Continuous-flow, palladium-catalysed alkoxycarbonylation reactions using a prototype reactor in which it is possible to load gas and heat simultaneously

Article abstract of DOI:10.1039/c1ob05808h

A prototype tube-in-tube reactor in which it is possible to load gas and heat simultaneously has been used in a continuous-flow approach to alkoxycarbonylation reactions of aryl iodides. In the stainless steel coil, liquid flows on the outside of a gas-permeable membrane. The coil can be heated and the temperature can be measured accurately via a probe touching the outer steel surface. A range of aryl iodides can be transformed to the corresponding esters in excellent conversion by reaction at 120 °C using 0.5 mol% palladium acetate as the catalyst with no additional ligand required. Small-scale optimization and substrate screening runs were followed by scale-up.

Full text of DOI:10.1039/c1ob05808h

View Article Online / Journal Homepage / Table of Contents for this issue
Organic &
Biomolecular
Chemistry
Dynamic Article Links
Cite this: Org. Biomol. Chem., 2011, 9, 6575
www.rsc.org/obc
PAPER
Continuous-flow, palladium-catalysed alkoxycarbonylation reactions using a
prototype reactor in which it is possible to load gas and heat simultaneously†
Michael A. Mercadante and Nicholas E. Leadbeater*
Received 23rd May 2011, Accepted 1st July 2011
DOI: 10.1039/c1ob05808h
A prototype tube-in-tube reactor in which it is possible to load gas and heat simultaneously has been
used in a continuous-flow approach to alkoxycarbonylation reactions of aryl iodides. In the stainless
steel coil, liquid flows on the outside of a gas-permeable membrane. The coil can be heated and the
temperature can be measured accurately via a probe touching the outer steel surface. A range of aryl
◦
iodides can be transformed to the corresponding esters in excellent conversion by reaction at 120 C
using 0.5 mol% palladium acetate as the catalyst with no additional ligand required. Small-scale
optimization and substrate screening runs were followed by scale-up.
Introduction
membrane tubing while the gas fills the PTFE outer tubing
and diffusion transfers the gas into the liquid stream. This
approach has been successful for numerous reactions including
Continuous-flow processing on micro- and mesoscale is currently
a technology of considerable interest in the synthetic chemistry
9
the ozonolysis of alkenes, carboxylation of Grignard reagents
1,2
community. It is possible to control reaction parameters very
precisely and to perform chemistry across a wide temperature
range very effectively. One class of reactions that has attracted
is that involving reactive gases. Hydrogenation reactions can be
10
with carbon dioxide and both homogeneous and heterogeneous
11
hydrogenation reactions. All reactions were performed at room
temperature. In the case of ozonolysis, carboxylation and homoge-
neous hydrogenation, the reagents are flowed through the tube-in
tube unit, loading gas and allowing the reaction to proceed. When
performing heterogeneous hydrogenation reactions, the reagents
are passed through the tube-in-tube unit to load the solution with
hydrogen before passing into a cartridge of palladium on carbon.
To convert all the starting material to product it was necessary to
recycle the reagent mixture through the entire system a number
3
,4
performed using the now ubiquitous H-cube system. Gases such
5
6
7
as oxygen, fluorine, and dimethylsulfide have also been used in
flow. One of the key considerations when performing reactions
involving gaseous reagents, is the ability to introduce the gas into
the reaction mixture effectively. Typically this is achieved by plug
flow or mixing of two phases.
Our laboratory recently took the plug flow approach to the
palladium-mediated carbonylation of aryl iodides in ethanol to
prepare ethyl esters. Palladium acetate (0.5 mol%) was used as
11
of times. This illustrates one of the limitations of the tube-in-
tube approach in its current configuration. When passing reagents
through either a supported catalyst/reagent bed or through a
heated zone, the reaction is limited by the quantity of gas that can
be loaded into the solution in the tube-in-tube unit beforehand.
This necessitates running reactions at more dilute concentrations
or else performing multiple passes through the entire system.
Recently we have had access to a prototype tube-in-tube reactor
in which it is possible to load gas into a heated reaction as it
is consumed. In a proof-of-concept study we have used this to
perform our palladium-catalysed alkoxycarbonylation reactions
and present our results here.
8
the catalyst with no additional ligand required. We introduced the
carbon monoxide gas through a back-pressure regulator and at an
◦
angle of 90 to the liquid flow. The gas/liquid flow then entered a
◦
coiled tube reactor heated to 120 C. An issue associated with this
approach was the necessity to control very carefully the gas flow
in order to obtain satisfactory product yields. In addition, a large
excess of toxic carbon monoxide gas was used in the reaction.
Recently, a way to alleviate some operational issues with
performing reactions involving gaseous reagents has been reported
9–11
by Ley and co-workers.
They use a “tube in tube” approach
(
comprising of an outer PTFE tube in which a gas-permeable
Teflon AF2400 tube is inserted) to load gas into either a solvent
or reagent stream. This allows the flow of liquid within the
Results and discussion
Our prototype reactor comprises of a coil of stainless steel tubing
through which liquid flows on the outside of a gas-permeable
inner membrane. The fact that the coil is made of stainless steel
and that the liquid is in contact with the outer wall means that the
coil (a) is rugged, (b) can be heated or cooled and (c) the liquid
temperature can be measured accurately via a probe touching the
Department of Chemistry, University of Connecticut, 55 North Eagleville
Rd, Storrs, CT, 06269-3060, USA. E-mail: nicholas.leadbeater@uconn.edu;
Fax: +1 860 486 2981; Tel: +1 860 486 5076
†
1
Electronic supplementary information (ESI) available. See DOI:
0.1039/c1ob05808h
This journal is © The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 6575–6578 | 6575

View Article Online
Table 1 Optimisation of conditions for the alkoxycarbonylation reaction
a,b
b
Entry
Reaction conditions
Conversion (%)
Fig. 1 Optimised flow configuration.
Pass through one coil at 0.5 mL min-
1
38
46
60
97
1
2
3
4
-
1
Pass through one coil at 0.25 mL min-
Pass through two coils at 0.5 mL min
a
Table 2 Alkoxycarbonylation of aryl halides
1
-
1
Pass through two coils at 0.25 mL min
a
Reactions were performed on the 5 mmol scale at 1 M concentration,
.5 mol% Pd(OAc) as catalyst, 1.1 eq DBU as base, a gas pressure of
0
1
2
-1
80 psi and a reagent flow rate of 0.25 mL min through two ~15 mL
◦
liquid volume coils heated to 120 C and exiting through a 250 psi back-
b
1
pressure regulator. Determined using H NMR spectroscopy.
b,c
Entry
1
Aryl halide
Alcohol
EtOH
Conversion (%)
91
outer steel surface. The reactor has a working liquid volume of
approximately 15 mL when the inner gas tube is inflated. Gas and
reagents enter the coil in a counter-current manner, the gas coming
in where the reagents exit and vice versa. Our first step was to
find suitable conditions for the alkoxycarbonylation reaction, the
results being shown in Table 1. We initially passed ethanol through
the coil and out through a 250 psi back-pressure regulator at a flow
2
3
EtOH
EtOH
98
95
rate of 0.5 mL/min. We gradually increased the gas pressure to
◦
1
80 psi and then heated the coil to 120 C. Once stabilized we saw
4
5
6
EtOH
EtOH
EtOH
98
98
95
a constant stream of bubbles after the back-pressure regulator
but none before it indicating the carbon monoxide was indeed
dissolving in the ethanol as desired. We then flowed the reagents
through the coil at 0.5 ml min maintaining the gas pressure at
80 psi. We used 4-iodotoluene (5 mmol) as our aryl halide starting
material, ethanol as the nucleophile and solvent (5 mL), DBU (1.1
eq) as base and 0.5 mol% Pd(OAc) as catalyst. Following the
-
1
1
2
stream of reagents we again passed ethanol through the reactor.
As the product mixture began to emerge through the back-pressure
regulator at the end of the reactor, we observed that the gas bubbles
initially subsided and then stopped entirely. Towards the time that
the product stream was fully exiting the coil the bubbles started
appearing again after the back-pressure regulator. This told us
that the reaction was gas-limited and this was exemplified by the
7
8
n-PrOH
i-PrOH
99
99
a
Reactions were performed on the 5 mmol scale at 1 M concentration,
as catalyst, 1.1 eq DBU as base, a gas pressure of
3
8% conversion to product that we obtained (Table 1, entry 1).
0.5 mol% Pd(OAc)
80 psi and a reagent flow rate of 0.25 mL min through two ~15 mL
2
-1
1
We performed a second trial, this time reducing the flow rate
to 0.25 mL/min. While this resulted in a somewhat improved
conversion of 46%, the reaction was still gas limited (Table 1, entry
◦
liquid volume coils heated to 120 C and exiting through a 250 psi back-
b
1
c
pressure regulator. Determined using H NMR spectroscopy. Our earlier
12
8
work, both using batch microwave heating and continuous plug-flow has
shown that product conversion is a very close measure for isolated product
yields for these alkoxycarbonylation reactions.
2
). We turned next to running two gas reactors in series. Working
-
1
at 0.5 mL min we obtained a 60% conversion but running at
0
.25 mL min- increased this to 97% (Table 1, entries 3 and 4). In
1
addition, we saw bubbles of CO after the back-pressure regulator
throughout the run, showing us that the reaction was no longer
gas-limited. Thus, our optimised reaction conditions were: 1 M
of 1 M, a range of aryl iodides could be converted to the ethyl
esters including ortho-substituted examples (Table 2, entries 4
and 6). A representative heteroaromatic substrate, 3-iodopyridine,
gave a good conversion of the desired ester (Table 2, entry 3).
Both 1-propanol and 2-propanol can be used as the alcohol for
the reaction (Table 2, entries 7 and 8). Product conversions were
as good, and in many cases better, than those obtained in our
aryl iodide in ethanol, 0.5 mol% Pd(OAc)
2
as catalyst, 1.1 eq
DBU as base, a gas pressure of 180 psi and a reagent flow rate of
-
1
0
1
.25 mL min , through two ~15 mL liquid volume coils heated to
◦
20 C and exiting through a 250 psi back-pressure regulator. A
8
pictorial representation of our optimised configuration is shown
in Fig. 1.
With the optimised conditions in hand, we screened a range of
aryl iodides to probe substrate scope. Our results are shown in
Table 2. On the 5 mmol scale, working at a reagent concentration
previous plug-flow approach.
Our next objective was to determine the scalability of the
reaction. It occurred to us that as we were passing relatively
small portions of reagents through the coils during our reaction
optimization and substrate screening studies, there may be some
6
576 | Org. Biomol. Chem., 2011, 9, 6575–6578
This journal is © The Royal Society of Chemistry 2011

View Article Online
considerable dispersion in the length of the two reactors. Therefore,
although we were passing reagents into the flow reactor at a
concentration of 1 M, it may be that steady-state is not achieved
and that the concentration of the reaction mixture as it passes
through the length of the two coils may be lower. To probe this,
we performed the reaction using 4-iodotoluene on the 25 mmol
scale, therefore passing 50 mL of reaction mixture through the
system. As the product mixture exited the flow reactor, we collected
fractions of 10 mL. Immediately noticeable was that in the middle
of the run we saw no gas bubbles emerging beyond the back-
pressure regulator, this suggesting that the reaction may have
become gas limited when the reagents had reached a steady-state
concentration in the reactors. Analysis of representative fractions
from the product stream confirmed this. While at the start of the
run and towards the end we obtained conversion of 98%, during
the middle this dropped to 77%. We repeated the 50 mL run but
this time inputting a reagent solution that was 0.75 M in iodoarene.
Again at the start and the end of the run we obtained near-
quantitative conversion. An improvement was observed during
the middle of the run (86% conversion), but the reaction was still
gas limited. Performing the 50 mL run inputting a reagent solution
that was 0.5 M in iodoarene led to a 98% conversion throughout
and gas bubbles were clearly seen after the back-pressure regulator.
This shows that, due to dispersion factors, while smaller-scale trails
can be performed at higher input reagent concentrations, scale-up
may require operating at more dilute conditions.
10 s of venting, the gas flow was turned off while continuing to
vent into the hood. The “gas out” plugs were then closed. Next,
the system pump was set to “solvent” and the flow rate set to
0.250 mL min while heating each reactor to 120 C. Once the
system pressure reached ~60 psi (~4 bar) the gas was turned on at
180 psi for the duration of the experiment. The system was then
allowed to continue to pressurize from liquid flow, until it reached
250–275 psi (~19 bar). At this pressure, gas bubbles are present
after the back pressure regulator. Once a constant bubble stream
is seen, the system is ready for loading the reagents.
-
1
◦
General procedure for alkoxycarbonylation reactions on the
5 mmol scale: The ethoxycarbonylation of 4-iodotoluene
To a 10 mL test tube was added anhydrous ethanol (5 mL),
followed by DBU (0.820 mL, 5.5 mmol, 1.1 equiv.). The solution
was thoroughly mixed before adding Pd(OAc) (5.0 mg, 0.022
2
mmol). The solution was again mixed thoroughly. If upon
examining the test tube particulate matter remained, the mixture
was gently heated until no particles remained.
With this solution prepared, the flow reactor was readied. Once
this was complete, 4-iodotoluene (1.09 g, 5 mmol) was added to
the test tube containing the other reaction components and mixed
thoroughly. The reagent line was then placed inside the prepared
solution, reaching the bottom of the tube. The line was secured
to the test tube and the reactor pump switched from “solvent”
to “reagent”. Product collection was commenced immediately
after this switch. After the reaction mixture had been completely
inputted into the reactor, the reactor pump was turned back to
“solvent”.
Experimental section
General experimental
Tothe product mixture was added diethylether (50mL)followed
by a volume (about 80 mL) of brine and the biphasic mixture
transferred to a 500 mL separatory funnel. The layers were
separated and the aqueous layer was re-extracted with diethyl ether
All aryl iodides were obtained from commercial suppliers and
used without further purification. For the alcohol substrates:
ethanol used was 200 proof (anhydrous); commercially available
anhydrous 1-propanol and 2-propanol were used. Reactions were
3
times (equal volumes, about 80 mL each). The combined organic
1
run without the need for exclusion of air. H-NMR spectra were
layers were then mixed with an equal volume (about 300 mL) of
hexanes. This facilitated removal of ethanol from the diethyl ether
layer. After waiting about 10 min the ethanol layer was removed.
recorded at 293 K on either a 300 MHz or 400 MHz spectrometer,
products being characterized by comparison with literature data
or in-house samples.
The diethyl ether layer was then dried with Na
2
SO
4
and filtered.
The solvent was removed by rotary evaporation and the product
conversion determined by H-NMR spectroscopy.
1
14,15
Apparatus
13
Experiments were performed on a Vapourtec R series. The
system was equipped with two gas loading reactor coils. The
General procedure for alkoxycarbonylation reactions on the
5 mmol scale: The ethoxycarbonylation of 4-iodotoluene
“
reagent out” port on the first reactor coil was connected to the
2
second reactor coil “reagent in” port using a 32 mm length of
tubing. The “reagent out” port of the second reactor was equipped
with a 250 psi back pressure regulator after which a length of
tubing led to a waste or collection flask. To load the reagent gas
to each reactor, the tubing from the gas tank was equipped with a
T-piece and connected to the “gas in” ports on each reactor coil.
The “gas out” ports were plugged. At this point the system was
then primed and flushed with anhydrous ethanol for ~5 min at
To a 125 mL Erlenmeyer flask was added anhydrous ethanol
(50 mL), followed by DBU (4.1 mL, 27.5 mmol, 1.1 equiv.). The
solution was thoroughly mixed before adding Pd(OAc) (28 mg,
2
0.125 mmol, 0.5 mol%). The solution was again mixed thoroughly.
If upon examining the test tube particulate matter remained, the
mixture was gently heated until no particles remained.
With this solution prepared, the flow reactor was readied. Once
this was complete, 4-iodotoluene (5.45 g, 25 mmol) was added
to the flask containing the other reaction components and mixed
thoroughly. The reagent line was then placed inside the prepared
solution, reaching the bottom of the flask. The line was secured
to the flask and the reactor pump switched from “solvent” to
“reagent”. Aliquots of 10 mL exiting the reactor were collected
-
1
8
mL min . The pump was turned off to prime the gas line. The
plugs on the “gas out” were loosened and the carbon monoxide
tank, was fitted with a regulator capable of delivering pressures of
up to 250 psi, was opened at 180 psi. CAUTION: Make sure the
“
gas out” apertures vent into a fumehood since carbon monoxide
is toxic and may be fatal if inhaled in significant quantities. After
This journal is © The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 6575–6578 | 6577

View Article Online
and then the product mixture isolated and analyzed using the
same procedure as for the smaller-scale trials.
C. O. Kappe, Chem. Asian J., 2010, 5, 1274; (c) D. Mark, S. Haeberle,
G. Roth, F. von Stetten and R. Zengerle, Chem. Soc. Rev., 2010, 39,
1
153; (d) N. Kockmann and D. M. Roberge, Chem. Eng. Technol., 2009,
2, 1682; (e) C. Wiles and P. Watts, Eur. J. Org. Chem., 2008, 1655.
3
3
4
http://www.thalesnano.com/products/h-cube.
Conclusion
For selected recent examples see: (a) M. Baumann, I. R. Baxendale and
S. V. Ley, Synlett, 2010, 5, 749; (b) B. Clapham, N. S. Wilson, M. J.
Michmerhuizen, D. P. Blanchard, D. M. Dingle, T. A. Nemcek, J. Y.
Pan and D. R. Sauer, J. Comb. Chem., 2008, 10, 88; (c) H. H. Horvath,
G. Papp, C. Csajagi and F. Joo, Catal. Commun., 2007, 8, 442; (d) G.
Szoellosi, B. Herman, F. Fulop and M. Bartok, React. Kinet. Catal.
Lett., 2006, 88, 391; (e) V. Franckevicius, K. R. Knudsen, M. Ladlow,
D. A. Longbottom and S. V. Ley, Synlett, 2006, 6, 889; (f) R. V. Jones,
L. Godorhazy, N. Varga, D. Szalay, L. Urge and F. Darvas, J. Comb.
Chem., 2006, 8, 110.
For selected recent examples see: (a) X. Ye, M. D. Johnson, T. Diao,
M. H. Yates and S. S. Stahl, Green Chem., 2010, 12, 1180; (b) S. H u¨ bner,
U. Bentrup, U. Budde, K. Lovis, T. Dietrich, A. Freitag, L. K u¨ pper and
K. J a¨ hnisch, Org. Process Res. Dev., 2009, 13, 952; (c) B. N. Zope and
R. J. Davis, Top. Catal., 2009, 52, 269; (d) A. Lapkin, B. Bozkaya and
P. K. Plucinski, Ind. Eng. Chem. Res., 2006, 45, 2220.
In summary, a prototype tube-in-tube reactor in which it is possible
to load gas and heat simultaneously has been used for continuous-
flow alkoxycarbonylation reactions of aryl iodides. A range of aryl
iodides can be transformed to the corresponding esters in excellent
◦
conversion by reaction at 120 C using 0.5 mol% palladium acetate
as the catalyst with no additional ligand required. Small-scale
optimization and substrate screening runs were followed by scale-
up. Due to dispersion factors, while smaller-scale trails can be
performed at 1 M input reagent concentrations, scale-up required
operating at 0.5 M. The methodology using the heated tube-in-
tube reactor was found to be operationally far superior to a plug-
flow approach. No longer was it necessary to painstakingly opti-
mize bubble size and use a significant excess of carbon monoxide
gas. An additional benefit was that the significant build up of
palladium black at the point of exit through the back-pressure
regulator that occurred with the plug-flow approach due to excess
CO poisoning the catalyst was no longer an issue when using the
heated tube-in-tube reactor. Work is now underway to use the
reactor for other chemical transformations, including expanding
the substrate scope of the alkyoxycarbonylation methodology
to aryl bromides and chlorides by using a phosphine-ligated
palladium complex as the catalyst.
5
6
(a) K. Jahnisch, M. Baerns, V. Hessel, W. Ehrfeld, V. Haverkamp, H.
Lowe, C. Wille and A. Guber, J. Fluorine Chem., 2000, 105, 117; (b) C. B.
McPake, C. B. Murray and G. Sandford, Tetrahedron Lett., 2009, 50,
1
674.
7
8
9
A. Kumar, T. Chilongo, J. Dewulf, S. J. Ergas and H. Van Langenhove,
Bioresour. Technol., 2010, 101, 8955.
C. B. Kelly, C. Lee, M. A. Mercadante and N. E. Leadbeater, Org.
Process Res. Dev., 2011, 15, 717–720.
M. O’Brien, I. R. Baxendale and S. V. Ley, Org. Lett., 2010, 12, 1596.
0 A. Polyzos, M. O’Brien, T. P. Petersen, I. R. Baxendale and S. V. Ley,
Angew. Chem., Int. Ed., 2011, 50, 1190.
11 M. O’Brien, N. Taylor, A. Polyzos, I. R. Baxendale and S. V. Ley, Chem.
Sci., 2011, 2, 1250–1257.
1
1
2 (a) C. M. Kormos and N. E. Leadbeater, Org. Biomol. Chem., 2007,
, 65; (b) M. D. Bowman, J. R. Schmink, C. M. McGowan, C. M.
Kormos and N. E. Leadbeater, Org. Process Res. Dev., 2008, 12, 1078;
5
Acknowledgements
(
c) M. Iannelli, F. Bergamelli, C. M. Kormos, S. Paravisi and N. E.
Leadbeater, Org. Process Res. Dev., 2009, 13, 634.
3 http://www.vapourtec.co.uk/.
Vapourtec is thanked for equipment support. David Griffin and
Duncan Guthrie are particularly thanked for their input. Funding
from the National Science Foundation (CAREER award CHE-
1
1
4 Literature data for compounds prepared in this study – Table 2, entries
1–6: R. Shang, Y. Fu, J. B. Li, S. L. Zhang, Q. X. Guo and L. Liu,
J. Am. Chem. Soc., 2009, 131, 5738 Table 2, entry 8: Q. Liu, G. Li, J.
He, J. Liu, P. Li and A. Lei, Angew. Chem. Int. Ed. Engl, 2010, 49, 3371.
0847262) is acknowledged.
1
1
5 Table 2, entry 7 – Propyl-4-methylbenzoate: H NMR (400 MHz,
Notes and references
CDCl
3
) d (ppm) 1.03 (t, J = 7.51 Hz, 3 H) 1.79 (m, J = 7.17 Hz, 2
H) 2.40 (s, 3 H) 4.26 (t, J = 6.66 Hz, 2 H) 7.23 (d, J = 7.85 Hz, 1 H)
13
1
For an overview see: Chemical Reactions and Processes under Flow
Conditions, S. V. Luis and E. Garcia-Verdugo, ed., Royal Society of
Chemistry, Cambridge UK, 2010.
7.94 (d, J = 8.19 Hz, 2 H). C NMR (101 MHz, CDCl ) d (ppm) 10.78,
3
21.88, 22.39, 66.59, 128.06, 129.28, 129.81, 143.66, 167.01. MS (EI):
178 (M+, 6%), 136 (71%), 119 (100%), 91 (45%), 65 (15%), 39 (4%).
2
For recent reviews see: (a) C. Wiles and P. Watts, Chem. Commun.,
HRMS (ESI+), calcd for C11H O [M]+: 179.1072, found 179.1055.
FTIR (cm , neat, ATR) 3039 (w), 2965 (m), 2878 (w), 1713 (s), 1613
(m), 1461 (w), 1265 (vs), 1174 (m), 1104 (s), 748 (s).
14 2
-1
2
011, 47, 6512, DOI: 10.1039/c1cc00089f; (b) J. Wegner, S. Ceylan and
A. Kirschning, Chem. Commun., 2011, 47, 4583–4592; T. Razzaq and
6
578 | Org. Biomol. Chem., 2011, 9, 6575–6578
This journal is © The Royal Society of Chemistry 2011