Ruthenium-catalysed codimerisation of myrcene with methyl acrylate: Catalyst screening and mechanistic discussions

Article abstract of DOI:10.1016/j.apcata.2012.12.020

The codimerisation of 1,3-dienes and acrylic compounds is a feasible strategy for the synthesis of functionalised alkenes. In this work, a direct approach to novel C₁₃ esters for the flavour and fragrance industry was performed by codimerisatio

Full text of DOI:10.1016/j.apcata.2012.12.020

Applied Catalysis A: General 453 (2013) 204–212
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Ruthenium-catalysed codimerisation of myrcene with methyl
acrylate: Catalyst screening and mechanistic discussions
Arno Behr∗, Leif Johnen, Nils Rentmeister
Technische Universität Dortmund, Lehrstuhl für Technische Chemie A, Emil-Figge-Str. 66, D-44227 Dortmund, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
The codimerisation of 1,3-dienes and acrylic compounds is a feasible strategy for the synthesis of function-
alised alkenes. In this work, a direct approach to novel C13 esters for the flavour and fragrance industry
was performed by codimerisation of myrcene and methyl acrylate. Thus, a further contribution was
demonstrated for the use of terpenes as a biobased feedstock for fine chemicals. Different homogeneous
ruthenium catalysts showed high activities in the target reaction. Systematic optimisation of the reaction
parameters were carried out, resulting in high regioselectivities and yields. A comparison of codimeri-
sation products described in the literature and the observed C13 esters led to mechanistic discussions.
Received 15 October 2012
Received in revised form
11 December 2012
Accepted 13 December 2012
Available online xxx
Keywords:
Codimerisation
Homogeneous catalysis
Renewable resources
Ruthenium
© 2013 Published by Elsevier B.V.
Terpenoids
1. Introduction
amine derivatives obtained by hydroamination [11] or telomerisa-
tion [12] and on odorous methyl ethers via direct hydroalkoxylation
In recent years, the pursuit of sustainable development has
started to attract attention to chemical products derived from
renewable, biological feedstocks. Current investigations are mainly
focused on the conversion of carbohydrates, fats and oils, and lignin
to develop a new biobased product platform [1–4]. Terpenes are
rarely taken into consideration in the ever-present discussion on
renewable resources, although they are a large and useful class of
hydrocarbons. Perfumes, fragrances, and resins are the most com-
mon applications of pure or crude terpenes, but they are also used
as starting materials in pharmaceutical syntheses [5–7]. Particu-
larly, the monoterpenes (C10 compounds) with their unsaturated
hydrocarbon skeleton are able to provide a chemistry similar to the
well-established petrochemistry. Unlike other described renew-
able resources, laborious defunctionalisations of monoterpenes are
not necessary. Interesting examples on catalytic conversions of ter-
penes to high-value fine chemicals were recently reviewed by Swift
and Monteiro [8,9].
[13]. In this contribution, we report on the codimerisation of
myrcene with methyl acrylate leading to odorous C13 esters.
The codimerisation of 1,3-dienes with acrylic compounds is a
feasible strategy to synthesise longer-chained alkenes with func-
tional groups in one step. Numerous review articles covering
different aspects of dimerisation and codimerisation reactions have
been published [14–19].
Codimerisation reactions of 1,3-dienes (butadiene, isoprene,
and myrcene) with acrylates can be performed by means of a
classic Ziegler-Natta catalyst system containing a complex of Ni,
Fe, or Co, an organometallic compound from a main group metal
(mostly AlEt3), and a ligand (usually PPh3) [20–26]. For example,
Klein et al. performed the first codimerisation of myrcene with an
acrylic compound using Ni(acac)2 and Et2AlOEt as a catalyst system
[25]. However, the yield and regioselectivity were only poor. Later,
Zakharkin and Petrushkina increased the yield to 23% in the pres-
ence of catalytic amounts of Co(acac)2 and Et3Al [26]. Additionally,
Grenouillet et al. reported on the codimerisation of isoprene with
methyl acrylate catalysed by a cationic Pd/phosphine complex
[27]. Despite the high selectivity towards one codimer, Diels–Alder
adducts were significantly obtained. In recent years, homogeneous
catalysts from ruthenium have proven to be capable of various
codimerisation reactions [28–30]. The working group of Mitsudo
reported on the particular versatility and activity of Ru(0) com-
plexes, e.g. (1,5-cyclooctadiene)(1,3,5-cyclooctatriene)ruthenium
or (ꢀ6-1,3,5-cyclooctatriene)bis(ꢀ2-dimethyl fumarate)ruthenium.
Our on-going investigations aim to use the industrially available
monoterpene myrcene (Fig. 1) as a renewable feedstock [10] and to
find further catalytic and atom-economical options for adding value
to the unsaturated hydrocarbon skeleton. Recently, we reported on
∗
Corresponding author. Tel.: +49 231 7552310; fax: +49 231 755 2311.
E-mail address: behr@bci.tu-dortmund.de (A. Behr).
0926-860X/$ – see front matter © 2013 Published by Elsevier B.V.
http://dx.doi.org/10.1016/j.apcata.2012.12.020

A. Behr et al. / Applied Catalysis A: General 453 (2013) 204–212
205
3. Results and discussion
3.1. Screening of different ruthenium catalysts
Fig. 1. Structure of myrcene.
Based on the literature results, investigations into the codimeri-
sation of myrcene with methyl acrylate were initiated with a
primary screening of different ruthenium catalysts. The screening
was performed under standard conditions with a catalyst con-
centration of 0.5 mol% and an equimolar ratio of myrcene/methyl
acrylate. Methanol was a very suitable solvent in the codimerisa-
tion reaction and was therefore selected as the solvent. An argon
pressure of 5 bar was applied to keep the volatile methyl acrylate
in solution. All primary screening experiments were carried out
in a 10 ml multiplex reactor developed by our group [38]. This set
of six parallel reactors allows efficient high-throughput screening
under inert and cost-effective conditions. As shown in Scheme 1,
different products and competing reactions were observed under
the investigated conditions.
Compounds 1–4 were the major products of the linear
codimerisation and were isolated and characterised via NMR
(Supplementary Data). The esters 1, 3, and 4 resulted from con-
jugated codimerisation where the myrcene molecule reacted as a
typical 1,3-diene (in the following, referred to as “1,4-codimers”).
Codimerisation at the tail of the diene unit led to the esters 1 and
4. Compound 3 was obtained from codimerisation at the head of
the diene unit. Obviously, the C C of methyl acrylate could remain
at the initial position, e.g. compound 4, or could switch to the new
connection, e.g. compound 1. The branched ester 2 resulted from
codimerisation in which one double bond of the diene unit did not
take part (in the following, referred to as “1,2-codimer”). The odour
of ester 2 can be described as green, fatty and heavy. A typical mix-
ture of esters 1 and 3 (70:30) smelled oily, soapy, and woody. These
olfactory characterisations were carried out in a 10% dilution in
dipropylene glycol after 4 h. Other codimers that were formed in
much lower amounts were verified by GC/MS coupling. As shown
in the following, the formation of the different codimers could be
well-controlled by the choice of the catalyst.
The homodimerisation of the involved coupling partners is
generally a competing reaction with codimerisation. Under the
investigated conditions, the dimerisation of two methyl acrylate
molecules did not occur. Nevertheless, in most cases, dimers of
myrcene (verified by GC/MS coupling) were observed at yields
up to 10%, which commonly resulted from thermal stress to the
highly reactive myrcene molecule. However, in this case, these
dimers were predominantly generated by the catalytically active
ruthenium complex. In contrast, the Diels–Alder reaction between
the conjugated 1,3-diene and the electron-poor dienophile methyl
acrylate was clearly the bigger challenge. For example, previous
experiments without any transition metal complex under the stan-
dard conditions described above showed that codimerisation did
not occur and Diels–Alder adducts 5 and 6 were generated with
yields up to 20%. The cyclic adducts 5 and 6 could be separately
synthesised under typical conditions for the Diels–Alder reaction
(110 ◦C, 10 mol% AlCl3, 6 h), isolated, and characterised via NMR
(Supplementary Data). Among these side reactions, the isomeri-
sation of myrcene (verified by GC/MS coupling) played a minor
role.
The primary screening showed that codimerisation of myrcene
and methyl acrylate could be performed with different ruthenium
catalysts (Table 1). The selection of catalysts was oriented towards
the results described in the literature and commercial availability.
The most active catalyst was Ru(cod)(2-methylallyl)2, which
gave a codimer yield of 42% after 4 h (entry 2). Additionally,
RuCl3 and the polymeric complex [Ru(cod)Cl2]n were catalytically
active in the codimerisation. Yields of 9% and 10%, respectively,
were obtained (entries 4 and 7). Obviously, these complexes
The codimerisation of dihydrofuranes [31], substituted alkynes
[32], isoprene [33], and N-vinylamides [34], with ␣,␤-unsaturated
esters could be performed with high yields and regioselectivi-
ties. The reduction of a ruthenium(III) complex with zinc also
creates an electron-rich ruthenium(0) species in situ which is
able to catalyse the codimerisation of 2-norbornene and methyl
acrylate [35]. In contrast, Fujiwhara et al. used a ruthenium(II)
complex in the codimerisation of myrcene with vinyl acetate
[36]. The catalyst chloro(pentamethylcyclopentadienyl)(1,5-
cyclooctadiene)ruthenium(II)
resulted
in
very
high
regioselectivities (>95%) and yields up to 83%. An unusual
(Z)-isomer was preferentially obtained, which could be explained
by the strong steric hindrance of the ruthenium complex. Recently,
Hirano et al. described ruthenium-promoted stoichiometric and
catalytic chemoselective codimerisations between butadiene and
methyl acrylate via an oxidative coupling mechanism [37].
2. Experimental
2.1. Reagents and product analysis
All preparations and manipulations were performed under a
dry, oxygen-free argon atmosphere using standard Schlenk tech-
niques. All non-aqueous solvents used in this work were purchased
dryly from Acros Organics (Geel, Belgium). All were ≥99% pure.
Other chemicals were purchased from commercial suppliers and
were of the highest purity available. They were used as received
without further purification. Selected Ru catalysts were obtained
from Strem Chemicals Inc. (Kehl, Germany), from Umicore (Hanau,
Germany) and from Sigma–Aldrich (St. Louis, United States). Argon
gas (99.998%) was purchased from AIR LIQUIDE Deutschland GmbH
(Düsseldorf, Germany). 1H and 13C spectra were recorded on a
Bruker model DPX500 spectrometer at room temperature. 1H and
13C NMR chemical shifts were reported on the ı-scale (ppm) rel-
ative to Me4Si as an external standard. The assignment of carbon
atoms was based on 2D experiments. Routine gas chromatographic
analyses were done on a HP 6890 instrument (Hewlett-Packard
GmbH, Waldbronn, Germany) equipped with a FI-detector and
an HP5 capillary column (30 m, diameter 0.25 mm, film thickness
0.25 ␮m) connected to an autosampler. GC–MS analyses were car-
ried out on a Hewlett-Packard 5973 (70 eV).
2.2. General procedure for codimerisation experiments
Reactions were performed in a homemade multiplex reac-
tor [38]. In
a typical run, Ru(cod)(2-methylallyl)2 (7.2 mg,
0.0225 mmol) was dissolved in anhydrous methanol (6.3 g, 8.0 ml).
The mixture was treated with ultrasound for 10 min. Myrcene
(204.3 mg, 1.50 mmol) and methyl acrylate (129.1 mg, 1.50 mmol)
were added. The reaction mixture was transferred into the evacu-
ated autoclave. The reactor was pressurised with 5 bar argon. The
reaction setup was placed in an oil bath at a temperature of 80 ◦C
and a magnetic stirrer was accelerated to 1200 rpm. After 4 h, the
reaction was stopped by reducing the stirring velocity and rapidly
cooling to room temperature using an ice bath. A sample was taken
and analysed by gas chromatography with di-n-butyl ether as an
internal standard and isopropanol as an additional solvent.

206
A. Behr et al. / Applied Catalysis A: General 453 (2013) 204–212
Scheme 1. Desired products and observed side reactions in the codimerisation reaction of myrcene with methyl acrylate.
predominantly led to the 1,4-codimers 1, 3, and 4. In contrast, the
1,2-codimer 2 was exclusively generated with a yield of 12% in
the presence of [RuCp*(MeCN)3]PF6 (entry 5). In most cases, the
yield of the Diels–Alder adducts 5 and 6 remained nearly constant
at 13–17%. The ruthenium chloride compounds served as Lewis
acids, so that in this case, the yield of the undesired cyclic adducts
was increased to 23–25% (entries 4 and 7). Preliminary selection
criteria for active ruthenium catalysts can be taken from these
primary experiments: ruthenium complexes with an oxidation
number 2+ or 3+ were active (entries 2 and 7). In the presence
of the ruthenium(0) complex, no codimers were observed (entry
1). Obviously, only complexes with good leaving groups, e.g. 2-
methylallyl, 1,5-cyclooctadiene, and acetonitrile, were catalytically
active among the ruthenium(II) complexes (entries 2–5). Accord-
ingly, the ruthenium complex Ru(cod)(2-methylallyl)2 showed
a relative high activity, in particular compared to the polymeric
[Ru(cod)Cl2]n. Complexes having a weak leaving group and a bulky
ligand sphere, e.g. RuH2(CO)(PPh3)3, were not able to catalyse
the desired reaction under the investigated conditions. The three
most effective catalysts were individually optimised in terms
of yields and selectivities. In the following only selected results
are demonstrated. All optimisation experiments are described in
detail in the Supplementary Data.
3.2. Ru(cod)(2-methylallyl)2 as the catalyst
Table 2 shows results selected of our systematic optimisa-
tion experiments. First, the influence of different solvents using
methanol, tetrahydrofurane, acetonitrile and toluene was investi-
gated. A trend concerning the polarity and coordinating properties
of the solvents was not found. The codimerisation reaction was
favoured using methanol (entry 1). With the aim of increasing the
selectivity towards codimers, the influence of the catalyst concen-
tration was investigated. Higher catalyst concentrations resulted
in faster formation of codimers so that the majority of myrcene
molecules react to form codimers and less substrate should remain
to react in the pathway of a Diels–Alder reaction. The yield of
Diels–Alder adducts was actually low (3%) using a catalyst con-
centration of 1.5 mol% and a further increase in the amount of the
catalyst had only a slight influence on conversion and yield (entries
2 and 3). A variation of the substrate ratio showed that an increase
in the methyl acrylate percentage led to higher yields of Diels–Alder
adducts and to lower yields of the linear codimers (entries 1 and
4).
In the experiments using higher catalyst concentration, par-
ticularly 2.5 mol%, no trend could be found and the yields of the
Diels–Alder adducts were very low, with a maximum of 3% (entries
Table 1
Results of the primary screening.a
Entry
Catalystb
Xc [%]
Yields (side products) [%]
Yields (codimers) [%]
Isomers
Dimers
5 + 6
1
2
3
4
Others
˙
1
2
3
4
5
6
7
Ru3(CO)12
24
66
22
38
33
24
39
5
3
5
2
3
4
2
2
7
3
4
2
4
2
17
14
14
23
13
16
25
–
13
–
3
–
–
–
–
–
12
–
–
13
–
3
–
–
8
–
–
–
–
–
–
8
–
3
3
–
4
–
42
–
9
15
–
Ru(cod)(2–methylallyl)2
RuH2(CO)(PPh3)3
[Ru(cod)Cl2]n
[RuCp*(MeCN)3]PF6
Ru(acac)3
–
3
–
3
RuCl3
–
10
a
Conditions: 0.5 mol% catalyst, T = 100 ◦C, t = 4 h, c(myrcene) = 0.19 mol l−1, myrcene/methyl acrylate 1/1, 8 ml methanol, p(argon) = 5 bar.
cod = 1,5-cyclooctadiene, Cp* = pentamethylcyclopentadienyl, acac = acetylacetonate.
Conversion of myrcene.
b
c

A. Behr et al. / Applied Catalysis A: General 453 (2013) 204–212
207
Table 2
Influence of selected reaction parameters in the presence of Ru(cod)(2-methylallyl)2.a
Entry
T [◦C]
c [mol%]
Myr/acr ratio
Xb [%]
Yields (side products) [%]
Yields (codimers) [%]
Isomers
Dimers
5 + 6
1
2
3
4
Others
˙
1
2
3
4
5
6
7
100
100
100
100
100
100
80
0.50
1.50
2.50
0.50
1.50
2.50
2.50
1/1
1/1
1/1
1/4
1/4
1/4
1/1
66
98
98
90
98
98
97
3
6
2
3
–
–
1
7
10
14
8
10
19
10
14
3
2
45
6
1
13
32
36
10
34
36
39
–
–
–
–
–
–
–
13
19
13
10
22
20
27
8
6
2
7
8
2
9
8
22
29
7
22
20
10
42
79
80
34
82
78
85
1
a
Conditions: Ru(cod)(2-methylallyl)2, t = 4 h, c(myrcene) = 0.19 mol l−1, 8 ml methanol, p(argon) = 5 bar.
Conversion of myrcene.
b
3 and 6). Obviously, the codimerisation reaction was completed
combination of both important parameters, a high amount of
catalyst (2.5 mol%) and a high acrylate ratio (1/8) did not lead
to a better yield (entry 5). Finally, the effect of temperature was
verified. As expected, fewer Diels–Alder adducts were generated
at lower temperatures (entries 6 and 7). At a temperature of 60 ◦C,
the catalyst activity was still sufficient, so that codimers were
obtained with a yield of 58% (entry 7). Diels–Alder adducts were
not observed under these conditions.
very quickly at high catalyst concentrations and, as a consequence,
only a few Diels–Alder adducts could be generated. Thus, an
increase in the amount of acrylate had no influence on the forma-
tion of the cyclic adducts.
In all experiments, the sum of codimer yields remained nearly
constant at 78–82%. However, about 5–9% more myrcene dimers
were obtained in the presence of 2.5 mol% catalyst than when using
a lower amount of 1.5 mol%. Obviously, the catalyst system also
catalysed the dimerisation of myrcene, which appeared more and
more as a competing side reaction with increasing catalyst concen-
trations.
With the aim of effectively preventing the thermally induced
Diels–Alder and dimerization reactions, the reaction temperature
was reduced to 80 ◦C. The yield of Diels–Alder adducts remained
nearly constant at 1%, however, the myrcene dimer amount
decreased slightly (entries 3 and 7). Advantageously, the activity
of the catalyst system remained high and after 4 h a promising
codimer yield of 85% was obtained.
3.4. Comparison of both ruthenium complexes
It was shown so far that the codimerisation reaction of
myrcene and methyl acrylate could be successfully performed
with homogeneous ruthenium complexes. In particular, the ruthe-
nium complexes Ru(cod)(2-methylallyl)2 and [RuCp*(MeCN)3]PF6
showed promising catalytic activity. Interestingly, both catalysts
led to entirely different products. The different products and
the different influence of the amount of acrylate indicated that
the codimerisation reaction proceeded via two different reaction
mechanisms.
3.3. [RuCp*(MeCN)3]PF6 as the catalyst
The major codimer 1 generated by the catalyst Ru(cod)(2-
methylallyl)2 had a comparable structure as the codimer of
butadiene and methyl acrylate described by Mitsudo (Fig. 2) [33].
The analogy of the structures suggested that the complex
Ru(cod)(2-methylallyl)2 used in our investigations and Mitsudo’s
ruthenium complex catalyse the reaction via the same mechanism.
In his report, Mitsudo postulated a ruthenium hydride complex
as the catalytically active species in the codimerisation reaction
of butadiene and methyl acrylate, which is generated in the tau-
tomeric equilibrium of the cyclooctadiene moiety (Scheme 2, step
1). After the formation of a hydride complex, the butadiene coordi-
nates to the metal (steps 1 and 2). In the course of the catalyst cycle,
the addition of the metal hydride results in an anti-methyl-ꢁ-allyl-
ruthenium complex (step 3). The coupling partner methyl acrylate
is inserted in the sterically less hindered ruthenium–carbon bond
which yields a ruthenium-alkyl complex (step 4). ␤-hydride elim-
ination and isomerisation occur and the major product is liberated
(steps 5–7).
As an analogue to the catalyst Ru(cod)(2-methylallyl)2, the influ-
ence of the most important reaction parameters was investigated
in the presence of the catalyst [RuCp*(MeCN)3]PF6 (Table 3). Again,
the codimerisation reaction was favoured in the solvent methanol,
15% codimers were obtained (entry 1).
Despite the same solvent affinity, entirely different products
were obtained using [RuCp*(MeCN)3]PF6 as the catalyst. In this
case, the 1,2-codimer 2 was preferentially generated, which was
not observed with the Ru(cod)(2-methylallyl)2 catalyst system. For
a further comparison, the effect of the amount of acrylate was
investigated. Again, an increase in activity with higher amounts
of acrylate was observed. The yield of Diels–Alder adducts also
increased with increasing acrylate concentration; however, one
important distinction to the previous catalyst system attracted
attention: using [RuCp*(MeCN)3]PF6 as the catalyst, an increasing
amount of acrylate resulted in a higher codimer yield. Obviously,
the substrate ratio of 1/8 appeared to be optimal concerning the
codimer yield, as 66% codimers and 25% Diels–Alder adducts were
obtained (entry 2).
Furthermore, it was observed that significantly more codimers
were generated using 1.5 mol% [RuCp*(MeCN)3]PF6 as the cat-
alyst; the codimer yield increased from 15% to 85%, thus the
selectivity was increased from 45% to 87% (entry 3). Additionally,
the three-fold higher amount of catalyst reduced the Diels–Alder
adducts from 13% to 5%. At the same time, myrcene dimerisation
was also catalysed; however, the dimer yield increased only
moderately to 8%. Greater amounts of the catalyst (over 1.5 mol%)
did not result in higher codimer yields (entry 4). Overall, codimer
2 was still the major product in the codimerisation catalysed
by [RuCp*(MeCN)3]PF6, with a selectivity of 72%. The obvious
Ru(cod)(cot)
Ru(cod)(2-methylallyl)2
COOCH3
COOCH3
codimer
1
described by Mitsudo
Fig. 2. Comparison of the codimer described by Mitsudo and the major codimer
observed in the presence of Ru(cod)(2-methylallyl)2 [33].

208
A. Behr et al. / Applied Catalysis A: General 453 (2013) 204–212
Table 3
Influence of selected reaction parameters in the presence of [RuCp*(MeCN)3]PF6.a
Entry
T [◦C]
c [mol%]
Myr/acr ratio
Xb [%]
Yields (side products) [%]
Yields(codimers) [%]
Isomers
Dimers
5 + 6
1
2
3
4
Others
˙
1
2
3
4
5
6
7
100
100
100
100
100
80
0.50
0.50
1.50
2.50
2.50
2.50
2.50
1/1
1/8
1/1
1/1
1/8
1/1
1/1
33
91
98
98
98
98
69
3
2
–
1
1
–
–
2
3
8
7
7
13
25
5
6
5
–
–
–
–
–
–
–
12
53
71
62
63
66
51
–
–
1
1
1
1
–
–
–
–
–
–
–
–
3
13
13
21
21
17
7
15
66
85
84
85
84
58
10
11
4
–
60
a
Conditions: [RuCp*(MeCN)3]PF6, t = 4 h, c(myrcene) = 0.19 mol l−1, 8 ml methanol, p(argon) = 5 bar.
Conversion of myrcene.
b
[Ru0]
COOCH3
1
2
7
H-[RuII]
COOCH3
[Ru]-H
6
H-[RuII]
COOCH3
[Ru]-H
3
5
[RuII]
COOCH3
COOCH3
[Ru]
4
Scheme 2. Catalytic cycle of the codimerisation reaction using Ru(cod)(cot) as the catalyst postulated by Mitsudo [33].
Interestingly, the complex Ru(cod)(2-methylallyl)2 used in our
ruthenium-hydride mechanism. However, the structure of the
branched and non-conjugated codimer 2 was analogous to the
codimers described in Fujiwhara’s protocol (Fig. 3) [36].
investigations has methylallyl ligands and could be an intermediate
in the catalytic cycle described by Mitsudo as shown in Scheme 3.
The postulated codimer A which resulted from the reac-
tion of the methylallyl ligand and methyl acrylate could be
verified in trace amounts by means of GC/MS coupling using
2.5 mol% Ru(cod)(2-methylallyl)2 as the catalyst. Thus, the gen-
erated ruthenium-hydride complex can start the catalysis cycle
postulated by Mitsudo.
Fujiwhara postulated
a
reaction mechanism via
a
ruthena(IV)cycle in his protocol on the codimerisation reac-
tion of myrcene with vinyl acetate (Scheme 4). The postulation
was based on Itoh’s protocol for a dimerisation reaction of
1,3-dienes [39,40]. Fujiwhara transferred the mechanism to the
codimerisation reaction of myrcene and vinyl acetate in the pres-
ence of RuCp*Cl(cod) as the catalyst. In contrast to the previously
described mechanism, a branched and non-conjugated product
The formation of codimer 2 from the experiments using
[RuCp*(MeCN)3]PF6 as the catalyst can not be explained by the
COOCH3
COOCH3
COOCH3
[Ru]
[RuII]
Ru(cod)(2-methylallyl)2
[Ru]-H
H-[RuII]
start of
codimerisation reaction
COOCH3
A
Scheme 3. Reaction pathway of Ru(cod)(2-methylallyl)2 in the catalyst cycle described by Mitsudo.

A. Behr et al. / Applied Catalysis A: General 453 (2013) 204–212
209
RuII
Cl
R
+
1
OOCH3
R
R
RuII
OOCH3
Cl
2
H3COO
R
+
OOCH3
4
H
RuIV
H
RuIV
R
R
Cl
Cl
H3COO
H3COO
3
Scheme 4. Catalytic cycle of the codimerisation reaction using RuCp*Cl(cod) as the catalyst postulated by Fujiwhara [36].
is generated. In this case, the cyclooctadiene moiety appears as
leaving group and no vinyl-hydride complex is obtained. After the
leaving group has been replaced from the coordination sphere by
the involved starting compounds (step 1), oxidative ring closure
occurs (step 2). After ␤-hydride elimination, a ruthenium-alkyl
complex is formed which results in the product after a reductive
elimination (step 4). Fujiwhara explained the formation of the
major product mainly by the steric hindrance of the Cp* ligand and
the observed regioselectivity by electronic effects.
form a ruthenium-hydride species, e.g. Ru(cod)(2-methylallyl)2,
results in the formation of conjugated 1,4-codimers. In this catalytic
cycle, ruthenium has the oxidation number 2+. Branched and non-
conjugated codimers are obtained if a ruthenium(II) catalyst is used
whose moieties work as leaving groups, e.g. [RuCp*(MeCN)3]PF6.
The intermediary metal cycle contains ruthenium with an oxidation
number of +IV.
In both cases, the observed products differed from those
observed by Hirano in the codimerisation of butadiene and methyl
acrylate promoted by a ruthenium(0) complex. Obviously, dif-
ferent mechanistic pathways giving other products are possible
[37].
Obviously, the selection of the catalyst appeared to be an elegant
opportunity to control the complex selectivities in the codimeri-
sation reaction. The use of a ruthenium catalyst, which is able to
3.5. RuCl3 as the catalyst
[RuCp*(MeCN)3]PF6
RuCp*Cl(cod)
Another opportunity to create a ruthenium-hydride complex
was described by McKinney et al. in their investigations on
the dimerisation of methyl acrylate, where RuCl3 was used as
precursor and reduced in situ to zero-valent ruthenium by zinc
and methanol [41]. In the presence of an electron-poor alkene,
in this case methyl acrylate, the reduced metal is stabilised by
complexation. The generated complex exists in the tautomeric
equilibrium between a ꢁ- and a vinyl-hydride complex.
Based on this work, it was expected that in the presence of zinc
and methanol, RuCl3 would produce the same products as the cat-
alyst Ru(cod)(2-methylallyl)2 described above. Thus, the catalyst
costs could be significantly reduced by using the cheap ruthenium
OCOCH3
codimer described
by Fujiwhara
COOCH3
2
Fig. 3. Comparison between codimer described by Fujiwhara and the major codimer
2 observed in the presence of [RuCp*(MeCN)3]PF6 [36].

210
A. Behr et al. / Applied Catalysis A: General 453 (2013) 204–212
Table 4
Effect of zinc on codimer yield in the presence of RuCl3.a
Entry
Zinc [eq.]
Xb [%]
Yields (side products) [%]
Yields (codimers) [%]
Isomers
Dimers
5 + 6
1
2
3
4
Others
˙
c
1
2
3
4
–
39
98
98
98
2
6
4
4
2
2
14
16
25
5
6
3
5
30
29
–
–
–
–
3
4
14
13
–
1
2
2
4
3
28
28
10
13
74
72
–
3
10
6
a
Conditions: 2.5 mol% RuCl3, T = 100 ◦C, t = 4 h, c(myrcene) = 0.19 mol l−1, myrcene/methyl acrylate 1/1, 8 ml methanol, p(argon) = 5 bar.
Conversion of myrcene.
0.5 mol% RuCl3, experiment of the primary screening (cf. Table 1).
b
c
salt RuCl3. Indeed, a catalytic activity and the same 1,4-codimers
ligand, zero-valent ruthenium species, and methyl acrylate. More
importantly, a highly space-occupying ligand sphere does not seem
favourable.
have already been observed in the primary screening using RuCl3 as
the catalyst, even without zinc. As a consequence, the influence of
zinc on catalyst activity was investigated in more detail (Table 4).
Based on the previous results, a relative high amount of catalyst
(2.5 mol%) was used. Comparing entries 1 and 2, it can be seen that
codimerisation using RuCl3 as the catalyst without zinc could be
carried out only with codimer yields up to 13%, but the results were
not satisfactory in spite of a five-fold greater amount of catalyst.
In the presence of 2.5 mol% catalyst, the mass balance showed
that oligomers or polymers were generated, which were not
detectable by gas chromatography (entry 2). As expected, the addi-
tion of zinc had a significant influence on the catalyst activity, e.g.
3 eq. of the reduction agent caused an increase in the codimer
yield from 13% to 74% (entry 3). The codimer distribution remained
identical. Myrcene dimers were generated between 14% and 16%
after the addition of zinc, because the nearly ligand-free ruthenium
species obviously catalysed the linking of two myrcene molecules.
Further addition of zinc had no influence on the reaction. Obvi-
ously, the reduction of ruthenium plays a very important role in
the mechanism of the codimerisation reaction.
Table 5 shows the influence of selected reaction parameters. At
lower temperatures, the conversion and the codimer yield fell, e.g.
the yield was decreased by 42% when the temperature was low-
ered from 100 ◦C to 60 ◦C (entries 1 and 3). Nevertheless, fewer
Diels–Alder adducts were obtained, as expected. A comparison with
Ru(cod)(2-methylallyl)2 at 80 ◦C showed that the catalyst system
RuCl3·xH2O/Zn was less active (entry 2). Instead of 84%, only 47%
codimers were generated.
Obviously, the catalytically active complex was favourably gen-
erated at higher temperatures. This is another important indication
that the codimerisation reaction proceeded via an allyl complex.
The catalyst Ru(cod)(2-methylallyl)2 inherently contains an allyl
unit and can therefore initiate the catalyst cycle even at lower tem-
peratures.
Using higher catalyst amounts of 1.5 or 2.5 mol%, quantitative
conversions and codimer yields up to 72% were achieved (entries
5 and 6). Again, a lower amount of catalyst (0.5 mol%) did not
lead to an acceptable codimer yield (13%, entry 4). Thus, compa-
rable results to using Ru(cod)(2-methylallyl)2 as the catalyst were
obtained, but the catalyst system RuCl3·xH2O/Zn was entirely less
active.
As previously discussed, the product distribution of these cata-
lyst systems suggested two different reaction mechanisms. Under
the current conditions, the catalysis of RuCl3·xH2O proceeded via a
ruthenium-hydride mechanism yielding 1,4-codimers. In contrast,
branched and non-conjugated codimers were only approachable
with catalysts which are able to form a ruthena(IV)cycle. In their
protocol on the oligomerisation of butadiene, Itoh et al. could
generate a cationic and coordinatively unsaturated ruthenium
species by converting RuCp*(ꢀ4-1,3-butadiene)Cl with silver tri-
flate [39,40]. If the formation of the branched 1,2-codimer 2 is
mainly attributed to the formation of a ruthenacycle, a codimeri-
sation reaction should also be possible with simple ruthenium
chloride and silver triflate (Scheme 5). It seems entirely logical that
in the catalyst system containing zinc, the oxidation number 4+
can scarcely be achieved and therefore the formation of codimer 2
is very improbable.
The catalyst costs could be still reduced by using the hydrate
complex of RuCl3. However, the codimer yield was slightly
decreased (−6%). Nevertheless, we decided to use the hydrate com-
plex in the following experiments due to the lower costs.
The nearly ligand-free ruthenium species resulted in promising
catalytic activity in the codimerisation reaction; the possibil-
ity of controlling product formation was additionally inves-
tigated by the use of different phosphine ligands. Among
the monodentate triphenylphosphines (PPh3), the bidentate
phosphine ligands 1,2-bis(diphenylphosphino)ethane (DPPE) and
1,4-bis(diphenylphosphino)butane (DPPB) were added in different
ruthenium–phosphine ratios. In all experiments, the conversion
decreased when a phosphine ligand was added (see Supplementary
Data). Additionally, the use of a phosphine ligand did not improve
the selectivity towards codimers. Hence, the ligand sphere of the
catalyst has to be obviously configured for a sufficient codimerisa-
tion reaction in such a way that either a good leaving group for
the formation of a ruthenacycle or appropriate conditions for a
ruthenium-hydride complex are present, e.g. cyclooctadiene as the
Table 5
Influence of selected reaction parameters in the presence of RuCl3·xH2O/Zn.a
Entry
T [◦C]
c [mol%]
Xb [%]
Yields (side products) [%]
Yields (codimers) [%]
Isomers
Dimers
5 + 6
1
2
3
4
Others
˙
1
2
3
4
5
6
100
80
60
100
100
100
2.50
2.50
2.50
0.50
1.50
2.50
98
60
31
33
98
98
9
5
3
2
5
9
16
5
2
2
13
16
7
3
2
16
8
7
26
16
9
5
25
24
–
–
–
–
–
–
11
16
9
5
16
14
2
6
2
–
9
7
27
9
4
3
22
21
66
47
24
13
72
66
a
Conditions: RuCl3·xH2O, Ru/Zn 1/10, T = 100 ◦C, t = 4 h, c(myrcene) = 0.19 mol l−1, myrcene/methyl acrylate 1/1, 8 ml methanol, p(argon) = 5 bar.
b
Conversion of myrcene.

A. Behr et al. / Applied Catalysis A: General 453 (2013) 204–212
211
RuCl3·xH2O
Zn / MeOH
AgOTf
?
COOCH3
1
via ruthenium hydride
2
COOCH3
via ruthenacycle
[RuCp*(MeCN)3]PF6
Ru(cod)(2-methylallyl)2
Scheme 5. Possibilities for controlling product selectivities in the codimerisation reaction with RuCl3·xH2O as the catalyst.
Table 6
Effect of AgOTf on the product distribution in the presence of the RuCl3·xH2O/Zn catalyst system.a
Entry
AgOTf/Zn [eq.]
Xb [%]
Yields (side products) [%]
Yields (codimers) [%]
Isomers
Dimers
5 + 6
1
2
3
4
Others
˙
1
2
3
4
5
1/–
2/–
3/–
3/10
–/10
98
98
98
97
98
3
–
1
10
9
2
3
3
11
16
5
11
11
7
3
–
–
26
26
4
–
–
–
12
11
–
–
–
2
2
7
17
16
26
27
14
34
34
69
66
17
18
3
7
–
a
Conditions: 2.5 mol% RuCl3·xH2O, T = 100 ◦C, t = 4 h, c(myrcene) = 0.19 mol l−1, myrcene/methyl acrylate 1/1, 8 ml methanol, p(argon) = 5 bar.
b
Conversion of myrcene.
As consequence, it was attempted to convert RuCl3·xH2O in
concentration, and temperature were investigated. Only by means
of the correct catalyst selection it was possible to prepare par-
ticular codimers with high selectivity. Competing side reactions
were mainly the Diels–Alder reaction of the two coupling part-
ners and the homodimerisation of myrcene. Owing to the different
product distributions with these catalysts, the mechanism of the
codimerisation reaction was considered in more detail. In partic-
ular, the selection criteria for capable ruthenium complexes and
mechanistic considerations could be derived from comparisons
with codimerisation reactions described in the literature. In addi-
tion to the oxidation state of the central atom, the ligand is an
important guide for the mechanism which either leaves the com-
plex and opens up the possibility to form a ruthena(IV)cycle or
is directly involved in the formation of a ruthenium(II)-hydride
species. Finally, mechanistic investigations using cost-efficient
RuCl3·xH2O as the catalyst showed that the observed product dis-
tributions can be attributed predominantly to the mechanism of
the codimerisation reaction.
a cationic and coordinatively unsaturated ruthenium compound
with different amounts of silver triflate (Table 6). This cationic
complex should yield the branched, non-conjugated 1,2-codimer
2 according to the Ru(II)–Ru(IV) mechanism, instead of the 1,4-
codimers 1, 3, and 4 which resulted from the ruthenium-hydride
mechanism.
High myrcene conversions of 97% to 98% were observed in all
experiments. The mass balance showed that oligomers or polymers
were generated which were not detectable by gas chromatogra-
phy. After adding 1 equivalent of silver triflate, only a low yield
of 14% codimers was obtained (entry 1). The addition of another
equivalent actually led to the favoured formation of the branched
1,2-codimer 2 (entry 2), and further addition had no influence on
activity and product formation. As soon as 10 eq. zinc were added
to the system, 3 eq. silver triflate no longer had an influence and
the formation of the products resulted from the ruthenium-hydride
mechanism (entry 5).
The considerations of the reaction mechanisms were herewith
confirmed. It was found that a single coordinating chloride ligand
at the ruthenium(II) species did not impair the course of the reac-
tion. Furthermore, it can be seen that the pathway towards the
1,2-codimer 2 via the oxidation states 2+ and 4+ was not possible in
the presence of the reduction agent zinc. Obviously, product selec-
tivity towards codimer 2 does not predominantly depend on the
Cp* ligand, but is mainly determined by the reaction mechanism.
Nevertheless, the Cp* ligand plays a role in product selectivity.
Acknowledgements
The financial support from Deutsche Forschungsgemeinschaft
(DFG) is gratefully acknowledged. We are also very grateful to Umi-
core AG & Co. KG for supplying the catalysts.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/
j.apcata.2012.12.020.
4. Conclusions
In summary, we have presented the codimerisation of myrcene
and methyl acrylate, providing a direct approach to novel C13 esters
for the flavour and fragrance industry. In particular, homogeneous
catalysts containing ruthenium appeared to be very active. System-
atic optimisations concerning the solvent, substrate ratio, catalyst
References
[1] P. Gallezot, Green Chem. 9 (2007) 295–302.
[2] P. Gallezot, Catal. Today 121 (2007) 76–91.

212
A. Behr et al. / Applied Catalysis A: General 453 (2013) 204–212
[3] H. van Bekkum, in: M. Graziani, P. Fornasiero (Eds.), Renewable Resources and
Renewable Energy, CRC Press, London, 2007.
[23] H. Bönnemann, C. Grard, W. Kopp, W. Pump, K. Tanaka, G. Wilke, Angew. Chem.
85 (1973) 1024–1035.
[4] A. Behr, A.J. Vorholt, L. Johnen, Nachr. Chem. 57 (2009) 757–761.
[5] A.A. Newman, Chemistry of Terpenes and Terpenoids, Academic Press Inc.,
London, 1972.
[24] K.S. Feldman, K.C. Grega, J. Organomet. Chem. 381 (1990) 251–260.
[25] E. Klein, F. Thoemel, H. Struwe, P. Heimbach, H. Schenkluhn, Liebigs Ann. Chem.
2 (1976) 352–366.
[6] J. Verghese, Terpene Chemistry, Tata McGraw-Hill Publishing Company, New
Delhi, 1982.
[7] E. Breitmaier, Terpenes, Wiley-VCH, Weinheim, 2006.
[8] K.A.D. Swift, Top. Catal. 27 (2004) 143–155.
[26] L.I. Zakharkin, E.A. Petrushkina, Zh. Org. Khim. 20 (1984) 490–493.
[27] P. Grenouillet, D. Neibecker, I. Tkatchenko, J. Chem. Soc., Chem. Commun.
(1983) 542–543.
[28] S. Murahashi, Ruthenium in Organic Synthesis, Wiley-VCH, Weinheim, 2004.
[29] B.M. Trost, F.D. Toste, A.B. Pinkerton, Chem. Rev. 101 (2001) 2067–2096.
[30] T. Naota, H. Takaya, S. Murahashi, Chem. Rev. 98 (1998) 2599–2660.
[31] H. Tsujita, Y. Ura, K. Wada, T. Kondo, T. Mitsudo, Chem. Commun. (2005)
5100–5102.
[32] T. Mitsudo, S. Zhang, M. Nagao, Y. Watanabe, J. Chem. Soc., Chem. Commun.
(1991) 598–599.
[33] T. Mitsudo, S. Zhang, T. Kondo, Y. Watanabe, Tetrahedron Lett. 33 (1992)
341–344.
[9] J.L.F. Monteiro, C.O. Veloso, Top. Catal. 27 (2004) 169–180.
[10] A. Behr, L. Johnen, ChemSusChem 2 (2009) 1072–1095.
[11] A. Behr, L. Johnen, N. Rentmeister, Adv. Synth. Catal. 352 (2010) 2062–2072.
[12] A. Behr, L. Johnen, A.J. Vorholt, ChemCatChem 2 (2010) 1271–1277.
[13] A. Behr, L. Johnen, P. Neubert, Catal. Sci. Technol. 2 (2012) 88–92.
[14] S.M. Pillai, M. Ravindranathan, S. Sivaram, Chem. Rev. 86 (1986) 353–399.
[15] J. Skupinska, Chem. Rev. 91 (1991) 613–648.
[16] H. Olivier-Bourbigou, L. Saussine, H. van Bekkum, in: B. Cornils, W.A. Herrmann
(Eds.), Applied Homogeneous Catalysis with Organometallic Compounds, vol.
1, Wiley-VCH, New York, 2002, pp. 253–265.
[34] H. Tsujita, Y. Ura, S. Matsuki, K. Wada, T. Mitsudo, T. Kondo, Angew. Chem. 119
(2007) 5252–5255.
[17] H. Olivier-Bourbigou, A. Forestière, L. Saussine, L. Magna, F. Favre, F. Hughes,
Erdöl, Erdgas, Kohle 126 (2010) 97–102.
[35] Y. Ura, H. Tsujita, K. Wada, T. Kondo, T. Mitsudo, J. Org. Chem. 70 (2005)
6623–6628.
[18] W. Keim, A. Behr, M. Röper, in: G. Wilkinson (Ed.), Comprehen-
sive Organometallic Chemistry, vol. 8, Pergamon Press, Oxford, 1982,
pp. 371–462.
[19] G.L. Tembe, A.R. Bandyopadhyay, P.A. Ganeshpure, S. Satish, Catal. Rev. Sci. Eng.
38 (1996) 299–327.
[20] H. Singer, W. Umbach, M. Dohr, Synthesis (1971) 265–266.
[21] H. Singer, W. Umbach, M. Dohr, Synthesis (1972) 42–44.
[22] M.A. El Amrani, A. Mortreux, F. Petit, Tetrahedron Lett. 30 (1989)
6515–6516.
[36] M. Fujiwhara, T. Nishikawa, Y. Hori, Org. Lett. 1 (1999) 1635–1637.
[37] M. Hirano, Y. Arai, N. Komine, S. Komiya, Organometallics 29 (2010) 5741–5743.
[38] A. Behr, G. Henze, L. Johnen, S. Reyer, J. Mol. Catal. A: Chem. 287 (2008)
95–101.
[39] K. Masuda, N. Katsumasa, F. Takahiko, H. Nagashima, K. Itoh, J. Organomet.
Chem. 428 (1992) C21–C24.
[40] K. Itoh, K. Masuda, T. Fukahori, K. Nakano, K. Aoki, H. Nagashima,
Organometallics 13 (1994) 1020–1029.
[41] R.J. McKinney, M.C. Colton, Organometallics 5 (1986) 1080–1085.