Accelerating reactions with microreactors at elevated temperatures and pressures: Profiling aminocarbonylation reactions

Article abstract of DOI:10.1002/anie.200604175

Squeeze play: Pressurized microreactor systems greatly expand the range of reaction conditions and accelerate gas-liquid mass transfer. Heck aminocarbonylation reactions (see scheme) exemplify the potential for the quick and safe scanning of reagents and

Full text of DOI:10.1002/anie.200604175

Communications
DOI: 10.1002/anie.200604175
Microreactors
Accelerating Reactions with Microreactors at Elevated Temperatures
and Pressures: Profiling Aminocarbonylation Reactions**
Edward R. Murphy, Joseph R. Martinelli, Nikolay Zaborenko, Stephen L. Buchwald,* and
Klavs F. Jensen*
The use of microreactors can greatly accelerate scanning and
optimization of reaction conditions because of enhanced heat
and mass transfer, reduced reaction volumes, and the ability
to run several experiments within a sealed system (minimizing
contamination by oxygen and water). These advantages have
been demonstrated in several studies over the past decade at
conditions typically used in bench-scale synthesis.[1–10]
A
major advantage of microreactors is the ability to perform
reactions under conditions that cannot be easily realized in
conventional glassware, such as the use of standard solvents
(e.g. toluene) at elevated temperatures and pressures.[4,9]
Reactions performed under these conditions is one of the
major benefits of microwave synthesis.[11,12] Microreactors
offer many of the advantages of microwave reactors and have
the additional advantages of continuous flow and that they do
not require a microwave generator.
The realization of high pressure in glass-based micro-
reactors is complicated by difficulties in interfacing with fluid
inlet and exit tubes. For example, typical compression sealing
techniques of glass and silicon devices are cumbersome, and
they are typically limited to moderate pressures (ca. 10 bar) to
avoid breaking the device. Herein we use a recently devel-
oped solder-based sealing technique[5] to construct micro-
reactors capable of reaching pressures exceeding 100 bar[13]
and to demonstrate the potential advantages of operating
above the boiling point of toluene in Heck aminocarbonyla-
tion[14] reactions (Scheme 1). The use of microreactors
enables rapid evaluation of the effects of modifications to
the reaction conditions on yield and selectivity, as well as the
use of experiments with several reagent solutions in rapid
succession. The ability to rapidly change reactants and
conditions would be a powerful strategy for the high-
Scheme 1. Pd-catalyzed Heck aminocarbonylation.
throughput synthesis of a diverse array of compounds as
well as for catalyst screening. Chemical parameters such as
functional groups, ligand properties, and base strength could
be rapidly varied along with reaction conditions. The con-
tinuous operation and scanning of reaction parameters
provides information that can lead to improved insight over
typical batch- or array-based processes, which are often
limited to one-variable-at-a-time experimentation. The case
study also illustrates the advantages of a closed system in
handling elevated pressures of a toxic gas (carbon monoxide)
and air-sensitive Pd catalysts.
The microreactor (Figure 1) is formed in silicon by
defining the mixer and channel layout by lithography and
then etching channels in silicon.[4] Subsequent oxidation of
silicon forms a glass layer on the surface so that when the
channels are capped by bonding a Pyrex glass wafer to the
oxidized silicon device, the reaction channels become func-
tionally equivalent to a glass reaction vessel. Moreover, the
top glass layer provides visual access to the reaction medium,
[*] E. R. Murphy, N. Zaborenko, Prof. K. F. Jensen
Department of Chemical Engineering
Massachusetts Institute of Technology
77 Massachusetts Ave., Cambridge, MA 02139 (USA)
Fax: (+1)617-258-8224
E-mail: kfjensen@mit.edu
J. R. Martinelli, Prof. S. L. Buchwald
Department of Chemistry
Massachusetts Institute of Technology
77 Massachusetts Ave., Cambridge, MA 02139 (USA)
[**] S.L.B. thanksthe National Institutesof Health (GM46059) for
funding and isalso indebted to Merck, Amgen, and Boehringer-
Ingelheim for unrestricted support. K.F.J. thanks the MicroChemical
Systems Technology Center for funding. We are also grateful to
Engelhard for a generousdonation of palladium acetate.
Supporting information for thisarticle isavailable on the WWW
under http://www.angewandte.org or from the author.
Figure 1. Pressurized microreactor setup.
1734
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 1734 –1737

Angewandte
Chemie
which is particularly useful in monitoring gas–liquid contact
and detecting the formation of solid by-products.
In the initial studies with the setup shown in Figure 1, a
Xantphos-Pd catalyst[15] was used with a silicon microreactor,
which we had previously designed for general organic syn-
thesis[5] [Xantphos = 4,5-bis(diphenylphosphino)-9,9-dime-
thylxanthene]. Because of the small dimensions of the reactor
channels, the generation of precipitates needed to be con-
trolled and their potential to clog the reactor minimized. The
solid by-products of chief concern were palladium black,
palladium carbonyl complexes,[16] and protonated amine salts.
Although some precipitation was observed, the formation of
solid materials was minimal and occurred primarily along the
reactor walls. The microreactor was rinsed daily with meth-
anol, which was sufficient to avoid any blockage of the flow.
The first Pd-catalyzed aminocarbonylation reaction
explored was that of 3-iodoanisole (1) and morpholine (2)
(Scheme 2). The carbonylation was performed by delivering
CO with a syringe pump at 7.9 bar. The pulsation of the
Figure 2. Aminocarbonylation product yieldsfor sampleswith greater
than 90% conversion of 3-iodoanisole. Each data point represents an
individual experiment.
To further demonstrate the ability of the microreactor to
rapidly scan pressures and temperatures and to explore their
effect on a-ketoamide formation, the selectivity of the
aminocarbonylation reaction of 4-bromobenzonitrile (5) was
profiled. Though carbonylation reactions of aryl bromides are
more challenging than those of aryl iodides, the greater
diversity and the lower cost of commercially available
substrates make the aryl bromide reactions more attractive.
To accommodate the longer reaction times required when
using the less reactive aryl bromides, the selectivity study was
performed by using the original microreactor design and a
volume extension consisting of a 300-mm length of stainless
steel tubing (1.17-mm internal diameter). The volume exten-
sion was connected to the reactor through the same solder-
bonding technique used for the other connections and it was
submerged alongside the reactor into the oil bath, to give a
total heated reactor volume of 400 mL. Rather than using the
syringe pump with its complicating pulsating flow, carbon
monoxide was delivered directly from the tank and controlled
by using needle valves (Scheme 3).
Reaction data were collected at temperatures between 98
and 1608C and carbon monoxide pressures of 4.5, 7.9, and
14.8 bar. For each sample, the overall selectivity for a-
ketoamide 7 was calculated as the ratio of a-ketoamide to
amide and plotted in Figure 3. The expected increase in
selectivity for amide formation with an increase in temper-
ature was observed, along with enhanced selectivity for the
production of a-ketoamide with increasing pressure, which
corresponds to larger amounts of dissolved carbon monoxide
in the solvent.
Scheme 2. Aminocarbonylation of 3-iodoanisole.
syringe pump induced minor changes in the residence times,
but the average reaction times (ca. 4 min) could still be
determined from the active reactor volume (78 mL) divided by
the averaged volumetric flow rate (QL = 4 mLminÀ1 per
syringe; QG = 12 mLminÀ1).
In contrast to conventional benchtop reactions at near-
atmospheric pressure, aminocarbonylation in the microreac-
tor at elevated CO pressures produced significant quantities
of a-ketoamide 4, in addition to amide 3.[14] This shift in
product distribution is attributed to the superior gas–liquid
contact area, the resulting improvement in mass transfer, and
the greater amount of CO in solution.[17] Figure 2 shows
measured yields of 3 and 4 for samples with greater than 90%
conversion of 3-iodoanisole and an average reaction time of
four minutes. The yield of amide increases with an increase in
temperature. In particular, temperatures above the normal
boiling point of the solvent (toluene, 1108C) produce a
significant shift towards amide formation, which under-
scores the advantage of being able to conduct experiments
at elevated pressures. These initial results suggest that in
addition to achieving faster reaction rates at elevated
temperatures, it could be possible to manipulate the
relative yields of amide and a-ketoamide by varying the
temperature and pressure. An increase in the temperature
favors amide formation whereas elevated CO pressures
will enhance a-ketoamide formation.
In addition to using the microreactor to rapidly scan the
effect of reaction conditions on yield and selectivity, several
reagent solutions could be tested in rapid succession by
Scheme 3. Aminocarbonylation of 4-bromobenzonitrile.
Angew. Chem. Int. Ed. 2007, 46, 1734 –1737
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1735

Communications
use of injection valves on the inlet lines offers the possibility
of using a wide range of substrates to efficiently produce a
library of products. Furthermore, with the microreactor
system, the reaction conditions themselves are improved.
The significantly greater mass-transfer area resulting from
segmented gas–liquid flow enables very rapid reaction times
from the accelerated mass transfer.[18]
The carbonylation case study demonstrates the consid-
erable potential of continuous-flow, microreactor-based
experiments at conditions not easily achieved in conventional
benchtop experiments. In particular, the technique provides a
useful tool for quickly and safely scanning reaction conditions
and reagents. By using this technique, we were able to test
multiple aryl halides over a wide range of temperatures and
pressures much more rapidly than could be accomplished
with batch experiments. For instance, in the study of the
effects of pressure and temperature on the aminocarbonyla-
tion of 4-bromobenzonitrile (5) up to 36 samples were
collected and analyzed in a single day. Table 1 summarizes
the conditions at which the maximum amide and a-ketoamide
yields for each reaction were observed. These results reveal
the general trend for an increase in the yield of amide with an
increase in temperature and an increase in the selectivity of a-
ketoamide at lower temperatures and higher pressures.
Figure 3. Product ratio of a-ketoamide (7) to amide (6) for the
aminocarbonylation of 4-bromobenzonitrile (5). Each data point repre-
sents an individual experiment.
installing injection valves onto the liquid reagent inlet lines.
Thus, if one desired to run several combinations of substrates
under the same conditions, this could be accomplished by
simply varying the solutions loaded into the inlet-sample
loops. This ability to rapidly change reactants and conditions
is a powerful strategy for fast synthesis of a diverse array of
compounds as well as for catalyst screening.
We demonstrated this ability by switching between 4-
bromobenzonitrile and 4-bromoanisole without stopping the
flow into the reactor. The system was brought to pressure with
toluene loaded in each syringe and the sample loops were
loaded with reagent stock solutions: Pd(OAc)2, Xantphos,
and toluene in one and Ar-Br, DBU (1,8-diazabicyclo-
[5.4.0]undec-7-ene), dodecane (internal standard), and mor-
pholine in the other. This system was successfully used to test
aminocarbonylation of 4-bromoanisole (Table 1, entry 5).
The use of a pressurized microreactor system greatly
expands the range of reaction conditions available to the
bench chemist. In this study, pressures from 4.5 to 14.8 bar
and temperatures from 98 to 1608C were examined with
greater flexibility, in terms of loading and sampling, than
would be possible with traditional high-pressure chemical
equipment such as a Parr bomb or autoclave. In addition, the
Experimental Section
Aminocarbonylation with carbon monoxide gas delivery directly
from gas tank: The system pressure was controlled by three needle
valves (Upchurch P-445). The cylinder outlet was split and each
branch was connected to a needle valve. One branch was connected to
the microreactor gas inlet and the second branch was connected to the
pressure bomb (pressurized collection vessel) makeup inlet
(Figure 1). The third needle valve was connected to the bomb
outlet tubing such that there was a controlled constant leak from the
headspace of the pressure bomb.
After the bomb was pressurized, all needle valves were closed.
The cylinder delivery pressure was then raised an additional 5% to
compensate for pressure drops in the delivery system. The liquid
reagent flow rate was next set to the desired reaction conditions, and
the reactor gas inlet was partially opened to allow gas flow and begin
slug flow equilibration. Additionally, the leak valve was slightly
opened to allow the microsystem to come to steady state. During
operation, these two valves were used to control the flow rate as
observed in the reactor by measuring the speed of the slugs. The
overall flow rate was varied between 30 and 150 mLminÀ1, which
corresponds to reaction times of 13–2.5 min, respectively. The bomb
inlet valve remained fully closed unless the pressure in the bomb
dropped below the desired operating pressure, in which case the valve
was opened to allow repressurization.
Table 1: Maximum yieldsfor variouscarbonylation reactions.
1
P [bar] T [8C] Av. time Conv. Yield P1 Yield P2
[min.]
[%]
[%]
Sample analysis: The samples were analyzed by gas chromatog-
raphy with an Agilent 6890 Series gas chromatograph and an FID
detector. The samples were injected by an Agilent 7683 automatic
liquid sampler into a 10-meter Agilent HP-1 capillary column
(internal diameter 200 mm; film thickness 0.11 mm) with a flow rate
of nitrogen 1 mLminÀ1 . The oven temperature was raised from 70 to
2408C over 6.5 minutes. Sample peak areas were normalized to the
response of the internal standard to determine the sample concen-
trations. Compounds were isolated by chromatography on the residue
7.9
7.9
146
116
3.3
4.2
100
100
68
35
28
65
2.7
14.8
160
109
7.1
6.6
100
99
83
32
0
57
2.7
150
12.7
48
35
0
1736
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 1734 –1737

Angewandte
Chemie
resulting from combining several samples from a series of experi-
ments by using the same stock solutions and removing the solvent.
[7] P. W. Miller, N. J. Long, A. J. de Mello, R. Vilar, J. Passchier, A.
Gee, Chem. Commun. 2005, 546.
[8] H. Pennemann, P. Watts, S. J. Haswell, V. Hessel, H. Löwe, Org.
Process Res. Dev. 2004, 8, 422.
[9] V. Hessel, C. Hofmann, P. Löb, J. Löhndorf, H. Löwe, A. Ziogas,
Org. Process Res. Dev. 2005, 9, 479.
Received: October 11, 2006
Revised: December 1, 2006
Published online: January 19, 2007
[10] B. K. H. Yen, A. Günther, M. A. Schmidt, K. F. Jensen, M. G.
Bawendi, Angew. Chem. 2005, 117, 5583; Angew. Chem. Int. Ed.
2005, 44, 5447.
[11] C. O. Kappe, Angew. Chem. 2004, 116, 6408; Angew. Chem. Int.
Ed. 2004, 43, 6250.
[12] R. E. Tundel, K. A. Anderson, S. L. Buchwald, J. Org. Chem.
2006, 71, 430.
[13] M. T. Timko, K. A. Smith, R. L. Danheiser, J. I. Steinfeld, J. W.
Tester, AIChE J. 2006, 52, 1127.
[14] R. Schoenberg, R. F. Heck, J. Org. Chem. 1974, 39, 3327.
[15] J. R. Martinelli, D. M. M. Freckmann, S. L. Buchwald, Org. Lett.
2006, 8, 4843.
[16] T. A. Stromnova, I. I. Moiseev, Russ. Chem. Rev. 1998, 67, 485.
[17] F. Ozawa, H. Soyama, H. Yanagihara, I. Aoyama, H. Takino, K.
Izawa, T. Yamamoto, A. Yamamoto, J. Am. Chem. Soc. 1985,
107, 3235.
Keywords: carbon monoxide · carbonylation · microreactors·
palladium · synthesis design
.
[1] W. Ehrfeld, V. Hessel, H. Löwe, Microreactors: New Technology
for Modern Chemistry, Wiley-VCH, Weinheim, 2000.
[2] K. Jähnisch, V. Hessel, H. Löwe, M. Baerns, Angew. Chem. 2004,
116, 410; Angew. Chem. Int. Ed. 2004, 43, 406.
[3] P. D. I. Fletcher, S. J. Haswell, E. Pombo-Villar, B. H. Warring-
ton, P. Watts, S. Y. F. Wong, X. L. Zhang, Tetrahedron 2002, 58,
4735.
[4] K. F. Jensen, Mater. Res. Soc. Bull. 2006, 31, 101.
[5] D. H. Ratner, E. R. Murphy, M. Jhunjhunwala, D. A. Snyder,
K. F. Jensen, P. H. Seeberger, Chem. Commun. 2005, 578.
[6] C. Mauger, O. Buisine, S. Caravieilhes, G. Mignani, J. Organo-
met. Chem. 2005, 690, 3627.
[18] A. Günther, K. F. Jensen, Lab Chip 2006, DOI: 10.1039/b6098.
Angew. Chem. Int. Ed. 2007, 46, 1734 –1737
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1737