A general continuous flow method for palladium catalysed carbonylation reactions using single and multiple tube-in-tube gas-liquid microreactors

Article abstract of DOI:10.1002/ejoc.201402804

A series of continuous flow chemistry processes that facilitate the palladium-catalysed carbonylation of aryl and vinyl iodides and aryl bromides with a range of alkoxy, hydroxy and amino nucleophiles is reported. Harnessing a semipermeable Teflon AF-2400 Tube-in-Tube assembly, these reactors permit the controlled transport of carbon monoxide into solution at elevated pressure to generate homogeneous flow streams, avoiding some potential issues associated with segmented flow gas-liquid reactors. As the volume of pressurised gas contained within the device is low, the hazards associated with this are potentially mitigated relative to comparable batch processes. We also show how the incorporation of a second in-line gas-flow reactor allows for the sequential introduction of two gases (carbon monoxide and a gaseous nucleophile) into the reaction stream. A Teflon AF-2400 Tube-in-Tube reactor facilitates the use of carbon monoxide in continuous-flow C-C bond forming reactions providing esters, amides, carboxylic acids, lactones and lactams. Two reactors can be used in series to permit the sequential addition of two reactive gases (CO and Me₂NH) into the flow stream.

Full text of DOI:10.1002/ejoc.201402804

FULL PAPER
DOI: 10.1002/ejoc.201402804
A General Continuous Flow Method for Palladium Catalysed Carbonylation
Reactions Using Single and Multiple Tube-in-Tube Gas-Liquid Microreactors
Ulrike Gross,[a] Peter Koos,[a] Matthew O’Brien,*[a,b] Anastasios Polyzos,[a,c] and
Steven V. Ley[a]
Keywords: Flow chemistry / Gas-liquid microreactor / Palladium catalysis / Carbonylation / C–C bond formation
A series of continuous flow chemistry processes that facilitate
the palladium-catalysed carbonylation of aryl and vinyl iod-
ides and aryl bromides with a range of alkoxy, hydroxy and
amino nucleophiles is reported. Harnessing a semipermeable
flow gas-liquid reactors. As the volume of pressurised gas
contained within the device is low, the hazards associated
with this are potentially mitigated relative to comparable
batch processes. We also show how the incorporation of a
Teflon® AF-2400 Tube-in-Tube assembly, these reactors per- second in-line gas-flow reactor allows for the sequential in-
mit the controlled transport of carbon monoxide into solution
at elevated pressure to generate homogeneous flow streams,
avoiding some potential issues associated with segmented
troduction of two gases (carbon monoxide and a gaseous nu-
cleophile) into the reaction stream.
Introduction
Ubiquitous in synthesis chemistry, the carbonyl group
plays a key role in a vast array of important bond forming
processes.[1] Consequently, over many decades, a great deal
of research effort has been invested into devising efficient
routes to compounds containing this most versatile func-
tional group. Perhaps one of the most successful methods
is the transition-metal-catalysed reaction of aryl and vinyl
halides with carbon monoxide (CO) and a suitable nucleo-
phile (Scheme 1).[2] From a standpoint of chemical effi-
ciency, the process has a number of merits. Due to the many
industrial processes by which CO can be produced from
cheap precursors, this gas is readily available and extremely
inexpensive.[3] Indeed, as the gas can be furnished by pro-
cessing renewable biomass feedstocks, it will continue to be
readily available long after traditional fossil fuel precursors
have been depleted.[4] The use of a catalytic amount of tran-
sition metal (which can either be recycled directly or reco-
vered for reprocessing) also adds to the overall efficiency of
the transformation.
Scheme 1. General transition-metal-catalysed carbonyl insertion re-
action.
Perhaps one of the most appealing aspects of using a
gaseous reagent is the ease of product purification, as the
controlled addition or removal of the gas can be achieved
simply by controlling pressure (assuming the reaction prod-
ucts are non-gaseous). Notwithstanding these considera-
tions, there exists a widespread aversion to the use of CO
gas in research laboratory settings, and this is not without
justification. CO binds very strongly to haem iron, thus in-
terfering with oxygen transport in the body. This makes the
gas extremely toxic, it is harmful to human health at con-
centrations as low as 100 ppm.[5] To compound matters fur-
ther, the gas is colourless, tasteless and odourless and, as
such, a victim may not even be aware that they are being
exposed until it is too late. Also, for chemical reactions to
take place at a reasonable rate, the gas often has to be press-
[a] Whiffen Laboratory, Department of Chemistry, University of urised to achieve the concentration required. Pressurised
Cambridge,
containment itself leads to hazards associated with the en-
Lensfield Road, Cambridge, CB2 1EW, UK
ergy released (which is proportional to volume) in the event
of mechanical failure. Obviously, the toxicity and flamm-
ability of CO would exacerbate the severity of any acciden-
tal release. Of course, for specific processes carried out on
an industrial scale, many of these problems can be over-
come by investing time and resources in bespoke engineer-
ing solutions and this area has seen a great deal of innova-
[b] School of Physical and Geographical Sciences, Lennard-Jones
Building, Keele University,
Staffordshire, ST5 5BG, UK
E-mail: m.obrien@keele.ac.uk
http://www.keele.ac.uk/chemistry/
[c] CSIRO, Materials Science and Engineering,
Bayview Avenue, Clayton South, VIC 3169, Australia
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201402804.
6418
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 6418–6430

General Continuous Flow Method for Pd-Catalysed Carbonylations
tion. However, for the constantly changing environment of material to deliver ozone, hydrogen, carbon dioxide, oxy-
the research laboratory, there is a clear need for dynamic gen, ethylene, syngas, and ammonia into solutions in order
and flexible processes that facilitate the safe, practical and to carry out chemistry in continuous flow.[15] Recently,
scalable use of CO and other reactive gasses. Over the last other groups have also reported the use of similar gas-liquid
decade or so, Flow Chemistry has established itself as a flow reactors,[16] many of which are based on Teflon® AF-
new enabling technology for chemical synthesis.[6] The key 2400. Ryu and co-workers have used the material in carb-
principle of flow chemistry is that substrates and reagents onylation reactions, generating CO from the acid promoted
pass continuously through a relatively small reaction space. dehydration of formic acid.[17] Kappe and co-workers have
In this way, only a tiny portion of the total reaction mixture made use of the material’s permeability to diazomethane.[18]
is being processed at any instant. Compared with batch Kunz and co-workers have carried out continuous flow re-
mode processes, this can often provide a number of benefits. actions where formaldehyde and trifluoroacetic acid perme-
As reactions are scaled over time (or through parallelis- ated through a Teflon® AF-2400 membrane.[19] A Teflon®
ation) rather than over dimension, they can be scaled up AF tube-in-tube reactor was used by Kim and co-workers
without adversely affecting the safety profile.[7] This is espe- to oxygenate flow streams in a recent flow synthesis of qui-
cially important for reactions that generate hazardous inter- nazolinones.[20] A microfluidic variant has also been utilised
mediates or which require extremes of temperature or pres- by van den Broek and co-workers.[21] Recently, a mathemat-
sure. In addition, as the (often small) reaction dimensions ical treatment of the gas permeation of this material in flow
are kept constant, phase-transfer and heat-exchange phe- has been carried out by Jensen and Yang.[22] Despite dis-
nomena (which are strongly affected by factors such as sur- playing poorer chemical compatibility with a variety of or-
face-area/volume ratios) are invariant, leading to highly reli- ganic solvents,[23] PDMS (silicone) is also highly permeable
able and scalable processes.[8] Furthermore, flow chemistry to many gases and this property has recently been har-
naturally lends itself to inline purification using solid-sup- nessed by Kim and co-workers in a series of continuous
ported reagents for scavenging/phase-switching.[9] Given flow reactions.[24] Leadbeater and co-workers have reported
these observations, we sought to examine the use of flow using a commercial reactor very similar in design to our
chemistry methods in the context of CO mediated processes tube-in-tube reactor to carry out carbonylations and hydro-
in organic synthesis. An important consideration for any genations, although they did not specify the material used
gas-liquid flow process is the method of interfacial contact. for the gas-permeable membrane.[25]
The majority of examples of the use of reactive gases in
flow chemical processes have largely been based on bi-
Results and Discussion
phasic/segmented flow.[10] In order to increase the interfa-
In this work we describe in detail our investigations[26]
cial surface area, and thereby increase the availability of gas
into the use of Teflon® AF-2400 tube-in-tube flow reactors
to the liquid reaction phase, various approaches to mechan-
to facilitate palladium-catalysed carbonylation reactions
ical mixing have been developed.[11] As the reaction rate is
using a variety of hydroxy, alkoxy and amino nucleophiles.
strongly influenced by the concentration of the reactive gas
We also report on the development of an extended multi-
compound in the liquid phase, it is important to gain con-
gas system wherein two tube-in-tube devices connected in
trol of this crucial parameter. However, this is not always
series permit the sequential delivery of two gases to the flow
straightforward. Many types of flow regime are possible
stream, in this case carbon monoxide and the gaseous nu-
(e.g. annular-flow, slug-flow, segmented-flow etc.) de-
cleophile dimethylamine. The gas-liquid flow reactor used
pending on gas and liquid flow rates, as well as other fac-
in this work involves an inner Teflon® AF-2400 semiperme-
tors such as channel dimension. In addition, even within a
able membrane tube (1.0 mm OD, 0.8 mm ID) through
particular flow regime, the distribution of the gas and liquid
which passes the reaction solution (Figure 1). This inner
phases depend strongly on the morphology of the mixing
tubing is contained within an outer tubing (thick walled
device/junction.[12] The relationships between the many
1/8ЈЈ PTFE) which is connected, via a pressure regulator, to
variables are often non-linear. Whilst there have been sev-
the CO cylinder. In this way, the inner tube is surrounded
eral innovative engineering approaches to this challenge, we
by a pressurised jacket of CO gas. It is also possible to re-
sought to develop a straightforward and more easily con-
verse this arrangement and have the gas contained within
trolled system. In particular, we believed that a system cap-
the inner tube with the liquid flowing through the outer
able of rapidly and reliably generating homogeneous solu-
tube.[15g]
tions of gas in a controlled manner would be especially use-
ful, as its response to parameters such as flow-rate/pressure
might well be more linear and predictable than systems em-
ploying biphasic flow. To this end, we explored the idea
of using gas-permeable polymer membranes to achieve gas-
liquid contact in continuous flow synthesis. We have found
that the amorphous fluoropolymer Teflon® AF-2400,[13]
Figure 1. Schematic of the tube-in-tube gas-liquid flow reactor.
which is highly permeable to a range of gasses but has very
low permeability to liquids, performs this task extremely
[CAUTION: Carbon monoxide is extremely toxic and
well.[14] We have successfully used reactors based on this flammable. This reactor is a research prototype and has not
Eur. J. Org. Chem. 2014, 6418–6430
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6419

U. Gross, P. Koos, M. O’Brien, A. Polyzos, S. V. Ley
FULL PAPER
undergone full safety testing. The reactor was used in an sample loops using a Vapourtec R2+ unit as a pumping
efficient fumehood and was monitored at all times. Excess device although any suitable similar pumping unit can be
CO released was vented into the exhaust of the fumehood, used in the experiment. To keep constant pressure in the
the sash of which was kept down whenever possible. Com- system and to prevent out-gassing of the dissolved CO, a
mercial CO detectors (which were regularly tested) were 6.9 bar (100 psi) back-pressure regulator was installed at the
placed both inside and outside the fumehood at all times.]
end of the flow stream. This is vital to the proper operation
In the reactor (Figure 1), the Teflon® AF-2400 (1.0 mm of the system as it maintains homogeneous solutions of gas
OD, 0.8 mm ID) tubing and the outer PTFE tubing are upstream of the back-pressure regulator at all times, avoid-
connected at the ends by stainless steel T-pieces (Swagelok) ing segmented flow. A base was introduced via a second
which connect them to the liquid flow streams and the gas stream to obviate early reduction of PdII to Pd0 which can
regulator respectively, permitting the liquid and gas streams sometimes result in blockage of the valves on the pumping
to be individually controlled in terms of pressure and flow device. A T-piece was used to join both reagent streams.
rate. The dimensions of the inner and outer tubing are such Initial reactions were carried out at room temperature at
that there is only about 1.5 mL of pressurised CO per meter different CO pressures (3, 5, 7, 10 bar).
of tubing. As the potential energy stored in a container due
The product, methyl ester 2a was afforded in only very
to pressurisation is proportional to its volume (at a particu- low conversions (ca 10%), as determined from integration
1
lar pressure), minimising this volume diminishes the kinetic of peaks in the H NMR spectrum of the crude material.
energy that would be released in the event of mechanical Excess CO could be observed out-gassing downstream of
failure.
the back-pressure regulator, indicating that insufficient tem-
We began our study by evaluating the methoxycarb- perature and/or reaction time (rather than a lack of CO)
onylation of m-iodoanisole 1a (Scheme 2).
were responsible for the low conversion. Unfortunately, we
also experienced severe issues with (presumably) Pd0 pre-
cipitation in the tube-in-tube device itself, which necessi-
tated its repeated dismantling. It was clear that this arrange-
ment would not be suitable for our purposes. In order to
increase the temperature and reaction time, and to eliminate
blockages of the tube-in-tube device with precipitated Pd0,
we devised a different reactor configuration (SETUP 2),
whereby the gas is introduced via a third separate solvent
stream (Figure 3). Here, while the tube-in-tube device
(3.8 m length) still performs the task of enriching the flow
stream with CO, it does not come into contact with the
substrate and catalyst streams. These are pumped into the
system, brought together at a T-piece and then united with
a third gas rich solvent stream from the outlet of the tube-
in-tube device. A back-pressure regulator was installed after
the gas-liquid reactor to prevent out-gassing of the dis-
solved CO. The flow stream then passes into the heating
coils and then through an optional scavenger cartridge to
Scheme 2. Initial palladium-catalysed methoxycarbonylations of 1a
in flow.
Initial reactions were performed using (dppf)Pd(OAc)2
as a soluble catalyst as this has had been previously re-
ported[27] to be a highly reactive system for palladium-cata-
lysed carbonylation reactions. In addition, 5% of DMF co-
solvent was used as this was found to preclude the forma-
tion of any observable Pd0 precipitates as the reaction pro-
ceeded. The flow arrangement (SETUP 1) for the methoxy-
carbonylation of m-iodoanisole 1a is shown in Figure 2.
Figure 2. Schematic of the flow reactor setup using tube-in-tube
reactor (2.0 m of AF-2400 tubing) (SETUP 1).
In this reactor arrangement the entire reaction stream
directly passes through the gas-flow reactor (2.0 m of AF-
2400) where it becomes enriched with CO. The reagents
Figure 3. Schematic of the gas-flow reactor setup using a third
were introduced to the gas-flow reactor via 2 mL PEEK stream (SETUP 2).
6420
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 6418–6430

General Continuous Flow Method for Pd-Catalysed Carbonylations
remove impurities before passing finally through the ter- the Palladium. Pleasingly, we found that only a slight excess
minal back-pressure regulator. This alternative reactor of the expensive xantphos ligand, relative to Palladium, was
setup was first evaluated using the carbomethylation of m- necessary to maintain high conversion (entry 6). Reducing
iodotoluene 1b as a test reaction (Scheme 3).
the reaction temperature to 70 °C resulted in a lower con-
version (entry 5). A slight increase in conversion to 90%
could be achieved by increasing the CO pressure to 10 bar
(Table 1, entry 8). Additional ligand screening experiments
showed only a negligible difference in conversion when
using either xantphos or the dppf ligand.
Having established conditions for high conversion, we
wished to investigate the scope of the reaction. For safety
reasons, considering the extremely toxic nature of CO, we
opted to carry out these reactions at 7 bar initially, al-
though the tube-in-tube device had been used at higher
pressures with no sign of mechanical failure. As the connec-
tors and ferrules used are conventionally used to swage
compression joints between like materials (e.g. metal to
metal), it was not initially known whether they would be
suitable for use with the PTFE tubing used in the device.
When, after significant use, the connections had proven
robust, reactions were carried out at pressures up to 15 bar.
To avoid a large amount of CO being released in the event
of a mechanical failure, we took the precaution of only
using the main cylinder regulator to pressurise the connect-
ing hose periodically. This acted as a reservoir which main-
tained constant pressure for several hours. For added con-
venience, a small gas sample canister connected between the
cylinder and the reactor could be used as a reservoir in-
stead, allowing the reactor to be disconnected from the
main cylinder completely. We have since tested these reac-
tors with various gasses pressurised up to 35 bar for pro-
longed periods and they have proven robust under these
conditions, although we must stress again that these reac-
tors are research prototypes and have not been fully tested
for safety. Table 2 shows the results obtained using the con-
ditions from Table 1, entry 6 with a range of aryl/alkenyl
iodides. In general, high conversions and isolated yields
were obtained in reactions using electron-deficient aryl iod-
ides (Table 2, entries 3, 4, 7 and 8) which furnished the
product methyl esters with high conversions and high iso-
lated yields after chromatographic purification, yields for
the relatively electron rich aromatic iodides 1a and 1b were
somewhat diminished in comparison (Table 2, entries 1 and
2). Interestingly, the reaction of substrate 1e, which has a
ketone group in close-proximity to the iodine-bearing car-
bon (Table 2, entry 5) proceeded with only a very low con-
version.
Scheme 3. Palladium-catalysed methoxycarbonylation reaction of
aryl iodide 1b.
In this case, the experimental conditions were slightly ad-
justed from those used with the initial configuration, in or-
der to minimise the precipitation that had been encoun-
tered. After some experimentation, we found that a toluene/
MeOH/DMF solvent mixture significantly increased the
solubility of the Pd complexes. Using SETUP 2, substrate
1b and triethylamine were premixed and loaded into a 2 mL
PEEK sample loop. A second 2 mL loop was filled with a
solution of palladium acetate and the ligand xantphos. Fi-
nally, a polymer supported thiourea (Quadrapure TU) car-
tridge was inserted into the flow line in order to scavenge
palladium residues from the outlet stream. We considered
key variables in this process to include flow rate, CO pres-
sure, heating coil length and temperature. The results of a
series of experiments investigating these conditions are
shown in Table 1. The first experiment employed a CO pres-
sure of 7 bar, 2.5 mol-% PdII, 5% xanthphos, equal flow
rates (0.1 mL/min) for the three pumps with a 20 mL reac-
tor coil set to a temperature of 100 °C (Table 1, entry 1).
Although this combination of conditions afforded the prod-
uct in only a low 18% conversion, a significant jump in
conversion (to 73%) could be obtained by increasing the
ratio of the CO flow stream to the substrate/catalyst
streams from 1:1 to 6:1 (Table 1, entry 2), clearly indicating
that at these higher temperatures and longer reaction times,
the amount of CO is becoming limiting. By increasing the
reaction loop volume the residence time was further ex-
tended, further augmenting the conversion to 86% (entry
3). Initially, a 2:1 ratio of xantphos to PdII was used to
ensure that there was sufficient ligand available to complex
Table 1. Results for investigation of reaction conditions for the con-
version of 1b to 2b.[a]
Entry PdII/xantphos CO Pumps [mL/min] Reactor Temp. Conv.
It is possible that the ortho-ketone functional group was
somehow interfering with the catalytic cycle, as suggested
by the significant amount of precipitation (presumably Pd0)
observed using this substrate. All of the other iodobenzene
derivatives reacted very cleanly and no significant side-
products could be observed under the conditions of the re-
action. For the vinylsilane substrate 1f, however, the reac-
tion did not proceed so smoothly. A modest isolated yield
of 38% was obtained only after reducing the temperature
to 50 °C. Even at this lower temperature, significant by-
products could be observed in the crude product mixture
[mol-%]
[bar] A = B
C
coil [mL] [°C] [%][b]
1
2
3
4
5
6
7
8
2.5/5.0
2.5/5.0
2.5/5.0
2.5/5.0
2.5/3.0
2.5/3.0
2.5/3.0
2.5/3.0
7
7
7
7
7
7
7
0.10
0.10
0.10
0.075
0.10
0.10
0.10
0.10
0.60
0.60
0.60
0.60
0.60
0.60
0.60
20
20
30
30
30
30
40
30
100
100
100
100
70
100
100
100
18
73
86
82
80
86
85
90
10 0.10
[a] Conditions: SETUP 2 (Figure 3), Ar–I (1 mmol, 0.5 m), QP-TU
1
scavenger. [b] Based on the H NMR spectrum.
Eur. J. Org. Chem. 2014, 6418–6430
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6421

U. Gross, P. Koos, M. O’Brien, A. Polyzos, S. V. Ley
FULL PAPER
Table 2. Palladium-catalysed methoxycarbonylation of aryl iodides gave rise to a low conversion and an isolated yield of 43%
in flow.[a]
(entry 10), the thiophene derivative 1i reacted very
smoothly, furnishing the methyl ester product in 71% yield
after column chromatography (entry 9). After establishing
that the system and conditions were broadly suitable for the
methoxycarbonylation of aryl/alkenyl iodides, we wished to
extend the scope to aryl bromides which, despite being less
reactive, might serve as more useful substrates due to their
greater availability. Using the previously established condi-
tions, only negligibly low conversions were obtained but,
after investigating alternative solvents, the reaction of p-
nitrobromobenzene 3a in dioxane/MeOH at the higher
pressure of 15 bar of CO did afford some of the corre-
sponding methyl ester, albeit with a low conversion of 34%
(Scheme 4).
Scheme 4. Palladium-catalysed methoxycarbonylation reaction of
p-nitro-bromobenzene 3a.
Using this as a starting point, we carried out further op-
timisation studies based around the existing flow chemistry
arrangement (SETUP 2, Figure 3, A = B = 0.1 mL/min, C
= 0.6 mL/min). The results are shown in Table 3. During a
brief screen of additives, it was found that using a 1:1 diox-
ane/methanol solvent system and an increased CO pressure
of 15 bar, a much improved conversion of 65% could be
achieved with the addition of 30 mol-% hydrazine (com-
monly used to efficiently reduce PdII species to Pd0) added
as a 1.0 m THF solution (Table 3, entry 6).[28] Substitution
of hydrazine for NH3 or PhB(OH)2 gave similar results of
50% and 53% respectively (Table 3, entries 3, 4 and 5),
while a large excess of NH3 facilitated fast formation of
Pd0 precipitates. Hydrazine was therefore chosen for further
optimisation, as it can be more easily removed from the
product compared with PhB(OH)2. Increasing the loading
of the catalyst provided the product in 90% conversion and
in 77% isolated yield after purification (Table 3, entry 7).
The use of a higher reaction temperature did effect an in-
crease in conversion (from substrate) but this was ac-
companied by a noticeable reduction in isolated yield, it
seems the higher temperature also promotes decomposition
[a] Conditions based on Table 1, entry 6. [b] Isolated yield after
reactions (Table 3, entry 8). Having established these new
chromatography on silica gel. [c] Based on the 1H NMR spectrum.
conditions (Table 3, entry 7) for the reaction of p-nitro-
bromobenzene 3a, we then evaluated the scope of this pro-
cess using other aryl bromides as precursors. The results of
methoxycarbonylations of different aryl bromides are
shown in Table 4. From the results in this brief study it is
[d] Substrate and base were introduced in DMF. [e] Rapid Pd0 pre-
cipitation observed. [f] Reaction performed at 50 °C, significant by-
products observed.
1
(by H NMR analysis). We also wished to investigate ex- clear that, as a general trend, electron deficient aromatic
panding the substrate scope by testing these conditions with bromides are transformed to products in higher conversion
heteroaromatic iodides. Although the pyrazole substrate 1j compared with more electron rich substrates.
6422
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 6418–6430

General Continuous Flow Method for Pd-Catalysed Carbonylations
Table 3. Results for investigation of reaction conditions for the con- tive methyl ester products in 62% yield (Table 5, entries 3
version of 3a to 2c.[a]
and 4). Substrate 1k, containing both iodine and bromine
substituents, reacted only at the iodo position (Table 5, en-
try 2). Perhaps surprisingly, we discovered that the use of
polymer supported thiourea was deleterious for some reac-
tions using these new conditions (where it led to lower iso-
lated yields) and in some cases this was replaced with a
scavenger cartridge of densely packed glass wool to filter
out precipitates. Encouraged by these results, we were hope-
ful that these new conditions could be applied to the reac-
tion of vinylsilane compound 1f. Unfortunately, the product
of this reaction, 2f, was isolated in low yield due to the
preferential formation of an unidentified side product. This
led to the re-optimization of conditions for these more reac-
tive vinyl iodides. After a brief investigation we discovered
that the vinylsilane compound 2f was incompatible with
both the polymer supported thiourea and also the xantphos
ligand itself.
Entry Pd(OAc)2 Ligand
[mol-%] [mol-%]
Additive
[mol-%]
Temp. Yield Conv.
[°C]
[%] [%][b]
1
2
3
2.5
5.0
2.5
xantphos (3)
PnBu(Ad)2 (15)
xantphos(3)
–
–
NH3
100
100
100
–
–
–
34
14
50
(10 equiv.)
PhB(OH)2 (3)
H2NNH2 (3)
H2NNH2 (30) 100
H2NNH2 (30) 100
4
5
6
7
8
2.5
2.5
2.5
5.0
5.0
xantphos(3)
xantphos(3)
xantphos(3)
xantphos(6)
xantphos(6)
100
100
–
–
–
77
72
53
55
65
90
94
H2NNH2 (3)
120
[a] Conditions: SETUP 2 (Figure 3), Ar–Br (1 mmol, 0.5 m), QP-
TU scavenger. [b] Based on the H NMR spectrum.
1
Table 4. Palladium-catalysed methoxycarbonylation of aryl brom-
ides in flow.[a]
Table 5. Palladium-catalysed methoxycarbonylation of aryl iodides
in flow.[a]
[a] Conditions: SETUP 2, Figure 3, Pump A = B (0.1 mL/min),
pump C (0.6 mL/min), dioxane/MeOH = 1:1, Pd(OAc)2 (5 mol-%),
xantphos (6 mol-%), CO (15 bar), hydrazine (30 mol-%) (1.0 m
THF solution), Et3N (1.1 equiv.), 100 °C, 30 mL reactor coil
(37.5 min), Ar–Br (0.6 mmol). [b] Isolated yield after chromatog-
1
raphy on silica gel. [c] Based on the H NMR spectrum. [d] Side
products observed. [e] Reaction performed at 120 °C.
[a] Conditions: SETUP 2, Figure 3, Pump A = B (0.1 mL/min),
pump C (0.6 mL/min), dioxane/MeOH = 1:1, Pd(OAc)2 (2.5 mol-
%), xantphos (3 mol-%), CO (15 bar), hydrazine (30 mol-%) (1.0 m
THF solution), Et3N (1.1 equiv.), 30 mL reactor coil (37.5 min),
The highest conversion was observed during the reaction
of p-nitrobromobenzene (3a) and the product 2c was iso- 100 °C, Ar–I (0.6 mmol), scavenger cartridge (glass wool). [b] Iso-
1
lated yield after chromatography on silica gel. [c] Based on the H
NMR spectrum. [d] QP-TU cartridge used.
lated in 77% yield (Table 4, entry 1). The reaction of m-
nitrobromobenzene (3c) (Table 4, entry 3) and p-(trifluoro-
methyl)bromobenzene (3b) (Table 4, entry 2) provided the
corresponding products but with lower conversions. Unsur-
Consequently, this reaction was carried out without li-
prisingly, the flow methoxycarbonylation of p-tolyl bromide gand and at room temperature. Pleasingly, these changes
(3d) gave only a 12% conversion to the product 2l (Table 4, provided an increased yield of 83% (Table 6, entry 1). En-
entry 4). Despite these disappointing results, we decided to couraged by this result, the reactions of several vinyl iodides
try these new conditions with iodopyrazole 1j that had pre- were examined (Scheme 5). The results are shown in
viously reacted with low conversion. (Table 5). Pleasingly, Table 6. Interestingly, E-vinyl iodide 1n gave almost no con-
these conditions afforded the corresponding methyl ester version in the absence of xantphos. It reacted very smoothly
product 2j in a much improved isolated yield (65%, Table 5 in the presence of the ligand at room temperature with full
vs. 43%, Table 2). Additionally, the gas-flow carbonylation conversion to the product 2p which was isolated in a very
of nitrogenous heterocycles 1l and 1m afforded their respec- respectable yield of 81% (Table 6, entry 2).
Eur. J. Org. Chem. 2014, 6418–6430
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6423

U. Gross, P. Koos, M. O’Brien, A. Polyzos, S. V. Ley
FULL PAPER
Table 6. Palladium-catalysed methoxycarbonylation of vinyl iod-
ides in flow.[a]
Table 7. Palladium-catalysed intramolecular carbonylations of aryl
iodides in flow.[a]
[a] Conditions: SETUP 2, Figure 3, Pump A = B (0.1 mL/min),
pump C (0.6 mL/min), dioxane, Pd(OAc)2 (5 mol-%), xantphos
(6 mol-%), CO (15 bar), if hydrazine then 30 mol-% loading (1.0 m
THF solution), Et3N (1.1 equiv.), 30 mL reactor coil (37.5 min),
100 °C, Ar–I (0.6 mmol), scavenger cartridge (glass wool). [b] Iso-
lated yield after column chromatography on silica gel. [c] Calcu-
[a] Conditions: SETUP 2, Figure 3, Pump A = B (0.1 mL/min),
pump C (0.6 mL/min), dioxane/MeOH = 1:1, Pd(OAc)2 (5 mol-%),
if xantphos then 6 mol-% loading, CO (15 bar), Et3N (1.1 equiv.),
30 mL reactor coil (37.5 min), room temp., substrate (0.6 mmol),
scavenger cartridge (glass wool). [b] Isolated yield after chromatog-
1
lated from H NMR of crude product. [d] Hydrazine used as pro-
moter.
1
raphy on silica gel. [c] Based on the H NMR spectrum of crude
product. [d] No QP-TU cartridge used. [e] Xanthphos ligand used.
[f] Products were not isolated due to their instability to chromatog-
raphy on silica gel.
Scheme 6. Intramolecular palladium-catalysed methoxycarb-
onylations of vinyl iodides in flow.
Scheme 5. Palladium-catalysed methoxycarbonylations of vinyl
iodides in flow.
In the reactions, the intramolecular interception of all
three aryl iodides 1q, 1r and 1s proceeded cleanly and the
corresponding lactone and lactam products were isolated in
Similarly, reaction of the cis vinyl iodide 1o provided high yields. Interestingly, the carbonylation reactions of aryl
cleanly the product 2q in good yield (Table 6, entry 3). The iodides 1q and 1r without hydrazine as an additive resulted
profound influence of the xantphos ligand in these palla- in lower conversion highlighting the importance of hydraz-
dium-catalysed methoxycarbonylations is illustrated clearly ine in the reaction system (Table 7, entries 1 and 2).
by the reaction of 2-bromovinyl iodide (1p) (Table 6, entries
Water is perhaps the simplest and most readily available
4 and 5). The reaction without xantphos proceeded only at nucleophile available for these reactions and, given the syn-
the iodo position, while the second insertion into the car- thetic versatility of the resulting carboxylic acid products
bon–bromine bond was accomplished only in the presence we were keen to use our system to exploit this possibility.
of the added ligand (Table 6, entry 5). Having established
As dioxane had been a good solvent for previous reac-
that the reactor system could be used for the methoxycarb- tions, and it is also miscible with water, we chose to employ
onylation of aryl/vinyl/heteroaryl iodides and bromides we it here. A 10:1 ratio of dioxane to water allowed excess nu-
sought to extend its use by varying the nature of the nucleo- cleophile to be present whilst keeping all the reaction com-
philic trap, beginning with an investigation of intramolec- ponents in solution. The other conditions were essentially
ular nucleophilic functional groups. As shown in Table 7, the same as those used for the intramolecular cyclisations
intramolecular reactions could be carried out with amino- (Scheme 7). The hydroxycarbonylations were examined
and hydroxy-aromatic iodides when the reaction was car- using two possible flow configurations. In an effort to har-
ried out in the absence of a nucleophilic solvent (Scheme 6). ness the reactivity of the carboxylic acid functional group
6424
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 6418–6430

General Continuous Flow Method for Pd-Catalysed Carbonylations
as a means of purification, we modified the flow system Table 8. Palladium-catalysed hydroxycarbonylations of aryl iodides
in flow.[a]
(SETUP 3) to follow a catch and release isolation protocol.
Although the flow system remained very similar to that
used in previous reactions, polymer-supported reagent was
introduced to permit in-line work-up and purification. The
exiting stream from the reactor was directed through a car-
tridge packed with polymer-supported ammonium hydrox-
ide (Ambersep 900, OH-form) to trap the acid, as its carb-
oxylate anion, via a “catch-and-release” protocol.
Scheme 7. Palladium-catalysed hydroxycarbonylations of aryl iod-
ides in flow.
Following a washing step with a dioxane/water mixture
to remove any unwanted impurities, a third stream (0.5 mL/
min) was introduced through an external pump to release
the carboxylic acid from the resin. A solution of dioxane/
formic acid (9:1) was sufficient to promote the complete
“release” of the carboxylic acids from the resin (Figure 4,
SETUP 3). For the second option, the flow reactor configu- [a] Conditions: A (SETUP 3, Figure 4, dioxane/H2O = 10:1); B
[SETUP 2, Figure 3, pump A = B (0.1 mL/min), pump C (0.6 mL/
ration remained the same as in SETUP 2 (Figure 3) and the
min)]; general: Pd(OAc)2 (5 mol-%), xantphos (6 mol-%), CO
carboxylic acids were purified by acid/base extraction by
(15 bar), Et3N (1.1 equiv.), 30 mL reactor coil (37.5 min), 100 °C,
adjusting the pH of the aqueous phase in a batch isolation
Ar–I (0.6 mmol).
process. Table 8 shows the yields of carboxylic acids that
were obtained from a series of aryl iodides using the two
different isolation methods.
yields (not shown) than using the acid/base extraction pro-
cess (Table 8, entries 3, 4 and 5). We next examined the
possibility of preparing amides using these optimised flow
protocols.
The reaction of the electron-rich m-iodotoluene (1b) with
varying amounts of n-propylamine was chosen as a test re-
action to evaluate conditions for the preparation of amides
in flow (Scheme 8). In this case, SETUP 2 was used and the
amine was loaded into the loop directly together with the
substrate, base and hydrazine. The amount of both n-prop-
ylamine and the hydrazine additive were varied and the re-
sults are shown in Table 9. Although n-propylamine itself is
inexpensive, many amine coupling partners would be much
more valuable.
Figure 4. Schematic of the flow reactor setup incorporating an in-
line “catch and release” purification protocol (SETUP 3).
The conversion of electron deficient aryl iodides 1c, 1d,
1g and 1h to the corresponding carboxylic acids proceeded
smoothly and the products were isolated in very good yields
(Table 8, entries 2, 3, 4 and 5). The reaction of electron rich
aryl iodide 1b yielded product in 58% yield after a catch-
and-release isolation process (Table 8, entry 1). It should be
Scheme 8. Palladium-catalysed aminocarbonylation reaction of
aryl iodide 1b.
We were pleased to observe, therefore, that decreasing the
pointed out that use of the catch and release isolation pro- amount of nucleophile from 7 to 1.5 equiv. resulted in only
tocol in the case of 1d, 1g and 1h yielded products in lower a small change in conversion (Table 9, entries 1, 2 and 3).
Eur. J. Org. Chem. 2014, 6418–6430
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6425

U. Gross, P. Koos, M. O’Brien, A. Polyzos, S. V. Ley
Table 9. Palladium-catalysed aminocarbonylation of aryl iodide 1b. amines were primary nucleophiles we also considered the
FULL PAPER
use of the secondary amine dimethylamine as a nucleophile.
However, the reaction of m-iodoanisole 1a with a 2.0 m
THF solution of HNMe2 (4 equiv.) using optimised condi-
tions (Table 9, entry 4) disappointingly proceeded in only
40% conversion. As the HNMe2 is a gaseous reagent, the
concentration of HNMe2 in solution after storage is likely
to be severely diminished. We therefore examined the pos-
sibility of direct introduction of a second gas into the flow
system. For this purpose two flow configurations (SETUP 4
and 5) involving a second gas-flow-reactor were constructed
(Figure 5). In both configurations, the actual gas-flow-reac-
tors for introduction of gaseous HNMe2 into the flow
stream were positioned before the CO gas-reactor. As the
first gas-flow reactor was pressurized only up to 2 bar of
dimethylamine, a common back-pressure regulator was
placed immediately after the output to avoid any movement
of CO-rich solvent into the less pressurized part of the flow
system. The reaction mixture could either be flowed directly
through the gas-flow reactors (SETUP 5) or mixed at a T-
piece with a solvent stream that was previously passed
through the gas-flow reactors to charge it with gas (SETUP
4).
Entry nPrNH2 [equiv.] H2NNH2 [mol-%] Yield Conv. [%][b]
1
2
3
4
5
7.0
3.5
1.5
1.5
1.5
30
30
30
10
0
–
–
–
93
70
100
100
97
93
86
[a] Conditions: SETUP 2, Figure 3, pump A = B (0.1 mL/min),
pump C (0.6 mL/min), dioxane, reactor coil 30 mL (37.5 min),
100 °C, gas-flow reactor (3.8 m of AF-2400), Ar–I 1b (0.6 mmol)
(scavenger: glass wool). [b] Isolated yield after column chromatog-
raphy on silica gel. [c] Based on 1H NMR spectrum of crude prod-
1
uct. [d] Cleaner by H NMR than product obtained in entry 3.
A bigger influence on the conversion was observed upon
reducing the hydrazine to 0 mol-% (Table 9, entry 5). In all
cases, no side products resulting from concurrent nucleo-
philic attack by hydrazine were observed, even when
30 mol-% of hydrazine was used. Using 10 mol-% of
hydrazine, and 1.5 equiv. of amine (Table 9, entry 4) af-
forded the amide product in 93% isolated yield. The scope
of this flow process using m-iodotoluene 1b and a range of
amines was then examined using these conditions. The re-
sults of these reactions are shown in Table 10. In all cases
the reaction proceeded well and the electronic nature and/
or steric hindrance of the amine did not overly influence
the outcome of the reaction. The corresponding amides 4c–
g from the flow amino carbonylations were isolated in high
yields (Table 10, entries 1–5). Since all previously examined
Table 10. Palladium-catalysed aminocarbonylations of 1b in con-
tinuous flow.[a]
Figure 5. Schematic of reactor setup with direct connection of a
second gas-flow reactor (SETUP 4 and 5), GFR = gas flow reactor.
The aminocarbonylation reaction of m-iodoanisole (1a)
using varying flow conditions was then examined
(Scheme 9). As the permeability of HNMe2 gas through the
AF-2400 membrane (and take-up in the solvent) might well
be different compared to CO we also varied the membrane
length in the gas-flow reactor. The results from these opti-
mization reactions using both flow setups are shown in
Table 11. As the reaction of m-iodoanisole (1a) at 7 bar of
CO and 0.6 bar of Me2NH (1.0 m length of membrane tub-
[a] Conditions: SETUP 2, Figure 3, pump A = B (0.1 mL/min),
pump C (0.6 mL/min), dioxane, Pd(OAc)2 (2.5 mol-%), xantphos
(3 mol-%), CO (15 bar), Et3N (1.1 equiv.), NH2R (1.5 equiv.),
hydrazine (10 mol-%) (1.0 m THF solution), 30 mL reactor coil
(37.5 min), 100 °C, m-iodotoluene 1b (0.6 mmol). [b] Isolated yield
after chromatography on silica gel.
Scheme 9. Optimization of palladium-catalysed aminocarb-
onylation of aryl iodide 1a using HNMe2 as a nucloephile.
6426
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 6418–6430

General Continuous Flow Method for Pd-Catalysed Carbonylations
ing) using SETUP 4 proceeded only with 17% conversion before first tube-in-tube device (CO). A back-pressure regu-
(Table 11, entry 1) the reaction using SETUP 5 (Table 11, lator between these two gas-flow reactors was necessary to
entry 2) was studied. By changing the flow setup, the reac- keep the CO gas in solution, thus avoiding the segmented
tion (50 min at 0.3 mL/min) proceeded with a higher con- flow. The results for the aminocarbonylation of several aryl
version of 35% although the reaction stream was 3 times iodides with dimethylamine are shown in Table 12. The re-
more concentrated (0.25 mol/L) than previously (Table 11, action of electron rich aryl iodides 1a and 1t, gave the corre-
entry 1 vs. 2).
sponding products after isolation in very high yields
(Table 12, entries 1 and 3). The electron deficient aryl iodide
1h reacted very rapidly and quantitatively to furnish the
amide product 4i which was isolated in high yield (Table 12,
Table 11. Results for investigation of reaction conditions for the
conversion of 1a to 4h.[a]
Entry CO Me2NH Setup Reaction time
Conc.
H2NNH2 Conv.
[bar] [m][b]
[min][c]
[mol/L][d] [mol-%][e] [%][f]
1
2
3
4
5
6
7
7
7
7
7
7
1.0
1.0
1.0
2.0
2.0
2.0
2.0
4
5
5
5
5
5
4
38
50
50
50
50
50
38
0.08
0.25
0.08
0.08
0.08
0.08
0.08
–
–
–
17
35
60
75
80
–
30
30
30
10
10
93
100
[a] Gas-flow reactor: length of membrane tubing (CO): 3.8 m. [b]
Length of membrane tubing (in metres), P(Me2NH) = 0.6 bar. [c]
Retention time in the heating coils. [d] Concentration of iodide
when it enters the heating coil. [e] H2NNH2 (1 m in THF). [f] Deter-
1
mined by H NMR spectroscopy.
In order to increase the relative proportion of gaseous
components in the reaction, we lowered the concentration
of the substrate to 0.08 mol/L. This led to an increased con-
version of 60% (Table 11, entry 3). This probably indicates
that the amount of at least one of the two gases was partly
limiting. As we knew from previous experiments that the
concentration of CO was adequate to provide full conver-
sion to a carbonylated product, we decided to vary the
membrane length in the first reactor in an attempt to in-
crease the concentration of HNMe2 in the reaction mixture.
By employing a longer membrane tubing (2.0 m) in the first
gas-flow reactor (introduction of HNMe2) the reaction pro-
vided product 4h in a much higher conversion of 75%
(Table 11, entry 4). We also examined the influence of
hydrazine on the conversion of the reaction. Consequently,
the reaction using hydrazine led to increased conversion of
80% (Table 11, entry 5). In addition, by increasing the CO
pressure (10 bar) the conversion was further increased to
93% (Table 11, entry 6).
Figure 6. General schematic of flow setup for palladium-catalysed
aminocarbonylation of aryl iodides using second gas-flow reactor
for introduction of dimethylamine (SETUP 4).
Table 12. Palladium-catalysed dimethylaminocarbonylations in
continuous flow.[a]
Since streaming of the palladium catalyst through the
membrane tubing of the gas-flow reactors could lead to
clogging, we examined once again the alternative flow con-
figuration (SETUP 4) by employing optimised conditions.
In this case, the reaction proceeded with full conversion
(Table 11, entry 7).
Therefore, for further aminocarbonylations of aryl iod-
ides the conditions from Table 11, entry 7 with a three
stream flow configuration (SETUP 4) were used. The com-
plete schematic of the flow reactor configuration (SETUP
4) used in the aminocarbonylation of several aryl iodides is
shown in Figure 6. As stated above, the optimal flow system
remained almost identical to the configuration of SETUP 2,
[a] Conditions: Setup 4, Figure 6, 2 m (Me2NH) and 3.8 m (CO) of
AF-2400, dioxane, Pd(OAc)2 (2.5 mol-%), xantphos (3 mol-%), CO
(10 bar), Et3N (1.1 equiv.), Me2NH (0.6 bar), hydrazine (30 mol-%)
(1.0 m THF solution), 30 mL reactor coil (37.5 min), 100 °C, Ar–I
but an additional tube-in-tube device (HNMe2) was placed (0.6 mmol). [b] Isolated yield after chromatography on silica gel.
Eur. J. Org. Chem. 2014, 6418–6430
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6427

U. Gross, P. Koos, M. O’Brien, A. Polyzos, S. V. Ley
FULL PAPER
time in the heated flow coils was 37.5 min. The installation of back-
pressure regulators (2ϫ 100 psi) before the second T-piece (located
before the heating coils) was to ensure unidirectional flow through
the heating coils. On exiting the heating coils, the product flow
stream was directed through a glass Omnifit column (15 mm
i.d.ϫ150 mm length) packed with polymer supported thiourea
[QP-TU (5.0 g)] to scavenge palladium residues. A backpressure
regulator (100 psi) was placed immediately after the glass Omnifit
column to prevent out-gassing of the dissolved CO from the solvent
mixture. The product stream was then collected into a round-bot-
entry 2). These results demonstrate that the simple imple-
mentation of a second gas-flow reactor in the flow Scheme
can provide an enriched mixture with two different reactive
gases in an efficient manner for a single flow process.
Conclusions
Harnessing the properties of the amorphous fluoropoly-
mer Teflon® AF-2400, these tube-in-tube gas-liquid flow re- tomed flask and concentrated under reduced pressure to provide
the methyl carboxylates. Final product purification was achieved
by column chromatography on silica gel.
actors enable the efficient, reliable and controllable genera-
tion of homogeneous solutions of CO in continuous flow.
They have been successfully employed in the palladium-cat-
alysed carbonylation reactions of aryl iodides and brom-
ides, coupling with range of hydroxy, alkoxy and amino nu-
cleophiles to furnish amides, esters and carboxylic acids.
Furthermore, intramolecular reactions afforded lactams
and lactones in high yield. We have also shown that the
Typical Experimental Procedure for Methoxycarbonylation of Aryl
Iodides with Solvent System 2 (dioxane/MeOH, 1:1): A solution of
the aryl iodide (1.0 equiv., 0.6 mmol), Et3N (1.1 equiv.) and
H2NNH2 (1 m in THF, 0–30 mol-%, amount as indicated in tables
for each individual reaction) in a mixture of dioxane/MeOH (1:1)
(2 mL) was loaded into a 2 mL PEEK sample loop and a solution
system can be further extended to allow the use of two of xantphos (3 mol-%) and Pd(OAc)2 (2.5 mol-%) in dioxane
(2 mL) was loaded into a second 2 mL PEEK sample loop. These
mixtures were pumped (flow rate 0.1 mL/min) and mixed at a T-
piece. The lengths of tubing between these sample loops and the
T-piece were such that the two 2 mL sample streams arrived at the
T-piece simultaneously. The combined streams were united with a
third solvent stream (dioxane/MeOH, 1:1) (0.6 mL/min) enriched
with CO (15 bar) from the gas-liquid reactor, via a second T-piece.
The combined substrate, catalyst and CO streams were directed to
the heating coils (3ϫ 10 mL, 100 °C) of the Vapourtec R4 system.
The total residence time in the heated flow coils was 37.5 min. The
installation of backpressure regulators (2ϫ 100 psi) before the sec-
ond T-piece (located before heating coils) was used to ensure uni-
directional flow through the heating coils. On exiting the heating
coils, the product flow stream was directed through a glass Omnifit
column (15 mm i.d. ϫ150 mm length) packed with polymer sup-
ported thiourea [QP-TU (5.0 g)] to scavenge palladium residues. In
some cases the glass Omnifit column was filled with glass wool
instead of QP-TU as indicated for each individual reaction. A
backpressure regulator (200 psi) was placed immediately after the
glass Omnifit column to prevent out-gassing of the dissolved CO
from the solvent mixture. The product stream was then collected
into a round-bottomed flask and concentrated under reduced pres-
sure to give the methyl carboxylates. Final product purification was
achieved using column chromatography on silica gel.
tube-in-tube gas-flow devices, permitting the use of gaseous
amine nucleophiles. By only using relatively small volumes
of pressurised gas, these continuous flow reactor systems
allow carbon monoxide to be used at relatively high pres-
sures with a potentially enhanced safety profile relative to
comparable batch mode systems.
Experimental Section
General: All solvents and commercially available reagents were pur-
chased from either Sigma–Aldrich, Alfa-Aesar, Fisher Scientific or
Acros, were reagent grade and were used as supplied without fur-
ther purification. Merck 9385 grade silica gel (purchased from
Sigma–Aldrich) was used for column chromatography.
Further details of apparatus and spectroscopic data for each com-
pound can be found in the Supporting Information.
Caution! Special precautions must be taken when using CO as it is
a profoundly toxic and flammable gas (R61, R12, R26, R48/23). All
experiments were performed in an efficient fumehood and the levels
of CO were checked constantly using regularly tested CO detectors
(both inside and outside the fumehood).
Gas pressure gauge readings correspond to pressure above the am-
bient pressure of the laboratory.
Experimental Procedure for the Synthesis of Carboxylic Acids 5a–
5e in Flow: A solution of the aryl iodide (1.0 equiv., 0.6 mmol) and
Et3N (1.1 equiv.) in a mixture of dioxane/H2O (10:1) (2 mL) was
Typical Experimental Procedure for Methoxycarbonylation with
with Solvent System 1 [toluene/MeOH/DMF (45:45:10)]: A solution loaded into a 2 mL PEEK sample loop and a solution of xantphos
of the aryl iodide (1.0 equiv., 1.0 mmol) and Et3N (1.1 equiv.) were
dissolved in a mixture of toluene/MeOH/DMF (45:45:10) (2 mL)
which was loaded into a 2 mL PEEK sample loop and a solution of
(6 mol-%) and Pd(OAc)2 (5 mol-%) in dioxane (2 mL) was loaded
into a second 2 mL PEEK sample loop. These mixtures were
pumped (flow rate 0.1 mL/min) and mixed at a T-piece. The lengths
xantphos (3 mol-%) and Pd(OAc)2 (2.5 mol-%) in toluene/MeOH/ of tubing between these sample loops and the T-piece were such
DMF (45:45:10) (2 mL) was loaded into a second 2 mL PEEK
sample loop. These mixtures were pumped (flow rate 0.1 mL/min)
and mixed at a T-piece. The lengths of tubing between these sample
loops and the T-piece were such that the two 2 mL sample streams
arrived at the T-piece simultaneously. The combined streams were
that the two 2 mL sample streams arrived at the T-piece simulta-
neously. The combined streams were united with a third solvent
stream (dioxane/H2O, 10:1) (0.6 mL/min) enriched with CO
(15 bar) from the gas-liquid reactor, via a second T-piece. The com-
bined substrate, catalyst and enriched CO streams were directed to
the heating coils (3ϫ 10 mL, 100 °C) of the Vapourtec R4 system.
The total residence time in the heated flow coils was 37.5 min. The
installation of backpressure regulators (2ϫ 100 psi) before the sec-
ond T-piece (located before the heating coils) was used to ensure
unidirectional flow through the heating coils.
united with
a third solvent stream (toluene/MeOH/DMF =
45:45:10) (0.6 mL/min) enriched with CO (7 bar) from the gas-li-
quid reactor, via a second T-piece. The combined substrate, catalyst
and enriched CO streams were directed to the heating coils (3ϫ
10 mL, 100 °C) of the Vapourtec R4 system. The total residence
6428
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 6418–6430

General Continuous Flow Method for Pd-Catalysed Carbonylations
Purification Method A. “Catch and Release”: The flow stream, exit-
ing the heating coil was then directed through a glass Omnifit col-
(0.6 bar) via a second T-piece. The enrichment was achieved by two
gas-liquid reactors in series: The first gas-liquid reactor was at-
umn (15 mm i.d.ϫ150 mm length) containing Ambersep A-900 tached to the Me2NH cylinder, the second one to the CO cylinder.
hydroxide resin (2.4 g, 4 equiv.). The A-900 column was washed
with a stream of dioxane/water (10:1) for 40 min to remove any
unwanted by-products. A third stream driven by an external pump
(Knauer Smartline-100) was used to introduce a solution of formic
acid/dioxane (1:9) via a T-piece placed in between the Omnifit col-
umn and the heating coil. The acid stream was directed through
the A-900 column for 1 h (0.5 mL/min). Finally, the collected sol-
vent was concentrated in vacuo to provide the carboxylic acid prod-
uct without further purification.
A backpressure regulator (100 psi) was placed in between the two
gas-liquid reactors. The combined substrate, catalyst and enriched
CO/Me2NH streams were directed to the heating coils (3ϫ 10 mL,
100 °C) of the Vapourtec R4 system. The total residence time in
the heated flow coils was 37.5 min. The installation of backpressure
regulators (2ϫ 100 psi) before the second T-piece (located before
the heating coils) was to ensure unidirectional flow through the
heating coils. On exiting the heating coils, the product flow stream
was directed through
a glass Omnifit column (15 mm i.d.
ϫ150 mm length) packed with glass wool. A backpressure regula-
tor (100 psi) was placed immediately after the glass Omnifit column
to prevent out-gassing of the dissolved CO from the solvent mix-
ture. The product stream was then collected into a round-bottomed
flask and concentrated under reduced pressure to furnish the crude
dimethyl amide products. Final product purification was achieved
by column chromatography on silica gel.
Purification Method B. Extraction: The flow stream, exiting the
heating coil was then directed through a glass Omnifit column
(15 mm i.d.ϫ150 mm length) containing glass wool. The reaction
mixture was concentrated under reduced pressure. The crude mix-
ture was then redissolved in ethyl acetate and water. The aqueous
layer was basified with NaOH solution (10% in water) to approxi-
mately pH 12–14 and was subsequently extracted twice with ethyl
acetate. Then, the aqueous layer was acidified with HCl (3 n) to
approximately pH 1, whereupon it was extracted six times with
ethyl acetate. The combined organic extracts (from the acidic aque-
ous solution) were dried with MgSO4 and the solvent was removed
under reduced pressure to afford the carboxylic acid products.
Supporting Information (see footnote on the first page of this arti-
1
cle): H and 13C NMR spectra for all products and additional ex-
perimental information.
Acknowledgments
Experimental Procedure for the Synthesis of Amides 4c–4 g in Flow:
A
solution of the aryl iodide (1.0 equiv., 0.6 mmol), Et3N
The authors thank the following for financial support: Georganics
Ltd. (grant to P. K.), the German Academic Exchange Service
(DAAD) (grant to U. G.), Sanofi-Aventis, CSIRO Material Science
and Engineering, Clayton, Australia (grant to A. P.), Engineering
and Physical Sciences Research Council (EPSRC) (grant to
M. O.’B) and the BP 1702 Endowment (grant to S. V. L.).
(1.1 equiv.) and H2NNH2 (1 m in THF, 10 mol-%) in dioxane
(2 mL) was loaded into a 2 mL PEEK sample loop and a solution
of xantphos (3 mol-%) and Pd(OAc)2 (2.5 mol-%) in dioxane
(2 mL) was loaded into a second 2 mL PEEK sample loop. These
mixtures were pumped (flow rate 0.1 mL/min) and mixed at a T-
piece. The lengths of tubing between these sample loops and the
T-piece were such that the two 2 mL sample streams arrived at the
T-piece simultaneously. The combined streams were united with a
third solvent stream of dioxane (0.6 mL/min) enriched with CO
(15 bar) from the gas-liquid reactor, via a second T-piece. The com-
bined substrate, catalyst and enriched CO streams were directed to
the heating coils (3ϫ 10 mL, 100 °C) of the Vapourtec R4 system.
The total residence time in the heated flow coils was 37.5 min. The
installation of backpressure regulators (2ϫ 100 psi) before the sec-
ond T-piece (located before heating coils) was to ensure unidirec-
tional flow through the heating coils. On exiting the heating coils,
the product flow stream was directed through a glass Omnifit col-
umn (15 mm i.d. ϫ150 mm length) packed with glass wool. A
backpressure regulator (200 psi) was placed immediately after the
glass Omnifit column to prevent out-gassing of the dissolved CO
from the solvent mixture. The product stream was then collected
into a round-bottomed flask and concentrated under reduced pres-
sure to afford the crude amides. Final product purification was
achieved by column chromatography on silica gel.
[1] J. Otera (Ed.), Modern Carbonyl Chemistry, Wiley-VCH,
Weinheim, Germany, 2000.
[2] a) A. Brennfuhrer, H. Neumann, M. Beller, Angew. Chem. Int.
Ed. 2009, 48, 4114–4133; Angew. Chem. 2009, 121, 4176; b) L.
Kollar, in: Modern Carbonylation Methods, Wiley-VCH,
Weinheim, Germany, 2008; c) R. Grigg, S. P. Mutton, Tetrahe-
dron 2010, 66, 5515–5548.
[3] W. Keim, J. Organomet. Chem. 1989, 372, 15–23.
[4] K. Goransson, U. Soderlind, J. He, W. N. Zhang, Renewable
Sustainable Energy Rev. 2011, 15, 482–492.
[5] a) L. D. Prockop, R. I. Chichkova, J. Neurol. Sci. 2007, 262,
122–130; b) C. L. Townsend, R. L. Maynard, Occup. Environ.
Med. 2002, 59, 708–711.
[6] a) J. C. Pastre, D. L. Browne, S. V. Ley, Chem. Soc. Rev. 2013,
42, 8849–8869; b) S. V. Ley, Chem. Rec. 2012, 12, 378–390; c)
J. I. Yoshida, Chem. Rec. 2010, 10, 332–341; d) R. L. Hartman,
J. P. McMullen, K. F. Jensen, Angew. Chem. Int. Ed. 2011, 50,
7502–7519; Angew. Chem. 2011, 123, 7642; e) S. G. Newman,
K. F. Jensen, Green Chem. 2013, 15, 1456–1472; f) J. Wegner,
S. Ceylan, A. Kirschning, Chem. Commun. 2011, 47, 4583–
4592; g) G. Jas, A. Kirschning, Chem. Eur. J. 2003, 9, 5708–
5723; h) D. Webb, T. F. Jamison, Chem. Sci. 2010, 1, 675–680;
i) D. T. McQuade, P. H. Seeberger, J. Org. Chem. 2013, 78,
6384–6389; j) X. Y. Mak, P. Laurino, P. H. Seeberger, Beilstein
J. Org. Chem. 2009, 5; k) V. Hessel, D. Kralisch, N. Kockmann,
T. Noel, Q. Wang, ChemSusChem 2013, 6, 746–789; l) L. Ma-
let-Sanz, F. Susanne, J. Med. Chem. 2012, 55, 4062–4098; m)
B. Ahmed-Omer, J. C. Brandt, T. Wirth, Org. Biomol. Chem.
2007, 5, 733–740; n) I. R. Baxendale, J. J. Hayward, S. Lanners,
S. V. Ley, C. D. Smith, Microreactors in Organic Synthesis and
Catalysis (Ed.: T. Wirth), Wiley, New York 2008; o) J. P.
Knowles, L. D. Elliott, K. I. Booker-Milburn, Beilstein J. Org.
Chem. 2012, 8, 2025–2052; p) B. D. A. Hook, W. Dohle, P. R.
General Description of the Synthesis of Amides 4h–4j in Flow Using
Gaseous Dimethylamine with a Second Gas-Liquid Reactor: A solu-
tion of the aryl iodide (1.0 equiv., 0.6 mmol), Et3N (1.1 equiv.) and
H2NNH2 (1 m in THF, 30 mol-%) in dioxane (2 mL) was loaded
into a 2 mL PEEK sample loop and a solution of xantphos (3 mol-
%) and Pd(OAc)2 (2.5 mol-%) in dioxane (2 mL) was loaded into
a second 2 mL PEEK sample loop. These mixtures were pumped
(flow rate 0.1 mL/min) and mixed at a T-piece. The lengths of tub-
ing between these sample loops and the T-piece were such that the
two 2 mL sample streams arrived at the T-piece simultaneously.
The combined streams were united with a third solvent stream of
dioxane (0.6 mL/min) enriched with CO (10 bar) and Me2NH
Eur. J. Org. Chem. 2014, 6418–6430
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6429

U. Gross, P. Koos, M. O’Brien, A. Polyzos, S. V. Ley
FULL PAPER
Hirst, M. Pickworth, M. B. Berry, K. I. Booker-Milburn, J.
Org. Chem. 2005, 70, 7558–7564; q) C. Wiles, P. Watts, Chem.
Commun. 2011, 47, 6512–6535; r) M. Brasholz, B. A. Johnson,
J. M. Macdonald, A. Polyzos, J. Tsanaktsidis, S. Saubern, A. B.
Holmes, J. H. Ryan, Tetrahedron 2010, 66, 6445–6449; s) C. F.
Carter, I. R. Baxendale, M. O’Brien, J. B. J. Pavey, S. V. Ley,
Org. Biomol. Chem. 2009, 7, 4594–4597.
[14] Biogeneral Inc., 9925 Mesa Rim Road, San Diego, California,
USA; www.biogeneral.com. We are grateful for complimentary
samples provided by Dr Ilia Koev of Biogeneral Inc.
[15] a) M. O’Brien, I. R. Baxendale, S. V. Ley, Org. Lett. 2010, 12,
1596–1598; b) M. O’Brien, N. Taylor, A. Polyzos, I. R. Baxend-
ale, S. V. Ley, Chem. Sci. 2011, 2, 1250–1257; c) A. Polyzos, M.
O’Brien, T. P. Petersen, I. R. Baxendale, S. V. Ley, Angew.
Chem. Int. Ed. 2011, 50, 1190–1193; Angew. Chem. 2011, 123,
1222; d) T. P. Petersen, A. Polyzos, M. O’Brien, T. Ulven, I. R.
Baxendale, S. V. Ley, ChemSusChem 2012, 5, 274–277; e) P. B.
Cranwell, M. O’Brien, D. L. Browne, P. Koos, A. Polyzos, M.
Pena-Lopez, S. V. Ley, Org. Biomol. Chem. 2012, 10, 5774–
5779; f) D. L. Browne, M. O’Brien, P. Koos, P. B. Cranwell, A.
Polyzos, S. V. Ley, Synlett 2012, 23, 1402–1406; g) J. C. Pastre,
D. L. Browne, M. O’Brien, S. V. Ley, Org. Process Res. Dev.
2013, 17, 1183–1191; h) S. L. Bourne, P. Koos, M. O’Brien, B.
Martin, B. Schenkel, I. R. Baxendale, S. V. Ley, Synlett 2011,
2643–2647; i) S. Kasinathan, S. L. Bourne, P. Tolstoy, P. Koos,
M. O’Brien, R. W. Bates, I. R. Baxendale, S. V. Ley, Synlett
2011, 2648–2651; j) S. L. Bourne, M. O’Brien, S. Kasinathan,
P. Koos, P. Tolstoy, D. X. Hu, R. W. Bates, B. Martin, B.
Schenkel, S. V. Ley, ChemCatChem 2013, 5, 159–172; k) S. L.
Bourne, S. V. Ley, Adv. Synth. Catal. 2013, 355, 1905–1910; l)
S. Newton, S. V. Ley, E. C. Arce, D. M. Grainger, Adv. Synth.
Catal. 2012, 354, 1805–1812.
[7] a) T. Razzaq, C. O. Kappe, Chem. Asian J. 2010, 5, 1274–1289;
b) J. C. Brandt, T. Wirth, Beilstein J. Org. Chem. 2009, 5, 30;
c) C. Spiteri, J. E. Moses, Synlett 2012, 23, 1546–1548; d) L.
Malet-Sanz, J. Madrzak, S. V. Ley, I. R. Baxendale, Org. Bio-
mol. Chem. 2010, 8, 5324–5332.
[8] a) M. O’Brien, P. Koos, D. L. Browne, S. V. Ley, Org. Biomol.
Chem. 2012, 10, 7031–7036; b) D. X. Hu, M. O’Brien, S. V.
Ley, Org. Lett. 2012, 14, 4246–4249; c) S. Kuhn, R. L. Hart-
man, M. Sultana, K. D. Nagy, S. Marre, K. F. Jensen, Lang-
muir 2011, 27, 6519–6527; d) J. G. Kralj, H. R. Sahoo, K. F.
Jensen, Lab. Chip 2007, 7, 256–263; e) O. K. Castell, C. J.
Allender, D. A. Barrow, Lab. Chip 2009, 9, 388–396; f) R. H.
Atallah, J. Ruzicka, G. D. Christian, Anal. Chem. 1987, 59,
2909–2914.
[9] a) M. O’Brien, R. Denton, S. V. Ley, Synthesis 2011, 1157–
1192; b) S. V. Ley, I. R. Baxendale, R. N. Bream, P. S. Jackson,
A. G. Leach, D. A. Longbottom, M. Nesi, J. S. Scott, R. I.
Storer, S. J. Taylor, J. Chem. Soc. Perkin Trans. 1 2000, 3815–
4195.
[10] a) S. Saaby, K. R. Knudsen, M. Ladlow, S. V. Ley, Chem. Com-
mun. 2005, 2909–2911; b) T. Fukuyama, T. Rahman, N. Kam-
ata, I. Ryu, Beilstein J. Org. Chem. 2009, 5; c) M. T. Rahman,
T. Fukuyama, N. Kamata, M. Sato, I. Ryu, Chem. Commun.
2006, 2236–2238; d) E. R. Murphy, J. R. Martinelli, N. Zabor-
[16] a) T. Noel, V. Hessel, ChemSusChem 2013, 6, 405–407; b) Z. Q.
Jia, J. Chem. Technol. Biotechnol. 2013, 88, 340–345; c) C. X.
Sun, Y. Y. Zhao, J. M. Curtis, Anal. Chim. Acta 2013, 762, 68–
75.
[17] C. Brancour, T. Fukuyama, Y. Mukai, T. Skrydstrup, I. Ryu,
Org. Lett. 2013, 15, 2794–2797.
enko, S. L. Buchwald, K. F. Jensen, Angew. Chem. Int. Ed. [18] a) V. D. Pinho, B. Gutmann, L. S. M. Miranda, R. de Souza,
2007, 46, 1734–1737; Angew. Chem. 2007, 119, 1764; e) C. Csa-
jagi, B. Borcsek, K. Niesz, I. Kovacs, Z. Szekelyhidi, Z. Bajko,
L. Urge, F. Darvas, Org. Lett. 2008, 10, 1589–1592; f) P. W.
Miller, L. E. Jennings, A. J. de Mello, A. D. Gee, N. J. Long,
R. Vilar, Adv. Synth. Catal. 2009, 351, 3260–3268; g) R. D.
Chambers, M. A. Fox, G. Sandford, J. Trmcic, A. Goeta, J.
Fluorine Chem. 2007, 128, 29–33; h) R. Abdallah, V. Meille, J.
Shaw, D. Wenn, C. de Bellefon, Chem. Commun. 2004, 372–
C. O. Kappe, J. Org. Chem. 2014, 79, 1555–1562; b) F. Mas-
tronardi, B. Gutmann, C. O. Kappe, Org. Lett. 2013, 15, 5590–
5593.
[19] A. E. Buba, S. Koch, H. Kunz, H. Lowe, Eur. J. Org. Chem.
2013, 4509–4513.
[20] S. Sharma, R. A. Maurya, K. I. Min, G. Y. Jeong, D. P. Kim,
Angew. Chem. Int. Ed. 2013, 52, 7564–7568; Angew. Chem.
2013, 125, 7712.
373; i) J. Kobayashi, Y. Mori, K. Okamoto, R. Akiyama, M. [21] J. J. F. van Gool, S. van den Broek, R. M. Ripken, P. J. Nieuw-
Ueno, T. Kitamori, S. Kobayashi, Science 2004, 304, 1305–
1308; j) M. D. Roydhouse, W. B. Motherwell, A. Constantinou,
A. Gavriilidis, R. Wheeler, K. Down, I. Campbell, RSC Adv.
2013, 3, 5076–5082; k) M. D. Roydhouse, A. Ghaini, A. Con-
land, K. Koch, F. Rutjes, Chem. Eng. Technol. 2013, 36, 1042–
1046.
[22] L. Yang, K. F. Jensen, Org. Process Res. Dev. 2013, 17, 927–
933.
stantinou, A. Cantu-Perez, W. B. Motherwell, A. Gavriilidis, [23] J. N. Lee, C. Park, G. M. Whitesides, Anal. Chem. 2003, 75,
Org. Process Res. Dev. 2011, 15, 989–996; l) M. Hamano, K. D.
Nagy, K. F. Jensen, Chem. Commun. 2012, 48, 2086–2088; m)
S. Hubner, U. Bentrup, U. Budde, K. Lovis, T. Dietrich, A.
Freitag, L. Kupper, K. Jahnisch, Org. Process Res. Dev. 2009,
13, 952–960; n) M. Irfan, T. N. Glasnov, C. O. Kappe, Org.
Lett. 2011, 13, 984–987; o) T. Fukuyama, T. Totoki, I. Ryu,
Green Chem. 2014, 16, 2042–2050.
6544–6554.
[24] a) R. A. Maurya, C. P. Park, J. H. Lee, D. P. Kim, Angew.
Chem. Int. Ed. 2011, 50, 5952–5955; Angew. Chem. 2011, 123,
6074; b) R. A. Maurya, C. P. Park, D. P. Kim, Beilstein J. Org.
Chem. 2011, 7, 1158–1163; c) C. P. Park, D. P. Kim, J. Am.
Chem. Soc. 2010, 132, 10102–10106.
[25] a) M. A. Mercadante, N. E. Leadbeater, Org. Biomol. Chem.
2011, 9, 6575–6578; b) M. A. Mercadante, C. B. Kelly, C. Lee,
N. E. Leadbeater, Org. Process Res. Dev. 2012, 16, 1064–1068.
[26] P. Koos, U. Gross, A. Polyzos, M. O’Brien, I. Baxendale, S. V.
Ley, Org. Biomol. Chem. 2011, 9, 6903–6908.
[11] a) N. Steinfeldt, R. Abdallah, U. Dingerdissen, K. Jähnisch,
Org. Process Res. Dev. 2007, 11, 1025–1031; b) Y. Wada, M. A.
Schmidt, K. F. Jensen, Ind. Eng. Chem. Res. 2006, 45, 8036–
8042.
[12] A. Gunther, M. T. Kreutzer, in: Micro Process Engineering: A
Comprehensive Handbook, vol. 1 (Eds.: V. Hessel, A. Renken,
J. C. Schouten, J. I. Yoshida), Wiley-VCH, 2009, p. 3–40.
[13] a) H. Zhang, S. G. Weber, in: Fluorous Chemistry, Topics in
Current Chemistry, vol. 308 (Ed.: I. T. Horvath), Springer,
2011, p. 307–338; b) I. Pinnau, L. G. Toy, J. Membr. Sci. 1996,
109, 125–133; c) A. Polyakov, Y. Yampolskii, Desalination
2006, 200, 20–20; d) M. K. Yang, R. H. French, E. W. Tokar-
sky, J. Micro-Nanolithogr. MEMS MOEMS 2008, 7, 9.
[27] J. Albaneze-Walker, J. A. Murry, A. Soheili, S. Ceglia, S. A.
Springfield, C. Bazaral, P. G. Dormer, D. L. Hughes, Tetrahe-
dron 2005, 61, 6330–6336.
[28] a) C. Paal, C. Amberger, Ber. Detsch. Chem. Ges. 1904, 37,
124–139; b) L. Malatesta, M. Angoletta, J. Chem. Soc. (Re-
sumed) 1957, 1186–1188; c) D. R. Coulson, L. C. Satek, S. O.
Grim, Inorg. Synth. 1972, 13, 121.
Received: June 23, 2014
Published Online: September 4, 2014
6430
www.eurjoc.org
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2014, 6418–6430