Continuous flow hydrogenation using an on-demand gas delivery reactor

Article abstract of DOI:10.1021/op300019w

A continuous-flow approach to the hydrogenation of alkenes utilizing Wilkinson's catalyst is reported. The approach relies on a newly developed coil design in which it is possible to load gas and heat the reaction mixture simultaneously. The hydrogenation of various substrates has been performed successfully on small scale and can be scaled up substantially.

Full text of DOI:10.1021/op300019w

Article
pubs.acs.org/OPRD
Continuous Flow Hydrogenation Using an On-Demand Gas Delivery
Reactor
Michael A. Mercadante, Christopher B. Kelly, Christopher (Xiang) Lee, and Nicholas E. Leadbeater*
Department of Chemistry, University of Connecticut, 55 North Eagleville Road, Storrs, Connecticut 06269, United States
*
S Supporting Information
ABSTRACT: A continuous-flow approach to the hydrogenation of alkenes utilizing Wilkinson’s catalyst is reported. The
approach relies on a newly developed coil design in which it is possible to load gas and heat the reaction mixture simultaneously.
The hydrogenation of various substrates has been performed successfully on small scale and can be scaled up substantially.
INTRODUCTION
■
Micro- and mesofluidic flow devices have proven to be
powerful enabling technologies in organic synthesis. Inherent
in these devices are aspects of enhanced safety, ease of scale-up,
and efficient mixing of reagents. One area of flow chemistry
that is currently of considerable interest is interfacing reactive
gases with a stream of liquid reagents. One approach is to use
mechanical mixing of the two phases, which is employed by
1,2
3
instruments such as the H-cube (for hydrogenation), O-cube
4
5
(
for ozonolysis), and X-cube (for carbonylation). Alter-
natively, a semipermeable membrane can be used to introduce
gas into the reagent stream, a technique pioneered by Ley and
Figure 1. Schematic of the gas reactor unit used.
6
−9
co-workers.
In their “tube-in-tube” design, a liquid reagent
stream flows within a gas-permeable Teflon AF-2400 mem-
brane and is housed in a PTFE outer tubing filled with a
gaseous reagent. This allows for diffusion-mediated gas transfe6r
into the liquid stream. It has been used for hydrogenation,
and, when two are interfaced with the flow unit, pictorially in
Figure 2.
RESULTS AND DISCUSSION
■
7
8
carbonylation, and ozonolysis as well as reactions involving
Our initial objective was to optimize the reaction conditions
for catalytic homogeneous hydrogenation using the on-demand
gas reactor. To achieve this, we chose to use anethole (p-
methoxyphenylpropene, 1) as a test substrate and RhCl(PPh )
9
other gases.
One of the drawbacks of these current designs is the inability
to load gas and heat the reaction mixture simultaneously.
Current approaches rely on saturation of either the reagent or
solvent flow prior to heating. This often necessitates running
reactions at low concentrations or performing multiple passes
through the entire system, thus limiting throughput. Recently,
we have had access to a prototype “tube-in-tube” reactor in
which it is possible to load gas and heat simultaneously. In a
proof-of-concept study we have used this for palladium-
3
3
11
(
Wilkinson’s catalyst ) as the catalyst for hydrogenation.
Compound 1 was chosen judicially as it would be much more
difficult to hydrogenate than terminal or electron-deficient
alkenes and therefore, would provide optimal conditions for
most substrates. Our optimization data are presented in Table
1
. We decided to employ the same configuration as used for our
alkoxycarbonylation methodology, namely two gas coils in
series, both heated to the desired reaction temperature and
10
catalyzed alkoxycarbonylation reactions. Here we report an
extension of the application of this reactor to catalytic homo-
geneous hydrogenation. The reactor design incorporates a
modified “tube-in-tube” model where the inner tube contains
the gaseous reagent, while the liquid reaction flows around it
contained in an outer stainless steel reactor coil. The inner
diffusion membrane is constructed from PTFE-type fluoropol-
ymer, offering excellent chemical compatibility, combined with
good high-temperature performance. The stainless steel outer
tube, as well as being robust, eliminates loss of dissolved gas. As
the liquid is in contact with this thin-walled stainless steel outer
tube, excellent heat transfer and therefore temperature control
are possible. The gas reactor is shown schematically in Figure 1
12
both individually fed with gas. A schematic representation of
the setup is shown in Figure 3. Working on a 5 mmol scale,
using 1 mL dichloromethane and employing a catalyst loading
of 0.5 mol %, we passed the reaction mixture at a flow rate of
0
.5 mL/min through the two gas coils heated to 70 °C and
obtained a 28% conversion to 1a (Table 1, entry 1). Doubling
the catalyst loading to 1 mol % almost doubled the conversion
Special Issue: Continuous Processes 2012
Received: January 20, 2012
Published: April 4, 2012
©
2012 American Chemical Society
1064
dx.doi.org/10.1021/op300019w | Org. Process Res. Dev. 2012, 16, 1064−1068

Organic Process Research & Development
Article
was significantly lower than the first-generation coil. This was
reflected by a decrease in conversion when compared to an
identical run in the first-generation coils (Table 1, entry 5). To
overcome this drop in conversion we increased the catalyst
loading to 1.2 mol % (Table 1, entry 6) and then reduced the
flow rate to 0.25 mL/min, leading to a 97% conversion (Table
1
, entry 7). When operating at this lower flow rate, we observed
the formation of small quantities of an insoluble precipitate
which we envisaged could lead to blockage of the back-pressure
regulator located at the end of the heated zone. We believe
that the precipitate is formed as a result of the formation of
HRhCl (PPh ) (due to the reaction being performed in a
2
3 2
chlorinated solvent) or due to dimerization of the catalyst over
1
3
the now extended reaction time. While insoluble in
dichloromethane, the precipitate readily dissolves in acetone.
To avoid blockage of the back-pressure regulator we used an
14
approach previously reported by us. We intercept the product
stream with a flow of a suitable organic solvent upon exiting the
heated zone, thus solubilizing the product and allowing it to
pass through the back-pressure regulator unimpeded. In this
case, when we intercepted the product stream with a flow of
acetone just before it enters the back-pressure regulator, we
were able to perform the hydrogenation reaction without any
blockage issues.
Figure 2. Gas reactor interfaced with the flow apparatus.
Table 1. Optimization of reaction conditions for the
a
continuous-flow hydrogenation of anethole
With the optimized conditions and coil configuration in
hand, we moved on to screening a variety of substrates (Table 2).
We were pleased to find that, in all but one case, near
quantitative conversions were obtained. The alkane derived
from 1 was obtained in excellent isolated yield under the
optimized conditions (entry 1). We next examined two styrene
derivatives, knowing that Wilkinson’s catalyst is very effective in
hydrogenating terminal phenyl-substituted alkenes (Table 2,
entries 2−3). The slightly lower isolated yield of the alkane
derived from 2 can be attributed to the volatility of the resulting
product, whereas we see a marked increase in yield with the
alkane from the less volatile benzodioxole, 3 (Table 1, entry 3).
We next explored the reduction of two representative electron-
deficient alkenes (Table 2, entries 4 and 5), which afforded
excellent yields of their corresponding alkanes. Less conjugated
alkene systems, namely 1-dodecene, 6, and eugenol, 7, were
also screened (Table 2, entries 6 and 7), and both furnished the
desired product in excellent yield. The alkane derived from
flow rate
temp.
catalyst loading
(mol %)
conversion
(%)
entry
(mL/min)
(°C)
b
1
0.5
0.5
0.5
0.5
0.5
0.5
0.25
70
70
0.5
1.0
1.0
1.0
1.0
1.2
1.2
28
50
70
92
72
89
97
b
2
13
b
3
100
125
125
125
125
b
4
c
5
c
6
c
7
a
Reaction conducted on a 5 mmol scale using 1 mL dichloromethane
as solvent in each case. Using a first-generation coil design. Using the
second-generation coil design.
b
c
(
allyloxy)benzene, 8, was isolated in slightly lower yield, likely
for the same rationale as that in the case of 2. Knowing that
terminal alkenes are more susceptible to hydrogenation than
internal or other more substituted alkenes, we decided to
screen dihydrocarvone, 9, and diethyl fumarate, 10 (Table 2,
entries 9 and 10). Both substrates could be hydrogenated
effectively and in high yield.
Figure 3. Schematic of the flow configuration used for the
hydrogenation of alkenes.
Not all substrates are compatible with our optimized reaction
conditions. While most were tolerated, several functionalities
proved problematic (Table 2, entries 11−14). Both the dia-
cetate and alcohol likely failed to undergo complete hydro-
genation due to competitive side reactions, causing subsequent
catalyst deactivation (Table 2, entries 11 and 12). We believe
that, in the case of the alcohol, a chelated metal complex is
formed, evidenced by the change in color (green) and the
(
Table 1, entry 2). By increasing the temperature to 100 °C,
the conversion increased to 70% (Table 1, entry 3).
Performing the reaction using 5 mmol of anethole in just
1
mL dichloromethane, while simultaneously increasing the
temperature slightly to 125 °C, led to a 92% conversion
(
Table 1, entry 4).
Due to the stress of operating at high temperatures for
1
broadening of peaks in the H NMR. In the case of the
prolonged periods of time, we became concerned about the
integrity of the inner gas membrane. We therefore transitioned
from the prototype reactor to a newer, more durable produc-
tion version of the reactor. As a consequence of this switch, we
found that the gas permeability of the second-generation coil
diacetate, we observed a Tsuji−Trost type reaction (elimination
of one acetate moiety) followed by polymerization. The high
temperatures of our optimized conditions likely led to
competitive polymerization of N-vinylpyrrolidinone (Table 2,
entry 13). Also, Wilkinson’s catalyst is known to react with
1
065
dx.doi.org/10.1021/op300019w | Org. Process Res. Dev. 2012, 16, 1064−1068

Organic Process Research & Development
Article
a
Table 2. Continuous-flow hydrogenation of alkenes using the on-demand gas delivery reactor
a
Reactions conditions unless otherwise indicated: 5 mmol substrate scale, 1 mL CH Cl , 1.2 mol % catalyst loading, 125 °C coil temperature, 0.25
2
2
b
1
c
mL/min flow rate. A quantitative conversion obtained for all reactions except entry 5 (93% conversion by H NMR analysis). Reaction performed
on a 90 mmol scale using 30 mL of CH Cl .
2
2
aldehydes to give the decarbonylated product, which indeed
was the case with trans-2-undecenal, 14 (Table 2, entry 14).
The final objective of our study was to assess the scalability of
our methodology. We previously observed during our
15
1
066
dx.doi.org/10.1021/op300019w | Org. Process Res. Dev. 2012, 16, 1064−1068

Organic Process Research & Development
Article
alkoxycarbonylation work that when small portions of reagents
were passed through the coils, there is considerable dispersion
mixed using a stir bar until it became a completely
homogeneous, clear, dark-red solution. The flow reactor was
readied using the equipment manufacturer’s suggested start-up
sequence, followed by heating the reactor coils to 125 °C. The
reaction mixture was then loaded into the reactor. Product
collection was commenced immediately after this switch. After
the reaction mixture had been completely loaded into the
reactor, the reactor pump was set back to pumping dichloro-
methane. After the product had been fully discharged from the
reactor coils, the resulting clear-yellow solution could then be
purified.
10
in the length of the two reactors. We therefore decided to
process materials using a 3 M concentration, corresponding to a
−1
throughput of 45 mmol h under continuous operation. We
selected trans-chalcone, 4, as the representative alkene for scale-
up. We obtained a quantitative conversion and 99% isolated
yield of the desired alkane (Table 2, entry 15).
EXPERIMENTAL SECTION
Description of the Apparatus. Experiments were
performed on a Vapourtec R series. The system was equipped
with two gas-loading reactor coils. The “reagent out” port on
the first reactor coil was connected to the “reagent in” port on
second reactor coil using a 32 mm length of tubing. The
■
16
CH Cl and acetone were removed using rotary evapora-
2
2
tion,leaving the crude product. The crude product was dis-
solved in a small amount of the elution solution (8:2 hexane/
EtOAc) and loaded onto a plug of silica. The plug was rinsed
thoroughly with eluting solution, and the solvent was stripped
in vacuo in a 50 °C water bath. This process was repeated until
“reagent out” port of the second reactor was linked into a
T-piece which allowed a flow of acetone to mix with the
reaction mixture departing the second reactor coil. The T-piece
was finally equipped with a 250 psi back-pressure regulator after
which was a length of tubing leading to a waste or collection
flask.
The system was initially primed using the equipment
manufacturer’s suggested startup sequence. CAUTION: Make
sure the “gas out” apertures vent into a fumehood since hydrogen is
highly flammable. After priming the unit, the reactor coils were
each heated to 125 °C. Once at temperature, the system was ready
for loading the reagent solution.
constant weight was obtained. Pure 5a, was obtained as an off-
1
white powdery solid (18.910 g, 99%). H NMR (CDCl , 500
3
MHz) δ 0.99 (t, J = 7.25 Hz, 3 H) 1.67 (sxt, J = 7.60 Hz, 2 H)
2
.58 (t, J = 7.60 Hz, 2 H) 3.83 (s, 3 H) 6.88 (d, J = 8.83 Hz, 2
13
H) 7.14 (d, J = 8.20 Hz, 2H) C NMR (CDCl ,125 MHz) δ
3
1
4.02 (CH ), 25.06 (CH ), 37.41 (CH ), 55.44 (CH ), 113.88
3 2 2 3
(
(
(
CH), 128.53(CH), 135.02 (C), 157.91 (C) GC−MS (EI) 151
+ +
[M + 1] , 5%), 150 ([M] , 44%), 122 (18%), 121 (100%), 91
16%), 78 (16%), 77 (17%), 65 (6%).
CONCLUSION
General Procedure for Small-Scale Hydrogenation
Reactions (Hydrogenation of Anethole). A 10 mL test
tube was charged with Wilkinson’s catalyst (0.0555 g,
■
In summary, we report a continuous flow approach to the homo-
geneous catalytic hydrogenation of alkenes, utilizing Wilkinson's
catalyst. The new reactor coil employed in this methodology
enabled the continuous input of gas while simultaneously allowing
for heating of the reaction mixture. A range of alkenes was hydro-
genated using this approach. The methodology was amenable
to significant scale-up; we processed approximately 0.1 mol of
material over the period of 2 h. This would correspond to
production of around 2 mol of product per day. A number of flow
units running in parallel could increase this throughput further.
Comparison of our approach with that of others using continuous-
flow processing shows that our system is able to process reaction
mixtures at significantly higher substrate concentrations and
hence throughput. This is important when considering the current
drive towards process intensification. Work is underway to use this
apparatus for other reactions using gaseous reagents.
0
1
.060 mmol 1.2 mol %), anethole, 1 (0.741 g, 5 mmol), and
mL of dichloromethane. The solution was thoroughly mixed
using sonication until it became a completely homogeneous
clear, dark-red solution. The flow reactor was readied using
the equipment manufacturer’s suggested start-up sequence
followed by heating the reactor coils to 125 °C. The reaction
mixture was then loaded into the reactor. Product collection
was commenced immediately after this switch. After the
reaction mixture had been completely loaded into the reactor,
the reactor pump was set back to pumping dichloromethane.
After the product had been fully discharged from the reactor
coils, the resulting clear-yellow solution could then be purified.
Dichloromethane and acetone were removed using rotary
evaporation, leaving the crude product. The crude product was
dissolved in a small amount of the elution solution (9:1
hexane/EtOAc) and loaded onto a plug of silica. The plug was
rinsed thoroughly with eluting solution, and the solvent was
ASSOCIATED CONTENT
Supporting Information
■
*
S
17
stripped in vacuo in a room temperature water bath. This
process was repeated until constant weight was obtained. Pure
Experimental procedures, characterization data, and spectra of
the compounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
1
8
1
-methoxy-4-propylbenzene, 1a, was obtained as a clear,
1
colorless oil (0.668 g, 89%). H NMR (CDCl , 500 MHz) δ
3
0
.99 (t, J = 7.25 Hz, 3 H) 1.67 (sxt, J = 7.60 Hz, 2 H) 2.58 (t,
AUTHOR INFORMATION
Corresponding Author
nicholas.leadbeater@uconn.edu.
■
J = 7.60 Hz, 2 H) 3.83 (s, 3 H) 6.88 (d, J = 8.83 Hz, 2 H) 7.14
1
3
(
(
d, J = 8.20 Hz, 2H) C NMR (CDCl ,125 MHz) δ 14.02
CH ), 25.06 (CH ), 37.41 (CH ), 55.44 (CH ), 113.88 (CH),
3
*
3
2
2
3
Notes
1
1
7
28.53(CH), 135.02 (C), 157.91 (C) GC−MS (EI) 151 ([M +
+ +
The authors declare no competing financial interest.
] , 5%), 150 ([M] , 44%), 122 (18%), 121 (100%), 91 (16%),
8 (16%), 77 (17%), 65 (6%).
ACKNOWLEDGMENTS
General Procedure for Larger-Scale Hydrogenation
■
Reactions (Hydrogenation of trans-Chalcone). A 50 mL
conical flask was charged with Wilkinson’s catalyst (1.000 g,
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-0847262) is acknowledged, and the University of
1
.08 mmol 1.2 mol %), trans-chalcone, 5 (19.33 g, 92.8 mmol),
and dichloromethane (30 mL). The solution was thoroughly
1
067
dx.doi.org/10.1021/op300019w | Org. Process Res. Dev. 2012, 16, 1064−1068

Organic Process Research & Development
Article
Connecticut Office for Undergraduate Research is thanked for
supplementary funding (C.L.).
REFERENCES
■
(
1) (a) Wiles, C.; Watts, P. Micro Reaction Technology in Organic
Synthesis; CRC Press, Boca Raton, 2011. (b) Luis, S. V.; Garcia-
Verdugo, E., Eds. Chemical Reactions and Processes under Flow
Conditions; Royal Society of Chemistry: Cambridge, 2010.
(
2) For recent reviews, see: (a) Wiles, C.; Watts, P. Chem. Commun.
011, 47, 6512. (b) Wegner, J.; Ceylan, S.; Kirschning, A. Chem.
Commun. 2011, 47, 4583. (c) Razzaq, T.; Kappe, C. O. Chem. Asian J.
010, 5, 1274. (c) Mark, D.; Haeberle, S.; Roth, G.; von Stetten, F.;
Zengerle, R. Chem. Soc. Rev. 2010, 39, 1153.
3) For a review, see: Irfan, M.; Glasnov, T. N.; Kappe, C. O.
ChemSusChem 2011, 4, 300.
2
2
(
(
4) Irfan, M.; Glasnov, N. T.; Kappe, C. O. Org. Lett. 2011, 13, 984.
(
5) Csajagi, C.; Borcsek, B.; Niesz, K.; Kovacs, I.; Szekelyhidi, Z.;
Bajko, Z.; Urge, L.; Darvas, F. Org. Lett. 2008, 10, 1589.
(
6) O’Brien, M.; Baxendale, I. R.; Ley, S. V. Org. Lett. 2010, 12, 1596.
(
7) Koos, P.; Gross, U.; Polyzos, A.; O’Brien, M.; Baxendale, I. R.;
Ley, S. V. Org. Biomol. Chem. 2011, 9, 6903.
8) O’Brien, M.; Taylor, N.; Polyzos, A.; Baxendale, I. R.; Ley, S. V.
Chem. Sci. 2011, 2, 1250.
9) (a) Polyzos, A.; O’Brien, M.; Petersen, T. P.; Baxendale, I. R.;
(
(
Ley, S. V. Angew. Chem., Int. Ed. 2011, 50, 1190. (b) Bourne, S. L.;
Koosa, P.; O’Brien, M.; Martin, B.; Schenkel, B.; Baxendale, I. R.; Ley,
S. V. Synlett 2011, 2643. (c) Kasinathan, S.; Bourne, S. L.; Tolstoy, P.;
Koos, P.; O’Brien, M.; Bates, R. W.; Baxendale, I. R.; Ley, S. V. Synlett
2
011, 2648.
10) Mercadante, M. A.; Leadbeater, N. E. Org. Biomol. Chem. 2011,
, 6575.
11) For background on the use of Wilkinson’s catalyst in
hydrogenation reactions, see: Jardine, J. F. Prog. Inorg. Chem. 1981,
8, 63.
12) A screen using only one coil gave a conversion of 36% under the
(
9
(
2
(
conditions shown in Table 1, entry 1. Adding a second coil allowed for
a longer reaction time, and the drop in conversion to 28% (Table 1,
entry 1) before further optimization is due to the overall dilution of
the reaction mixture (via dispersion) as it proceeded through the
length of the reactor.
(
13) Birch, A. J.; Williamson, D. H. Homogeneous Hydrogenation
Catalysts in Organic Synthesis. Organic Reactions; Wiley-VCH:
Weinheim, 1976; pp 1−186.
(
14) Kelly, C. B.; Lee, C.; Leadbeater, N. E. Tetrahedron Lett. 2011,
52, 263.
(
15) Ohno, K.; Tsuji, J. J. Am. Chem. Soc. 1968, 90, 99.
(
16) See http://www.vapourtec.co.uk.
(
17) Note that it was necessary to keep the bath water at room
temperature in most cases due to the volatility of the alkane products
18) Dhakshinamoorthy, A.; Sharmila, A.; Pitchumani, K. Chem.
Eur. J. 2010, 16, 1128.
(
1
068
dx.doi.org/10.1021/op300019w | Org. Process Res. Dev. 2012, 16, 1064−1068