A family of chiral metal-organic frameworks

Article abstract of DOI:10.1002/chem.201002568

Chiral metal-organic frameworks with a three-dimensional network structure and wide-open pores (>30) were obtained by using chiral trifunctional linkers and multinuclear zinc clusters. The linkers, H₃ChirBTB-n, consist of a 4,4′,4′′-benzene-1,3

Full text of DOI:10.1002/chem.201002568

FULL PAPER
DOI: 10.1002/chem.201002568
A Family of Chiral Metal–Organic Frameworks
Kristina Gedrich,[a] Maja Heitbaum,[b] Andreas Notzon,[b] Irena Senkovska,[a]
Roland Frçhlich,[b] Jꢀrgen Getzschmann,[a] Uwe Mueller,[c] Frank Glorius,*[b] and
Stefan Kaskel*[a]
Abstract: Chiral metal–organic frame-
works with a three-dimensional net-
work structure and wide-open pores
(>30 ꢀ) were obtained by using chiral
trifunctional linkers and multinuclear
zinc clusters. The linkers, H3ChirBTB-
n, consist of a 4,4’,4’’-benzene-1,3,5-
triyltribenzoate (BTB) backbone deco-
rated with chiral oxazolidinone sub-
stituents. The size and polarity of these
substituents determines the network
topology formed under solvothermal
synthesis conditions. The resulting
chiral MOFs adsorb even large mole-
cules from solution. Moreover, they
are highly active Lewis acid catalysts in
the Mukaiyama aldol reaction. Due to
their chiral functionalization, they
show significant levels of enantioselec-
tivity, thereby proving the validity of
the modular design concept employed.
Keywords: aldol reaction · chirali-
ty · dyes/pigments · enantioselectivi-
ty · metal–organic frameworks
Introduction
employs MOFs was achieved with POST-1, which was used
to catalyze the kinetic resolution of rac-1-phenyl-2-propanol
by means of a transesterification process.[3] Later, some priv-
ileged chiral ligands of asymmetric catalysis were modified
and employed as linkers for MOF formation.[4–6] 1,1’-Bi-2-
naphthol (BINOL)-based ligands in particular were exten-
sively studied by Lin et al., recently resulting in the forma-
tion of the first isoreticular series of chiral MOFs.[7] The re-
sulting highly enantioselective catalysts were used with suc-
cess; however, these catalysts use the established mode of
action of the corresponding homogeneous catalysts, and in
addition, the coordination polymers were not highly crystal-
line.[5] Thus, the full structural characterization of chiral
MOFs remains a challenge.[4,5]
Recently, metal–organic frameworks (MOFs) have emerged
as a new class of porous materials with extraordinarily large
accessible surface areas and pore sizes that exceed those of
traditional adsorbents such as zeolites and activated car-
bons.[1] The crystallinity renders these materials to be prom-
ising specific catalysts due to the possibility of localizing the
active sites in the structure, but also due to monomodal
pore-size distribution available to achieve size-selective ef-
fects.[2] The rational building block approach of MOFs is
ideally suited for the integration of functionalities into the
framework, particularly for the design of porous chiral MOF
catalysts or adsorbents.
Asymmetric catalysis is of increasing importance, since
the demand for enantiomerically pure compounds is rising.
An initial breakthrough in the area of chiral catalysis that
MOFs represent a three-dimensional platform that allows
the design of completely new catalyst architectures. They
offer a family of building blocks, namely, nodes and linkers
with a wide range of different connectivity. In addition, the
pore diameter can be adjusted by changing the length of the
linker.[8] However, whereas linker functionality determines
rather rigorously its connectivity, the formation of multi-
[a] K. Gedrich, Dr. I. Senkovska, Dr. J. Getzschmann,
Prof. Dr. S. Kaskel
Department of Inorganic Chemistry
Dresden University of Technology
Bergstrasse 66, 01069 Dresden (Germany)
Fax : (+49)351-46337287
AHCTUNGTREGnNNUN uclear clusters used as MOF building blocks (nodes) is
often hard to predict. Thus, in zinc carboxylates, di-, tri-,
and tetranuclear clusters are observed as the predominant
species that serve as 4- or 6-connecting nodes. Another am-
biguity in MOF design is interpenetration.[9] Even though
rules have been found that rationalize interpenetration,[10]
for more complex and substituted linkers, the packing is dif-
ficult to predict and solvent–linker interaction will affect the
degree of interpenetration.
E-mail: stefan.kaskel@chemie.tu-dresden.de
[b] Dr. M. Heitbaum, A. Notzon, Dr. R. Frçhlich, Prof. Dr. F. Glorius
Organisch-Chemisches Institut
Westfꢁlische Wilhelms-Universitꢁt Mꢂnster
Corrensstrasse 40, 48149 Mꢂnster (Germany)
Fax : (+49)251-8333202
E-mail: glorius@uni-muenster.de
[c] Dr. U. Mueller
Institute Soft Matter and Functional Materials
BESSY-MX Group
Helmholtz-Zentrum Berlin fꢂr Materialien und Energie
Albert-Einstein-Strasse 15, 12489 Berlin (Germany)
Besides such limitations, in our view, the rational build-
ing-block approach is applicable to the design of porous
chiral MOF catalysts, and a new strategy seems to be
attractACHTNUTRGNEiUNG ve: the attachment of chiral groups/auxiliaries to
linkers successfully used for MOF synthesis, thereby leading
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201002568.
Chem. Eur. J. 2011, 17, 2099 – 2106
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2099

to the formation of unprecedented chiral MOF catalysts
with no homogenous analogue. Moreover, these chiral auxil-
iaries, placed in close vicinity to the open accessible metal
nodes as an integral part of the MOF structure, could
induce enantioselectivity in Lewis acid or metal-catalyzed
reactions. A related synthetic approach was pursued by
Fujita et al. by decorating molecular Pd12L24 cages (L=1,3-
bis(pyridin-4-ylethynyl)benzene) with amino acids or pep-
tides.[11] However, the obtained material was not tested in
asymmetric catalysis. For MOF catalysts, one requirement is
the presence and accessibility of active metal centers. Such
active sites are found in M2 paddle-wheel units (M=Cu, Zn,
Ni, Co, Mo) and can be used as Lewis acid catalysts.[12]
Three-dimensional networks with very large pores in which
accessible sites are not blocked are formed in combination
with tricarboxylates. The combination of square and triangu-
lar building blocks allows for the formation of two different
topologies,[13] the twisted boracite topology as is found in
formation of wide-open structures for accessibility of the
active sites. At the same time, chiral oxazolidinone substitu-
tion with varying steric demand in close vicinity to the coor-
dinating carboxylate group introduces a new modular con-
cept for the synthesis of enantioselective MOF catalysts.
Results and Discussion
Synthesis of chirally substituted 4,4’,4’’-benzene-1,3,5-triben-
zoic acids: The chiral acids were synthesized in five steps
from commercially available 4-iodoacetophenone
1
(Scheme 1). The 4-iodoacetophenone was brominated selec-
tively in the 3-position, and then the iodide was converted
into ethyl ester 2 by palladium-catalyzed carbonylation. The
chiral oxazolidinone was introduced by a copper-catalyzed
coupling of the aryl bromide. Finally, acetophenone deriva-
tives 3a and 3b were trimerized with SiCl4 and the ester was
cleaved under basic conditions to give the chiral H3BTB de-
rivatives 5a and 5b.
HKUST-1 (Cu
late)[14] and in [Cu
azine-2,4,6-triyltribenzoate),[15] and the Pt3O4 topology with
(BTC)2; BTC=1,3,5-benzene tricarboxy-
3A(TATB)2A(H2O)3] (TATB=4,4’,4’’-s-tri-
CHTUNGTRENNUNG
BTB (Cu
G
Synthesis and crystal structure of metal–organic frame-
works: Crystals of [Zn3(ChirBTB-1)2] (6) were obtained by
treating H3ChirBTB-1 (5a) with an excess amount of zinc
nitrate in diethylformamide (DEF) at 1008C for 20 h. The
structure analysis of the resulting cubic structure of 6 (space
group F432, a=47.515(2) ꢀ, Figure 1c) is a challenge on its
own due to the large cell volume and extremely open struc-
ture with meso- and micropores. Finally, single-crystal syn-
chrotron data provided the basis for a successful refinement
of the complex structure. Zinc paddle-wheels in 6 are sur-
rounded by four ChirBTB-1 ligands (Figure 1a). The axial
sites of the paddle-wheel are occupied by oxygen atoms of
solvent molecules. The BTB is not planar and rather screw-
like. As expected, the structure of 6 has a twisted boracite
topology analogous to HKUST-1[13] (see Figure S3a in the
Supporting Information). Three
zoate) in MOF-14.[16,17]
Consequently, to introduce chiral groups in the vicinity of
the connecting and catalytically active paddle-wheels, we
prepared chiral tricarboxylic acids with a BTB backbone
substituted by chiral oxazolidinones (H3ChirBTB-n): 1,3,5-
tri{4-[2-(4-isopropyl-2-oxooxazolidin-3-yl)]benzoate}benzene
(5a, H3ChirBTB-1) and 1,3,5-tri{4-[2-(4-benzyl-2-oxooxazo-
lidin-3-yl)]benzoate}benzene
(Scheme 1). We reasoned that oxazolidinones, established
chiral auxiliaries in stereoselective synthesis,[18] might be a
good group because these rigid cyclic system exhibit a signif-
icant dipole moment and should be tolerant of the condi-
tions of MOF formation. On the other hand, the rigid and
extended nature of the BTB ligand is ideally suited for the
types of pores are present (Fig-
ure 1c). A larger pore (LP) ap-
proximately 33.7 ꢀ in diameter
À
(Zn Zn atom center distance)
exhibits accessible Zn sites that
point towards the center of the
pore. The LP has windows of
12.8 ꢀ and is polar in character.
The small pore (ꢀ13.7 ꢀ in di-
ameter) is formed by four
ChirBTB-1 linkers. In addition,
a third type of pore only slight-
ly smaller than the large pore is
present. This pore is also con-
fined by ChirBTB-1 linkers and
represents the interstitial space
between the large and the small
pores (this pore is not being
shown in Figure 1c).
Scheme 1. Synthesis of the chiral H3BTB-derivatives 5a and 5b starting from 4-iodoacetophenone (1).
DMEDA=N,Nꢄ-dimethylethylenediamine.
2100
www.chemeurj.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 2099 – 2106

Chiral Metal–Organic Frameworks
FULL PAPER
the Supporting Information)
match the patterns calculated
from the crystal structure.
Slight differences can be as-
signed to the fact that the en-
closed solvent molecules could
not be taken into account for
the calculation of the theoreti-
cal patterns.
The composition of the
frameworks was further con-
firmed by thermal analysis (see
Figures S8 and S9 in the Sup-
porting Information) combined
with elemental analysis. In both
cases, the solvent molecules
bound to the metal centers
could not be removed by sol-
vent exchange with CH2Cl2 and
subsequent drying in vacuum.
In addition, several molecules
of water are included, thereby
leading to the final framework
compositions of [Zn3(ChirBTB-
Figure 1. a) Paddle-wheel unit within the crystal structure of [Zn3(ChirBTB-1)2] (6). The oxazolidinone subunit
is disordered over two positions. b) Zn3ACHTNUTRGNE(UNG RCOO)2 cluster in [Zn3(ChirBTB-2)2] (7). For reasons of clarity the
carbon atoms of both linkers are represented in two different colors. c) Crystal structure of 6 exhibiting two
different pores, represented in orange. d) Channels in the crystal structure of 7, represented in orange. Only
one oxazolidinone subunit is shown. Zn: green, O: red, N: blue, C: (dark) gray.
1)
[Zn3(ChirBTB-2)
AHCTUNGTRENNG(NU H2O)3] for 7. The presence of
2A
R
6 and
CTHUNGTRENNUNG
Even though similar reaction conditions were used for the
preparation of [Zn3(ChirBTB-2)2] (7), the crystal structure
differs completely from that found for [Zn3(ChirBTB-1)2]
(6). Compound 7 crystallizes in the tetragonal, chiral space
DEF and water was also verified by the IR spectra of com-
pounds 6 and 7 (see Figures S10 and S11), indicated by
strong bands around 1643 cmÀ1 (IR spectrum of 6) and
1657 cmÀ1 (IR spectrum of 7) for the C=O valence vibration
of DEF and broad bands between 3000 and 3700 cmÀ1
(H2O), respectively. Furthermore, bands caused by symmet-
ric valence vibrations of the carboxylate groups were identi-
fied (6: 1402 and 1433 cmÀ1, 7: 1400, 1435 and 1456 cmÀ1).
The presence of the oxazolidinone subunits is proven by
very strong bands at 1732 and 1761 cmÀ1 (6) and 1734 and
1763 cmÀ1 (7), respectively, which can be assigned to the C=
O stretching vibration of the oxazolidinone ring. The bands
group P43212 (Figure 1d). Trinuclear M3ACHTNUGTRNEUG(N RCO2)6 (M=Zn,
Mg) clusters, a common secondary building unit (SBU) in
porous MOFs such as MOF-3 and TUDMOF-2 and -3[19]
form the nodes of the network. The three zinc atoms of the
SBU form a nearly linear array (Zn-Zn-Zn-angle 175.8(3)8;
Figure 1b). The central zinc atom has octahedral coordina-
tion, whereas the terminal zinc atoms are distorted tetrahe-
dral, thereby resulting in a 6-connecting node. The SBUs
are connected by the chiral linker to form a chiral 3D net
with cys topology (every net has a name consisting of three
characters, in this case “cys”. The different network topolo-
gies were collected by the Center for Reticular Chemistry
(http://rcsr.anu.edu.au/), see Figure S3b in the Supporting
Information). The linkers are stacked pairwise (Figure 1b)
due to the strong p–p interactions between two face-to-face
stacked ligands,[20] thus resulting in a relatively dense struc-
ture with two types of channels that run through all three di-
rections of the crystals. The rectangular channels along the
c-axis are 18ꢅ18 ꢀ (as measured from atom center to atom
center) and are partially occupied by the benzyl groups of
the oxazolidinone substituents (Figure 1d). The solvent-ac-
cessible volume calculated by using PLATON[21] is 60.3% of
the unit cell volume for 6 and 50.7% for 7.
between 2800 and 3000 cmÀ1 caused by aliphatic C H vibra-
À
tions support this finding, too. In the case of MOF 7, addi-
tional bands at 3028, 3062, and 3086 cmÀ1 are observed (aro-
matic valence vibrations) that are caused by the presence of
the benzyl group.
Due to the internal stress caused upon the drying of such
highly porous structures, thus resulting in pore collapse, gas
physisorption that requires activation in vacuum was not the
method of choice. According to Monte Carlo integration,
the accessible surface area for 6 and 7 is 2278 and
1990 m2 gÀ1, respectively.[22] Even though nitrogen or argon
physisorption is usually used for determination of porosity
of the materials, for catalytic applications in fine chemical
production, not the access of gases but rather pore access
for large molecules in solution is essential. Thus, to prove
the permanent porosity and accessibility for larger mole-
cules of 6 and 7, three different dye molecules were ad-
sorbed in solution (Table 1). Both metal–organic frame-
Host–guest systems: The powder X-ray diffraction patterns
of freshly made samples of 6 and 7 (see Figures S4 and S5 in
Chem. Eur. J. 2011, 17, 2099 – 2106
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2101

F. Glorius, S. Kaskel et al.
Reichardtꢄs dye
Table 1. Adsorption of different dyes within the pores of 6 and 7.
Fluorescein
Merocyanine
structure
ACHTUNGTRENNUNG[Zn3(ChirBTB-1)2] (6)
ACHTUNGTRENNUNG[Zn3(ChirBTB-2)2] (7)
[a] Fluorescein@MOF under UV irradiation.
works are able to adsorb fluorescein (8) and merocyanine 9
from solution. The adsorption behavior differs, however,
when Reichardtꢄs dye (10),[23] a very large molecule (1.26ꢅ
1.01ꢅ0.81 nm3)[24] is used. Dye 10 was previously employed
to probe the enormous pore size in MOF-177[16] and is com-
pletely adsorbed by 6 (see Figure S12 in the Supporting In-
formation). As determined by 1H NMR spectroscopy (see
Figures S13 and S14), one molecule of Reichardtꢄs dye (10)
tional long-term stability if one judges from the powder X-
ray diffraction pattern of merocyanine 9@Zn3(ChirBTB-2)2
(7) obtained two years after loading (see Figure S5d).
Application in Mukaiyama aldol reactions: Since the pores
of 6 and 7 are accessible even for large molecules, the Lewis
acidic metal sites can be used for the catalytic conversion of
organic substrates. Compounds 6 and 7 were tested in the
Mukaiyama aldol reaction of benzaldehyde (11a) and 1-
naphthaldehyde (11b) with 1-methoxy-1-(trimethylsiloxy)-2-
methyl-1-propene (12) (Scheme 2; for details, see the Exper-
is adsorbed per formula unit [Zn3(ChirBTB-1)2] (6), whereas
1
MOF-177 adsorbs only = molecules of Reichardtꢄs dye (10)
8
per formula unit Zn4OACTHNUTRGNEUNG
(BTB)2.[16] Compound 7 is not able to
adsorb the large Reichardtꢄs dye (10, Table 1). Though the
crystal structure of 7 exhibits rectangular channels (Fig-
ure 1d), these channels are not accessible for the larger dye
since they are occupied by benzyl groups as described
above. Furthermore, the solvatochromic Reichardtꢄs dye
(10) shows a color change from green in the less-polar sol-
vent CH2Cl2 towards violet in the MOF (see Figure S12),
thus indicating adsorption in the hydrophilic LP of 6, which
exposes polar functional groups such as Zn sites and amide
groups. Besides that, dye@MOF materials exhibit an excep-
Scheme 2. Mukaiyama aldol reactions with 1-methoxy-1-(trimethylsilyl-
oxy)-2-methyl-1-propene (12) and benzaldehyde (11a) or 1-naphthalde-
hyde (11b).
2102
www.chemeurj.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 2099 – 2106

Chiral Metal–Organic Frameworks
FULL PAPER
imental Section and Supporting Information) that required
relatively strong Lewis acid catalysts.[25] The application of
metal–organic frameworks in Mukaiyama aldol reactions
was first investigated by Long and co-workers. They used a
MOF with exposed manganese sites in the catalytic conver-
sion of different aldehydes and silyl enol ethers.[25] After
four daysꢄ reaction time, they obtained a moderate yield of
63% by using benzaldehyde (11a) and silyl enol ether 12. In
this case, the metal atoms that formed the nodes of the
framework were themselves used as catalytic active sites.
Cohen et al. successfully employed a postsynthetic approach
to synthesize an active MOF catalyst by fixation of an iron
complex at the framework of UMCM-1.[26] The catalyst thus
formed was able to convert the bulky aldehydes 1-naphthal-
dehyde (11b) and mesitylaldehyde. But, up to now, the suc-
cessful application of chiral MOF catalysts in the enantiose-
lective Mukaiyama aldol reaction has not been reported to
the best of our knowledge.
The catalytic performance of 6 and 7 was compared with
that of zinc nitrate tetrahydrate (15), a homogeneous ana-
logue; MOF-177 (16), a well-known zinc-based MOF with a
large accessible pore volume;[16] and MIL-101 (17),[27] an
MOF that exhibits high catalytic activity in Lewis acid medi-
ated reactions.[28] For reasons of comparison, the quantity of
MOF employed was calculated in a way that the number of
metal sites with regard to the amount of aldehyde was the
same. The conversion of benzaldehyde (11a) in the Mukai-
reaction time and reached by 7 after seven days. The perfor-
mance of [ZnACHTNURTGENNUG(NO3)2ACHTUNTGREN(NUNG H2O)4] (15) is comparable to that of 7
and thus unexpectedly low, probably due to the low solubili-
ty of zinc nitrate in n-heptane. The poorest performance in
the Mukaiyama aldol reaction was with benzaldehyde (11a)
and MOF-177 (16). Since this material has only coordina-
tively saturated zinc atoms, this finding seems to be rational.
The catalytic activity of 16 can be explained by defects in
the crystal structure[29] or a short lifetime expansion of the
coordination number, as was already observed for DUT-7, a
Zn4O-based chiral MOF.[30] Judging from the observed con-
version rates, the new chiral MOFs 6 and 7 show competi-
tive catalytic performances, especially with regard to MIL-
101 (17).
Besides the conversion of the starting material 11a, the
nature and enantiomeric excess (ee) of the formed products
is of greater interest. For all catalysts that exhibit open
metal sites, a considerable amount of 14a (the desilylated
form of 13a) is formed (6: 22%, 7: 20%, 17: 27%). This di-
minishes the yield of 13a expected from the conversion
rates (Figure 2c, solid lines) to 77% for 6 (conversion:
99%), 74% for 7 (conversion: 98%), and 69% for MIL-101
(17, conversion: 98%). But on the whole, both compounds
13a and 14a are products of the Mukaiyama aldol reaction
(Figure 2c, dotted lines), thereby resulting in good product
selectivities for the new chiral MOF materials 6 and 7. This
finding is also supported by the small amount of unidenti-
fied side products formed during the reaction (ꢁ1% for 6
and ꢀ4% for 7; see Figure S15 in the Supporting Informa-
tion). The yield of 13a observed during the performance
studies is also supported by the isolated yields after column
ACHTUNGTRENNUNGyama aldol reaction with silyl enol ether 12 by the afore-
mentioned catalysts is presented in Figure 2a. As expected,
the highest initial catalytic performance is found for MIL-
101 (17) but it is surpassed by that of 6 after only one day of
chromatography
(Table 2).
MOF-177 (16) is the best cata-
lyst in terms of the selectivity
of formed product 13a, since
only 4% of desilylated product
14a and 2% further side prod-
ucts are formed. This finding is
in accordance with the poor
performance in terms of the
conversion of benzaldehyde
(11a).
As discussed in the foregoing
sections, [Zn3(ChirBTB-1)2] (6)
is an efficient catalyst for the
Mukaiyama aldol reaction that
uses benzaldehyde (11a). Inter-
estingly, the enantioselectivity
of 6 as catalyst during catalysis
depends on the solvent used. If
n-heptane is used as solvent, an
enantiomeric excess of 9% is
detected (Table 2, entry 7) but
no influence on the enantiose-
lectivity of the formed product
could be observed when di-
Figure 2. Catalytic performances of [Zn3(ChirBTB-1)2] (6) and [Zn3(ChirBTB-2)2] (7) relative to [ZnACHTNUGTRNEG(UN NO3)2-
ACHTUNGTRENNUNG(H2O)4] (15), MOF-177 (16), and MIL-101 (17) in n-heptane at room temperature. a) Conversion of benzalde-
hyde (11a). b) Conversion of 1-naphthaldehyde (11b). c) Yield of 13a. The dotted gray line represents the
sum of the yields of 13a and 14a. d) Yield of 13b. Catalyst 6: triangles; catalyst 7: squares with crosshair; cata-
lyst 15: circles; catalyst 16: squares; catalyst 17: diamonds.
chloromethane
was
used
Chem. Eur. J. 2011, 17, 2099 – 2106
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2103

F. Glorius, S. Kaskel et al.
Table 2. Results of Mukaiyama aldol reactions using dichloromethane or
n-heptane as solvent.
of the conversion rates for all catalysts (Figure 2b). Com-
pound 7 showed no conversion at all. Again, compound 6
shows the best performance (conversion: 97%), but in con-
trast to the conversion of benzaldehyde (11a), the gap be-
Entry
Aldehyde
MOF
Solvent
Yield [%]
ee value [%]
/Reaction time
dichloromethane
tween 6 and the other catalysts—[ZnACTHNUTRGENN(UG NO3)2ACHTUGNTRNE(NUGN H2O)4] (15,
1
2
3
4
5
6
11a
11b
11b
11a
11a
11b
6
6
6
7
7
7
48 h
17 d
17 d
13 d
13 d
83[a]
31[a]
22[a]
66[a]
54[a]
0[a]
0
40[b,c]
10[b,d]
8[e]
conversion: 65%), MOF-177 (16, conversion: 61%), and
MIL-101 (17, conversion: 60%)—is significant. As already
found for 11a, a large number of side products are formed if
catalysts with open metal sites (6, 17) are employed (see
Figure S16 in the Supporting Information), thus leading to
decreased yields in comparison to the observed conversions
(6: 86%, 17: 36%). In contrast to the use of benzaldehyde
(11a), in the case of 1-naphthaldehyde (11b) none of the
peaks in the chromatogram could be assigned to the desily-
lated product 14b, thus the effective yields are possibly
higher. The isolated yield of 13b after column chromatogra-
phy is significantly lower (31%, Table 2, entry 2) than that
observed during the kinetic measurements (Figure 2d and
Table 2, entry 8), but a significant enantiomeric excess of
40% was observed in the first run, which decreased to 10%
in a second one (both dichloromethane as solvent, Table 2,
entries 2 and 3). If n-heptane is used as solvent, an enantio-
meric excess of 16% is detected (Table 2, entry 8).
The different behavior of 6 and 7 in the Mukaiyama aldol
reaction is consistent with their crystal structures that exhib-
it completely different pore systems. A filtration test proved
the assumed heterogeneity of the reaction mechanism. The
framework of all MOF catalysts remains intact on the basis
of the powder X-ray diffraction patterns before and after
the catalytic reactions, since they do not show significant
changes (see Figure S4–S7 in the Supporting Information).
8[e,d]
–
14 d
n-heptane
48 h
48 h
7 d
48 h
7
8
11a
11b
6
6
77 (98)[f]
77[f]
9[e]
16[b]
84[f]
9
11a
7
43 (82)[f]
74 (94)[f]
6[e]
12 d
[a] Isolated yield of 13a after column chromatography. [b] Determined
by chiral HPLC. [c] However, a different batch of material gave a 10%
ee in the first cycle. [d] Second cycle. [e] Determined by chiral GC.
[f] Yield of 13a determined during kinetic measurements; yield of 13a+
14a in parentheses.
(Table 2, entry 1). [Zn3(ChirBTB-2)2] (7) catalyzes the reac-
tion of benzaldehyde (11a) and the silyl enol ether 12 with a
decreased reaction rate and a small but reproducible enan-
tiomeric excess of the product 13a of 6% (n-heptane) and
8% (dichloromethane), respectively, was measured (Table 2,
entries 4, 5, and 9).
Since the catalytically active sites of 6 and 7 are blocked
by solvent molecules, which cannot be removed in vacuum
without a collapse of the framework, the influence of these
coordinated solvent molecules was studied for MIL-101
(17). A portion of 17 equivalent to the amount used in the
standard reaction described in the Experimental Section was
completely soaked with an excess amount of dry DEF for
one day. Afterwards, the supernatant DEF was removed
and the sample treated with dry dichloromethane to elimi-
nate DEF molecules occluded in the pores of MIL-101 but
not the ones coordinated to chromium centers. After activa-
tion in vacuum at room temperature, the obtained material
denoted as DEF@MIL-101 (18) was used in the Mukaiyama
aldol reaction with benzaldehyde (11a) and silylenol ether
12 (Scheme 2). As expected, the initial reaction rate drops
significantly (see Figure S17 in the Supporting Information).
In contrast to pure MIL-101, the kinetic curve does not
reach a plateau after three daysꢄ reaction time but a lower
rate is observed until it reaches the same conversion after
12 days. This steady increase in conversion indicates the suc-
cessive replacement of the coordinated DEF molecules by
benzaldehyde (11a). This result supports the assumption
that the catalytic activity of 6 and 7 arises from the open
zinc centers at the nodes of the frameworks and not from
defects in the crystal structure. In comparison to MIL-101
(17), the use of DEF@MIL-101 (18) suppresses the forma-
tion of the desilylated product 14a (see Figure S18) and fur-
ther side products (see Figure S19).
Conclusion
In the present contribution, we describe the successful syn-
thesis of metal–organic frameworks derived from two new
chiral organic tricarboxylates of the 4,4’,4’’-benzene-1,3,5-tri-
benzoic acid (H3BTB) family. Two new metal–organic
frameworks, [Zn3(ChirBTB-1)2] (6) and [Zn3(ChirBTB-2)2]
(7), were obtained by the reaction of the aforementioned
acids with an excess amount of zinc nitrate in diethylform-
AHCTUNGTREGaNNNU mide. Depending on the nature of the chiral oxazolidinone,
either substituted by an isopropyl or by a benzyl group, two
completely different framework topologies were observed.
The isopropyl-substituted oxazolidinone ligand causes the
formation of a structure analogous to that of Cu3ACTHNUTRGNEU(GN BTC)2. In-
troduction of the benzyl-substituted oxazolidinone, on the
other hand, has a striking effect on the crystal structure
formed. The [Zn3ACHTNUTRGNEUNG(RCOO)6] nodes identified in the crystal
structure of 7 are interconnected by ChirBTB-2 linkers
stacked pairwise due to p–p interactions. The crystal struc-
tures of 6 and 7 show large pore systems accessible even for
large substrates proven by dye adsorption from solution.
Compound 6 is even capable of adsorbing the extraordinari-
ly large Reichardtꢄs dye (10) since the pore size is in the
mesopore range (2–50 nm). In contrast to this finding, the
Using the larger 1-naphthaldehyde (11b) in the Mukai-
ACHTUNGTRENNUNGyama aldol reaction with silylenol ether 12 results in a drop
2104
www.chemeurj.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 2099 – 2106

Chiral Metal–Organic Frameworks
FULL PAPER
ther collected by filtration under argon and dried in vacuum at room
temperature. Yield: 64 mg (71% referred to the amount of H3ChirBTB-
2). IR: n˜ =656 (w), 700 (m), 729 (m), 756 (m), 796 (m), 825 (m), 864 (w),
1001 (w), 1038 (m), 1082 (m), 1115 (m), 1144 (s), 1240 (m), 1265 (m),
1311 (m), 1400 (s), 1435 (s), 1456 (s), 1498 (m), 1552 (m), 1614 (s), 1657
(s), 1734 (s), 1763 (s), 2912 (m), 2978 (w), 3028 (w), 3062 (w), 3086 (w),
2700–3700 cmÀ1 (br); elemental analysis calcd (%) for Zn3(ChirBTB-2)2-
space in the channels found in the crystal structure of 7 is
not sufficient for the adsorption of 10, probably because
these channels are occupied by a third of the oxazolidinone
moieties. Both MOFs exhibit metal sites accessible for po-
tential substrates, since they are occupied by solvent mole-
cules. These metal sites are surrounded by the chiral oxazo-
lidinone subunits fixed at the BTB linker (see Figures S1
and S2 in the Supporting Information). In the Mukaiyama
aldol reactions catalyzed by 6 and 7, enantiomeric excesses
of up to 40% were obtained. Thus the preparation of chiral
MOF catalysts by the incorporation of auxiliaries known
from homogeneous catalysis into linkers not yet established
in MOF synthesis has been proven to be successful. Even
though the selectivity in these few test reactions was low,
the enantioselectivity is a proof of the validity of this new
concept. Further development and understanding of linker–
substrate interactions is required for enhancing enantiose-
lectivity and yield.
In summary, we have presented the synthesis of the two
new chiral linkers H3ChirBTB-1 (5a) and H3ChirBTB-2
(5b) and their successful use for the synthesis of chiral
MOF catalysts for asymmetric catalysis. Although com-
pound 6 has the same topology as HKUST-1, slight changes
in the oxazolidinone substituent lead to a completely differ-
ent topology in 7. The pore systems of both materials are ac-
cessible to even large molecules. Both MOFs are active
Lewis acid catalysts and show enantioselectivity in the Mu-
kaiyama aldol reaction.
(H2O)3:[31] C 62.7, H 4.75, N 4.72, O 19.5, Zn 8.26; found: C
ACHUTNRGEN(NUG DEF)2ACHTUNGTRENNUGN
(62.6Æ0.5), H (4.97Æ0.05), N (4.82Æ0.08), O (19.2Æ0.2), Zn (7.92Æ
0.04).
Crystal data: Zn3(C57H42O12N3)2·2ACTHNUTRGNEUNG
(C5H11NO); Mr =2320.38 gmolÀ1; tet-
ragonal, P43212 (no. 96); a=28.3927(4), c=20.3756(3) ꢀ; V=
16425.7(4) ꢀ3; Z=4; 1calcd =0.920 gcmÀ3; synchrotron l=0.88561 ꢀ; T=
208C; qmax =28.418; reflections collected/unique 38101/10603; Rint
=
0.0561, R1 =0.0963, wR2 =0.2580; Flack parameter x=À0.09(2); largest
diff. peak and hole: 0.085 and À0.011 eꢀÀ3
.
CCDC-735821 (6) and 735822 (7) contain the supplementary crystallo-
graphic data for 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.
Dye adsorption from solution: Freshly prepared samples of 6 and 7, re-
spectively, were transferred to concentrated solutions of fluorescein (8)
in DEF/CH2Cl2, merocyanine 9 in DEF/MeOH, and Reichardtꢄs dye (10)
in CH2Cl2. After several days, the crystals were washed with fresh DEF
to remove dye molecules not soaked into the MOF.
To determine the amount of adsorbed 10, sample 6 was placed in a solu-
tion of dye in CH2Cl2 for 6 d and washed thoroughly with CH2Cl2 (see
Figure S7 in the Supporting Information). The resulting dye@MOF was
subsequently hydrolyzed with aqueous HCl. The 1H NMR of the hydro-
lyzed MOF shows a ratio of 2:1 (linker/dye) (see Figures S13 and S14).
Performance of Mukaiyama aldol reactions: The Mukaiyama aldol reac-
tions were performed by using 6 and 7 as catalysts, as well as by applying
MIL-101 (17), MOF-177 (16), and [ZnACHTNURTGENNG(U NO3)2ACHUTNTGREN(NUGN H2O)4] (15) as reference
catalysts. The amount of catalyst employed was calculated in a way that
15 mol% accessible metal centers were present in reference to the
amount of benzaldehyde (11a) or 1-naphthaldehyde (11b). Since 6 and 7
cannot be generated in a solvent-free manner, these materials were treat-
ed in the following way prior to use. Freshly prepared samples of 6 and 7
were washed with fresh DEF and treated with absolute ethanol to ex-
change the occluded high-boiling solvent DEF with a solvent miscible
with n-heptane. The mass of the MOF samples was determined using a
pycnometer and the ethanol was replaced by dry n-heptane three times.
The other catalysts were treated as described in the supplementary part.
After transferring the respective catalyst to the reaction vessel, dry n-
heptane (10 mL) was introduced. Under stirring at room temperature,
benzaldehyde (319 mg, 11a, 3 mmol) or 1-naphthaldehyde (469 mg, 11b,
3 mmol), respectively (each freshly distilled), and silyl enol ether
(1.046 g, 12, 6 mmol) were added. The reaction was monitored by GC-
MS analysis using n-octane as standard. For the determination of the iso-
lated yields of 13a and 13b, the reaction was conducted in dichlorome-
thane and the reaction mixture was filtered through a short pad of silica
(pretreated with NEt3), washed with EtOAc, and concentrated under re-
duced pressure. The crude product was purified by flash chromatography
(silica pretreated with NEt3). The results of the catalytic test reactions
are summarized in Table 2.
Experimental Section
Synthesis of H3ChirBTB-1 and -2: A detailed description of the synthesis
route is given in the Supporting Information.
Synthesis and crystal structure of [Zn3(ChirBTB-1)2]·3DEF (6):
H3ChirBTB-1 (5a, 85.0 mg 0.104 mmol) and zinc nitrate tetrahydrate
(82.0 mg, 0.314 mmol, Merck 98.5%) were dissolved in DEF (2.5 mL).
The solution was heated in a Pyrex tube at 1008C for 20 h. The resulting
yellowish crystals were washed with fresh DEF and subsequently ex-
changed with dichloromethane within three days. The product was fur-
ther collected by filtration under argon and dried in vacuum at room
temperature. Yield: 58 mg (52% referred to the amount of H3ChirBTB-
1). IR: n˜ =658 (w), 725 (w), 769 (w), 798 (m), 825 (w), 850 (w), 877 (w),
976 (w), 1014 (w), 1053 (w), 1082 (w), 1119 (m), 1149 (m), 1236 (m), 1263
(w), 1402 (s), 1433 (s), 1643 (s), 1732 (s), 1761 (s), 2875 (w), 2931 (m),
2960 (m), 2700–3700 cmÀ1 (br); elemental analysis calcd (%) for
Zn3(ChirBTB-1)2ACHTUNGTRENNUNG(DEF)3ACHTUNGTRENNUNG
(H2O)5:[28] C 56.7, H 5.76, N 5.67, O 23.0, Zn
8.82; found: C (56.2Æ0.2), H (5.98Æ0.09), N (5.72Æ0.07), O (23.5Æ0.4),
Zn (8.35Æ0.03).
For the filtration test, the addition of the enol ether 12 to 11a with 6 was
run again in the presence of mesitylene as an internal standard and the
MOF was removed by filtration after 18 h. The ratio of product (13a) to
standard was measured directly after the filtration, after 30 h, 54 h and
7 d by means of GC-MS. No more product was formed after filtration.
Crystal data: Zn3(C45H42O12N3)2·3C5H11NO; Mr =2133.27 gmolÀ1; cubic,
F432 (no. 209); a=47.515(2) ꢀ; V=107271(9) ꢀ3; Z=16; 1calcd
=
0.527 gcmÀ3; synchrotron l=0.88561 ꢀ; T=208C; qmax =32.308; reflec-
tions collected/unique 71285/8014; int =0.0534, R1 =0.0686, wR2 =
R
0.2214; Flack parameter x=À0.13(2); largest diff. peak and hole: 0.397
and À0.895 eꢀÀ3
.
Synthesis and crystal structure of [Zn3(ChirBTB-2)2]·2DEF (7):
H3ChirBTB-2 (5b, 75.3 mg, 0.078 mmol) and zinc nitrate tetrahydrate
(61.5 mg, 0.236 mmol, Merck 98.5%) were dissolved in DEF (2.1 mL).
The solution was heated in a Pyrex tube at 1008C for 20 h. The resulting
yellowish crystals were washed with fresh DEF and subsequently ex-
changed with dichloromethane within three days. The product was fur-
Acknowledgements
The authors thank Dr. G. Auffermann (Max Planck Institute for Chemi-
cal Physics of Solids, Dresden) for the elemental analyses and P. Woll-
Chem. Eur. J. 2011, 17, 2099 – 2106
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2105

F. Glorius, S. Kaskel et al.
mann for measurement of IR spectra. Financial support by the German
Research Foundation within the priority program “Porous Metal–Organic
Frameworks” (SPP 1362, MOFs) is gratefully acknowledged. In addition,
we thank Prof. C. Reichardt for the kind donation of Reichardtꢄs dye and
Prof. Spange for the donation of merocyanine.
Blake, A. Dailly, C. M. Brown, J. M. Simmons, M. Zoppi, G. S.
Walker, K. M. Thomas, T. J. Mays, P. Hubberstey, N. R. Champness,
M. Schroder, J. Am. Chem. Soc. 2009, 131, 2159–2171.
[9] a) S. R. Batten, R. Robson, Angew. Chem. 1998, 110, 1558–1595;
Angew. Chem. Int. Ed. 1998, 37, 1460–1494; b) S. R. Batten, Crys-
tEngComm 2001, 3, 67–72.
[10] a) V. A. Blatov, L. Carlucci, G. Ciani, D. M. Proserpio, CrystEng-
Comm 2004, 6, 378; b) I. A. Baburin, V. A. Blatov, L. Carlucci, G.
Ciani, D. M. Proserpio, J. Solid State Chem. 2005, 178, 2452–2474.
[11] K. Suzuki, M. Kawano, S. Sato, M. Fujita, J. Am. Chem. Soc. 2007,
129, 10652–10653.
[12] K. Schlichte, T. Kratzke, S. Kaskel, Microporous Mesoporous Mater.
2004, 73, 81–88.
[13] O. M. Yaghi, M. OꢄKeeffe, N. W. Ockwig, H. K. Chae, M. Eddaoudi,
J. Kim, Nature 2003, 423, 705–714.
[1] a) S. Kitagawa, R. Kitaura, S.-I. Noro, Angew. Chem. 2004, 116,
2388–2430; Angew. Chem. Int. Ed. 2004, 43, 2334–2375; b) G.
Fꢆrey, Stud. Surf. Sci. Catal. 2007, 170 A, 66–86; c) U. Mueller,
M. M. Schubert, O. M. Yaghi in Handbook of Heterogeneous Cataly-
sis, Vol. 1, 2nd ed. (Eds.: G. Ertl, H. Knçzinger, F. Schꢂth, J. Weit-
kamp), Wiley-VCH, Weinheim, 2008, pp. 247–262; d) S. Kaskel in
Handbook of Porous Solids, Vol. 2 (Eds.: R. Schꢂth, K. S. W. Sing, J.
Weitkamp),Wiley-VCH, Weinheim, 2002, pp. 1190–1249.
[14] S. S. Y. Chui, S. M. F. Lo, J. P. H. Charmant, A. G. Orpen, I. D. Wil-
liams, Science 1999, 283, 1148–1150.
[2] A. Corma, H. Garcia, F. X. Llabres i Xamena, Chem. Rev. 2010,
110, 4606–4655.
[15] D. Sun, S. Ma, Y. Ke, D. J. Collins, H.-C. Zhou, J. Am. Chem. Soc.
2006, 128, 3896–3897.
[3] J. S. Seo, D. Whang, H. Lee, S. I. Jun, J. Oh, Y. J. Jeon, K. Kim,
Nature 2000, 404, 982–986.
[16] H. K. Chae, D. Y. Siberio-Perez, J. Kim, Y. B. Go, M. Eddaoudi,
A. J. Matzger, M. OꢄKeeffe, O. M. Yaghi, Nature 2004, 427, 523–
527.
[17] B. Chen, M. Eddaoudi, S. T. Hyde, M. OꢄKeeffe, O. M. Yaghi, Sci-
ence 2001, 291, 1021–1023.
[18] Y. Gnas, F. Glorius, Synthesis 2006, 1899–1930.
[19] a) H. Li, C. E. Davis, T. L. Groy, D. G. Kelley, O. M. Yaghi, J. Am.
Chem. Soc. 1998, 120, 2186–2187; b) I. Senkovska, S. Kaskel, Eur. J.
Inorg. Chem. 2006, 4564–4569; c) I. Senkovska, J. Fritsch, S. Kaskel,
Eur. J. Inorg. Chem. 2007, 5475–5479.
[20] C. Janiak, J. Chem. Soc., Dalton Trans. 2000, 3885–3896.
[21] A. L. Spek, J. Appl. Crystallogr. 2003, 36, 7–13.
[22] T. Dꢂren, F. Millange, G. Ferey, K. S. Walton, R. Q. Snurr, J. Phys.
Chem. C 2007, 111, 15350–15356.
[23] K. Dimroth, C. Reichardt, T. Siepmann, F. Bohlmann, Justus Liebigs
Ann. Chem. 1963, 661, 1–37.
[24] N. O. Mchedlov-Petrossyan, N. A. Vodolazkaya, A. A. Kornienko,
E. L. Karyakina, C. Reichardt, Langmuir 2005, 21, 7090–7096.
[25] S. Horike, M. Dinca, K. Tamaki, J. R. Long, J. Am. Chem. Soc. 2008,
130, 5854–5855.
[26] K. K. Tanabe, S. M. Cohen, Angew. Chem. 2009, 121, 7560–7563;
Angew. Chem. Int. Ed. 2009, 48, 7424–7427.
[27] O. I. Lebedev, F. Millange, C. Serre, G. Van Tendeloo, G. Ferey,
Chem. Mater. 2005, 17, 6525–6527.
[4] a) Z. Wang, G. Chen, K. Ding, Chem. Rev. 2009, 109, 322–359;
b) J. Y. Lee, O. K. Farha, J. Roberts, K. A. Scheidt, S. B. T. Nguyen,
J. T. Hupp, Chem. Soc. Rev. 2009, 38, 1450–1459; c) L. Ma, C.
Abney, W. Lin, Chem. Soc. Rev. 2009, 38, 1248–1256; d) K. Ding, Z.
Wang, X. Wang, Y. Liang, X. Wang, Chem. Eur. J. 2006, 12, 5188–
5197.
[5] M. Heitbaum, F. Glorius, I. Escher, Angew. Chem. 2006, 118, 4850–
4881; Angew. Chem. Int. Ed. 2006, 45, 4732–4762.
[6] a) L. Ma, W. Lin, J. Am. Chem. Soc. 2008, 130, 13834–13835; b) R.
Vaidhyanathan, D. Bradshaw, J.-N. Rebilly, J. P. Barrio, J. A. Gould,
N. G. Berry, M. J. Rosseinsky, Angew. Chem. 2006, 118, 6645–6649;
Angew. Chem. Int. Ed. 2006, 45, 6495–6499; c) D. N. Dybtsev, A. L.
Nuzhdin, H. Chun, K. P. Bryliakov, E. P. Talsi, V. P. Fedin, K. Kim,
Angew. Chem. 2006, 118, 930–934; Angew. Chem. Int. Ed. 2006, 45,
916–920; d) S.-H. Cho, B. Ma, S. T. Nguyen, J. T. Hupp, T. E. Al-
brecht-Schmitt, Chem. Commun. 2006, 2563–2565; e) X. Wang, L.
Shi, M. Li, K. Ding, Angew. Chem. 2005, 117, 6520–6524; Angew.
Chem. Int. Ed. 2005, 44, 6362–6366; f) C.-D. Wu, A. Hu, L. Zhang,
W. Lin, J. Am. Chem. Soc. 2005, 127, 8940–8941; g) Y. Liang, Q.
Jing, X. Li, L. Shi, K. Ding, J. Am. Chem. Soc. 2005, 127, 7694–
7695; h) A. Hu, H. L. Ngo, W. Lin, Angew. Chem. 2003, 115, 6182–
6185; Angew. Chem. Int. Ed. 2003, 42, 6000–6003; i) S. Takizawa, H.
Somei, D. Jayaprakash, H. Sasai, Angew. Chem. 2003, 115, 5889–
5892; Angew. Chem. Int. Ed. 2003, 42, 5711–5714; j) A. Hu, H. L.
Ngo, W. Lin, J. Am. Chem. Soc. 2003, 125, 11490–11491; k) O. R.
Evans, H. L. Ngo, W. Lin, J. Am. Chem. Soc. 2001, 123, 10395–
10396; l) L. Ma, C.-D. Wu, M. M. Wanderley, W. Lin, Angew.
Chem., 2010, 122, 8420–8424; Angew. Chem. Int. Ed. 2010, 49,
8244–8248.
[28] A. Henschel, K. Gedrich, R. Kraehnert, S. Kaskel, Chem. Commun.
2008, 4192–4194.
[29] U. Ravon, M. E. Domine, C. Gaudillere, A. Desmartin-Chomel, D.
Farrusseng, New J. Chem. 2008, 32, 937–940.
[30] K. Gedrich, I. Senkovska, I. A. Baburin, U. Mueller, O. Trapp, S.
Kaskel, Inorg. Chem. 2010, 49, 4440–4446.
[31] Additional water molecules can be assigned to sample handling
under an air atmosphere prior to elemental or thermal analysis.
[7] L. Ma, J. M. Falkowski, C. Abney, W. Lin, Nat. Chem. 2010, 2, 838–
846.
[8] a) M. Eddaoudi, J. Kim, N. Rosi, D. Vodak, J. Wachter, OꢄKeeffe,
O. M. Yaghi, Science 2002, 295, 469–472; b) X. Lin, J. Jia, X. Zhao,
K. M. Thomas, A. J. Blake, G. S. Walker, N. R. Champness, P. Hub-
berstey, M. Schroeder, Angew. Chem. 2006, 118, 7518–7524; Angew.
Chem. Int. Ed. 2006, 45, 7358–7364; c) X. Lin, I. Telepeni, A. J.
Received: September 3, 2010
Published online: January 17, 2011
2106
www.chemeurj.org
ꢃ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 2099 – 2106