Heterogeneous asymmetric catalysis with homochiral metal-organic frameworks: Network-structure-dependent catalytic activity

Article abstract of DOI:10.1002/anie.200602099

It's all in the preparation: Heterogeneous asymmetric catalysts based on two homochiral highly porous metal-organic frameworks formed from the same building blocks show remarkably different activity. Although both catalysts are based on twofold interpenetrated structures, one catalyzes the addition of diethylzinc to aromatic aldehydes with high enantioselectivity (up to 90% ee), while the other is inactive. (Chemical Equation Presented).

Full text of DOI:10.1002/anie.200602099

Angewandte
Chemie
DOI: 10.1002/anie.200602099
Heterogeneous Asymmetric Catalysis
Heterogeneous Asymmetric Catalysis with Homochiral Metal–Organic
Frameworks: Network-Structure-Dependent Catalytic Activity**
Chuan-De Wu and Wenbin Lin*
The prospect of generating functional solids by rational
design has spurred a recent explosive growth of interest
in research into metal–organic frameworks (MOFs).[1,2]
A particular focus in this research area has been placed
upon porous MOFs because of their intrinsic zeolite-
mimicking, size- and shape-selective properties which
make them potentially useful in a number of applica-
tions such as catalysis,[3] gas storage,[4] separation,[5] and
ion-exchange processes.[6] By taking advantage of the
mild conditions that are typically used for the synthesis
of MOFs, we and others have successfully synthesized
catalytically active homochiral porous MOFs by incor-
porating chiral constituent building blocks that con-
tained orthogonal functionalities.[7] The single-crystal-
Scheme 1. Synthesis of chiral MOFs 1 and 2.
À
À
line nature of MOFs allows detailed structural interrogation,
which makes it possible to examine the relationship between
the structure and the catalytic activity of homochiral porous
MOFs in great detail. A detailed understanding of the
structure–activity relationship is key to further fine-tuning
the catalytic performance of MOFs by judicious choice of
building blocks. Herein we report two homochiral porous
MOFs built from the same chiral bridging ligand and metal
connecting point. Their drastically different catalytic activ-
ities are readily rationalized on the basis of their different
framework structures.
atom (Cd O = 2.299 Š), and two pyridyl groups (Cd N =
2.286–2.330 Š) of two L ligands that are cis to each other.
The L ligands thus link the first type of CdII centers into a 2D
square grid of dimensions 20.305 ” 20.305 Š that is lying in
2
the ab plane (Figure 1A), and link the second type of CdII
centers to form 1D zigzag polymeric chains that run along the
110 or 1À10 direction (Figure 1B). The 2D grids and 1D
zigzag polymeric chains are joined to each other by the
bridging nitrate groups to form a 3D framework structure
Colorless crystals of [Cd3L4(NO3)6]·7MeOH·5H2O (1)
were synthesized by slow diffusion of diethyl ether into a
mixture of Cd(NO3)2·4H2O and (R)-6,6’-dichloro-2,2’-
dihydroxy-1,1’-binaphthyl-4,4’-bipyridine (L)[8] in DMF/
CHCl3/MeOH at RT (Scheme 1). Compound 1 crystallizes
in the tetragonal P4122 space group with 1.5 CdII centers in the
asymmetric unit.[9] The first (half-occupied) CdII center is
À
octahedrally coordinated to four L ligands (Cd N = 2.310–
2.317 Š) in the equatorial positions and to two nitrate oxygen
À
atoms (Cd O = 2.356 Š) in the axial positions, while the
second CdII center is heptacoordinated to two bidentate
À
nitrate groups (Cd O = 2.317–2.552 Š), one nitrate oxygen
[*] Dr. C.-D. Wu, Prof. W. Lin
Department of Chemistry, CB no. 3290
University of North Carolina
Chapel Hill, NC 27599 (USA)
Fax: (+1)919-962-2388
E-mail: wlin@unc.edu
Homepage: http://www.chem.unc.edu/people/faculty/linw/
wlindex.html
Figure 1. Crystal structure of 1. A)The 2D square grid in 1. B)The 1D
zigzag polymeric chain in 1. C)Schematic representation of the 3D
framework of 1. D)Schematic representation of the twofold interpene-
tration of 1. E)Space-filling model of 1 as viewed down the c axis
[**] We acknowledge financial support from the NSF (CHE-0512495).
W.L. is a Camille Dreyfus Teacher-Scholar. We thank Dr. Lin Zhang
for experimental help.
2
showing the chiral 1D channels of 13.5”13.5 Š in dimensions. Color
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
scheme for A), B), and E): cyan: Cd, green: Cl, red: O, blue: N, gray:
C, and light gray: H.
Angew. Chem. Int. Ed. 2007, 46, 1075 –1078
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1075

Communications
(Figure 1C). The 3D frameworks in 1 form twofold inter-
penetrated structures, presumably to avoid the formation of
extraordinarily large open space (Figure 1D). The interpene-
trated 3D networks are held together by strong p···p
interactions between the pyridyl and adjacent naphthyl
rings with shortest C···C distances of 3.399 Š. Even after
twofold interpenetration, compound 1 still possesses very
large interconnected channels that are occupied by the
solvent molecules. The approximate dimensions of these
2
channels are 4.9 ” 13.1 Š along the a and b axes and 13.5 ”
2
13.5 Š along the c axis.
Calculations performed using PLATON[10] showed that
the effective volume for inclusion is about 10837.1 Š3 per unit
cell, which is about 53.0% of the crystal volume. Thermo-
gravimetric analysis (TGA) of 1 showed a weight loss of
10.2% between 20 and 1458C which corresponds to the loss of
solvent molecules (calculated: 10.3%). The X-ray powder
diffraction (XRPD) spectrum of a sample of 1 that had been
ground and heated at 758C for 12 h under vacuum showed a
sharp diffraction pattern similar to that of the pristine sample.
This result indicated that the neutral chiral porous framework
was maintained after removal of all the solvent molecules.
CO2 adsorption measurements showed that an evacuated
sample of 1 possessed permanent porosity, with a Langmuir
Figure 2. A)A view of the 2D rhombic grids of [CdL 2(H2O)2]n in the
crystal structure of 2. B)Schematic representation showing the inter-
penetration of mutually perpendicular 2D rhombic grids in 2.
C)Space-filling model showing the open channels running down the c
axis. Rectangular channels with dimensions of approximately
1.2”1.5 nm run along the c axis. D)Schematic representation of the
steric congestion around the chiral dihydroxy groups of L ligands
(orange spheres)arising from the interpenetration of mutually perpen-
dicular 2D rhombic grids through p···p stacking interactions. Color
scheme in A), C), and D): cyan: Cd, green: Cl, red: O, blue: N, gray:
C, and light gray: H.
surface area of 772.3 m2 gÀ1 and
0.25 mLgÀ1.[11]
a
pore volume of
Diffusion of diethyl ether into
a
mixture of
Cd(ClO4)2·6H2O and L in DMF and MeOH led to a 3D
homochiral MOF with the formula [CdL2(H2O)2]-
[ClO4]2·DMF·4MeOH·3H2O (2; Scheme 1). Compound 2
crystallizes in the tetragonal P43212 space group with each
CdII center coordinating to four L pyridyl groups in the
equatorial plane and two water ligands in the axial positions
to form 2D rhombic grids lying along the (110) planes
(Figure 2A). Interestingly, another set of crystallographically
equivalent 2D rhombic grids lying along the (1À10) planes
form a highly interpenetrated 3D network with the set of 2D
grids lying along the (110) planes, presumably to avoid the
formation of extremely large cavities (Figure 2B). Such a
mode of interpenetration was previously observed in simpler
iron–bipyridine MOFs.[12] A space-filling model shows that 2
possesses 1D channels with dimensions of approximately 1.2 ”
1.5 nm2, even after the interpenetration, which are filled by
perchlorate anions and disordered DMF, MeOH, and H2O
solvent molecules (Figure 2C). PLATON calculations indi-
organic transformations that are typically catalyzed by Lewis
acids.[13] Treatment of 1 with excess Ti(OiPr)4 in toluene
indeed led to an active catalyst (1·Ti) for the addition of
diethylzinc to aromatic aldehydes to afford chiral secondary
alcohols upon hydrolytic workup. 1·Ti generated chiral
secondary alcohols in very high yields and enantioselectivities
(Table 1). Specifically, 1·Ti catalyzed the addition of di-
ethylzinc to 1-naphthaldehyde to afford (R)-1-(1-naphthyl)-
propanol with complete conversion and 90.0% ee (entry 1).
Table 1: Addition of diethylzinc to aromatic aldehydes.[a]
3
cate that 59.4% of the crystal volume of 2 (6064.1 Š out of
10211.2 Š3) is accessible to solvents and anions. Thermogra-
vimetric analysis showed a weight loss of 16.2% for 2 in the
20–1258C range which corresponds to the loss of all the
solvent molecules (calculated: 16.1%). CO2 adsorption
measurements indicated that an evacuated sample of 2 had
a specific surface area of 370 m2 gÀ1 and a pore volume of
0.16 mLgÀ1.
We have attempted to generate heterogeneous asymmet-
ric catalysts by activating Lewis acidic metal centers (namely,
Ti(OiPr)4) with the chiral dihydroxy groups that are present
as the orthogonal secondary functionalities in the porous
solids 1 and 2. It is well established in homogeneous systems
that TiIV–binolate complexes are active catalysts for a range of
Entry
Ar
Ligand
Loading [%]
Conv. [%]
ee [%]
1
2
3
4
5
6
7
8
1-naphthyl
4-CH3Ph
4-CH3Ph
Ph
1
1
1
1
L
1
1
1
12
12
25
12
20
12
12
12
>99
>99
>99
>99
>99
>99
>99
>99
90.0
84.2
84.9
81.9
78.6
71.0
60.2
45.0
Ph
3-BrPh
4-ClPh
4-CF3Ph
[a] All reactions were conducted at room temperature for 15 h with 1 or L
and excess amounts of Ti(OiPr)4 in toluene. Conversion and ee values
were determined by GC on a Supelco b-Dex 120 column for all of the
secondary alcohols.
1076
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 1075 –1078

Angewandte
Chemie
DMF (1 mL) and MeOH (3 mL) gave colorless crystals of 2 after one
1·Ti also catalyzed the addition of diethylzinc to a range of
other aromatic aldehydes with complete conversions and high
ee values (entries 2–4 and 6–8). The conversions and ee values
are only slightly affected by catalyst loadings (entries 2 and 3).
No secondary alcohol product was obtained when 1-naph-
thaldehyde was treated with diethylzinc in the presence of the
supernatant from a mixture of 1 and Ti(OiPr)4 under
otherwise identical conditions. This control experiment
demonstrates the heterogeneous nature of the present
catalyst system.
week. These crystals were filtered, washed with MeOH and then
Et2O, and dried at room temperature. Yield: 5 mg (67%). Elemental
analysis calcd for 2 (%): C 50.76, H 4.13, N 4.42; found: C 49.21, H
4.63, N 3.91. IR (KBr pellet): n˜ = 1662s, 1609m, 1544w, 1494m, 1420w,
1383s, 1255w, 1220w, 1186w, 1101s, 951m, 870m, 843w, 772w, 655m,
623m, 533m, 460w cmÀ1
.
Asymmetric additions of diethylzinc to aromatic aldehydes using
solid catalyst 1·Ti: Solid 1 was placed in a 25-mL Schlenk flask and
dried under vacuum at 708C for 4 h. Toluene (1 mL) was added by
syringe, followed by Ti(OiPr)4 (48 mL, 0.14 mmol) under argon. After
the mixture had been stirred for 30 min, 1-naphthaldehyde (7 mL,
50 mmol) and diethylzinc (0.154 mmol) were added. This mixture was
stirred under argon at RT for 15 h and then quenched with dilute
hydrochloric acid. The organic phase was separated and passed
through a short column of silica gel. An aliquot was analyzed by GC
on a chiral stationary phase to give the conversion and ee values.
Interestingly, however, a mixture of 2 and Ti(OiPr)4 under
identical conditions did not catalyze the addition of di-
ethylzinc to aromatic aldehydes, even though 2 possesses
permanent porosity and chiral dihydroxy groups, just as in 1.
A closer examination of the structure of 2 reveals that the
pyridyl and naphthyl rings from mutually perpendicular,
interpenetrating 2D rhombic grids form strong p···p inter-
actions, with a nearest C···C separation of 3.273 Š. As a result,
all the chiral dihydroxy groups of the L ligands are held very
close to the {Cd(py) (H O) } hinges (Figure 2D). We believe
Received: May 25, 2006
Revised: October 9, 2006
Keywords: asymmetric catalysis · chirality · functional materials ·
.
2
2
2
heterogeneous catalysis · metal–organic frameworks
that the lack of catalytic activity with the 2/Ti(OiPr)4 system is
a result of the steric congestion around these chiral dihydroxy
groups which prevents the substitution of two isopropoxide
groups by the binolate functionality. We recently proved the
inability of congested chiral dihydroxy groups to undergo
substitution reactions with Ti(OiPr)4 in a related homoge-
neous metallocyclophane system.[14]
[1] a) S. Lee, A. B. Mallik, Z. Xu, E. B. Lobkovsky, L. Tran, Acc.
Chem. Res. 2005, 38, 251; b) K. S. Suslick, P. Bhyrappa, J.-H.
Chou, M. E. Kosal, S. Nakagaki, D. W. Smithenry, S. R. Wilson,
Acc. Chem. Res. 2005, 38, 283; c) P. Feng, X. Bu, N. Zheng, Acc.
Chem. Res. 2005, 38, 293.
[2] a) B. Moulton, M. J. Zaworotko, Chem. Rev. 2001, 101, 1629;
b) M. Eddaoudi, D. B. Moler, H. Li, B. Chen, T. M. Reineke, M.
OꢀKeeffe, O. M. Yaghi, Acc. Chem. Res. 2001, 34, 319; c) S. L.
James, Chem. Soc. Rev. 2003, 32, 276; d) C. Janiak, Dalton Trans.
2003, 2781; e) M. Oh, G. B. Carpenter, D. A. Sweigart, Acc.
Chem. Res. 2004, 37, 1; f) O. R. Evans, W. Lin, Acc. Chem. Res.
2002, 35, 511; g) S. Kitagawa, R. Kitaura, S.-i. Noro, Angew.
Chem. 2004, 116, 2388; Angew. Chem. Int. Ed. 2004, 43, 2334.
[3] a) M. Fujita, Y.-J. Kwon, S. Washizu, K. Ogura, J. Am. Chem.
Soc. 1994, 116, 1151; b) J. S. Seo, D. Wand, H. Lee, S. I. Jun, J.
Oh, Y. Jeon, K. Kim, Nature 2000, 404, 982; c) L. Pan, H. Liu, X.
Lei, X. Huang, D. H. Olson, N. J. Turro, J. Li, Angew. Chem.
2003, 115, 560; Angew. Chem. Int. Ed. 2003, 42, 542; d) B.
Kesanli, W. Lin, Coord. Chem. Rev. 2003, 246, 305.
The fact that complete conversions were observed for all
of the small aromatic aldehydes used in the heterogeneous
reactions demonstrates that 1·Ti represents a unique highly
enantioselective asymmetric catalyst derived from an inter-
penetrated homochiral MOF. However, the lack of catalytic
activity with the 2/Ti(OiPr)4 system provides indirect proof
for the heterogeneous nature of the 1/Ti(OiPr)4 system. The
drastically different catalytic activities observed for the 1/
Ti(OiPr)4 and 2/Ti(OiPr)4 systems is remarkable since 1 and 2
were built from exactly the same building blocks, and this
finding highlights the important role of the framework
structure in determining the catalytic performance.
[4] a) S. Noro, S. Kitagawa, M. Kondo, K. Seki, Angew. Chem. 2000,
112, 2161; Angew. Chem. Int. Ed. 2000, 39, 2081; b) N. L. Rosi, J.
Eckert, M. Eddaoudi, D. T. Vodak, J. Kim, M. OꢀKeeffe, O. M.
Yaghi, Science 2003, 300, 1127; c) G. FØrey, M. Latroche, C.
Serre, F. Millange, T. Loiseau, A. Percheron-GuØgan, Chem.
Commun. 2003, 2976; d) B. Kesanli, Y. Cui, M. Smith, E. Bittner,
B. Bockrath, W. Lin, Angew. Chem. 2005, 117, 74; Angew. Chem.
Int. Ed. 2005, 44, 72.
In summary, we successfully synthesized a highly enantio-
selective asymmetric catalyst derived from an interpenetrated
homochiral MOF. We have observed remarkable dependency
of the catalytic activity on the framework structure for the
heterogeneous asymmetric catalysts based on the two homo-
chiral porous MOFs with the same building blocks.
[5] a) K. S. Min, M. P. Suh, Chem. Eur. J. 2001, 7, 303; b) K. Uemura,
S. Kitagawa, M. Kondo, K. Fukui, R. Kitaura, H.-C. Chang, T.
Mizutani, Chem. Eur. J. 2002, 8, 3586; c) M. P. Suh, J. W. Ko, H. J.
Choi, J. Am. Chem. Soc. 2002, 124, 10976; d) D. Bradshaw, T. J.
Prior, E. J. Cussen, J. B. Claridge, M. J. Rosseinsky, J. Am. Chem.
Soc. 2004, 126, 6106.
[6] a) B. F. Hoskins, R. Robson, J. Am. Chem. Soc. 1990, 112, 1546;
b) K. S. Min, M. P. Suh, J. Am. Chem. Soc. 2000, 122, 6834.
[7] a) C.-D. Wu, A. Hu, L. Zhang, W. Lin, J. Am. Chem. Soc. 2005,
127, 8940; b) A. Hu, H. L. Ngo, W. Lin, J. Am. Chem. Soc. 2003,
125, 11490; c) S. H. Cho, B. Ma, S. T. Nguyen, J. T. Hupp, T. E.
Albrecht-Schmitt, Chem. Commun. 2006, 2563.
Experimental Section
Synthesis of 1: Slow diffusion of diethyl ether into a mixture of
Cd(NO3)2·4H2O (15.4 mg, 0.05 mmol) and L (25.5 mg, 0.05 mmol) in
a mixed solvent of DMF (2 mL), CHCl3 (3 mL), and MeOH (2 mL)
afforded colorless crystals after one week. The crystals were filtered,
washed with MeOH and then Et2O, and dried at room temperature.
Yield: 30 mg (79% based on L). Elemental analysis calcd for 1 (%): C
49.83, H 3.62, N 6.41; found: C 50.14, H 3.39, N 6.88. IR (KBr pellet):
n˜ = 1663s, 1611s, 1544w, 1494w, 1425w, 1384m, 1290m, 1218w, 1180m,
1146w, 1090s, 1020w, 951m, 870w, 841m, 815 m, 772w, 655 m, 609w,
532w, 478w cmÀ1
.
[8] S. J. Lee, W. Lin, J. Am. Chem. Soc. 2005, 127, 4554.
[9] Crystal-structure determinations of 1 and 2 were performed on a
Siemens SMART CCD. The data were collected using graphite-
Synthesis of 2: Slow diffusion of diethyl ether into a mixture of
Cd(ClO4)2·6H2O (4.2 mg, 0.01 mmol) and L (5.1 mg, 0.01 mmol) in
Angew. Chem. Int. Ed. 2007, 46, 1075 –1078
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1077

Communications
monochromatic MoKa radiation (l = 0.71073 Š) at 173 K. The
this paper. These data can be obtained free of charge from The
Cambridge Crystallographic Data Centre via www.ccdc.cam.
ac.uk/data_request/cif.
data sets were corrected by using the SADABS program.[15] The
structures were solved by direct methods and refined by the full-
matrix least-squares method with the SHELXTL-97 program
package.[16] Crystallographic data for 1: tetragonal, space group
P4122, a = 20.305(1), c = 49.641(4) Š, V= 20466(2) Š3, Z = 4,
[10] A. L. Spek, PLATON, A Multipurpose Crystallographic Tool,
Utrecht University, Utrecht, The Netherlands, 2001.
[11] K. Kaneko, K. Murata, K. Shimizu, S. Camara, T. Suzuki,
Langmuir 1993, 9, 1165.
1calcd = 0.993 gcmÀ3
,
m(MoKa) = 0.469 mmÀ1
.
Least-squares
refinements based on 16912 reflections with I > 2s(I) and 866
parameters led to convergence, with a final R1 = 0.0639, wR2 =
0.1798, Flack parameter= 0.00(3), and GOF = 1.190. Crystallo-
graphic data for 2: tetragonal, space group P43212, a = 17.962(1),
[12] a) G. J. Halder, C. J. Kepert, B. Moubaraki, K. S. Murray, J. D.
Cashion, Science 2002, 298, 1762; b) J. A. Real, E. AndrØs, M. C.
Muꢁoz, M. Julve, T. Granier, A. Bousseksou, F. Varret, Science
1995, 268, 265.
[13] L. Pu, H.-B. Yu, Chem. Rev. 2001, 101, 757.
[14] H. Jiang, W. Lin, Org. Lett. 2004, 6, 861.
c = 31.649(1) Š, V= 10211.2(6) Š3, Z = 4, 1calcd = 1.031 gcmÀ3
,
m(MoKa) = 0.423 mmÀ1
. Least-squares refinements based on
2254 reflections with I > 2s(I) and 201 parameters led to
convergence, with a final R1 = 0.1240, wR2 = 0.3090, Flack
parameter= 0.06(14), and GOF = 1.111. CCDC-609008 (1) and
-261745 (2) contain the supplementary crystallographic data for
[15] G. M. Sheldrick, SADABS, Siemens Analytical X-ray Instrument
Division, Madison, WI, 1995.
[16] G. M. Sheldrick, Program for Structure Refinement: University
of Gꢂttingen, Germany, 1997.
1078
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 1075 –1078