Powerful Continuous-Flow Hydrogenation by using Poly(dimethyl)silane-Supported Palladium Catalysts

Article abstract of DOI:10.1002/cctc.201500973

We developed poly(dimethyl)silane-supported Pd catalysts that are readily prepared from Pd(OAc)₂, poly(dimethyl)silane, and Al₂O₃. The immobilization was achieved for the first time with a support that does not contain benzene rings. The Pd catalyst thus prepared was found to have higher hydrogenation activity than Pd/C. Furthermore, the catalyst was used in continuous-flow hydrogenation with various substrates, including simple liquid substrates (neat) and dissolved solid substrates. Vegetable oils, squalenes, and phosphatidylcholine were successfully hydrogenated on gram to kilogram scales.

Full text of DOI:10.1002/cctc.201500973

DOI: 10.1002/cctc.201500973
Communications
Powerful Continuous-Flow Hydrogenation by using
Poly(dimethyl)silane-Supported Palladium Catalysts
[
a]
[a]
[a, b]
[a]
Sh u¯ Kobayashi,* Mikiko Okumura, Yuichi Akatsuka,
Hiroyuki Miyamura,
[
a]
[a]
Masaharu Ueno, and Hidekazu Oyamada
We developed poly(dimethyl)silane-supported Pd catalysts that
Poly(methylphenyl)silane was chosen as an analogue of poly-
[
4]
are readily prepared from Pd(OAc) , poly(dimethyl)silane, and
styrene, because in microencapsulated and polymer-incarcer-
2
[
5]
Al O . The immobilization was achieved for the first time with
ated catalysts with polystyrenes as polymer backbones, it is
assumed that the benzene rings of the backbone are impor-
tant for the immobilization of the metal catalysts through elec-
tronic interactions (p electrons). This reasoning led us to
assume that benzene rings would also be important for immo-
bilization of polysilane-supported catalysts, and poly(methyl-
phenyl)silane was chosen accordingly. However, the availability
of poly(methylphenyl)silane is relatively limited. In contrast,
2
3
a support that does not contain benzene rings. The Pd catalyst
thus prepared was found to have higher hydrogenation activi-
ty than Pd/C. Furthermore, the catalyst was used in continu-
ous-flow hydrogenation with various substrates, including
simple liquid substrates (neat) and dissolved solid substrates.
Vegetable oils, squalenes, and phosphatidylcholine were suc-
cessfully hydrogenated on gram to kilogram scales.
[
6]
poly(dimethyl)silane is readily available; therefore, we exam-
ined the use of poly(dimethyl)silane in place of polystyrenes or
poly(methylphenyl)silane as a backbone for supported cata-
lysts.
Catalytic hydrogenation is one of the most important methods
[
1]
for organic synthesis in both academia and industry. It is
widely used for the synthesis of natural products, biologically
important compounds, active pharmaceutical ingredients, and
many intermediate compounds. The hydrogenation of fats,
phospholipids, and squalenes is performed to prevent oxida-
tion of their unsaturated bonds, which causes coloration or
We began our study by exploring the preparation of a poly(-
dimethyl)silane-supported palladium/alumina hybrid catalyst
[Pd/(DMPSi-Al O )] according to the method shown in
2
3
Scheme 1. This method was based on a procedure that was
[
2]
odor. The hydrogenated products thus prepared are used as
emulsifiers, compounding agents, and so on. Hydrogenation of
lecithin is an important process for the preparation of emulsifi-
ers. Catalytic hydrogenation by using Pd/C has been used in
a batch system for this process; however, a relatively high
loading of Pd/C is required and subsequent removal of Pd by
filtration is time consuming. Moreover, the recovered Pd has
low activity. To address these issues, alternative catalyst sys-
tems have been investigated, and among the various catalysts
tested, polysilane-supported Pd (Pd/PSi) was found to have
high activity. We developed this methodology further, and
herein, we describe a continuous-flow hydrogenation process
by using a novel Pd/PSi system as the catalyst.
2 3
Scheme 1. Preparation of Pd/(DMPSi-Al O ).
[
3]
used previously to prepare the Pd/MPPSi catalyst. Unexpect-
À1
edly, the Pd loading of the catalyst (56.6 mmolg ), which was
We previously prepared poly(methylphenyl)silane-supported
determined by inductively coupled plasma (ICP) analysis, was
almost the same as that of Pd/MPPSi. Thus, Pd was successfully
loaded onto poly(dimethyl)silane, which does not contain ben-
[
3]
Pd (Pd/MPPSi) and used it as a catalyst for several reactions.
zene rings. Pd/(DMPSi-Al O ) thus prepared was then character-
[
a] Prof. Dr. S. Kobayashi, M. Okumura, Y. Akatsuka, Dr. H. Miyamura,
Dr. M. Ueno, Dr. H. Oyamada
Department of Chemistry, School of Science
The University of Tokyo
Hongo, Bunkyo-ku
Tokyo, 113-0033 (Japan)
2
3
ized by scanning transmission electron microscopy (STEM) and
energy-dispersive X-ray spectroscopy (EDS). From the results of
these analyses, Pd nanoparticles with an average size of (6.4Æ
3
.4) nm (Figure 1a) were found to exist on silicon-rich areas
E-mail: shu_kobayashi@chem.s.u-tokyo.ac.jp
rather than on aluminum- or oxygen-rich areas (Figure 1b).
This observation suggested that these particles were formed
through reduction by the polysilane support, as observed in
[
b] Y. Akatsuka
Present address: Nikko Chemicals Co. Ltd.
Nasu Factory
[3]
our previous reports, and were simultaneously immobilized
1
844 Kamiishigami
on the polymer through electronic interactions.
Ohtawara, Tochigi, 324-0037 (Japan)
We used Pd/(DMPSi-Al O ), thus prepared, as a catalyst for
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201500973.
2
3
[
7]
the reduction of phosphatidylcholine in a batch system and
ChemCatChem 2015, 7, 4025 – 4029
4025
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
this flow setup, approximately
9
00 mg of the catalyst was
packed in a sus column (5 mm
diameter, 50 mm length), which
was then placed in an aluminum
heating block to control the re-
action temperature. The sub-
strate, either neat or in solution,
was introduced into the sus
column packed with the catalyst
by using a peristaltic pump and,
simultaneously, H gas was intro-
2
duced at atmospheric pressure
into the catalyst column through
a mass flow controller, which
regulated the amount of H2 as
required. After passing both the
substrate and H2 gas through
the column, the resulting liquid
was collected and analyzed.
First, we examined the contin-
uous-flow hydrogenation of
ethyl cinnamate (1a) as a model
reaction. Upon introducing this
compound to the catalyst
column at
a
flow rate of
À1
5
0 mLmin
under neat condi-
2 3
Figure 1. a) Particle-size distribution and b) STEM and EDS mapping images of Pd/(DMPSi-Al O ); green: Pd,
orange: Si, pink: Al, blue: O.
tions with H gas (2.0 equiv.) at
2
3
08C, the reaction proceeded
quantitatively over 5 h (Table 2,
compared the results with those obtained by using Pd/C,
which is commonly used in many laboratories. As shown in
Table 1, both Pd/(DMPSi-Al O ) and Pd/C showed high activity
entry 1). In this reaction, product 2a was obtained in pure
form (>99%), without any purification, by collecting the liquid
eluted downstream of the catalyst column directly. Decreasing
2
3
in the first run. After filtration, we then examined the second
use of the catalysts. Interestingly, whereas the activity of Pd/C
decreased significantly, the reduction proceeded smoothly
without loss of activity with Pd/(DMPSi-Al O ). Notably, the
the amount of H gas to 1.5 equivalents did not affect the re-
2
action outcome (Table 2, entry 2). However, if the flow rate of
À1
the substrate was increased to 100 mLmin , a small amount
(<5%) of the starting material was detected in the product
(Table 2, entry 3). In an attempt to improve the conversion, the
reaction temperature was increased to 508C, and again, quan-
2
3
high activity of the polymer-supported catalyst remained until
at least the fifth run.
Table 1. Catalyst recycling for the reduction of phosphatidylcholine (PC)
[
a]
with hydrogen in a batch system.
Table 2. Continuous-flow hydrogenation of ethyl cinnamate (1a).
[
b]
Catalyst
Yield [%]
Run 1
98
98
Run 2
97
78
Run 3
98
–
Run 4
98
–
Run 5
97
–
Pd/(PMPSi-Al
2
O
3
)
[c]
[a]
[b]
5
% Pd/C
Entry Substrate flow
H2
T
Conv.
[%]
Yield
[%]
TON/TOF
À1
À1
[
mLmin
50
]
[equiv.] [8C]
[h
–
–
–
–
]
[
2
a] Reaction conditions: Pd (0.2 mol%), H (0.5 MPa), 708C, 2 h. PC was
1
1
2
3
4
5
6
2.0 30 >95
1.5 30 >95
1.5 30 >95
1.5 50 >95
1.5 50 >95
1.5 50 >95
quant.
quant.
–
purchased from Nikko Chemicals Co., Ltd. [b] Determined by H NMR
spectroscopy. [c] Purchased from Wako Pure Chemical Industries, Ltd.
50
[
c]
100
100
150
200
quant.
quant.
–
[d]
5024/1004
–
[e]
[c]
We then turned our attention to continuous-flow hydroge-
nation by using Pd/(DMPSi-Al O ). Flow systems have several
1
2
3
[a] Determined by H NMR spectroscopy. No starting material was detect-
ed. [b] Yield of isolated product. [c] A small amount of the starting materi-
al was detected by H NMR spectroscopy (<5%). [d] 46.4 g. [e] Reaction
advantages over batch systems in terms of environmental
1
[
8,9]
impact, efficiency, and safety.
A schematic diagram of the
time: 1 h.
continuous-flow reactor that was used is shown in Figure 2. In
ChemCatChem 2015, 7, 4025 – 4029
www.chemcatchem.org
4026
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
Solid compounds could also
be submitted to this reaction by
using an appropriate solvent;
this was exemplified by the hy-
drogenation of diphenylacety-
lene (3b) and cholesterol (1g),
as shown in Scheme 2. Upon dis-
solving 3b in hexane, the reac-
tion proceeded quantitatively to
give dibenzyl (4b; 13.7 g) with
À1
a flow rate of 500 mLmin and
Figure 2. Schematic diagram of the continuous-flow reactor. PEEK=polyetheretherketone.
3.0 equivalents of H2 at 308C.
Cholesterol (1g), on the other
hand, was dissolved in a mixture
of iPrOH/hexane (3:1); in this
titative yield was obtained over 5 h (Table 2, entry 4). Indeed,
case, full conversion was obtained with a flow rate of
À1
at this temperature, the reaction proceeded quantitatively over
350 mLmin and 4.0 equivalents of H at 808C. In this reaction,
2
À1
5
h even if the flow rate of 1a was increased to 150 mLmin ,
92% of the desired hydrogenated cholestanol (2g) was ob-
which provided 46.4 g of product 2a in total (Table 2, entry 5).
tained; however, the remaining material was obtained in the
[
10]
The turnover number (TON) and turnover frequency (TOF) of
more oxidized form as cholestan-3-one (7). We believe that
the hydroxy group of either 1g or 2g underwent aerobic oxi-
dation catalyzed by Pd/(PMPSi-Al O ) with a trace amount of
À1
the Pd catalyst were calculated to be 5024 and 1004 h , re-
spectively. The finding that such a large amount of product
could be obtained with excellent TON and TOF by such
a simple and safe experimental procedure demonstrates the
remarkable efficiency of Pd/(DMPSi-Al O )-catalyzed continu-
2
3
[
4]
contaminating molecular oxygen.
Furthermore, a selection of vegetable oils were successfully
hydrogenated to afford the desired saturated oils, which are
commercial products, with high efficiency (Table 4). A sus
column (10 mm diameter, 150 mm length) was used in all
cases except for castor oil (Table 4, entry 3), for which a larger
column was used (10 mm diameter, 300 mm length). One of
the advantages of continuous-flow systems is the ease with
which the systems can be scaled up. Indeed, the production of
2
3
ous-flow hydrogenation. A further increase in the flow rate of
À1
1
a to 200 mLmin failed to provide full conversion; therefore,
we chose the conditions summarized in entry 5 (Table 2) as the
optimized conditions for 1a. It is noted that, under these con-
ditions, no leaching of palladium was detected by ICP analysis.
We then explored the scope of this flow system with
a series of alkenes 1 and 1-octyne (3a) under neat
conditions (Table 3). For monosubstituted alkenes
1
b and 1c, the hydrogenation proceeded quanti-
Table 3. Continuous-flow hydrogenation of alkenes 1b–f and 1-octyne (3a).
À1
tatively over 5 h with a flow rate of 300 mLmin
for the substrate, with 1.5 equivalents of H gas at
2
3
08C (Table 3, entries 1 and 2). In both cases, ap-
[
a]
Entry Substrate
Substrate
H
2
T
Product
Yield TON/TOF
proximately 90 g of products 2b and 2c was ob-
tained, respectively. Disubstituted alkene 1d was
also hydrogenated smoothly to give product 2d
in quantitative yield over 5 h by decreasing the
flow
À1
À1
[
mLmin
]
[equiv.] [8C]
[g]
[h
]
[
b]
1
300
300
100
50
1.5
1.5
1.5
2.0
30
30
30
70
91.8
16475/3295
16839/3368
6016/1203
2516/498
À1
flow rate to 100 mLmin with 1.5 equivalents of
1b
2b
2c
2d
2e
H gas at 308C (Table 3, entry 3). In the case of tri-
substituted alkene 1e, although a decrease in the
flow rate of the substrate, a higher temperature,
[b]
2
2
86.5
26.5
12.7
1
c
[
[
b]
b]
3
4
and an increase in the amount of H gas equiva-
2
lent were required, we could also obtain quantita-
tive yield over 5 h with this substrate (Table 3,
entry 4). Squalene (1 f), with six trisubstituted
alkene units, was also successfully hydrogenated
1d
1
e
[
c]
[b]
(
(
Table 3, entry 5). The terminal alkyne 1-octyne
5
6
40
15.0
3.0
120
30
6.5
304/101
3a) was also hydrogenated quantitatively to give
À1
1 f
2 f
4a
4
a with a flow rate of 100 mLmin for the sub-
[b]
100
23.2
3951/790
strate by using 3.0 equivalents of H gas at 308C
(
2
Table 3, entry 6). No leaching of palladium was
3
a
detected by ICP analysis in any of the cases exam-
ined.
[a] Yield of isolated product. [b] Quant. [c] Reaction time: 3 h.
ChemCatChem 2015, 7, 4025 – 4029
www.chemcatchem.org
4027
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
Experimental Section
Preparation of Pd/DMPSi-Al O
2
3
Pd(OAc) (13.5 g, 60.0 mmol) was added to a suspension of poly(di-
2
methyl)silane (100 g) and Al O (900 g) in toluene at 08C under an
2
3
argon atmosphere. After stirring the mixture for 50 min, MeOH
100 mL) was added and stirring was continued for 70 min at 08C.
(
The resulting gray solid was collected by filtration; washed with
MeOH, water, and acetone; and then dried under vacuum at 808C
to afford Pd/(PMPSi-Al O ) (1035.5 g). The loading of Pd was deter-
2
3
mined by ICP analysis after heating the catalyst in a mixture of sul-
À1
furic acid and nitric acid at 2008C (Pd loading: 56.6 mmolg ).
General procedure for the continuous-flow hydrogenation
of alkenes by Pd/DMPSi-Al O
2
3
Scheme 2. Continuous-flow hydrogenation of diphenylacetylene (3b) and
cholesterol (1g); in the latter reaction, product 2g and the corresponding
ketone (8%) were obtained.
Pd/PMPSi-Al O3 (916.0 mg) was packed in the catalyst column,
2
which was then heated to 508C. Neat ethyl cinnamate (1a) was
À1
fed into the column by using a peristaltic pump (150 mLmin ),
and H gas was introduced simultaneously by using a mass flow
2
À1
controller by downflow (30.2 mLmin , 1.5 equiv. relative to 1a).
The system was allowed to stabilize for 20 min, and then the prod-
Table 4. Continuous-flow hydrogenation of vegetable oils.
uct was collected over 5 h. Product 2a (46.4 g) was obtained, the
1
[
a]
purity of which was assessed by H NMR spectroscopy analysis.
Entry Substrate
Substrate flow
H
2
T
Yield
TOF
À1
À1
[
mLmin
]
[equiv.]
[8C] [%]
[h
]
[
b]
1
2
olive oil
macadamia 1.0
1.0
9.0–9.5 120 quant. 161–169
5.9–6.1 120 quant. 166–173
Acknowledgements
[
c]
nut oil
castor oil
[
d]
3
4
0.25
1.0
10
4.0
140 quant.
120 quant. 224
21
This work was partially supported by a Grant-in-Aid for Scientific
Research from the Japan Society for the Promotion of Science
[
e]
jojoba oil
(
JSPS), the Japan Science and Technology Agency (JST), and the
Ministry of Education, Culture, Sports, Science and Technology
MEXT). We thank Mr. Noriaki Kuramitsu (The University of Tokyo)
[
a] Determined by NMR spectroscopy. [b] Triglyceride in which the fatty
acid chains are oleate (60~80%), linolate (4~16%), and palmitate (9~
8%). [c] Triglyceride in which the fatty acid chains are oleate (60~80%)
1
(
and palmitoleate (15~30%). [d] Triglyceride in which the fatty acid chain
is mainly ricinoleate. [e] Ester shown below:
for STEM and EDS analyses.
Keywords: continuous flow · flow synthesis · heterogeneous
catalysis · hydrogenation · polysilanes
À1
castor oil could be increased to 1.5 kgh by using a large sus
[1] a) “Hydrogenation—heterogeneous”: in P. Gallezot, Encyclopedia of Cat-
alysis, Vol. 4 (Ed.: I. T. Horvath), Wiley-Interscience, New York, 2003,
pp. 17–55; b) S. L. Kiperman, in Studies in Surface Science and Catalysis,
Vol. 27 (Ed.: L. Cerveny), Elsevier, Amsterdam, 1986; c) P. Rylander in The
Catalytic Hydrogenation in Organic Syntheses (Ed.: P. Rylander), Academ-
ic Press, Waltham, MA, 1979.
[11]
column (100 mm diameter, 1000 mm length).
was found to be stable for more than one month.
The system
In summary, we developed poly(dimethyl)silane-supported
Pd catalysts as novel heterogeneous Pd catalysts that can be
readily prepared from commercially available materials. It is
noted that no benzene rings were required on the support to
immobilize Pd. Moreover, the Pd catalyst showed higher activi-
ty than Pd/C. The catalyst was used in the continuous-flow hy-
drogenation of a range of substrates, including simple liquid
substrates (neat) and dissolved solid substrates. Furthermore,
the production of vegetable oils, squalenes, and phosphatidyl-
choline was conducted on gram to kilogram scales. We expect
that this novel catalyst and powerful continuous-flow hydroge-
nation methodology will be useful in academic research labo-
ratories and may also be applied in industry.
[2] C. N. Satterfield, Heterogeneous Catalysis in Industrial Practice, McGraw-
Hill, New York, 1991.
[
3] a) H. Oyamada, R. Akiyama, H. Hagio, T. Naito, S. Kobayashi, Chem.
Commun. 2006, 4297–4299; b) H. Oyamada, T. Naito, S. Miyamoto, R.
Akiyama, H. Hagio, S. Kobayashi, Org. Biomol. Chem. 2008, 6, 61–65.
4] a) R. Akiyama, S. Kobayashi, Angew. Chem. Int. Ed. 2001, 40, 3469–3471;
Angew. Chem. 2001, 113, 3577–3579; b) R. Akiyama, S. Kobayashi,
Chem. Rev. 2009, 109, 594–642.
[
[
5] a) R. Akiyama, S. Kobayashi, J. Am. Chem. Soc. 2003, 125, 3412–3413;
b) K. Okamoto, R. Akiyama, H. Yoshida, T. Yoshida, S. Kobayashi, J. Am.
Chem. Soc. 2005, 127, 2125–2135; c) S. Kobayashi, H. Miyamura, Chem.
Rec. 2010, 10, 271–290; d) S. Kobayashi, H. Miyamura, Aldrichimica Acta
2013, 46, 3–19.
[
[
6] Available from JGC C&C.
7] a) S. Lan, Z. Min, L. Liang, Food Sci. 2008, 29, 350–352;b) T. S. V. R. Nee-
harika, K. N. P. Rani, T. P. Kumar, M. S. L. Karunya, R. B. N. Prasad, Int. Food
Res. J. 2014, 21, 1097–1101.
[8] a) G. Jas, A. Kirschning, Chem. Eur. J. 2003, 9, 5708–5723; b) R. V. Jones,
L. Godorhazy, N. Varga, D. Szalay, L. Urge, F. Darvas, J. Comb. Chem.
2006, 8, 110–116; c) M. Irfan, T. N. Glasnov, C. O. Kappe, ChemSusChem
ChemCatChem 2015, 7, 4025 – 4029
www.chemcatchem.org
4028
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communications
2011, 4, 300–316; d) M. O’Brien, N. Taylor, A. Polyzos, I. R. Baxendale,
[10] M. S. Kwon, N. Kim, C. M. Park, J. S. Lee, K. Y. Kang, J. Park, Org. Lett.
S. V. Ley, Chem. Sci. 2011, 2, 1250–1257.
2005, 7, 1077–1079.
[11] The reduced phosphatidylcholine was produced with yields >10 kgh
À1
[
9] a) J. Kobayashi, Y. Mori, K. Okamoto, R. Akiyama, M. Ueno, T. Kitamori, S.
Kobayashi, Science 2004, 304, 1305–1308; b) J. Kobayashi, Y. Mori, S. Ko-
bayashi, Chem. Commun. 2005, 2567–2568; c) M. Ueno, T. Suzuki, T.
Naito, H. Oyamada, S. Kobayashi, Chem. Commun. 2008, 1647–1649;
d) H. Oyamada, T. Naito, S. Kobayashi, Beilstein J. Org. Chem. 2011, 7,
.
7
35–739; e) M. Ueno, Y. Morii, K. Uramoto, H. Oyamada, Y. Mori, S. Ko-
Received: September 1, 2015
bayashi, J. Flow Chem. 2014, 4, 160–163.
Published online on October 23, 2015
ChemCatChem 2015, 7, 4025 – 4029
www.chemcatchem.org
4029
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim