The first continuous flow hydrogenation of amides to amines

Full text of DOI:10.1002/cctc.201300431

CHEMCATCHEM
COMMUNICATIONS
DOI: 10.1002/cctc.201300431
The First Continuous Flow Hydrogenation of Amides to
Amines
[
a]
[b]
[b]
Jacorien Coetzee, Haresh G. Manyar, Christopher Hardacre, and
[a]
David J. Cole-Hamilton*
Amines are a versatile class of compounds with applications
ranging from dyes, solvents and detergents to the pharma-
ceutical industry. Amine functionalities are often introduced by
initial amide formation followed by a reduction step using re-
ducing agents such as LiAlH4 or boranes. These reducing
agents are, however, hazardous and difficult to handle, particu-
larly on large scale, and their use often involves complex and
wasteful workup procedures. Thus, routes to selective and en-
vironmentally benign amide reduction are sought after, and
for this reason, the American Chemical Society Green Chemis-
try Institute (GCI) and members of the Pharmaceutical Round
Table have shortlisted amide hydrogenation as one of their
geneous catalyst for amide hydrogenation is currently limited
to nonaromatic substrates owing to the complication of un-
wanted ring hydrogenation, the ease of catalyst separation as-
sociated with such systems is a great advantage to their imple-
mentation in continuous flow systems for industrial applica-
tions. Herein, we report the first selective catalytic hydrogena-
tion of amides to amines in continuous flow by using a bi-
[10]
metallic TiO -supported Pt–Re-based catalyst.
2
The air-stable 4%Pt–4%Re/TiO2 catalyst employed in this
study was first reported by some of us to be catalytically active
[10]
towards amide hydrogenation. A range of bimetallic Pt–Re-
based catalysts supported on CeZrO , TiO2 and Al O were
4
2
3
[
1]
three most desirable reactions for development. Although
a number of successful amide reductions through catalytic hy-
tested for the selective hydrogenation of N-methylpyrrolidin-2-
one to N-methylpyrrolidine in hexane under an atmosphere of
hydrogen (20 bar, 1 bar=100 kPa) at 1208C. Under these con-
[
2]
drosilylation have been reported, catalytic hydrogenation
represents a promising alternative, as water is generated as
ditions, 4%Pt–4%Re/TiO displayed the highest activity and
2
[
3]
[4]
the only by-product. The groups of Saito, Milstein, and Ber-
gave almost full conversion after 24 h. For this system, the
nature of the solvent was shown to play an important role,
and the rate of the reaction was found to decrease in the
order: hexane>tetrahydrofuranꢀdiethyl ether>methanol>
[
5]
gens reported homogeneously catalysed hydrogenation of
amides, but in all cases, the reduction proceeds either through
CÀN bond cleavage to give amines and alcohols or through
monohydrogenation to give hemiaminals. We recently report-
ed the first successful homogeneous Ru-based catalyst capable
of selectively hydrogenating amides to amines without CÀN
bond cleavage, even in the presence of aromatic ring sys-
[10]
methyl tert-butyl ether.
To assess the performance of this catalyst under continuous
flow conditions, we developed a versatile flow reactor. The re-
actor is vertical with an upwards flow and has the facility to
[
6]
tems. The scope of this reaction is, however, currently limited
to substrates containing a phenyl ring directly attached to the
N atom and to primary amides. A number of bimetallic hetero-
geneous hydrogenation catalysts have been reported by the
flow liquids, gases and CO simultaneously through the verti-
2
cal-packed bed reactor containing the catalyst. The flowing
stream is then decompressed, and the products are collected
free from the catalyst and other impurities apart from side
products, solvent, and unreacted substrates. A schematic of
the reactor is shown in Figure 1.
[
7]
[8]
groups of Fuchikami and Whyman to give good conver-
sions of amides into amines, but they generally require fairly
harsh operating conditions. More recently, promising bimetallic
Prior to testing 4%Pt–4%Re/TiO2 in continuous flow, the
catalyst was tested in batch mode with an aromatic substrate
to determine the tolerance of this catalyst towards arene
functionalities. Acetanilide, a substrate that performed very
graphite-supported Pd–Re and TiO -supported Pt–Re based
2
catalysts capable of promoting amide hydrogenations under
mild reaction conditions were reported independently by the
[
9]
[10]
groups of Breit and Hardacre. Although the use of a hetero-
well with the homogeneous [Ru(acac) ]/triphos [acac=
3
acetylacetonate, triphos=1,1,1-tris(diphenylphosphinomethyl)-
ethane] system (Table 1, entry 1), was used as a test substra-
[
a] Dr. J. Coetzee, Prof. Dr. D. J. Cole-Hamilton
EastChem, School of Chemistry, North Haugh
University of St. Andrews
[6a]
te. Acetanilide (5 mmol) in hexane was heated at 1208C in
the presence of 4%Pt–4%Re/TiO (1.6 mol% in metal) under
2
St. Andrews, Fife
an atmosphere of hydrogen (20 bar, RT) for 16 h (Table 1,
entry 4). Although this resulted in hydrogenation of 36% of
the amide functionality, ring hydrogenation occurred for both
the final amine as well as the remaining substrate to give
a mixture of N-cyclohexyl-N-ethylamine (2) and N-cyclohexyl-
acetamide (3) as the final products (Scheme 1). In addition,
minor amounts of N-cyclohexylamine (4) and ethanol (5) were
produced as a result of CÀN bond cleavage.
KY16 9ST, Scotland (United Kingdom)
Fax: (+44)0-1334-463808
E-mail: djc@st-andrews.ac.uk
[
b] Dr. H. G. Manyar, Prof. Dr. C. Hardacre
CenTACat, School of Chemistry and Chemical Engineering
Queen’s University
Stranmillis Road, Belfast
BT9 5AG, Northern Ireland (United Kingdom)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201300431.
ꢀ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 2843 – 2847 2843

CHEMCATCHEM
COMMUNICATIONS
www.chemcatchem.org
mentary to our previously re-
ported homogeneous system in
which aliphatic substrates such
as N-methylpyrrolidin-2-one and
N-methylpropanamide
per-
formed rather poorly and gave
either no conversion (Table 1,
entry 2) or poor selectivity
[6a]
(
Table 1, entry 3).
Conversely,
hydrogenation of N-methylpyrro-
lidone or N-methylpropanamide
in batch mode with 4%Pt–
4
%Re/TiO2 under an atmo-
sphere of H2 (20 bar) at 1208C
gave 92–100% conversion with
75–100% selectivity towards the
desired amine (Table 1, entries 5
and 6).
Using the flow reactor de-
scribed in Figure 1 and 4%Pt–
4
%Re/TiO2 as the catalyst,
N-methylpyrrolidin-2-one could
be hydrogenated smoothly
Figure 1. Schematic drawing of the continuous flow reactor.
under continuous flow to pro-
duce N-methylpyrrolidine as the
only product. Upon passing a so-
lution of N-methylpyrrolidin-2-
one in hexane (0.33m, solution
Table 1. A comparison between the homogeneous and heterogeneous catalytic systems in the hydrogenation
of amides to amines in batch mode.
Entry
Substrate
Catalyst
t [h]
Product
Conv. [%]
100
Sel. [%]
À1
flow rate=0.06 mLmin ) over
[
a]
Homogeneous system
a fixed catalyst bed of 4%Pt–
4
%Re/TiO2 at 1208C, together
1
[Ru(acac)
3
]/triphos
16
92
0
with hydrogen (flow rate=
À1
1
90 mLmin ) under a total pres-
2
3
[Ru(acac)
[Ru(acac)
3
3
]/triphos
]/triphos
16
16
0
sure of 20 bar, full conversion
with 100% selectivity towards
N-methylpyrrolidine was ach-
[
b]
67
45
ieved over 8 h on stream
[c]
[11]
Heterogeneous system
(Figure 2,
*).
Pumping
N-methylpyrrolidin-2-one either
neat (Figure 2, ) or as an emul-
4
4%Pt–4%Re/TiO
2
16
100
0
&
sion in supercritical CO2 (total
pressure in this case increased to
5
6
4%Pt–4%Re/TiO
4%Pt–4%Re/TiO
2
2
16
16
100
97
100
74
1
00 bar; Figure 2, ~) led to no
catalytic activity. In the first in-
stance, this was presumably due
to the polarity of neat N-methyl-
pyrrolidin-2-one, and in the
second instance, this could be
due to the low solubility of the
[
a] Conditions: Amide (5 mmol), [Ru(acac)
MSA; 1.5 mol%), H (15 bar), 2208C, Hastelloy autoclave. [b] The main side products were methyldipropyl-
(1.6 mol% in
3
] (1 mol%), triphos (2 mol%), THF (10 mL), methanesulfonic acid
(
2
amine (25%) and dimethylpropylamine (12%). [c] Conditions: Amide (5 mmol), 4%Pt–4%Re/TiO
metal), hexane (15 mL), H (20 bar), 1208C, Hastelloy autoclave.
2
2
substrate in supercritical CO . If
2
the original N-methylpyrrolidin-
2-one solution in hexane (0.33m)
Despite the fact that this catalyst is not suitable for the hy-
drogenation of amide substrates containing arene functionali-
ties, it is very effective in catalysing the hydrogenation of ali-
phatic amides such as N-methylpyrrolidin-2-one and N-methyl-
propanamide. This catalytic system can be viewed as comple-
was again passed over the same catalyst bed (Figure 2, *),
a decrease in the activity was observed, and this suggests that
the use of supercritical CO or neat substrate leads to some de-
2
activation of the catalyst over time. This is attributable to
strong adsorption of the amide on the surface, which has also
ꢀ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 2843 – 2847 2844

CHEMCATCHEM
COMMUNICATIONS
www.chemcatchem.org
Scheme 1. Catalytic hydrogenation of acetanilide in batch mode with the
use of 4%Pt–4%Re/TiO .
2
Figure 3. Temperature-programmed oxidation (a) and reduction (b) profiles
for fresh and used 4%Pt–4%Re/TiO as the catalyst.
2
Figure 2. Hydrogenation of N-methylpyrrolidin-2-one with 4%Pt–4%Re/TiO
as the catalyst under continuous flow with hexane (*), neat (
), supercritical
2
&
[
11]
CO
2
(~) and hexane again (*).
been demonstrated in batch experiments by pre-treating the
catalyst with N-methylpyrrolidin-2-one in hexane for 3 h prior
[
10]
to the addition of H2.
The temperature-programmed oxidation (TPO) profile (Fig-
ure 3a) of used 4%Pt–4%Re/TiO shows an additional broad
2
peak (300–5008C) in comparison to the profile of the fresh cat-
alyst, which is likely to be attributed to adsorbed organic mole-
cules on the surface. The temperature-programmed reduction
(TPR) profile of the fresh catalyst shows low-temperature re-
duction of ReO (ꢀ2608C), which can be attributed to the
x
Figure 4. Hydrogenation of N-methylpyrrolidin-2-one with 4%Pt–4%Re/TiO
2
strong electronic perturbation of Re by Pt. On the contrary, the
TPR profile of the used catalyst indicates decreased interaction
between Pt and Re, which gives rise to an additional higher
as the catalyst under continuous flow with variation of the substrate flow
[11]
rate. In each case, the selectivity was 100% to N-methylpyrrole.
temperature reduction peak for ReO (4708C).
x
For a 0.67m solution of N-methylpyrrolidin-2-one in hexane
under H2 flow of 190 mLmin , a temperature of 1208C and
of N-methylpropanamide in hexane, the substrate was passed
over the catalyst bed at 1208C as an emulsion (0.94 mL sub-
strate in 85 mL hexane; substrate emulsion flow rate=
À1
a total pressure of 20 bar, the substrate solution flow rate
À1
À1
À1
could be increased to 0.12 mLmin without any drop in con-
0.12 mLmin ) together with H2 (flow rate=190 mLmin )
under a total pressure of 20 bar. Under these conditions, high
conversion (75–99%) and selectivity towards the secondary
amine (80–86%) were achieved continuously over 8 h on
stream (Figure 5); 1-propanol was obtained as the major side
product. Notably, given that N-methylpropanamide was fed as
an emulsion in hexane, there were some difficulties in main-
taining a constant and homogeneous substrate feed despite
version or selectivity. If the substrate solution flow rate was
À1
again doubled (0.24 mLmin ) to give even shorter residence
times, a sharp drop in the conversion was observed (Figure 4).
Using the same slightly deactivated catalyst bed as that
used during earlier continuous flow experiments with super-
critical CO , N-methylpropanamide could also be hydrogenated
2
successfully under continuous flow. Owing to the low solubility
ꢀ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 2843 – 2847 2845

CHEMCATCHEM
COMMUNICATIONS
www.chemcatchem.org
time on stream, and the greatest deactivation (observed as
a sharp drop in conversion) occurred within the first four hours
on stream (Figure 6, *).
A number of samples taken at different time intervals were
analysed for metal content by using inductively coupled
plasma mass spectrometry (ICP-MS). In all cases, metal leaching
was negligible, and the values measured were generally below
the detection limit (1 ppm; Table S1, Supporting Information).
Under batch conditions, recycling of the catalyst also resulted
[10]
in significant deactivation. This was attributed to the loss of
Pt–Re interactions as observed in the TPR profile, probably as
a result of the exposure of the catalyst to air upon recycling,
which caused an exotherm on the partially reduced surface.
Under flow conditions, the catalyst is maintained under reduc-
ing conditions at all times Therefore, the observed deactivation
is likely to be caused by the presence of a strongly adsorbed
material blocking the surface sites, as observed in the TPO
profile.
Figure 5. The conversion (*) and selectivity (*) for the catalytic hydrogena-
2
tion of N-methylpropanamide with 4%Pt–4%Re/TiO as the catalyst under
continuous flow over an 8 h period on stream by using a recycled catalyst
[
11]
In conclusion, we have demonstrated the first catalytic hy-
drogenation of amides to amines under continuous flow con-
ditions. Aliphatic amides such as N-methylpyrrolidin-2-one and
N-methylpropanamide were hydrogenated successfully under
mild conditions in continuous flow by using the heteroge-
neous bimetallic catalyst 4%Pt–4%Re/TiO2.
bed.
vigorous stirring. Owing to this complication, the conversion
for this reaction varies slightly over time on stream.
The low boiling point of hexane resulted in poor mass re-
covery for continuous flow reactions with this solvent. Thus,
decane was employed as a less volatile and more environmen-
tally benign alternative. With decane as the solvent, a mass re-
covery of greater than 90% could be maintained for the hy-
drogenation of N-methylpyrrolidin-2-one under flow over
a 10 h period on stream (Figure 6, ~). To evaluate the stability
Acknowledgements
We thank the European Commission for a Fellowship (J.C.). Re-
search leading to these results received funding from the Europe-
an Community’s Seventh Framework Programme [FP7/2007-
of the catalyst over time, the catalytic hydrogenation of
N-methylpyrrolidin-2-one in decane was performed under low
conversion with a shorter residence time (substrate solution
2
2
013] for SYNFLOW under grant agreement no. NMP2-LA-2010-
46461.
À1
À1
flow rate=0.24 mLmin , H flow=190 mLmin ). Under these
2
conditions, gradual catalyst deactivation was observed with
Keywords: amides · amines · continuous flow · heterogeneous
catalysis · hydrogenation
[
1] D. J. C. Constable, P. J. Dunn, J. D. Hayler, G. R. Humphrey, J. L. Leazer, Jr.,
R. J. Linderman, K. Lorenz, J. Manley, B. A. Pearlman, A. Wells, A. Zaks,
T. Y. Zhang, Green Chem. 2007, 9, 411–420.
[
2] a) S. Hanada, E. Tsutsumi, Y. Motoyama, H. Nagashima, J. Am. Chem. Soc.
J. Am. Chem. Soc 2009, 131, 15032–15040; b) S. Das, S. Zhou, D. Addis,
S. Enthaler, K. Junge, M. Beller, Top. Catal. 2010, 53, 979–984; c) S. Das,
B. Join, K. Junge, M. Beller, Chem. Commun. 2012, 48, 2683–2685; d) S.
Das, B. Wendt, K. Mçller, K. Junge, M. Beller, Angew. Chem. 2012, 124,
1
694–1698; Angew. Chem. Int. Ed. 2012, 51, 1662–1666.
3] T. Miura, I. E. Held, S. Oishi, M. Naruto, S. Saito, Tetahedron Lett. 2013,
4, 2674–2678.
4] E. Balaraman, B. Gnanaprakasam, L. J. W. Shimon, D. Milstein, J. Am.
Chem. Soc. 2010, 132, 16756–16758.
[
[
[
[
5
5] J. M. John, S. H. Bergens, Angew. Chem. 2011, 123, 10561–10564;
Angew. Chem. Int. Ed. 2011, 50, 10377–10380.
6] a) J. Coetzee, D. L. Dodds, J. Klankermayer, S. Brosinski, W. Leitner,
A. M. Z. Slawin, D. J. Cole-Hamilton, Chem. Eur. J. 2013, DOI: 10.1002/
chem.201204270; b) A. A. NfflÇez Magro, G. R. Eastham, D. J. Cole-Hamil-
ton, Chem. Commun. 2007, 3154–3156; c) D. L. Dodds, J. Coetzee, J.
Klankermayer, S. Brosinski, W. Leitner, D. J. Cole-Hamilton, Chem.
Commun. 2012, 48, 12249–12262 (published additions & corrections).
Figure 6. The stability of the 4%Pt–4%Re/TiO
2
catalyst during the hydroge-
nation of N-methylpyrrolidin-2-one over a 10 h period on stream showing
the conversion (*), mass recovery (~) and selectivity (*).
[7] C. Hirosawa, N. Wakasa, T. Fuchikami, Tetahedron Lett. 1996, 37, 6749–
[
11]
6752.
ꢀ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 2843 – 2847 2846

CHEMCATCHEM
COMMUNICATIONS
www.chemcatchem.org
[
8] a) G. Beamson, A. J. Papworth, C. Philipps, A. M. Smith, R. Whyman, J.
Catal. 2010, 269, 93–102; b) G. Beamson, A. J. Papworth, C. Philipps,
A. M. Smith, R. Whyman, J. Catal. 2011, 278, 228–238.
9] M. Stein, B. Breit, Angew. Chem. 2013, 125, 2287–2290; Angew. Chem.
Int. Ed. 2013, 52, 2231–2234.
out at beginning of the reaction until the reactor has filled up, with the
fill-up time dependant on the substrate solution flow rate. For reactions
performed with a recycled catalyst bed, the fill-up times were shorter
despite prior washing of the catalyst bed with hexane and thorough
[
purging under a flow of H
2
for 1–2 hrs.
[
[
10] R. Burch, C. Paun, X. M. Cao, P. Crawford, P. Goodrich, C. Hardacre, P. Hu,
L. McLaughlin, J. Sµ, J. M. Thompson, J. Catal. 2011, 283, 89–97.
11] Please note that no conversion is shown over the first 2 hrs on stream
for reactions done with a fresh catalyst bed, and 1 h for reactions per-
formed with a recycled catalyst bed. This is because no products flow
Received: June 5, 2013
Published online on August 19, 2013
ꢀ
2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2013, 5, 2843 – 2847 2847