Continuous-flow Heck synthesis of 4-methoxybiphenyl and methyl 4-methoxycinnamate in supercritical carbon dioxide expanded solvent solutions

Article abstract of DOI:10.3762/bjoc.9.325

The palladium metal catalysed Heck reaction of 4-iodoanisole with styrene or methyl acrylate has been studied in a continuous plug flow reactor (PFR) using supercritical carbon dioxide (scCO₂) as the solvent, with THF and methanol as modifiers.

Full text of DOI:10.3762/bjoc.9.325

Continuous-flow Heck synthesis of
4-methoxybiphenyl and methyl 4-methoxycinnamate
in supercritical carbon dioxide expanded
solvent solutions
Phei Li Lau1,2, Ray W. K. Allen1 and Peter Styring*1
Full Research Paper
Open Access
Address:
Beilstein J. Org. Chem. 2013, 9, 2886–2897.
1Department of Chemical & Biological Engineering, Sir Robert
Hadfield Building, The University of Sheffield, Sheffield S7 2GA,
United Kingdom and 2Department of Chemical and Environmental
Engineering, The University of Nottingham, Malaysia Campus, Jalan
Broga, 43500 Semenyih, Selangor Darul Shsan, Malaysia
doi:10.3762/bjoc.9.325
Received: 22 August 2013
Accepted: 20 November 2013
Published: 17 December 2013
Email:
This article is part of the Thematic Series "Chemistry in flow systems III".
Guest Editor: A. Kirschning
Peter Styring* - p.styring@sheffield.ac.uk
* Corresponding author
© 2013 Lau et al; licensee Beilstein-Institut.
License and terms: see end of document.
Keywords:
continuous flow; flow chemistry; Heck; palladium; supercritical carbon
dioxide
Abstract
The palladium metal catalysed Heck reaction of 4-iodoanisole with styrene or methyl acrylate has been studied in a continuous plug
flow reactor (PFR) using supercritical carbon dioxide (scCO2) as the solvent, with THF and methanol as modifiers. The catalyst
was 2% palladium on silica and the base was diisopropylethylamine due to its solubility in the reaction solvent. No phosphine
co-catalysts were used so the work-up procedure was simplified and the green credentials of the reaction were enhanced. The reac-
tions were studied as a function of temperature, pressure and flow rate and in the case of the reaction with styrene compared against
a standard, stirred autoclave reaction. Conversion was determined and, in the case of the reaction with styrene, the isomeric product
distribution was monitored by GC. In the case of the reaction with methyl acrylate the reactor was scaled from a 1.0 mm to 3.9 mm
internal diameter and the conversion and turnover frequency determined. The results show that the Heck reaction can be effectively
performed in scCO2 under continuous flow conditions with a palladium metal, phosphine-free catalyst, but care must be taken when
selecting the reaction temperature in order to ensure the appropriate isomer distribution is achieved. Higher reaction temperatures
were found to enhance formation of the branched terminal alkene isomer as opposed to the linear trans-isomer.
Introduction
The use of cross-coupling reactions between organometallic the past three decades. The development of these cross-coupling
reagents and organic halides as a straightforward method of methods has undoubtedly revolutionalised the protocols for the
carbon–carbon bond formation has gained much popularity over construction of natural products, building blocks for supra-
2886

Beilstein J. Org. Chem. 2013, 9, 2886–2897.
molecular chemistry, self-assembly of organic materials and rally this triggers the question about the possibility of using
polymers, and lead compounds in medicinal chemistry from continuous reactions, an issue which this paper addresses. For
simpler entities [1]. These include the Suzuki, Kumada, and continuous flow systems the catalyst stability is important as it
Heck reactions. Among these, the palladium-catalysed Heck must be used over numerous runs without catalysts degradation,
reaction [2] is one of the most explored families of such reac- metal leaching or reduction in activity through impurity deposi-
tions. In the Heck reactions, a vinylic hydrogen is replaced by a tion [12,15]. Virtually all forms of palladium have been used as
vinyl, aryl or benzyl group through reaction with a halide com- catalysts for the Heck reaction. A wide array of combinations
pound. This reaction has been recognised as an indispensably depending on the components of the catalytic system is
simple yet effective method for molecular structure elaboration. possible, for example the catalyst can present as nanoparticles,
The Heck reaction is a useful component of the chemist’s ligand-less, unsupported or supported among other forms [16].
toolkit as it allows carbon chain extension, with the addition of
an alkene motif, using catalytic C–C bond formation chemistry. Practically speaking, heterogeneous catalysts provide benefits
Traditionally, Heck reactions have been performed using homo- such as ease of separation and recovery from the reaction mix-
geneous palladium catalysts with added phosphines, either as ture. In general, heterogeneous catalysts tend to be more robust
free reagents, coordinated discrete ligands or built into the and stable under harsh reaction conditions, such as elevated
ligand system itself. In recent years there has been a drive to temperatures and can be easily stored over longer periods,
develop phosphine-free catalysts and systems to ease sep- which in turn provides the possibility of by-passing the use of
aration and reduce both cost and product contamination [3]. expensive, sensitive ligands and inert atmospheres [17]. The
Further efforts have focused on the development of heteroge- catalysts can be reused and recycled either as they are or after
neous catalysts or immobilised homogeneous catalysts, particu- some form of regeneration [17]. The types of heterogeneous
larly for use in continuous flow reactor systems [3-6].
catalyst can be further categorised by their support, with exam-
ples including polymers, dendrimers, mesoporous materials,
The yield and selectivity of Heck reactions are profoundly zeolites, metal oxides, and carbon.
influenced by a number of variables, such as the nature of the
base, solvent, catalyst and operating conditions. The addition of Reactions using palladium supported on mesoporous materials
base is necessary to neutralise the halide acid formed in the have been reported by numerous authors. Among them, Zhao et
reaction, which would otherwise become detrimental to the al. [18] studied the coupling of non-activated aryl iodides with
reaction system. A wide array of bases, organic and inorganic, different alkenes using Pd-MCM-41-NH2. Other examples
has been investigated. Among them, tertiary amines such as include Pd-FSM 16 (functionalised with pyridine-carboimine or
triethylamine (Et3N), tributylamine (Bu3N) and alkali salts like quinoline-carboimine) [19], oxime carbapalladacycle anchored
carbonates, acetates or phosphates play crucial roles [7]. Solu- on several mesoporous materials [20], and a palladium
bility of the base and its basicity in the appropriate solvents are bipiyridyl complex anchored on nanosized MCM-41 [21].
criteria that need to be considered.
Furthermore, a myriad of zeolites have been applied in Heck
reactions, with examples including: modernite, HY or beta
Traditionally, most Heck reactions are carried out in polar [22,23], montmorillonite [24], layered double hydroxides
aprotic solvents, such as N,N-dimethylformamide (DMF), (LDH) [25]. Heterogeneous catalysts also lend themselves
1-methyl-2-pyrrolidone (NMP) and N,N-dimethylacetamide particularly well to continuous processing. One such example is
(DMA), though other organic solvents, such as toluene, hexane, reported by Karbass et al. [26], who carried out continuous flow
methanol and ethanol have been utilised [6]. In addition, the reaction using Pd(0) derived from [Pd3(OAc)6] supported on
reactions have been carried out in biphasic mode using a mix- polymeric monoliths containing methylimidazole, an ionic-
ture of solvents, such as ethylene glycol and toluene [8]. The liquid (IL) moiety in near-critical ethanol. It is thought that the
drive to carry out reactions in environmentally benign media resins act as catalysts to release active palladium-species into
also led to studies being carried out in “green” (environmen- solution. It also appeared that the IL-like unit was capable of
tally friendly) solvents, such as aqueous media, some supercrit- capturing and stabilising the catalytically active soluble palla-
ical fluids (SCFs), ionic liquids and even “solventless” systems dium species that are generated during the reaction. The reac-
[9-12]. Pioneering work using SCFs as reaction media was tion system eliminated the requirement of an inert atmosphere.
carried out independently in the late 1990s [13,14] using a Reactions between iodobenzene and methyl acrylate achieved
palladium complex with fluorinated phosphine ligands. Most of 85% yield and 100% selectivity to trans-methyl cinnamate at
the published results utilising supercritical fluid media have 200 °C and 80 bar. The reaction system was then extended to
been based on homogeneously-catalysed stirred reactions, using acrylonitrile as the olefinic substituent, and achieved a yield of
custom-built constant volume high pressure autoclaves. Natu- 66% and 70% trans-selectivity under similar conditions [26].
2887

Beilstein J. Org. Chem. 2013, 9, 2886–2897.
Other examples of carbon–carbon cross-coupling reactions util- Results and Discussion
ising continuous flow systems were also reported. However, Two reactions were studied using continuous flow chemistry as
conventional organic solvents were required either during the shown in Scheme 1. In both cases 4-iodoanisole was used as the
reaction or for the extraction progress [3,4,26-31].
aryl halide. The alkene used was either styrene to yield the stil-
benes 1 or methyl acrylate to yield the cinnamate ester 2. For
The majority of the examples of catalysts discussed employ each reaction there are three possible regioisomers. The thermo-
phosphines as electron-donating ligands but this is undesirable dynamically favoured trans- (t) isomer is most common in stan-
as they are toxic, air-sensitive, and moisture-sensitive [32]. An dard organic solvents although under certain conditions the cis-
excess amount of phosphine and palladium are normally (c) isomer can be formed under kinetic control. The third possi-
required under the Heck reaction conditions as they are suscep- bility is the geminal or branched (b) isomer which is usually
tible to decomposition. Unfortunately, an excess of phosphine formed where a cationic intermediate forms that facilitates
reduces the reaction rate, while higher quantities of palladium migration of the σ-Pd–C bond from the terminal alkenic 1-pos-
result in potentially significant increased production costs [33]. ition to the 2-position. In the majority of publications, the for-
Palladium on a magnesium–lanthanum mixed oxide support mation of the branched isomer is either not observed or not
(Pd/MgLaO) was reported to be an efficient, thermally stable reported. Certainly for activated alkenes the branched isomer
and highly active heterogeneous catalyst in the Heck reactions rarely forms [12], however we have reported previously that it
of styrene and halides, including iodides, chlorine substituents, can be observed in both homogeneous and heterogeneously
and both activated and non-activated bromine [32]. It has been catalysed reactions and also under conditions of continuous
rationalised that the MgLaO support itself appears to act as a flow [5].
base, which in turn acted as an electron-donating group relative
Synthesis stilbenes under stirred tank and
to Pd, thereby eliminating the requirement of the phosphine
ligand. The catalyst can also aid the reaction without requiring continuous flow conditions
an inert atmosphere. The synthesis of 4-methoxystilbene (1) was carried out under
continuous flow conditions and also in a stirred autoclave
In this paper we report the Heck reaction between 4-iodo- reactor to compare the processes. In the autoclave reaction
anisole and either styrene or methyl acrylate in a continuous (Figure 1) a single composition was investigated corresponding
flow reactor. The solvent system used was a mixture of scCO2 to the concentrations in the continuous flow reaction set for a
and THF/methanol which contained the alkene and the organic scCO2/organic solution ratio of 5:1. The pressure of the auto-
base, DIPEA. In order to reduce the complexity of the system, clave was set as close to 200 bar as possible, however because
palladium metal supported on silica was used as catalyst and no no back pressure regulator was present this was achieved by
phosphines were added. While this was likely to reduce the setting the applied pressure at 167 bar then increasing the
activity of the system over homogeneously catalysed reactions temperature to 145 °C to attain reaction temperature and a pres-
using complexes with elaborate ligand design, it would also sure close to 200 bar. Because the autoclave is a sealed system
make purification simpler.
it was difficult to take samples on a regular basis. Therefore, the
Scheme 1: General Heck reaction showing the possible isomers that can be produced.
2888

Beilstein J. Org. Chem. 2013, 9, 2886–2897.
reactor was run for 24 hours and the sample then analysed after Continuous flow reactions
de-pressurising the system. A total conversion of 94% was All reactions were performed in 300 or 100 cm stainless steel
obtained (GC analysis relative to the limiting reagent 4-iodo- tubular plug flow reactors (PFRs) of 1 mm internal diameter. A
anisole), with 77% conversion to the t isomer and 17% to the b larger diameter (3.9 mm) PFR was also used as a comparison in
isomer. No c isomer was detected. This gave a t/g ratio of 4.5 to one example (Figure 2). Initial studies used flow rates
1. The remaining analysis corresponded to unreacted 4-iodo- commensurate with the concentration used in the autoclave
anisole.
reactor.
Figure 1: Schematic diagram of the stirred autoclave reactor. Pickel’s pump NWA PM101 was used to achieve supercritical pressure.
Figure 2: Schematic diagram of the continuous flow system. The reactor shown is the 3.9 mm i.d. PFR. For the 1 mm i.d. reactor this was a coil of the
300 cm or 100 cm tube according to the specific reaction.
2889

Beilstein J. Org. Chem. 2013, 9, 2886–2897.
Residence time in the continuous flow
reactors
were not collected over the initial 2 hour period to allow transit
of the fluids through the reactor system. The reactor was oper-
The volume of the unpacked 1 mm diameter reactors in the ated over an eight hour period taking the first sample at 3 hours
styrene and methyl acrylate reactions were 2.36 and 0.79 mL unless otherwise shown and then at regular intervals.
respectively. The reactors containing catalyst with packing were
found to have a porosity of 0.80, so neglecting packing In the initial flow studies, the scCO2 flow rate was set at
swelling, the total volumes available for flow for the 300 cm 0.1 mL min−1 and the organic solution flow rate at 0.02 mL
and 100 cm length reactors were 1.89 and 0.63 mL respectively. min−1. The temperature was controlled to ±0.1 °C using a
Unless otherwise stated, the scCO2 flow rate was 0.1 mL min−1 Memmert oven. Three reaction temperatures were chosen: 145,
while the organic flow rate was precisely varied over the range 155 and 165 °C. The first samples were analysed 3 hours after
shown in Table 1 to give the total flow rates shown. The range the organic flow was started to allow complete flushing of the
of possible residence times on each reactor are also given in complete reactor system with reagents. Samples were then
Table 1, based on the total flow rates and effective reactor taken at regular intervals over an eight hour period. Figure 3
volumes.
shows the total conversions obtained for combined t and b
isomers. No c isomer was observed in any of the reactions
carried out.
Table 1: Correlation of reaction flow rates to reactor residence times.
Flow rates, mL/min
Residence times, min
The first feature of note is that the initial conversion at all
temperatures is low. The system takes time to reach equilib-
rium as the catalyst becomes conditioned so therefore requires a
start-up procedure before reaching full conversion capacity.
300 cm
reactor
100 cm
reactor
scCO2
Organic
Total
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.033
0.025
0.02
0.133
0.125
0.12
14.2
15.1
15.7
16.1
16.5
16.7
17.0
4.7
5.0
5.2
5.4
5.5
5.6
5.7
At 145 °C, conversion of 47% is achieved with a t/b ratio of 6:1
although the system is far from running at equilibrium conver-
sion. Curve fitting shows that after 24 hours approximately 60%
conversion would be achieved. Therefore on comparable
timescale the continuous flow reactor gives a lower conversion
but with a slightly improved selectivity for the conversion to the
t isomer. Increasing the temperature to 155 °C resulted in 75%
0.017
0.014
0.013
0.011
0.117
0.114
0.113
0.111
Because of the piping pre- and post-reactor there was a volume conversion after 8 hours with a t/b ratio of 4:1. At 165 °C the
through which flow was possible but without reaction. There- conversion was 86% after 8 hours although the t/b ratio has
fore, after the commencement of organic solution flow, samples decreased to 3:1.
Figure 3: Total conversion of 4-iodoanisole as a function of reactor run time for three reaction temperatures.
2890

Beilstein J. Org. Chem. 2013, 9, 2886–2897.
The effect of overall flowrate through the reactor was deter- Increasing the pressure from 200 to 250 bar had a dramatic
mined by changing the organic flow rate while keeping that of effect on conversion. The continuous flow reaction was
the scCO2 constant at 0.1 mL min−1. Therefore, the concentra- repeated at 145 °C and flow rate of 0.12 mL min−1 but at
tion of the organic reagents increased with increased flow. The 250 bar scCO2 pressure. The results are shown in Figure 5.
temperature was maintained at 155 °C and the pressure at While the conversion increased to 48% at 200 bar with
200 bar throughout. Figure 4 shows the reaction profile for the continued use until the reaction was stopped, there was less than
rates studied. Optimum conversion of 76% was observed at 5% conversion in all cases when the pressure was increased to
0.12 mL min−1. At faster flows the conversion decreased with 250 bar. At this time we cannot account for this almost
the lowest conversion (1%) was observed for the fastest flow complete inhibition of the reaction at increased pressure so we
rate of 0.133 mL min−1. At slower flow rates the conversion are carrying out further studies to elucidate the precise mecha-
decreased away from the optimum value showing the process to nism. We are also investigating the effect of pressure over an
be a compromise between flow rate and reaction rate.
extended range to consider solutions at sub- and near-critical
Figure 4: Total conversion of 4-iodoanisole at 155 °C and 200 bar as a function of reactor run time for different combined flow rates and therefore
residence times. The values of the total flow rates given in the legend are mL min−1.
Figure 5: Total conversion of 4-iodoanisole as a function of reactor run time at 145 °C and at 200 (●) and 250 (ο) bar CO2 pressure.
2891

Beilstein J. Org. Chem. 2013, 9, 2886–2897.
pressures to determine the optimum conditions for conversion under the same operational conditions. The conversion reached
and selectivity in continuous flow systems. Furthermore, it was equilibrium after only 3.5 hours with a conversion of 4-iodo-
noted that while there was a dramatic change in conversion, the anisole of 90 ± 3%. Comparisons between methyl acrylate and
t/b ratio remained unchanged at 6:1.
styrene also shows that the latter requires a reaction run time of
about 6 hours to achieve comparable yields and conversion. The
increased reactivity of methyl acrylate over styrene can be
explained in terms of the increased resonance stabilisation of
Synthesis of methyl 4-methoxycinnamate
under continuous flow conditions
An activated alkene, methyl acrylate was reacted with 4-iodo- the intermediate through electron delocalisation over the ester
anisole under catalytic PFR conditions and the conversion and carbonyl motif.
product distribution monitored. A column containing the 2%
Pd/SiO2 catalyst with Chromosorb packing (1:3) was Another aspect of interest is the product selectivity in the reac-
constructed with dimensions 100 cm × 1 mm by cutting fresh tions involving 4-iodoanisole and methyl acrylate. The terminal
sections of the 300 cm tubes used for the reactions previously product, methyl trans-methoxycinnamate (2t) was obtained
described using styrene. Each reactor therefore contained exclusively, whereas in the system using styrene, a mixture of
15.7 μmol Pd dispersed throughout the volume. The pressure the trans-linear (1t) and terminal-branched (1b) isomers were
was maintained at 200 bar and 155 °C throughout and the total observed with the composition dependent on the system
flow rate was set to be 0.12 mL min−1 using a scCO2 flow rate temperature. This may be explained due to a steric effect [34].
of 0.1 mL min−1. The residence time on the reactor was calcu- In the case of methyl acrylate, proper alignment of the double
lated as 5.2 minutes. Including flow in the upstream and bond for migratory insertion is hindered by the presence of the
downstream pipework then the total time on the system was methyl group, thereby favouring the formation of the terminal
84 minutes so the first sample was taken after 90 minutes.
arylation product. The presence of the methyl group on methyl
acrylate might have a more profound impact on steric limita-
Figure 6 shows that reaction between 4-iodoanisole and methyl tion than the phenyl group of styrene. The results shown here
acrylate proceeded more rapidly than the reaction with styrene are consistent to published literature utilising 4-iodoanisole and
Figure 6: Conversion of 4-iodoanisole to methyl 4-methoxycinnamate (●) at 155 °C and 200 bar as a function of reactor run time. The open symbols
are the conversions to the different isomeric products for the reaction of 4-iodoanisole with styrene for comparison.
2892

Beilstein J. Org. Chem. 2013, 9, 2886–2897.
methyl acrylate in the presence of ruthenium- and platinum- with the packed catalyst, the total available volume for fluid
complex catalysts, in which complete regioselectivity towards flow in the 3.9 mm reactor was 0.88 mL with a residence time
the trans-isomer was achieved [19].
at a combined flow rate of 0.12 mL min−1 of 7.3 minutes.
Having obtained encouraging conversions in the 1 mm diam- Figure 7 shows that the yields obtained in the 1 mm PFR are
eter reactor, it was decided to scale the reaction out to a 3.9 mm slightly lower than those achieved in the 3.9 mm reactor,
internal diameter PFR. The 9 cm long reactor was packed with however the mass of catalyst was also lower. Furthermore, the
the catalyst and packing material and the reaction carried out 3.9 mm PFR gave complete conversion of the starting material
under identical reaction conditions to those used in the 1 mm after only 4 hours flow and this conversion was maintained
reactor. Accounting for the 0.80 porosity of the reactor filled while flow continued. In order to study the effects of the cata-
Figure 7: Comparison of conversion as a function of reactor run time for the reaction of methyl acrylate in the 1 mm × 300 cm (ο) and 3.9 mm × 9 cm
(●) plug flow reactors over the Pd-SiO2 catalyst.
Figure 8: TOF values for the 1 mm (ο) and 3.9 mm (●) PFRs at a total flow rate of 0.12 mL min−1, 155 °C and 200 bar for the reaction of 4-iodo-
anisole with methyl acrylate.
2893

Beilstein J. Org. Chem. 2013, 9, 2886–2897.
Figure 9: Turnover frequencies of the Heck reactions at 155 °C using styrene (ο) as combined isomeric products, and methyl acrylate (●) as a func-
tion of reactor run time for the 1 mm internal diameter plug flow reactors.
lyst amount, the turnover frequency (TOF) was also calculated temperature and pressure of the reaction and the flow rate of the
and this is shown in Figure 8. It is clear that within experi- reagents over the fixed bed catalyst. It was found that increasing
mental error that the TOFs in the two reactor configurations are the temperature of the reaction between 4-iodoanisole and
essentially identical which is encouraging in terms of the ability styrene resulted in an increased conversion but that the distribu-
to both scale up and scale out the reactions. The TOF is defined tion of isomers was affected with more of the branched,
as the number of moles of substrate converted per mol of cata- germinal addition isomer being produced at higher tempera-
lyst per unit time. For a continuous flow system this can be tures. In the case of reaction with methyl acrylate as the alkene,
defined in terms of the molar or volumetric flow rates (uM or only the single trans-substituted isomer was produced.
uV), the conversion as a mol fraction (x) and the number of mol Increasing the pressure from 200 to 250 bar in the PFR had the
of substrate at the start of the reaction (nS) and the number of dramatic effect of almost completely inhibiting the reaction.
mol of catalyst (nCAT) as shown in Equation 1.
The turnover frequency is, as expected, considerably higher
when methyl acrylate is used instead of styrene. However, there
is very little change in the TOF for methyl acrylate when the
internal diameter of the PFR was increased from 1.0 to 3.9 mm.
This is encouraging for the design of continuous flow reactor
(1)
The TOFs were also calculated for the reactions of styrene and systems as the TOF is essentially insensitive to diameter scaling
methyl acrylate with 4-iodoanisole in the 1 mm PFRs in order over the length scales studied and hence there are good
to normalise with respect to catalyst concentration as the prospects for scaling the flow to handle bigger substrate inven-
lengths of the reactor and therefore residence times for the same tories.
flow rate differed three-fold. The results are shown in Figure 9
Experimental
and emphasise the more reactive nature of the methyl acrylate
as a substrate. Indeed the methyl acrylate is almost four times Reagents and instruments
more reactive than styrene under identical conditions of All reagents were purchased from Sigma-Aldrich unless other-
temperature and pressure.
wise stated and were used as received without further purifica-
tion. The supported catalyst, 2% palladium on silica was
purchased from Johnson Matthey.
Conclusion
Supercritical carbon dioxide (scCO2) can be effectively used as
a reaction solvent for the palladium metal-catalysed Heck reac- Gas chromatography–mass spectra GC–MS analyses (EI and CI
tion between an aryl halide and an alkene under continuous modes) were performed using a PerkinElmer GC–MS with a
flow conditions in a plug flow reactor (PFR). The efficiency of 30 m × 0.25 mm × 0.25 μm Phenomenex-ZB5 column by Jane
the reaction depends on the nature of the alkene used, the Stanbra and Simon Thorpe in the Department of Chemistry, The
2894

Beilstein J. Org. Chem. 2013, 9, 2886–2897.
University of Sheffield. The oven temperature programme was (1 equivalent, 3.5105 g, 15 mmol), styrene (2 equivalents,
60–260 °C at 10 °C/min with a final temperature isothermal 3.1245 g, 30 mmol), N,N-diisopropylethylamine (DIPEA,
hold for 10 minutes, corresponding to a total run time of 2 equivalents, 5.23 mL, 30 mmol) and dissolved in THF
30 minutes. The MS mass limit was set at between 50 and (15 mL) and methanol (15 mL each) was prepared and stirred
450 Da. GC analyses were carried out using a Varian 3900 GC thoroughly until completely mixed. The ratio of organic to
fitted with an autosampler (AS 8400) and a flame ionisation scCO2 used in the autoclave corresponded to the 5:1 volumetric
detector (FID) on a 15 m × 0.25 mm × 0.25 μm CP-Sil 5CB ratio used in the continuous flow reactions. The tetrahydrofuran
Chrompak Capillary Column. The oven temperature and methanol were used to dissolve the organo iodide and act as
programme was set to ramp from an initial temperature of 50 °C a miscible, polar modifier to enhance the solvating properties of
(held for 1 minute) to 250 °C (held for 0.5 minutes) at a rate of the scCO2. The 60 mL autoclave was charged with 2.4 mL of
20 °C/minute, totalling to a run time of 11.5 minutes. The the reaction solution, corresponding to 4% of the autoclave
temperatures of the injector and the FID were set at 200 and volume. 2% Palladium on a silica support catalyst (20.4 mg,
250 °C respectively.
2.463 mmol) was then added and the autoclave sealed and
inserted into a pre-heated aluminium heating block (external
Inductively coupled plasma-atomic emission (ICP-AE) analyses temperature 145 °C, internal temperature 120 °C). After a
were performed on a Spectro Ciros CDD instrument (Spectro- period of 30 minutes the remaining reactor volume (57.6 mL)
Analytical, UK) by Alan Cox and Neil Bramall, Department of was pressurised with scCO2 at 167 bar using a Pickel’s pump
Chemistry, The University of Sheffield. The catalyst particles NWA PM-101 operating in constant pressure mode. The reac-
were rinsed with organic solvents (diethyl ether, THF and tion was stirred for 24 hours at a rate of 145 rpm. To work-up
methanol, 2 × 5 mL each), followed by distilled water the reaction mixture, the autoclave was allowed to cool to room
(2 × 5 mL) and dried under vacuum prior to analysis. During temperature and then slowly depressurised by venting the CO2.
the sample preparation step, the catalyst sample was quantita- The reaction mixture remaining in the autoclave was then
tively transferred to a clean (acid-washed) tube and 6 mL of dissolved in additional tetrahydrofuran, before extracting into
concentrated hydrochloric acid and 2 mL of concentrated nitric diethyl ether and washing with water. The resulting organic
acid were added. The sample was placed in a dri-block which mixture (top layer) was then withdrawn and dried over magne-
was thermostatically controlled with the temperature gradually sium sulfate then analysed by gas chromatography using a
raised to 150 °C. The sample was then maintained at this Varian 3900 GC with reference to standard solutions of trans-4-
temperature for 1 hour. After this digestion period, the sample methoxystilbene.
was left to cool and quantitatively transferred to a 50 mL tube
and topped up to 50 mL with 1% nitric acid. The element of Continuous flow reactions
interest (palladium) was measured against a calibrated sample Particles of the catalyst, 2% Pd on silica, (250 mg) thoroughly
prepared from known standards. The sample was introduced to mixed with the packing material (Chromosorb 750;
the instrument using a flow regulated pump. On reaching the 100–120 mesh; 750 mg) were packed into standard HPLC
nebuliser, the liquid sample met argon gas, generating a spray columns (length 300 cm, internal diameter 1 mm). The columns
of droplets. The heavy droplets were captured by the spray were laid out straight and a small pressure (1 bar) was applied
chamber walls and were pumped to a waste container. The fine to fill the material into the column. No vacuum was used. The
droplets (ca. 2% of the sample) passed through the spray catalysts were dispersed in a uniform manner through the whole
chamber and into the argon plasma (ca. 7000 °C) via the torch. length of the column, and finally the ends plugged with 25 μm
The resulting atoms and ions were excited and emitted light of stainless steel frits. For the reaction with methyl acrylate, the
characteristic wavelengths (Pd measured at 340.458 nm). The column was cut into 100 cm sections. The unpacked volume of
light was detected by charged coupled devices (CCD chips) the 300 cm long reactor was 2.36 mL and the 100 cm long
mounted on the spectrometer. The concentration of Pd was then reactor 0.79 mL. In all cases the columns were then coiled a
calculated using the instrument software. ICP–AE analysis gave small volume within the cavity of a Memmert oven. The setup
the concentration of palladium metal on the silica gel support as for the continuous flows is shown schematically in Figure 2.
0.02 g or 0.189 mmol Pd/g. This was later diluted with Chro-
mopac for use in the flow reactors at a ratio of 1:3 catalyst/ The reactor was connected to the fluid delivery system using
packing.
standard Swagelok fittings, and housed in a Memmert oven
with custom fitted inlet and outlet ports. Supercritical carbon
dioxide was generated by passing the feed from a CO2 cylinder
Stirred autoclave reactions
The autoclave reaction set-up is shown schematically in (BOC, liquefied at 15 °C, 50 bar pressure) through a Jasco
Figure 1. A standard reaction solution made from 4-iodoanisole PU-1580-CO2 supercritical pump. The system was first satu-
2895

Beilstein J. Org. Chem. 2013, 9, 2886–2897.
rated with CO2 at high flowrates (1 to 2 mL) until the desired particular Professor Martyn Poliakoff, FRS, Dr Jason Hydes
pressure was reached, followed by gradual reduction of the CO2 and Dr Ben Walsh for her supervision on site.
flowrate to the target value. The system was heated up when the
References
1. de Meijere, A.; Bräse, S. J. Organomet. Chem. 1999, 576, 88–110.
reactor was saturated with CO2 at 200 bar. When the desired
temperature was reached, the system was left to equilibrate for a
doi:10.1016/S0022-328X(98)01087-0
further period of 15 minutes. Pumping the organic reaction mix-
2. Heck, R. F.; Nolley, J. P., Jr. J. Org. Chem. 1972, 37, 2320–2322.
ture (4-iodoanisole, 15 mmol; styrene, 30 mmol; N,N-diiso-
doi:10.1021/jo00979a024
propylethylamine, 30 mmol; tetrahydrofuran, 15 mL; methanol,
3. Phan, N. T. S.; Brown, D. H.; Adams, H.; Spey, S. E.; Styring, P.
15 mL) was achieved using a Jasco PU-1580 HPLC pump set at
Dalton Trans. 2004, 1348–1357. doi:10.1039/b316553c
constant flow mode. The organic reaction mixture was pumped
at a variety of flowrates while maintaining constant scCO2
flowrate of 0.1 mL/min for all runs. The entire system pressure
was controlled by a Jasco PU-1580-81 back pressure regulator
(BPR). Samples were collected at regular intervals using vials
placed underneath the BPR outlet port. The CO2 evaporated
immediately on de-pressurisation leaving a solid residue
containing the mixture of products. The samples were then
analysed by gas chromatography (GC) and GC-mass spectrom-
etry (GC–MS) against known calibration standards of the
starting material and products. At the end of reaction, the
system tubing and reactor were flushed with CO2 and the
organic solvent mixture (THF and methanol) for a further
period of 6 hours to remove any products, byproducts and unre-
acted organic reagent.
4. Phan, N. T. S.; Brown, D. H.; Styring, P. Tetrahedron Lett. 2004, 45,
7915–7919. doi:10.1016/j.tetlet.2004.08.153
5. Phan, N. T. S.; Brown, D. H.; Styring, P. Green Chem. 2004, 10,
526–532. doi:10.1039/b405203j
6. Diederich, F.; Stang, P. J., Eds. Metal-Catalysed Cross-Coupling
Reactions; Wiley-VCH: Weinheim, 1998. doi:10.1002/9783527612222
7. Köhler, K.; Heidenreich, R. G.; Krauter, J. G. E.; Pietsch, M.
Chem.–Eur. J. 2002, 8, 622–631.
doi:10.1002/1521-3765(20020201)8:3<622::AID-CHEM622>3.0.CO;2-
0
8. Bhanage, B. M.; Zhao, F.-G.; Shirai, M.; Arai, M. Tetrahedron Lett.
1998, 39, 9509–9512. doi:10.1016/S0040-4039(98)02225-4
9. Mikami, K. Green Reaction Media in Organic Synthesis; Blackwell
Publishing: Oxford, 2005. doi:10.1002/9780470988770
10.Jessop, P. G.; Leitner, W., Eds. Chemical Synthesis Using
Supercritical Fluids; Wiley-VCH: Weinheim, 1999.
doi:10.1002/9783527613687
11.Leitner, W. Acc. Chem. Res. 2002, 35, 746–756.
doi:10.1021/ar010070q
A larger continuous flow reactor with 3.9 mm internal diameter
of length 9 cm was also used as a comparison to assess the
effects of scaling up the flow reactor. This was constructed
from stainless steel tubing with 25 μm steel frits at each end,
connected to the solvent delivery system using standard ¼ inch
Swagelok fittings. The reactor body was filled with the 1:3
(w/w) mixture of the 2% palladium on silica catalyst and Chro-
mosorb packing material. The unpacked volume of the reactor
was 1.075 mL.
12.Raynor, C. M. Org. Process Res. Dev. 2007, 11, 121–132.
doi:10.1021/op060165d
13.Carroll, M. A.; Holmes, A. B. Chem. Commun. 1998, 1395–1396.
doi:10.1039/a802235f
14.Morita, D. K.; Pesiri, D. R.; David, S. A.; Glaze, W. H.; Tumas, W.
Chem. Commun. 1998, 1397–1398. doi:10.1039/a802621a
15.Styring, P.; Grindon, C.; Fisher, C. M. Catal. Lett. 2001, 77, 219–225.
doi:10.1023/A:1013209202418
16.Alonso, F.; Beletskaya, I. P.; Yus, M. Tetrahedron 2005, 61,
11771–11835. doi:10.1016/j.tet.2005.08.054
17.Kohler, K.; Prockl, S. S.; Kleist, W. Curr. Org. Chem. 2006, 10,
1585–1601. doi:10.2174/138527206778249630
Sample analysis
18.Zhao, F.; Shirai, M.; Ikushima, Y.; Arai, M. J. Mol. Catal. A: Chem.
2002, 180, 211–219. doi:10.1016/S1381-1169(01)00436-8
19.Horniakova, J.; Raja, T.; Kubota, Y.; Sugi, Y. J. Mol. Catal. A: Chem.
2004, 217, 73–80. doi:10.1016/j.molcata.2004.02.019
20.Corma, A.; García, H.; Leyva, A. Tetrahedron 2004, 60, 8553–8560.
doi:10.1016/j.tet.2004.06.121
To work up the reaction, the raw sample was diluted with
diethyl ether (2 × 3 mL), then washed with an equivalent
volume of distilled water to dissolve the halide salt formed.
Diethyl ether (2 mL) was added to a GC vial and 5 drops of
solution from the organic top layer of the extraction mixture
was withdrawn using a Pasteur pipette and added into the GC
vial. The solution was thoroughly mixed and analysed by GC
and/or GC–MS.
21.Tsai, F.-Y.; Wu, C.-L.; Mou, C.-Y.; Chao, M.-C.; Lin, H.-P.; Liu, S.-T.
Tetrahedron Lett. 2004, 45, 7503–7506.
doi:10.1016/j.tetlet.2004.08.065
22.Dams, M.; Drijkoningen, L.; De Vos, D.; Jacobs, P. Chem. Commun.
2002, 1062–1063. doi:10.1039/b201180h
23.Corma, A.; García, H.; Leyva, A.; Primo, A. Appl. Catal., A 2003, 247,
41–49. doi:10.1016/S0926-860X(03)00060-7
Acknowledgements
We are grateful to the Engineering and Physical Sciences
Research Council (GR/R41668/01) and the University of
Sheffield (Fees) for the funding of a Ph.D. studentship to P.L.L.
We also thank the University of Nottingham School of Chem-
istry for hosting P.L.L. as part of the Ph.D. programme and in
24.Poyatos, M.; Márquez, F.; Peris, E.; Claver, C.; Fernandez, E.
New J. Chem. 2003, 27, 425–431. doi:10.1039/b204911b
25.Choudary, B. M.; Madhi, S.; Chowdari, N. S.; Kantam, M. L.;
Sreedhar, B. J. Am. Chem. Soc. 2002, 124, 14127–14136.
doi:10.1021/ja026975w
2896

Beilstein J. Org. Chem. 2013, 9, 2886–2897.
26.Karbass, N.; Sans, V.; Garcia-Verdugo, E.; Burguete, M. I.; Luis, S. V.
Chem. Commun. 2006, 3095–3097. doi:10.1039/b603224a
27.Phan, N. T. S.; Khan, J.; Styring, P. Tetrahedron 2005, 61,
12065–12073. doi:10.1016/j.tet.2005.07.109
28.Phan, N. T. S. Polymer-Supported Catalysts for Greener Synthesis.
Ph.D. Thesis, The University of Sheffield, 2005.
29.Lee, C. K. Y.; Holmes, A. B.; Ley, S. V.; McConvey, I. F.; Al-Duri, B.;
Leeke, G. A.; Santos, R. C. D.; Seville, J. P. K. Chem. Commun. 2005,
2175–2177. doi:10.1039/b418669a
30.Liu, S.; Fukuyama, T.; Sato, M.; Ryu, I. Org. Process Res. Dev. 2004,
8, 477–481. doi:10.1021/op034200h
31.Solodenko, W.; Wen, H.; Leue, S.; Stuhlmann, F.;
Sourkouni-Argirusi, G.; Jas, G.; Schönfeld, H.; Kunz, U.; Kirschning, A.
Eur. J. Org. Chem. 2004, 3601–3610. doi:10.1002/ejoc.200400194
32.Cwik, A.; Hell, Z.; Figueras, F. Adv. Synth. Catal. 2006, 348, 523–530.
doi:10.1002/adsc.200505225
33.Hillier, A. C.; Grasa, G. A.; Viciu, M. S.; Lee, H. M.; Yang, C.;
Nolan, S. P. J. Organomet. Chem. 2002, 653, 69–82.
doi:10.1016/S0022-328X(02)01154-3
34.Beletskaya, I. P.; Cheprakov, A. V. Chem. Rev. 2000, 100, 3009–3066.
doi:10.1021/cr9903048
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.9.325
2897