Construction of open metal-organic frameworks based on predesigned carboxylate isomers: From achiral to chiral nets

Article abstract of DOI:10.1002/chem.200501340

The four-connected carboxylate ligand N,N,N′, N′-tetrakis(4- carboxyphenyl)-1,4-phenylenediamine (TCPPDA) exists as three stereoisomers: a pair of enantiomers (^δD₂- and ^λD₂-TCPPDA) and a diastereomer (C _2h<

Full text of DOI:10.1002/chem.200501340

DOI: 10.1002/chem.200501340
Construction of Open Metal–Organic Frameworks Based on Predesigned
Carboxylate Isomers: From Achiral to Chiral Nets
[
a]
[a]
[a]
[b]
Daofeng Sun, David J. Collins, Yanxiong Ke, Jing-Lin Zuo, and
[
a]
Hong-Cai Zhou*
Abstract: The four-connected carb-
oxylate ligand N,N,N’,N’-tetrakis(4-car-
boxyphenyl)-1,4-phenylenediamine
formamide). Complexes 1 and 2 are su-
pramolecular isomers, in which all the
ligands adopt pseudotetrahedral (both
pseudotetrahedral and rectangular geo-
metries. D -TCPPDA connects the bi-
2
nuclear neodymium units to generate a
two-dimensional layer, further linked
d
l
(
TCPPDA) exists as three stereoisom-
D - and D -TCPPDA) and rectangu-
2 2
d
ers: a pair of enantiomers ( D - and
lar (C -TCPPDA) geometries, respec-
by C -TCPPDA to create a three-di-
2
2h
2h
l
D -TCPPDA) and
a
diastereomer
tively. Both compounds connect pad-
dlewheel secondary building units
(SBUs) to form three-dimensional
porous networks possessing PtS and
NbO nets, respectively. In 3, all ligands
mensional open framework. In 5, all
the ligands possess pseudotetrahedral
2
(
C -TCPPDA). TCPPDA was prede-
2h
signed for the construction of isomeric
coordination networks. Reactions of
geometry (D -TCPPDA), as found in 1
2
and 3. However, all the TCPPDA li-
d
M
A
C
H
T
R
E
U
N
G
(NO )
(M=Cu, Zn, Co) or
gands in 5 appear as either the D or
3
2
2
d
l
Nd(NO) with TCPPDA under solvo-
possess pseudotetrahedral (both D2-
the D form, thus, the whole structure
3
2
l
thermal conditions gave rise to five
novel porous metal–organic frame-
and D -TCPPDA) geometry and link
is homochiral. Complex 5 crystallizes
2
hourglass SBUs to form a three-dimen-
in the I4 32 space group and the octa-
1
works:
[Cu (D -tcppda)-
sional porous framework. Compound 4
hedral SBU in 5 is connected by the
enantiopure TCPPDA to generate a
three-dimensional porous network pos-
sessing the corundum Al O net. Com-
2
2
A
C
H
T
R
E
U
N
G
(H O) ]·2DMSO·6H O (1), [Cu (C -
contains all three stereoisomers (C -,
2
2
2
2
2h
2
d
l
tcppda)
A
C
H
T
R
E
U
N
G
(H O) ]·2DMSO·6H O
(2),
D -, and D -TCPPDA), thus, has both
2 2
2
2
2
[
Co (D -Htcppda) ]·4DEF·5H O (3),
3 2 2 2
2
3
[
Nd (D -tcppda)(C -tcppda) -
plexes 1, 2, and 5 possess permanent
porosity, and 4 and 5 exhibit strong lu-
minescence at lmax =423 and 424 nm,
2
2
2h
0.5
Keywords: carboxylate ligands
·
A
C
H
T
R
E
U
N
G
(DMSO) ]·3DMSO·5H O (4), and
3
2
chiral networks · crystal engineer-
ing · organic–inorganic hybrid com-
posites · supramolecular chemistry
[Zn O(D -tcppda) ]·DMF·H O
(5)
4
2
1.5
2
(
DMSO=dimethyl sulfoxide, DEF=
respectively,
268.5 nm.
upon
excitation
at
diethylformamide,
DMF=dimethyl-
Introduction
properties, etc.) and topologies (perovskite, rutile, SrSi , PtS,
2
[3]
NbO, etc.). The geometries of the organic ligand and the
metal or metal-cluster secondary building unit (SBU) are
two important factors that always influence the topology of
the final structure. By predesigning and selecting the organic
ligand and the SBU, it is possible to construct novel coordi-
nation networks possessing a desired topology based on the
geometries of the ligand and the SBU. However, recent de-
velopments in supramolecular isomerism and the interpene-
tration of structures have introduced more elements of com-
plexity to the construction of MOFs and coordination net-
works, and have illuminated the difficulty in accurately pre-
Crystal engineering provides a powerful tool for the con-
struction of metal–organic frameworks (MOFs).
secondary building units and organic ligands have been spe-
cifically designed for the synthesis of MOFs with desirable
characteristics (porosity, chirality, fluorescence, magnetic
[1,2]
Various
[
a] Dr. D. Sun, D. J. Collins, Dr. Y. Ke, Prof. Dr. H.-C. Zhou
Department of Chemistry and Biochemistry
Miami University, Oxford, Ohio 45056 (USA)
Fax : (+1)513-529-8091
E-mail: zhouh@muohio.edu
[4,5]
dicting final structures.
[
b] Dr. J.-L. Zuo
Supramolecular isomerism often results from the exis-
tence of several different building units with little or no dif-
ference in formation energy. Supramolecular isomers in-
Coordination Chemistry Institute
State Key Laboratory of Coordination Chemistry
Nanjing University, Nanjing 210093 (China)
3768
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3768 – 3776

FULL PAPER
[
1,6]
clude structural isomers,
in which atom connectivities
have endeavored to carry out a systematic study of this
ligand by employing different metal ions and clusters and by
varying reaction conditions. Here, we describe the details of
the syntheses, characterizations, and properties of MOFs
based on this predesigned carboxylate ligand and its iso-
mers.
differ, and stereoisomers, in which atom connectivities are
identical. The latter encompasses diastereomers and enan-
[7]
tiomers. The study of supramolecular stereoisomerism can
lead potentially to the generation of chiral coordination pol-
ymers from achiral ligands; however, the body of literature
[
8]
in this field is still limited.
Controlled synthesis of MOFs remains a significant chal-
lenge to chemists, because the final structure is influenced
by numerous factors, including ligand geometry, solvent, and
Results and Discussion
[9]
reaction temperature. In general, even seemingly minor
changes in any of these factors can result in the formation
of a different structural topology. Most ligands used in the
study of supramolecular isomerism are derivatives of pyrrole
The tetracarboxylate ligand TCPPDA was synthesized in
two steps through an amination reaction and subsequent hy-
drolysis (Scheme 1). We selected TCPPDA as a target
ligand in the study of supramolecular stereoisomerism and
construction of MOFs primarily for the isomeric possibilities
it presents. TCPPDA has three stereoisomers (a pair of
enantiomers and a diastereomer); the diastereomer has C2h
symmetry with three phenyl rings oriented as left- and right-
handed propellers around the
[5,10]
or pyridine,
carboxylate ligand, the tetra-anion of N,N,N’,N’-tetrakis(4-
carboxyphenyl)-1,4-phenylenediamine (TCPPDA)
Figure 1), which was chosen intentionally to lead to poten-
however, we recently designed a new tetra-
(
two nitrogen atoms. A plane of
symmetry exists through the
central phenyl ring, reflecting
one nitrogen-centered propeller
with respect to the other. All
four carboxylate carbon atoms
lie within the same plane and
define a rectangle; in a coordi-
nation network, the C2h isomer
will be a four-connected unit
similar to a square-planar con-
nector. The pair of enantiomers
possess D2 symmetry. In one
Figure 1. The designed tetracarboxylate ligand TCPPDA.
tial supramolecular isomers.
This ligand has three stereo-
isomers: a pair of enantiomers
(
D2 symmetry) and a diaster-
eomer (C2h symmetry). Despite
our early preliminary work with
this ligand, which revealed tem-
perature-dependent supramo-
[11]
lecular stereoisomerism,
the
potential of this ligand to form Scheme 1. The synthesis of H TCPPDA.
4
novel supramolecular structures
remained largely unexplored;
conceivable structures could contain both C2h and D iso-
case, both nitrogen-centered propellers are right-handed
2
d
mers, or possibly only one type of D isomer (d or l). Fur-
( D ), and in the other case, both nitrogen-centered propel-
2
2
l
thermore, the size, shape, and nonplanarity of TCPPDA
may hinder the formation of interpenetrating networks,
which makes it a good choice for the construction of these
networks.
lers are left-handed ( D ). The conversion of the isomer
2
from C2h to D will take place if the direction of one propel-
2
ler is inverted by rotating two nitrogen–carboxyphenyl
bonds on the same side of the central phenyl ring. The inter-
d
l
In the past decade, a large number of coordination net-
works or MOFs have been constructed from various organic
conversion between the D and D isomers can occur by
2 2
inversion of both propellers, through rotation of all four ni-
trogen–carboxyphenyl bonds. The four carboxylate carbon
atoms occupy the vertices of a stretched tetrahedron; in a
[12]
ligands with different geometries and functionalities; how-
ever, the designed construction of MOFs that exploits ligand
[9]
isomerism is quite rare. To extend our previous work, we
coordination network, the D enantiomers will be four-con-
2
Chem. Eur. J. 2006, 12, 3768 – 3776
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3769

H.-C. Zhou et al.
Figure 2. The coordination modes of TCPPDA in 1–5.
nected units similar to tetrahedral connectors. The energy
barrier for the inversion of a propeller should be low
enough that TCPPDA can exist in all three forms in solution
at room temperature. However, in a coordination network,
it is possible that one of the three isomers will be favored
over the others. If the pseudotetrahedral and rectangular ge-
ometries of the ligand can independently connect a rigid
SBU, supramolecular isomers can be expected; furthermore,
TCPPDA remains protonated, indicated by a CꢀO bond
length of 1.283 Š, and by IR data, in which a strong absorp-
ꢀ
1
tion peak at around 1700 cm for ꢀCOOH is observed. As
shown in Figure 2, each TCPPDA molecule is seven-coordi-
nate in the following manner: two carboxylate groups adopt
bidentate bridging modes linking two cobalt atoms each;
one carboxylate oxygen bridges two cobalt atoms; the proto-
nated carboxylate group coordinates one cobalt atom in a
monodentate coordination mode. Carboxylate groups from
six different TCPPDA ligands connect three cobalt atoms
arranged linearly to form a trinuclear “hourglass” SBU, as
shown in Figure 3. Note that the central cobalt atom is octa-
d
if D -TCPPDA appears in a structure as solely the D or
2
2
l
D form, a chiral MOF will result.
2
With these considerations in mind, we first attempted to
incorporate this ligand into a coordination network with the
copper(ii) ion, due to its known propensity to form paddle-
wheel SBUs with carboxylate ligands. The solvothermal re-
action of H TCPPDA and Cu(NO ) ·2.5H O in dimethyl-
A
H
R
U
G
4
3 2 2
sulfoxide (DMSO) at 1208C resulted in yellow crystals of 1,
Cu (D -tcppda)(H O) ]·2DMSO·6H O, which possesses the
[
ACHTREUNG
2
2
2
2
2
PtS network in which all ligands adopt the pseudotetrahe-
dral geometry (D -TCPPDA). However, the same reaction
2
carried out at 1158C results in burgundy-red crystals of 2,
[
Cu (C -tcppda)(H O) ]·2DMSO·6H O, which possesses
A
C
H
T
R
E
U
N
G
2
2h
2 2 2
the NbO network in which all ligands adopt rectangular ge-
[
13]
ometry (C -TCPPDA). As indicated above, 1 and 2 were
2
h
[
11]
Figure 3. The hourglass SBU in 3. Cobalt atoms are represented as poly-
hedra.
reported previously; they will, therefore, be discussed fur-
ther only in relation to the additional studies described
herein.
Heating a mixture of H TCPPDA and Co(NO ) ·6H O in
A
H
R
U
G
4
3 2 2
diethylformamidine (DEF) to 1208C for four days resulted
in the formation of red-violet crystals of 3, [Co (D -
hedrally coordinated, whereas the two terminal cobalt
atoms are tetrahedrally coordinated. This hourglass SBU is
quite different from another hourglass SBU characterized
previously in our laboratory, in which all the carboxylate
groups adopt a bidentate bridging mode and two coordinat-
3
2
Htcppda) ]·4DEF·5H O. X-ray diffraction analysis revealed
2
2
that 3 has a three-dimensional porous network containing a
trinuclear cobalt cluster. One of the carboxylate groups of
3770
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3768 – 3776

Construction of Open Metal–Organic Frameworks
FULL PAPER
[
14]
ed water molecules occupy the terminal positions.
The
noninterpenetrated porous framework. The lack of a strong
ꢀ
1
average Co–O distance is 2.037(6) Š. Each of the TCPPDA
ligands in 3 has D symmetry and is, therefore, chiral. How-
ever, because the ratio of D - to D -TCPPDA in this net-
work is 1:1, the coordination network is racemic overall.
IR absorption peak at 1700 cm (for ꢀCOOH) indicates
that all carboxylates are deprotonated during the reaction
and formation of the network. Within the structure of 4,
there are 1.5 crystallographically independent ligands, and
2
d
l
2
2
The D -TCPPDA in 3 can be considered to be an elongat-
two different geometries of the ligand are present, C - and
2
2h
ed tetrahedral connector: the dihedral angle between the
two planes, each defined by a nitrogen atom and the accom-
panying carboxylate carbon atoms, is 74.58. Each TCPPDA
ligand links four hourglass SBUs, and every SBU can be
considered as an eight-connected node. Thus, the TCPPDA
ligand connects hourglass SBUs to generate an eight-con-
nected network structure with open channels (Figure 4). The
D -TCPPDA. Both TCPPDA ligands have similar coordina-
2
tion modes: two carboxylate groups adopt a bidentate che-
lating mode, chelating one neodymium atom each, and the
other two carboxylate groups adopt a chelating-bridging
mode, connecting two neodymium atoms each, as shown in
Figure 2. As in 3, D -TCPPDA forms an elongated tetrahe-
2
dral connector, with a dihedral angle between the two car-
boxylate-nitrogen-carboxylate planes of 63.68. In contrast,
all the carboxylate carbon atoms are coplanar in C -
2h
TCPPDA. As in 3, the coordination network is racemic
d
l
overall, due to a 1:1 ratio of the D - to D -TCPPDA enan-
2
2
tiomers.
In 4, two crystallographically independent neodymium
atoms, Nd1and Nd2, are connected by one TCPPDA ligand
to form a binuclear Nd unit. Both neodymium atoms adopt
2
a nine-coordinate geometry: Nd1is coordinated by two
oxygen atoms from C -TCPPDA, six from D -TCPPDA,
2h
2
and one from a coordinated DMSO molecule; Nd2 is coor-
dinated by three oxygen atoms from C -TCPPDA, four
2h
from D -TCPPDA, and two from coordinated DMSO. Each
2
C -TCPPDA connects four Nd2 atoms to generate a one-di-
2h
mensional chain, in which the Nd2···Nd2 distance is
3.440 Š, as shown in Figure 5. Each D -TCPPDA links four
1
2
Figure 4. a) The three-dimensional porous framework of 3 viewed along
the b axis. b) A schematic representation of 3 simplifying the organic
ligand as a pseudotetrahedral connector and the hourglass SBU as an
eight-connected node.
dimensions of the channels are 7.149”10.832 Š, in which
free solvates reside. The structure is noninterpenetrated
with a total potential solvent-accessible volume of 58%
Figure 5. The one-dimensional chain in 4 connected by C2h-TCPPDA.
3
3
[15]
(
1519.1 Š per 2619.6 Š ), as calculated by PLATON.
In 1, 2, and 3, all the TCPPDA ligands adopt only a single
Nd1atoms to give rise to a two-dimensional layer frame-
work, in which the Nd1···Nd1 distance is 13.109 Š
(Figure 6). The one-dimensional chains and two-dimensional
layers are further connected by the carboxylate groups of
C - or D -TCPPDA in a chelating or bridging mode to give
geometry, either pseudotetrahedral or rectangular. This is
probably a result of the coordination geometry imposed by
the first-row transition-metal ions, which typically adopt a
low coordination number. The solvothermal reaction of
2h
2
H TCPPDA and Nd
A
H
R
U
G
rise to a three-dimensional, noninterpenetrated porous
framework, as shown in Figure 7. The Nd1···Nd2 distance
along the a axis is 3.850 Š. The dimensions of the channels
are 10.558”19.316 Š, in which free solvates reside. The sol-
4
3
3
2
48 h resulted in light-yellow rodlike crystals of 4, [Nd (D -
tcppda)(C -tcppda) (DMSO) ]·3DMSO·5H O. X-ray dif-
2h
0.5 3 2
fraction analysis reveals that 4 also has a three-dimensional,
Chem. Eur. J. 2006, 12, 3768 – 3776
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3771

H.-C. Zhou et al.
1
, 3, and 4 all have a racemic network, due to the 1:1 ratio
d
l
of D - to D -TCPPDA within the network. However, if a
2
2
synthesis results in all the D -TCPPDA ligands appearing as
2
d
l
either the D or the D form, then the overall network will
2
2
be chiral rather than achiral. The solvothermal reaction of
Zn(NO ) ·6H O and H TCPPDA in dimethylformamide
DMF) at 1208C results in the formation of red block crys-
tals of 5, [Zn O(D -tcppda) ]·DMF·H O. Although the bulk
AHCTREUNG
3
2
2
4
(
4
2
1.5
2
sample of 5 is racemic, individual crystals self-assemble into
separate chiral enantiomeric structures with ligands in the
d
l
form of either all D or all D .
2
2
All the carboxylate groups of TCPPDA are deprotonated
during the reaction, with each carboxylate connecting two
zinc atoms in a bridging mode. Four zinc atoms are linked
by six carboxylate groups from different TCPPDA ligands
and one m -oxygen atom to form a Zn O(OOCR) unit, a
A
H
R
U
G
4
4
6
basic carboxylate octahedral SBU (Figure 8a). Two crystal-
2
Figure 6. The two-dimensional layer in 4 connected by D -TCPPDA. a)
Viewed along the c axis. b) Viewed along the a axis. c) and d) Schematic
representations of the same views, simplifying the organic ligand as a
pseudotetrahedral connector.
Figure 8. a) The basic carboxylate octahedral SBU, Zn
4
A
H
R
U
G
6
O(OOCR) , in 5.
b) Each TCPPDA ligand connects four octahedral SBUs.
lographically independent D -TCPPDA units appear in the
2
structure; the carboxylate-nitrogen-carboxylate dihedral
angles are 77.1and 85 8. Each TCPPDA ligand connects four
octahedral SBUs (Figure 8b) and each octahedral SBU at-
taches to six TCPPDA ligands to give a three-dimensional
porous framework with two kinds of channels (Figure 9).
The dimensions of the channels are 7.007”7.007 Š (A) and
9
.510”9.510 Š (B), in which DMF and water solvates
reside. The structure is noninterpenetrating, due to the non-
planarity of the ligand and the rigidity of the SBU, and has
3
a total solvent-accessible volume of 74.7% (42068.2 Š per
3
[15]
5
6342.0 Š ), as calculated by PLATON.
Although it is difficult to combine tetrahedral organic li-
gands and octahedral metal SBUs in the construction of
MOFs, it has been shown that the best compromise between
Figure 7. The three-dimensional porous framework of 4 viewed along he
a axis. a) The three-dimensional porous framework of 4 viewed along the
a axis. b) A schematic representation of 4, simplifying the organic ligand
as pseudotetrahedral and rectangular connectors.
these two forms is found in the corundum form of Al O (a-
2
3
alumina), in which both aluminum and oxide ions are dis-
[16,17]
torted from their ideal geometries.
To further under-
stand the topology of 5 as an analogue of the corundum top-
ology, simplified models of the ligand and the SBU cluster
can be considered: as discussed above, the organic ligand
3
3
vent-accessible volume is 67.9% (4059.0 Š per 5977.3 Š ),
calculated by PLATON.
[15]
3772
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3768 – 3776

Construction of Open Metal–Organic Frameworks
FULL PAPER
dum topology constructed from a D -symmetric pyridine-
2d
based ligand (3,3’,5,5’-tetrakis(4-pyridyl)bimesityl) and a
[18]
single metal ion.
The achiral zinc and cadmium frame-
works formed possess lower porosity (44 and 46%, respec-
tively) than 5, due to the presence of several methyl groups
in the 3,3’,5,5’-tetrakis(4-pyridyl)bimesityl ligand, which can
protrude into the channels. The considerably greater porosi-
ty found in the TCPPDA-based framework 5 can be attrib-
uted to the lack of these methyl groups and the larger over-
all size of the ligand.
Unlike frameworks 1 and 3, all the D -TCPPDA ligands
2
d
l
in 5 appear as either the D or the D form, thus, the
2
2
whole structure is enantiopure and homochiral; the crystals
have the I4 32 space group. To the best of our knowledge,
1
complex 5 represents the first chiral MOF possessing corun-
dum topology. Preliminary second harmonic generation
(
SHG) measurements showed that 5 displays a response in
the powder state that is three times greater than that of
urea; this is a slightly stronger SHG response than that of
some chiral MOFs constructed from chiral organic li-
Figure 9. The three-dimensional porous framework of 5. The two unique
types of channels are labeled A and B. Zinc atoms are represented as tet-
rahedra.
[17,19]
gands.
The results of thermogravimetric analysis (TGA) for com-
plexes 3–5 measured under N are shown in Figure 11. For
2
can be modeled as a tetrahedral connector, and the SBU
cluster can be modeled as an octahedral node with the m4-
oxygen atom at the center. The resulting corundum topology
is shown in Figure 10. Recently, Moorthy et al. reported zinc
and cadmium coordination networks possessing the corun-
Figure 11. TGA plot of 3, 4, and 5.
3
, the weight loss of 27% from 50 to 3508C corresponds to
the loss of four free DEF molecules and five free water mol-
ecules (calcd: 26.8%). For 4, the weight loss of 30% from
5
0 to 2458C corresponds to the loss of three free DMSO
molecules, five free water molecules, and three coordinated
DMSO molecules (calcd: 32%). There is no further weight
loss from 245 to 4008C. TGA shows that the framework of 5
is stable up to 4108C: from 50 to 3008C, there is a gradual
weight loss of 7.7%, corresponding to the loss of one DMF
and one water solvate trapped within the lattice (calcd:
7
.3%), with no further weight loss from 300 to 4108C.
Figure 10. a) View of the network of
5
along [111]; b) space-filling
Gas sorption measurements for N2 and H2 were per-
model: in both (a) and (b), zinc atoms are indicated in aqua, nitrogen
atoms in blue, carbon atoms in gray, and oxygen atoms in red; c) 5 is sim-
plified as the corundum network by modeling the ligand as a tetrahedral
connector (gray) and the octahedral SBU as a six-connected node
formed on compounds 1, 2, and 5, and are shown in Fig-
ures 12 and 13, respectively. Complex 4 was unable to retain
its framework after solvate removal. As-synthesized samples
of 1 and 2 were exchanged with methanol to remove DMSO
(
aqua).
Chem. Eur. J. 2006, 12, 3768 – 3776
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3773

H.-C. Zhou et al.
highly porous MOFs with open metal sites for hydrogen
storage has, thus, become a target for future work.
Photoluminescence measurements for 4 and 5 at room
temperature show that these MOFs exhibit strong lumines-
cence at lmax =425 and 423 nm, respectively, upon excitation
at 276 nm (Figure 14). The emissions of complexes 4 and 5
2
Figure 12. N gas sorption isotherms of 1, 2, and 5 measured at 78 K.
Figure 14. Emission spectra of 4 (lmax =425 nm), 5 (lmax =423 nm), and
TCPPDA ligand (lmax =498 nm) in DMF solutions, upon excitation at
2
76 nm.
can be assigned to an intraligand p*!p transition, because
the free TCPPDA possesses similar emission (l =498 nm)
max
in solution, although a sizable blue-shift (approx. 75 nm) is
observed in the complexes. Photoluminescent emission by 4
and 5 is considerably more intense than that of the free
ligand; this may be explained in terms of ligand rigidity. In
solution at room temperature, TCPPDA exists in all three
forms and is much more flexible than the forms within a
MOF; this rigidity of the coordinated ligand effectively re-
duces the loss of energy, thereby increasing fluorescent effi-
2
Figure 13. H gas sorption isotherms of 1, 2, and 5 measured at 78 K.
and water molecules, then dried under vacuum overnight.
N sorption at 77 K (Figure 12) shows that complexes 1 and
2
2
possess Type I isotherms with Langmuir surface areas of
2
ꢀ1
[11]
[23]
626 and 504 m g , respectively. The hydrogen-storage ca-
ciency.
ꢀ
1
ꢀ1
pabilities of both compounds (14 mgg for 1, 1 2 mgg for
) are comparable to reported values for “isoreticular”
2
[20]
MOFs with high reported surface areas. The difference in
the nitrogen- and hydrogen-storage capacities of the two iso-
mers is attributed to the variation in porosity, consistent
with structures and calculated potential solvent-accessible
volumes. An as-synthesized sample of 5 was exchanged with
chloroform to remove DMF and water solvates, then dried
Conclusion
We have solvothermally synthesized five novel porous
metal–organic frameworks based on a rationally designed
tetracarboxylate ligand, TCPPDA, which exhibits stereoiso-
merism. This ligand can adopt two different geometries:
pseudotetrahedral (D -TCPPDA) and rectangular (C -
under vacuum overnight. N sorption at 77 K shows that 5
2
2
2h
also has a Type I isotherm with a Langmuir surface area of
TCPPDA). By varying the reaction temperature and/or
metal species, porous frameworks containing either a single
geometry (1, 2, 3, and 5) or mixed geometries (4) of the
ligand can be successfully isolated. Most interestingly, a
2
ꢀ1
2095 m g , indicating that 5 possesses permanent porosity.
Although 5 possesses much higher porosity than 1 and 2, hy-
drogen sorption measurement showed that 5 has very low
ꢀ
1
hydrogen-storage capability (8 mgg ). The higher hydro-
gen-storage capacity of 1 and 2 may be attributable to the
existence of open metal sites in these networks after remov-
al of axially coordinated water molecules in the paddlewheel
SBUs; these open metal sites may have a high affinity for
chiral porous framework possessing the corundum topology
d
can be constructed in a zinc complex (5), in which only D -
2
l
or D -TCPPDA are present in each crystal. These results
2
demonstrate that a pseudorigid ligand that can adopt differ-
ent geometries has significant advantages over a rigid one in
the construction of coordination networks: such a ligand can
adopt different geometries, according to the reaction condi-
tions, to meet the coordination requirements of the metal
[21]
H molecules, and are absent in 5. This result also indi-
2
cates that the degree of porosity is not the sole factor in de-
[22]
termining gas adsorptivity.
The design and synthesis of
3774
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3768 – 3776

Construction of Open Metal–Organic Frameworks
FULL PAPER
ion or metal-containing SBU. This provides a new strategy
in the synthesis of MOFs with various structural topologies.
In the future, the design and synthesis of additional organic
ligands that have this advantageous mix of structural rigidity
and conformational flexibility will be a powerful tool in the
construction of MOFs with desirable topologies and func-
tionalities.
N
2
at 213 K. Frames were collected with 0.38 intervals in f and w for 30 s
per frame, such that a hemisphere of data was collected. Raw data collec-
tion and cell refinement were achieved by using SMART; data reduction
was performed by using SAINT+ and corrected for Lorentz and polari-
[24]
zation effects.
Absorption corrections were applied by using the
SADABS routine. Structures were solved by direct methods by using
SHELXTL and were refined by full-matrix least-squares on F by using
SHELX-97.
placement parameters during the final cycles. Hydrogen atoms were
placed in calculated positions with isotropic displacement parameters set
to 1.2”Ueq of the attached atom. The solvent molecules in 3, 4, and 5 are
highly disordered, and attempts to locate and refine the solvent peaks
were unsuccessful. Contributions from these solvent molecules to scatter-
ing were removed by using the SQUEEZE routine of PLATON; struc-
2
[
25]
Non-hydrogen atoms were refined with anisotropic dis-
Experimental Section
[
15]
tures were then refined again by using the data generated.
Compound 3: 0.21”0.20”0.15 mm, dark red-violet
C H Co N O , M =1347.85 gmol triclinic, P1, a=13.703(3) b=
14.120(3) c=15.810(4) Š, a=105.069(4) b=103.355(4) g=109.052(5)8,
All starting materials were obtained commercially and were used without
further purification. Thermal gravimetric analyses (TGA) were per-
block,
ꢀ
1
¯
formed under N
2
by using a Perkin–Elmer TGA 7 analyzer. A Beckman
,
6
8
42
3
2
16
r
Coulter SA 3100 surface-area analyzer was used to measure gas adsorp-
tion. Photoluminescence spectra were obtained by using a Perkin–Elmer
LS 50B luminescence spectrometer. Solution NMR spectra were record-
ed by using a Bruker 200 MHz spectrometer. Elemental analyses (C, H,
N) were performed by Canadian Microanalytical Services.
3
ꢀ3
ꢀ1
V=2619.6(10) Š , Z=1, 1calcd =0.854 gcm
687, total reflections=7120, independent reflections=6303, R =0.0515,
,
m=0.513 mm
,
F(000)=
A
C
H
T
R
E
U
N
G
1
wR =0.1325, (R =0.1059, wR =0.1707 before SQUEEZE), GOF=
2
1
2
0.801, residuals based on [I>2s(I)], D1max,min =0.279, ꢀ0.468 eŠꢀ
3
.
Synthesis of L1: (Scheme 1) 1,4-benzenediamine (0.5 g, 4.6 mmol), tert-
butyl-4-bromobenzoate (5.0 g, 19 mmol), palladium acetate (45 mg,
Compound 4: 0.2”0.2”0.2 mm, light-yellow block, C H N Nd O S ,
57 48 3 2 15 3
¯
triclinic, P1, a=15.156(3) b=15.657(2) c=
ꢀ
1
M
r
=1399.68 gmol
26.967(4) Š, a=81.603(16)
5977.3(15) Š , Z=1, 1
,
0
.20 mmol), tri-tert-butyl-phosphine (32 mg, 0.33 mmol), sodium tert-but-
oxide (2.3 g, 24 mmol), and toluene (20 mL) were mixed in an N drybox.
The reaction mixture was removed from the drybox and heated under N
b=78.042(12)
=0.389 gcm , 1=0.472 mm , F
g=73.529(12)8,
V=
3
ꢀ3
ꢀ1
2
calcd
ACHTREUNG
2
total reflections=30569, independent reflections=25476,
wR =0.1837, (R =0.1762, wR
0.749, residuals based on [I>2s(I)], D1max,min =2.533, ꢀ2.866 eŠ
Compound 5: 0.48”0.41” 0.60 mm, red block, C102
R
1
at 758C for 24 h with constant stirring. The resulting brown mixture was
poured into a saturated aqueous solution of ammonium chloride, and the
aqueous layer was extracted with toluene. The combined organic phases
2
1
2
ꢀ
3
.
60 6 8 r
H N O26Zn , M =
were dried with anhydrous MgSO
4
. Crude product was concentrated and
ꢀ1
2
5
308.71gmol , cubic, I4
6342(2) Š , Z=8,
1
32, a=b=c=38.3364(9) Š, a=b=g=908, V=
purified by flash chromatography to give L1 as a light yellow solid (1.8 g,
3
ꢀ3
ꢀ1
1
calcd =0.544 gcm
,
m=0.697 mm
,
F
ACHTREUNG
1
4
(
8%). H NMR (200 MHz, CD
3
Cl): d=1.62 (s, 36H), 7.08 (s, 4H), 7.12
total reflections=77336, independent reflections=3273,
wR
R
1
1
3
d, 8H), 7.95 ppm (d, 8H); C NMR (200 MHz, CD Cl): d=28.66, 81.22,
3
2
=0.0624, (R
1
=0.1399, wR
2
=0.2763 before SQUEEZE), GOF=
1
22.92, 126.69, 131.32, 150.88, 165.75 ppm.
ꢀ3
0
.793, residuals based on [I>2s(I)], D1max,min =0.075, ꢀ0.145 eŠ , Flack
Synthesis of H
nylenediamine): L1 (0.25 g) was dissolved in CH
4
TCPPDA (N,N,N’,N’-tetrakis(4-carboxyphenyl)-1,4-phe-
Cl (20 mL), to which
parameter 0.09(2).
2
2
CCDC-287825, CCDC-287826, and CCDC-270804 contain the supple-
mentary crystallographic 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.
trifluoroacetic acid (0.5 mL) was added; the resulting yellow solution was
allowed to stir at RT overnight. The yellow precipitate that formed was
filtered, washed with CH
TCPPDA as yellow solid (0.175 g, 96%). H NMR (200 MHz,
CD Cl) d=7.08 (s, 4H), 7.12 (d, 8H), 7.95 ppm (d, 8H).
Synthesis of 3: A mixture of [Co(en) ]I ·H O (15 mg, 0.024 mmol, en=
ethylene diamine) and H TCPPDA (2.5 mg, 0.043 mmol) in N,N-diethyl-
2 2
Cl , and dried under vacuum to give
1
H
4
a
3
3
3
2
4
Acknowledgements
formamide (DEF) (1.5 mL) was sealed in a thick-walled glass tube under
vacuum. The tube was heated at 1208C for 5 d, then cooled to RT at a
ꢀ
1
This work was supported by the National Science Foundation (CHE-
rate of 0.18Cmin . The resulting red-violet crystals were washed with
DEF to give 3. Composition of 3 was determined by crystal structure
analysis and TGA.
0
449634), Miami University, and the donors of the American Chemical
Society Petroleum Research Fund. H.C.Z. also acknowledges the Re-
search Corporation for a Research Innovation Award and a Cottrell
Scholar Award. The diffractometer was funded by NSF grant EAR-
Synthesis of 4: A mixture of Nd
TCPPDA (2.5 mg, 0.043 mmol) in dimethylsulfoxide (DMSO)
1.0 mL) and pyridine (0.5 mL) was sealed in a thick-walled glass tube
under vacuum. The tube was heated at 908C for 24 h, then cooled to RT
A
H
R
U
G
3 3 2
(NO ) ·6H O (20 mg, 0.046 mmol) and
H
4
0
003201.
(
ꢀ
1
at a rate of 0.18Cmin . The resulting yellow rod crystals were washed
with DMSO to give 4. Elemental analysis calcd (%) for 4: C 43.89, H
[
[
1] T. L. Hennigar, D. C. MacQuarrie, P. Losier, R. D. Rogers, M. J. Za-
worotko, Angew. Chem. 1997, 109, 1044–1046; Angew. Chem. Int.
Ed. Engl. 1997, 36, 972–973.
4
.44, N 2.44; found: C 43.63, H 3.91, N 2.44.
Synthesis of 5: A mixture of Zn(NO ·6H
TCPPDA (2.5 mg, 0.043 mmol) in N,N-dimethylformamide (DMF)
1.5 mL) was sealed in a thick-walled glass tube under vacuum. The tube
A
H
R
U
G
3
)
2
2
O (20 mg, 0.067 mmol) and
2] a) P. J. Hagrman, D. Hagrman, J. Zubieta, Angew. Chem. 1999, 111,
H
(
4
2
798–2848; Angew. Chem. Int. Ed. 1999, 38, 2639–2684; b) S. Kita-
gawa, R. Kitaura, S.-I. Noro, Angew. Chem. 2004, 116, 2388–2430;
Angew. Chem. Int. Ed. 2004, 43, 2334–2375; c) E. Lee, J. Heo, K.
Kim, Angew. Chem. 2000, 112, 2811–2813; Angew. Chem. Int. Ed.
ꢀ
1
was heated to 1208C for 24 h and cooled to RT at a rate of 0.18Cmin
.
The resulting red block crystals were washed with DMF to give 5
2.5 mg, 71%). Elemental analysis calcd (%) for 5: C 52.08, H 3.16, N
.50; found: C 51.49, H 3.15, N 4.79.
(
4
2
000, 39, 2699–2701; d) T. J. Prior, D. Bradshaw, S. J. Teat, M. J.
Rosseinsky, Chem. Commun. 2003, 500–501.
Crystal structure determination: Single-crystal X-ray diffraction was per-
formed by using a Bruker Apex D8 CCD diffractometer equipped with a
fine-focus sealed-tube X-ray source (MoKa radiation, l=0.71073 Š,
graphite monochromated) operating at 45 kV and 40 mA. Crystals of 3,
[3] a) L. Carlucci, G. Ciani, D. M. Proserpio, A. Sironi, Angew. Chem.
1995, 107, 2037–2040; Angew. Chem. Int. Ed. Engl. 1995, 34, 1895–
1898; b) S. R. Batten, B. F. Hoskins, R. Robson, J. Chem. Soc.
Chem. Commun. 1991, 445–447; c) R. W. Gable, B. F. Hoskins, R.
Robson, J. Chem. Soc. Chem. Commun. 1990, 762–763; d) B. F.
4
, and 5 were mounted on glass fibers and maintained under a stream of
Chem. Eur. J. 2006, 12, 3768 – 3776
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3775

H.-C. Zhou et al.
Hoskins, R. Robson, N. V. Y. Scarlett, Angew. Chem. 1995, 107,
317–1318; Angew. Chem. Int. Ed. Engl. 1995, 34, 1203–1204.
4] D.-L. Long, A. J. Blake, N. R. Champness, C. Wilson, M. Schroder,
[13] a) J. C. Stephens, M. A. Khan, R. P. Houser, Inorg. Chem. 2001, 40,
5064–5065; b) M. Eddaoudi, J. Kim, M. OꢁKeeffe, O. M. Yaghi, J.
Am. Chem. Soc. 2002, 124, 376–377.
1
[
[
[
[
Chem. Eur. J. 2002, 8, 2026–2033.
5] L. R. MacGillivray, J. L. Reid, J. A. Ripmeester, Chem. Commun.
[14] Y. Ke, D. J. Collins, D. Sun, H.-C. Zhou, Inorg. Chem. 2006, ASAP
ic051900p.
[15] A. L. Spek, J. Appl. Crystallogr. 2003, 36, 7–1 3.
[16] M. OꢁKeeffe, M. Eddaoudi, H. Li, T. Reineke, O. M. Yaghi, J. Solid
State Chem. 2000, 152, 3–20.
[17] Z.-R. Qu, H. Zhao, Y.-P. Wang, X.-S. Wang, Q. Ye, Y.-H. Li, R.-G.
Xiong, B. F. Abrahams, Z.-G. Liu, Z.-L. Xue, X.-Z. You, Chem. Eur.
J. 2004, 10, 53–60.
[18] R. Natarajan, G. Savitha, P. Dominiak, K. Wozniak, J. N. Moorthy,
Angew. Chem. 2005, 117, 2153–2157; Angew. Chem. Int. Ed. 2005,
44, 2115–2119.
2
001, 1034–1035.
6] H. Abourahma, B. Moulton, V. Kravtsov, M. J. Zaworotko, J. Am.
Chem. Soc. 2002, 124, 9990–9991.
7] a) A. J. Blake, N. R. Brooks, N. R. Champness, M. Crew, D. H. Greg-
ory, P. Hubberstey, M. Schroder, A. Deveson, D. Fenske, L. R.
Hanton, Chem. Commun. 2001, 1432–1433; b) E.-Q. Gao, Z.-M.
Wang, C.-S. Liao, C.-H. Yan, New J. Chem. 2002, 26, 1096–1098.
8] a) W. Bi, R. Cao, D. Sun, D. Yuan, X. Li, Y. Wang, X. Li, M. Hong,
Chem. Commun. 2004, 2104–2105; b) Y. Kim, D.-Y. Jung, Chem.
Commun. 2002, 908–909.
[
[
[19] S. P. Anthony, T. P. Radhakrishnan, Chem. Commun. 2004, 1058–
1059.
[20] J. L. C. Rowsell, A. R. Millward, K. S. Park, O. M. Yaghi, J. Am.
Chem. Soc. 2004, 126, 5666–5667.
9] a) E.-Y. Choi, K. Park, C.-M. Yang, H. Kim, J.-H. Son, S. W. Lee,
Y. H. Lee, D. Min, Y.-U. Kwon, Chem. Eur. J. 2004, 10, 5535–5540;
b) S. R. Batten, J. Solid State Chem. 2005, 178, 2475–2479.
[
10] a) I. S. Lee, D. M. Shin, Y. K. Chung, Chem. Eur. J. 2004, 10, 3158–
165; b) C.-Y. Su, A. M. Goforth, M. D. Smith, H.-C. zur Loye,
[21] B. Chen, N. W. Ockwig, A. R. Millward, D. S. Contreras, O. M.
Yaghi, Angew. Chem. 2005, 117, 4823–4827; Angew. Chem. Int. Ed.
2005, 44, 4745–4749.
[22] a) L. Pan, M. B. Sander, X. Huang, J. Li, M. Smith, E. Bittner, B.
Bockrath, J. K. Johnson, J. Am. Chem. Soc. 2004, 126, 1308–1309;
b) J. L. C. Rowsell, O. M. Yaghi, Angew. Chem. 2005, 117, 4748–
4758; Angew. Chem. Int. Ed. 2005, 44, 4670–4679.
[23] L.-Z. Cai, W.-T. Chen, M.-S. Wang, G.-C. Guo, J.-S. Huang, Inorg.
Chem. Commun. 2004, 7, 611–613.
3
Inorg. Chem. 2003, 42, 5685–5692; c) J.-P. Zhang, Y.-Y. Lin, X.-C.
Huang, X.-M. Chen, Chem. Commun. 2005, 1258–1260; d) N. Mas-
ciocchi, S. Bruni, E. Cariati, F. Cariati, S. Galli, A. Sironi, Inorg.
Chem. 2001, 40, 5897–5905; e) I. Boldog, J. Sieler, K. V. Domase-
vitch, Inorg. Chem. Commun. 2003, 6, 769–772.
[
11] D. Sun, Y. Ke, T. M. Mattox, B. A. Ooro, H.-C. Zhou, Chem.
Commun. 2005, 5447–5449.
[
12] a) C. Janiak, Dalton Trans. 2003, 2781–2804; b) C. N. R. Rao, S. Na-
tarajan, R. Vaidhyanathan, Angew. Chem. 2004, 116, 1490–1521;
Angew. Chem. Int. Ed. 2004, 43, 1466–1496; c) M. Eddaoudi, D. B.
Moler, H. Li, B. Chen, T. M. Reineke, M. OꢁKeeffe, O. M. Yaghi,
Acc. Chem. Res. 2001, 34, 319–330.
[24] SAINT+, Bruker Analytical X-Ray Systems, Madison, WI, 2001.
[25] G. M. Sheldrick, SHELX-97, Bruker Analytical X-Ray Systems,
Madison, WI, 1997.
Received: October 28, 2005
Published online: February 23, 2006
3776
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3768 – 3776