A microfluidic approach to the rapid screening of palladium-catalysed aminocarbonylation reactions

Article abstract of DOI:10.1002/adsc.200900563

The evaluation and selection of the most appropriate catalyst for a chemical transformation is an important process in many areas of synthetic chemistry. Conventional catalyst screening involving batch reactor systems can be both time-consuming and expensive, resulting in a large number of individual chemical reactions. Continuous flow microfluidic reactors are increasingly viewed as a powerful alternative format for reacting and processing larger numbers of small-scale reactions in a rapid, more controlled and safer fashion. In this study we demonstrate the use of a planar glass microfluidic reactor for performing the three-component palladium-catalysed aminocarbonylation reaction of iodobenzene, benzylamine and carbon monoxide to form N-benzylbenzamide, and screen a series of palladium catalysts over a range of temperatures. N-Benzylbenzamide product yields for this reaction were found to be highly dependent on the nature of the catalyst and reaction temperature. The majority of catalysts gave good to high yields under typical flow conditions at high temperatures (150°C), however the palladium(II) chloride-Xantphos complex [PdCl₂ (Xantphos)] proved to be far superior as a catalyst at lower temperatures (75-120°C). The utilised method was found to be an efficent and reliable way for screening a large number of palladium-catalysed carbonylation reactions and may prove useful in screening other gas/liquid phase reactions.

Full text of DOI:10.1002/adsc.200900563

FULL PAPERS
DOI: 10.1002/adsc.200900563
A Microfluidic Approach to the Rapid Screening of Palladium-
Catalysed Aminocarbonylation Reactions
Philip W. Miller,a,* Lucy E. Jennings,a Andrew J. deMello,a Antony D. Gee,a,b
Nicholas J. Long,a and Ramon Vilara
a
Department of Chemistry, Imperial College London, South Kensington, London, SW7 2AZ, U.K.
Fax : (+44)-20–7594–5804; e-mail: philip.miller@imperial.ac.uk
GlaxoSmithKline, Clinical Imaging Centre, Imperial College, Hammersmith Hospital, London, W12 0NN, U.K.
b
Received: August 10, 2009; Published online: December 10, 2009
Abstract: The evaluation and selection of the most lysts over a range of temperatures. N-Benzylbenz-
appropriate catalyst for a chemical transformation is
ACHTUNGTRENaNGNU mide product yields for this reaction were found to
an important process in many areas of synthetic be highly dependent on the nature of the catalyst
chemistry. Conventional catalyst screening involving and reaction temperature. The majority of catalysts
batch reactor systems can be both time-consuming gave good to high yields under typical flow condi-
and expensive, resulting in a large number of individ- tions at high temperatures (1508C), however the pal-
ual chemical reactions. Continuous flow microfluidic ladium(II) chloride-Xantphos complex [PdCl2
reactors are increasingly viewed as a powerful alter-
ACHTUNGTREN(NGNU Xantphos)] proved to be far superior as a catalyst at
native format for reacting and processing larger lower temperatures (75–1208C). The utilised method
numbers of small-scale reactions in a rapid, more was found to be an efficent and reliable way for
controlled and safer fashion. In this study we demon- screening a large number of palladium-catalysed car-
strate the use of a planar glass microfluidic reactor bonylation reactions and may prove useful in screen-
for performing the three-component palladium-cata- ing other gas/liquid phase reactions.
lysed aminocarbonylation reaction of iodobenzene,
benzylamine and carbon monoxide to form N-ben- Keywords: carbonylation; catalysis; catalyst screen-
zylbenzamide, and screen a series of palladium cata- ing; microfluidics; micro-reactor; palladium
Introduction
field. Firstly, there are the obvious safety concerns of
handling a poisonous gas such as carbon monoxide
Since its initial discovery,[1–3] the palladium-catalysed which may additionally need to be used at high pres-
carbonylation reaction of aryl, heteroaryl, vinyl and sures. This may also necessitate the use of specialised
benzyl halides has developed into a highly versatile and expensive equipment. Secondly, carbonylation re-
and effective synthetic route to an array of organic actions are not as well characterised as other palladi-
molecules possessing a carbonyl functional group.[4–6] um-catalysed reactions such as cross-coupling reac-
À
À
À
Carboxylic acid derivatives, ketones, aldehydes, carbo- tions for C C, C N and C O bond formation. Yields
nates, carbamates and ureas may be readily prepared are highly dependent on the aryl substrates and nucle-
in one pot via the three-component reaction of an or- ophiles and, moreover, can vary widely depending on
ganic electrophile, nucleophile and carbon monoxide the CO pressure, solvent and reaction temperature. In
in the presence of a suitable palladium catalyst. Car- addition, the nature of the catalyst is known to have a
bonylation reactions are of high interest in the fine particularly noticeable effect on the yield and reaction
chemical and pharmaceutical sectors where they are rates of particular carbonylation reactions.[10] Conse-
viewed as an atom-efficient method for preparing val- quently much emphasis has been placed on catalyst
uable carbonyl compounds from readily available aryl development in recent years to permit reactions
and vinyl halide precursors.[7–9] Although carbonyla- under milder temperatures and pressures in addition
tion reactions have attracted significant interest in to allowing more challenging aryl chloride or tosylate
recent years from academic groups, especially with re- substrates to be coupled.[7,11–18] The characteristics of
gards to catalyst discovery and development, there the ligand bound to the metal centre are of significant
are however several factors which have prevented the importance for catalyst development. The most
wider exploitation of this method within the academic widely investigated ligand systems with regards to car-
3260
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 3260 – 3268

A Microfluidic Approach to the Rapid Screening
FULL PAPERS
bonylation reactions have been tertiary phosphines al-
Multiphase gas-liquid reactions such as hydrogena-
though carbene ligands are receiving much current in- tions,[34–36] oxidations,[37,38] fluorinations[39–41] and more
terest.[19,20] Generally, however, chelating diphosphines recently carbonylations[42–45] are well suited for adap-
are viewed as having ideal characteristics for control- tation to a microfluidic format. This is because, in ad-
ling geometry at the palladium centre and for stabilis- dition to the advantages stated above, significant
ing the active catalyst at higher temperatures by pre- gains in interfacial gas-liquid contact areas can lead to
venting the formation of Pd-black and Pd-carbonyl enhanced reactivity. This is particularly appealing for
clusters which are thought to inhibit the reaction.[21,22] carbonylations because the low solubility and resul-
Determining the reaction conditions that achieve tant poor reactivity of carbon monoxide may be en-
the highest chemical yields in the shortest possible hanced without the need for excessively high pres-
times (space-time yields) usually involves striking the sures. Palladium-catalysed carbonylation reactions are
right balance between the main variables in the reac- proposed to proceed via three distinct steps in the cat-
tion.[23] This is often an arduous and iterative process, alytic cycle: oxidative addition of the aryl/vinyl halide
and can inevitably lead to delays and bottlenecks in to the active Pd catalyst; migratory insertion of
process development. For carbonylation reactions CO carbon monoxide to form an acyl-Pd intermediate;
pressure, temperature and catalyst are the key varia- nucleophilic attack followed quickly by reductive
bles. Optimal pressures are normally in the range 5– elimination to give the carbonyl compound and re-
20 atm while temperatures range from 50 to 1508C generate the active Pd species. At lower carbon mon-
depending on the solvent. However, there is an obvi- oxide pressures the migratory insertion of CO is
ous desire to reduce both the pressures and tempera- thought to be rate-limiting. Thus there is the real pos-
tures of these reactions. Selecting the best catalyst can sibility of enhancing the rate-limiting migratory inser-
be more demanding because related ligands with tion step of carbon monoxide at lower pressures
structures that are seemingly very similar may pro- which is both technically appealing and safer. We re-
duce contrasting results. Take the wide range of palla- cently reported the first example of a palladium-cata-
dium diphosphine complexes, for example, where lysed carbonylation reaction on
a
microfluidic
ligand structures and bite angles can vary by a few device[42] which showed significant advantages in
atoms or degrees with the result of profound changes terms of product yield and reaction rates over compa-
in catalytic activities.[24]
rable batch reactions for the synthesis of secondary
Parallel reactor systems, where a number of reac- amides. In this study we specifically investigated the
tions are run in tandem, are most commonly used for effects of aryl halide substrates and liquid flow rates
rapidly conducting catalyst and reaction condition on the product yields and distributions. Interestingly
screening. This equipment, extensively used in indus- we found that two amide products were formed in the
try, is rather specialised and can be prohibitively ex- reaction; the expected single carbonyl amide and the
pensive for smaller academic groups or more isolated double carbonyl alpha-ketoamide in varying amounts.
investigations. An alternative to these batch reactors The distribution of these products is known to be de-
are continuous flow microfluidic systems which are pendent on the catalyst and reaction conditions, with
particularly adept at handling, reacting and processing temperature playing a dominant role.[45] Herein we
extremely small amounts of substrates and/or cata- wish to extend these studies and report a method for
lysts on the mg to low mg scale.[25,26] In recent times, the screening of a range of catalysts for a model ami-
microfluidic devices have gained recognition for effi- nocarbonylation reaction using microfluidic technolo-
ciently performing a wide range of chemical reactions gy with the aim of rapidly identifying the best catalyst
on extremely small scales.[27–30] Additionally, micro- and reaction conditions.
fluidic reactors have a number of advantages over
commonly used conventional laboratory batch reac-
tors. These include the obvious reduction in reagent Results and Discussion
consumption which can be important when dealing
with high value or precious substrates and catalysts. The well characterised model aminocarbonylation re-
The unique environment within microfluidic channels action of iodobenzene, carbon monoxide and benzyl-
can provide other less obvious advantages which in-
clude greater reaction control and enhancement of
mass transport between reagent streams. Heat trans-
fer into or out of the microchannels is also much im-
proved because of the small convective distances and
large surface-to-volume areas involved. Such advan-
tages have been proven to have a positive impact on
the rates, yields and/or selectivities of chemical reac-
tions.[31–33]
ACHTUNGTRENaNUNG mine to form N-benzylbenzamide (Scheme 1) was
Scheme 1. The aminocarbonylation reaction of iodobenzene
with carbon monoxide gas and benzylamine forming N-ben-
zylbenzamide.
Adv. Synth. Catal. 2009, 351, 3260 – 3268
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
3261

Philip W. Miller et al.
Bite angle
FULL PAPERS
selected to assess a number of palladium-based cata-
lysts over a range of temperatures. The catalysts se-
lected for testing include a range of palladium mono-
dentate phosphine, bidentate phosphine, carbene and
palladium dimer complexes (Table 1) which may
allow us to infer the effect that catalyst stability, elec-
tronic effects, sterics and bite angles could have on
these types of carbonylation reactions. Our microflui-
dic testing rig (Figure 1) is based around a soda-lime
glass device (Figure 2) consisting of two inlets (for in-
troducing liquid reagents and carbon monoxide), a T-
Table 1. (Continued)
Catalyst structure
Catalyst
1,3-bis(2,6-diisopropyl-
phenyl)imidazolinium
chloride+Pd acetate
–
{Pd[P
(t-Bu3)]2}
–
–
Table 1. Catalyst screen.
Catalyst structure
Catalyst
Bite angle
T
AHCTUNGTRENNUNG
2ACHTUNGTREN(GNUN m-Cl)ClACHUTNGTREG[NNUN m-PACHTUNGTREN(NUGN t-Bu)2
A
928
junction to allow mixing, a long continuous serpentine
channel with dimensions of l=5 m, w=200 mm, d=
75 mm and an outlet port for product collection. From
our previous studies it was found that this long ser-
pentine channel was necessary to retain the liquid re-
agents long enough on the device to react under an
annular-type flow regime. Annular flow results when
a gas stream flows through the centre of a channel or
pipe and forces the liquid to adhere and flow along its
internal walls.[46] This type of flow provides the largest
gas-liquid contact surface area and therefore would
be expected to give the highest chemical yields in the
shortest time.
G
U
938
[PdCl
1088
An injector port with a 50 mL sample loop was in-
corporated between the solvent pump and the micro-
fluidic reactor to allow a measured volume of cata-
lyst/iodobenzene/benzylamine solution to be injected
into a continuous flow of solvent and onto the device.
Carbon monoxide was introduced and regulated into
[PdCl
E
1048
998
ACHUTNERGN[NUG PdCl2HCATUNGTRENN(UGN dppf)]
Rac-[PdCl
G
938
A
E
918
A
E
–
[Pd
G
–
Figure 1. Schematic drawing of the reactor set-up.
3262
asc.wiley-vch.de
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 3260 – 3268

A Microfluidic Approach to the Rapid Screening
FULL PAPERS
lyst/reagent solution was introduced into a continuous
flow of THF solvent driving the mixture on to the
chip where it mixed and reacted with carbon monox-
ide. Each 50 mL slug of reagent solution contained
10.2 mg of iodobenzene and approximately 1 mg of
the catalyst. Although it could be expected that much
smaller quantities of reagents and catalysts could be
used we found these quantities convenient for testing.
The residence time of the liquid reagents on the chip,
which can be taken as the reaction time, was previ-
ously measured[42] to be approximately 2 min while
the total processing time for each independent reac-
tion was 12 min from initial injection to collection of
the reacted product. In total 12 different catalysts
were tested and approximately 100 reactions per-
formed at temperatures ranging from 758C to 1508C.
The reaction temperature was found to have a pro-
nounced effect on the yields of N-benzylbenzamide.
As may be expected, chemical yields were found to
increase with temperature, however, it should be
noted this is not always the case with carbonylation
reactions since higher temperatures can degrade the
Figure 2. Glass fabricated microfluidic device used to per-
form gas-liquid aminocarbonylation reactions.
the chip device using a gas flow controller and me- catalyst.[47] Yields were significantly increased when
tered at a constant flow rate. A pressure gauge was the temperatures were increased from 1208C to
used to monitor the back pressure in the system. The 1508C compared to changes from 1008C to 1208C.
chip device was heated using a heating plate with an The higher temperature regimes also resulted in the
associated thermocouple to ensure constant tempera- exclusive formation of N-benzylbenzamide with no
tures.
Reaction solutions consisted of a 1M solution of io- product previously observed at lower temperatures.[42]
dobenzene in benzylamine, which has the role of re- The reaction catalysed by [PdCl2A(dppf)] (Figure 3),
evidence of the double carbonylation a-ketoamide
CTHUNGTRENNUNG
agent, base and solvent, and 2 mol% of palladium cat- for example, gave average yields of 14% at 1008C
alyst. In each reaction a 50 mL injection of the cata- and 36% at 1208C, however when the temperature
Figure 3. Results of a catalyst screen of 4 palladium diphosphine complexes. Each data point represents the average yield
(GC). Error bars show the standard deviation about the mean value (n=3).
Adv. Synth. Catal. 2009, 351, 3260 – 3268
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
3263

Philip W. Miller et al.
FULL PAPERS
Figure 4. Results of a catalyst screen of 4 palladium complexes. Each data point represents the average yield (GC). Error
bars show the standard deviation about the mean value (n=3).
was increased to 1508C yields were almost quantita- structural similarity to the biphenphos and synphos li-
tive. This indicates that some of the catalysts, such as gands, and previous reports of highly successful me-
[PdCl2ACHTUNGTRENNUNG(dppf)] and [PdACHUTGNTREN(NUGN PPh3)4], are much more active thoxycarbonylation reactions of heteroaryl chlor-
at higher temperatures than might otherwise be ex- ides.[13] However, a recent report showed the binap
pected on simple extrapolation of the lower tempera- ligand to be completely ineffective for the preparation
ture data. Reaction yields were generally good to ex- of N-methoxy-N-methylamides (Weinreb amides) via
cellent at higher temperatures despite the short resi- the aminocarbonylation of aryl bromides.[47] At lower
dence times (~2 min) of the liquid reagents on the temperatures (1008C and 1208C) moderate to low
device. Normally, such carbonylation reactions are yields of amide were obtained for all the diphosphine
considered sluggish taking several hours to reach ligand complexes except for the Xantphos complex
completion at higher pressures. However, under the which showed remarkable conversions at both 1008C
right reaction conditions and with the appropriate cat- and 1208C. Average yields of amide in excess of 94%
alysts, they can be surprisingly fast and high yielding (Figure 4) were obtained for the Pd-Xantphos catalyst
as seen with our system, albeit with a relatively reac- which greatly surpassed any of the other catalysts we
tive aryl iodide substrate.
tested at these lower temperatures. The Pd-Xantphos
Diphosphine-based catalysts have been widely used complex still showed remarkable activity giving an
À À
for a range of catalytic reactions. The effect of P M
P bite angle of the diphosphine complexes has been ature was lowered to 758C.
average yield of 83% even when the reaction temper-
extensively investigated and known to be important
Tetrakis(triphenylphosphine)palladium is often
in determining the selectivities and rates of catalytic used as a benchmark catalyst for cross-coupling reac-
reactions.[48,49] In this study we tested seven palladium tions. Carbonylation reactions using catalysts based
diphosphine catalysts which have bite angles ranging on monodentate phosphines have often been ob-
from 928 to 1088. Ligands with wider bite angles, such served to degrade at higher temperatures to catalyti-
as Xantphos, have been generally recognised as cally inactive Pd-black or Pd-CO clusters. Under our
having a positive effect on hydroformylation, hydro- reaction conditions [PdACHTNUTRGNE(UNG PPh3)4] performed on-par
cyanation, CO/alkene copolymerisation, cross-cou- with the diphosphine catalysts at the higher tempera-
pling and carbonylation reactions.[24,50,51] At higher ture of 1508C achieving average yields >75%. At this
temperatures (1508C) all the palladium diphosphine temperature, however, palladium black was observed
catalysts gave good yields (>75%, Figure 3 and to coat the channels of the microfluidic chip.
Figure 4) of N-benzylbenzamide, the only exception The
being the [PdCl2A(binap)] complex which gave a mod- [PdCl
erate yield of 45%. This was surprising considering its significant quantities of Pd-black at 1208C. The palla-
other
triphenylphosphine-based
catalyst,
C
R
2ACHTUNGERTN(NUNG PPh3)2], gave lower yields and also produced
3264
asc.wiley-vch.de
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 3260 – 3268

A Microfluidic Approach to the Rapid Screening
FULL PAPERS
Figure 5. Results of a catalyst screen of 4 palladium complexes. Each data point represents the average yield (GC). Error
bars show the standard deviation about the mean value; n=3 for [PdCl2A(PPh3)2] and [Pd(OAc)2carbene]; n=2 for {Pd[P(t-
Bu)2]} and [Pd2A(m-Cl)Cl{m-P(t-Bu)2-(Bph-H)].
C
R
ACHTUNGTRENNUNG
C
R
ACHTUNGTRENNUNG
dium dimer complex [Pd
of particular interest because of recent reports of its bis(2,6-diisopropylphenyl)imidazolinium
high catalytic activity for cross-coupling reactions at Yields of N-benzylbenzamide obtained using this cat-
low temperatures.[52] Unfortunately [Pd
(m-Cl)Cl(t-Bu- alyst system at 1008C were very low (<10%) but im-
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
biphen)]was only sparingly soluble in the iodoben- proved to >75% when the temperature was increased
zene/benzylamine solution and therefore the reaction to 1508C (Figure 5). At this higher temperature, how-
was conducted in DMF. Disappointingly low yields of ever, a significant amount of palladium black was ob-
amide (<30%, Figure 5) were obtained at 1008C and served in the microchannels of the device suggesting
1208C with significant amounts of palladium black that the catalyst was degrading.
obscuring the microchannels of the reactor. The
The bidentate phosphine catalysts appear to have a
{Pd[P(t-Bu3)2]} complex bearing the highly bulky and greater thermal stability, producing less palladium
ACHTUNGTRENNUNG
electron rich tri-tert-butylphosphines has also been re- black, than the monophosphine or NHC complexes
garded as a highly active catalyst for cross-coupling and are therefore better suited to higher temperature
and carbonylation reactions[10,53] and was therefore of aminocarbonylation reactions. The Xantphos catalyst
high interest to test. A low average yield of 22% was was a clear outlier, performing significantly better
obtained at 1008C, while at 1208C an encouraging than all the other catalysts we tested under these re-
yield of 43% was obtained which is the highest of the action conditions. The wide bite angle (1088) and flex-
monodentate ligand catalysts at this temperature.
ible nature of the Xantphos ligand has been widely
N-heterocyclic carbene (NHC) ligands have attract- reported to have a beneficial effect on a number of
ed much attention as alternative ligand systems to catalytic reactions compared to other diphosphine li-
phosphines, and have consequently been tested for a gands.[50] It is suggested to provide a dynamic and
range of catalytic reactions including, quite recently, flexible coordination environment which is important
carbonylation reactions.[19] NHC ligands have widely for stabilising the metal centre in various oxidation
been considered to supersede phosphines under more states and geometries. Although DPEphos and dppf
harsh reaction conditions because of the degradation have similar bite angles to Xantphos they gave much
À
of P C bonds in phosphine-based catalyst systems at less impressive yields at lower temperatures which
higher temperatures. Pd-NHC complexes are also suggests that the flexibility of the Xantphos ligand
often less sensitive to oxygen and moisture, in addi- may play an important role in these reactions. Xant-
tion to having a higher thermal stability, which poten- phos has, in fact, recently been reported to be an
tially makes them more catalytically robust than Pd- active ligand for aminocarbonylation reactions giving
phosphine catalysts. Unlike the previous preformed impressive conversions when other bidentate and
Pd-phosphine catalysts, the Pd-NHC catalyst we used monodentate phosphine ligand systems failed com-
Adv. Synth. Catal. 2009, 351, 3260 – 3268
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
3265

Philip W. Miller et al.
FULL PAPERS
pletely.[47] However, in another report involving a pal- frames. The method we have described for the amino-
ladium-catalysed aminocarbonylation/amidation cas- carbonylation of aryl halides possesses certain advan-
cade reaction for the one-pot synthesis of quinolones, tages over conventional batch reactor systems for
the Xantphos catalyst system did not perform as well these types of reactions. It is a safer approach using
as the other catalysts that were screened.[54] Although only minute quantities of reagents and small volumes
the positive effects of the Xantphos ligand are well of poisonous carbon monoxide. Although the quanti-
recognised with regards to Rh-catalysed hydroformy- ties of reagent used for our amminocarbonylation re-
lation reactions, its effect on palladium-catalysed ami- actions were small by normal laboratory standards
nocarbonylation, as seen with our results and (10.2 mg iodobenzene, ~1 mg catalyst) these reactions
others,[47] are particularly encouraging.
could be conducted on a much smaller scale (at least
A key feature of our system is that reactions are an order of magnitude smaller) without the need to
performed in microfluidic channels and in continuous modify our existing experimental set-up. The equip-
flow whereby the annular gas/liquid arrangement pro- ment is relatively unsophisticated, inexpensive and
vides a high surface area for enhancing reactivity. Al- commercially available; standard and bespoke glass
though we have found that this leads to much im- fabricated microfluidic devices are now readily avail-
proved space-time yields when compared to equiva- able from a number of commercial suppliers. In addi-
lent batch carbonylation reactions, the system is limit- tion to these technical advantages, performing reac-
ed by the fact that reaction (or residence) times are tions within a continuous flow microfluidic format
very short (~2 min). While such short reaction times also often gives enhanced reactivities compared to
proved sufficient to achieve excellent yields of amides batch systems. Under our reaction conditions moder-
via the aminocarbonylation of aryl iodides, they may ate to excellent yields of N-benzylbenzamide were ob-
prove too short to provide good yields for more chal- tained within extremely short reaction times, albeit at
lenging carbonylations such as the reaction of deacti- high temperatures. The notable exception to this was
vated aryl bromides or aryl chloride substrates. How- the Xantphos ligand system which showed excellent
ever, one straightforward solution to this would be to conversions at much lower temperatures. Interestingly
impose a different flow regime on the gas and liquid this was in agreement with a previous carbonylation
mixture such as alternate slugs of gas and liquid (i.e., study using Xantphos as a ligand[47] and further high-
a segmented flow). Although this method can give lights the beneficial effect of such wide bite angle di-
more controlled residence times of the gas and liquid phosphine ligands. As mentioned above, the short re-
mixtures (that would be suitable for longer reaction action times imposed by the flow regimes in this
times), this may be at the expense of surface area ef- system may be considered as a limitation to its wider
fects seen for the annular flow regime. An added fea- applicability. Nevertheless this system has proven ex-
ture of our reaction system is the speed and reliability tremely useful for rapidly identifying the most suita-
with which reactions could be performed, with each ble reaction conditions for a given aminocarbonyla-
reaction taking only 12 min. In total 12 different cata- tion reaction.
lysts were tested at various temperatures, with 2 or 3
reactions being run at each temperature to ensure re-
producibility. Although no attempt was made to Experimental Section
scale-up any of the reactions under study, the data ob-
tained clearly suggest that the Xantphos ligand General Considerations
system at a minimum temperature of 1008C is best
for the aminocarbonylation of iodobenzene. This
All preparations were carried out using standard Schlenk
line techniques under an inert atmosphere of N2 unless oth-
rapid methodology for screening reactions may there-
fore provide a useful ꢁsnap-shotꢂ of catalyst activity,
and may prove important in identifying the best cata-
lysts and reaction conditions at an early stage of de-
velopment prior to in-depth catalyst screening or
scale-up using standard batch reaction methods.
erwise stated. Solvents were dried over standard drying
agents and freshly distilled under nitrogen before use. All
starting materials and ligands were of reagent grade and
purchased from either Sigma–Aldrich Chemical Company
1
or Strem Chemicals. H and 31P{1H} NMR spectra were re-
corded on Bruker Av-400, DRX-400, Av-500 spectrometers.
Chemical shifts are reported in ppm using the residual
proton impurities in the solvents.
Conclusions
Catalyst Preparation
Microfluidic reactors are playing an increasingly im-
portant role in chemical synthesis especially in areas
such as high throughput synthesis and analysis where
there is a drive to perform large numbers of reactions
Preformed diphosphine catalysts were prepared by the reac-
tion of 1:1 stoichiometry amounts of the appropriate diphos-
phine ligand and [PdCl
(COD)] in dichloromethane. Cata-
2A(PPh3)2] were purchased from
lysts [Pd(PPh3)4] and [PdCl CHTUGNTRENNUNG
G
on smaller volumes of material within shorter time commercial sources. The palladium N-heterocyclic carbene
3266
asc.wiley-vch.de
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 3260 – 3268

A Microfluidic Approach to the Rapid Screening
FULL PAPERS
complex was prepared in situ by the 1:1 stoichiometric reac-
tion of 1,3-bis(2,6-diisopropylphenyl)imidazolinium chloride
and palladium acetate. The palladium dimer complex [Pd2
Acknowledgements
PWM is grateful to the EPSRC for the award of a Life Scien-
ces Interface fellowship (EP/E039278/1) and to GSK for
funding. Katherine Gerry and James Hall are acknowledged
for their assistance with GC analysis.
A
ACHTUNGTRENNUNG{m-PACHTUNGTRENNUNG(t-Bu)2ACHTUNGTRENNUNG(Bph-H)] (Bph-H=biphenyl) was pre-
AHCTUNGTRENNUNG
Bu)3]2} complex was prepared by a modified procedure to
that previously reported in the literature.[56] To a suspension
of [Pd2ACHTUNGTRENNUNG(dba)3] (1.0 mmol, 0.916 g) in toluene (80 mL) was
added tri-tert-butylphosphine (4.95 mmol, 1.0 g) and the re-
action mixture stirred for 48 h at ambient temperature. The
solution was then filtered into a new flask via cannula and
toluene removed under reduced pressure. This crude solid
was then dissolved in dry diethyl ether (30 mL) and placed
in a freezer at À358C overnight to yield white crystals. The
supernatant was removed using a cannula and the crystals
washed with dry diethyl ether (2ꢃ4 mL) and dried under re-
duced pressure; yield. 0.56 g (55%) 1H NMR (benzene-d6,
400 MHz): d=1.65–1.62 (m, 54H), (3JP, H =5.7 Hz);
31P{1H} NMR (benzene-d6, 162 MHz): d=84.9 (s).
References
[1] A. Schoenberg, I. Bartoletti, R. F. Heck, J. Org. Chem.
1974, 39, 3318.
[2] A. Schoenberg, R. F. Heck, J. Am. Chem. Soc. 1974, 96,
7761.
[3] A. Schoenberg, R. F. Heck, J. Org. Chem. 1974, 39,
3327.
[4] A. Brennfuhrer, H. Neumann, M. Beller, Angew.
Chem. 2009, 121, 4176; Angew. Chem. Int. Ed. 2009, 48,
4114.
[5] C. F. J. Barnard, Org. Process Res. Dev. 2008, 12, 566.
[6] C. F. J. Barnard, Organometallics 2008, 27, 5402.
[7] M. Beller, W. Magerlein, A. F. Indolese, C. Fischer,
Synthesis 2001, 1098.
[8] H. U. Blaser, A. Indolese, F. Naud, U. Nettekoven, A.
Schnyder, Adv. Synth. Catal. 2004, 346, 1583.
[9] A. Zapf, M. Beller, Top. Catal. 2002, 19, 101.
[10] L. Ashfield, C. F. J. Barnard, Org. Process Res. Dev.
2007, 11, 39.
Microfluidic Reactor Set-Up
A detailed description of the microfluidic reactor structure
and design has been previously reported.[42] Briefly, the
device was fabricated from a soda-lime glass substrate with
channels being etched into the glass using an HF/NH4F solu-
tion. External holes were drilled at the reagent entry and
exit points prior to thermal bonding of the glass cover plate.
Nanoports (part no. N-333) were bonded to the entrance
and exit holes of the device which allowed external tubing
to be easily connected. The flow of carbon monoxide gas
was regulated directly from the gas cylinder using a Sierra
100 series Smart-Trak mass flow controller and metered at 2
sccm into the chip device. A Harvard syringe pump (pump
11) was used to continuously infuse THF into the chip
device at a flow rate of 10 mL/min via a six-way Rheodyne
injector (part no. 7725i) complete with a 50 mL sample loop.
[11] D. A. Watson, X. X. Fan, S. L. Buchwald, J. Org. Chem.
2008, 73, 7096.
[12] J. R. Martinelli, T. P. Clark, D. A. Watson, R. H.
Munday, S. L. Buchwald, Angew. Chem. 2007, 119,
8612; Angew. Chem. Int. Ed. 2007, 46, 8460.
[13] J. Albaneze-Walker, C. Bazaral, T. Leavey, P. G.
Dormer, J. A. Murry, Org. Lett. 2004, 6, 2097.
[14] C. X. Cai, N. R. Rivera, J. Balsells, R. R. Sidler, J. C.
McWilliams, C. S. Shultz, Y. K. Sun, Org. Lett. 2006, 8,
5161.
Microfluidic Reactions
Reaction mixtures were prepared by dissolving the required
catalyst (2 mol%) in a 1M iodobenzene/benzylamine solu-
tion. In a typical reaction 0.02 mmol of catalyst was dis-
solved in iodobenzene/benzylamine solution (1 mL, 1M).
The reaction solution was loaded into the 50 mL loop and
the injector switched to redirect the flow of THF into the
sample loop which drives the reaction mixture into the chip
device where it mixes and reacts with CO gas. The device
was heated to the required temperature using a heating
block and the product solution collected at the device outlet
into a sample vial. After a processing time of 10 min the
flow rate of the THF from the pump was increased to flush
off any residual product on the device and collected into the
sample vial. The collected samples were allowed to evapo-
rate and then diluted using a dichoromethane/phenyl ether
standard solution. GC samples were run on an HP5890
series II fitted with a Supelco Omegawax column (dimen-
sions; 30 m, 0.25 mm, and 250 mm film thickness) using
helium as the carrier gas. Quantitative analysis was conduct-
ed by gas chromatography using the peak area of the prod-
uct N-benzylbenzamide normalised in response to the inter-
nal diphenyl ether standard.
[15] B.-D. Yehoshua, M. Portnoy, D. Milstein, J. Am. Chem.
Soc. 1989, 111, 8742.
[16] W. Mꢄgerlein, A. F. Indolese, M. Beller, Angew. Chem.
2001, 113, 2940; Angew. Chem. Int. Ed. 2001, 40, 2856.
[17] W. Mꢄgerlein, A. F. Indolese, M. Beller, in: Internation-
al Symposium on Interactions of pi-Systems with
Metals, Elsevier Science Sa, Heidelberg, Germany,
2001, p 30.
[18] C. Jimenez-Rodriguez, G. R. Eastham, D. J. Cole-Ham-
ilton, Dalton Trans. 2005, 1826.
[19] A. S. Veige, Polyhedron 2008, 27, 3177.
[20] V. Calo, P. Giannoccaro, A. Nacci, A. Monopoli, J. Or-
ganomet. Chem. 2002, 645, 152.
[21] K. Kudo, M. Hidai, Y. Uchida, J. Organomet. Chem.
1971, 33, 393.
[22] M. Hidai, M. Kokura, Y. Uchida, J. Organomet. Chem.
1973, 52, 431.
[23] H. Neumann, A. Brennfuhrer, P. Grob, T. Riermeier, J.
Almena, M. Beller, Adv. Synth. Catal. 2006, 348, 1255.
[24] M. N. Birkholz, Z. Freixa, P. van Leeuwen, Chem. Soc.
Rev. 2009, 38, 1099.
[25] A. J. deMello, Nature 2006, 442, 394.
Adv. Synth. Catal. 2009, 351, 3260 – 3268
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
3267

Philip W. Miller et al.
FULL PAPERS
[26] K. Jꢄhnisch, V. Hessel, H. Lowe, M. Baerns, Angew.
Chem. 2004, 116, 410; Angew. Chem. Int. Ed. 2004, 43,
406.
[27] P. Watts, C. Wiles, Chem. Commun. 2007, 443.
[28] H. R. Sahoo, J. G. Kralj, K. F. Jensen, Angew. Chem.
2007, 119, 5806; Angew. Chem. Int. Ed. 2007, 46, 5704.
[29] D. Belder, M. Ludwig, L. W. Wang, M. T. Reetz,
Angew. Chem. 2006, 118, 2523; Angew. Chem. Int. Ed.
2006, 45, 2463.
[30] P. W. Miller, J. Chem. Technol. Biotechnol. 2009, 84,
309.
[31] M. Brivio, W. Verboom, D. N. Reinhoudt, Lab Chip
2006, 6, 329.
[42] P. W. Miller, N. J. Long, A. J. deMello, R. Vilar, J.
Passchier, A. Gee, Chem. Commun. 2006, 546.
[43] P. W. Miller, N. J. Long, A. J. deMello, R. Vilar, H. Au-
drain, D. Bender, J. Passchier, A. Gee, Angew. Chem.
2007, 119, 2933; Angew. Chem. Int. Ed. 2007, 46, 2875.
[44] M. T. Rahman, T. Fukuyama, N. Kamata, M. Sato, I.
Ryu, Chem. Commun. 2006, 2236.
[45] E. R. Murphy, J. R. Martinelli, N. Zaborenko, S. L.
Buchwald, K. F. Jensen, Angew. Chem. 2007, 119, 1764;
Angew. Chem. Int. Ed. 2007, 46, 1734.
[46] A. Gunther, K. F. Jensen, Lab Chip 2006, 6, 1487.
[47] J. R. Martinelli, D. M. M. Freckmann, S. L. Buchwald,
Org. Lett. 2006, 8, 4843.
[32] B. P. Mason, K. E. Price, J. L. Steinbacher, A. R.
Bogdan, D. T. McQuade, Chem. Rev. 2007, 107, 2300.
[33] J. Yoshida, A. Nagaki, T. Iwasaki, S. Suga, Chem. Eng.
Technol. 2005, 28, 259.
[48] P. C. J. Kamer, P. W. N. van Leeuwen, J. N. H. Reek,
Acc. Chem. Res. 2001, 34, 895.
[49] P. van Leeuwen, P. C. J. Kamer, J. N. H. Reek, P. Dier-
kes, Chem. Rev. 2000, 100, 2741.
[50] Z. Freixa, P. van Leeuwen, Dalton Trans. 2003, 1890.
[51] L. A. van der Veen, P. H. Keeven, G. C. Schoemaker,
J. N. H. Reek, P. C. J. Kamer, P. van Leeuwen, M. Lutz,
A. L. Spek, Organometallics 2000, 19, 872.
[52] U. Christmann, D. A. Pantazis, J. Benet-Buchholz, J. E.
McGrady, F. Maseras, R. Vilar, J. Am. Chem. Soc.
2006, 128, 6376.
[34] J. Kobayashi, Y. Mori, S. Kobayashi, Adv. Synth. Catal.
2005, 347, 1889.
[35] J. Kobayashi, Y. Mori, K. Okamoto, R. Akiyama, M.
Ueno, T. Kitamori, S. Kobayashi, Science 2004, 304,
1305.
[36] A. Iles, M. Habgood, A. J. deMello, R. C. R. Wootton,
Catal. Lett. 2007, 114, 71.
[37] A. Leclerc, M. Alame, D. Schweich, P. Pouteau, C. De-
lattre, C. de Bellefon, Lab Chip 2008, 8, 814.
[38] R. C. R. Wootton, R. Fortt, A. J. deMello, Org. Process
Res. Dev. 2002, 6, 187.
[53] T. Doi, S. Kamioka, S. Shimazu, T. Takahashi, Org.
Lett. 2008, 10, 817.
[54] A. C. Tadd, A. Matsuno, M. R. Fielding, M. C. Willis,
Org. Lett. 2009, 11, 583.
[39] R. D. Chambers, M. A. Fox, G. Sandford, Lab Chip
[55] U. Christmann, R. Vilar, A. J. P. White, D. J. Williams,
2005, 5, 1132.
Chem. Commun. 2004, 1294.
[56] P. Frederic, J. Patt, J. F. Hartwig, Organometallics 1995,
14, 3030.
[40] R. D. Chambers, R. C. H. Spink, Chem. Commun. 1999,
883.
[41] K. Jahnisch, M. Baerns, V. Hessel, W. Ehrfeld, V. Ha-
verkamp, H. Lowe, C. Wille, A. Guber, J. Fluorine
Chem. 2000, 105, 117.
3268
asc.wiley-vch.de
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 3260 – 3268