Isoreticular chiral metal-organic frameworks for asymmetric alkene epoxidation: Tuning catalytic activity by controlling framework catenation and varying open channel sizes

Article abstract of DOI:10.1021/ja1069773

A family of isoreticular chiral metal-organic frameworks (CMOFs) of the primitive cubic network topology was constructed from [Zn₄(μ ₄-O)(O₂CR)₆] secondary building units and systematically elongated dicarboxyla

Full text of DOI:10.1021/ja1069773

Published on Web 10/11/2010
Isoreticular Chiral Metal-Organic Frameworks for Asymmetric
Alkene Epoxidation: Tuning Catalytic Activity by Controlling
Framework Catenation and Varying Open Channel Sizes
Feijie Song, Cheng Wang, Joseph M. Falkowski, Liqing Ma, and Wenbin Lin*
Department of Chemistry, CB#3290, UniVersity of North Carolina, Chapel Hill,
North Carolina 27599, United States
Received August 4, 2010; E-mail: wlin@unc.edu
Abstract: A family of isoreticular chiral metal-organic frameworks (CMOFs) of the primitive cubic network
topology was constructed from [Zn4(µ4-O)(O2CR)6] secondary building units and systematically elongated
dicarboxylate struts that are derived from chiral Mn-Salen catalytic subunits. CMOFs 1-5 were synthesized
by directly incorporating three different chiral Mn-Salen struts into the frameworks under solvothermal
conditions, and they were characterized by a variety of methods, including single-crystal X-ray diffraction,
1
PXRD, TGA, and H NMR. Although the CMOFs 1 vs 2 and CMOFs 3 vs 4 pairs were constructed from
the same building blocks, they exhibit two-fold interpenetrated or non-interpenetrated structures, respectively,
depending on the steric sizes of the solvents that were used to grow the MOF crystals. For CMOF-5, only
a three-fold interpenetrated structure was obtained due to the extreme length of the Mn-Salen-derived
dicarboxylate strut. The open channel and pore sizes of the CMOF series vary systematically, owing to the
tunable dicarboxylate struts and controllable interpenetration patterns. CMOFs 1-5 were shown to be highly
effective catalysts for asymmetric epoxidation of a variety of unfunctionalized olefins with up to 92% ee.
The rates of epoxidation reactions strongly depend on the CMOF open channel sizes, and the catalytic
activities of CMOFs 2 and 4 approach that of a homogeneous control catalyst. These results suggest that,
although the diffusion of bulky alkene and oxidant reagents can be a rate-limiting factor in MOF-catalyzed
asymmetric reactions, the catalytic activity of the CMOFs with large open channels (such as CMOFs 2 and
4 in the present study) is limited by the intrinsic reactivity of the catalytic molecular building blocks. The
CMOF catalysts are recyclable and reusable and retain their framework structures after epoxidation reactions.
This work highlights the potential of generating highly effective heterogeneous asymmetric catalysts via
direct incorporation of well-defined homogeneous catalysts into framework structures of MOFs.
Introduction
characterized by X-ray crystallography, which facilitates detailed
elucidation of their structure-property relationships and fine-
tuning of their functions. Such an ability to combine organic
synthetic methodologies and crystal engineering in MOF
Metal-organic frameworks (MOFs) have attracted a great
deal of recent interest owing to the ability to fine-tune their
properties at the molecular level.1-5 Unlike traditional inorganic
materials, MOFs are typically synthesized under mild conditions,
allowing for the incorporation of constituent building blocks
with desired functionalities, leading to numerous functional
MOFs that have shown promise for a number of applications,
such as gas storage,6-9 chemical sensing,10-14 catalysis,15-18
biomedical imaging,19,20 and drug delivery.21,22 MOFs also tend
to adopt ordered, crystalline structures that can be readily
(10) Allendorf, M. D.; Houk, R. J.; Andruszkiewicz, L.; Talin, A. A.;
Pikarsky, J.; Choudhury, A.; Gall, K. A.; Hesketh, P. J. J. Am. Chem.
Soc. 2008, 130, 14404.
(11) Chen, B.; Wang, L.; Xiao, Y.; Fronczek, F. R.; Xue, M.; Cui, Y.;
Qian, G. Angew. Chem., Int. Ed. 2009, 48, 500.
(12) Lan, A.; Li, K.; Wu, H.; Olson, D. H.; Emge, T. J.; Ki, W.; Hong,
M.; Li, J. Angew. Chem., Int. Ed. 2009, 48, 2334.
(13) Xie, Z.; Ma, L.; deKrafft, K. E.; Jin, A.; Lin, W. J. Am. Chem. Soc.
2010, 132, 922.
(1) Bradshaw, D.; Warren, J. E.; Rosseinsky, M. J. Science 2007, 315,
977.
(14) White, K. A.; Chengelis, D. A.; Gogick, K. A.; Stehman, J.; Rosi,
N. L.; Petoud, S. J. Am. Chem. Soc. 2009, 131, 18069.
(15) Kesanli, B.; Lin, W. B. Coord. Chem. ReV. 2003, 246, 305.
(16) Ingleson, M. J.; Barrio, J. P.; Bacsa, J.; Dickinson, C.; Park, H.;
Rosseinsky, M. J. Chem. Commun. 2008, 1287.
(17) Lee, J.; Farha, O. K.; Roberts, J.; Scheidt, K. A.; Nguyen, S. T.; Hupp,
J. T. Chem. Soc. ReV. 2009, 38, 1450.
(18) Ma, L.; Abney, C.; Lin, W. Chem. Soc. ReV. 2009, 38, 1248.
(19) deKrafft, K. E.; Xie, Z.; Cao, G.; Tran, S.; Ma, L.; Zhou, O. Z.; Lin,
W. Angew. Chem., Int. Ed. 2009, 48, 9901.
(20) Lin, W.; Rieter, W. J.; Taylor, K. M. Angew. Chem., Int. Ed. 2009,
48, 650.
(21) Rieter, W. J.; Pott, K. M.; Taylor, K. M.; Lin, W. J. Am. Chem. Soc.
2008, 130, 11584.
(22) Horcajada, P.; et al. Nat. Mater. 2010, 9, 172.
(2) Evans, O. R.; Lin, W. Acc. Chem. Res. 2002, 35, 511.
(3) Ferey, G.; Mellot-Draznieks, C.; Serre, C.; Millange, F. Acc. Chem.
Res. 2005, 38, 217.
(4) Kitagawa, S.; Kitaura, R.; Noro, S. Angew. Chem., Int. Ed. 2004, 43,
2334.
(5) Ockwig, N. W.; Delgado-Friedrichs, O.; O’Keeffe, M.; Yaghi, O. M.
Acc. Chem. Res. 2005, 38, 176.
(6) Rowsell, J. L.; Yaghi, O. M. Angew. Chem., Int. Ed. 2005, 44, 4670.
(7) Kesanli, B.; Cui, Y.; Smith, M. R.; Bittner, E. W.; Bockrath, B. C.;
Lin, W. B. Angew. Chem., Int. Ed. 2005, 44, 72.
(8) Dinca, M.; Long, J. R. Angew. Chem., Int. Ed. 2008, 47, 6766.
(9) Zhao, D.; Yuan, D. Q.; Zhou, H. C. Energy EnViron. Sci. 2008, 1,
222.
9
15390 J. AM. CHEM. SOC. 2010, 132, 15390–15398
10.1021/ja1069773  2010 American Chemical Society

Isoreticular Chiral MOFs for Asymmetric Alkene Epoxidation
A R T I C L E S
synthesis provides a unique opportunity for rational design and
development of versatile MOFs as a new class of functional
materials.
structing isoreticular MOFs. These sterically demanding SBUs have
been shown to sustain non-interpenetrated MOF structures, but
the tendency to form interpenetrated networks increases as the
bridging ligands become longer. Such interpenetrations can
severely reduce or even eliminate the interior void space of the
MOFs. A few reports have recently appeared on controlling
framework catenation by varying reaction temperature or reagent
concentration,36,44 templating with oxalic acid,45 introducing
space-filling groups on the ligands,46 and employing solvent
molecules of different sizes as templates.39
In this work, we hypothesize that isoreticular chiral MOFs
(CMOFs) of tunable open channel sizes can be constructed from
[Zn4(µ4-O)(O2CR)6] SBUs and tunable Mn-Salen-derived di-
carboxylic acids. Chiral Salen ligands, such as (R,R)-1,2-
cyclohexanediamino-N,N′-bis(3-tert-butyl-salicylidene), have
been established as one of the few privileged ligand systems
for asymmetric catalysis.47 Mn-Salen complexes have, for
example, been shown to be highly effective in catalyzing
asymmetric epoxidation reactions of unfunctionalized olefins.48-54
Several reports have used Salen-type ligands to construct
MOFs.55-57 For example, Kitagawa et al. synthesized an achiral
Zn-MOF using M-salphdc (M ) Cu(II), Co(II), and Ni(II);
salphdc ) N,N′-phenylenebis(salicylideneimine)dicarboxylic
acid) building blocks,58 whereas Chen et al. studied hydrogen
uptake with mixed metal MOFs constructed from the achiral
Cu(Pyen) structure (Pyen-H2 ) 5-methyl-4-oxo-1,4-dihydro-
pyridine-3-carbaldehyde).59 Mirkin et al. studied the intercon-
version between amorphous and crystalline microparticles that
are built from Ni-salen-dicarboxylic acid and an excess amount
of Ni(OAc)2 ·4H2O.56 Cui et al. carried out chiral recognition
and separation using a 2-D coordination polymer built from
unsymmetrical chiral Schiff base metal complexes.60 Hupp et
al. constructed a chiral MOF based on the Mn-Salen-derived
Because of their mild synthetic conditions, MOFs are
particularly suited for immobilizing well-defined molecular
catalysts, leading to a new generation of solid catalysts with
uniform catalytic sites and open channel structures for shape-,
size-, chemo-, and enantioselective reactions. The ability to
easily recover and reuse such MOF-based heterogeneous
catalysts is also highly desirable for reducing processing and
waste disposal costs in large-scale reactions. Compared to other
immobilized catalytic systems, MOFs can have well-defined,
single-crystalline solid structures, unprecedentedly high catalyst
loadings, more uniform and accessible catalytic centers, and
enhanced catalytic activity by eliminating multimolecular
catalyst deactivation pathways.15,17,18
Although a large number of MOFs have been examined as
heterogeneous catalysts, most of these studies rely on the
intrinsic catalytic activity (e.g., weak Lewis acidity) of the metal-
connecting points.15,17,23-25 A more rational strategy than such
an “opportunistic” approach is to introduce well-defined mo-
lecular catalysts into the MOF structures,15,18,23,25 which can
lead to a new generation of heterogeneous catalysts that are
capable of catalyzing more advanced and value-added reactions,
such as enantioselective reactions. Post-synthesis modification
(PSM) has recently been successfully used to generate highly
active and selective MOF catalysts,26-28 but this method tends
to reduce the open channel sizes during the PSM process and
hence negatively impacts the performance of MOF catalysts by
slowing down substrate and product diffusion through the
channels. Direct incorporation of a well-defined homogeneous
catalyst (or precatalyst) into the framework of a MOF represents
an attractive alternative that has so far been under-explored.29-31
We report here the rational design of a family of isoreticular
chiral MOFs based on Mn-Salen-derived bridging ligands and
the applications of these solid catalysts in highly enantioselective
alkene epoxidation reactions.
(40) Lin, X.; Telepeni, I.; Blake, A. J.; Dailly, A.; Brown, C. M.; Simmons,
J. M.; Zoppi, M.; Walker, G. S.; Thomas, K. M.; Mays, T. J.;
Hubberstey, P.; Champness, N. R.; Schroder, M. J. Am. Chem. Soc.
2009, 131, 2159.
(41) Wu, S.; Ma, L.; Long, L. S.; Zheng, L. S.; Lin, W. Inorg. Chem.
2009, 48, 2436.
Isoreticular MOFs with the same framework topology but
tunable pore and channel sizes have been constructed by linking
metal-coordinated secondary building units (SBUs) with bridging
ligands of varied length.32-43 Among the many metal-coordinated
SBUs, [Zn4(µ4-O)(O2CR)6],36,37 [Cu2(sol)2(O2CR)4],38-41 and
[Zr6(O)4(OH)4(O2CR)12]42,43 are most commonly used for con-
(42) Cavka, J. H.; Jakobsen, S.; Olsbye, U.; Guillou, N.; Lamberti, C.;
Bordiga, S.; Lillerud, K. P. J. Am. Chem. Soc. 2008, 130, 13850.
(43) Guillerm, V.; Gross, S.; Serre, C.; Devic, T.; Bauer, M.; Ferey, G.
Chem. Commun. 2010, 46, 767.
(44) Zhang, J.; Wojtas, L.; Larsen, R. W.; Eddaoudi, M.; Zaworotko, M. J.
J. Am. Chem. Soc. 2009, 131, 17040.
(45) Ma, S.; Sun, D.; Ambrogio, M.; Fillinger, J. A.; Parkin, S.; Zhou,
H. C. J. Am. Chem. Soc. 2007, 129, 1858.
(23) Lin, W. B. J. Solid State Chem. 2005, 178, 2486.
(24) Wang, Z.; Chen, G.; Ding, K. L. Chem. ReV. 2009, 109, 322.
(25) Lin, W. B. MRS Bull. 2007, 32, 544.
(46) Farha, O. K.; Malliakas, C. D.; Kanatzidis, M. G.; Hupp, J. T. J. Am.
Chem. Soc. 2010, 132, 950.
(47) Yoon, T. P.; Jacobsen, E. N. Science 2003, 299, 1691.
(48) Palucki, M.; Finney, N. S.; Pospisil, P. J.; Guler, M. L.; Ishida, T.;
Jacobsen, E. N. J. Am. Chem. Soc. 1998, 120, 948.
(49) Zhang, W.; Loebach, J. L.; Wilson, S. R.; Jacobsen, E. N. J. Am.
Chem. Soc. 1990, 112, 2801.
(26) Wu, C. D.; Hu, A.; Zhang, L.; Lin, W. J. Am. Chem. Soc. 2005, 127,
8940.
(27) Banerjee, M.; Das, S.; Yoon, M.; Choi, H. J.; Hyun, M. H.; Park,
S. M.; Seo, G.; Kim, K. J. Am. Chem. Soc. 2009, 131, 7524.
(28) Tanabe, K. K.; Cohen, S. M. Angew. Chem., Int. Ed. 2009, 48, 7424.
(29) Hu, A.; Ngo, H. L.; Lin, W. J. Am. Chem. Soc. 2003, 125, 11490.
(30) Hu, A.; Ngo, H. L.; Lin, W. Angew. Chem., Int. Ed. 2003, 42, 6000.
(31) Cho, S. H.; Ma, B.; Nguyen, S. T.; Hupp, J. T.; Albrecht-Schmitt,
T. E. Chem. Commun. 2006, 2563.
(50) Katsuki, T. Synlett 2003, 281–297.
(51) Venkataramanan, N. S.; Kuppuraj, G.; Rajagopal, S. Coord. Chem.
ReV. 2005, 249, 1249.
(52) Baleizao, C.; Garcia, H. Chem. ReV. 2006, 106, 3987–4043.
(53) Canali, L.; Sherrington, D. C. Chem. Soc. ReV. 1999, 28, 85.
(54) Gupta, K. C.; Sutar, A. K.; Lin, C. C. Coord. Chem. ReV. 2009, 253,
1926.
(32) Lin, W. B.; Evans, O. R.; Xiong, R. G.; Wang, Z. Y. J. Am. Chem.
Soc. 1998, 120, 13272.
(33) Evans, O. R.; Xiong, R. G.; Wang, Z. Y.; Wong, G. K.; Lin, W. B.
Angew. Chem., Int. Ed. 1999, 38, 536.
(55) Oh, M.; Mirkin, C. A. Angew. Chem., Int. Ed. 2006, 45, 5492.
(56) Jeon, Y. M.; Heo, J.; Mirkin, C. A. J. Am. Chem. Soc. 2007, 129,
7480.
(34) Evans, O. R.; Lin, W. B. Chem. Mater. 2001, 13, 3009.
(35) Evans, O. R.; Lin, W. B. Chem. Mater. 2001, 13, 2705.
(36) Eddaoudi, M.; Kim, J.; Rosi, N.; Vodak, D.; Wachter, J.; O’Keeffe,
M.; Yaghi, O. M. Science 2002, 295, 469.
(57) Jung, S.; Oh, M. Angew. Chem., Int. Ed. 2008, 47, 2049.
(58) Kitaura, R.; Onoyama, G.; Sakamoto, H.; Matsuda, R.; Noro, S.;
Kitagawa, S. Angew. Chem., Int. Ed. 2004, 43, 2684.
(59) Chen, B.; Zhao, X.; Putkham, A.; Hong, K.; Lobkovsky, E. B.;
Hurtado, E. J.; Fletcher, A. J.; Thomas, K. M. J. Am. Chem. Soc.
2008, 130, 6411.
(37) Li, Q.; Zhang, W.; Miljanic, O. S.; Sue, C. H.; Zhao, Y. L.; Liu, L.;
Knobler, C. B.; Stoddart, J. F.; Yaghi, O. M. Science 2009, 325, 855.
(38) Ma, L.; Lee, J. Y.; Li, J.; Lin, W. Inorg. Chem. 2008, 47, 3955.
(39) Ma, L.; Lin, W. J. Am. Chem. Soc. 2008, 130, 13834.
(60) Yuan, G.; Zhu, C.; Xuan, W.; Cui, Y. Chemistry 2009, 15, 6428.
9
J. AM. CHEM. SOC. VOL. 132, NO. 43, 2010 15391

A R T I C L E S
Song et al.
Scheme 1. Synthesis of Ligands L1-H2-L3-H2
bipyridine bridging ligand and demonstrated its asymmetric
catalytic activity in alkene epoxidation reactions, albeit at lower
enantioselectivity than the homogeneous counterpart.31 How-
ever, no systematic study on constructing isoreticular Mn-Salen-
based MOFs with tunable open channels has appeared in the
literature to date.
tive alkene epoxidation catalytic activity on framework catena-
tion and open channel sizes. Our work provides solid evidence
that the catalytic activity of the CMOFs with large open channels
can rival that of an analogous homogeneous catalyst and is
limited by the intrinsic reactivity of the catalytic molecular
building blocks.
Systematic tuning of highly porous MOFs for gas storage
applications elucidated the relationships between channel open-
ing sizes/framework interpenetrations and gas-uptake capacities
of MOFs.7,45,59,61,62 On the other hand, there is only one study
on the dependence of MOF catalytic properties on the open
channel sizes.63 In this paper, we not only develop a family of
isoreticular CMOFs based on [Zn4(µ4-O)(O2CR)6] SBUs and
tunable Mn-Salen-derived dicarboxylic acids but also demon-
strate the control over framework catenation using solvents of
different steric bulk and the dependence of MOF enantioselec-
Results and Discussion
Synthesis and Characterization. Enantiopure Mn-Salen di-
carboxylic acid ligands, L1-H2 to L3-H2, were synthesized by
Schiff base condensation reactions between (R,R)-cyclohex-
anediamine and corresponding 2-hydroxybenzaldehyde deriva-
tives with pendant carboxylic acid groups (L1-al to L3-al),
followed by metalation with Mn(OAc)2 ·4H2O and in situ
oxidation in air to afford Mn(III) complexes (Scheme 1). The
2-hydroxybenzaldehyde derivatives L1-al and L3-al are known
compounds, whereas L2-al was synthesized by a Pd-catalyzed
carbon-carbon coupling reaction between the bromo compound
and methyl acrylate followed by base-promoted hydrolysis.
More detailed synthetic schemes and procedures are shown in
(61) Ryan, P.; Broadbelt, L. J.; Snurr, R. Q. Chem. Commun. 2008, 4132.
(62) Wang, Z.; Tanabe, K. K.; Cohen, S. M. Chemistry 2010, 16, 212.
(63) Ma, L.; Falkowski, J. M.; Abney, C.; Lin, W. Nat. Chem. 2010, 2,
838–846.
9
15392 J. AM. CHEM. SOC. VOL. 132, NO. 43, 2010

Isoreticular Chiral MOFs for Asymmetric Alkene Epoxidation
A R T I C L E S
Scheme 2. Synthesis of CMOFs 1-5
the Supporting Information. All of the new compounds were
characterized by NMR spectroscopy and mass spectrometry.
We hypothesized that Mn-Salen-derived dicarboxylate ligands
L1-L3 can link octahedral [Zn4(µ4-O)(O2CR)6] SBUs to form
isoreticular CMOFs of the primitive cubic network (pcu) to-
pology with tunable pore/channel sizes and shapes. As shown
in Scheme 2, reactions of Zn(NO3)2 ·6H2O with the Mn-Salen
complex L1-H2 in dimethylformamide (DMF)/EtOH at 80-90
°C for 96 h afforded dark brown single crystals of [Zn4(µ4-
O)(L1)3]·20DMF·2H2O (CMOF-1). CMOF-1 crystallizes in the
trigonal R32 space group, as revealed by single-crystal X-ray
crystallography (Figures 1 and 2). The asymmetric unit of the
CMOF-1 framework contains two L1 ligands, two-thirds of
Zn4(µ4-O) clusters which are composed of two Zn atoms of full
occupancy and two Zn atoms of one-third occupancy, and two
O atoms of one-third occupancy. As expected, the carboxylate
groups from six adjacent L1 ligands coordinate to the four Zn
centers to form [Zn4(µ4-O)(carboxylate)6] SBUs, which link L1
ligands to form a 3D network of the pcu topology. Because of
the elongated L1 ligands, CMOF-1 adopts a two-fold interpen-
etrated structure with 61.1% of void space as calculated by
PLATON that is filled by the DMF and water molecules.64 The
precise solvent content cannot be determined by X-ray crystal-
lography owing to their disordered nature. The solvent contents
were instead established by a combination of 1H NMR studies
and thermogravimetric analysis (Supporting Information). The
largest dimensions of open channels measure 0.8 × 0.6 nm2,
and the largest cavities inside the interpenetrated networks
measure 1.4 nm in diameter in CMOF-1 (Figure 3).
which act as space-filling templates. Diethylformamide (DEF)
is a larger molecule than DMF and thus can favor the formation
of non-interpenetrated structure as a result of the solvent tem-
plating effects. When the reaction between Zn(NO3)2 ·6H2O and
L1-H2 was carried out in DEF/EtOH at 60-90 °C, the non-
interpenetrated 3D CMOF-2 was obtained, with the formula
determined to be [Zn4(µ4-O)(L1)3]·22DEF·4H2O (Scheme 2).
Single-crystal X-ray crystallography showed that CMOF-2 has
cell parameters very similar to those of CMOF-1 but crystallizes
in the space group R3. The framework of CMOF-2 is exactly
identical to one of the two interpenetrating nets in CMOF-1,
with the same metal-ligand connectivity and network topology
(Figure 2). The non-interpenetrated framework enjoys much
larger open channel dimensions of 2.0 × 1.6 nm2 and a cavity
size of 2.6 nm (Figure 3), with 80.2% of void space as calculated
by PLATON. This huge enhancement of accessible internal
volume and open channel size in non-interpenetrated structures
is crucial for enhanced catalytic performance by facilitating the
diffusion of reactants and products (see below).
Similar reactions of Zn(NO3)2 ·6H2O with L2-H2 in DMF/
EtOH and DEF/EtOH resulted in single crystals of [Zn4(µ4-
O)(L2)3] · 42DMF (CMOF-3) and [Zn4(µ4-O)(L2)3]·37DEF·
23EtOH·4H2O (CMOF-4), respectively (Scheme 2). Just as in
the CMOF-1/2 pair, CMOF-3 and CMOF-4 adopt two-fold
interpenetrated and non-interpenetrated structure, respectively
(Figure 2). The two-fold interpenetrated CMOF-3 possesses
channel sizes of 1.5 × 0.7 nm2 and an open cavity of 2.0 nm in
diameter, with 76.8% of void space, while the non-interpen-
etrated CMOF-4 exhibits the largest channel dimensions of 2.5
× 2.3 nm2 and a cavity diameter of 3.2 nm (Figure 3), reaching
88.4% of void space as calculated by PLATON. The synthesis
of CMOF-1 vs -2 or CMOF-3 vs -4 is dependent on the steric
bulk of the templating solvents but does not appear to be
sensitive to the reaction/crystallization temperatures. For ex-
ample, CMOF-4 was obtained in pure phase in the 60-90 °C
temperature range, but as expected, longer reaction time was
needed for the reactions carried out at lower temperatures.
As we have shown previously,39 catenation isomerism can
be controlled by introducing solvent molecules of different sizes
As ligand L3 gets even longer, the reactions between L3-H2
and Zn(NO3)2 in both DMF/EtOH and DEF/EtOH yield the
same interpenetrated framework, CMOF-5 (Scheme 2). Similar
reactions in even more sterically demanding solvents such as
N,N′-dibutylformamide and N,N′-diisopropylformamide also
failed to produce non-interpenetrated CMOF, implying a delicate
balance between framework catenation and solvent templating
effects. CMOF-5 grown in DMF at 80 °C has the formula of
[Zn4(µ4-O)(L3)3]·38DMF ·11EtOH. Unlike CMOF-1 or CMOF-
3, CMOF-5 adopts a three-fold interpenetrated framework that
crystallizes in the R3 space group (Figure 2). Detailed analysis
of the single-crystal X-ray diffraction data revealed that two of
the three interpenetrating nets are fully occupied and pack
closely to each other, while the third fold randomly occupies
Figure 1. Ball-and-stick and polyhedra presentations of Mn-Salen-derived
dicarboxylate bridging ligands L1 (a), L2 (b), and L3 (c) and the
[Zn4O(carboxylate)6] SBUs in CMOFs 1-5.
9
J. AM. CHEM. SOC. VOL. 132, NO. 43, 2010 15393

A R T I C L E S
Song et al.
Figure 2. Stick/polyhedra models showing the connectivity of Mn-Salen-derived dicarboxylate bridging ligands (L1-L3) and [Zn4O(carboxylate)6] SBUs
and the interpenetration of the pcu networks and space-filling models as viewed perpendicular to the (1 0 -2) plane in (a) CMOF-2 (one-fold), (b) CMOF-4
(one-fold), (c) CMOF-1 (two-fold), (d) CMOF-3 (two-fold), and (e) CMOF-5 (three-fold).
three possible positions with a one-third occupancy (Supporting
Information). The assignment of three-fold interpenetration is
fully supported by the H NMR and TGA solvent content
match between the experimental and simulated powder X-ray
diffraction (PXRD) patterns.
1
Although interpenetration is frequently encountered in the
literature, few methods other than time-consuming single-crystal
X-ray crystallography can tell the difference between interpen-
etrated and non-interpenetrated structures. PXRD patterns of
interpenetrated and non-interpenetrated structures often do not
differ much, and cannot be used as a reliable indicator for
framework catenation (see below). The density of MOF crystals
measurements in combination with the void space predicted
from the structural model. In a simplified model of three closely
packed and fully occupied interpenetrating nets, CMOF-5
possesses open channel (1.1 × 0.8 nm2) and cavity (1.8 nm)
dimensions that are intermediate between those of CMOF-1 and
CMOF-3 (Figure 3). Phase purity of the bulky crystalline
samples of all the five CMOFs was confirmed by an excellent
9
15394 J. AM. CHEM. SOC. VOL. 132, NO. 43, 2010

Isoreticular Chiral MOFs for Asymmetric Alkene Epoxidation
A R T I C L E S
Figure 3. Different cavity sizes shown as red balls in CMOFs 1-5: (a) CMOF-1 (1.4 nm); (b) CMOF-3 (2.0 nm); (c) CMOF-5 (1.8 nm); (d) CMOF-2 (2.6
nm); (e) CMOF-4 (3.2 nm).
Figure 4. Powder X-ray diffraction patterns for (a) CMOF-1 simulated from CIF file (black), CMOF-1 as synthesized (red), CMOF-2 simulated from CIF
file (green), CMOF-2 as synthesized (blue); (b) CMOF-3 simulated from CIF file (black), CMOF-3 as synthesized (red), CMOF-4 simulated from CIF file
(green), CMOF-4 as synthesized (blue).
can be a potentially useful indicator for framework catenation,65
but when the channel sizes of MOFs are very large, as in this
CMOF series, solvent molecules used for density measurements
can exchange with the native solvent molecules to alter the
crystal density and to render this technique useless. The
determination of included solvent molecules inside MOF
channels, although not very precise, provides a simple and
reliable way to identify different interpenetration patterns in the
bulk samples and to provide supporting evidence to the single-
crystal X-ray diffraction results.
From single-crystal structures, the relative positions of the
two interpenetrating nets in CMOF-1 and CMOF-3 are different.
In CMOF-1, each net sits right in the middle of the other,
whereas in CMOF-3, the second fold of the nets resides off the
center of the first one (Figure 2). The relative positions of the
two nets in the interpenetrated structures of CMOF-1 and
CMOF-3 are supported by the PXRD studies. Comparing the
PXRD patterns of CMOF-1 and CMOF-2, the first peak around
2θ ) 4.3° in the powder pattern of CMOF-2 (corresponding to
the Miller indices 1 0 -2) is absent in that of CMOF-1 (Figure
4a). The absence of the 1 0 -2 peak can occur only when the
second fold of the nets sits right in the middle of the first fold,
which generates a pseudo 43 screw axis for the [Zn4(µ4-
O)(carboxylate)6] SBUs and hence introduces additional sys-
tematic absence. In contrast, comparing the PXRD patterns of
CMOF-3 and CMOF-4, no additional absence is observed for
CMOF-3, while peak intensities do change with respect to that
of CMOF-4 (Figure 4b). It is important to note that the relative
positions of the two interpenetrating nets also significantly
modify the maximum open channel sizes and accessible cavity
volumes in the structures.
High porosity, which is needed to efficiently transport reagent
and product molecules to facilitate asymmetric catalysis, was
expected for all the CMOFs 1-5 on the basis of the crystal
structures. However, nitrogen adsorption measurements indi-
cated that the CMOFs exhibit negligible surface areas. The
discrepancy between expected and experimental surface areas
(64) Spek, A. L. J. Appl. Crystallogr. 2003, 36, 7.
(65) Farha, O. K.; Mulfort, K. L.; Thorsness, A. M.; Hupp, J. T. J. Am.
Chem. Soc. 2008, 130, 8598.
9
J. AM. CHEM. SOC. VOL. 132, NO. 43, 2010 15395

A R T I C L E S
Song et al.
Table 1. Key Crystallographic and Dye Uptake Data for CMOFs 1-5
CMOF-1
CMOF-2
CMOF-3
CMOF-4
CMOF-5
framework formula
space group
[Zn4O(L1)3]
R32
[Zn4O(L1)3]
R3
[Zn4O(L2)3]
R32
[Zn4O(L2)3]
R3
[Zn4O(L3)3]
R3
cell dimensions (Å)
interpenetration
29.7 × 29.7 × 72.7 29.6 × 29.6 × 72.5 35.9 × 35.9 × 87.8 35.9 × 35.9 × 87.9
41.9 × 41.9 × 100.3
two-fold
61.1
one-fold
80.2
two-fold
76.8
one-fold
88.4
three-fold
75.9
void space % calcd by PLATON
framework density (g/cm3)
largest channel sizes (nm)
largest cavity (nm)
0.758
0.382
0.462
1.5 × 0.7
2.0
0.230
0.502
0.8 × 0.6
2.0 × 1.6
2.5 × 2.3
1.1 × 0.8
1.4
2.6
3.2
1.8
solvent contenta
20DMF·2H2O
41.3
22DEF ·4H2O
53.2
15.4
42DMF
55.5
6.3
37DEF ·23EtOH·4H2O 38DMF·11EtOH
solvent weight loss (%)b
68.1
37.7
152.1
53.0
16.3
125.0
dye uptake (wt % of the framework) 2.1
effective dye concentration
inside MOF channels (mM)
31.1
90.0
45.6
a Solvent contents were determined by a combination of 1H NMR spectroscopy and TGA. b The solvent weight loss happened in the 25-250 °C
temperature range.
probably results from the framework distortion upon solvent
removal, a phenomenon commonly observed for MOFs with
large open channels.66-69 This was confirmed by PXRD studies
of the evacuated sample of CMOF-2, showing the loss of
ordered framework structure after removing the solvents. On
the other hand, the test for the accessibility of MOF channels
to solvent or other large molecules is more relevant to screening
heterogeneous asymmetric catalysts, as enantioselective organic
transformations are typically carried out in solution. A dye
uptake assay recently developed in our laboratory63 was thus
employed to quantify the intrinsic porosity of the CMOFs as
well as the capability of the channels in transporting large
molecules. By soaking the CMOFs in a solution of 24.2 mM
Brilliant Blue R-250 (BBR-250) dye in methanol for 16 h,
significant fractions of the BBR-250 can be absorbed into the
internal channels of the CMOFs. The dye solution was then
decanted, and the CMOFs were washed with water several times
to remove dye molecules adsorbed on the external surfaces of
the crystals. The dye-loaded CMOFs were then digested with
disodium ethylenediaminetetraacetate (Na2EDTA), and the
amounts of released BBR-250 were quantified by ultraviolet-
visible spectroscopy. As shown in Table 1, the CMOF series
exhibit dye uptake capacity (ranging from 2.1 to 37.7 wt % of
the framework) dependent upon the open channels, correspond-
ing to an effective dye concentration of 31.1-152.1 mM in the
channels of the CMOFs, which is equivalent to 1.3-6.3 times
the original dye concentration in the MeOH solution. These
results unambiguously prove the accessibility of the open
channels of the CMOFs to large reagents typically used for
enantioselective catalysis, as the BBR-250 molecule has an
estimated cross section of 1.8 × 2.2 nm2. Moreover, the drastic
difference in dye uptake capacities between interpenetrated and
non-interpenetrated CMOFs (2.1 wt % for 1 vs 15.4 wt % for 2
and 6.3 wt % for 3 vs 37.7 wt % for 4) provides another important
means for identifying framework catenation.
competent bridging ligands.63 With the ease of changing the
length of two-connected linear dicarboxylate bridging ligands,
the channel and cavity sizes of the resulting MOFs can be
systematically tuned. Controlling interpenetrations with the help
of templating solvents provides an additional means to alter the
open channel sizes. The successful incorporation of molecular
Mn-Salen catalytic building blocks into the 3D frameworks and
the tunable open channel and pore sizes of this CMOF series
open up a great opportunity toward delineating their structure/
catalytic activity relationships in enantioselective alkene epoxi-
dation reactions.
Enantioselective Catalysis. Catalytic activities of CMOFs 1-5
toward asymmetric epoxidation of unfunctionalized olefins were
evaluated with 1H-indene (6a) as substrate and 2-(tert-butyl-
sulfonyl)iodosylbenzene (7) as oxidant. As shown in Table 2
(entries 1-3), CMOFs 1, 3, and 5 are highly effective catalysts
for asymmetric epoxidation of 6a to generate (1R,2S)-indene
oxide in 47-64% ee. In comparison, the homogeneous control,
L3-Me2, gave (1R,2S)-indene oxide in 64% ee. The conversions
of 6a range from 54% to 80% for CMOFs 1, 3, and 5, which
compare favorably with that of the homogeneous control L3-
Me2 (60%, Table 1, entry 4). CMOF-4 and CMOF-5 have also
been shown to be highly effective in asymmetric epoxidation
of a variety of unfucntionalized alkenes to afford chiral epoxides
in good to excellent yields and ee’s (Table 2), demonstrating
generality of the asymmetric epoxidation activities of CMOF
catalysts. The level of enantioselectivity observed for CMOF-
catalyzed alkene epoxidation is comparable to that of the
homogeneous control catalyst L3-Me2. Asymmetric epoxidation
of 2,2-dimethyl-2H-chromene (6b) by CMOF-5, for example,
gave the (R,R)-chromene oxide in 92% ee, a value that is the
same as the homogeneous result for L3-Me2 and higher than
the ee value (82%) reported earlier for a chiral MOF built from
(R,R)-1,2-cyclohexanediamino-N,N′-bis(3-tert-butyl-5-(4-py-
ridyl)salicylidene)MnIIICl.31
Isoreticular MOFs based on [Zn4(µ4-O)(O2CR)6] SBUs
represent the most well-established family of isoreticular porous
materials. The CMOF series represents the first isoreticular
MOFs that are built from direct incorporation of catalytically
The isoreticular nature of CMOFs 1-5 provides an interesting
platform for probing the dependence of epoxidation reaction
rates on the open channel dimensions. As shown in Figure 5,
the time-dependent conversion curves for the epoxidation of
6b greatly depend on the channel sizes. The conversion rate
increases in the order of CMOF-1 < CMOF-5 < CMOF-3 <
CMOF-2 < CMOF-4, consistent with the increase of open
channel sizes in this series (Table 1). This trend suggests that
the diffusion of the bulky alkene and oxidant reagents and the
epoxide product can be a significant factor in determining the
reaction rates for CMOF-catalyzed asymmetric epoxidation
(66) Serre, C.; Millange, F.; Thouvenot, C.; Nogues, M.; Marsolier, G.;
Louer, D.; Ferey, G. J. Am. Chem. Soc. 2002, 124, 13519.
(67) Ghoufi, A.; Maurin, G. J. Phys. Chem. C. 2010, 114, 6496.
(68) Llewellyn, P. L.; Maurin, G.; Devic, T.; Loera-Serna, S.; Rosenbach,
N.; Serre, C.; Bourrelly, S.; Horcajada, P.; Filinchuk, Y.; Ferey, G.
J. Am. Chem. Soc. 2008, 130, 12808.
(69) Neimark, A. V.; Coudert, F. X.; Boutin, A.; Fuchs, A. H. J. Phys.
Chem. Lett. 2010, 1, 445.
9
15396 J. AM. CHEM. SOC. VOL. 132, NO. 43, 2010

Isoreticular Chiral MOFs for Asymmetric Alkene Epoxidation
A R T I C L E S
Table 2. CMOF-Catalyzed Enantioselective Epoxidation of
product diffusion through open channels is not the rate-limiting
factor. With open channel sizes of 2.0 × 1.6 nm2 and 2.5 ×
2.3 nm2 in CMOFs 2 and 4, respectively, the epoxidation rate
appears to be limited by the intrinsic activity of the Mn-Salen
catalytic units. The reactivity studies in the present isoreticular
CMOF series thus make a convincing case that CMOFs can be
effective heterogeneous asymmetric catalysts for a number of
important organic transformations. The open channel size-
dependent conversions observed for CMOFs 1-5 also provide
indirect evidence for the reagents accessing active catalytic sites
in the interiors of CMOF crystals and highlight the opportunity
for size- and enantioselective catalysis with CMOFs.
Alkenesa
It is difficult to obtain quantitative kinetic parameters from
the present study since the reaction rate depends on the MOF
crystal sizes as well as the stirring rate. Since most MOFs,
including the present CMOF series, are mechanically unstable,
they tend to fracture into smaller particles by the magnetic stir
bar. The contribution from the surface catalytic sites can be a
significant factor, which can lead to an underestimate of the
dependence of the conversion on the CMOF open channel size.
To eliminate the possibility of the catalytic activity all arising
from surface catalytic sites of very small particles, epoxidation
of 6d was carried out with large crystals of CMOF-4 (∼0.4 ×
0.4 × 0.4 mm3 on average, 0.1 mol % catalyst loading) without
stirring or shaking. A 35% conversion was obtained after 1 h,
and the epoxide product exhibited 84% ee, clearly demonstrating
the ability for the substrate to enter the interior of the MOF
crystals, since such large crystals possess too few surface
catalytic sites to account for the reaction. In comparison,
epoxidation of 6d with CMOF-1 crystals (∼0.3 × 0.3 × 0.3
mm3 on average, 0.1 mol % catalyst loading) under the same
conditions gave 10% conversion at 78% ee, consistent with the
less accessible catalytic sites inside smaller open channels of
CMOF-1. These results provide unambiguous evidence for the
alkenes and oxidants to access the catalytic Mn-Salen sites in
the interiors of these MOF crystals.
We have also carried out a number of control experiments
to demonstrate the heterogeneity of CMOF catalysts. First, a
supernatant of CMOF-4 after sonication was used to catalyze
the epoxidation of 6d with 1 equiv of the oxidant, but no epoxide
was observed. Second, we tested the reactivity of the supernatant
after epoxidation reaction in a crossover experiment. Substrate
6e was epoxidized in the presence of the CMOF-4 catalyst with
gentle shaking (instead of stirring) in 1 h to afford the
epoxidation product in 57% conversion and 74% ee. The solid
was removed from the reaction mixture by filtering through
Celite, and a second substrate, 6d, and 2 equiv of the oxidant
were then added to the supernatant. After 3 h, no epoxidation
was detected by GC. This lack of epoxidation in the crossover
experiment strongly supports the heterogeneous nature of the
CMOF catalysis.
a 0.5 equiv of oxidants added in 10 portions at 15 min intervals, and
the reaction continued for another 0.5 h after the oxidant addition.
b Determined by GC. c Determined by chiral GC with a Supelco ꢀ-DEX
120 capillary column or by chiral HPLC with a Chiralcel AD column.
We have also examined the recyclability of CMOF-5 and
CMOF-4 for the asymmetric epoxidation of 6b and 6c, res-
pectively. As shown in Table 3, CMOF-4/CMOF-5 was readily
recovered from the catalytic reaction via centrifugation, and the
recovered catalyst showed only slight deterioration in conversion
and enantioselectivity. Furthermore, the solid catalyst recovered
from the catalytic epoxidation reaction catalyzed by CMOF-5
exhibited the same PXRD as the pristine solid of CMOF-5
(Figure 6), unambiguously supporting the stability of the CMOF
framework during the catalytic reactions. Inductively coupled
plasma mass spectrometric (ICP-MS) analysis of the supernatant
showed that less than 7.5% of the manganese had leached into
Figure 5. Plots of percent conversions versus time for epoxidation of 2,2-
dimethyl-2H-chromene catalyzed by CMOF-1 (black), CMOF-2 (red),
CMOF-3 (green), CMOF-4 (blue), CMOF-5 (cyan), and L3-Me2 (purple).
The epoxidation reactions were carried out with 3 equiv of 2-(tert-
butylsulfonyl)iodosylbenzene in the presence of 0.1 mol % catalyst.
reactions when the CMOF channels are small (as in interpen-
etrated CMOFs 1, 3, and 5). On the other hand, CMOF-2 and
CMOF-4 gave conversion rates comparable to that of the ho-
mogeneous catalyst L3-Me2, suggesting that the substrate and
9
J. AM. CHEM. SOC. VOL. 132, NO. 43, 2010 15397

A R T I C L E S
Song et al.
Table 3. Recycling of CMOF-4 and CMOF-5 for the Asymmetric
this CMOF series were systematically tuned by introducing
spacers of different lengths, as well as by successful control
over framework catenation. These isoreticular chiral MOFs were
demonstrated to be highly active in catalyzing enantioselective
alkene epoxidation reactions. The isoreticular nature of this
CMOF series also allows probing the dependence of epoxidation
reaction rates on the open channel dimensions. The rates of
epoxidation reactions strongly depend on the CMOF open
channel sizes, providing solid evidence that larger open channel
dimensions can increase reaction rate by facilitating the diffusion
of reactant and product molecules. The catalytic activities of
CMOFs 2 and 4 rival that of a homogeneous control catalyst,
suggesting the catalytic activity of the CMOFs with large open
channels is limited by the intrinsic reactivity of the catalytic
molecular building blocks. The CMOF catalysts are recyclable
and reusable and retain their framework structures after epoxi-
dation reactions. The modular nature of this synthetic approach
should allow the design of other metal-Salen-based MOF
catalysts for many important asymmetric organic transformations.
Epoxidation
run
conversion, %b
ee, %c
CMOF-4
CMOF-5
1
2
3
4
1
2
3
78
80
79
72
72
70
62
76
72
71
69
86
81
80
a The reactions were carried out with 0.5 mol % catalyst and 0.5
equiv of oxidants. b Determined by GC. c Determined by chiral capillary
column supelco ꢀ-DEX 120.
Experimental Section
General experimental section and detailed procedures for ligand
and CMOF synthesis and characterization are provided in the
Supporting Information. Only key representative experimental
procedures are shown below.
A typical procedure for crystal growth: Zn4O(L2)3 ·(DEF)37 ·
(EtOH)23 ·(H2O)4 (CMOF-4): L2-H2 (10 mg, 0.015 mmol) and
Zn(NO3)2 ·6H2O (10 mg, 0.034 mmol) were added to a 2-dram,
screw-capped vial. DMF (3 mL) and EtOH (2 mL) were then added.
The vial was capped and placed into an oven at 60 °C for 12 days.
Black cubic crystals were obtained after filtration (19.1 mg, 54%
yield). Solvent content calculated from proposed formula: DEF,
52.5%; EtOH, 14.8%; H2O, 1.0%. Determined by 1H NMR/TGA:
DEF, 52.3%; EtOH, 14.9%; H2O, 0.9%.
General procedure for the enantioselective epoxidation reaction
of alkene: CMOF-5 (1.08 mg, 0.0006 mmol) was charged into a
2-dram, screw-capped vial, washed with CH2Cl2 three times, and
sonicated for 10 min. To this vial were added alkene (0.6 mmol),
undecane (0.063 mL, 0.3 mmol), and CH2Cl2 (0.5 mL). 2-(tert-
Butylsulfonyl)iodosylbenzene (0.010 g, 0.03 mmol) was then added.
The same amount of oxidant was added nine more times at 15 min
intervals. Aliquots of the reaction solution were taken (10 µL),
diluted with EtOAc and filtered through a syringe filter. The filtrate
was analyzed by GC to give the conversion and enantioselectivity.
Slightly modified procedures are used for time-dependent con-
version, heterogeneity test, crossover catalysis, CMOF recycle and
reuse, catalyst framework stability, and catalyst leaching studies.
All of these details are provided in the Supporting Information.
Figure 6. PXRD patterns of CMOF-5: simulated from CIF file (black),
pristine CMOF-5 (red), and CMOF-5 recovered from the epoxidation
reaction (blue).
the reaction mixture when the epoxidation was carried out in
the presence of 0.1 mol % CMOF-5 catalyst.
MOFs based on the [Zn4O(carboxylate)6] SBUs have been
shown to be hydrolytically unstable.70 This intrinsic instability
sets the limit of the applications of the present CMOFs in
enantioselective alkene epoxidation reactions. More commonly
used oxidants m-CPBA and bleach were attempted for the
epoxidation reactions, but the CMOFs were found to be unstable
under the reaction conditions required for these oxidants. The
discussion of the stability of a certain catalyst has to be specific
to reaction conditions. We have demonstrated that CMOFs 1-5
function as efficient heterogeneous asymmetric alkene epoxi-
dation catalysts with 2-(tert-butylsulfonyl)iodosylbenzene as
oxidant.
Acknowledgment. We thank NSF (CHE-0809776) for financial
support and Ms. Kathryn deKrafft for experimental help.
Supporting Information Available: General experimental
section; procedures for ligand synthesis; synthesis and charac-
terization of isoreticular CMOFs; single-crystal X-ray structure
determinations; representative procedure for dye uptake mea-
surements; CMOF-catalyzed enantioselective epoxidation reac-
tions of unactivated alkenes. This material is available free of
charge via the Internet at http://pubs.acs.org.
Conclusion
A family of isoreticular chiral MOFs was rationally designed
and synthesized by direct incorporation of well-defined Mn-
Salen catalysts into the frameworks. The open channel sizes in
(70) Low, J. J.; Benin, A. I.; Jakubczak, P.; Abrahamian, J. F.; Faheem,
S. A.; Willis, R. R. J. Am. Chem. Soc. 2009, 131, 15834.
JA1069773
9
15398 J. AM. CHEM. SOC. VOL. 132, NO. 43, 2010