The development and evaluation of a continuous flow process for the lipase-mediated oxidation of alkenes

Article abstract of DOI:10.3762/bjoc.5.27

We report the use of an immobilised form of Candida antarctica lipase B, Novozym 435, in a preliminary investigation into the development of a continuous flow reactor capable of performing the chemo-enzymatic oxidation of alkenes in high yield and purity,

Full text of DOI:10.3762/bjoc.5.27

The development and evaluation of a continuous flow
process for the lipase-mediated oxidation of alkenes
Charlotte Wiles*,1,2, Marcus J. Hammond1 and Paul Watts*,1
Full Research Paper
Open Access
Address:
Beilstein Journal of Organic Chemistry 2009, 5, No. 27.
1Department of Chemistry, The University of Hull, Cottingham Road,
Hull, HU6 7RX, UK and 2Chemtrix BV, Burgemeester Lemmensstraat
358, 6163JT, Geleen, The Netherlands
doi:10.3762/bjoc.5.27
Received: 27 February 2009
Accepted: 13 May 2009
Published: 02 June 2009
Email:
Charlotte Wiles* - c.wiles@chemtrix.com; Paul Watts* -
p.watts@hull.ac.uk
Guest Editor: A. Kirschning
* Corresponding author
© 2009 Wiles et al; licensee Beilstein-Institut.
License and terms: see end of document.
Keywords:
Candida antarctica lipase B; continuous flow; epoxidation; hydrogen
peroxide; lipase; micro reactor; Novozym® 435; peracids
Abstract
We report the use of an immobilised form of Candida antarctica lipase B, Novozym® 435, in a preliminary investigation into the
development of a continuous flow reactor capable of performing the chemo-enzymatic oxidation of alkenes in high yield and purity,
utilising the commercially available oxidant hydrogen peroxide (100 volumes). Initial investigations focussed on the lipase-medi-
ated oxidation of 1-methylcyclohexene, with the optimised reaction conditions subsequently employed for the epoxidation of an
array of aromatic and aliphatic alkenes in 97.6 to 99.5% yield and quantitative purity.
Introduction
In addition to their synthetic value as intermediates in the large scale. Many of the epoxides produced industrially are
preparation of diols, alcohols, hydroxyesters and alkenes, epox- synthesised using the chlorohydrin method, or via in situ gener-
ides are a key raw material in many industrial processes, finding ated peracids derived from formic acid or acetic acid (1)/
application in adhesives, polymers, coatings and paints [1,2], hydrogen peroxide (2). As H2O2 (2) is itself not sufficiently
with some epoxides even exhibiting biological activity [3,4]. As electrophilic to epoxidise a non-conjugated double bond
such, over the years, convenient and efficient techniques have directly, its use in the formation of a peracid has afforded a
been sought for their preparation.
route to the epoxidation of alkenes in the presence of a strong
mineral acid [10,11]. Along with the numerous safety issues
Within the research laboratory, epoxidation of alkenes is that these methods present, the generation of large quantities of
achieved using organic peroxides and metal catalysts [5] or chlorinated by-products and acidic waste make these
peracids [6] such as m-chloroperbenzoic acid (m-CPBA) [7,8]. approaches undesirable; not only from a disposal stand point,
The hazardous nature of these techniques and the potential to but also with respect to product quality, with reduced chemose-
hydrolyse the epoxide [9] however precludes their use on a lectivity observed in strongly acidic media.
Page 1 of 12
(page number not for citation purposes)

Beilstein Journal of Organic Chemistry 2009, 5, No. 27.
As a means of addressing these problems, the past 20 years has extended period of time (4 h) affording the target epoxides in
seen a series of authors investigate the use of a chemo- conversions ranging from 15% to quantitative depending on the
enzymatic process for the oxidation of olefins, based on the alkene employed.
enzymatic formation of peracids from carboxylic acids and
oxidants such as H2O2 (2) and urea–hydrogen peroxide (UHP, Törnvall et al. [20] subsequently conducted a detailed investiga-
3) [12].
tion into the source of enzyme deactivation, concluding that
high temperature (60 °C) and high concentrations of H2O2 (2, 6
For this transformation, the biocatalysts of choice are lipases to 12 M) resulted in rapid loss of enzyme activity, again illus-
(E.C. 3.1.1.3), which are a group of water soluble enzymes that trating that careful dosage of H2O2 (2) increased the opera-
catalyse the hydrolysis of lipid substrates, i.e. triglycerides and tional lifetime of the biocatalyst. The authors also noted that the
fats, in biological systems and are a subclass of the esterases presence of EtOH and urea resulted in moderate deactivation of
[13]. Changes in enzymatic behaviour observed when lipases the enzyme. In 2006, Olivo and co-workers [21] reported an
are employed in organic solvent systems have led to their use in environmentally benign method for alkene epoxidation via the
the synthesis of lipids, carbohydrate esters, amines and amides, chemo-enzymatic perhydrolysis of ethyl acetate using
along with finding industrial application in the synthesis of Novozym® 435 (4), immobilised Candida antarctica lipase B,
pharmaceuticals, fine chemicals and cosmetics [14]. In these and UHP (3). They utilised the high specificity of Candida
reactions, the lipases have been shown to exhibit excellent antarctica lipase B towards ester hydrolysis to generate acetic
regio- and stereo-selectivity, enabling a facile route to the acid (1) in situ and subsequently synthesise the peracid 5
synthesis of optically active compounds [15-17].
(Scheme 2). The authors found that compared to the use of aq
H2O2 (2), the use of UHP (3) enabled controlled release of
Chemo-enzymatic epoxidation provides an environmentally anhydrous H2O2 (2) and enabled the enzyme to be recycled
benign approach to the synthesis of epoxides, with Björkling with no observable effect on conversion over two catalytic
and co-workers [18,19] initially reporting the generation of cycles. Using this approach, Olivo et al. demonstrated the
epoxides under mild conditions using seven lipases. As Scheme synthesis of twelve epoxides with yields ranging from 73% for
1 illustrates, the biocatalytic epoxidation involves the lipase 1-hexene to quantitative for 1-methylcyclohexene (6) and reac-
catalysed formation of a peroxy acid, from a carboxylic acid tion times of 166 to 2 h respectively. Along with Björkling and
and an oxidant, such as H2O2 (2), which donates oxygen to the co-workers [18,19], the authors found the use of aq H2O2 (2)
double bond, affording the respective epoxide or oxirane and resulted in a reduction in conversion after two cycles (60%)
regenerating the carboxylic acid.
affording complete deactivation of the enzyme on the third
cycle; yields in the range of 73 to 85% were however obtained
upon adding the oxidant in aliquots over a period of 24 h.
Scheme 1: Illustration of the chemo-enzymatic epoxidation of an
alkene; involving the biocatalytic perhydrolysis of a carboxylic acid
followed by a peroxy acid epoxidation.
Initial investigations illustrated the highest conversions in
solvents such as hexane and toluene, with the lowest in dioxane
and acetonitrile, following the lipase trend where the biocata-
lysts typically perform better in water immiscible solvents.
Scheme 2: Illustration of the chemo-enzymatic epoxidation of
1-methylcyclohexene (6) to 1-methylcyclohexene oxide (7), using
EtOAc as both a reactant and a solvent.
Moderate deactivation of the enzyme was however observed, Although the chemo-enzymatic epoxidation of alkenes is an
attributed to the use of high concentrations of H2O2 (2). This attractive alternative to the hazardous reagents currently
was overcome by adding aliquots of the oxidant over an employed, finding use in the synthesis of the aroma linalool
Page 2 of 12
(page number not for citation purposes)

Beilstein Journal of Organic Chemistry 2009, 5, No. 27.
proposed that by conducting the lipase-mediated epoxidation
under continuous flow it would be possible to reduce the reac-
tion times required to oxidise synthetically useful alkenes,
whilst increasing the biocatalyst’s lifetime. With this in mind,
our initial investigation into the development of a continuous
process for the synthesis of epoxides focussed on the incorpora-
tion of immobilised Candida antarctica lipase B, Novozym®
435 (4), into a flow reactor and the use of urea–hydrogen
Scheme 3: Model reaction used to compare the continuous flow epox-
idation strategy with the conventional batch technique.
oxide and epoxidised soybean oil [16], the reaction times peroxide (3) as the oxidising agent.
employed for the transformations (< 166 h) and observed reduc-
tions in enzyme activity in the presence of H2O2 (2) do not In addition to reports made by Olivo and co-workers, UHP (3)
currently make the technique a practical alternative to tradi- [29] was selected as the oxidant as it is a cheap, commercially
tional techniques. With this in mind, we report herein the devel- available, source of anhydrous H2O2 (2) which, in addition to
opment and evaluation of continuous flow technique for the its use in the synthesis of epoxides [30,31], has found wide-
enzyme-mediated synthesis of epoxides, employing the immob- spread application in the conversion of nitriles to amides [32],
ilised Candida antarctica lipase B, Novozym® 435 (4). This aldehydes to acids [33], and sulfides to sulfones [34]. To enable
was recently found to be the most efficient biocatalyst with comparison of the method developed here with batch investiga-
respect to peracid formation, facilitating a series of oxidations tions previously conducted, the oxidation of 1-methylcyclo-
including alkenes [21] and ketones via the Baeyer–Villiger hexene (6) to 1-methylcyclohexene oxide (7) was selected as a
reaction [22] and retaining the broad substrate specificity of model reaction (Scheme 3) for investigation within the reaction
Candida antarctica lipase B.
set-up illustrated in Figure 1.
In the past decade micro reactors, and more generically Micro reactor set-up
continuous flow reactors, have been shown to offer many As Figure 1 depicts, the flow reactor consisted of a borosilicate
advantages to the synthetic chemist, such as reduced reactions glass capillary (3.0 mm (i.d.) × 3.6 cm (long) packed with 100
times, increased catalytic efficiency, increased product purity mg of Novozym® 435 (4), with reactions conducted by
and atom efficiency [23-26]. More recently, authors have begun pumping a pre-mixed solution of the alkene 6 and UHP (3)
to incorporate biocatalysts into flow reactors as a means of through the packed-bed using a syringe pump.
increasing productivity, biocatalyst lifetimes and exploring the
potential of employing immobilised enzymes in industrial The reaction products were collected in a sample vial
processes. Recent examples include the continuous flow containing EtOAc (1 ml) over a period of 10 min, and analysed
enantioselective acetylation of a series of racemic secondary by GC-MS in order to quantify the conversion of 1-methylcyc-
alcohols [27] and the continuous flow synthesis of alkyl esters lohexene (6) to 1-methylcyclohexene oxide (7). Using this
[28].
approach enabled rapid screening of the reaction conditions to
be conducted: if residual alkene 6 was detected the reaction was
simply repeated employing an increased reaction time, which
Results and Discussion
Combining our experience of micro reaction methodology with was achieved by reducing the flow rate of the reactant stream.
the observations made by Olivo and co-workers [21,22], it was Upon changing the reaction conditions, the system was allowed
Figure 1: Schematic of the reaction set-up used to evaluate the continuous flow chemo-enzymatic epoxidation of 1-methylcyclohexene (6) via the
perhydrolysis of acetic acid (1).
Page 3 of 12
(page number not for citation purposes)

Beilstein Journal of Organic Chemistry 2009, 5, No. 27.
to equilibrate for 30 min prior to sampling, thus ensuring equi- conducting the reactions under continuous flow, where at any
libration of the reactor. In addition, to ensure that meaningful one time the quantity of oxidant 2 in contact with the enzyme
data was obtained, all reaction conditions were monitored for a would be low, continued operation and efficient recycling
minimum of 5 samples (n) and the average result reported along would be feasible.
with the relative standard deviation (% RSD).
As the aim of the investigation was to develop a continuous
Evaluation of UHP as an oxidant
flow process for the synthesis of epoxides in high purity, it was
When adapting or developing reaction methodology for a decided at this stage to employ 5 μl min−1 as the minimum flow
continuous flow application, it is imperative to ensure that all rate and increase the reactant concentrations by a factor of four;
reactants, by-products and products remain soluble over the the 2:1 oxidant to alkene ratio was however maintained. With
course of the reaction, so as to prevent blockage formation these factors in mind, flow reactions were performed by
within the fluidic system which can lead to the build-up of pres- pumping a pre-mixed solution of 1-methylcyclohexene (6, 0.1
sure followed by leaking from any interconnections. M) and H2O2 (2, 0.2 M), in ethyl acetate, through the reactor
Consequently, due to the limited solubility of UHP (3) in those (0.10 g Novozym® 435 (4)) at a flow rate of 50 μl min−1 (0.52
organic solvents found to promote the epoxidation of alkenes min residence time). After a short equilibration time, 30 min,
(typically ethyl acetate and DCM) the use of conditions the reaction products were once again collected at ten minute
analogous to those reported within the literature (~0.3 M) were intervals and analysed by GC-MS in order to determine the
not feasible within the flow reactor. As such, an ethyl acetate percentage conversion of the alkene 6 to 1-methylcyclohexene
solution of reduced concentration [alkene 6 (2.5 × 10−2 M) and oxide (7). As Figure 2 illustrates, at room temperature a
UHP (3) (5.0 × 10−2 M)] was employed and initial investiga- moderate conversion of 10.1% was obtained, this was however
tions focussed on evaluating the effect of reaction time on the improved upon by systematically reducing the flow rate, which
epoxidation of 1-methylcyclohexene (6). Employing a flow rate had the effect of increasing the residence time from 0.52 to 5.20
of 10 μl min−1, hence a residence time of 2.6 min, it was grati- min and afforded an optimal conversion of 46.2% at a flow rate
fying to observe that under continuous flow conditions, of 5 μl min−1 (5.20 min).
enzymatic hydrolysis of ethyl acetate to acetic acid (1) followed
by formation of peracetic acid (5) and finally oxidation of
1-methylcyclohexene (6) had occurred, albeit in low conver-
sion (5.1%). Varying the flow rate of the reactant solution
through the packed bed between 0.1 and 50 μl min−1, at 27 °C,
enabled us to identify 0.1 μl min−1 (4.3 h residence time) as the
optimum flow rate for the synthesis of 1-methylcyclohexene
oxide (7, 89.0%).
Using such a low flow rate is however impractical as it would
only generate 6.0 × 10−3 mg h−1 of 1-methylcyclohexene oxide
(7), thus providing no practical advantage over a conventional,
stirred batch reaction. Although a facile approach to increasing
the throughput of flow reactions is to increase the concentration
of the reactant stream, in this case the limited solubility of UHP
Figure 2: Graph illustrating the effect of flow rate (hence residence
(3) precludes this as a viable option.
time) on the conversion of 1-methylcyclohexene (6) to 1-methylcyclo-
hexene oxide (7) at room temperature (27 °C); employing 2.0 equiv of
H2O2 (2).
H2O2 as an alternative oxidant
As a means of increasing the reactor throughput, an alternative
oxidant was sought. H2O2 (2) was initially considered as it To ensure that the results obtained were as a result of
presented the distinct advantage over UHP (3) that it is soluble employing a packed-bed reactor and to exclude the possibility
in the selected reaction solvent, ethyl acetate, enabling it to be that the reaction was continuing within the collection vial, as
employed at a higher concentration if required. Although batch only the hydrolysis and perhydrolysis steps were enzyme cata-
investigations had highlighted problems associated with the use lysed, five samples were analysed immediately and then
of H2O2 (2) in the chemoenzymatic epoxidation of alkenes, subjected to re-analysis 5 h later. Analogous results were
with it found to deactivate Candida antarctica lipase B [35] and obtained within experimental error (± 0.1%), confirming that
prevent recycling of the biocatalyst, it was proposed that by the reaction occurred only within the packed bed.
Page 4 of 12
(page number not for citation purposes)

Beilstein Journal of Organic Chemistry 2009, 5, No. 27.
Effect of reaction temperature on the oxida-
Although up to this point we had failed to oxidise the alkene 6
completely, the use of H2O2 (2) was found to be advantageous tion reaction
compared to UHP (3) as it removed the issues associated with Novozym® 435 (4) has been shown to retain activity when
reagent insolubility and enabled flow reactions to be conducted heated to 60 °C, however in the presence of H2O2 (2) the
at increased alkene concentrations, with product isolation enzyme has been reported to lose activity, with a half life of 3.9
simplified as the only by-product from H2O2 (2) is water. In h at 60 °C [6.0 M H2O2 (2)] [20]. It was therefore proposed that
addition, compared to an analogous batch reaction, it can be due to the short contact times between the enzyme and oxidant
seen that equivalent results can be obtained within the flow 2 within the flow reactor, denaturation of the enzyme by the
reactor with reaction time of 1.3 min (22.9%) compared to 5 h oxidant 2 would be avoided; consequently the effect of reactor
(23.4%) in batch and improved conversions attainable with a temperature on the epoxidation of 1-methylcyclohexene (6) was
reaction time of 5.2 min (46.2%). This approach is therefore investigated.
advantageous as it provides a facile means of continuously
adding the oxidant 2, therefore removing the need for the time When heating a flow reactor, it is important to consider the
consuming practise of adding aliquots of the oxidant 2 in order boiling point of the reactants and solvent system as any changes
to maintain the enzymes activity.
in viscosity, and even boiling, of the components can lead to
irreproducible residence times within the reactor; at 5 μl min−1
and 60 °C, 64.5 ± 14.5% (n = 5) 1-methylcyclohexene oxide (7)
Figure 3: Graph illustrating the effect of (a) flow rate and (b) residence time on the conversion of 1-methylcyclohexene (6) to 1-methylcyclohexene
oxide (7) at various reaction temperatures (27 to 70 °C).
Page 5 of 12
(page number not for citation purposes)

Beilstein Journal of Organic Chemistry 2009, 5, No. 27.
Figure 4: Graph illustrating the effect of oxidant stoichiometry on the conversion of 1-methylcyclohexene (6) to 1-methylcyclohexene oxide (7) at a
residence time of 2.6 min (70 °C).
was obtained. To ensure that a uniform residence time was favourably to the 46.2% 7 obtained prior to heating the biore-
obtained, a back-pressure regulator (100 psi) was inserted into actor; confirming enzyme activity was maintained.
the reaction set-up (between the packed-bed and the collection
Effect of oxidant concentration on enzyme
stability
tube), enabling heating of the reactor above the boiling point of
the solvent − in this case 77 °C. To supply heat to the biore-
actor, a commercially available HPLC column heater was
employed and the effect of reaction temperature on the rate of
epoxidation assessed (40 to 70 °C) over a range of flow rates.
Incorporation of a BPR into the reaction system reduces the
error associated with the process (± 0.1%) and is a facile means
of obtaining stable, plug flow at elevated reaction temperatures.
Although it was pleasing to observe stability of the Novozym®
435 (4) at various reaction temperatures, with literature
precedent reporting enzyme deactivation at an oxidant concen-
tration of 5.0 × 10−2 M the next step of the investigation was
aimed at evaluating the enzyme’s tolerance to H2O2 (2) in the
micro reactor. To achieve this, the concentration of alkene 6
was maintained at 0.1 M and the oxidant 2 concentration varied
Although the percentage of epoxide 7 synthesised was found to between 5 × 10−2 and 0.3 M (0.5 to 3.0 equiv with respect to
increase with reducing flow rate, when comparing the data alkene 6) and the reactions conducted at 10 μl min−1 (2.6 min)
obtained at room temperature with that collected at elevated and 70 °C, with samples taken after 1 and 3 h of continuous
temperatures (40 and 70 °C), it can be seen that the oxidation of operation. As Figure 4 illustrates, when employing 0.5 equiv of
1-methylcyclohexene (6) increased most notably with reaction oxidant 2 an average of 32.5% conversion was obtained. This
temperature; whereby quantitative conversion was obtained at rose to 42.0% when employing equimolar quantities of H2O2
70 °C, cf. 38.1% 7 at 27 °C (10 μl min−1) (Figure 3). Under the (2), whereas employing a slight excess afforded 65.3% conver-
optimised conditions stated, 1-methylcyclohexene oxide (7) was sion (1.5 equiv) and 2 equiv again affording quantitative
synthesised at a throughput of 6.7 mg h−1 from a single reactor conversion to 1-methylcyclohexene oxide (7). Interestingly
with post reactor processing consisting simply of an aqueous when a concentration of 2.5 M H2O2 (2) was employed, rapid
extraction to remove water and acetic acid (1) from the epoxide deactivation of the enzyme was observed, illustrated by a reduc-
7.
tion in conversion to 29.5% after 1 h and 7.6% after 3 h. To
confirm that the enzyme had undergone irreversible damage,
Having subjected the same portion of enzyme 4 to a prolonged the reactor was purged with ethyl acetate (2.5 ml) and the reac-
thermal regime (21 h), it was imperative to ensure that the tion utilising 2 equiv of H2O2 (2) repeated at 70 °C, whereby
enzyme had not lost activity; to evaluate this a flow reaction baseline formation of the epoxide was observed. Prior to
was performed using the initial reaction conditions of 5 μl conducting investigations at 0.3 M, the Novozym® 435 (4) was
min−1 at room temperature. Using this approach, 46.3% removed and a second portion of the biocatalyst packed into the
1-methylcyclohexene oxide (7) was obtained, which compared reactor (0.10 g), again however deactivation was observed, this
Page 6 of 12
(page number not for citation purposes)

Beilstein Journal of Organic Chemistry 2009, 5, No. 27.
Figure 5: Illustration of the enzyme 4 stability to H2O2 (2) for the conversion of 1-methylcyclohexene (6) to 1-methylcyclohexene oxide (7) (residence
time = 2.6 min) at 27 °C and 70 °C over a 24 h period.
time after only 1 h. Although deactivation of the enzyme was of 1-methylcyclohexane (6), attaining quantitative conversion to
observed at 2.5 and 3.0 M H2O2 (2), the biocatalyst 4 was found 7 with a reactor temperature of 70 °C and a reaction time of 2.6
to be tolerant to four times the oxidant concentration (0.2 M) min, the investigation was extended to evaluate the generality of
compared to stirred and shaken batch investigations (0.05 M), the methodology developed. Employing the aforementioned
illustrating another advantage associated with micro reaction conditions, the epoxidation of a series of alkenes was conducted
technology.
and the results obtained are summarised in Table 1.
Monitoring enzyme stability over an extended
period of operation
In the case of styrene (9), incomplete conversion was obtained
with a residence time of 2.6 min hence the flow rate was
Having confirmed that the optimum conditions for the synthesis reduced to 5 μl min−1 (5.2 min) whereby quantitative conver-
of 1-methylcyclohexene oxide (7) were a residence time of 2.6 sion to styrene oxide was obtained; continuous operation of the
min, a reactor temperature of 70 °C and an oxidant 2/alkene 6 reactor, followed by an aqueous extraction, afforded the target
ratio of 2:1, the stability of the reaction system was monitored compound in 99.2% isolate yield. Cyclohexene (10) was readily
over a period of 24 h using fresh biocatalyst (0.10 g 4). In order converted to cyclohexene-1,2-oxide at 70 °C, 2.6 min residence
to highlight any fluctuations in the system, along with demon- time, and was isolated in 97.6% yield. In addition to evaluation
strating continuous operation of analytically pure 7 under the of aromatic and aliphatic alkenes discussed, the epoxidation of
optimised conditions, the system was also operated at the sub- cis-stilbene (11) and trans-stilbene (12) was investigated
optimal temperature of 27 °C. Samples were collected every 10 (Scheme 4).
min for 7.5 h, then once an hour for 9 h, after an 8 h break in
sampling; for visual clarity, over the initial 7.5 h every other Due to the molecule’s symmetry, the configuration of the epox-
data point is plotted.
ides synthesised could not be characterised by 1H NMR spec-
troscopy; however the compounds did possess significantly
As Figure 5 illustrates, in both cases stable operation was different retention times (tR) when analysed by GC-MS; cis-1,2-
observed over the 24 h period, with no fluctuations in the purity diphenyloxirane (13, tR = 9.81 min) and trans-1,2-diphenylox-
of the epoxide 7 synthesised at 70 °C and 0.08% RSD observed irane (14, tR = 10.28). Consequently, GC-MS in conjunction
for the reaction conducted at room temperature.
with a synthetic standard of trans-1,2-diphenyloxirane (14) was
used to confirm the geometric isomer synthesised. In both
cases, the optimal reaction conditions were 70 °C and a resid-
Effect of substrate
Having demonstrated the ability to rapidly optimise the reac- ence time of 5.2 min, affording the cis-isomer 13 in 99.5% yield
tion conditions required for the chemo-enzymatic epoxidation as colourless oil and the trans-isomer 14 in 99.1% yield as a
Page 7 of 12
(page number not for citation purposes)

Beilstein Journal of Organic Chemistry 2009, 5, No. 27.
Table 1: Summary of the reaction conditions employed for the lipase-mediated epoxidation of an array of alkenes under continuous flow.
Alkene
Temperature (°C)
70
Residence Time (min)b Conversion (%)
Yield (%)b
99.1 (6.7)c
2.6 (2)
100.0
70
70
2.6
5.2 (33)
57.2
100.0
–
99.2 (3.6)
70
2.6 (40)
100.0
97.6 (5.9)
70
70
2.6
5.2
31.9
100.0
–
99.5 (5.9)
70
70
2.6
5.2
32.1
100.0
–
99.1 (5.9)
The number in parentheses represents athe reaction time required under batch conditions (h) [21], bthe isolated yield obtained after continuous opera-
tion of the reactor for 24 h and cthe throughput (mg h−1) using the optimised conditions.
white crystalline solid. In all cases, the epoxides were synthes- tion of Candida antarctica lipase B was observed with 25% less
ised in excellent yield and purity with no hydrolysis products oxidant 2. Consequently, the oxidation of styrene (9) can be
observed.
performed at a throughput of 3.6 mg h−1 compared to 0.1 mg
h−1 in batch. Furthermore, the continuous reactor capacity can
In comparison to the reaction times employed in analogous be increased by employing additional packed-bed reactors,
batch reactions, Table 1, the examples reported herein serve to operated under identical reaction conditions, compared to batch
illustrate the dramatic reductions in reaction time attainable where increases in reactor volume/capacity will result in the
through the use of micro reaction technology; in the case of the need for longer dosing times for the oxidant 2.
epoxidation of styrene (9), a 381-fold reduction in reaction time
is obtained by employing H2O2 (2) under continuous flow. In Conclusion
addition to the reduced reaction times, the continuous flow In conclusion, we have demonstrated the successful transfer of a
methodology also enables larger quantities of oxidant 2 to be chemo-enzymatic batch process to a packed-bed flow reactor,
used, compared to conventional batch systems where deactiva- enabling a dramatic reduction in reaction time to be obtained. In
addition, the use of a flow reactor enabled higher quantities of
H2O2 (2) to be employed over prolonged periods of time (> 24
h) without deactivation of the biocatalyst. Using this approach
has enabled the development of a robust and operationally
simple technique for the chemo-enzymatic epoxidation of
olefins.
In the examples reported herein, we focussed on the use of ethyl
acetate as it acted as both a solvent and reagent with in situ
hydrolysis providing a source of acetic acid (1), from which the
peracid 5 could be generated. It is however acknowledged that
other solvents may be required and future work will therefore
Scheme 4: Illustration of the reaction products obtained when
conducting the continuous flow epoxidation of cis-stilbene (11) and
trans-stilbene (12).
focus on the use of alternative solvents systems, employing a
catalytic quantity of a carboxylic acid additive to generate the
Page 8 of 12
(page number not for citation purposes)

Beilstein Journal of Organic Chemistry 2009, 5, No. 27.
respective peracid, enabling greater system flexibility, along containing stainless steel blocks (Jones Chromatography) and
with evaluating the methodology towards other synthetically the reactor temperature monitored using a thermocouple (206-
useful oxidations.
3750, RS). When heating the reactor in excess of 50 °C, reac-
tions were conducted under pressure, using a back-pressure
regulator cartridge (100 psi) housed within a stainless steel
holder (Upchurch Scientific, IDEX Corporation), to prevent
Experimental
Materials
1-Methylcyclohexene (6, 99%, Aldrich), styrene (9, Reagent- boiling of the reactants and solvent system.
Plus®, ≥ 99%, Aldrich), cis-stilbene (11, 96%, Aldrich), trans-
General procedure for the epoxidation of
stilbene (12, 96%, Aldrich), cyclohexene (10, 99%, Aldrich),
urea–hydrogen peroxide (3, Aldrich), hydrogen peroxide (2, olefins in batch
30%, 100 volume, Fisher Scientific), Novozym® 435 (4) To a solution of Novozym® 435 (4, 0.10 g) and UHP (3, 1.0
(Lipase acrylic resin from Candida antarctica, ≥ 10 000 U g−1, mmol) in ethyl acetate (5 ml), was added the alkene (0.5 mmol)
Nordisk), deuterated chloroform (+0.03% TMS, < 0.01% H2O, and the reaction mixture shaken on a carousel reactor at room
Euriso-top) and ethyl acetate (Reagent grade, Fisher Scientific). temperature. Upon completion of the reaction, as indicated by
In all cases, materials were used as received.
TLC, the reaction mixture was filtered through a plug of cotton
wool and the organic solvent evaporated prior to dissolution of
the residue in DCM (20 ml). The organic layer was then washed
Instrumentation
Unless otherwise stated, nuclear magnetic resonance (NMR) with an aq solution of saturated NaHCO3 (2 × 20 ml) to remove
spectra were obtained at room temperature as solutions in any acetic acid (1) and residual urea (8). The organic layer was
deuterated chloroform (CDCl3) using a Jeol GX400 spectro- then dried using MgSO4, filtered and concentrated in vacuo to
meter; in the case of known compounds, all spectra obtained afford the target epoxide. The reaction products were then
were consistent with the literature. The following abbreviations analysed by GC-MS, 1H and 13C NMR spectroscopy and the
are used to report NMR spectroscopic data; s = singlet, d = spectroscopic data compared with those reported within the
doublet, t = triplet, br s = broad singlet, q = quartet, dd = double literature.
doublet, dt = doublet of triplets, m = multiplet and C0 =
General procedure for the epoxidation of
quaternary carbon. Analysis of samples by Gas Chromato-
graphy-Mass Spectrometry (GC-MS) were performed using a olefins in batch using H2O2 (2)
Varian GC (CP-3800) coupled to a Varian MS (2100T) with a To a stirred solution of Novozym® 435 (4, 0.10 g) and alkene
CP-Sil 8 (30 m) column (Phenomenex, UK) and ultra high (0.5 mmol) in ethyl acetate (5 ml) was added H2O2 (2, 1.0
purity helium (99.9999%, Energas, UK) as the carrier gas. mmol) in aliquots of 0.1 mmol 15 min−1, dispensed from an
Reactions employing 1-methylcyclohexene (6), cyclohexene autopipettor. Upon completion of the reaction, as indicated by
(10), styrene (9), cis-stilbene (11) and trans-stilbene (12) were TLC, the reaction mixture was filtered through a plug of cotton
analysed using the following method; injector temperature 200 wool and the organic solvent evaporated prior to dissolution of
°C, carrier gas flow rate 1.00 ml min−1, oven temperature 50 °C the residue in DCM (20 ml). The organic layer was then dried
for 4 min then ramped to 150 °C at 30 °C min−1 and held at 150 using MgSO4, filtered and concentrated in vacuo to afford the
°C for 6.17 min, with a 2.5 min filament delay. Mass spectro- target epoxide. The reaction products were then analysed by
metry data was obtained using a Shimadzu QP5050A instru- GC-MS, 1H and 13C NMR spectroscopy and the spectroscopic
ment with an EI ionisation source. Batch reactions were data compared with those reported within the literature.
conducted using a carousel reactor/rotator (SB2, Stuart) to
General procedure for the optimisation of
reduce mechanical degradation of the immobilised enzyme.
Flow reactions were conducted using a displacement pump olefin epoxidation under continuous flow
(MD-1001, Bioanalytical Systems Inc.), capable of delivering To optimise a continuous flow epoxidation, a gas-tight syringe
three solutions at flow rates between 0.1 and 100 μl min−1 (2.5 ml) was filled with a solution containing the alkene under
(calibrated for a 1 ml gas-tight syringe) and solutions delivered investigation (0.1 M), oxidant (0.2 M) in ethyl acetate and
to the flow reactor using a 2.5, 10 or 20 ml gas-tight syringe connected to the Omnifit reactor cartridge using PTFE tubing
(Hamilton, UK). The packed-bed consisted of a borosilicate [150 μm (i.d.) × 10 cm (long)] and a PEEK luer connector
glass column (3.0 mm (i.d.) × 3.6 cm (long)) (Omnifit), the coupled to a 1/16″ HPLC connector. To the pre-weighed reactor
immobilised enzyme was held in place using 25 μm poly- was added Novozym® 435 (4, 0.10 g, 26 μl total volume) and
ethylene frits and fluidic connections made using commercially the reactor outlet formed from a length of PTFE tubing [150 μm
available PEEK and PTFE connectors (Supelco). The flow (i.d.) × 5 cm (long)] to enable efficient sampling. The reaction
reactor was heated using a HPLC programmable column heater, mixture was then pumped through the packed bed (5 to 50 μl
Page 9 of 12
(page number not for citation purposes)

Beilstein Journal of Organic Chemistry 2009, 5, No. 27.
min−1) for a period of 30 min, in order to equilibrate, prior to 37.9% (% RSD = 0.08) was obtained and at 70 °C, quantitative
collection of the reaction products every 10 min for a minimum conversion was obtained over the 24 h period.
of 50 min. The resulting reaction products were analysed by
GC-MS in order to determine the conversion of alkene to 1-Methylcyclohexene oxide (7)
epoxide. If residual alkene was detected, the reaction was A pre-mixed solution of 1-methylcyclohexene (6 ,0.1 M) and
repeated at a slower flow rate to increase the reaction time (0.52 30% H2O2 (2, 0.2 M) was pumped through the enzyme filled
to 5.2 min).
reactor, at a flow rate of 10 μl min−1 and a reaction temperature
of 70 °C, for a period of 24 h to afford 1-methylcyclohexene
To increase the rate of epoxidation, reactions were also oxide (7) as a colourless oil (0.160 g, 99.1% yield) after an
conducted at elevated reaction temperatures (27 to 70 °C), with aqueous extraction; 1H NMR (400 MHz, CDCl3) δ 1.49–1.68
heating achieved using a stainless steel column heater and the (m, 4H), 1.65 (s, 3H), 1.89–2.20 (m, 4H) and 2.91 (m, 1H); 13C
reaction temperature monitored using a thermocouple. When NMR (100 MHz, CDCl3) δ 17.6 (CH3), 19.9 (CH2), 20.1
employing reaction temperatures > 50 °C, a back-pressure regu- (CH2), 25.2 (CH2), 26.1 (CH2), 52.1 (CH) and 62.2 (C0); m/z
lator (100 psi) was incorporated into the system at the reactor (EI) 113 (M++1, 20%), 112 (10), 97 (77), 85 (15), 95 (100), 83
outlet in order to prevent boiling of the reactants, products and (15), 69 (15) and 55 (25); GC-MS tR = 5.26 min. The spectro-
solvent. Samples were again taken at intervals of 10 min and scopic data obtained for 1-methylcyclohexene oxide (7) was
analysed off-line by GC-MS to determine the percentage of consistent with that reported within the literature [36].
epoxide synthesised.
1,2-Epoxy-1-phenylethane (styrene oxide)
A pre-mixed solution containing styrene (9, 0.1 M) and 30%
General protocol used to synthesise epox-
ides under continuous flow
H2O2 (2, 0.2 M) in ethyl acetate was pumped through the
In order to deliver a constant stream of reactants through the enzyme filled reactor at a flow rate of 5 μl min−1 and a reaction
packed-bed, a 20 ml gas-tight syringe (10 ml) was filled with a temperature of 70 °C for a period of 24 h. The resulting reac-
pre-mixed solution containing the alkene under investigation tion products were concentrated in vacuo and the residue
(0.1 M) and 30% hydrogen peroxide (2, 0.2 M) in ethyl acetate. subjected to an aqueous extraction, affording 1,2-epoxy-1-
The reaction mixture was then pumped through the packed-bed, phenylethane as a colourless oil (0.086 g, 99.2% yield); 1H
containing Novozym® 435 (4, 0.10 g, 26 μl total volume), NMR (400 MHz, CDCl3) δ 2.75 (dd, J = 2.8, 5.6, 1H), 3.08 (dd,
under the previously optimised reaction conditions for 12 h. The J = 4.1, 5.6, 1H), 3.80 (dd, J = 2.8, 4.1, 1H) and 7.24–7.33 (m,
reaction mixture was then concentrated in vacuo to remove the 5H); 13C NMR (100 MHz, CDCl3) δ 51.0 (CH2), 51.1 (CH),
reaction solvent, prior to dissolution of the residue in DCM (20 125.3 (2 × CH), 128.0 (CH), 128.2 (2 × CH) and 137.5 (C); m/z
ml). The organic layer was then washed with an aq solution of (EI) 121 (M++1, 5%), 120 (20), 119 (50), 92 (25), 91 (100), 89
saturated NaHCO3 (2 × 20 ml) to remove any acetic acid and (40) and 77 (20); GC-MS tR = 5.37 min. The spectroscopic data
residual peracetic acid (5). The organic layer was then dried obtained for 1,2-epoxy-1-phenylethane was consistent with that
using MgSO4, filtered under suction and the filtrate concen- reported within the literature [37].
trated in vacuo to afford the target epoxide. The reaction
cis-1,2-Diphenyloxirane (cis-stilbene oxide,
product was then analysed by GC-MS, 1H and 13C NMR spec-
troscopy in order to characterise the epoxide and determine 13)
product purity.
A solution containing cis-stilbene (11, 0.1 M) and 30% H2O2
(2, 0.2 M) in ethyl acetate was pumped through the packed-bed
Enzyme stability
reactor (70 °C) at a flow rate of 5 μl min−1 and the reactor
To confirm the Novozym® 435 (4) was sufficiently stable to be effluent collected over a period of 24 h. The reaction products
used for the continuous flow synthesis of epoxides over were concentrated in vacuo and subjected to an aqueous extrac-
extended periods of operation, two experiments were conducted tion to afford cis-stilbene oxide (13) as a colourless oil (0.140 g,
over a period of 24 h. In both cases a 20 ml solution of 99.5% yield); 1H NMR (400 MHz, CDCl3) δ 4.31 (s, 2H),
1-methylcyclohexene (6, 0.1 M) and hydrogen peroxide (2, 0.2 7.09–7.21 (m, 10H); 13C NMR (100 MHz, CDCl3) δ 59.9 (2 ×
M) in ethyl acetate was pumped through the reactor at a flow CH), 125.9 (4 × CH), 127.2 (2 × CH), 127.5 (4 × CH) and 135.6
rate of 10 μl min−1, with one reaction performed at 27 °C and (2 × C); m/z (EI) 197 (M++1, 100%), 196 (75), 195 (76), 167
the other at 70 °C. Samples were taken every 10 min for 7.5 h (75), 152 (10), 105 (50), 89 (25) and 77 (10); GC-MS tR = 9.81
and then every 1 h for 9 h after a break in sampling of 8 h; oper- min. The spectroscopic data obtained for cis-1,2-diphenylox-
ation was maintained for these 8 h but no samples taken. At 27 irane (13) was consistent with that reported within the literature
°C, an average conversion to 1-methylcyclohexene oxide (7) of [38].
Page 10 of 12
(page number not for citation purposes)

Beilstein Journal of Organic Chemistry 2009, 5, No. 27.
trans-1,2-Diphenyloxirane (trans-stilbene
oxide, 14)
7. Vogel, A. I.; Furniss, B. S.; Hannaford, A. J.; Smith, P. W. G.;
Tatchell, A. R. Vogel’s Textbook of Practical Organic Chemistry, 5th
ed.; Longmann Scientific and Technical: New York, 1989; pp
1132–1135.
A pre-mixed solution of trans-stilbene (12, 0.1 M) and 30%
H2O2 (2, 0.2 M) was pumped through the enzyme filled reactor,
at a flow rate of 5 μl min−1 and a reaction temperature of 70 °C,
for a period of 24 h to afford trans-stilbene oxide (14) as a
white crystalline solid (0.140 g, 99.5% yield); 1H NMR (400
MHz, CDCl3) δ 3.87 (s, 2H) and 7.25–7.59 (m, 10H); 13C NMR
(100 MHz, CDCl3) δ 62.8 (CH), 125.5 (4 × CH), 128.3 (2 ×
CH), 128.5 (4 × CH) and 137.1 (2 × C); m/z (EI) 197 (M++1,
70%), 196 (75), 195 (73), 180 (65), 167 (100), 165 (50), 105
(25), 89 (20) and 77 (10); GC-MS tR = 10.28 min. The spectro-
scopic data obtained for trans-1,2-diphenyloxirane (14) was
consistent with that reported within the literature [39].
8. Fieser, L. F.; Fieser, M. Reagents for Organic Synthesis;
Wiley-Interscience: New York, 1967; Vol. 1, pp 135 ff.
9. Rüsch gen. Klaas, M.; Steffens, K.; Patett, N.
J. Mol. Catal. B: Enzym. 2002, 19, 499–505.
doi:10.1016/S1381-1177(02)00204-7
10.Häberlein, H.; Korbanka, H.; Nowy, G. Verfahren zur Herstellung von
Oxiranverbindungen. Ger. Offen. DE 2436817 A1, February 12, 1976.
11.Siegmeier, R. Verfahren zur Herstellung von Epoxiden. Ger. Patent
3723843 C2, April 20, 1989.
12.Piazza, G. J.; Foglia, T. A.; Nuñez, A.
Biotechnol. Lett. 2000, 22, 217–221. doi:10.1023/A:1005614427586
13.Martinelle, M.; Holmquist, M.; Hult, K.
Biochim. Biophys. Acta, Lipids Lipid Metab. 1995, 1258, 272–276.
doi:10.1016/0005-2760(95)00131-U
1,2-Epoxycyclohexane (cyclohexene oxide)
A pre-mixed solution of cyclohexene (10, 0.1 M) and 30%
H2O2 (2, 0.2 M) was pumped through the packed-bed reactor at
a flow rate of 5 μl min−1 and a reaction temperature of 70 °C for
a period of 24 h. The reaction products were concentrated in
vacuo and the residue subjected to an aqueous work up,
affording 1,2-epoxycyclohexane as a colourless oil (0.138 g,
97.6% yield); 1H NMR (400 MHz, CDCl3) δ 1.15–1.27 (m,
2H), 1.31–1.49 (m, 2H), 1.65–1.99 (m, 4H) and 3.11 (s, 2H);
13C NMR (100 MHz, CDCl3) δ 19.9 (2 × CH2), 25.1 (2 × CH2)
and 52.2 (2 × CH); m/z (EI) 99 (M++1, 5%), 98 (7), 97 (20), 83
(100), 81 (80), 69 (20) and 55 (25); GC-MS tR = 4.876 min.
The spectroscopic data obtained for 1,2-epoxycyclohexane was
consistent with that reported within the literature [40].
14.Hung, T.-C.; Giridhar, R.; Chiou, S.-H.; Wu, W.-T.
J. Mol. Catal. B: Enzym. 2003, 26, 69–78.
doi:10.1016/S1381-1177(03)00167-X
15.Bornscheuer, U. T.; Kazlauskas, R. J. Hydrolases in Organic
Synthesis: Regio- and Stereoselective Biotransformations, 2nd ed.;
Wiley VCH: Weinheim, 2006.
16.Chen, C.-S.; Sih, C. J. Angew. Chem., Int. Ed. 1989, 28, 695–707.
doi:10.1002/anie.198906951
17.Adelhorst, K.; Björkling, F.; Godtfredsen, S. E.; Kirk, O.
Synthesis 1990, 112–115. doi:10.1055/s-1990-26802
18.Björkling, F.; Godtfredsen, S. E.; Kirk, O.
J. Chem. Soc., Chem. Commun. 1990, 1301–1303.
doi:10.1039/C39900001301
19.Björkling, F.; Frykman, H.; Godtfredsen, S. E.; Kirk, O.
Tetrahedron 1992, 48, 4587–4592.
doi:10.1016/S0040-4020(01)81232-1
20.Törnvall, U.; Orellana-Coca, C.; Hatti-Kaul, R.; Adlercreutz, D.
Enzyme Microb. Technol. 2007, 40, 447–451.
Acknowledgments
doi:10.1016/j.enzmictec.2006.07.019
The authors would like to acknowledge the Department of
Chemistry at The University of Hull for financial support of the
final year undergraduate research project (M. J. H), the results
of which led to this investigation.
21.Ankudey, E. G.; Olivo, H. F.; Peeples, T. L.
Green Chem. 2006, 8, 923–926. doi:10.1039/b604984b
22.Ríos, M. Y.; Salazar, E.; Olivo, H. F. Green Chem. 2007, 9, 459–462.
doi:10.1039/b618175a
23.Yoshida, J.-i. Flash Chemistry: Fast Organic Synthesis in
Microsystems; Wiley: Chichester, U.K., 2008.
References
24.Wiles, C.; Watts, P. Eur. J. Org. Chem. 2008, 1655–1671.
doi:10.1002/ejoc.200701041
1. Gan, L. H.; Ooi, K. S.; Goh, S. H.; Gan, L. M.; Leong, Y. C.
Eur. Polym. J. 1995, 31, 719–724. doi:10.1016/0014-3057(95)00031-3
2. Witnauer, L. P.; Knight, H. B.; Palm, W. E.; Koos, R. E.; Ault, W. C.;
Swern, D. Ind. Eng. Chem. 1955, 47, 2304–2311.
25.Kirschning, A.; Solodenko, W.; Mennecke, K.
Chem.–Eur. J. 2006, 12, 5972–5990. doi:10.1002/chem.200600236
26.Mason, B. P.; Price, K. E.; Steinbacher, J. L.; Bogdan, A. R.;
McQuade, D. T. Chem. Rev. 2007, 107, 2300–2318.
doi:10.1021/cr050944c
doi:10.1021/ie50551a034
3. LeBlanc, R.; Dickson, J.; Brown, T.; Stewart, M.; Pati, H. N.;
VanDerveer, D.; Arman, H.; Harris, J.; Pennington, W.; Holt, H. L., Jr.;
Lee, M. Bioorg. Med. Chem. 2005, 13, 6025–6034.
27.Csajági, C.; Szatzker, G.; Tőke, E. R.; Ürge, L.; Darvas, F.; Poppe, L.
Tetrahedron: Asymmetry 2008, 19, 237–246.
doi:10.1016/j.bmc.2005.06.028
doi:10.1016/j.tetasy.2008.01.002
4. Bhat, B. A.; Puri, S. C.; Qurishi, M. A.; Dhar, K. L.; Qazi, G. N.
Synth. Commun. 2005, 35, 1135–1142. doi:10.1081/SCC-200054225
5. Sharpless, K. B.; Verhoeven, T. R. Aldrichimica Acta 1979, 12, 63–74.
6. Swern, D. Org. React. 1953, 7, 378–433.
28.Woodcock, L. L.; Wiles, C.; Greenway, G. M.; Watts, P.; Wells, A.;
Eyley, S. Biocatal. Biotransform. 2008, 26, 466–472.
doi:10.1080/10242420802456571
29.Taliansky, S. Synlett 2005, 12, 1962–1963. doi:10.1055/s-2005-871968
30.Cooper, M. S.; Heaney, H.; Newbold, A. J.; Sanderson, W. R.
Synlett 1990, 533–535. doi:10.1055/s-1990-21156
Page 11 of 12
(page number not for citation purposes)

Beilstein Journal of Organic Chemistry 2009, 5, No. 27.
31.Marigo, M.; Franzén, J.; Toulsen, T. B.; Zhuang, W.; Jørgensen, K. A.
J. Am. Chem. Soc. 2005, 127, 6964–6965. doi:10.1021/ja051808s
32.Balicki, R.; Kaczmarek, Ł. Synth. Commun. 1993, 23, 3149–3155.
doi:10.1080/00397919308011173
33.Heaney, H.; Newbold, A. J. Tetrahedron Lett. 2001, 42, 6607–6609.
doi:10.1016/S0040-4039(01)01332-6
34.Balicki, R. Synth. Commun. 1999, 29, 2235–2239.
doi:10.1080/00397919908086223
35.Moreira, M. A.; Bitencourt, T. B.; da Graça Nascimento, M.
Synth. Commun. 2005, 35, 2107–2114. doi:10.1081/SCC-200066705
36.Robinson, P. L.; Barry, C. N.; Kelly, J. W.; Evans, S. A., Jr.
J. Am. Chem. Soc. 1985, 107, 5210–5219. doi:10.1021/ja00304a030
37.Easton, N. R., Jr.; Anet, F. A. L.; Burns, P. A.; Foote, C. S.
J. Am. Chem. Soc. 1974, 96, 3945–3948. doi:10.1021/ja00819a035
38.Davies, S. G.; Whitham, G. H.
J. Chem. Soc., Perkin Trans. 2 1975, 861–863.
doi:10.1039/P29750000861
39.Aggarwal, V. K.; Ford, J. G.; Jones, R. V. H.; Fieldhouse, R.
Tetrahedron: Asymmetry 1998, 9, 1801–1807.
doi:10.1016/S0957-4166(98)00028-7
40.Murphy, A.; Dubois, G.; Stack, T. D. P.
J. Am. Chem. Soc. 2003, 125, 5250–5251. doi:10.1021/ja029962r
License and Terms
This is an Open Access article under the terms of the
Creative Commons Attribution License
(http://creativecommons.org/licenses/by/2.0), which
permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.
The license is subject to the Beilstein Journal of Organic
Chemistry terms and conditions:
(http://www.beilstein-journals.org/bjoc)
The definitive version of this article is the electronic one
which can be found at:
doi:10.3762/bjoc.5.27
Page 12 of 12
(page number not for citation purposes)