Facile hydrogenation of carbon-carbon double bonds using catalytic noble nanoparticles immobilized in microfluidic reactors

Article abstract of DOI:10.1016/j.catcom.2011.12.002

Facile hydrogenation of carbon-carbon double bonds was achieved in a poly(dimethylsiloxane) (PDMS) microfluidic reactor with immobilized noble nanoparticles, using the conversion of 6-bromo-1-hexene to 1-bromo-hexane as a model reaction. The microreactor was used to measure the intrinsic catalytic activity and turnover frequencies (TOF) of palladium (Pd), platinum (Pt) and ruthenium (Ru) nanocatalysts. The TOF of reactions run with immobilized nanocatalysts in the microfluidic reactor were hundreds of times larger than those measured in identical reactions in batch reactors. The combination of well-defined nanocatalysts and microfluidics significantly enhances hydrogen diffusion to catalytic sites, thus eliminating mass transfer limitations and enabling evaluation of the intrinsic catalytic activity. The system provides an excellent platform for high throughput screening of catalysts, and for conducting mechanistic studies of reaction kinetics.

Full text of DOI:10.1016/j.catcom.2011.12.002

Catalysis Communications 18 (2012) 168–175
Contents lists available at SciVerse ScienceDirect
Catalysis Communications
journal homepage: www.elsevier.com/locate/catcom
Short Communication
Facile hydrogenation of carbon-carbon double bonds using catalytic noble
nanoparticles immobilized in microfluidic reactors
a
a
a
b
a,
Rui Lin , Xianfeng Ma , Thomas R. Fielitz , Sherine O. Obare , Robert Y. Ofoli ⁎
a
Department of Chemical Engineering and Materials Science, Michigan State University, East Lansing, MI 48824, USA
Department of Chemistry, Western Michigan University, Kalamazoo, MI 49008, USA
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Facile hydrogenation of carbon-carbon double bonds was achieved in a poly(dimethylsiloxane) (PDMS)
Received 30 September 2011
Received in revised form 25 November 2011
Accepted 1 December 2011
Available online 8 December 2011
microfluidic reactor with immobilized noble nanoparticles, using the conversion of 6-bromo-1-hexene to
1-bromo-hexane as a model reaction. The microreactor was used to measure the intrinsic catalytic activity
and turnover frequencies (TOF) of palladium (Pd), platinum (Pt) and ruthenium (Ru) nanocatalysts. The
TOF of reactions run with immobilized nanocatalysts in the microfluidic reactor were hundreds of times larger
than those measured in identical reactions in batch reactors. The combination of well-defined nanocatalysts
and microfluidics significantly enhances hydrogen diffusion to catalytic sites, thus eliminating mass transfer
limitations and enabling evaluation of the intrinsic catalytic activity. The system provides an excellent
platform for high throughput screening of catalysts, and for conducting mechanistic studies of reaction
kinetics.
Keywords:
Nanocatalysis
High throughput catalyst screening
Triphase hydrogenation
Catalyst recycling
©
2011 Elsevier B.V. All rights reserved.
1
. Introduction
particle sizes and narrow size distributions also have a high percentage
of their catalytically active atoms on the surface, which enables greatly
enhanced reactivities. Stabilizing ligands are required to prevent particle
agglomeration, to maintain catalytic activity. For example, n-dodecyl
sulfide (NDS) has been shown to be effective as a stabilizing ligand for
synthesizing uniform palladium (Pd), platinum (Pt) and ruthenium
(Ru) nanoparticles in the 2–5 nm size range [6]. The sulfur atom in the
NDS molecule coordinates well with noble metal atoms, and the long
carbon chains prevent particle agglomeration through steric hindrance.
With a surface area to volume ratio at least 100 times larger than
in conventional batch reactors, microfluidic reactors have been used
extensively for heterogeneous catalytic hydrogenation [7–10]. Mass
transfer limitations are greatly reduced by the increased interfacial
areas and short diffusion distances [7–9]. A microfluidic system also
enables high throughput and high space-time yield, low consumption
of reagents and energy as well as precise control over various reaction
parameters [11]. Microfluidic systems fabricated from poly(dimethyl-
siloxane) (PDMS) have been used widely in biomedical research
because of their biocompatibility, easy and inexpensive assembly,
versatility of surface functionalization, and high hydrogen permeability
compared to those constructed from silicon, glass or stainless steel. As a
result of these attributes, they have been drawing increasing attention
in synthetic chemistry and catalysis [7–12]. In spite of their low
Triphasic hydrogenation with solid catalysts, molecular hydrogen,
and liquid substrates plays an important role in many academic and
industrial laboratories. However, mass transfer between the gas,
liquid and solid phases can often be inefficient, making it difficult to
get sufficient hydrogen to catalytic sites for the required reactions.
As a result, triphasic hydrogenation is usually carried out under
high temperature and high hydrogen pressure over long periods of
time. Such harsh reaction conditions and long reaction times can
lead to problems that include undesirable side reactions, lack of
operational safety and high energy consumption [1,2]. It can also
hinder assessment of the intrinsic reaction kinetics. An effective
solution to these problems is to combine catalytic nanoparticles
(
(
which provide significantly enhanced surface areas) with microfluidics
which reduces the diffusion length for hydrogen transport). The
resulting improvement in multiphase mass transfer enables high
yield and selectivity under mild reaction conditions, in an efficient
and environmentally friendlier manner.
Noble metallic colloidal nanoparticles under 10 nm in diameter have
been shown to greatly improve catalytic efficiency for multiphase
hydrogenation, due to their large surface area to volume ratios compared
to their bulk counterparts [3–5]. Nanoparticles with well-controlled
g
tolerance for temperature (T ~125 °C) and tendency to swell in
non-polar organic solvents, PDMS microfluidic reactors are suitable
for hydrogenation in aqueous or polar organic media under mild
reaction conditions [13]. Catalysts can be immobilized evenly and
robustly on the channel walls to allow catalyst reuse and recycling,
and to avoid significant pressure drop across the reactor [11,14,15].
⁎
Corresponding author. Tel./fax: +1 517 432 1575.
E-mail address: Ofoli@egr.msu.edu (R.Y. Ofoli).
1566-7367/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.catcom.2011.12.002

R. Lin et al. / Catalysis Communications 18 (2012) 168–175
Table 1
169
In the present work, we report the combination of nanocatalysts
Conditions used for synthesis of palladium (Pd), platinum (Pt) and ruthenium (Ru)
nanoparticles, along with the resulting average particle sizes.
and microfluidics for improved reaction efficiency in a model triphasic
hydrogenation reaction. We used a PDMS microfluidic reactor with
immobilized metallic noble nanoparticles for high throughput
assessment of intrinsic catalytic activity. The simple and robust
immobilization protocol is based on our previously reported scheme
Sample Stabilizing Precursor
ligand
Molar ratio
of ligand to (°C) (h)
precursor
T
Time Solvent Particle
size (nm)
NDS-Pd [CH
3
Pd(OAc)
2
5:1
5:1
5:1
95
3
1
1
Toluene 3.0± 0.3
[
1
10]. The model reaction, hydrogenation of 6-bromo-1-hexene to
(
CH
2
3
)
)
)
11
]
2
2
2
S
S
S
-bromo-hexane (Scheme 1), is important in the synthesis of pharma-
NDS-Pt [CH
Pt(C
5
H
7
O
2
)
2
180
60
Phenyl 3.3± 0.7
ether
Toluene 3.5± 0.4
ceutical intermediates where the selective reduction of carbon–
carbon double bonds can be carried out without affecting the halogen
atom.
(
CH
NDS-Ru [CH3
(CH
2
11
]
Ru(cod)
(cot)
2
11
]
2
. Experimental
2
.1. Synthesis of NDS-stabilized nanoparticles
20-V, Ruurlo, Nederland) and a syringe pump (Harvard apparatus
PHD 2200, Holliston, MA). The gas and liquid flow rates were adjusted
to produce an annular flow pattern in which the gas travelled in the
Pd, Pt and Ru nanoparticles were synthesized by thermal decom-
position, using the parameters given in Table 1. Briefly, a 1:5 molar
ratio of the metal precursors and stabilizing ligands were mixed and
heated to the respective decomposition temperature, followed by
acetone washing and centrifugation [6]. NDS is highly soluble in
toluene, so it was used as the solvent in the synthetic procedures
for Pd and Ru. Since Pt precursors have a higher decomposition
temperature than both Pd and Ru, we used phenyl ether as the
solvent for Pt synthesis because of its higher boiling point (259 °C vs.
a
UV light
High-resolution
transparency
Patterns
Photoresist
Si
1
11 °C for toluene).
Master
2
.2. Design and fabrication of microfluidic reactors
Si
A PDMS microfluidic reactor with a Y-shaped inlet was fabricated
by soft lithography [8,16] using the protocol shown in Fig. 1a. A
serpentine reactor pattern on a silicon wafer was transferred to a
PDMS block, and irreversibly bonded to a microscope glass slide to
form rectangular channels with dimensions (W, H, L) of 500 μm,
PDMS
Si
1
00 μm and 30 cm, respectively (Fig. 1b). To prevent swelling of the
Channels
PDMS, the nanoparticles were redispersed in compatible solvents
following synthesis. In situ nanoparticle immobilization was achieved
by sequential infusion of a mixture of 3-aminopropyltrimethoxysilane
PDMS
Glass
(
APTMS, Sigma, St. Louis, MO) and nanoparticles into the oxygen
plasma-treated microfluidic reactors at a flow rate of 0.025 mL/min
for several hours, followed by curing at 60 °C for 1 h in an oven. The
aminosilanization process deposits primary amine groups (−NH )
2
b
on all microfluidic channel surfaces already rich in hydroxyl groups
following oxygen plasma treatment. The sequential introduction of
Pd, Pt or Ru nanoparticles into the channel enables coordination
with the primary amine groups on the channel surfaces [10], thus
inducing immobilization (Fig. 2). A similar immobilization procedure
was carried out on 1 cm by 1 cm pieces of glass and PDMS surfaces,
for the purpose of characterization.
Channel width 500 µm
Channel height 100 µm
2
.3. Hydrogenation of 6-bromo-1-hexene in batch and microfluidic reactors
Hydrogenation of 6-bromo-1-hexene was conducted in both
blank and microfluidic reactors with immobilized catalysts. The
microfluidic reactors were pretreated by flowing hydrogen through
the system at 100 °C for 1 h before reactions. All reactions were
conducted at room temperature (25 °C). Hydrogen at a pressure of
Fig. 1. (a) Step-by-step soft lithography procedure for fabrication of microfluidic channels
in PDMS [8,16]. The channel pattern was designed by Adobe illustrator and printed on a
dark high resolution transparency on which only the pattern is transparent. The pattern
was first transferred to a silicon wafer coated with photoresist by exposure to UV light,
creating a positive master. A mixture of PDMS precursor and cross-linker was poured
onto the positive master and cured to form a PDMS block containing the negative pattern.
Sealed microfluidic channels were then created by using oxygen plasma treatment to
irreversibly bond the PDMS block to the microscope slide. (b) Serpentine microfluidic
channel with a Y-shaped inlet for independent transport of hydrogen and reactants into
the microreactor. The SEM image on the lower left is the top view of the microfluidic
channel. The SEM image on the right is the cross-section of the microfluidic channel,
with the PDMS block to the left and the glass slide to the right. The channel width and
height are 500 μm and 100 μm, respectively. The channel length in this setup is 30 cm,
and can be controlled by the number of loops.
1
atm. and reactant solutions of 1, 5, and 10 mM were introduced
independently into the microfluidic channel through the Y-shaped
inlet using a mass flow controller (Bronkhorst F-201CV-020-AGD-
Scheme 1. The hydrogenation of 6-bromo-1-hexene to 1-bromo-hexane was used as a
model reaction to assess catalytic activity.

170
R. Lin et al. / Catalysis Communications 18 (2012) 168–175
NH2
NH2
NH2
Si
O
Si
O
Si
O
O
O
O
O
OH OH OH OH OH OH OH
APTMS: 3-aminopropyltrimethoxysilane
NDS-NPs
NDS: N-dodecyl sulfide
Pd, Pt or RuNPs
NH2
NH2
NH2
Si
O
Si
O
Si
O
O
O
O
O
Fig. 2. Scheme used for immobilization of NDS-stabilized nanoparticles in the microfluidic reactor. The reactor surfaces were first treated with oxygen plasma to introduce hydroxyl
groups, then aminosilanized by APTMS to install primary amine groups to which Pd, Pt or Ru nanoparticles can attach through their stabilizing ligands to complete the
immobilization.
center and the liquid along the channel walls. Such a flow pattern
allows hydrogen to easily diffuse across the thin layer of liquid reactants
to the catalytic surfaces, thus eliminating mass transfer limitations and
facilitating optimized interaction among the gas, liquid substrate and
solid catalyst [1,11,17]. After a stable flow pattern was obtained, based
on visual inspection, the reaction product was collected continuously
for 10 min in a single vial at the outlet. At the end of each run, the entire
product was diluted 100 times prior to analysis by gas chromatography
and shape (Fig. 3 a–f). The average diameter of the spherical Pd, Pt
and Ru nanoparticles were 3.0± 0.3 nm, 3.3± 0.7 nm and 3.5±
0.4 nm, respectively.
3.2. Assessment of the effectiveness of particle immobilization on surfaces
X-ray photoelectron spectroscopy (XPS) results confirmed successful
catalyst immobilization on both glass and PDMS surfaces (Table 2).
Compared to the data acquired on blank surfaces, we observed sharp
increases in C and N on the aminosilanized surfaces, along with
decreases in Si and O. The increased C and N contents are due to the
molecular structure of APTMS with a 3-carbon arm and a primary
amine group at the end (Fig. 2). The “loss” in the surface concentrations
of Si and O is due to coverage of the glass and PDMS surfaces following
aminosilanization. The characteristic metal elements from the nanopar-
ticles were detected by XPS on all surfaces with immobilized catalysts,
with atomic concentrations ranging from 0.1% to 1.0%. The XPS peaks
of the noble metals (Fig. 4 a–c) provided evidence that the majority
existed in elemental catalytic form. However, since the samples
were exposed to air during characterization, it was difficult to
avoid oxidation of the metals, which resulted in Pd(II), Pt(VI),
Ru(IV) and Ru(VI) on the reactor surfaces [18]. Thus, it was
necessary to pre-reduce each of the catalysts prior to running the
hydrogenation reactions.
(
Agilent 6890 with a flame ionization detector, and a DB1dimethylsi-
loxane column, 0.250 mm I.D. and 30 m in length). Each microfluidic
reactor was rinsed with methanol between runs, and purged with
hydrogen for 15 min. Replicate runs were conducted under similar
reaction conditions in at least two different microfluidic reactors with
identical immobilization schemes. To test for catalyst leaching, the
as-collected reaction mixture was dried following GC analysis, digested
in aqua regia, and its metallic content analyzed by ICP-OES with a detec-
tion limit of 80 parts per billion (ppb).
For comparison, we conducted homogeneous reactions in batch
reactors maintained at the temperature and pressure used in the
microreactors. A round-bottomed flask containing 6-bromo-1-hexene
(
(
0.05 mL, 0.37 mmol) and NDS-stabilized colloidal Pd, Pt and Ru NPs
0.5 mL, 3 mmol) in methanol was put under vacuum for a few
2
seconds, then connected to a rubber bladder containing H at 1 atm.
The solution was stirred at room temperature for 12 h, after which
the rubber bladder was disconnected from the flask. An aliquot
of the reaction product was centrifuged to separate the particles
from the solution. The particles remained as solids at the bottom of
the centrifuge tube, and an aliquot of the supernatant was analyzed
by GC (Agilent 6890) to calculate the conversion.
3.3. Conversion, selectivity and intrinsic turnover frequency (TOF)
All reactions attained 100% selectivity and various degrees of
conversion, again confirming the effectiveness of the immobilization
scheme. Control experiments in blank reactors produced no reactions.
Over 90% of carbon balance was achieved in each reaction. The average
residence time was 1.9 s, calculated on the basis of the liquid holdup,
which is linked directly to the gas and liquid flow rates. Since the
flow pattern in the reactor is annular, we used the Lockhart-
Martinelli-Chisholm correlation for two-phase flow [19] to calculate
the liquid holdup.
At a substrate concentration of 1 mM, both Pd and Pt converted
over 90% of the substrate to the desired product (1-bromo-hexane)
upon first use of the catalysts, approximately 2.5 times higher than
Ru (Fig. 5). These results are in agreement with previous reports
3
. Results and discussions
3
.1. Characterization of nanoparticles
High resolution transmission electron microscopy (HRTEM) images
were taken with a JEOL 2200 (Tokyo, Japan) at an acceleration voltage
of 200 KV. The average particle size was determined by measuring the
diameters of 100 randomly selected images of particles and calculating
the mass average. The results showed that NDS is an effective ligand for
synthesizing stable nanoparticles and controlling their size, uniformity

R. Lin et al. / Catalysis Communications 18 (2012) 168–175
171
8
7
6
5
4
3
2
1
0
0
0
0
0
0
0
0
0
a
c
e
b
d
f
Pd size distribution
1
2
3
4
5
Diameter (nm)
3
2
2
1
1
0
5
0
5
0
5
0
Pt size distribution
2
2.5
3
3.5
4
4.5
5
5.5
Diameter (nm)
5
4
4
3
3
2
2
1
1
0
5
0
5
0
5
0
5
0
5
0
Ru size distribution
2
2.5
3
3.5
4
4.5
Diameter (nm)
Fig. 3. HRTEM images and histographs of NDS-stabilized Pd (a–b), Pt (c–d) and Ru (e–f) nanoparticles at a magnification of 100 K (scale bar=20 nm). The average diameters of the
spherical Pd, Pt and Ru nanoparticles were 3.0± 0.3 nm, 3.3± 0.7 nm and 3.5± 0.4 nm, respectively.
that Pd and Pt are effective in reducing carbon-carbon double bonds
20]. Though Ru is better known for reducing carbonyl bonds than
carbon-carbon double (olefenic) bonds, we still observed a first
time conversion of 38%. We examined the effects of substrate concen-
tration on the overall conversion for each of the three catalysts. The
conversion levels for Pd and Pt dropped steadily from low to high
[

172
R. Lin et al. / Catalysis Communications 18 (2012) 168–175
Table 2
Atomic concentrations of different elements on glass and PDMS surfaces following immobilization of the nanoparticles in sealed microfluidic reactors (XPS characterization).
Atomic concentrations on glass, %
Atomic concentrations on PDMS, %
Elements
Blank
Aminosilanization
NDS-Ru
NDS-Pd
NDS-Pt
Blank
Aminosilanization
NDS-Ru
NDS-Pd
NDS-Pt
C
N
O
Si
S
Ru
Pd
Pt
22.7
1.0
53.0
10.6
24.0
12.0
N/A
N/A
N/A
N/A
55.5
6.2
24.0
13.9
0.1
0.3
N/A
N/A
54.2
5.0
24.3
16.2
0.2
N/A
0.1
N/A
57.3
8.7
23.7
9.9
45.7
N/A
28.4
25.9
N/A
N/A
N/A
N/A
45.8
6.0
47.0
1.7
30.0
20.0
0.2
0.1
N/A
N/A
47.7
1.9
46.2
1.8
32.0
18.9
0.1
N/A
N/A
1.0
50.7
25.6
N/A
N/A
N/A
N/A
29.0
19.0
N/A
N/A
N/A
N/A
28.3
21.9
0.05
N/A
0.1
0.3
N/A
N/A
0.1
N/A
substrate concentration (Fig. 6), which may be an indication of
physical limits on catalyst performance over the extremely short
residence time (approximately two seconds) over which these
one-pass reactions were conducted in the microfluidic reactor.
By contrast, the conversion achieved with Ru increased significantly
at the highest substrate concentration over the nearly identical levels
observed at the two lower concentrations. The reason for this is not
entirely clear to us at this point. One could speculate that this may be
an indication that Ru has a higher performance saturation ceiling for
conversion of the substrate than either Pd or Pt, even though its overall
conversion is lower. It is also possible that the rate of reaction is a much
stronger function of concentration for Ru than for the other catalysts.
For example, the total amount of substrate converted (estimated simply
as the product of reactant concentration and percent conversion)
increased from low to high substrate concentrations for all three
catalysts, but much more significantly for Ru (Table 3). We are in the
process of acquiring the kinetic data required to fully address these
issues, by conducting similar experiments in Parr reactors. Overall, the
catalyst effectiveness for this reaction may be characterized as
Pd>Pt>Ru, in agreement with other reports in the published literature
We investigated catalyst reactivity upon recycling at all three
substrate concentrations, with each catalyst reused a minimum of
three times. Pd achieved a complete conversion over three consecutive
runs at a substrate concentration at 1 mM (Fig. 8 a). However, the
conversion decreased significantly by the third recycle at substrate
concentrations of 5 and 10 mM. Pt activity remained relatively stable
at a substrate concentration of 5 mM, although there were decreased
conversions after the first run at the other two substrate concentrations
(Fig. 8 b). Ru was recycled three times without significant loss in activity
at substrate concentrations of 1 and 5 mM, but the conversion level
dropped by about 50% after the first run at the substrate concentration
of 10 mM (Fig. 8 c). One possible reason for these decreases is that the
immobilization scheme was unstable, and that the catalyst was leaching
from the interface. To investigate this issue, we did an ionic coupled
plasma-optical emission spectroscopy (ICP-OES) characterization of all
the reactor effluents. We did not detect any metals in the product
mixtures, indicating zero or negligible catalyst leaching and demon-
strating the robustness of the immobilization protocol. Since we ruled
out catalyst leaching as the cause of the observed reductions in catalytic
activity, we are now investigating the possibility of fouling of catalyst
sites and/or catalyst poisoning.
[18,20–23].
We assessed the effect of substrate flow rates on conversion by
keeping the gas flow rate constant at 2.4 mL/min and doubling the
liquid flow rate to 0.2 mL/h, using Pt at substrate concentration of
We observed partial agglomeration of nanoparticles following
their re-dispersion in methanol after synthesis. The formation of
nanoparticle clusters would have the effect of decreasing the overall
surface area and, correspondingly, the number of available catalytic
sites. This would decrease both the reaction rates and conversion
levels. This would mean that the TOF values we estimated for the
microfluidic reactor are low which, in a sense, provides further
evidence of the high efficiency at which the system operates. We
are currently working on improving the transfer of nanoparticles
from synthesis media to solvents friendly to PDMS, to avoid the
possibility of nanoparticle aggregation. We are also investigating the
use of ligands that provide better NP stability upon transfer to
PDMS-friendly solvents, to ensure improved particle dispersion
upon immobilization in the microreactors. This platform significantly
reduces the time required for complete reactions, making the system
a convenient and inexpensive tool for high throughput screening of
catalysts for a variety of purposes.
1
0 mM as an example. We still observed the desired annular flow
pattern, with H in the middle of the channel and the reactant along
2
the walls. However, as would be expected, the liquid film thickness
along the channels increased. We observed a decrease in substrate
conversion at the higher liquid flow rate (Fig. 7), which we attributed
to the shorter residence time in the channel.
We calculated the turnover frequency (TOF), defined as the moles
of product per mole of active catalyst per hour, for each catalyst, and
compared the results to reactions with the same catalysts in a batch
reactor under the same reaction conditions. We calculated the metal
dispersion using particle sizes estimated from HRTEM images, assuming
spherical particles under closed-packed conditions. In all situations, the
TOF values were hundreds of times higher in the microfluidic reactor
than in the batch reactions we ran in flasks at the same temperature
and hydrogen pressure (Table 4). We recognize that, while temperature
and hydrogen pressure were identical in both reactors, the reactant
concentration at any location in the microreactor was essentially
constant with respect to time, whereas it continuously changed in the
flask. So, this TOF comparison is not to imply that the reaction proceeds
in an identical fashion in the two systems. Rather, it is an indication of a
much more efficient system for catalyst evaluation, in terms of both
materials and time. In essence, the much higher TOF value provides
validation that the high surface area to volume ratios afforded by the
combination of nanocatalysts and microfluidics promotes much better
access to catalytic sites. Coupled with the much shorter residence
times required for complete reactions (on the order of seconds rather
than hours), the system has great potential for high throughput screening
of catalysts.
4. Conclusions
We have integrated catalytic nanoparticles with microfluidics, and
developed a reliable protocol for immobilization of nanocatalysts in
sealed microreactors. The immobilization procedure is reproducible
and robust, and promotes catalyst recovery and reuse. We have
demonstrated that the system can be used with great efficiency to
evaluate intrinsic catalytic activity for multiphase hydrogenation
reactions, and to assess kinetic parameters such as catalyst turnover
frequencies. The integration of well-defined nanocatalysts and micro-
fluidics provides an excellent platform with great potential for high
throughput screening of catalysts and for mechanistic studies of reaction
kinetics.

R. Lin et al. / Catalysis Communications 18 (2012) 168–175
173
a
8
7
6
5
4
3
000
000
000
000
000
000
Pos.
Sep. %Area
35.46 0.00 49.81
37.77 2.31 10.07
3
3
3
3
Pd 3d
40.73 5.27 33.37
43.04 7.58
6.75
Pd (0)
PdO
3
55
350
345
340
335
330
Binding Energy (eV)
Fig. 5. Conversion of 6-bromo-1-hexene to 1-bromohexane upon first use of Pd, Pt and
Ru catalysts, respectively, at a substrate concentration of 1 mM. All reactions were
conducted in PDMS microfluidic reactors at room temperature and hydrogen pressure
at 1 atm. Gas flow rate was set at 2.4 mL/min by a mass controller, and the liquid flow rate
at 0.1 mL/h by a syringe pump. Each reaction lasted 10 min, and the reaction products
were collected during this time and analyzed by GC.
b
3
2
2
1
1
000
500
000
500
000
Pos.
Sep. %Area
7
7
7
7
1.81 0.00 47.88
Pt 4f
3.35 1.54 10.60
5.14 3.33 34.00
6.68 4.87
7.52
Pt (0)
PtO2
5
00
9
0
85
80
75
70
65
Binding Energy (eV)
c 4
200
100
000
900
800
700
600
500
400
Ru 3p3
4
4
3
3
3
3
3
3
Pos.
Sep. %Area
4
4
4
61.96 0.00 66.43
63.49 1.53 22.18
65.42 3.46 11.40
Fig. 6. Effect of substrate concentration on the conversion of 6-bromo-1-hexene upon
first use og Pd, Pt and Ru catalysts, respectively, at substrate concentrations of 1, 5
and 10 mM. The reaction conditions are the same as described in Fig. 5. The solid
lines are for the purpose of guiding the eye; they do not represent model fits of the
data.
Ru (0)
RuO2
RuO3
Acknowledgements
This project is funded in part by the Michigan Economic Develop-
ment Corporation (MEDC) and the National Science Foundation
Table 3
Total amount of substrate (6-bromo-1-hexene) converted by hydrogenation catalyzed by
NDS-Pd, Pt and Ru nanoparticles in the microfluidic reactors at substrate concentrations of
1, 5 and 10 mM.
4
78 476 474 472 470 468 466 464 462 460 458
Binding Energy (eV)
Catalysts
Total amount of substrate converted
Fig. 4. (a) to (c): XPS spectra of immobilized Pd (a), Pt (b), and Ru (c) nanoparticles on
microreactor surfaces. While most of the metals are in elemental form, exposure of the
samples to air during characterization resulted in some oxidation, leading to Pd(II), Pt
−
8
(×10
mol)
Substrate concentration (mM)
NDS-Pd
NDS-Pt
NDS-Ru
(VI), Ru(IV) and Ru(VI). All catalysts were reduced prior to use in reactors.
1
5
1.67
5.68
8.32
1.59
2.92
5.10
0.65
3.08
8.67
10

174
R. Lin et al. / Catalysis Communications 18 (2012) 168–175
a
Fig. 7. Assessment of the effect of liquid flow rates on the conversion of 6-bromo-1-
hexene catalyzed by Pt at a substrate concentration of 10 mM. The reaction conditions
are the same as described in Fig. 5, except that two liquid flow rates were used: 0.1 mL/
h and 0.2 mL/h. Pt was recycled 5 times at each liquid flow rate. The solid lines are for
the purpose of guiding the eye; they do not represent model fits of the data.
b
Table 4
Comparison of turnover frequencies (TOF) for reactions run in batch and microfluidic
reactors under identical conditions (25 °C, hydrogen pressure of 1 atm).
System
TOF (h−1)
NDS-Pd
NDS-Pt
NDS-Ru
Microfluidic reactor
Batch reactor
9540
41
6540
9.1
2010
3.5
(
through S.O. Obare's CAREER award CHE 1005456). We acknowledge
XPS characterization by Dr. Per Askeland from the Michigan State
University Composite Materials and Structures Center (CMSC). We
also appreciate the many helpful discussions with Professor James
Jackson of the Department of Chemistry at Michigan State University.
c
References
[
[
[
1] J. Kobayashi, Y. Mori, S. Kobayashi, Multiphase organic synthesis in microchannel
reactors, Chemistry, an Asian Journal 1 (1–2) (2006) 22–35.
2] M. Irfan, T.N. Glasnov, C.O. Kappe, Heterogeneous catalytic hydrogenation reac-
tions in continuous-flow reactors, Chemsuschem 4 (3) (2011) 300–316.
3] H. Bonnemann, R.M. Richards, Nanoscopic metal particles — synthetic methods
and potential applications, European Journal of Inorganic Chemistry 10 (2001)
2455–2480.
[4] F. Raimondi, et al., Nanoparticles in energy technology: examples from electro-
chemistry and catalysis, Angewandte Chemie International Edition 44 (15)
(2005) 2190–2209.
[5] J.M. Thomas, et al., High-performance nanocatalysts for single-step hydrogena-
tions, Accounts of Chemical Research 36 (1) (2003) 20–30.
[
6] M. Ganesan, R.G. Freemantle, S.O. Obare, Monodisperse thioether-stabilized pal-
ladium nanoparticles: synthesis, characterization, and reactivity, Chemistry of
Materials 19 (14) (2007) 3464–3471.
[
7] T. McCreedy, Rapid prototyping of glass and PDMS microstructures for micro total
analytical systems and micro chemical reactors by microfabrication in the general
laboratory, Analytica Chimica Acta 427 (1) (2001) 39–43.
[
8] J.C. McDonald, G.M. Whitesides, Poly(dimethylsiloxane) as a material for fabricat-
ing microfluidic devices, Accounts of Chemical Research 35 (7) (2002) 491–499.
9] G.M. Whitesides, A.D. Stroock, Flexible methods for microfluidics, Physics Today
[
54 (6) (2001) 42–48.
[10] R. Lin, et al., In-situ immobilization of palladium nanoparticles in microfluidic re-
actors and assessment of their catalytic reactivity, Nanotechnology (2010)
Fig. 8. Conversion of 6-bromo-1-hexene as a function of the number of catalyst recycles
at substrate concentrations of 1, 5 and 10 mM. Reaction (a) was catalyzed by Pd, (b) by
Pt, and (c) by Ru. All reactions were conducted at room temperature and hydrogen
pressure at 1 atm. The liquid and gas flow rates were 0.1 mL/h and 2.4 mL/min, respectively.
The solid lines are for the purpose of guiding the eye; they do not represent model fits of the
data.
325605.
[
11] A. Iles, et al., A simple technique for microfluidic heterogeneous catalytic hydro-
genation reactor fabrication, Catalysis Letters 114 (1–2) (2007) 71–74.
12] T. McCreedy, Fabrication techniques and materials commonly used for the pro-
duction of microreactors and micro total analytical systems, Trac-Trends in Ana-
lytical Chemistry 19 (6) (2000) 396–401.
[

R. Lin et al. / Catalysis Communications 18 (2012) 168–175
175
[
13] J.N. Lee, C. Park, G.M. Whitesides, Solvent compatibility of poly(dimethylsilox-
[19] D. Chisholm, A theoretical basis for the Lockhart–Martinelli correlation for two-phase
flow, International Journal of Heat and Mass Transfer 10 (12) (1967) 1767–1778.
[20] M. Ruta, N. Semagina, L. Kiwi-Minsker, Monodispersed Pd nanoparticles for acet-
ylene selective hydrogenation: particle size and support effects, Journal of Physi-
cal Chemistry C 112 (35) (2008) 13635–13641.
ane)-based microfluidic devices, Analytical Chemistry 75 (23) (2003)
6544–6554.
[
[
[
[
[
14] L. Kiwi-Minsker, A. Renken, Microstructured reactors for catalytic reactions, Ca-
talysis Today 110 (1–2) (2005) 2–14.
15] E. Klemm, et al., Microstructured reactors in heterogeneous catalysis, Chemical
Engineering and Technology 30 (12) (2007) 1615–1621.
16] Y.N. Xia, G.M. Whitesides, Soft lithography, Angewandte Chemie International
Edition 37 (5) (1998) 551–575.
17] E.T. Castellana, et al., Direct writing of metal nanoparticle films inside sealed
microfluidic channels, Analytical Chemistry 78 (1) (2006) 107–112.
18] M. Zawadzki, J. Okal, Synthesis and structure characterization of Ru nanoparticles
[21] R.X. Liu, et al., Selective hydrogenation of citral catalyzed with palladium nano-
particles in CO
2
-in-water emulsion, Green Chemistry 11 (7) (2009) 979–985.
by one-step flame
[22] O. Mekasuwandumrong, et al., Preparation of Nano-Pd/SiO
2
spray pyrolysis and its hydrogenation activities: comparison to the conventional
impregnation method, Industrial and Engineering Chemistry Research 48 (6)
(2009) 2819–2825.
[23] P. Weerachawanasak, et al., Effect of strong metal-support interaction on the cat-
stabilized by PVP or gamma-Al
111–3121.
2
O
3
, Materials Research Bulletin 43 (11) (2008)
alytic performance of Pd/TiO
lacetylene, Journal of Catalysis 262 (2) (2009) 199–205.
2
in the liquid-phase semi hydrogenation of pheny-
3