Sequential reactions directed by core/shell catalytic reactors

Article abstract of DOI:10.1002/smll.200902336

Millimeter-sized reactor particles made of permeable polymer doped with catalysts arranged in a core/shell fashion direct sequences of chemical reactions (e.g., alkyne coupling followed by hydrogenation or hydrosilylation followed by hydrogenation). Spati

Full text of DOI:10.1002/smll.200902336

Sequential Reactions Directed by Core/Shell Catalytic Reactors
Sequential catalysis
Sequential Reactions Directed by Core/Shell Catalytic
Reactors
Yanhu Wei, Siowling Soh, Mario M. Apodaca, Jiwon Kim, and
Bartosz A. Grzybowski*
M
illimeter-sized reactor particles made of permeable polymer doped with
Keywords:
catalysts arranged in a core/shell fashion direct sequences of chemical
reactions (e.g., alkyne coupling followed by hydrogenation or
ꢁ
ꢁ
ꢁ
ꢁ
core/shell reactors
multifunctional catalysts
nanoparticles
hydrosilylation followed by hydrogenation). Spatial compartmentalization
of catalysts coupled with the diffusion of substrates controls reaction order
and avoids formation of byproducts. The experimentally observed yields of
reaction sequences are reproduced by a theoretical model, which accounts
for the reaction kinetics and the diffusion of the species involved.
sequential reactions
1
. Introduction
products between these regions that determines the outcome of
the overall reaction sequence. For example, amino acids are
The ability to perform multiple reactions in one reaction translated into proteins in the endoplasmic reticulum (ER) but
vessel and in a well-defined sequence can simplify and their further ‘‘packaging’’ takes place in the Golgi apparatus
[
7]
accelerate multiple-step syntheses, and can translate into from which they later migrate to final destinations. Another
significant economic savings by reducing the amount of manifestation of cellular compartmentalization is that of
[
1]
byproduct
and by eliminating intermediate purification multienzyme complexes for metabolic channeling (‘‘metabo-
[
8]
steps, which account for as much as 60% of the total synthetic lons’’) whereby biosynthetic intermediates migrate and react
cost. To date, most research on sequential reactions has sequentially between catalytic sites without diffusing into the
focused on the control of intramolecular structure and bulk of the cell.
[
2]
[
6]
reactivity in the so-called cascade (also known as tandem or
Beyond these and other biological examples, several
[
2a–c,3]
domino) reactions,
on identification of compound research groups have reported non-biological compartmenta-
[
4]
collections suitable for multicomponent reactions (in which, lized systems for multifunctional catalysis. One strategy that
[
4c,5]
however, the exact sequence of steps is rarely known
), and has been developed relies on the immobilization of mutually
[
2b]
[6]
on the design of multifunctional
which reaction sequence is controlled by the arrangement of cular supports. For example, Avnir’s group pioneered the
catalytic subunits.
separate entrapment of catalysts and catalyst-poisoning
While the elegance and practicality of many of these reagents in sol–gel matrices, which enable the simultaneous
or modular catalysts in ‘‘incompatible’’ reagents/catalysts onto solid or macromole-
[
9]
approaches cannot be disputed, biology teaches us that use, in one pot, of these chemicals. This concept was
[
10]
[11]
arguably a simpler strategy can be used to control reaction demonstrated in acid–base,
redox,
In another approach, Hawker et al. attached
which the consecutive reactions take place at different regions/ acidic and basic residues to the interior of star polymers such
‘compartments’’. It is then the transport of reaction substrates/ that they did not quench each other, which allowed both acid-
and even enzymatic
[
12]
sequences – namely, that of spatial compartmentalization in chemistries.
‘
[
13]
and base-catalyzed reactions in the same reaction vessel.
] Prof. B. A. Grzybowski, Dr. Y. Wei, S. Soh, M. M. Apodaca, J. Kim Fr e´ chet also reported a system in which ‘‘polarity-directed’’
[ꢀ
Department of Chemical and Biological Engineering
Department of Chemistry, Northwestern University
sequential reactions occur in a biphasic medium comprising oil
[
14]
droplets suspended in aqueous phase.
Spatial separation of
2145 Sheridan Road, Evanston, IL 60208 (USA)
E-mail: grzybor@northwestern.edu
catalysts has also been combined with system geometry,
notably in the form of planar multilayer architectures. For
example, Alfonso et al. fabricated two-layer MgO/PtSnAl
:
Supporting Information is available on the WWW under http://
www.small-journal.com or from the author.
[
15]
membranes, which under constant flow carried consecutive
oxidative and non-oxidative dehydrogenation of butane; on
DOI: 10.1002/smll.200902336
small 2010, 6, No. 7, 857–863
ß 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
857

full papers
B. A. Grzybowski et al.
exit, this system produced a mixture of butene isomers,
butadiene, carbon monoxide, carbon dioxide, unreacted
butane, and several other products. Two-layered membranes
[
16]
were also used in Rothenberg’s cat-in-cup system, where the
catalyst was immobilized on one of the membranes while the
reaction substrates and products could diffuse freely in and out.
In another study, Bowden et al. reported an elegant use of
poly(dimethylsiloxane) (PDMS) membranes to spatially
separate polar and non-polar reagents and appropriate
[
17]
catalysts,
such that consecutive reactions could occur on
different sites of the membrane. In parallel to these studies,
there has been significant interest in driving sequential reac-
tions using multifunctional/multilayered particles rather than
membranes (whose incorporation into flow/diffusion systems
can be cumbersome). For example, Eiser et al. synthesized Pd
or Au catalytic nanoclusters surrounded by polyelectrolyte
shells and suggested that these systems could be used for
sequential catalysis (though only one reaction, Sonogashira
[
18]
cross coupling, was demonstrated).
In another interesting
example, Tsubaki’s group described catalytic particles compris-
[19]
2 3 2
ing a zeolite layer surrounding a Co/Al O or Co/SiO core.
These particles were used to direct synthesis of isoparaffins from
syngas, and it was suggested that upon modifications they could be
used to direct sequential reactions. The last two examples suggest
that while preparation of core/shell architectures is certainly
feasible, the extension of these systems to sequential catalysis
remains an experimental challenge requiring careful control of
reaction rates versus the speed of migration of the substrates/
intermediates through the supporting matrix.
In this work, we address this challenge and set out to design
compartmentalized ‘‘reactor’’ particles, in which sequences of
catalytic reactions can be controlled by the system’s geometry
and by the diffusion of the reagents/intermediates through the
regions containing different catalysts. Specifically, we describe
systems in which two catalysts (C and C ) are occluded in a
1
2
permeable polymer matrix and are arranged in a core/shell
manner (Figure 1a). When these composite reactors are placed
in the solution containing reaction substrates, the substrates
first react with C
diffusing to the inner core, where reaction R
Provided the dimensions of the core and of the shell are
appropriate, only the sequence of reactions {R , R } takes place
and the alternative sequence {R , R } is avoided (Figure 1b).
1
(in reaction R
1
) in the outer shell before
2
is effected by C .
2
1
2
2
1
The experimentally observed yields of sequential reactions are
reproduced by a theoretical model, which accounts for the
reaction kinetics and the diffusion of the species involved.
Figure 1. a) Scheme of the fabrication protocol. Cured PDMS containing
catalystC
are then immersed in PDMS containing catalyst C
1
iscut(usinghomemadeprecisioncutter)intosmallcubesthat
. After curing, the core
2
and shell reactors are cut into millimeter-sized cubes. b) A sequential
reaction in a core/shell catalytic reactor transforms substrate A first into
2. Results and Discussion
intermediateB (reactioncatalyzedby C
1
). B diffusinginto thecoreis then
We fabricated (see Experimental Section) and tested two converted to C (by C ), which ultimately diffuses out of the reactor. In the
2
types of multicomponent, core/shell reactors based on the absence of the core/shell architecture, as in the case of PDMS particles
[
20]
carryingdifferentcatalysts(rightscheme),thereagentscanalsoreactfirst
with C and only then with C to produce byproducts D and E. c) The
PDMS
matrix permeable to gasses (e.g., H
2 2 2 4
, O , N , CH
[
21]
[22]
2
1
etc.)
as well as many common organic solvents including
species involved in the reaction sequence catalyzed by PdNP/Cu(OAc)
system. The core/shell architecture promotes formation of the 1,4-
diphenylbutane product and prevents direct hydrogenation to
phenylethane. d) Similarly, PdNP/AuNP reactors mediate
2
acetonitrile and toluene used in our experiments. In the first
system (Figure 1b), the PDMS core contained ꢂ4.8 mM (in
terms of metal atoms) of 5-nm Pd nanoparticles stabilized with
dodecylamine (DDA), which are known to be efficient hydrosilylation/hydrogenation sequence, but eliminate hydrogenation
[
23]
hydrogenation catalysts.
The outer PDMS shell contained by-product (4-methoxyethylbenzene).
ß 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
8
58 www.small-journal.com
small 2010, 6, No. 7, 857–863

Sequential Reactions Directed by Core/Shell Catalytic Reactors
Table 1. Comparison of experimental and modeled steady-state distri-
Table 2. Comparison of experimental and modeled steady-state distri-
butions (molar %) of species in sequential alkyne coupling-hydrogen- butions (molar %) of species in sequential hydrosilylation -hydrogen-
ation reactions in PdNP/Cu(OAc) reactors. ation reactions in PdNP/AuNP reactors.
2
[
a]
[a]
[a]
r
1
S
mm] [mm]
t
Run Time Conv%,B
[h]
Conv%,C
Exp/model Exp/model Exp/model
Conv%,D
[
#
2
2
2
1
2
2
2
1.5
1.1
0.8
1.5
1.1
1.1
1.1
1
1
1
1
231
187
111
240
144
144
144
14/8.7
12/6.8
10/4.9
17/15
13
86/89
88/90.6
88/90.1
83/84
87
0.0/2.3
0.0/2.6
1.9/4.9
0.0/0.8
0.0
[a]
[a]
Conv%
(G1 þ G2)
[a]
Conv% H
Exp/model
r
1
S
mm] [mm]
t
Run Time Conv%
#
[
[h]
(F1 þ F1)
.5
Exp/model Exp/model
[
[
[
b]
1
2
3
b]
b]
29
11
71
38
0.0
35
2
2
2
2
2
2.7
1.8
1.8
1.8
1.8
1
1
23
23
23
23
23
24.6/20.3 75.4/79.6
0.0/0.1
0.0/0.7
0.0
7.2/11.3
7.0
41
92.8/88.0
93
59
[
b]
1
1
1
1
a] The degree of conversion is calculated based on the integration of H
[b]
[
0.0
52.1
[b]
37.5
10.4
NMR signals (see Supporting Information, Section 3). [b] Experimental
1
data from H NMR integration; no theoretical data available since the
1
a] The degree of conversion is calculated based on the integration of H
[
concentration/aggregation of catalysis in the reused PDMS could not be
quantified.
NMR signals (see Supporting Information, Section 3), and the molar
1
ratio of two isomer products is about 5:2. [b] Experimental data from H
NMR integration; no theoretical data available since the concentration/
aggregation of catalysis in the reused PDMS could not be quantified.
ꢂ53 mM Cu(OAc)
2
, which served as an alkyne-coupling
[
24]
catalyst. When the PDMS reactor was placed in acetonitrile
containing a phenylacetylene substrate and under hydrogen
atmosphere, this substrate was first exposed to Cu(OAc) and
2
converted to 1,4-diphenylbutadiyne before it diffused to the
Whiletheyieldsforasingleuseofbothtypesofreactorwere
core where PdNPs catalyzed the hydrogenation of 1,4- high, the degree of conversion decreased markedly when these
b
b
diphenylbutadiyne to 1,4-diphenylbutane (Figure 1c). We reactors were used multiple times (entries marked Run 1 , 2 ,
b
emphasize that since both Cu(OAc)
towards phenylacetylene, the core/shell arrangement is essen- system this was likely due to the ‘‘leaking’’/diffusion of
tial to suppress the alternative reaction pathway leading to Cu(OAc) out of the PDMS matrix. In the case of AuNPs,
phenylethane. Indeed, for reactors comprising r
ꢂ 2-mm cores repeated swelling of the PDMS with toluene caused NP
surrounded by 1.1-mm-thick shells, the overall conversion (i.e., aggregation (evidenced by the color shift from red to violet
the yield of the two sequential reactions) to 1,4-diphenylbutane and decrease in the effective surface area of the catalytic NPs.
was as high as 88% (for this and other data, see Tables 1 and 2). The parameters that control the yield of reaction sequence
2 2
and PdNPs are reactive and 3 in Tables 1 and 2). In the case of the PdNP/Cu(OAc)
2
1
[
26]
)
Incontrast, controlexperimentsinwhichCu(OAc) andPdNPs include concentration of catalysts occluded in PDMS, diffusiv-
2
weredispersedinseparatePDMSpieces(cf. Figure1b)orinthe ities of the various species through PDMS, and the dimensions
same PDMS piece but without any spatial ordering always gave of thereactor’s core andshell. Theconcentrations weusedwere
a mixture of 1,4-diphenylbutane and phenylethane products, chosensuchastoloadthemaximalamountsofcatalystparticles
typically in with a ꢂ45:55 molar ratio.
in PDMS while preventing – or at least minimizing – their
The second system we tested comprised PdNPs occluded in clumping into large aggregates (this was monitored by optical
thecoreandAuNPsintheshellregions(Figure1d). Thesolvent microscopy or, for AuNPs, by UV/Vis spectra in which
in this case was toluene and the substrate, 4-methoxyphenyl- unaggregated NPs had an surface plasmon resonance (SPR)
acetylene. In the presence of Et
3
SiH, the shell Au particles adsorption peak near 520 nm). The diffusivities are determined
catalyzed the conversion of this substrate into two isomers, bytheporosityofthePDMSmatrixandthesizesofthediffusing
(
1-triethylsilyl-2-(4-methoxyphenyl)-ethene and 1-triethylsilyl-1- species. For small molecules, the typical diffusion coefficients
[
25]
ꢃ11
ꢃ10
2
ꢃ1
(
4-methoxyphenyl)-ethene), which were then hydrogenated are D ꢂ 10 –10 m s depending on whether PDMS is
[27]
by PdNPs into triethyl(2-(4-methoxyphenyl)ethyl)-silane and swollen by the solvent or not.
For larger NPs, the diffusion
28]
[
triethyl-(1-(4-methoxyphenyl) ethyl)-silane (with a ꢂ5:2 molar coefficients are negligibly small.
ratio). Inthis case, the conversionfor the reaction sequence was The diameter of the core and the thickness of the shell offer
as high as 93% (in reactors with r
ꢂ2-mm cores and most room for adjustment. Intuitively, for a given core size, the
.8-mm-wide shells, Table 2). At the same time, no 4- thicker the shell, the more complete the first of the two
methoxyethylbenzene byproducts were observed. Also, con- reactions, R , should be and the less unreacted substrate should
trol experiments with PDMS pieces doped uniformly with both reach the core to undergo reaction R therein. In other words,
1
1
1
2
types of NP or with separate pieces containing PdNPs and the thicker the shell, the higher the ratio of desired product (via
AuNPs, respectively, always gave a mixture of triethyl(2-(4- the {R , R } sequence) compared to the byproduct (via the {R ,
1
2
2
methoxyphenyl) ethyl)-silane, triethyl(1-(4-methoxyphenyl) R } sequence). Of course, this increased ‘‘selectivity’’ is at the
1
ethyl)-silane, and byproduct 4-methoxyethyl-benzene, typi- expense of the time required for the {R , R } conversion, since
1
2
cally in a 14:6:80 molar ratio.
diffusion through thicker shell takes more time (for very thick
small 2010, 6, No. 7, 857–863
ß 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.small-journal.com 859

full papers
B. A. Grzybowski et al.
The above qualitative reasoning can be quantified by a
model accounting for both the reaction and the diffusion
[
29]
(
RD) of the species involved through the core/shell reactors.
This model gives not only the steady-state concentrations at
long times (cf. Tables 1 and 2) but also the timecourse of the
reactions (Figure 2). Let us consider PdNP/Cu(OAc)
To simplify calculus, the reactor is approximated as a sphere
core radius r , overall radius r , shell thickness S –r ).
2
system.
(
1
2
t
¼ r
2
1
Denoting the starting phenylacetylene as A, the 1,4-diphenyl-
butadiyne intermediate as B, the desired product 1,4-
diphenylbutane as C, and the undesired products phenylethane
as D (cf. Figure 1c), the pertinent reactions and rate constants
k
k
k
3
1
2
are 2A ꢃ! B ꢃ! C and A ꢃ! D. The RD equations can then
be written as:
@
½Aꢄ
2
2
¼
¼
¼
¼
DAr ½Aꢄ ꢃ k1½Aꢄ HS ꢃ k3½AꢄHC
(1)
(2)
(3)
@
t
@
½Bꢄ
2
2
DBr ½Bꢄ þ k1½Aꢄ HS ꢃ k2½BꢄHC
@
t
@½Cꢄ
2
DCr ½Cꢄ þ k2½BꢄHC
@t
@½Dꢄ
2
DDr ½Dꢄ þ k3½AꢄHC
(4)
@t
Figure 2. The composition of the reaction mixture in the PdNP/Cu(OAc)
system as a function of time and for different reactor dimensions. Color
2
where square brackets denote concentrations and D
x
is the
diffusion coefficient of species x ¼ A, B, C, or D. Note that
coding: green ¼ phenylacetylene reactant; blue ¼ 1,4-
diphenylbutadiyne intermediate; red ¼ the desired 1,4-diphenylbutane because hydrogen is present in excess and its diffusivity
ꢃ8
2
ꢃ1
product, and brown ¼ phenylethane byproduct. Markers represent
through PDMS is ꢂ10 m s (roughly two to three orders of
1
experimentaldatadeterminedby HNMRanalysisofthereactionmixture
[21]
magnitude larger than of other solute species present ), its
(see Supporting Information, Section 3). Error bars are based on the
partial pressure can be treated as constant and can be
analysis of at least three samples for each condition. Solid lines are the
predictionsof the reaction-diffusion theoretical model. For(a–c)the core
diameter is 4 mm but the shell thickness increases from 0.8 mm in (a), to
effectively incorporated into the apparent rate constant k
Equations (2) and (3). In the reaction terms, functions H and
specify that these reactions take place in the shell and the
core regions, respectively. Mathematically, H ) and
¼ H(r–r
¼ H(r)–H(r–r ), where H(z) is the Heaviside step function
2
in
S
H
C
1.1 mm in (b), to 1.5 mm in (c). Note that for the thinnest shell (0.8 mm),
the sequence goes to completion the fastest but there is a substantial
amount of byproduct (brown line; also see data in Table 1). For thicker
shells, the reactions reach steady state at longer times but there is less
unwanted product. c,d) Compare the effects of changing core diameter
S
1
H
C
1
(
H(z) ¼ 0 for z < 0, H(z) ¼ 1 for z ꢅ 0). At the surface of the
cubes, r ¼ r , the diffusive flux of species x across this surface is
2
(4 mm in (c) versus 3 mm in (d)) while keeping the shell thickness the
equal to the change in the concentration of x in solution
same. For the smaller core, it takes longer for the reaction sequence to
reach completion. This is emphasized in the inset that overlays curves
from (c) (dashed lines) and those from (d) (solid lines). Other parameters
around the reactor:
ꢀ
ꢀ
ꢀ
ꢀ
used in experiments and simulations: [A]
0
¼ 25 mM and Vsol ¼ 18 mL for
@½xꢄ
@½xꢄ
sol
ꢃDxS
¼
Vsol
(5)
shell thickness S ¼ 33 mM and Vsol ¼ 15 mL for
t
¼ 1.5 mm; [A]
0
@
r
@t
S
S
t
¼ 1.1 mm; [A]
0
¼ 33 mM and Vsol ¼ 10 mL for S ¼ 0.8 mm. Raw
t
experimentaldataforplots(a–d)aregivenintheSupportingInformation,
Section 4.
where S is the surface area of the reactor and Vsol is the volume
of the solution per one reactor particle. The initial conditions
are such that A is present outside of the cubes at some uniform
shells, the first reaction, R
1
, will be complete but very little of its concentration [A]0 while [B] ¼ [C] ¼ [D] ¼ 0. The other
0
0
0
1
product will reach the core). The experimental data (by H model parameters are diffusion coefficients D ꢂ D ꢂ
A
B
ꢃ11
2
ꢃ1
NMR, see Supporting Information, Section 3) in Tables 1 and 2 D ꢂ D ꢂ 2 ꢆ 10
m
s for small molecules in PDMS
C
D
[
26a]
andinFigure2a–cconfirmbothoftheseconclusions. Regarding swollen only marginally by the acetonitrile solvent
and rate
ꢃ1 ꢃ1
ꢃ1
the dimensions of the core, it is reasonable to expect that for a constants k ¼ 0.18 s
M
, k ꢂ k ¼ 0.072 s (see Supporting
1
2
3
given shell thickness, the smaller this core is, the less time the Information, Section 1 for more details). With these pre-
diffusing chemicals have to complete reaction R therein and liminaries, the equations can be solved numerically using finite
the lower the yield of the {R , R } sequence is at a given reaction difference method, implemented in Matlab R2007b.
2
1
2
time. Again, these trendsare confirmedbyexperimentaldatain
Tables 1 and 2 and also by the plots in Figure 2c and d.
Solid lines in Figure 2a–d plot the modeled compositions of
the reaction mixture at different times and for different core/
8
60 www.small-journal.com
ß 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2010, 6, No. 7, 857–863

Sequential Reactions Directed by Core/Shell Catalytic Reactors
shell dimensions. These modeled dependencies agree well with
the experimental data indicated by solid markers. It is worth
emphasizing that this agreement is achieved with no free/fitting
parameters but only physical quantities determined experi-
mentally (see also Supporting Information, Section 1). Of
course, the modeled dependencies confirm the experimental
trends discussed earlier (better selectivity but slower conver-
sion for thicker shells; higher degree of conversion with larger
cores).
Another significant conclusion from the model is that it can
be used to estimate the ‘‘optimal’’ core diameters and/or shell
thicknesses for which the selectivity (here, ratio of the desired
products to byproducts, normalized to 1) is maximal. This is
illustrated in Figure 3 where the gray portions in the selectivity
plots indicate the core diameters maximizing the selectivity,
and shell thicknesses for which selectivities are above 0.9. The
arrows in the corresponding concentration plots indicate
the actual dimensions in our experiments – as can be seen,
the reactors we used had dimensions close to the optimal values
predicted theoretically.
Finally, we note that the model is applicable to both the
2
PdNP/Cu(OAc) and PdNP/AuNP systems. In the latter case,
the RD equations have to be modified to account for the fact
that the reactions in the shell involve both 4-methoxypheny-
lacetylene and Et SiH. With appropriate values of diffusion
3
and rate constants (for details see Supporting Information,
Section 2), the model reproduces the experimental distribution
of intermediates and products (Table 2).
3. Conclusions
In summary, we have demonstrated control of sequential
reactions with core/shell catalytic reactors. The principle of
compartmentalization underlying this work appears general
and can be extended to other types of sequential-reaction
systems, in which reaction centers are spatially disjoint but are
‘
‘coupled’’ by the transport of chemicals involved. Reaction-
diffusion modeling provides an accurate mathematical tool
with which to study and design such systems. The practical
problem to be solved is the decrease of catalytic activity when
reactors due to the leaking or aggregation of the catalysts when
the reactors are reused multiple times; these problems could
potentially be avoided either by the use of supporting matrices
thatlimitdiffusion(e.g., mesoporousaerogelsofzeolitesforNP
catalysts) or by covalent immobilization of the catalysts on the
polymer matrix.
Figure 3. Theeffectofa)shellthicknessandb)corediameteronthefinal
(at 250 h) distribution of various species in the alkyne coupling/
hydrogenationsequencecatalyzedbyPdNP/Cu(OAc) reactors.c,d)Plots
2
of the selectivity of desired product against byproduct (as defined in the
text)withshellthicknessandcorediameter,respectively.Grayportionsof
the curves indicate the close-to-optimal dimensions for the shell and
core. In (a), core diameter is kept constant at 4 mm, as in most
experiments. The optimal shell thickness (i.e., one for which the desired
4. Experimental Section
1
1
,4-diphenylbutanetosidephenylethaneproductsismaximal)isaround
mm, asshownin(c)(comparewithTable1). In(b), theshellthicknessis
Synthesis of Pd Nanoparticles: i) 2-nm Pd NP seeds: 0.1 mmol
palladium acetate, Pd(OAc)
ꢂ10 min) in a toluene solution (13.4 mL) of DDA (2.1 mmol) and
dodecyldimethylammonium bromide (DDAB, 1.1 mmol). In a
2
, was dissolved by sonication
keptconstantat1.5 mm,andtheoptimalcorediameterasshownin(d)is
above3.5 mm, asin experiments. The arrowsindicate conditions usedin
experiments.e–h)SimilarplotsforthePdNP/AuNPreactorsat23 h.In(e),
(
corediameteriskeptconstantat4 mm,andtheoptimalshellthicknessin separate vial, a solution of tetrabutylammonium borohydride
(g)isaround1.5 mm.In(f),shellthicknessiskeptconstantat1.8 mmand (TBAB, 0.7 mmol), DDAB (0.2 mmol), and toluene (4.9 mL) was
the optimal shell thickness is above ꢂ4.5 mm.
prepared by sonication. The two solutions were mixed and stirred
small 2010, 6, No. 7, 857–863
ß 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com 861

full papers
B. A. Grzybowski et al.
overnight at room temperature to generate 2-nm Pd NP seeds. d): 128.7, 113.9, 55.5, 29.3, 14.0, 7.7, 3.4. Triethyl(1-(4-
1
ii) 5 nm Pd NPs: Pd(OAc)
2
3
(0.5 mmol) was dissolved by sonication methoxyphenyl)ethyl)-silane, H NMR (400MHz, CDCl , d): 6.95
in a toluene solution (67 mL) of DDA (10.7 mmol), and DDAB (d, 2H), 6.76 (d, 2H), 2.21 (q, 1H), 1.31 (d, 3H), 0.80 (m, 9H), 0.46
1
3
(
3
5.3 mmol). To this solution the seed solution described above (q, 6H). C NMR (400MHz, CDCl , d): 128.0, 113.7, 55.5, 25.8,
was added. In a separate vial, a solution of DDAB (2.1 mmol) and 15.9, 7.7, 2.2.
toluene (21 mL) in anhydrous hydrazine (4.2 mmol) was prepared
by sonication. The hydrazine solution was added dropwise (over
ꢂ30 min) to the palladium solution and then stirred at room
temperature overnight to generate ꢂ5-nm Pd NPs.
Acknowledgements
Reactor fabrication: Pd NPs were precipitated from toluene by
the addition of methanol and were then redispersed in hexane.
PdNP/hexane solution was mixed into PDMS (Sylgard 184, Dow
Chemicals) prepolymer and crosslinker, followed by degassing and
curing in an oven at 70 8C overnight. After curing, the PDMS was cut
This work was supported by the Non-equilibrium Energy
Research Center (NERC), which is an Energy Frontier Research
Center funded by the U.S. Department of Energy, Office of
Science, Office of Basic Energy Sciences under Award Number
DE-SC0000989.
3
into blocks of dimensionsꢂ 4 ꢆ 4 ꢆ 4 mm . The PdNP-containing
PDMS cubes were immersed into a mixture of PDMS prepolymer,
crosslinker, and Cu(OAc)
methane). This mixture was then degassed and cured in an oven at
08C overnight. Afterwards, the PDMS slab was cut (using a
2
dissolved in THF (or AuNPs in dichloro-
7
[
1] a) L. F. Tietze, Chem. Rev. 1996, 96, 115; b) J. Shin, H. Matsushima,
J. W. Chan, C. E. Hoyle, Macromolecules 2009, 42, 3294;
c) R. A. Bunce, Tetrahedron 1995, 51, 13103.
homemade precision cutter) around the immersed PdNP cores to
give approximately cubic core/shell reactors (see Figure 1d).
Sequential alkyne coupling-hydrogenation reactions: The
PDMS reactors comprised a core containing PdNPs (ꢂ4.8 mM by
[2] a) J. Lee, Y. Na, H. Han, S. Chang, Chem. Soc. Rev. 2004, 33, 302;
b) F. X. Felpin, E. Fouquet, ChemSusChem 2008, 1, 718;
c) J. C. Wasilke, S. J. Obrey, R. T. Baker, G. C. Bazan, Chem. Rev.
Pd atoms) and a shell containing Cu(OAc)
2
(ꢂ53 mM). Such
2
2
005, 105, 1001; d) Y. Hong, W. Hong, H. N. Chang, Biotech. Lett.
000, 22, 871.
reactors (typically ꢂ50) were placed in a flask containing
phyenylacetylene substrate in CH CN (the specific concentrations
3
[
3] a) N. Hall, Science 1994, 266, 32; b) J. Barluenga, A. Dieguez,
A. Fernandez, F. Rodriguez, F. J. Fananas, Angew. Chem. 2006, 118,
2145; Angew. Chem. Int. Ed. 2006, 45, 2091; c) . Nishibayashi,
M. Yoshikawa, Y. Inada, M. Hidai, S. Uemura, J. Am. Chem. Soc.
differ for reactors of different dimensions and are given in the
main text), followed by stirring at room temperature and under
hydrogen atmosphere for up to 10 days. Every 12–24 h a small
2
004, 126, 16066; d) W. J. Zuercher, M. Hashimoto, R. H. Grubbs,
aliquot of the reaction mixture was withdrawn using a syringe and
1
analyzed by 400 MHz H NMR (in CDCl
J. Am. Chem. Soc. 1996, 118, 6634.
3
). When the distribution of
[
4] a) B. A. Seigal, C. Fajardo, M. L. Snapper, J. Am. Chem. Soc. 2005,
products reached steady state, the reactors were removed, the
solvent was evaporated, and the products were purified by
1
27, 16329; b) A. Pinto, L. Neuville, P. Retailleau, J. Zhu, Org. Lett.
2
006, 8, 4927; c) D. Enders, M. R. M. H u¨ ttl, C. Grondal, G. Raabe,
column chromatography (silica, hexane). The spectra were i) for
1
,4-diphenylbutadiyne, H NMR (400MHz, CDCl
Nature 2006, 441, 861.
1
3
, d): 7.56 (d, 2H),
[5] a) M. Rueping, A. Kuenkel, F. Tato, J. W. Bats, Angew. Chem. 2009,
121, 3754; Angew. Chem. Int. Ed. 2009, 48, 3699; b) A. Padwa,
S. K. Bur, Tetrahedron 2007, 63, 5341.
6] a) C. Khosla, AIChE Journal 2009, 55, 1926; b) C. M. Kao,
M. McPherson, R. N. McDaniel, H. Fu, D. E. Cane, C. Khosla,
J. Am. Chem. Soc. 1998, 120, 2478.
1
.38 (m, 3H); C NMR (400MHz, CDCl
3
7
3
, d): 132.7, 129.4, 128.6,
1
22.1, 81.8, 74.2., ii) for 1,4-diphenylbutane, H NMR (400MHz,
1
[
1
3
CDCl
3
, d): 7.22 (d, 2H), 7.14 (m, 3H), 2.6 (t, 2H), 1.63 (t, 2H);
, d): 128.6, 128.5, 125.9, 31.8, 22.9.
C
NMR (400MHz, CDCl
3
Sequential hydrosilylation–hydrogenation reactions: In a
typical procedure, the core of the PDMS reactors was loaded with
PdNPs (ꢂ4.8 mM by Pd atoms) and the shell with AuNPs (ꢂ3.5 mM
by Au atoms). A flask containing ꢂ50–120 PdNP/AuNP core/shell
cubes was dried by gas flame and cooled under Ar. To this dry
flask was added 40 mL of dry toluene, followed by the addition of
[7] H. Lodish, A. Berk, P. Matsudaira, C. A. Kaiser, M. Krieger,
M. P. Scott, S. L. Zipursky, J. Darnell, Molecular Cell Biology,
5th ed., W.H. Freeman and Company, New York 2004.
[
8] a) P. A. Srere, Annu. Rev. Biochem. 1987, 56, 89; b) B. S. J. Winkel,
Annu. Rev. Plant Biol. 2004, 55, 85; c) P. V. Reddy, A. B. Pardee,
Proc. Natl. Acad. Sci. USA 1980, 77, 331.
[9] F. Gelman, J. Blum, D. Avnir, D, J. Am. Chem. Soc. 2000, 122,
11999.
0
.3 mmol 4-methoxyphenylacetylene and 0.45 mmol of triethylsi-
lane. The reaction mixture was heated to 75 8C under hydrogen [10] F. Gelman, J. Blum, D. Avnir, Angew. Chem. 2001, 113, 3759;
Angew. Chem. Int. Ed. 2001, 40, 3647.
11] F. Gelman, J. Blum, D. Avnir, New J. Chem. 2003, 27, 205.
12] F. Gelman, J. Blum, D. Avnir, J. Am. Chem. Soc. 2002, 124, 14460–
atmosphere for up to 23 hours. Upon completion, the solution
[
[
was drawn out by syringe and the reactors were washed with
Cl . The crude products were obtained by combining
3
ꢆ 30 mL CH
2
2
14463.
all organic solution and evaporating solvent under vacuum. The
products containing two isomers (triethyl(2-(4-methoxyphenyl)-
ethyl)-silane and triethyl(1-(4-methoxyphenyl)ethyl)-silane in a
ꢂ5:2 molar ratio) were purified by column chromatography (silica,
[
13] B. Helms, S. J. Guillaudeu, Y. Xie, M. McMurdo, C. J. Hawker,
J. M. J. Fr e´ chet, Angew. Chem. Int. Ed. 2005, 44, 6384; Angew.
Chem. 2005, 117, 6542.
14] S. T. Scroggins, Y. Chi, J. M. J. Fr e´ chet, Angew. Chem. Int. Ed. 2009.
DOI: 10.1002/anie.200902945.
15] M. J. Alfonso, M. Mene’ndez, J. Santamaria, Ind. Eng. Chem. Res.
[
[
[
1
0% dichloromethane in hexane). Spectral characterization
1
triethyl(2-(4-methoxyphenyl)ethyl)-silane, H NMR (400MHz,
2001, 40, 1058.
CDCl , d): 7.08 (d, 2H), 6.79 (d, 2H), 3.76 (s, 3H), 2.53 (m, 2H),
3
16] A. V. Gaikwad, V. Boffa, J. E. ten Elshof, G. Rothenberg, Angew.
Chem. 2008, 120, 5487; Angew. Chem. Int. Ed. 2008, 47, 5407.
1
3
0
3
.90 (t, 9H), 0.83 (m, 2H), 0.52 (q, 6H); C NMR (400MHz, CDCl ,
8
62 www.small-journal.com
ß 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2010, 6, No. 7, 857–863

Sequential Reactions Directed by Core/Shell Catalytic Reactors
[
17] a) M. B. Runge, M. T. Mwangi, A. L. Miller, II, M. Perring,
N. B. Bowden, Angew. Chem. 2008, 120, 949; Angew. Chem.
[25] A. M. Caporusso, L. A. Aronica, E. Schiavi, G. Martra, G. Vitulli,
P. Salvadori, J. Organometallic Chemistry 2005, 690, 1063.
Int. Ed. 2008, 47, 935; b) A. L. Miller, N. B. Bowden, J. Org. Chem. [26] a) R. Klajn, K. J. M. Bishop, M. Fialkowski, M. Paszewski,
2
009, 74, 4834.
18] D. H. Turkenburg, A. A. Antipov, M. B. Thathagar, G. Rothenberg,
C. J. Campbell, T. P. Gray, B. A. Grzybowski, Science 2007, 316,
261; b) R. Klajn, P. J. Wesson, K. J. M. Bishop, B. A. Grzybowski,
Angew. Chem. 2009, 121, 7169; Angew. Chem. Int. Ed. 2009, 48,
1.
[
[
G. B. Sukhorukov, E. Eiser, Phys. Chem. Chem. Phys. 2005, 7,
2
237.
19] a) J. Bao, J. He, Y. Zhang, Y. Yoneyama, N. Tsubaki, Angew. Chem. [27] a) M. M. Santore, M. J. Kaufman, J. Polym. Sci. Polym. Phys. 1996,
2
008, 120, 359; Angew. Chem. Int. Ed. 2008, 47, 353; b) J. He,
Z. Liu, Y. Yoneyama, N. Nishiyama, N. Tsubaki, Chem. Eur. J. 2006,
2, 8296.
20] M. T. Mwangi, M. B. Runge, N. B. Bowden, J. Am. Chem. Soc. 2006,
28, 14434.
34, 1555; b) J. P. Robinson, E. S. Tarleton, K. Ebert, C. R. Millington,
A. Nijmeijer, Ind. Eng. Chem. Res. 2005, 44, 3238; c) T. Hajsz,
I. Csetneki, G. Filipcsei, M. Zrinyi, Phys. Chem. Chem. Phys. 2006,
8, 977.
1
[
[
[
[
1
[28] a) D. H. Cole, K. R. Shull, P. Baldo, L. Rehn, Macromolecules 1999,
32, 771; b) D. H. Cole, K. R. Shull, L. E. Rehn, P. M. Baldo, Nucl.
Instrum. Methods Phys. Res. Sect. B 1997, 283.
21] T. C. Merkel, V. I. Bondar, K. Nagal, B. D. Freeman, I. Pinna, J. Polym.
Sci.: Polym. Phys. 2000, 38, 415.
22] J. N. Lee, C. Park, G. M. Whitesides, Anal. Chem. 2003, 75, [29] a) B. A. Grzybowski, Chemistry in Motion: Reaction Diffusion
6
544.
Systems for Micro and Nanotechnology, Wiley, Chichester, UK
2009; b) W. M. Deen, Analysis of Transport Phenomena, Oxford
University Press, New York 1998.
23] a) L. Wu, B. L. Li, Y. Y. Huang, H. F. Zhou, Y. M. He, Q. H. Fan, Org.
Lett. 2006, 8, 3605; b) J. Alvarez, J. Liu, E. Roman, A. E. Kaifer,
Chem. Commun. 2000, 1151.
[
24] P. Siemsen, R. C. Livingston, F. Diederich, Angew. Chem. 2000,
Received: December 16, 2009
Published online: March 1, 2010
1
12, 2740; Angew. Chem. Int. Ed. 2000, 39, 2632.
small 2010, 6, No. 7, 857–863
ß 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com 863