of
0.9:1 was then tried in order to favor monoaddition; although it
gave a slight increase of regioselectivity (77%), the isolated yield
Characterization of Monomers. The terpenes used in this
study are pure enantiomers. The addition of one molecule of
thiol to the exocyclic double bond of 1 or 2 or to the double bond
of 3 generates a new stereogenic center and, thus, two different
diastereomers of each product. The addition of a second thiol
molecule to the endocyclic double bond of 1 or 2 generates two
additional stereogenic centers, increasing the number of possible
diastereomer to eight. However, the cyclic and bicyclic structures
of these terpenes, together with the fact that they are pure
enantiomers, favor the formation of certain diastereomers over
others as a result of preferential sides for the thiol approach
during the hydrogen abstraction process. Nevertheless, complex
7255
dx.doi.org/10.1021/ma201544e |Macromolecules 2011, 44, 7253–7262

Macromolecules
ARTICLE
NMR spectra were obtained for all addition products and
conventional two-dimensional NMR experiments (COSY,
HSQC, and HMBC) were necessary for full assignment.
see experimental part). The methyl groups of both diastereomers
appear as doublets of equal intensity, which merge in a false
triplet, indicating a 1:1 ratio between both diastereomers. A
different case can be observed in Figure 1b, which shows the 1H
NMR spectrum of 13 (addition of mercaptoethanol to (À)-β-
pinene). As can be seen, the signals of the methyl groups attached
to the bicyclic backbone present different chemical displace-
ments in both diastereomers. The integration of these signals
reveals a ratio of 5:1 between the diastereomers, which confirms
that the hydrogen abstraction is stereoselective. In order to
determine which one of the two possible diastereomers is the
major, we performed a NOE experiment (see Supporting
Information), which showed correlation between the protons
He and Hd of the major diastereomer. This confirms that the
approach of the second thiol, and thus the hydrogen abstraction,
preferentially takes place from the opposite face to the methyl
groups due to steric hindrance.
The presence of diastereomers, which could not be isolated by
column chromatography, was also observed in the GC-MS
chromatograms of the purified monoaddition products to 1, 2,
and 3, where (although not always) two peaks can be observed.
For instance, the chromatogram of 14 shows two peaks with
different intensities (see Supporting Information), indicating
that after the initial thiol addition the approach of a second thiol
to the carbon-centered radical, and thus the hydrogen abstrac-
tion, takes place preferably from one side. The GC chromato-
grams of the purified diaddition products to 1 and 2 present
several peaks of different intensities as a result of the partial
diastereoselectivity in the hydrogen abstraction (see Supporting
Information for chromatograms of 17 and 19).
The analysis of the NMR spectra of these products further
confirmed the presence of diastereomers. Moreover, in some
cases, the integration of the 1H NMR signals allowed calculating
the ratio of the diastereomers. For instance, Figure 1a shows the
1H NMR spectrum of 7 (addition of 2-mercaptoethanol to
R-(+)-limonene), in which the region of methyl groups bonded
to the new stereogenic center is highlighted (for full assignment
The 1H NMR spectrum of 17 (diaddition of methyl thiogly-
colate to R-(+)-limonene) shows a higher complexity due to the
higher number of possible diastereomers (Figure 2a). In this
case, the analysis of the 13C NMR is more helpful in order to
determine the number of diastereomers present in the diaddition
product and to estimate their ratio. Figure 2b shows the region
from 53 to 51 ppm, in which four different signals are observed
Scheme 3. Addition of Thiols 4 or 5 to Monoaddition
Products
Figure 2. (a) 1H NMR and (b) 13C NMR (region 53À51 ppm) of 17.
Figure 1. 1H NMR spectra of (a) 7 and (b) 13.
7256
dx.doi.org/10.1021/ma201544e |Macromolecules 2011, 44, 7253–7262

Macromolecules
ARTICLE
Scheme 4. Polymerization of Difunctional Monomers and Homopolymerization of Heterodifunctional Monomers
Table 3. Analytical Data of Synthesized Limonene-Based
Polyestersa
for Ca, indicating that the product mainly consists of a mixture of
four diastereomers.
Polymerization Studies. The difunctional monomers ob-
tained by diaddition of 2-mercaptoethanol or methyl thioglyco-
late to (R)-(+)-limonene or (S)-(À)-limonene were sub-
sequently used for polyester synthesis. The purity of these
monomers was >99% by GC analysis; no monofunctional
byproduct were detected that could act as chain-stoppers.
1,5,7-Triazabicyclo[4.4.0]dec-5-ene (TBD) was chosen as poly-
condensation catalyst based on our own experience and on
previous reports on its high tranesterification activity.38 The
removal of methanol strongly determines the efficiency of TBD-
catalyzed transesterification and can be carried out applying a
continuous stream of an inert gas such as nitrogen or constant
vacuum in the case of polymerizations. Thus, we performed the
polymerizations in presence of 5 mol % (related to ester groups)
TBD at 100 and 120 °C under continuous vacuum for 7 h.
Initially, diesters 17 and 19 were polymerized with diols 16 or 18
at 200 mbar for 7 h; however, oligomers of 3 kDa were obtained
(Scheme 4), probably due to the bulky cyclic structures of both
monomers, which could hinder the catalysts’ approach and thus
decreasing the polymerization efficiency. Moreover, longer reac-
tion times did not lead to an increase of the molecular weight.
Titanium isopropoxide was also tried in an attempt to reach
higher molecular weights. Titanium alkoxides are known to act as
efficient transesterification catalysts and have been successfully
used for polyester synthesis39 being compatible with a variety of
functional groups.40 Moreover, titanates are recognized as ex-
tremely powerful for ester-exchange reactions.41 However, using
5 mol % of titanium isopropoxide under continuous vacuum and
100 or 120 °C provided only slightly higher molecular weights
(up to 4 kDa) after 7 h (longer reaction times did not lead to
higher molecular weights). The polymerization of heterodifunc-
tional monomers 20À23 was then attempted in presence of
TBD (Scheme 4). Interestingly, despite having very similar
structures to the above-mentioned diesters and diols, the poly-
condensation of these monomers at 120 °C under continuous
vacuum for 7 h led to homopolymers with higher molecular
weights (Mn) between 8 and 10 kDa (see Table 3 for data
of precipitated polymers). During polycondensations, little
polymer (monomer)
Mn (kDa)
PDI
Tg (°C)b
P1 (20)
P2 (21)
P3 (22)
P4 (23)
9.3
7.7
1.79
1.66
1.89
1.65
À9.7
À9.2
À10.4
À9.9
10.5
8.2
a GPC and DSC data of precipitated polymers. b DSC data recorded at
10 °C/min; results from the second heating scan.
deviations from an ideal 1:1 ratio between monomers can
substantially decrease molecular weights through a chain-stopper
effect. Monomers 20À23, which do not need comonomers, are
thus more likely to reach high molecular weights. However, also
here, the sterical hindrance caused by the terpene core might
reduce the activity of the catalyst.
From these results, it can be concluded that in order to obtain
high molecular weight polyesters from limonene-based diesters
and diols, comonomers providing spacing between terpene units
might be needed. First, diesters 17 and 19 were polymerized with
1,3-propanediol and 1,6-hexanediol in the presence of 5 mol % of
TBD. The polymerizations, which were also performed at 100
and 120 °C and 200 mbar vacuum, only yielded oligomers below
5 kDa. Moreover, substituting TBD by titanium isopropoxide
(5 mol %) in the same reaction conditions provided only slightly
higher molecular weights (up to 7 kDa) and broad polydisper-
sities. All polymerizations were performed for 7 h, since longer
reaction times did not further increase the molecular weights.
These results suggested that comonomers with longer alkyl
chains between the reactive groups might be needed due to
steric reasons. To prove this, suitable diester (24 and 25,
Scheme 6) and diol (26 and 27, Scheme 6) monomers were
prepared via self-metathesis or thiolÀene homocoupling of
methyl-10-undecenoate and 10-undecen-1-ol. It is worth to
mention that both platform chemicals are castor oil derived,
and thus, the monomers obtained thereof are also renewable.
The presence of double bonds affects both crystallinity and
thermal stability of polymers, and for this reason, unsaturated
7257
dx.doi.org/10.1021/ma201544e |Macromolecules 2011, 44, 7253–7262

Macromolecules
ARTICLE
Scheme 5. Synthesis of Fatty Acid Derived 1,20-Diol 26 and Metathesis Catalysts Studied
diester 24 was reduced to its saturated counterpart (28,
Scheme 6) in order to compare the thermal properties of the
polyesters derived from them. Diesters 2442 and 2537 and diol
2737 were prepared as previously reported. The synthesis of diol
26 was carried out via self-metathesis of 10-undecen-1-ol
(Scheme 5). Olefins containing alcohol groups, such as 9-de-
cen-1-ol,43 have been reported to give moderate yields during
metathesis reactions due to formation of side products and/or
degradation of the catalyst.44À46 In this regard, alcohol protec-
tion was shown to be helpful in overcoming this drawback.47
Moreover, we have already shown that acetylation of oleyl
alcohol and subsequent cross-metathesis reactions were more
efficient, in terms of both catalyst load and produced waste, than
using oleyl alcohol directly.48 Therefore, we followed the same
strategy and protected 10-undecen-1-ol by reaction with acetic
anhydride to obtain 10-undecenyl acetate (29, Scheme 5). After
purification by simple filtration through silica gel, the behavior of
29 in self-metathesis was studied. Three different metathesis
catalysts were tested (Scheme 5), namely HoveydaÀGrubbs
second generation (C1), Zhan (C2), and Umicore 51 (C3), to
obtain the desired self-metathesis product 30 (Scheme 5). These
catalysts have been shown to give high conversions at loadings as
low as 0.05 mol % in self- and cross-metathesis reactions of fatty
acid derivatives.42 Moreover, the addition of 1 mol % of p-
benzoquinone to these reactions efficiently prevents double
bond isomerization,49 which can drastically decrease the yield
of the desired product.42 Initially, catalyst loadings of 0.5 mol %
were used, and the reactions were followed by GC-MS in order to
identify the most active catalyst. Vacuum (200 mbar) was applied
throughout the reaction to remove the released ethylene. Fast
conversions were observed for all three catalysts within the first
hour of reaction; however, the conversions did not further
increase appreciably. C1 gave the lowest conversion (48.12%
after 5 h) with very low double bond isomerization (1.3%). C2
and C3 performed similarly leading to conversions over 80%
with only slightly higher isomerization degrees around 2.5%
after 5 h reaction. On the basis of these results, C2 was selected
for further optimization. Thus, C2 loadings of 0.2 and 1 mol %
were also tested; however, using 0.2 mol % resulted in lower
conversion after 5 h reaction, and using 1 mol % gave a higher
isomerization degree despite an increase in conversion. From
these results, we could conclude that around 0.5 mol % C2 is a
suitable choice for the self-metathesis of 29. The deprotection
of the hydroxyl groups was carried out by direct reaction of the
metathesis reaction mixture with excess of methanol in pre-
sence of TBD at 65 °C, and the product (26) was purified by
recrystallization from methanol. It is worth to highlight the
simplicity of this synthetic route, which leads to 26 with an
overall yield of 67%, without chromatographic purification
steps (Scheme 5).
The polymerizations of limonene-based monomers 16À19
were then performed with 24, 25, 26, 27, and 28 as fatty acid-
based comonomers, in the presence of 5 mol % of TBD, and
under continuous vacuum (10 mbar), leading to polyesters
P5ÀP14 (Scheme 6). Table 4 summarizes the results of these
polymerizations at 120 °C for 7 h (data of precipitated poly-
mers), and Figure 3 shows GPC traces of the (R)-(+)-limonene-
derived polyesters for comparison. As expected, higher molecular
weight polymers (between 9 and 24 kDa and PDI between 1.75
and 2.47) could be obtained in all cases. This was further
confirmed by NMR analysis. As a representative example,
1
Figure 4 shows the H NMR spectra of diester 17 and P5,
revealing the transformation of methyl ester (Ha) to the back-
bone ester links (Hc).
When short-chain comonomers are used, the activity of the
catalyst can be reduced due to the steric hindrance caused by the
terpene rings. On the other hand, the introduction of long,
flexible alkyl chains can reduce the steric hindrance around the
terpene units, facilitating the approach of the catalyst. Interest-
ingly, when comparing the polymerization results obtained with
diesters 24 (unsaturated, P7, and P8) and 28 (saturated, P13,
and P14), higher molecular weights are obtained with the
unsaturated diester, probably due to a more flexible chain both
in the monomers and the resulting polymer.
Furthermore, a closer look at the polymerization results
reveals an interesting fact regarding the obtained molecular
weights. The molecular weights reached with diester monomers
containing the methyl thioglycolate moiety (17 and 19) are
clearly higher than those obtained with diesters which do not
have the sulfur in β-position to the carbonyl group (24 and 25).
7258
dx.doi.org/10.1021/ma201544e |Macromolecules 2011, 44, 7253–7262

Macromolecules
ARTICLE
Scheme 6. Synthesis of (R)-(+)- and (S)-(À)-Limonene/Fatty Acid-Based Polyesters
Table 4. Analytical Data of Synthesized (R)- and (S)-Limo-
nene/Fatty Acid-Based Polyesters
polymer (diester/diol)
Mna (kDa)
PDIa
Tg (°C)b/Tm (°C)b
P5 (17/26)
P6 (19/26)
P7 (24/16)
P8 (24/18)
P9 (17/27)
P10 (19/27)
P11 (25/16)
P12 (25/18)
P13 (28/16)
P14 (28/18)
18.9
21.3
15.3
15.5
24.7
23.1
12.4
14.0
9.7
1.97
2.00
1.82
1.75
2.47
2.47
2.40
2.06
2.22
2.22
À46.9c/À15.4c
À47.8c/À15.6c
À46.6d/17.1d
À45.3e/17.0e
À45.4f/28.2f
À46.3d/28.3d
À41.9d/47.5d
À43.0g/47.2g
h/50.8d
9.2
h/50.3d
Figure 3. GPC traces of (R)-(+)-limonene-based polyesters.
a GPC data of precipitated polymers. b DSC data recorded at 10 °C/min.
c 1 h annealing at À25 °C. d Second heating scan. e 1 h annealing at 16 °C.
f 1 h annealing at 20 °C. 1 h annealing at 10 °C. h Tg not observable
g
In order to establish whether the sulfur atom has an effect on the
course of the esterification reaction, we performed a simple
by DSC.
7259
dx.doi.org/10.1021/ma201544e |Macromolecules 2011, 44, 7253–7262

Macromolecules
ARTICLE
Figure 4. 1H NMR spectra of (a) monomer 17 and (b) product of
polycondensation of 17 with diol 26 (P5).
Figure 6. DSC traces of terpene-based polyesters P1ÀP4 (second
heating scan).
Figure 5. Study of the effect of sulfur in β-position to ester group on the
reactivity toward transesterification with 10-undecen-1-ol.
Figure 7. DSC traces of (R)-(+)-limonene (—) and (S)-(À)-limonene
(
) based polyesters P5ÀP14 (see Table 4 for experimental
3 3 3
conditions).
kinetic study in which 10-undecen-1-ol was reacted with diester
(31) containing both a methyl β-thioester and a methyl ester
(Figure 5). The reaction was performed under polymerization
conditions (5 mol % TBD related to ester groups, 120 °C, and 10
mbar) with a 1:1 molar ratio of the reactants, and the consump-
tion of both ester groups was monitored in time by integration of
characteristic 1H NMR signals (see Supporting Information for
details). The results of this competitive study, which are shown in
Figure 5 as a plot of ester conversion against time, confirm the
higher reactivity of the methyl β-thioester compared to the
unsubstituted one.
The effect of sulfur on the reactivity of the ester group might be
explained on the basis of the mechanism through which TBD
catalyzes transesterification reactions. It has been proposed that
in a first step a reversible amidation occurs by formation of a six-
membered ring intermediate between TBD and the ester group
followed by release of methanol.50 In a second step, a hydrogen
bond is established between the TBD moiety and the approach-
ing alcohol, favoring its approach and weakening the OÀH bond.
In this second step, the sulfur atom might establish a hydrogen
bond with the alcohol, thus assisting the process. This and other
hypotheses are currently under investigation in order to clarify
the role of sulfur in this reaction.
is thus very unlikely, and DSC analysis reveals amorphous structures
for all of them (see Figure 6) with glass transition temperatures
around À10 °C.
Limonene/fatty acid-based polyesters P5ÀP14 (Scheme 6)
display as a common structural feature the alternation of
relatively long aliphatic segments (18 and 20 carbon atoms)
and substituted cycloalkanes. On one hand, the aliphatic chains
are able to crystallize, but on the other hand, the 1,3,6-trisub-
stituted cyclohexane units are too bulky to crystallize. DSC traces
obtained for polyesters P5ÀP12 are shown in Figure 7, and the
extracted analytical data are summarized in Table 4. With the
exception of P7, P10, P13, and P14, the studied polyesters
displayed multiple melting transitions on the second heating
scan. Annealing studies were thus performed to determine
whether these multiple melting transitions were due to the
presence of metastable crystalline phases or actual crystalline
structures (see Table 4 for experimental details). Appropriate
annealing conditions could be found for P5, P6, P8, and P12,
which led to one, more or less broad, melting peak. However,
annealing studies performed on P9 and P11 did not lead to single
melting transitions, suggesting polymorphism. A first look at the
thermal data in Table 4 reveals that all the studied polyesters are
semicrystalline, displaying both glass transition (Tg) and melting
(Tm). Especially noteworthy is the fact that all studied polyesters
(except P13 and P14, for which the Tg could not be observed by
Thermal Analysis. Polyesters obtained from limonene-based
monomers 20À23 (P1ÀP4) consist of bulky 1,3,6-trisubstituted
cyclohexane units connected by very short segments. Crystallization
7260
dx.doi.org/10.1021/ma201544e |Macromolecules 2011, 44, 7253–7262

Macromolecules
ARTICLE
DSC) have basically the same glass transition temperature
despite their clearly differentiated Tm values. This data could
suggest that the polyester chains might crystallize independently
of the cyclic moieties (common in all polyesters), letting them
out of the crystallites as amorphous regions. The melting
behavior of polyesters P5ÀP14 can be explained attending to
the factors that enhance or hinder interchain interactions and to
their different chain-packing abilities. P5 (Tm = À15.4 °C) and
P7 (Tm = 17.1 °C) present well-differentiated melting points
despite having relatively similar structures. Both present bulky
main-chain groups that can impede crystallization, but in this
case, the spacing between the ester groups and the bulky
limonene rings, which is longer in P7, seems to favor interchain
interactions and/or chain packing compared to P5. The same
applies for the pairs P6/P8, P9/P11, and P10/P12. On the
other hand, the introduction of fatty acid-based comonomers
containing sulfur atoms leads to an increase of the melting
temperatures. Thus, polyesters P9ÀP12 display melting tem-
peratures which are up to 43 °C higher than for polyesters
P5ÀP8 (see e.g. P5 vs P9). The effect of sulfur on the Tm of
these polyesters can be correlated with a higher cohesion
energy due to stronger interchain interactions.51 Finally, satu-
rated polyesters P13 and P14 display sharp melting en-
dotherms, which are around 35 °C higher than those of their
unsaturated counterparts (P7 and P8). This effect is due to the
higherpackingabilityofthe saturatedalkylchainsandtothehigher
flexibility of the unsaturated ones, which results in the observed
higher melting points. No glass transition could be detected for
these polyesters, which also supports their higher crystallinity
compared to P7 and P8.
’ ACKNOWLEDGMENT
M.F. is thankful for a fellowship from the Indonesian Direc-
torate General of Higher Education.
’ DEDICATION
Dedicated to Professor Christian Bruneau on the occasion of his
60th birthday.
’ REFERENCES
(1) Christensen, C. H.; Rass-Hansen, J.; Marsden, C. C.; Taarning,
E.; Egeblad, K. ChemSusChem 2008, 1, 283–289.
(2) Metzger, J. O.; Eissen, M. C. R. Chim. 2004, 7, 569–581.
(3) Corma, A.; Iborra, S.; Velty, A. Chem. Rev. 2007, 107,
2411–2502.
(4) Speight, J. G. In Chemical Process and Design Handbook;
McGraw-Hill: New York, 2002.
(5) Burdock, G. A. In Fenaroli’s Handbook of Flavour Ingredients, 3rd
ed.; CRC Press: Boca Raton, FL, 1995.
(6) Nonino, E. A. Perfum. Flavor. 1997, 22, 53–58.
(7) Kerton, F. M. In Alternative Solvents for Green Chemistry; Clark,
J. H., Kraus, G. A., Eds.; RSC Green Chemistry: Cambridge, UK, 2009;
p 109.
(8) Thomas, A. F.; Bessiꢀere, Y. Nat. Prod. Rep. 1989, 291–309.
(9) Posner, T. Chem. Ber. 1905, 38, 646–657.
(10) Griesbaum, K. Angew. Chem. 1970, 82, 276–290.
(11) Walling, C.; Helmreich, W. J. Am. Chem. Soc. 1959,
81, 1144–1148.
(12) Kolb, H. C.; Finn, M. G.; Sharpless, K. B. Angew. Chem., Int. Ed.
2001, 40, 2004–2021.
(13) Hoyle, C. E.; Lee, T. Y.; Roper, T. J. Polym. Sci., Part A: Polym.
Chem. 2004, 42, 5301–5338.
(14) Killops, K. L.; Campos, L. M.; Hawker, C. J. J. Am. Chem. Soc.
2008, 130, 5062–5064.
’ CONCLUSIONS
(15) Marvel, C. S.; Olson, L. E. J. Polym. Sci. 1957, 26, 23–28.
(16) Tolstikov, G. A.; Kanzafarov, F. Y.; Sangalov, Y. A.; Dzhemilev,
U. M. Neftekhimiya 1979, 19, 425–429.
(17) Tolstikov, G. A.;Kanzafarov, F. Y.; Dzhemilev, U. M.; Kantyukova,
R. G.; Zelenova, L. M. Zh. Org. Khim. 1983, 19, 2075.
(18) Janes, J. F.; Marr, I. M.; Unwin, N.; Banthorpe, D. V.; Yusuf, A.
Flavour Fragr. J. 1993, 8, 289–294.
(19) Silvestre, A. J. D.; Gandini, A. In Monomers, Polymers and
Composites from Renewable Resources; Belgacem, M. N., Gandini, A.,
Eds.; Elsevier: Oxford, 2008; p 17.
Addition of alcohol and ester functionalized thiols to terpenes
takes place at room temperature in the absence of solvent and
radical initiator. These reactions are regioselective and can thus
be controlled to yield monofunctional, difunctional, or hetero-
difunctional monomers which are interesting renewable building
blocks. Moreover, it was found that the addition of thiols to
terpenes is also diastereoselective, reaching a ratio of 5:1 for this
radical reaction in the addition of 2-mercaptoethanol to (À)-β-
pinene. This effect was related to steric effects making hydrogen
abstraction more favorable from one of both terpene faces.
Difunctional monomers prepared from (R)-(+)-limonene (1)
and (S)-(À)-limonene (2) have been subjected to polyconden-
sation. Oligomers or low molecular weight polyesters were
obtained when these terpene-based monomers were homopoly-
merized or copolymerized with short-chain diols. However,
when long-chain fatty acid-based diesters and diols were used
as comonomers, polyesters with number-average molecular
weights up to 25 kDa (Mn) were obtained. Most polyesters
synthesized are semicrystalline with a common Tg around
À45 °C and melting points ranging from À15 to 50 °C.
(20) Xie, R.; Chu, L.-Y.; Deng, J.-G. Chem. Soc. Rev. 2008,
37, 1243–1263.
(21) Nakano, T.; Okamoto, Y. Chem. Rev. 2001, 101, 4013–4038.
(22) Verbiest, T.; Kauranen, M.; Persoons, A. J. Mater. Chem. 1999,
9, 2005–2012.
(23) Kane-Maguire, L. A. P.; Wallace, G. G. Chem. Soc. Rev. 2010,
39, 2545–2576.
(24) (a) Roberts, W. J.; Day, A. R. J. Am. Chem. Soc. 1950,
72, 1226–1230. (b) Barros, M. T.; Petrova, K. T.; Ramos, A. M. Eur. J.
Org. Chem. 2007, 8, 1357–1363. (c) Martinez, F. J. Polym. Sci., Part A:
Polym. Chem. 1984, 22, 673–677.
(25) Satoh, K.; Sugiyama, H.; Kamigaito, M. Green Chem. 2006,
8, 878–882.
(26) Lu, J.; Kamigaito, M.; Sawamoto, M. Macromolecules 1997,
30, 27–31.
’ ASSOCIATED CONTENT
(27) Ramos, A. M.; Lobo, L. S. Macromol. Symp. 1998, 128, 43–50.
(28) Li, A.-L.; Wang, Y.; Liang, H.; Lu, J. J. Polym. Sci., Part A: Polym.
Chem. 2006, 44, 2376–2387.
S
Supporting Information. Experimental details. This ma-
b
terial is available free of charge via the Internet at http://pubs.acs.org.
(29) Satoh, K.; Matsuda, M.; Nagai, K.; Kamigaito, M. J. Am. Chem.
Soc. 2010, 132, 10003–10005.
’ AUTHOR INFORMATION
(30) Barros, M. T.; Petrova, K. T.; Ramos, A. M. Eur. J. Org. Chem.
2007, 8, 1357–1363.
(31) Sharma, S.; Srivastava, A. K. Eur. Polym. J. 2004, 40, 2235–2240.
Corresponding Author
*E-mail: m.a.r.meier@kit.edu.
7261
dx.doi.org/10.1021/ma201544e |Macromolecules 2011, 44, 7253–7262

Macromolecules
ARTICLE
(32) Sharma, S.; Srivastava, A. K. J. Macromol. Sci., Pure Appl. Chem.
2003, A40, 593–603.
(33) Sharma, S.; Srivastava, A. K. Polym. Plast. Technol. Eng. 2003,
42, 485–502.
(34) Sharma, S.; Srivastava, A. K. Des. Monomers Polym. 2006,
9, 503–516.
(35) Aikins, J. A.; Williams, F. In Ring-Opening Polymerization;
McGrath, J. E., Ed.; ACS Symp. Ser.; American Chemical Society:
Washington, DC, 1985; Vol. 286, Chapter 24, p 335.
(36) Byrne, C. M.; Allen, S. D.; Lobkovsky, E. B.; Coates, G. W. J.
Am. Chem. Soc. 2004, 126, 11404–11405.
(37) T€ur€unc-, O.; Meier, M. A. R. Macromol. Rapid Commun. 2010,
31, 1822–1826.
(38) Schuchardt, U.; Sercheli, R.; Vargas, R. M. J. Braz. Chem. Soc.
1998, 9, 199–210.
(39) Okada, M.; Aoi, K.; Mizuno, W.; Ito, S. J. Polym. Sci., Part A:
Polym. Chem. 1993, 31, 1135–1140.
(40) Seebach, D.; Hungerbuhler, E.; Haef, R.; Schnurrenberger, P.
Synthesis 1982, 138–141.
(41) (a) Coullerez, G.; Lowe, C.; Pechy, P.; Kausch, H. H.; Hilborn,
J. J. Mater. Sci., Mater. Med. 2000, 11, 505–510. (b) Hiltunen, K.; Seppala,
J. V.; Harkonen, M. Macromolecules 1997, 30, 373–379.
(42) Djigouꢁe, G. B.; Meier, M. A. R. Appl. Catal., A 2009,
368, 158–162.
(43) Blackwell, H. E.; O’Leary, D. J.; Chatterjee, A. K.; Washenfelder,
R. A.; Bussmann, D. A.; Grubbs, R. H. J. Am. Chem. Soc. 2000, 122, 58–71.
(44) Schmidt, B.; Staude, L. J. Organomet. Chem. 2006, 691,
5218–5221.
(45) Schmidt, B. Eur. J. Org. Chem. 2004, 9, 1865–1880.
(46) Banti, D.; Mol, J. C. J. Organomet. Chem. 2004, 689, 3113–3116.
(47) Lantelme, B.; Dumon, M.; Mai, C.; Pascault, J. P. J. Non-Cryst.
Solids 1996, 194, 63–71.
(48) Rybak, A.; Meier, M. A. R. Green Chem. 2008, 10, 1099–1104.
(49) Hong, S. H.; Sanders, D. P.; Lee, C. W.; Grubbs, R. H. J. Am.
Chem. Soc. 2005, 127, 17160–17161.
(50) Coles, M. P. Chem. Commun. 2009, 3659–3676.
(51) Yoda, N. J. Polym. Sci., Part A: Gen. Pap. 1963, 1, 1323–1338.
7262
dx.doi.org/10.1021/ma201544e |Macromolecules 2011, 44, 7253–7262