Reductive amination by continuous-flow hydrogenation: Direct and scalable synthesis of a benzylpiperazine

Article abstract of DOI:10.1055/s-0032-1316550

A benzylpiperazine was prepared directly from the corresponding benzaldehyde and piperazine on a continuous-flow hydrogenation apparatus. The synthesis was protecting group free, safe, and environmentally friendly. Large-scale synthesis under flow hydrogenation conditions was also demonstrated. Georg Thieme Verlag Stuttgart · New York.

Full text of DOI:10.1055/s-0032-1316550

PAPER
▌2469
paper
Reductive Amination by Continuous-Flow Hydrogenation: Direct and
Scalable Synthesis of a Benzylpiperazine
Continuous-Flow Hydrogenation
Jing Liu,* Anne E. Fitzgerald, Neelakandha S. Mani
Janssen Research & Development, L.L.C., 3210 Merryfield Row, San Diego, CA 92121, USA
Fax +1(858)3203354; E-mail: jliu30@its.jnj.com
Received: 19.03.2012; Accepted after revision: 18.05.2012
screening within a wide range of pressure and tempera-
ture. Elimination of H2 storage and containment of cata-
lyst also alleviate many safety concerns. In a pioneering
study, Ley and co-workers demonstrated the reduction of
imines to amines on a prototype flow hydrogenation in-
strument.10 However, there were few further reports on re-
ductive amination by flow hydrogenation, particularly
with secondary amines.11 In one of our drug discovery
programs, benzylpiperazine 1 (Figure 1) was needed in
large quantity. Here we report the development, optimiza-
tion, and scale-up of a direct synthesis of 1 on a flow-
hydrogenation apparatus.
Abstract: A benzylpiperazine was prepared directly from the cor-
responding benzaldehyde and piperazine on a continuous-flow hy-
drogenation apparatus. The synthesis was protecting group free,
safe, and environmentally friendly. Large-scale synthesis under
flow hydrogenation conditions was also demonstrated.
Key words: amination, hydrogenation, green chemistry, ben-
zylation, palladium
The reductive amination reaction is a key method for C–
N bond formation. An aldehyde or ketone reacts with an
amine in the presence of a reducing agent to give a higher
order amine. The most popular reducing agents include
NaBH3CN1 and NaBH(OAc)3,2 thanks to their high selec-
tivity and excellent functional group compatibility. How-
ever, from a green chemistry perspective, they are far
from ideal. Reactions with NaBH3CN generate cyanide-
containing toxic waste, and less than 0.5 wt% of
NaBH(OAc)3 is incorporated into the product as a single
hydride.3 Other reducing agents, such as silanes,4 zinc,5
and tin hydrides,6 suffer from similar drawbacks.
O
O
F
F
HN
N
1
Figure 1 Compound of interest: benzylpiperazine 1
Benzylpiperazine 1 was initially prepared in gram quanti-
ties by reductive amination of benzaldehyde 2 followed
by Boc removal (Scheme 1). This was certainly a viable
way in small scale, but the inefficiencies became more
pronounced upon scale up. For example, a 200 g synthesis
would require 455 g of NaBH(OAc)3, with virtually all the
mass relegated to waste stream. Use of Boc protecting
group not only mandated an additional deprotection step,
but also increased the cost of goods (Boc-piperazine is al-
most 200 times more expensive than piperazine).12 The
route has a Process Mass Intensity (PMI)13 of 91, indicat-
ing significant adverse environmental impact for a rela-
tively small molecule.
In theory, H2 is the ideal reducing agent for reductive am-
ination. It is completely incorporated into product, and
water is the only by-product. In practice, however, H2 is
usually only used in the so-called ‘indirect’ reductive am-
inations with primary amines, where an imine is pre-
formed prior to reduction.7 Reactions with secondary
amines are much more challenging,7a since carbonyl
group is often reduced concomitantly to give alcohol side
product. In addition, storage and handling of H2 usually
require dedicated facility and stringent safety regulation –
significant practical inconvenience in a laboratory setting.
In recent years, continuous-flow hydrogenation reactors
have become commercially available.8,9 H2 is generated
on demand, mixed with reactants stream, and passed
through a catalyst cartridge. The design enables rapid
To design a more efficient synthesis of 1, we first replaced
NaBH(OAc)3 with H2 using a flow hydrogenation reac-
tor.14 The initial experiment afforded desired product 3 in
low yield (Scheme 2). As expected, the main product was
BocN
NH
O
O
HCl
O
F
F
F
F
BocN
1
1,4-dioxane
r.t., 4 h
N
NaBH(OAc)3
THF, 0 °C to r.t.
99%
O
OHC
75%
2
Scheme 1 Initial synthesis of 1
SYNTHESIS 2012, 44, 2469–2473
Advanced online publication: 02.07.2012
0
0
3
9
-
7
8
8
1
1
4
3
7
-
2
1
0
X
DOI: 10.1055/s-0032-1316550; Art ID: SS-2012-M0284-OP
© Georg Thieme Verlag Stuttgart · New York

2470
J. Liu et al.
PAPER
ed to the appropriate solvent and stirred at room tempera-
ture. At various time points, a portion was taken out and
subjected to hydrogenation. Product mixture was ana-
lyzed by quantitative HPLC (Figure 2). Overall, longer
aging time led to more product 1 and less alcohol 4. The
effect varied in different solvents, and methanol gave the
best results. It is likely that intermediates such as hemiam-
inal were formed during the aging period, and the nature
of intermediate(s) varies in different solvents.15
O
O
F
F
BocN
N
BocN
3 (19%)
NH
O
O
F
F
+
H2 (1 atm), Pd/C
EtOH, 50 °C
flow rate: 1 mL/min
OHC
2
O
F
F
HO
O
4 (59%)
Scheme 2 Synthesis of 1 employing a flow hydrogenation reactor
alcohol 4, a well-known problem when secondary amines
are used.
Instead of optimizing the above reaction, we decided to
pursue a more ambitious goal and replace N-Boc pipera-
zine with piperazine. This would give a direct, protecting
group free synthesis of 1. We expected it to be challenging
since multiple unproductive pathways were distinct possi-
bilities (Scheme 3). In addition to undesired aldehyde re-
duction, formation of 5 by double alkylation became
likely. Although we had yet to observe toluene 6, it could
also be formed under harsher conditions by reduction of
alcohol 4 or debenzylation of amine 1.
Figure 2 Effect of aging time in different solvents. Conditions: pi-
perazine (4 equiv), 0.05 M, 10% Pd/C cartridge, 1 atm, 50 °C, 1
a
b
mL/min flow rate. Based on quantitative HPLC. The time for
which the reaction mixture was stirred at room temperature prior to
hydrogenation.
O
O
F
F
Various transition-metal catalysts were then screened
(Table 1). The highest yield of 1 was obtained with 10%
Pd/C. No aldehyde reduction was observed with 20%
Pd(OH)2/C, but the reaction did not reach completion.
Other catalysts led to lower conversion and yield. Based
on these results, 10% Pd/C was used in further optimiza-
tion studies.
HN
H
+
NH
O
2
H2
O
O
O
O
F
F
F
F
HN
+
N
HO
Table 1 Screening of Catalystsa
1
4
Entry
Catalyst
Yield (%)b
O
O
O
F
F
F
N
+
1
2
4
5
6
N
F
O
1
2
3
4
5
6
Pd/C (10%)
Pt/C (10%)
Pd(OH)2/C (20%)
Raney Ni
76
61
72
23
33
0
3
15
13
0
6
5
5
6
8
5
0
0
0
0
0
0
5
21
23
71
59
95
O
F
F
+
O
0
6
Rh/C (5%)
0
Scheme 3 Possible reaction pathways and products
Ru/C (5%)
0
First, the experiment was performed under the same con-
ditions as in Scheme 2, but with excess piperazine (4
a General conditions: piperazine (4 equiv), 0.05 M in MeOH, 1 atm
H2, 50 °C, 1 mL/min flow rate; the reactants solution was left at r.t.
equiv). This effectively suppressed double alkylation (5, overnight prior to hydrogenation.
b Based on quantitative HPLC.
2%), but the desired product 1 was only obtained in 8%
yield. The major product was again alcohol 4 (89%). The
addition of acids such as AcOH (1–3 equiv) only slightly
Next, the effects of other reaction parameters including
increased the yield of 1. Interestingly, leaving the solution
for a period of time prior to hydrogenation led to more
product formation and less aldehyde reduction. A system-
atic study was carried out to quantify the effect of ‘aging’
in various solvents. Aldehyde 2 and piperazine were add-
temperature, pressure, flow rate, and concentration were
examined (Table 2). Increased pressure or decreased flow
rate (Table 2, entries 2, 3) led to significant formation of
alcohol 4 side product. On the other hand, higher temper-
ature (70 °C) led to clean formation of 1 (entry 4). The
Synthesis 2012, 44, 2469–2473
© Georg Thieme Verlag Stuttgart · New York

PAPER
Continuous-Flow Hydrogenation
2471
Table 2 Screening of Conditionsa
Entry
Temp ( °C)
Pressure (bar)
Flow (mL/min) [C] (M)
Piperazine (equiv) Yield (%)b
1
2
4
5
6
1
2
3
4
5
6
7
50
50
50
70
70
70
70
1
20
1
1.0
1.0
0.5
1.0
1.0
1.0
1.0
0.05
0.05
0.05
0.05
0.2
4
76
18
21
3
0
0
0
15
82
79
0
6
0
0
0
3
0
0
0
0
0
0
6
4
4
1
4
100
75
1
4
22
0
0
1
0.05
0.2
1.5
1.5
90
0
10
10
1
78
0
6
a General conditions: 10% Pd/C cartridge, MeOH as solvent; the reactants solution was left at r.t. overnight prior to hydrogenation.
b Based on quantitative HPLC.
concentration was then increased from 0.05 M to 0.2 M to gave very good conversion (entry 5). After further optimi-
gain higher material throughput, but incomplete conver- zation, excellent yield and satisfactory material through-
sion was observed (entry 5). Not surprisingly, decreasing put were achieved after a single pass (entry 7). The yield
the amount of piperazine led to formation of dimer 5 (en- could be further increased to 98% when the product solu-
tries 6, 7).
tion from entry 7 was hydrogenated again using the same
conditions and catalyst cartridge (entry 8).
The above conditions were successfully used to prepare
gram quantities of product. Further scale-up was limited The reaction mixture mainly contained product 1 and ex-
by the capacity of catalyst cartridge and relatively low cess piperazine. We took advantage of solubility differ-
material throughput. For synthesis on 200–300 gram ences between 1 and piperazine to develop a simple
scale, we employed the H-Cube Midi, which is designed isolation protocol. After solvent swap from methanol to
for larger-scale operations. It features higher flow rate, toluene, excess piperazine was easily removed by filtra-
larger catalyst cartridges, and increased H2 generation ca- tion and aqueous wash. Crystallization from heptanes af-
pacity. While the instrument operates under similar mech- forded 1 in good yield and excellent purity. The material
anisms as the smaller H-Cube, results sometimes vary due output for the flow hydrogenation was >100 grams per
to differences in catalyst particles sizes and batches.16 In day (Scheme 4). A total of 300 grams of 1 was prepared
our case, initial experiment on the H-Cube Midi gave low by this protocol.17 Compared to the original procedure
conversion (Table 3, entry 1). Lowering concentration (Scheme 1), the new synthesis was much more efficient
and flow rate increased conversion to some extent (entries and environmentally friendly. The PMI value was low-
2–4). On the other hand, reaction with 20% Pd(OH)2/C
Table 3 Screening of Conditions on H-Cube Midia
Entry
Catalyst
Flow (mL/min)
Concn (M)
Yield (%)b
1
2
4
5
6
1
2
3
4
5
6
7
8c
Pd/C (10 mol%)
10
10
5
0.25
0.15
0.15
0.10
0.15
0.30
0.50
28
38
62
56
84
92
93
98
72
61
36
41
2
0
0
0
0
1
0
0
0
0
1
2
3
3
4
2
2
0
0
0
0
Pd/C (10 mol%)
Pd/C (10 mol%)
Pd/C (10 mol%)
5
Pd(OH)2/C (20 mol%)
Pd(OH)2/C (20 mol%)
Pd(OH)2/C (20 mol%)
5
10
0
5
4
6
5
0
0
0
a General conditions: 70 °C, 1 atm H2, piperazine (4 equiv), MeOH as solvent; the reactants solution was left at r.t. overnight prior to hydroge-
nation.
b Based on quantitative HPLC.
c The product solution from entry 7 was hydrogenated again using the same conditions and catalyst cartridge.
© Georg Thieme Verlag Stuttgart · New York
Synthesis 2012, 44, 2469–2473

2472
J. Liu et al.
PAPER
ed to approximately 20 mL and a second batch of product 1 was ob-
tained (6.7 g, 5%). The combined yield was 115 g (83%).
ered from 92 to 26, and the cost-of-goods was reduced by
over 60%.
1H NMR (400 MHz, CDCl3): δ = 7.11 (d, J = 0.9 Hz, 1 H), 6.99 (dd,
J = 9.0, 0.9 Hz, 1 H), 6.95 (d, J = 9.0 Hz, 1 H), 3.45 (s, 2 H), 2.92–
2.83 (m, 4 H), 2.39 (s, 4 H).
13C NMR (101 MHz, CDCl3): δ = 143.89, 142.72, 134.68, 131.65
(t, JC,F = 255.3 Hz), 123.88, 110.13, 108.82, 63.10, 54.38, 46.07.
MS (ESI+): m/z calcd for C12H15F2N2O2 [M + H]+: 257.1; found:
256.9.
NH
HN
O
O
F
F
F
F
HN
(4 equiv)
N
H
O
O
0.5 M in MeOH
H-Cube Midi
20% Pd(OH)2/C cartridge
70 °C, 1 atm, 6 mL/min
2
1
O
115 g, 83%
Scheme 4
Supporting Information for this article is available online at
http://www.thieme-connect.com/ejournals/toc/synthesis.SnoIufproi
m
tgioSrantnugIifoop
r
itmnatr
The optimized conditions could be easily adopted for tra-
ditional batch hydrogenation (1 atm H2, 20% Pd(OH)2/C,
10 wt%, 60 °C, 3 h; 93% yield). This demonstrated that
flow hydrogenation was, at a minimum, a valuable tool
for rapid optimization. All reaction parameters could be
easily modified, a very small amount of starting materials
were consumed in each run, and results were obtained
within minutes. More importantly, for this reaction, flow
hydrogenation was also advantageous on larger scale
thanks to enhanced safety. Methanol is known to ignite
easily upon contact with palladium and air, especially at
elevated temperature as in our case.18 This is a significant
safety concern in large-scale batch operations, but much
less so in flow hydrogenation since catalyst is securely en-
closed and not exposed to air.
References
(1) Borch, R. F.; Bernstein, M. D.; Durst, H. D. J. Am. Chem.
Soc. 1971, 93, 2897.
(2) (a) Abdel-Magid, A. F.; Carson, K. G.; Harris, B. D.;
Maryanoff, C. A.; Shah, R. D. J. Org. Chem. 1996, 61, 3849.
(b) Abdel-Magid, A. F.; Mehrman, S. J. Org. Process Res.
Dev. 2006, 10, 971; and references cited therein.
(3) Trost, B. M. Science 1991, 254, 1471.
(4) Apodaca, R.; Xiao, W. Org. Lett. 2001, 3, 1745.
(5) Micovic, I. V.; Ivanovic, M. D.; Piatak, D. M.; Bojic, V. D.
Synthesis 1991, 1043.
(6) Suwa, T.; Sugiyama, E.; Shibata, I.; Baba, A. Synlett 2000,
556.
(7) For select examples, see: (a) Robichaud, A.; Ajjou, A. N.
Tetrahedron Lett. 2006, 47, 3633. (b) Xing, L.; Cheng, C.;
Zhu, R.; Zhang, B.; Wang, X.; Hu, Y. Tetrahedron 2008, 64,
11783. (c) Erathodiyil, N.; Ooi, S.; Seayad, A. M.; Han, Y.;
Lee, S. S.; Ying, J. Y. Chem.–Eur. J. 2008, 14, 3118.
(d) Bhor, M. D.; Bhanushali, M. J.; Nandurkar, N. S.;
Bhanage, B. M. Tetrahedron Lett. 2008, 49, 965.
(8) (a) Jones, R. V.; Godorhazy, L.; Varga, N.; Szalay, D.; Urge,
L.; Darvas, F. J. Comb. Chem. 2006, 8, 110.
In conclusion, we have developed a direct synthesis of
benzylpiperazine 1 from unprotected piperazine and benz-
aldehyde 2 using H2-mediated reductive amination. The
reaction was rapidly developed and scaled up under con-
tinuous-flow hydrogenation conditions. The process is
safe, environmentally friendly, and scalable. The com-
plete scope of the reaction is being studied and will be re-
ported in due course.
(b) Yoswathananont, N.; Nitta, K.; Nishiuchi, Y.; Sato, M.
Chem. Commun. 2005, 40. (c) O’Brien, M.; Taylor, N.;
Polyzos, A.; Baxendale, I. R.; Ley, S. V. Chem. Sci. 2011, 2,
1250.
(9) For select examples of flow-hydrogenation applications,
see: (a) Irfan, M.; Petricci, E.; Glasnov, T. N.; Taddei, M.;
Kappe, C. O. Eur. J. Org. Chem. 2009, 1327. (b) Knudsen,
K. R.; Holden, J.; Ley, S. V.; Ladlow, M. Adv. Synth. Catal.
2007, 349, 535. (c) Franckevicius, V.; Knudsen, K. R.;
Ladlow, M.; Longbottom, D. A.; Ley, S. V. Synlett 2006,
889.
1H and 13C NMR spectra were recorded on a 400 MHz NMR spec-
trometer. HRMS (ESI) was performed on a MicroTof apparatus. All
reactions were monitored by HPLC. MeOH, THF, toluene, and
DMF were dried by passage through two alumina columns. Other
solvents and reagents were purchased from commercial sources and
used without further purification.
1-(2,2-Difluorobenzo[1,3]dioxol-5-ylmethyl)piperazine (1)
A 2 L Erlenmeyer flask was charged with piperazine (185 g, 2.15
mol, 4.0 equiv), 2,2-difluorobenzo[1,3]dioxole-5-carbaldehyde (2;
100 g, 0.54 mol, 1.0 equiv), and MeOH (1.08 L). The solution was
stirred at r.t. for 18 h. The solution was passed twice through the H-
Cube Midi flow-hydrogenation instrument with a new 20%
Pd(OH)2/C MidiCart cartridge at the following settings: 70 °C, 1
atm pressure, 6 mL/min flow rate, and 10% excess H2 production.
After removal of MeOH, toluene (1.20 L) was added and the mix-
ture was stirred at r.t. for 18 h. The resulting white suspension was
filtered and the collected solid was rinsed with toluene (200 mL).
The combined filtrate was washed with H2O (2 × 300 mL), dried
(Na2SO4), filtered, and concentrated. The resulting oil was dis-
solved in heptane (100 mL). Upon the addition of small amount of
product crystals as seed crystals, the solution quickly became a
thick white suspension. It was cooled to 0 °C and filtered. The white
solid was dried under vacuum at 50 °C for 24 h to give the product
1 as a white solid (108 g, 78%). The heptane filtrate was concentrat-
(10) Saaby, S.; Knudsen, K. R.; Ladlow, M.; Ley, S. V. Chem.
Commun. 2005, 2909.
(11) At the time of manuscript preparation, another example of
reductive amination with a secondary amine on a flow-
hydrogenation apparatus was reported: Cooper, C. G. F.;
Lee, E. R.; Silva, R. A.; Bourque, A. J.; Clark, S.; Katti, S.;
Novorozhkin, V. Org. Process Res. Dev 2012, 16, 1090.
(12) The price difference is on a per mole basis. From the same
vendor (Sigma-Aldrich), N-Boc piperazine costs $335/25 g,
or $2500/mol; piperazine costs $16/100 g, or $14/mol at the
time the manuscript was written.
(13) Process Mass Intensity (PMI) is the total mass of raw
materials input divided by the mass of final Active
Pharmaceutical Ingredient (API) output. It is one of the
benchmarks to measure the environmental impact and
sustainability of chemical processes. For a review of such
benchmarks, see: Constable, D. J. C.; Curzons, A. D.;
Cunningham, V. L. Green Chem. 2002, 4, 521.
Synthesis 2012, 44, 2469–2473
© Georg Thieme Verlag Stuttgart · New York

PAPER
Continuous-Flow Hydrogenation
2473
(14) Reactions were conducted using a commercially available
instrument, the H-Cube (manufactured by ThalesNano
Nanotechnology Inc. in Budapest, Hungary).
(15) A preliminary NMR experiment was performed to monitor
formation of intermediate (s). Equal molar of 2 and
piperazine were dissolved in CD3OD and allowed to stand at
r.t. for 48 h. 1H NMR showed mainly three compounds in the
ratio of 1:1:0.2. The structures were tentatively assigned to
aldehyde 2, piperazine hemiaminal methyl ether, and
dimethyl ketal, respectively.
the nature of catalyst and the batches. For an excellent
analysis of various parameters that can impact cartridge
performances, see: Bryan, M. C.; Wernick, D.; Hein, C. D.;
Petersen, J. V.; Eschelbach, J. W.; Doherty, E. M. Beilstein
J. Org. Chem. 2011, 7, 1141.
(17) Keith, J. M.; Liu, J. PCT Int. Appl WO139951, 2011; Chem.
Abstr. 2011, 155, 637818.
(18) Bretherick’s Handbook of Reactive Chemical Hazards; Vol.
1, 7th ed; Urben, P. G., Ed.; Elsevier: Amsterdam, 2007,
1980.
(16) Cartridges for H-Cube and H-Cube Midi contain catalyst of
different particle sizes. The actual difference varies based on
© Georg Thieme Verlag Stuttgart · New York
Synthesis 2012, 44, 2469–2473