Anodic aluminium oxide catalytic membranes for asymmetric epoxidation

Article abstract of DOI:10.1039/b507541f

Catechol-functionalized (salen)Mn complexes can be supported on mesoporous anodized aluminium oxide disks to yield catalytic membranes that are highly active in the enantioselective epoxidation of olefins when being deployed in a forced-through-flow react

Full text of DOI:10.1039/b507541f

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Anodic aluminium oxide catalytic membranes for asymmetric
epoxidation{
So-Hye Cho, Nolan D. Walther, SonBinh T. Nguyen* and Joseph T. Hupp*
Received (in Berkeley, CA, USA) 27th May 2005, Accepted 30th August 2005
First published as an Advance Article on the web 26th September 2005
DOI: 10.1039/b507541f
combination of activity, selectivity, and recyclability has not been
Catechol-functionalized (salen)Mn complexes can be supported
on mesoporous anodized aluminium oxide disks to yield
catalytic membranes that are highly active in the enantioselec-
tive epoxidation of olefins when being deployed in a forced-
through-flow reactor.
achieved to date.
We have long been interested in the design of a catalytic system
that can perform both a chemical transformation and a separation
event within a single functionalized membrane material. To this
end, we have explored the use of commercially available AAO
membranes (Anodisc1, Whatman) as the supporting materials for
chiral (salen)Mn complexes. Being non-compressible and posses-
sing high chemical and thermal stability, the AAO membranes can
serve as a stationary phase suitable for incorporation into catalytic
membrane reactors. In addition, their mono-disperse pores provide
well-defined surfaces upon which site isolation of the supported
catalyst can be more carefully controlled than in cross-linked
polymers or other inorganic solid supports.
Catalytic membrane reactors have attracted much attention over
the last decade due to practical advantages over other reactor
designs. They can potentially reduce the size of conventional
reactors and cost of operation by combining two essential
processes, chemical reaction and separation of the resulting
products from the catalyst, in one stage.1 In particular, the
catalytic membrane reactor configuration confers a significant
advantage to oxidation reactions—the use of a catalytic membrane
can provide a reactive interface for the oxidation to take place
while avoiding long contact times of the desired product with
catalysts, thereby minimizing over-oxidation.2
Herein, we report the fabrication of a chiral (salen)Mn-
immobilized AAO membrane and demonstrate its use in a
catalytic membrane reactor for enantioselective epoxidation. As
anchoring groups for the (salen)Mn complexes, we have chosen
catechol (1,2-dihydroxyphenyl), which has great affinity for AlIII
ions9 and can adsorb strongly onto Al2O3.10
Recent developments in the synthesis of inorganic materials
have allowed chemists to create single-site catalysts and catalyst
supports that provide a uniform environment around each active
catalyst center.3 Among these inorganic materials, mesoporous
anodic aluminium oxide (AAO) membranes have received great
attention.4 They contain well-ordered, densely packed, nanoscale
pores that naturally form when aluminium films are anodized in
an acidic electrolyte. Although AAO membranes can be brittle,
they are used in a number of diverse applications such as bio-
reactors;5a sensors;5b templates for quantum dots,5c nanowires,5d
nanotubes,5e and nanofibers;5f and supports for metals5g and
catalysts.5h
Our catechol-functionalized unsymmetrical chiral salen ligand
can be prepared easily from (1R,2R)-diaminocyclohexane in good
yield and then metallated with MnCl2/LiCl in air, resulting in
complex 1. Immobilization of 1 was accomplished by stirring its
ethanolic solution with an AAO membrane{ at 70 uC for 24 h,
resulting in a brown membrane (1-AAO) (Fig. 1). The catalyst
loading was determined by inductively coupled plasma (ICP)
spectroscopy to be 4.1 mmol g21. The modified membrane was
characterized by FT-IR, diffuse reflectance UV-Vis, and X-ray
An attractive potential use for AAO membrane would be as a
support for the immobilization of asymmetric homogeneous
catalysts to generate an enantioselective catalytic membrane.
Heterogenized asymmetric catalysts are intrinsically more econom-
ical and convenient to use than their homogeneous counterparts
since they can provide a direct route to chiral products without
costly separations.6 Among asymmetric homogeneous catalysts,
chiral (salen)Mn complexes (Jacobsen–Katsuki catalysts)7 for the
asymmetric epoxidation of olefins comprise one of the most widely
immobilized classes of homogeneous catalysts.8 However, the ideal
Department of Chemistry and the Institute for Environmental Catalysis,
Northwestern University, 2145 Sheridan Rd., Evanston, IL 60208, USA.
E-mail: stn@northwestern.edu; j-hupp@northwestern.edu.;
Fax: 847-491-7713; Tel: 847-467-3347
{ Electronic supplementary information (ESI) available: Synthetic proce-
dures and characterization data for complexes 1–2; characterization data
for 1-AAO; general procedure for asymmetric epoxidations catalyzed by 1-
AAO. See http://dx.doi.org/10.1039/b507541f
Fig. 1 The enantioselective catalytic AAO membrane (1-AAO) coated
with chiral (salen)MnIII catecholate complex 1. The complex was not
shown in scale to the channel dimensions.
This journal is ß The Royal Society of Chemistry 2005
Chem. Commun., 2005, 5331–5333 | 5331

View Article Online
Table 1 Catalytic performance of 1-AAO vs. 2 in the asymmetric
epoxidation of 2,2-dimethyl-2H-chromenea
Entry
Catalyst
Yield [%]b
ee [%]c
s [%]d
1
2
3
4
5e
a
1-AAO
79
70
62
51
82
81
76
73
69
86
100
99
97
95
1-AAO 2nd cycle
1-AAO 3rd cycle
1-AAO 4th cycle
2
Fig. 2 Schematic diagram of the liquid-phase forced-through-flow
catalytic membrane reactor for the enantioselective epoxidation of olefins.
100
compatibility with high flux (maximum flux for water 5
4.9 mL min21 cm22),19 fluxes of up to 7.5 mL h21 cm22 were
easily obtained with reactant mixture in dichloromethane, which is
much less viscous than water. In this liquid phase forced-through-
flow membrane reactor,20 1-AAO exhibited extremely high activity
and excellent chemoselectivity (Table 2, entries 1–2, 4).
Furthermore, its enantioselectivity was similar to that of 2 under
homogenous conditions. When the 7.5 mL h21 cm22 flux was
applied, 1-AAO provided unprecedented high turnover frequency
(TOF) and one-pass conversion (Table 2, entry 4). Although not
directly comparable to our present data, PDMS-membrane-
occluded Jacobsen’s catalyst only affords a TOF of 0.0438 h21
for the styrene–NaOCl system21 and [Al-MCM-41]-immobilized
Jacobsen’s catalyst only gave a TOF of 26 h21 for the stilbene–
iodosylbenzene system.16a Moreover, when 1-AAO was reused
after one cycle, it was still active and selective although the TOF
had dropped off significantly16c (Table 2, entry 5). It is probable
that at 2000 TON (total TON after 2 cycles at 10 mL h21, Table 2,
entries 3–4) we are close to the limit of TON for 1-AAO in the
presence of 3.
Reaction performed in a shell vial under ambient conditions using
magnetic stirring. Molar ratio olefin/oxidant/catalyst 5 100/200/1.
h
b
GC yield after
Determined using
1
using undecane as an internal standard.
Supelco b–DEX 120 chiral GC column.
c
a
d
Product selectivity of epoxide over ketone as measured by 1H
e
NMR spectroscopy.
conditions using 1 mol% catalyst.
Control experiment under homogeneous
photoelectron spectroscopy to confirm the presence of surface-
attached catalyst (see Supplementary Information{).
To provide a benchmark for the catalytic activity and selectivity
of 1-AAO in the homogeneous state, complex 211 was also
prepared. Catalytic epoxidations of 2,2-dimethyl-2H-chromene
were carried out with both 1-AAO and 2 using 2-(tert-
butylsulfonyl)iodosylbenzene, 3, as the oxidant (Table 1).12
Although initial rates cannot be obtained,13 under optimized
batch reaction conditions 1-AAO afforded the epoxide in yield and
selectivities that are similar to those of 2 (Table 1, cf. entries 1 and
5).14 These data suggest that reactant molecules have easy access to
the supported active sites on the membrane, and that the sites
retain structures that are similar to that of the free catalyst. This
notion makes sense given the well-ordered, highly porous, and
rigid morphology of AAO membrane and the large channel sizes
(¢20 nm), which define a readily accessible and unconstrained
environment15 for the immobilized catalyst.
The product/reactant ratio can be further improved to 80%
conversion (from 60%) by passing the initial permeate through
another identical 1-AAO membrane (Table 2, cf. entries 2 and 3).
The leached-out Mn in the permeate side does not contribute to
the overall activity as only a minimal increase (y0.6%) in
conversion was observed when the permeate was allowed to stir
for an additional hour after being passed through the membrane.
The high activity and selectivity for asymmetric epoxidation
obtained in the catalytic membrane reactor system can be ascribed
to the two beneficial effects of the AAO membrane: (1) the well-
ordered and unconstrained cylindrical pore structure of the
membrane makes all active catalyst sites readily available to
reactant molecules and (2) short catalyst contact time allows
more turnovers while maintaining high selectivity and minimizing
over-oxidation.
Moreover, catalytic membrane 1-AAO can be recovered and
reused after a simple cleaning procedure (Table 1, entries 2–4).
Although the recycled 1-AAO membrane did exhibit a gradual
decrease in activity,16 it was still active with excellent chemoselec-
tivity and good enantioselectivity after four cycles (85% of the
original enantioselectivity). ICP analysis carried out on
a
membrane after the fourth cycle showed that 83–87% of the
manganese still remained in the membrane.17
Membrane 1-AAO can also be used effectively in a simple
catalytic membrane reactor (Fig. 2) where the substrate and
oxidant are introduced via a syringe pump, which also controls the
feed rate of reactants.18 The catalytic results, obtained with the
various fluxes, are presented in Table 2. Due to the membrane’s
In summary, we have demonstrated, for the first time, that
chiral catalysts can be immobilized onto mesoporous AAO
Table 2 Catalytic performance of 1-AAO in a liquid phase forced-through-flow reactor for the asymmetric epoxidation of 2,2-dimethyl-
2H-chromene
Cycle for
membrane
Cycle for
permeate
Entry
Flux [mL h21 cm22 a
]
Total TONb
Conversion [%]c
TOF [min21 d
]
ee [%]e s [%]f
1
2
3
4
5g
a
1
1
1
2
1
1
1.5
4.5
4.5
7.5
7.5
1174
1492
47
60
80
67
7
20
75
n/a
134
15
86
85
85
84
80
100
100
100
100
98
1
1h
1
1174 (1st) + 820 (2nd)
1670
179
2
Controlled by a syringe pump. b Total turnover number, based on conversion per catalyst. Molar ratio olefin/oxidant/catalyst 5 10,000/2,500/1.
Based on oxidant as the limiting reagent. Turnover frequency, based on conversion per catalyst per minute. Determined using a
c
d
e
f
Supelco b–DEX 120 chiral GC column. Product selectivity of epoxide over ketone, measured by NMR spectroscopy. After one cycle
g
h
(entry 3), 1-AAO was recovered, thoroughly washed with CH2Cl2, and reused with fresh reactants. The permeate of entry 2 was
passed through another freshly prepared 1-AAO.
5332 | Chem. Commun., 2005, 5331–5333
This journal is ß The Royal Society of Chemistry 2005

View Article Online
M. Boman and J. O. Carlsson, Mater. Sci. Eng., C, 2003, C23,
823–826; (h) K. N. Rai and E. Ruckenstein, J. Catal., 1975, 40, 117–23.
6 For recent reviews: (a) P. McMorn and G. J. Hutchings, Chem. Soc.
Rev., 2004, 33, 108–122; (b) L.-X. Dai, Angew. Chem., Int. Ed., 2004, 43,
5726–5729; (c) D. E. de Vos, I. F. J. Vankelecom and P. A. Jacobs,
Chiral Catalyst Immobilization and Recycling, Wiley-VCH, New York,
2000; (d) D. C. Sherrington and A. P. Kybett, Supported Catalysts and
Their Applications, Royal Society of Chemistry, Cambridge, 2001.
7 E. N. Jacobsen, in Catalytic Asymmetric Synthesis, ed. I. Ojima, VCH
Publishers Inc., Cambridge, 1993.
8 For recent reviews: (a) C. E. Song and S.-G. Lee, Chem. Rev., 2002, 102,
3495–3524; (b) D. C. Sherrington, Catal. Today, 2000, 57, 87–104.
9 (a) S. M. Cohen and K. N. Raymond, Inorg. Chem., 2000, 39,
3624–3631; (b) M. H. Chisholm, J. Gallucci, D. Navarro-Llobet and
H. Zhen, Polyhedron, 2003, 22, 557–561.
membranes, which can then be used effectively in enantioselective
catalytic membrane reactors. The supported catalytic membrane
provided easy catalyst separation and recycling with comparable
activity and selectivity to homogeneous counterparts. When used
in a forced-through-flow reactor, it offered great flexibility in the
control of substrate feed and product separation from the catalyst.
Under optimized conditions, unusually high activity (TON 5 1670,
TOF 5 135 min21) with high enantioselectivity and excellent
chemoselectivity can be observed for olefin epoxidation. This
strategy likely can be extended to other catalytic reactions,
allowing for straightforward inexpensive isolation of valuable
products as well as minimization of side reactions.
10 (a) P. A. Connor, K. D. Dobson and A. J. McQuillan, Langmuir, 1995,
11, 4193–4195; (b) Y. Ho, Y.-L. Lee and K.-Y. Hsu, J. Chromatogr., B:
Biomed. Appl., 1995, 665, 383–389.
The authors gratefully acknowledge support by the Institute of
Environmental Catalysis at Northwestern University (DOE grant
# DE-FG02-03ER15457).
11 Complex 2 is used in the control experiment instead of 1 because the
latter is not very soluble in CH2Cl2 due to the exposed hydroxyl groups
on the catechol moiety. In addition, the acidity of the phenolic protons
of 1 can also cause side reaction with the epoxide products.
12 The use of excess oxidant was necessary because (tert-butylsulfonyl)-
iodosylbenzene undergoes partial disproportionation to (tert-butylsulfo-
nyl)iodoxylbenzene and (tert-butylsulfonyl)iodobenzene under our
reaction conditions. See: D. Macikenas, E. Skrzypczak-Jankun and
J. D. Protasiewicz, J. Am. Chem. Soc., 1999, 121, 7164–7165.
13 Under the reaction conditions shown in Table 1, the first 40% of
conversion occurred within 1 min, rendering the measurement of initial
rate via manual sampling impossible. This fast reaction rate is due to the
high solubility of the oxidant (see ref. 12) as well as the high activity of
the (salen)MnvLO intermediate towards olefins.
14 Due to phase-induced kinetic differences, the activities of salen-derived
heterogenized catalysts for olefin epoxidation are often compared to
those of their homogeneous counterparts in terms of TON or TOF, and
not initial rate. See: (a) S. Keith and C.-H. Liu, Chem. Commun., 2002,
886–887; (b) F. Bigi, L. Moroni, R. Maggi and G. Sartori, Chem.
Commun., 2002, 716–717.
15 Unlike AAO or mesoporous inorganic oxides with large pores, cross-
linked polymer or small-pore zeolite supports may force the immobilized
catalysts to adopt geometries that are significantly different from those
of their homogenous counterparts due to the limited size of the
supporting cavities in these latter materials. See: (a) L. Canali and
D. C. Sherrington, Chem. Soc. Rev., 1999, 28, 85–93; (b) F. Bedioui,
Coord. Chem. Rev., 1995, 144, 39–68.
16 A similar tendency toward the decline in activity and selectivity over
recycling has been observed for other heterogenized epoxidation
catalysts. See: (a) P. Piaggio, P. McMorn, C. Langham, D. Bethell,
P. C. Bulman-Page, F. E. Hancock and G. J. Hutchings, New J. Chem.,
1998, 22, 1167–1169; (b) S. B. Ogunwumi and T. Bein, Chem. Commun.,
1997, 901–902; (c) J. M. Fraile, J. I. Garcia, J. Massam and
J. A. Mayoral, J. Mol. Catal. A: Chem., 1998, 136, 7–57. This decline
in activity and selectivity is possibly a consequence of the catalyst
leaching (see ref. 17), ligand decomposition, or site blockage by reactant
or product molecules.
17 Although ICP analysis can yield the amount of manganese present in a
membrane, not all of the Mn ions exist as supported (salen)Mn
complex. The 83–87% of the initial Mn still present in the membrane
after the 4th cycle represents an upper limit on the amount of active
(salen)Mn complex that has not been leached out, decomposed, or
blocked.
18 Direct comparison of the batch reactor (Table 1) and the forced-
through-flow reactor (Table 2) cannot be performed due to the
differences in reagent concentrations and mass transport (flow through
vs. diffusion).
19 http://www.whatman.com.
20 (a) R. Dittmeyer, K. Svajda and M. Reif, Top. Catal., 2004, 29, 3–27; (b)
M. Reif and R. Dittmeyer, Catal. Today, 2003, 82, 3–14.
21 I. F. J. Vankelecom, D. Tas, R. F. Parton, V. Van de Vyver and
P. A. Jacobs, Angew. Chem., Int. Ed. Engl., 1996, 35, 1346–1348.
Notes and references
{ Preparation of 1-AAO. Into a 15 mL Pyrex conical test tube (15 mm ID,
Corning product #8060-15 ) equipped with a vaned magnetic stirbar were
placed 20 membranes of Anodisc1 13 (Whatman, 13 mm diameter, 60 mm
thickness, 20 nm front-end pore), each separated from the others by thin
cylindrical polyethylene rings (12 mm ID 6 1.5 mm H 6 0.5 mm T). An
ethanol solution (12 mL) of 1 (12 mg, 1.8 6 1022 mmol) was added to the
tube. The tube was capped with a rubber septum and the reaction was
heated at 70 uC for 24 h. The membranes were then separated, thoroughly
washed with EtOH (5 6 15 mL), and sonicated in EtOH (3 6 15 mL) for
5 min each. The resulting brown-colored membranes were air-dried for
15 min and baked at 70 uC in an oven for 10 min. They were then stored in
the vacuum desiccator before use.
1 For recent reviews: (a) J. G. Sanchez Marcano and T. T. Tsotsis,
Catalytic Membranes and Membrane Reactors, Wiley-VCH, Weinheim,
2002; (b) S. P. Nunes and K. V. Peinemann, Membrane Technology in
the Chemical Industry, Wiley-VCH, New York, 2001; (c)
I. F. J. Vankelecom, Chem. Rev., 2002, 102, 3779–3810; (d)
T. Maschmeyer and J. C. Jansen, Top. Catal., 2004, 29, 1–92.
2 (a) G. Langhendries, G. V. Baron, I. F. J. Vankelecom, R. F. Parton
and P. A. Jacobs, Catal. Today, 2000, 56, 131–135; (b) A. A. Yawalkar,
V. G. Pangarkar and G. V. Baron, J. Membr. Sci., 2001, 182, 129–137;
(c) P. J. Gellings and H. J. M. Bouwmeester, Catal. Today, 2000, 58,
1–53.
3 For recent reviews: (a) A. T. Bell, Science, 2003, 299, 1688–1691;
D. R. Rolison, Science, 2003, 299, 1698–1702; (b) S. L. Scott,
C. M. Crudden and C. W. Jones, Nanostructured Catalysts, Kluwer
Academic, New York, 2003; (c) J. Coronas and J. Santamaria, Top.
Catal., 2004, 29, 29–44; (d) A. Julbe, D. Farrusseng and C. Guizard,
J. Membr. Sci., 2001, 181, 3–20.
4 For recent reviews: (a) D. Honicke and E. Dietzsch, in Handbook of
Porous Solids, ed. F Schu¨th, K. S. W. Sing and J. Weitkamp, Wiley-
VCH, Weinheim, 2002; (b) S. E. Park, Nanotechnology in
Mesostructured Materials (Proceedings of the 3rd International
Mesostructured Materials Symposium, Jeju, Korea), Elsevier, New
York, 2003; (c) S. Shingubara, J. Nanopart. Res., 2003, 5, 17–30; (d)
F. Li, L. Zhang and R. M. Metzger, Chem. Mater., 1998, 10,
2470–2480.
5 (a) W. Yang, H.-Y. Qu, H.-H. Yang and J.-G. Xu, Anal. Lett., 2004, 37,
1793–1809; (b) K. H. A. Lau, L.-S. Tan, K. Tamada, M. S. Sander and
W. Knoll, J. Phys. Chem. B, 2004, 108, 10812–10818; (c) Z. L. Wang,
Y. Liu and Z. Zhang, Handbook of Nanophase and Nanostructured
Materials, Kluwer Academic/Plenum, New York, 2003; (d) Q. Zhang,
Y. Li, D. Xu and Z. Gu, J. Mater. Sci. Lett., 2001, 20, 925–927; (e)
K. Delendik, I. Emeliantchik, A. Litomin, V. Rumyantsev and
O. Voitik, Nucl. Phys. B, Proc. Suppl., 2003, 125, 394–399; (f)
L. Zhang, B. Cheng and E. T. Samulski, Chem. Phys. Lett., 2004,
398, 505–510; (g) A. Johansson, T. Torndahl, L. M. Ottosson,
This journal is ß The Royal Society of Chemistry 2005
Chem. Commun., 2005, 5331–5333 | 5333