Artificial metalloenzymes via encapsulation of hydrophobic transition-metal catalysts in surface-crosslinked micelles (SCMs)

Article abstract of DOI:10.1039/c2cc33012a

Encapsulation of a hydrophobic rhodium catalyst in crosslinked micelles allowed nonpolar substrates to react in water with unusual selectivity. This journal is The Royal Society of Chemistry 2012.

Full text of DOI:10.1039/c2cc33012a

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
ChemComm
Cite this: Chem. Commun., 2012, 48, 9998–10000
www.rsc.org/chemcomm
COMMUNICATION
Artificial metalloenzymes via encapsulation of hydrophobic
transition-metal catalysts in surface-crosslinked micelles (SCMs)w
Shiyong Zhang and Yan Zhao*
Received 26th April 2012, Accepted 17th August 2012
DOI: 10.1039/c2cc33012a
Encapsulation of a hydrophobic rhodium catalyst in crosslinked
micelles allowed nonpolar substrates to react in water with
unusual selectivity.
Chemists have long marveled at the abilities of enzymes to catalyze
reactions under ambient conditions in water with extraordinary
selectivity. Water is an attractive green solvent due to its abundance,
low cost, nonflammability, and nontoxicity. Aqueous biphasic
Fig. 1 Preparation of the SCM-encapsulated rhodium catalyst and
catalysis, in particular, has the potential benefits of facile product
schematic representation of the biphasic catalytic hydrogenation of
isolation and catalyst recovery as a result of the phase-separated
organic products and the reaction medium.1 The process, however,
water-insoluble alkenes in water.
is limited industrially to a few reactions in which the organic
starting material has substantial solubility in water.
a hydrophobic fluorescent probe (pyrene) inside the SCMs and
found that the probe stayed entrapped for over 6 months.10 In this
work, we solubilized a commercially available bisphosphine
rhodium(I) complex (3) in water by surfactant 1 at [3]/[1] =
1/50. With an estimated micellar aggregation number of
B50,9,11 each SCM should contain one rhodium complex on
average. After crosslinking, we ‘‘terminated’’ the residual alkyne
groups of the SCMs by reacting with excess 2-azidoethanol and
removed the water-soluble impurities such as copper salts by
dialysis against water.12 The concentration of Rh was measured
by ICP-MS and translated to 0.92 Æ 0.03 rhodium atom per
SCM (see the ESIw for details).
It is not enough for the rhodium complex to be solubilized
by the SCM. To be useful in aqueous biphasic catalysis, the
complex has to be physically trapped within the crosslinked
micelle—this is the key difference between our method and
reported solubilization of organometallic catalysts in surfactant
micelles.13 Fig. 2 shows two aqueous solutions of 3 solubilized by
CTAB (i.e., cetryltrimethylammonium bromide) and SCMs,
respectively. When the solutions were placed on top of a chloro-
form layer, the yellow color of the rhodium complex stayed in the
The fundamental difficulty in expanding aqueous biphasic
catalysis to more hydrophobic substrates lies in the latter’s low
solubility in water.1,2 Although many organometallic catalysts
can be made water-soluble,2,3 a nonpolar reactant has
difficulty accessing the catalyst in the aqueous phase. Organic
co-solvents and surfactants may be added to ease the solubility
problem but, inevitably, compromises the phase separation of
the product. To solve the dilemma, chemists have investigated
many ligands and supports for the catalysts, including water-
soluble calixarenes,4 thermally regulated phase-transfer ligands,5
cyclodextrin derivatives,6 polystyrene lattices,7 and crosslinked
polymeric microreactors.8
How can we make organometallic catalysts soluble in water
without affecting their accessibility by nonpolar organic
molecules? We herein report a method to trap conventional
hydrophobic organometallic catalysts in crosslinked micelles
(Fig. 1). The SCMs contained hydrophobic sites near the
catalysts, which could bind nonpolar substrates and create
unusual selectivity for the catalysis. Moreover, the entrapment
protects the catalysts from deactivation pathways such as
dimerization.
The SCMs were prepared by crosslinking alkynylated
surfactant 1 in the micellar configuration with 1.5 equiv. of 2
in the presence of Cu(I) catalysts.9 We previously encapsulated
Department of Chemistry, Iowa State University, Ames,
Iowa 50011-3111, USA. E-mail: zhaoy@iastate.edu;
Fax: +1-515-294-0105; Tel: +1-515-294-5845
w Electronic supplementary information (ESI) available: Experimental
details for the syntheses and hydrogenation, and additional reaction
data (PDF). See DOI: 10.1039/c2cc33012a
Fig. 2 Comparison of rhodium complex 3 protected by CTAB (A)
and SCM (B) in the presence of CHCl3 (a) before, (b) after 2 min of
hand-shaking and 1 min of standing, and (c) after standing overnight
at room temperature.
c
9998 Chem. Commun., 2012, 48, 9998–10000
This journal is The Royal Society of Chemistry 2012

View Article Online
aqueous phase initially (Fig. 2a). After being hand-shaken and
left standing for 1 min, the CTAB-containing sample (A) turned
into an emulsion, whereas the SCM-containing sample (B) quickly
separated into two layers (Fig. 2b). Most importantly, upon
further standing, the rhodium complex migrated to the organic
phase completely in sample A but remained in the aqueous phase
in sample B (Fig. 2c). In fact, the yellow color of the SCM sample
stayed completely in the aqueous phase even if the solution was
stirred with a large amount of chloroform for several days.
Encouraged by the ease of phase-separation and the physical
containment of the rhodium complex, we examined the ability of
Rh(I)–[PPh3]2@SCM to catalyze hydrogenation of hydrophobic
alkenes. Catalyst optimization was performed with 1-octene and
1-decene as the model substrates. In general, the alkene was mixed
with the aqueous solution of SCM-encapsulated catalyst under
vigorous stirring at 800 psi H2. After 24 h at room temperature,
1-octene was converted to the alkane quantitatively but the yield
for 1-decene was only 23% (Table 1, entries 1 and 2).
Table 2 Catalytic hydrogenation with Rh(I)–[PPh3]2@SCM prepared
with 25 mol% 1-dodecanol as the additivea
Entry
Substrate
Yield (%)
1
2
3
4
5
6
7
8
1-Hexene
1-Octene
1-Decene
1-Dodecene
Styrene
Cyclohexene
Butyl acrylate
t-Butyl acrylate
Allyl alcohol
Pent-4-en-1-ol
2-trans-Octeneb
3-trans-Octeneb
4-trans-Octeneb
>95
>95
>95
21
78
22
>95
79
26
28
22
9
10
11
12
13
10
18
a
Catalytic hydrogenation was carried out with 0.2 mol% Rh catalyst
at 1200 psi H2 for 24 h. Yields were determined by 1H NMR
spectroscopy after the reaction mixture was extracted with dichlor-
b
omethane. Yields were determined by GC. GC-MS was used to
identify the identity of the products.
These results were encouraging to us. Although the longer
alkene had difficulty approaching the catalyst, 1-octene, being
highly hydrophobic, had no problem entering the SCM to react.
To improve the reactivity of the longer alkene, we prepared
Rh(I)–[PPh3]2@SCM in the presence of various additives
(Table 1). The hypothesis was that the additives would occupy
the space inside the micelles of 1 and, after crosslinking could be
removed to create hydrophobic sites near the catalyst. However,
neither cyclohexane, which was expected to be located in the
hydrophobic core of the micelle, nor CTAB, which occupied
space both on the surface and within the core of the micelle,
afforded better catalysts (Table 1, entries 3–6).
for 1-dodecene. Heating the reaction mixture at 60 1C gave
essentially the same results. The inherent chemical reactivity of
1-decene and 1-dodecene could not explain the large difference
in the yields. Neither substrate has significant solubility in
water. It seems that the crosslinked micelle, limited by the chain
length of the hydrophobic tail, could only accommodate hydro-
carbons with a certain chain length. 1-Dodecene was probably
too long to fit within the hydrophobic sites of the SCM (Fig. 3).
Exposure of a hydrophobic compound to water is unfavorable.
If a large section of 1-dodecene has to be exposed to water in
order to react with the entrapped rhodium, the hydrogenation is
expected to be disfavored by the hydrophobic effect.
1-Dodecanol turned out to be a much better additive. The
catalyst prepared with 25 mol% of the alcohol converted
1-octene and 1-decene quantitatively. Note that, if the additive
was not removed after the preparation of Rh(I)–[PPh3]2@SCM,
the hydrogenation was very sluggish (Table 1, entry 9). Clearly,
the additive was occupying hydrophobic sites inside the SCM
that were critical for the substrate to access the rhodium metal.
We then examined the reactivity of a number of alkenes to
understand the selectivity of the SCM-encapsulated catalyst. As
shown in Table 2, linear alkenes with 6 to 10 carbons had no
difficulty getting to the catalytic site (entries 1–3). An increase of
two additional carbons, however, reduced the yield to 21%
We also prepared an analogue of 1 with a C16 hydrocarbon
tail and used cetyl alcohol as the additive to prepare
Rh(I)–[PPh3]2@SCM. We were hoping that the longer hydro-
carbon tail might allow the longer alkene to be accommo-
dated. Unfortunately, the longer surfactant had difficulty
forming micelles at room temperature. Although the catalyst
entrapment could be performed at higher temperature (ca.
50 1C), no improvement in the hydrogenation of 1-dodecene
was observed. Since the catalysts prepared at different
temperatures could be different, other factors might have
contributed to the results (i.e., lack of improvement).
Table 1 Optimization of the preparation of Rh(I)–[PPh3]2@SCMa
It seems to us that, even if a surfactant with a longer
hydrocarbon tail is used successfully in the SCM preparation,
Entry
Additiveb
Substrate
Yield (%)
1
2
3
4
5
6
7
8
9
None
None
1-Octene
1-Decene
1-Octene
1-Decene
1-Octene
1-Decene
1-Octene
1-Decene
1-Decene
>95
23
>95
Trace
>95
22
>95
>95
35
100 mol% cyclohexane
100 mol% cyclohexane
10 mol% CTAB
10 mol% CTAB
25 mol% 1-dodecanol
25 mol% 1-dodecanol
25 mol% 1-dodecanolc
a
Catalytic hydrogenation was carried out with 0.2 mol% Rh catalyst
at 800 psi H2 for 24 h. Yields were determined by the 1H NMR spectroscopy
after the reaction mixture was extracted with dichloromethane (see ESI for
details). b The additive was added during the preparation of the
Rh(I)–[PPh3]2@SCM and removed by CH2Cl2 extraction after the cross-
linking. c 1-Dodecanol was not removed after the catalyst preparation.
Fig. 3 Schematic representation of the hypothesized chain-length
selectivity.
c
This journal is The Royal Society of Chemistry 2012
Chem. Commun., 2012, 48, 9998–10000 9999

View Article Online
samples in between the reactions,16 the rhodium catalyst was
extremely robust when encapsulated inside the SCM.
In conclusion, physical entrapment of a conventional hydro-
phobic transition metal catalyst within the water-soluble surface-
crosslinked micelle provided a hydrophobic microenvironment
around the catalyst. In comparison to other methods to prepare
water-soluble transition metal catalysts,1–3 our method requires no
structural modification of the catalyst and enables hydrophobic
substrates to access the catalyst in water. The activity of the
catalyst was still rather low, as the complete reaction required
24 h at room temperature.17 Although rudimentary in comparison
to the substrate selectivity found in biocatalysts, the chain-
length and terminal/internal selectivity displayed by the SCM-
encapsulated rhodium demonstrated the potential power of the
supramolecular confinement. Overall, the Rh(I)–[PPh3]2@SCM
catalyst has remarkable resemblance to natural metalloenzymes
with water-solubility, modifiable surface groups,9 an internal
catalytic site, and hydrophobic binding sites. Further modification
of these catalytic nanoparticles should endow them with additional
features, possibly creating useful, reusable catalysts for aqueous
biphasic catalysis.
Fig. 4 Reusing of Rh(I)–[PPh3]2@SCM in the hydrogenation of
1-octene.
the longer tail does not necessarily translate to a deeper location
of the catalyst. The hydrophobic tails of the surfactants need to
aggregate tightly to maximize hydrophobic interactions. In the
presence of a bulky rhodium complex, these tails would kink
and possibly wrap around the catalyst, diminishing the
potential ‘‘depth’’ of the catalyst in the SCM. Besides, there is
no reason for the catalyst to stay in the center of the SCM.
Possibly, the selectivity in the Rh(I)–[PPh3]2@SCM was caused
by small ‘‘crevices’’ in the crosslinked micelles formed after the
removal of the alcohol additive. Linear alkenes can squeeze into
these crevices to access the metal center. If the alkenes are too
bulky or too long, they will have difficulty fitting into these
crevices, making them less reactive.
We thank the U.S. Department of Energy—Office of Basic
Energy Sciences (grant DE-SC0002142) for supporting the
research.
Notes and references
1 B. Cornils and W. A. Herrmann, Aqueous-Phase Organometallic Cata-
lysis: Concepts and Applications, 2nd edn, Wiley-VCH, Weinheim, 2004.
2 K. H. Shaughnessy, Chem. Rev., 2009, 109, 643.
3 (a) C. J. Li, Chem. Rev., 2005, 105, 3095; (b) U. M. Lindstrom,
Chem. Rev., 2002, 102, 2751.
4 S. Shimizu, S. Shirakawa, Y. Sasaki and C. Hirai, Angew. Chem.,
Int. Ed., 2000, 39, 1256.
5 (a) C. Liu, J. Y. Jiang, Y. H. Wang, F. Cheng and Z. L. Jin, J. Mol.
Catal. A: Chem., 2003, 198, 23; (b) D. E. Bergbreiter and S. D. Sung,
Adv. Synth. Catal., 2006, 348, 1352.
6 J. Cabou, H. Bricout, F. Hapiot and E. Monflier, Catal. Commun.,
2004, 5, 265.
7 K. Kunna, C. Muller, J. Loos and D. Vogt, Angew. Chem., Int.
¨
Ed., 2006, 45, 7289.
8 Y. Lan, M. C. Zhang, W. Q. Zhang and L. Yang, Chem.–Eur. J.,
2009, 15, 3670.
Reactions for the other alkenes seemed to be consistent with
the above explanation. In general, linear, ‘‘slimmer’’ alkenes
were more reactive than bulkier ones (Table 2). Although the
difference was not large, butyl acrylate was clearly more
reactive than t-butyl acrylate. Hydrophilic alcohols (allyl
alcohol and pent-4-en-1-ol) were unreactive, fully in line with
the hydrophobic microenvironment around the catalyst.
The most interesting selectivity was the terminal versus internal
for the linear alkenes. Although the hydrogenation of 1-octene
proceeded smoothly (Table 2, entry 2), none of the internal
octenes gave good yields (entries 11–13). When the reactions
were performed in methanol homogenously, 1-octene was more
reactive than the other octenes by 2–3-fold (data not shown).
The highest selectivity among the octenes was >9 with the
SCM-encapsulated rhodium (entries 2 and 12), clearly due to
the supramolecular control of the reactivity. The rhodium catalyst
was confirmed to be physically trapped inside the SCM. In
addition to the chloroform extraction experiment (Fig. 2), we
analyzed the methylene chloride extract after the hydrogenation.
The concentration of Rh determined by ICP-MS was 1.05 Æ
0.08 ppb, which corresponded to ca. 0.01% of Rh leaching.
Catalytically active rhodium(I) species can be deactivated
easily in homogeneous solution by dimerization.14 Such deac-
tivation will be difficult with the catalyst protected by the
SCM. Indeed, the Rh(I)–[PPh3]2@SCM catalyst could be
reused many times in the biphasic catalysis (Fig. 4). Only in
the eighth cycle, a significant decrease in yield (to 77%)
occurred.15 The turn-over frequency (TOF) of the catalyst
stayed largely unchanged in the repeated reactions, either
at the end of 24 h when the reaction was near completion
(Table S1, ESIw) or at 6 h at relatively low conversions
(Table S2, ESIw). Considering the harsh treatment of the
9 S. Zhang and Y. Zhao, Macromolecules, 2010, 43, 4020.
10 S. Zhang and Y. Zhao, J. Am. Chem. Soc., 2010, 132, 10642.
11 H.-Q. Peng, Y.-Z. Chen, Y. Zhao, Q.-Z. Yang, L.-Z. Wu,
C.-H. Tung, L.-P. Zhang and Q.-X. Tong, Angew. Chem., Int. Ed.,
2012, 51, 2088.
12 The concentration of copper in the sample after the dialysis was
1.42 Æ 0.07 ppm according to ICP-MS.
13 (a) F. M. Menger, L. H. Gan, E. Johnson and D. H. Durst, J. Am.
Chem. Soc., 1987, 109, 2800; (b) T. Dwars, E. Paetzold and G. Oehme,
Angew. Chem., Int. Ed., 2005, 44, 7174; (c) B. H. Lipshutz and
S. Ghorai, Aldrichimica Acta, 2008, 41, 59.
14 J. F. Hartwig, Organotransition Metal Chemistry: From Bonding to
Catalysis, University Science Books, Sausalito, CA, 2010, pp. 588.
15 The reaction conditions were not optimized. The fluctuation in the
yields in the first several cycles of reaction most likely resulted from
variations in the stirring conditions resulted from the hetero-
geneous nature of the reaction.
16 After each cycle of hydrogenation, the aqueous phase (2 mL), where
the catalyst was, was extracted with methylene chloride (3 Â 3 mL)
and heated at 50 1C for B2 min to evaporate the residual methylene
chloride before the next cycle of hydrogenation. The extraction and
evaporation were performed in air without special protection.
17 The low reactivity was probably caused by steric hindrance around
the catalyst and possibly also the difficulty of the nonpolar alkene
to migrate in the aqueous phase.
c
10000 Chem. Commun., 2012, 48, 9998–10000
This journal is The Royal Society of Chemistry 2012