Flexibility and sorption selectivity in rigid metal-organic frameworks: The impact of ether-functionalised linkers

Article abstract of DOI:10.1002/chem.201002341

The functionalisation of well-known rigid metal-organic frameworks (namely, [Zn₄O(bdc)₃]_n, MOF-5, IRMOF-1 and [Zn ₂(bdc)₂(dabco)]_n; bdc=1,4-benzene dicarboxylate, dabco=diazabicyclo[2.2.2]octane) with additional alkyl ether groups of the type -O-(CH₂)_n-O-CH₃ (n = 2-4) initiates unexpected structural flexibility, as well as high sorption selectivity towards CO₂ over N₂ and CH₄ in the porous materials. These novel materials respond to the presence/absence of guest molecules with structural transformations. We found that the chain length of the alkyl ether groups and the substitution pattern of the bdc-type linker have a major impact on structural flexibility and sorption selectivity. Remarkably, our results show that a high crystalline order of the activated material is not a prerequisite to achieve significant porosity and high sorption selectivity. Just breathe! The implementation of linkers functionalised with flexible ether groups in known, but rigid metal-organic frameworks (MOFs), yields isoreticular frameworks that exhibit unexpected responsiveness towards small molecules and high sorption selectivity for CO₂ (see figure). Copyright

Full text of DOI:10.1002/chem.201002341

DOI: 10.1002/chem.201002341
Flexibility and Sorption Selectivity in Rigid Metal–Organic Frameworks:
The Impact of Ether-Functionalised Linkers
Sebastian Henke,[a] Rochus Schmid,[a] Jan-Dierk Grunwaldt,[b] and Roland A. Fischer*[a]
14296
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 14296 – 14306

FULL PAPER
Abstract: The functionalisation of well-
known rigid metal–organic frameworks
wards CO2 over N2 and CH4 in the
porous materials. These novel materials
respond to the presence/absence of
guest molecules with structural trans-
formations. We found that the chain
length of the alkyl ether groups and
the substitution pattern of the bdc-type
linker have a major impact on structur-
al flexibility and sorption selectivity.
Remarkably, our results show that a
high crystalline order of the activated
material is not a prerequisite to achieve
significant porosity and high sorption
selectivity.
(namely,
IRMOF-1 and [Zn
[Zn4O
G
MOF-5,
(dabco)]n;
U
bdc=1,4-benzene
dicarboxylate,
dabco=diazabicycloACTHNUTRGNEUGN[2.2.2]octane) with
Keywords: carbon dioxide separa-
tion · metal–organic frameworks ·
polymers · selective sorption · struc-
tural flexibility
additional alkyl ether groups of the
type -O-(CH2)n-O-CH3 (n = 2–4) ini-
tiates unexpected structural flexibility,
as well as high sorption selectivity to-
Introduction
(NH2-bdc=2-amino-1,4-benzene dicarboxylate; dabco=
diazabicyclo
[2.2.2]octane) initiated a breathing behaviour.[13]
Furthermore, the implementation of a special pillar linker,
featuring flexible 2-hydroxyethoxy groups, in [Cd2A(pzdc)2L]n
(pzdc=2,3-pyrazinedicarboxylate, L=2,5-bis(2-hydroxye-
AHCTUNGTRENNUNG
Metal–organic frameworks (MOFs) are an emerging class of
crystalline porous hybrid materials consisting of inorganic
building blocks (e.g., metal clusters or metal–oxo clusters),
which are interconnected by organic sub-units, the so-called
linkers.[1,2] Their high specific pore volumes and surface
areas together with tailored chemical functionality are inter-
esting for a huge variety of applications, such as gas stor-
age,[1,3] sensing,[4] separation,[5] catalysis[6] and medical appli-
cations.[7] In general, two families of MOF materials can be
distinguished. 1) Firstly, there are many MOFs that feature
robust and rigid networks. The full retention of the frame-
work structure after desorption of solvent molecules (activa-
tion) and subsequent reversible adsorption/desorption cycles
of other guests is a highly desired property.[1,3,8] 2) Secondly,
there are soft MOFs, which exhibit pronounced framework
dynamics. This feature can be associated with responsiveness
to adsorption of guest molecules with a reversible change in
lattice parameters, space group symmetry or other structural
and physical properties, such as spin states of metal ions.[9–11]
Most known flexible MOFs are layer-based systems, in
which two-dimensional layers of coordination polymers are
interconnected in the third dimension with a bidentate pillar
linker. However, flexible and responsive MOFs are relative-
ly rare in the literature compared with the wide variety of
known rigid frameworks. Recently, linker functionalisation
was shown to influence and trigger responsiveness in
MOFs.[12] In particular, the post-synthetic linker functionali-
CTHUNGTRENNUNG
thoxy)-1,4-bis(4-pyridyl)benzene) initiated high sorption se-
lectivity for small polar molecules.[14]
Herein, we present our related results of turning known
rigid and non-selective MOFs into responsive isoreticular
derivatives. Ultra-high sorption selectivity towards CO2 over
N2 and CH4 was initiated by choosing flexible ether side
groups of the type -O-(CH2)n-O-CH3 (n=2, 3 or 4), which
are covalently bound to the phenyl ring of the dicarboxylate
linker. We found that both the alkyl chain length and the
substitution pattern at the phenyl ring have a major impact
on flexibility as well as sorption selectivity of the MOFs. To
demonstrate the general applicability of our concept, we
have chosen two well-known MOFs, namely, [Zn4O
ACHTUNGTRENNUNG
(also named MOF-5, IRMOF-1)[8] and [Zn
CHTUNGTRENNUNG
[15]
G
(bdc=1,4-benzene dicarboxylate) as examples.
The novel ether-functionalised bdc-type linkers 2,5-bis(2-
methoxyethoxy)benzene dicarboxylic acid (H2L1), 2,5-bis(3-
methoxypropoxy)benzene dicarboxylic acid (H2L2), 2,5-
bis(4-methoxybutoxy)benzene dicarboxylic acid (H2L3) and
2-(2-methoxyethoxy)benzene dicarboxylic acid (H2L4) were
synthesised by Mitsunobu etherification[16] and then utilised
in the MOF syntheses instead of conventional H2bdc.
sation of the layer-based MOF [Zn2ACHTUNGTRNENUG(NH2-bdc)2CAHTUNGTREN(NUGN dabco)]n
[a] S. Henke, Dr. R. Schmid, Prof. Dr. R. A. Fischer
Chair of Inorganic Chemistry II
Organometallics and Materials Chemistry
Ruhr-University Bochum, Universitꢁtsstrasse 150
44780 Bochum (Germany)
Fax : (+49)234-32-14174
E-mail: roland.fischer@rub.de
[b] Prof. Dr. J.-D. Grunwaldt
Chair for Chemical Technology and Catalysis
Institute for Technical Chemistry and Polymer Chemistry
Karlsruhe Institute of Technology
Engesserstrasse 20, 76131 Karlsruhe (Germany)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201002341.
Chem. Eur. J. 2010, 16, 14296 – 14306
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14297

R. A. Fischer et al.
Results and Discussion
Syntheses and characterisation:
The ether-functionalised coordi-
nation polymers [Zn4O(LX)3-
A
and
[Zn2(LX)2-
A
ACHTUNGTRENNUNG
and 4) were synthesised in a
solvothermal synthesis from
ZnACHTUNGTRENNUNG(NO3)2·xH2O (x=4 and 6)
and the organic linkers in DMF
(Figure 1). The implemented
alkyl ether side-chain substitu-
ents are flexible and feature a
polar methoxy head group.
Hence, the functional groups
introduce ether solvent like
properties, but are fixed at the
backbone of the porous parent
network. The structural analogy
Figure 1. Solvothermal syntheses of ether-functionalised MOFs. Four different linkers, H2LX, were utilised in
MOF syntheses to achieve eight different ether-functionalised frameworks [Zn2(LX)
2ACHTUNTRGNE(NUG dabco)]n and
[Zn4O(LX)3]n (X=1, 2, 3 and 4). Zinc, oxygen, nitrogen and carbon atoms are shown in yellow, red, blue and
grey, respectively (hydrogen atoms and disordered solvent molecules were omitted for clarity).
of the functionalised frameworks was indicated by compar-
ing the powder X-ray diffraction (PXRD) patterns of the as-
synthesised materials with the calculated diffraction patterns
of the unmodified parent MOFs (Figures 2 and 7).[8,15] More-
over, single-crystal X-ray structures of the as-synthesised
materials [Zn4O(LX)
3ACHTUGNTRENN(GNU DMF)m]n (X=1, 2, 3 and 4) and
[Zn2(L1)
G
ACHTUNGTRENNUNG
molecular structure of the compounds in more detail (see
the Supporting Information). As expected, it was not yet
possible to refine the atom positions of the flexible ether
chains in the single-crystal structure analysis due to high dis-
order. In fact, the structural features of the ether-functional-
ised MOFs can be viewed as a crystalline solid-state materi-
al combined with disordered liquid-state solvent proper-
ties.[17]
The as-synthesised MOFs were activated by solvent ex-
change with chloroform and finally dried in vacuo to ach-
ieve the solvent-free materials [Zn4O(LX)3]n and [Zn2(LX)2-
ACHTUNGTRENNUNG(dabco)]n (X=1, 2, 3 and 4) according to standard proce-
dures for the parent MOFs. The activated MOFs were ana-
lysed with thermogravimetric analysis (TGA), FTIR spec-
troscopy and 13C-MAS-NMR (magic angle spinning)
spectroscopy. Evidence of the expected stoichiometric pres-
ence of the functional groups inside the pores was given by
elemental analysis, as well as high-resolution 1H and
13C NMR spectroscopy of digested MOF samples in DCl/
D2O/[D6]DMSO (see the Supporting Information).
Figure 2. Powder X-ray diffraction (PXRD) patterns of [Zn4O(LX)3]n
(X=1, 2, 3 and 4): a) [Zn4O(bdc)3]n reference (simulated from the X-ray
structure);[8] [Zn4O(L1)
b) as-synthesised 3A(DMF)m]n; c) dried
[Zn4O(L1)3]n; d) reinfiltrated DMF@[Zn4O(L1)3]n; e) as-synthesised
[Zn4O(L2) (DMF)m]n; f) dried [Zn4O(L2)3]n; g) reinfiltrated DMF@-
3A
[Zn4O(L2)3]n; [Zn4O(L3)
h) as-synthesised 3A(DMF)m]n; i) dried
[Zn4O(L3)3]n; j) reinfiltrated DMF@[Zn4O(L3)3]]n; k) as-synthesised
3A
[Zn4O(L4) (DMF)m]n; l) dried [Zn4O(L4)3]n and m) reinfiltrated DMF@-
[Zn4O(L4)3]n.
AHCTUNGTRENNUNG
CTHUNGTRENNUNG
Structural dynamics of [Zn4O(LX)3]n MOFs: Interestingly,
the novel functionalised MOFs behave differently upon re-
moval of the solvent guests from the pores. Surprisingly, the
PXRD pattern of dried [Zn4O(L1)3]n shows an almost X-ray
amorphous material (Figure 2c). Only the first reflex at
2q=6.78 is present as a very small peak. This effect is much
less pronounced for the other derivatives [Zn4O(LX)3]n (X=
2, 3 and 4). Moisture was rigorously ruled out by strictly
working under an inert gas atmosphere (dried Ar; Schlenk
AHCTUNGTRENNUNG
CTHUNGTRENNUNG
G
CHTUNGTRENNUNG
AHCTUNGTRENNUNG
CTHUNGTRENNUNG
AHCTUNGTRENNUNG
and glovebox techniques). However, exposing the amor-
phous material [Zn4O(L1)3]n to DMF vapour under reduced
pressure quantitatively yielded the reinfiltrated material
14298
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 14296 – 14306

Flexible Metal–Organic Frameworks
FULL PAPER
DMF@
(Figure 2d), similar to the initial as-synthesised material
[Zn4O(L1)
3A(DMF)m]n. It is well known that the formation of
crystalline IRMOFs can occur at room temperature in the
liquid phase.[18] However, the reordering process reported
herein is very different from such a room-temperature syn-
thesis. Instead the structural transformation occurs depend-
ing on the absence or presence of guest molecules adsorbed
in the pores of the framework.
Interestingly, the fully reversible crystalline-to-amorphous
transformation appears also by repetitive infiltration and re-
moval of other polar guest molecules such as N,N’-diethyl-
formamide (DEF) and THF in the gas phase (Figure 3b and
G
esting new phenomenon in IRMOF chemistry. A more or
less pronounced and fully reversible crystalline-to-amor-
phous phase transition can be triggered by the implementa-
tion of flexible methoxy-terminated alkyl groups and can be
tuned by the alkyl chain length and the substitution pattern
at the bdc-type linker.
In general, crystalline-to-amorphous phase transitions are
known from other MOFs.[11,19] Interestingly, such effects
were never reported for IRMOF-type structures, which are
particularly known for rigidity and do not show any flexibili-
ty, even in the case of a huge variety of previously known
linker-functionalised derivatives.[1,20] That rigidity is attribut-
ed to the symmetry of the inorganic sub-unit, which has less
flexibility to allow framework dynamics.[10,21] However, our
data surprisingly show that a suitably functionalised IRMOF
structure can be responsive to the nature of guest in the
pores.
Two different mechanisms are suggested to explain the re-
versible crystalline-to-amorphous phase transitions in
[Zn4O(L1)3]n: 1) In the absence of any polar guest molecule,
strong dipolar intermolecular interactions of the methoxy
groups with each other may lead to a major displacement of
the linkers and the [Zn4O]6+ framework nodes. 2) In the ab-
sence of any polar guest molecule, the methoxy groups coor-
dinate to the zinc centres at the framework nodes and
induce a random distortion of the [Zn4O]6+ cluster. Both
mechanisms may lead to a significant loss of long-range
order as a source of the phase transitions.
FTIR spectroscopy is a useful tool to analyse the coordi-
nation of the carboxylates to the zinc centres in the crystal-
line as well as in the amorphous states of [Zn4O(L1)3]n in
more detail.[22] A close-up view of the vibration bands be-
Figure 3. PXRD patterns of reinfiltrated samples of [Zn4O(L1)3]n:
a) DMF@
ACHTUNGTRENNUNG ACHTUNGTRENNUNG ACTHNUGTRENNGUN
[Zn4O(L1)3]n; b) DEF@[Zn4O(L1)3]n; c) THF@[Zn4O(L1)3]n;
d) pentane@ACHTUNGTRENNUNG ACHTNUGTRENNUNG
[Zn4O(L1)3]n and e) toluene@[Zn4O(L1)3]n.
tween 400 and 2000 cmÀ1 of crystalline [Zn4O(L1)
3ACHTNUTRGNE(UNG DMF)m]n,
c). However, the recovery of crystallinity does not work by
gas-phase infiltration of non-polar guest molecules, such as
pentane or toluene, as analysed by PXRD (Figure 3d and
e). PXRD analysis of the other ether-functionalised IRMOF
derivatives [Zn4O(LX)3]n (X=2, 3 and 4) in the course of ac-
tivation and solvent reinfiltration cycles reveals the huge
impact of the substitution pattern of the linker and the alkyl
chain length. The PXRD pattern of the dried material,
[Zn4O(L2)3]n, also indicates a substantial loss of crystallinity
of the material upon solvent removal (Figure 2 f). The X-ray
reflexes of low hkl values are still visible as broad peaks in
the diffraction pattern, suggesting that the loss in long-range
order is not as far-reaching as in [Zn4O(L1)3]n. Again, the in-
itial highly crystalline phase is recovered upon gas-phase in-
filtration of DMF in the pores (Figure 2g). Similar results
can be concluded from the diffraction patterns of
[Zn4O(L3)3]n (Figure 2h–j). The loss in crystallinity upon re-
moval of DMF from the pores is less apparent compared to
[Zn4O(L2)3]n. Remember, the only difference between
[Zn4O(L1)3]n, [Zn4O(L2)3]n, and [Zn4O(L3)3]n is the number
of CH2 groups in the alkyl chain. However, the fourth
IRMOF material [Zn4O(L4)3]n, which utilises linker L4 with
one ether side chain instead of two, displays no loss in crys-
tallinity at all (Figure 2k–m). These results indicate an inter-
amorphous [Zn4O(L1)3]n, pure DMF, protonated H2L1, and
the potassium salt K2L1 is shown in Figure 4. The spectra of
H2L1 and K2L1 represent the positions of the carboxylate
stretching vibrations, if they are protonated and non-coordi-
nating or if they are coordinating in a monodentate fashion
to metal centres. No significant changes or shifts of the car-
boxylate stretching vibrations are visible in the spectra of
the DMF-loaded MOF and the dried MOF. The additional
vibration bands in the spectrum of [Zn4O(L1)
3ACHTNUTRGNE(NUG DMF)m]n can
be assigned to DMF molecules in the pores of the as-syn-
thesised material. The splitting of the carbonyl vibration of
DMF is assigned to interactions between the adsorbed DMF
molecules and the framework. A comparison of the spec-
trum of [Zn4O(L1)3]n with the spectrum of H2L1 indicates
that no protonated carboxylates, similar to H2L1, are present
in the dried material. The small shoulder at the asymmetric
stretching vibration of the carboxylate groups (1548 cmÀ1) in
the spectrum of [Zn4O(L1)3] is in a similar position as for
the asymmetric stretching vibration of K2L1. Therefore,
some carboxylates of the dried MOF may only coordinate in
a monodentate fashion to the zinc centres. However, the
majority of the carboxylates still bind in the expected biden-
tate m2 coordination to zinc. Similar results can be concluded
from the FTIR spectra of the other functionalised IRMOF
Chem. Eur. J. 2010, 16, 14296 – 14306
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14299

R. A. Fischer et al.
Figure 5. Representations of the calculated model structures: A) Basic
zinc formiate [Zn4OCATHNUGTREN(NNGU O2CH)6] and B) [Zn4OAHCUTNRTGE(NGNUN O2CH)6ACHTNGUTRENN(NGU OMe2)2]. Zinc,
oxygen, carbon, and hydrogen atoms are shown in yellow, red, cyan, and
silver, respectively.
[Zn4OACHTNUGTRNEUNG(O2C)6]-based MOFs feature one zinc atom per
[Zn4O]6+ building unit with an octahedral coordination
sphere, whereas the other three zinc atoms of the inorganic
building block are tetrahedrally coordinated.[23] The addi-
tional ligands at the octahedrally coordinated zinc atoms of
the literature reference structures are DEF or water solvent
molecules from the MOF syntheses. Structure B is construct-
ed similar to the framework nodes of the reference struc-
tures. Structure A was optimised in the Td point group,
whereas structure B was optimised with no symmetry con-
straints. For determination of the binding energy of dimethyl
ether to basic zinc formiate, the structure of dimethyl ether
was optimised in C2v symmetry. To ensure that local minima
were located, vibrational frequencies were calculated for all
optimised structures.
Figure 4. Attenuated total reflectance (ATR) FTIR spectra of a) as-syn-
thesised [Zn4O(L1) (DMF)m]n, b) dried [Zn4O(L1)3]n, c) DMF, d) H2L1
3ACHTUNGTRENNUNG
and e) K2L1. The asymmetric stretching vibrations of the carboxylate
groups are marked with an asterisk (*).
derivatives utilising linker L2 and L3 (see the Supporting In-
formation).
A further hint for a reversible change in the zinc coordi-
nation sphere, depending on the presence or absence of
DMF guests, is given by preliminary X-ray absorption near-
edge structure (XANES) and extended X-ray absorption
fine structure (EXAFS) data (see the Supporting Informa-
tion). The XANES and EXAFS spectra are very similar
before drying and after reinfiltration of DMF guests. They
show a distortion of the [Zn4O]6+ tetrahedron upon solvent
removal (see the Supporting Information for details). How-
ever, it is not yet possible to get unambiguous evidence for
coordination of the flexible methoxy groups to the zinc
atoms of the inorganic building block of the IRMOF deriva-
tive from the spectroscopic information obtained so far.
To get more information about the possible coordination
of ether groups to the [Zn4O]6+ nodes of an IRMOF, we
performed DFT calculations of simple model structures.
The binding energy per dimethyl ether molecule to basic
zinc formiate was calculated to be 15.1 kJmolÀ1. Therefore,
a coordination of ether groups to the framework nodes of
the activated MOFs seems to be feasible. Evaluation of
structure B shows a remarkable distortion from the ideal oc-
tahedral coordination sphere (see the Supporting Informa-
tion for further details). This distortion may be statistically
initiated from several directions in the cubic structure,
which may lead to a lack in long-range order and therefore
an X-ray amorphous material. We suggest that such a distor-
tion of the [Zn4O]6+ building block, initiated by additional
coordination of some of the flexible methoxy groups of the
linker, is the major reason for the crystalline-to-amorphous
phase transition.
In the presence of polar guests, the ether side chains may
interact with the guest molecules and also compete with
them for weak coordination at Zn2+ sites (Figure 6). Clearly,
the covalent anchorage of the ether side groups of optimum
length at the linkers appears to be a prerequisite to trigger
the observed effect. In summary, we attribute the amor-
phous-to-crystalline transition to a peculiar case of frame-
work flexibility and not to an unspecific decomposition and
breakdown of the framework upon solvent removal.
Basic zinc formiate, [Zn4OACHTNUGTRENUNG(O2CH)6] (Figure 5, structure A),
was employed as a structural model for the framework node
of IRMOF-type structures. A distorted basic zinc formiate
model structure, consisting of [Zn4OACHTUNGRTNEUNG(O2CH)6] and two addi-
tional (CH3)2O molecules (Figure 5, structure B), which are
connected to one zinc atom of the [Zn4O]6+ cluster, is em-
ployed to get information about the binding of ether groups
to the zinc atoms in comparison to the conventional basic
zinc formiate model structure. We have chosen this distorted
model structure because well-known literature examples of
Structural dynamics of [Zn2(LX)
ble crystalline-to-crystalline transformation was observed in
2ACHTNUGTREN(NNUG dabco)]n MOFs: A reversi-
14300
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 14296 – 14306

Flexible Metal–Organic Frameworks
FULL PAPER
Figure 6. Representation of the reversible crystalline-to-amorphous tran-
sition in [Zn4O(L1)3]n. In response to the presence or absence of polar
guest molecules, the framework reversibly switches between two states.
Upon solvent removal some attached flexible methoxy groups may coor-
dinate to the zinc centres, which induces a random distortion of the
[Zn4OACHTUNGTRENNUNG(O2C)6]n building block leading to a vast deprivation in long-range
translational order.
the case of the functionalised derivatives of [Zn2ACTHNUTRGNEN(UG bdc)2-
ACHTUNGTRENNUNG HCTUTGNREN(NUNG dabco)]n (Fig-
(dabco)]n. The PXRD of dried [Zn2(L1)
2A
ure 7d) shows a completely different pattern than the as-
synthesised form [Zn2(L1)
2A(dabco)(DMF)m]n (Figure 7c).
C
ACHTUNGTRENNUNG
Some of the reflexes are shifted to higher 2q values, indicat-
ing a remarkable contraction of the framework upon solvent
or guest removal. This kind of framework breathing appears
to be similar to the recently reported breathing of post-syn-
Figure 7. PXRD patterns of [Zn2(LX)
a) [Zn2A(bdc)2A(dabco)(DMF)m]n reference (simulated from the X-ray
structure);[15] b) [Zn
2A(bdc)2A(dabco)]n reference (simulated from the X-ray
structure);[15] c) as-synthesised [Zn2(L1)
2A(dabco)(DMF)m]n; d) dried
[Zn2(L1) [Zn2(L1)
2A(dabco)]n; e) reinfiltrated DMF@ 2A(dabco)]n; f) reinfil-
trated EtOH@ 2A 2A(dabco)-
[Zn2(L1) (dabco)]n; g) as-synthesised [Zn2(L2)
(DMF)m]n; h) dried [Zn2(L2) [Zn2(L2)2-
2A(dabco)]n; i) reinfiltrated DMF@
(dabco)]n; j) as-synthesised 2A(dabco)(DMF)m]n; k) dried
[Zn2(L3)
[Zn2(L3) [Zn2(L3)
2A(dabco)]n; l) reinfiltrated DMF@ 2A(dabco)]n; m) as-syn-
thesised [Zn2(L4) (DMF)m]n; n) dried [Zn2(L4)
2A(dabco) 2A(dabco)]n and
o) reinfiltrated DMF@ 2A(dabco)]n.
[Zn2(L4)
2AHCTUNTGRNEG(NU dabco)]n (X=1, 2, 3 and 4):
thetic modified derivatives of [Zn2ACHTUNGRTENNUG(NH2-bdc)2HCATUNGTRNE(NUGN dabco)]n
(NH2-bdc=2-aminobenzene dicarboxylate).[14] Gas-phase in-
G
R
ACHTUNGTRENNUNG
G
CHTUNGTRENNUNG
filtration of DMF or EtOH molecules in the pores of
C
ACHTUNGTRENNUNG
[Zn2(L1)
2A
R
A
CHTUNGTRENNUNG
A
E
G
CHTUNGTRENNUNG
G
U
CHTUNGTRENNUNG
again feature the original open-pore phase (Figure 7e and
f). Due to the saturated coordination sphere of the zinc cen-
G
E
ACHTUNGTRENNUNG
E
C
ACHTUNGTRENNUNG
E
A
CHTUNGTRENNUNG
tres in the [Zn2(L1)
2ACHTUNGTRENNUNG(dabco)]n (in contrast to the above cases
G
N
CHTUNGTRENNUNG
[Zn4O(LX)3]n), the structural transformation must be based
on interactions between the flexible ether groups and the
framework backbone (similar to the structural transforma-
tion reported for the partially 2-propanol-loaded material
N
CHTUNGTRENNUNG
[Zn
G
N
[Zn2(L1)
infiltrated DMF@
(dabco)
E
2ACHTUNGTREN(NUNG dabco)]n, and re-
tion/uncoordination of the methoxy groups to the zinc cen-
tres. Interestingly, PXRD patterns of the dried materials
N
CHTUNGTRENNUNG
monoclinic space group C2/m. To analyse the breathing
[Zn2(L2)
G
N
mode in more detail, the cell parameters for the guest infil-
E
trated and dried states of [Zn2(L1)
ACTHNGUTERNN(UG dabco)]n were refined
(Figure 7g–o). The diffraction reflexes shift only slightly in
position relative to the reflexes of the as-synthesised materi-
als. Therefore, our results again illustrate the major impact
of the substitution pattern, as well as alkyl chain length, on
the breathing behaviour of the “jungle-gym” type
(Table 1; see the Supporting Information for details). The
(110) reflex at 2q=8.18 is shifted to 2q=9.58 when the guest
molecules were removed from the pores. Accordingly the
(200) reflex is shifted from 2q=10.58 to 2q=9.48, whereas
the (001) reflex does not change its position (Figure 8). Cor-
responding to the refined cell parameters a structural model
[Zn2(dicarboxylate)
As-synthesised
(dabco)]n frameworks.
[Zn2(L1)2-
ACHTUNGTRENNUNG(dabco)ACHTUNGTRENNUNG(DMF)m]n crystallises in
Table 1. Results of the cell refinement of PXRD patterns of the different forms of [Zn2(L1)
2ACHTUNGTRENNUNG
monoclinic space group C2/m.
(dabco)]n in the
the C2/m space group as al-
ready determined by single-
crystal diffraction (see the Sup-
porting Information). Accord-
ingly, the reflexes of the PXRD
a [ꢂ]
b [ꢂ]
c [ꢂ]
a [8]
b [8]
g [8] ]
V [ꢂÀ3
A
G
E
16.75(3)
18.90(4)
16.70(5)
14.20(3)
10.74(1)
13.98(3)
9.75(1)
9.66(1)
9.63(2)
90
90
90
91.8(7)
90.3(9)
91.6(24)
90
90
90
2316(8)
1958(4)
2247(9)
E
N
patterns
of
as-synthesised
Chem. Eur. J. 2010, 16, 14296 – 14306
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14301

R. A. Fischer et al.
for the dried material [Zn2(L1)2-
AHCTUNGRTEN(GNUN dabco)]n could be derived (see
the Supporting Information for
details).
The
open-pore phase
(dabco)ACTHUNGTRNENU(G DMF)m]n is
[Zn2(L1)
U
transformed to a narrow-pore
phase in [Zn2(L1)
CTHUNGTRENNUNG
which has only 85% of the cell
volume of the initial open form.
Subsequent re-adsorption of
DMF in the pores of the dried,
narrow-pore material transfers
the MOF back to the open-
pore structure DMF@
AHCTUNGTREG(NUNN dabco)]n (Figure 8).
[Zn2(L1)2-
ACHTUNGTRENNUNG
Gas sorption properties: The
separation of CO2 from N2 or
CH4 is of great relevance for in-
dustrial applications, for exam-
ple, the capture of CO2 from
flue gas or CO2/CH4 separation
for natural gas upgrading. To
analyse the sorption abilities,
N2 (77 K), CO2 (195 K) and
CH4 (195 K) sorption isotherms
of [Zn4O(LX)3]n and [Zn2(LX)2-
Figure 8. Close-up view of the PXRD patterns in the range from 2q=7–118 (bottom) and structural models
for the individual states of [Zn2(L1)
2A(dabco)]n derived from cell refinement of the PXRD patterns (top): a) as-
synthesised [Zn2(L1) (DMF)m]n; b) dried [Zn2(L1) [Zn2(L1)
2A(dabco) 2A(dabco)]n; c) reinfiltrated DMF@ 2A(dabco)]n.
The flexible 2-methoxyethoxy side chains were not included in the model structures.
CTHUNGTRENNUNG
G
R
G
G
CHTUNGTRENNUNG
AHCTUNTGREN(GNNU dabco)]n (X=1, 2 and 4) were
recorded (Figure 9). The mate-
rials [Zn4O(L1)3]n, [Zn4O(L2)3]n,
1
2
4
1
~
&
*
Figure 9. N2 (77 K, ), CO2 (195 K, ) and CH4 (195 K, ) sorption isotherms of a) [Zn4O(L )3]n; b) [Zn4O(L )3]n; c) [Zn4O(L )3]n; d) [Zn2(L )
e) [Zn2(L2) (dabco)]n and f) [Zn2(L4) (dabco)]n. CH4 sorption isotherms were only recorded for the materials [Zn4O(L1)3]n and [Zn2(L1)
2A 2A 2A(dabco)]n. The
adsorption and desorption branches are shown with closed and open symbols, respectively.
2ACHTUNGTRENNUNG(dabco)]n;
G
N
CHTUNGTRENNUNG
14302
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 14296 – 14306

Flexible Metal–Organic Frameworks
FULL PAPER
[Zn2(L1)
(dabco)]n and [Zn2(L2)
E
Furthermore, CH4 sorption isotherms recorded on the ma-
but N2 is not able to penetrate into the pores. In fact, there
is no uptake of nitrogen at all. This behaviour can only be
attributed to the presence of methoxy groups in the func-
tionalised materials. Note, that IRMOF-4 and IRMOF-5,
which are very similar to [Zn4O(L1)3]n and [Zn4O(L2)3]n, dis-
play permanent porosity towards N2 and no CO2 sorption
selectivity.[1] The only difference is the methoxy group at the
end of the alkyl chain in the functionalised MOFs instead of
a simple alkyl chain in IRMOF-4 and IRMOF-5. Therefore,
the reason for the high selectivity of the functionalised ma-
terials is not the reduction of the pore aperture window size
(e.g., a molecular sieving effect), but the very polar me-
thoxy-terminated pore surface.
terials [Zn4O(L1)3]n and [Zn2(L1)
2ACTHUNGRETNNU(G dabco)]n show interesting
results (Figure 9). CH4 is still able to penetrate in the pores
of [Zn4O(L1)3]n. Nevertheless, the uptake of CH4 is much
less compared to CO2 at the same temperature (selectivity
factor=VadsACHTUNTRGENNU(G CO2)/VadsCAHTUNGTREN(NUGN CH4)=3.78 at 1.0 bar). Because there
is no uptake of N2 in [Zn4O(L1)3]n, the material is promising
for highly selective separation of CH4 and N2, which is a
rarely reported feature of MOFs.[25] In contrast, [Zn2(L1)2-
AHCTUNGRTENNG(UN dabco)]n shows essentially no uptake of CH4 and N2, which
makes it highly selective for the adsorption of CO2
(Figure 10).
Compared with the loading with polar solvents, in situ
PXRD of amorphous [Zn4O(L1)3]n at 195 K in CO2 atmos-
phere (1 bar) does not show reordering of the structure and
recovery of the crystalline state (data not shown). This data
indicates that interesting sorption properties can very well
be achieved with suitable coordination polymers, even in
the case of an amorphous, non-crystalline phase. Thus, high
crystalline order of MOFs is not a stringent prerequisite for
porosity and high sorption selectivity.
Figure 10. Sketch of the selective sorption in [Zn2(L1)
2ACHTUNRGTNEUNG(dabco)]n. CH4 and
N2 cannot enter the pores in the closed state and will not be adsorbed in
the functionalised MOF, whereas CO2 as able to penetrate in the pores
and transform the structure to the open state.
The CO2 sorption isotherms of the materials [Zn4O(L1)3]n,
[Zn4O(L2)3]n and [Zn2(L2)
CHTUNGTRENNUNG
2A
behaviour, whereas the CO2 isotherm of [Zn2(L1)
CTHUNGTRENNUNG
is different. It displays a two-step uptake, which is typical
for flexible MOFs that undergo a crystal-to-crystal transition
upon adsorption. As already evident from the PXRD pat-
Conclusion
tern, the dried [Zn2(L1)
2ACHTNUTRGENNG(U dabco)]n exists in a narrow-pore
state, which has only minor porosity. In the pressure range
of p=0–0.23 bar a typical type I isotherm behaviour is pres-
ent. At the gate-opening pressure (p=0.23 bar), the struc-
ture switches to the open state and the pores are rapidly
filled with CO2. On desorption a significant hysteresis is
present with a gate-closing pressure of p=0.16 bar.
Our data demonstrate the potential of alkyl ether function-
alised linkers to tune flexibility and sorption properties of
MOFs. The alkyl ether chain length has a huge influence on
the structural flexibility of the materials. Both MOFs that
utilise linker L1 show remarkable flexibility and structural
transformations in response to adsorbed guest molecules
that are not so extensive in the related materials, which uti-
lise linkers with longer alkyl ether chains (L2 and L3). Fur-
thermore, the substitution pattern of the bdc-type linker has
a critical impact on the sorption properties of the presented
materials. In general, the materials utilising a bdc derivative
substituted in positions 2 and 5 (linkers L1 and L2) exhibit
high sorption selectivity towards CO2, whereas the materials
employing the bdc derivative only substituted in position 2
(linker L4) do not.
Clearly, the density of flexible ether groups in the frame-
work is of high significance. We suggest that the flexible and
polar groups act as molecular gates at the pore apertures of
the porous networks. At least two flexible ether chains per
dicarboxylate linker are necessary to lock the pore windows
effectively. Due to the polar nature of the ether chains,
polar molecules such as CO2 can easily penetrate through
the molecular gate, whereas N2 molecules cannot. We would
like to emphasise here that even an amorphous MOF phase,
such as [Zn4O(L1)3]n, may exhibit interesting sorption prop-
erties. Perfect long-range translational order of the MOF
material and/or reversible crystalline-to-crystalline transfor-
In contrast, the materials [Zn4O(L4)3]n and [Zn2(L4)2-
ACHTUNGTRENNUNG(dabco)]n do not possess any sorption selectivity (Figure 9c
and f, respectively). Both gases, N2 and CO2, are almost
equally adsorbed in these materials. Determination of the
specific surface area from the nitrogen sorption isotherms
according to the BET model yields surface areas of
1730 m2 gÀ1 for [Zn4O(L4)3]n (SLangmuir =2130 m2 gÀ1) and
810 m2 gÀ1 [Zn2(L4)
values are significantly lower than the surface areas of the
unmodified parent structures [Zn4O(bdc)3]n and [Zn2A(bdc)2-
ACHTUNGTRENNUNG
(dabco)]n.[8,15] This can be ascribed to the additional func-
(dabco)]n (SLangmuir =895 m2 gÀ1). These
2ACHTUNGTRENNUNG
A
U
tional groups localised in the framework voids, which in-
crease molecular weight and density, and decrease the acces-
sible free volume compared with the unmodified parent
structures. Remarkably, the CO2 sorption capacity of
2A
(dabco)]n (180 cm3 gÀ1 at 1 bar) is very similar. However, the
nitrogen sorption capacity is very different (almost 0 cm3 gÀ1
2A
for [Zn2(L1) (dabco)]n and 200 cm3 gÀ1 for [Zn2(L4)2-
(dabco)]n). This underlines the important influence of the
substitution pattern at the bdc-type linkers.
ACHTUNGTRENNUNG
CHTUNGTRENNUNG
ACHTUNGTRENNUNG
Chem. Eur. J. 2010, 16, 14296 – 14306
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14303

R. A. Fischer et al.
mations as in the case of [Zn2(L1)
2ACTHNUTRGNE(UNG dabco)]n is apparently
the PLATON program package to give an account for the electron densi-
ty associated to the disordered alkyl ether substituents, as well as disor-
dered and partially occupied solvent molecules (DMF) in the porous co-
ordination polymers. Powder X-ray diffraction (PXRD) patterns were re-
corded on a D8 Advance Bruker AXS diffractometer with CuKa radiation
(l=1.54178 ꢂ) and a Gçbel mirror in q–2q geometry with a position-sen-
sitive detector in a 2q range from 5–508 at a scan speed of 18minÀ1 at
298 K. a-Al2O3 was employed as an external standard. The powder
sample of the as-synthesised MOF was filled into glass capillaries (diame-
ter=1.5 mm) with a pipette in air, whereas the samples of the dried and
infiltrated materials were filled into glass capillaries (diameter=0.7 mm)
in a glovebox (Ar atmosphere). Each capillary was sealed prior to the
measurement. The thermogravimetric analyses were performed on a
Seiko TG/DTA 6300S11 instrument (sample weight approximately
10 mg) at a heating rate of 5 KminÀ1 in a temperature range from 300–
870 K. The measurement was performed at atmospheric pressure under
flowing nitrogen (99.9999%; flow rate=300 mLminÀ1). Sorption meas-
urements were performed by using a Quantachrome Autosorp-1 MP in-
strument and optimised protocols and gases of 99.9995% purity. The N2
measurements were performed at 77 K and CO2 and CH4 measurements
at 195 K. XANES and EXAFS spectra were recorded at beamline X1 at
HASYLAB, and Deutsches Elektronensynchrotron DESY, Hamburg,
not a general prerequisite for highly selective gas adsorption
for (functionalised) porous coordination polymers. Rather,
the sorption properties of the discussed functionalised
MOFs are mainly dependent on the particular linkers and
only weakly dependent on the details of the framework
structure. Therefore, it should in principle be possible to
tune the sorption properties of MOFs in general by rational
design of suitably functionalised linkers.
In summary, we have developed a methodology to inte-
grate responsiveness and flexibility accompanied with high
sorption selectivity towards CO2 in already known, other-
wise quite rigid, MOFs without changing the underlying re-
ticular structure type and topology. We use the quite abun-
dant bdc-type linkers, which were covalently modified with
flexible alkyl ether side chains that do not interfere with
MOF formation and are thus likely to be widely applicable
in the synthesis of other bdc-based MOF systems. According
to multivariate MOFs, which were recently introduced by
Deng et al., it should in principle be possible to fine-tune
sorption selectivity and flexibility of MOFs by utilising sev-
eral ether-functionalised and non-functionalised linkers in
varying ratios in MOF synthesis.[20b] In future work we will
examine if related and similarly functionalised carboxylate
linkers can establish sorption selectivity and structural flexi-
bility in such multivariant and other related MOF struc-
tures.
using a SiACTHNUTRGNE(NUG 111) double crystal monochromator (50% detuning for remov-
ing higher harmonics) and Zn-foil as reference. The data treatment was
performed with the Winxas 3.1. software.[26]
Synthesis of 2,5-bis(2-methoxyethoxy)-1,4-benzene dicarboxylic acid
(H2L1): The functionalised linkers were synthesised by Mitsunobu etheri-
fication.[16] Diethyl 2,5-dihydroxyterephthalate (1.12 g, 4.42 mmol), tri-
phenylphosphine (2.43 g, 9.30 mmol), and di-tert-butylazodicarboxylate
(2.14 g, 9.30 mmol) were suspended in dry THF (15 mL). 2-Methoxyetha-
nol (9.30 mmol) was added dropwise to the solution. Subsequently, the
reaction mixture was stirred and sonicated for 10 min and then washed
with hexane (2 mL). NaOH (0.30 g) in water (10 mL) was added to the
ether phase followed by 60 min sonication. Accordingly the water phase
was separated from the ether phase and extracted three times with ethyl
acetate. Finally the water phase was acidified with aqueous HCl (ꢀ10%)
and the precipitated white product was filtered, recrystallised from meth-
anol and dried in vacuo to give H2L1 (1.25 g, 3.98 mmol, 90%). 1H NMR
(250 MHz, [D8]THF, 258C, TMS): d=7.58 (s, 2H; Ar-H), 4.23 (t, 4H;
CH2), 3.71 (t, 4H; CH2), 3.38 ppm (s, 6H; CH3).
Experimental Section
Materials: All chemicals were purchased from commercial suppliers
(Sigma–Aldrich, Fluka, Alfa Aesar, and others) and used without further
purification. THF used in the linker synthesis was catalytically dried, de-
oxygenated, and saturated with argon using an automatic solvent purifi-
cation system from MBraun. The residual water content was determined
by Karl Fischer titration, exhibiting levels of 5 ppm. Before further ma-
nipulations all dried and activated MOF samples were stored under inert
gas atmosphere in a glovebox.
Synthesis of 2,5-bis(3-methoxypropoxy)-1,4-benzene dicarboxylic acid
(H2L2): Diethyl 2,5-dihydroxyterephthalate (1.12 g, 4.42 mmol), triphenyl-
phosphine (2.43 g, 9.30 mmol), and di-tert-butylazodicarboxylate (2.14 g,
9.30 mmol) were suspended in dry THF (15 mL). 3-Methoxypropanol
(9.30 mmol) was added dropwise to the solution. Subsequently, the reac-
tion mixture was stirred and sonicated for 10 min and then washed with
hexane (2 mL). NaOH (0.30 g) in water (10 mL) was added to the ether
phase followed by 60 min sonication. Accordingly the water phase was
separated from the ether phase and extracted three times with ethyl ace-
tate. Finally the water phase was acidified with aqueous HCl (ꢀ10%)
and the precipitated white product was filtered, recrystallised from meth-
anol and dried in vacuo to give H2L2 (1.29 g, 3.78 mmol, 85%). 1H NMR
(250 MHz, [D8]THF, 258C, TMS): d=7.51 (s, 2H; Ar-H), 4.14 (t, 4H;
CH2), 3.54 (t, 4H; CH2), 3.28 (s, 6H; CH3), 2.01 ppm (m, 2H; CH2).
Methods: Elemental analyses were performed in the Microanalytical
Laboratory of the Department of Analytical Chemistry at the Ruhr-Uni-
versity Bochum. Liquid-phase NMR spectra were recorded on a Bruker
Advance DPX-250 spectrometer (1H, 250.1 MHz; 13C, 62.9 MHz) at
293 K. 1H NMR spectra of the synthesised linker molecules were record-
ed in [D8]THF, whereas the 1H and 13C NMR spectra of the digested
MOFs were recorded in 0.5 mL [D6]DMSO and 0.1 mL DClACHTUNGRTNEUNG(20%)/D2O.
Chemical shifts are given relative to TMS and were referenced to the sol-
vent signals as internal standards. Solid-state 13C-MAS-NMR spectra
were recorded on a Bruker DSX-400 MHz spectrometer in ZrO2 rotors
of 2.5 mm diameter at 293 K. All spectra were measured by applying
pulse programs written by H.-J. Hauswald at the Department of Analyti-
cal Chemistry of the Ruhr-University Bochum. Several 13C-MAS-NMR
spectra show a rather weak signal at d=80 ppm, which is a measurement
artefact and can be assigned to the centre of the resonance experiment.
IR spectra were recorded inside a glovebox on a Bruker Alpha-P FTIR
instrument in the ATR geometry with a diamond ATR unit. Single-crys-
tal X-ray structures were measured on an Oxford Excalibur 2 diffractom-
eter in a nitrogen cold stream (100–110 K) using MoKa radiation (l=
0.71073 ꢂ). The structure was solved by direct methods using SHELXS-
97 and refined against F2 on all data by full-matrix least squares with
SHELXL-97 (SHELX-97 program package, Sheldrick, Universitꢁt Gçt-
tingen, 1997). The structures were treated with the “sqeeze” protocol in
Synthesis of 2,5-bis(4-methoxybutoxy)-1,4-benzene dicarboxylic acid
(H2L3): Diethyl 2,5-dihydroxyterephthalate (1.12 g, 4.42 mmol), triphenyl-
phosphine (2.43 g, 9.30 mmol), and di-tert-butylazodicarboxylate (2.14 g,
9.30 mmol) were suspended in dry THF (15 mL). 4-Methoxybutanol
(9.30 mmol) was added dropwise to the solution. Subsequently, the reac-
tion mixture was stirred and sonicated for 10 min and then washed with
hexane (2 mL). NaOH (0.30 g) in water (10 mL) was added to the ether
phase followed by 60 min sonication. Accordingly the water phase was
separated from the ether phase and extracted three times with ethyl ace-
tate. Finally the water phase was acidified with aqueous HCl (ꢀ10%)
and the precipitated white product was filtered, recrystallised from meth-
anol and dried in vacuo to give H2L3 (1.33 g, 3.59 mmol, 81%). 1H NMR
(250 MHz, [D8]THF, 258C, TMS): d=7.48 (s, 2H; Ar-H), 4.07 (t, 4H;
CH2), 3.39 (t, 4H; CH2), 3.26 (s, 6H; CH3), 1.85 ppm (m, 4H; CH2).
14304
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 14296 – 14306

Flexible Metal–Organic Frameworks
FULL PAPER
Synthesis of 2-(2-methoxyethoxy)-1,4-benzene dicarboxylic acid (H2L4):
The precursor 2-hydroxyterephthalic acid was synthesised according to a
literature report.[27] 2-Bromoterephthalic acid (5.00 g, 20.4 mmol), NaOH
(1.64 g, 40.8 mmol) and NaOAc (3.68 g, 44.8 mmol) were dissolved in
water (95 mL). Cu powder (0.026 g) and few drops of phenolphthalein
solution were added and the mixture was heated under reflux for 4 d. Oc-
casionally aqueous KOH (ꢀ10%) was added to keep the reaction mix-
ture alkaline. After cooling, the mixture was filtered, and the filtrate was
acidified with aqueous HCl (ꢀ10%). A white solid precipitated, was col-
lected by filtration and dried in vacuo to give 2-hydroxyterephthalic acid
(3.52 g, 19.3 mmol, 95%). The 2-hydroxyterephthalic acid was used with-
out further purification. 2-Hydroxyterephthalic acid (3.52 g, 19.3 mmol)
was dissolved in dry methanol (50 mL) and borontrifluoride ethyl ether-
ate (5 mL). Afterwards the solution was heated at reflux for 2 h and then
cooled to room temperature. The solvent was completely removed under
reduced pressure. The white residue was washed three times with water,
recrystallised from pyridine and dried in vacuo to give dimethyl 2-hy-
droxyterephthalate. Etherification of the hydroxyl group was again per-
formed according to the Mitsunobu method.[16] Dimethyl 2-hydroxytere-
phthalate (1.00 g, 4.76 mmol), triphenylphosphine (1.31 g, 5.00 mmol),
and di-tert-butylazodicarboxylate (1.15 g, 5.00 mmol) were suspended in
dry THF (10 mL). 2-methoxyethanol (5.00 mmol) was added dropwise to
the solution. The reaction mixture was stirred and sonicated for 10 min
and then extracted with hexane (2 mL). NaOH (0.40 g) in water (10 mL)
was added to the ether phase followed by 60 min sonication. Accordingly
the water phase was separated from the ether phase and extracted three
times with ethyl acetate. Finally the water phase was acidified with aque-
ous HCl (ꢀ10%) and the precipitated white product was filtered, recrys-
tallised from methanol and dried in vacuo to give H2L4 (0.989 g,
4.12 mmol, 87%). 1H NMR (250 MHz, [D8]THF, 258C, TMS): d=7.87
(d, 1H; Ar-H), 7.73 (s, 1H; Ar-H), 7.65 (d, 1H; Ar-H), 4.29 (t, 2H;
CH2), 3.76 (t, 2H; CH2), 3.41 ppm (s, 3H; CH3).
NMR spectroscopy and 1H and 13C NMR spectroscopy of digested sam-
ples in [D6]DMSO (0.5 mL) and DCl (0.1 mL)/D2O (20%). For analyses
of the sorption properties of the materials N2 and CO2 sorption isotherms
were recorded.
Gas-phase infiltration of solvent guests in the dried donor-functionalised
MOFs: In
a
typical infiltration experiment the dedicated material
2A(dabco)]n (X=1, 2, 3 and 4; 100 mg) and the
[Zn4O(LX)3]n or [Zn2(LX)
CTHUNGTRENNUNG
solvent guests (DMF, DEF, THF, EtOH, toluene or pentane; 1 mL) were
placed in two separate glass vials in a Schlenk tube. The tube was evacu-
ated for 1 min to achieve a static vacuum, sealed and kept at room tem-
perature (323 K for DMF, DEF, and toluene) for 16 h to achieve the
maximum loading of the material. The infiltrated materials guest@-
AHCTUNGTERNNGNU 2AHCTUNGTRENN(GUN dabco)]n (X=1, 2, 3 and 4) were
[Zn4O(LX)3]n and guest@Zn2(LX)
stored under an inert atmosphere. PXRD was carried out for characteri-
sation.
CCDC-788015, 788016, 788017, 788018 and 788019 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.
Acknowledgements
The authors would like to thank the “Research Centre 558” of the
German Research Foundation DFG for funding. S.H. is grateful to ac-
knowledge the Research School of the Ruhr-University Bochum for a
fellowship and the “Fonds of the German Chemical Industry” for a doc-
toral scholarship. We thank HASYLAB at DESY (Hamburg, Germany)
for beamtime allocation and the EU for financial support (contract RII3-
CT-2004-506008).
Synthesis of [Zn4O(LX)3]n (X=1, 2, 3 and 4): H2LX (X=1, 2, 3 and 4;
0.666 mmol) and ZnACHTUNGTRENNUNG(NO3)2·4H2O (0.756 g, 2.88 mmol) were dissolved in
[1] M. Eddaoudi, J. Kim, N. Rosi, D. Vodak, J. Wachter, M. OꢃKeeffe,
O. M. Yaghi, Science 2002, 295, 469–472.
[2] a) S. Kitagawa, R. Kitaura, S. Noro, Angew. Chem. 2004, 116, 2388–
2430; Angew. Chem. Int. Ed. 2004, 43, 2334–2375; b) R. Robson,
Dalton Trans. 2008, 5113–5131.
[3] a) A. R. Millward, O. M. Yaghi, J. Am. Chem. Soc. 2005, 127,
17998–17999; b) M. Dinca, J. Long, Angew. Chem. 2008, 120, 6870–
6884; Angew. Chem. Int. Ed. 2008, 47, 6766–6779.
[4] M. D. Allendorf, R. J. T. Houk, L. Andruszkiewicz, A. A. Talin, J.
Pikarsky, A. Choudhury, K. A. Gall, P. J. Hesketh, J. Am. Chem.
Soc. 2008, 130, 14404–14405.
[5] a) L. Alaerts, C. E. A. Kirschhock, M. Maes, M. A. van der Veen, V.
Finsy, A. Depla, J. A. Martens, G. V. Baron, P. A. Jacobs, J. F. M. De-
nayer, D. E. De Vos, Angew. Chem. 2007, 119, 4371–4375; Angew.
Chem. Int. Ed. 2007, 46, 4293–4297; b) J.-R. Li, R. J. Kuppler, H.-C.
Zhou, Chem. Soc. Rev. 2009, 38, 1477–1504, and references therein.
[6] J. Y. Lee, O. K. Farha, J. Roberts, K. A. Scheidt, S. T. Nguyen, J. T.
Hupp, Chem. Soc. Rev. 2009, 38, 1450–1459.
[7] P. Horcajada, C. Serre, G. Maurin, N. A. Ramsahye, F. Balas, M.
Vallet-Regꢄ, M. Sebban, F. Taulelle, G. Fꢅrey, J. Am. Chem. Soc.
2008, 130, 6774–6780.
DMF (50 mL). The reaction mixture was transferred into a 100 mL reac-
tion vessel, sealed and heated from room temperature to 373 K with a
heating rate of 2 KhÀ1. The target temperature was held for 20 h. Then
the vessel was cooled to room temperature with a rate of À1.5 KhÀ1
.
Subsequently, the mother liquor was decanted and the colourless, cubic
MOF crystals of [Zn4O(LX)
3ACHTNUGTRENUNG(DMF)m]n (X=1, 2, 3 and 4) were washed
three times with DMF and then stored in DMF (30 mL) until further ma-
nipulation was required. The as-synthesised materials were characterised
by single-crystal X-ray diffraction analysis, PXRD and IR spectroscopy.
The DMF was decanted off and the crystals were stirred for 3 d in
chloroform (50 mL), whereas every 24 h the solvent was replaced by
fresh CHCl3. Afterwards the materials were dried under reduced pres-
sure (ꢀ10À3 mbar) for 48 h at 400 K to achieve the dried MOFs
[Zn4O(LX)3]n (X=1, 2, 3 and 4), which were stored under inert gas at-
mosphere in a glovebox. The dried MOFs were characterised by elemen-
tal analysis, PXRD, N2 and CO2 sorption, thermogravimetric analysis, IR
and 13C-MAS-NMR spectroscopy. Further characterisation was per-
formed by digestion of a few milligrams in [D6]DMSO (0.5 mL) and DCl
1
(0.1 mL)/D2O (20%) and collection of H and 13C NMR spectra.
Synthesis of [Zn2(LX)
synthesised in analogy to the literature synthesis of [Zn
(dabco)]n.[15] Zn(NO3)2·6H2O (500 mg, 1.68 mmol), H2LX (X=1, 2, 3 and
(dabco)]n (X=1, 2, 3 and 4): The materials were
2A(bdc)2-
CTHUNGTRENNUNG
[8] H. Li, M. Eddaoudi, M. OꢃKeefe, O. M. Yaghi, Nature 1999, 402,
276–278.
A
ACHTUNGTRENNUNG
4; 1.68 mmol), and 1,4-diazabicyclooctane (0.094 g, 0.84 mmol) were sus-
pended in DMF (20 mL) and transferred into a 50 mL reaction vessel.
The mixture was heated to 393 K for 48 h to yield block-shaped colour-
[9] D. Maspoch, D. Ruiz-Molina, K. Wurst, N. Domingo, M. Cavallini,
F. Biscarini, J. Tejada, C. Rovira, V. Veciana, Nat. Mater. 2003, 2,
190–195.
[10] G. Fꢅrey, C. Serre, Chem. Soc. Rev. 2009, 38, 1380–1399.
[11] G. J. Halder, C. J. Kepert, B. Moubaraki, K. S. Murray, J. D. Cashion,
Science 2002, 298, 1762–1765.
less crystals of [Zn2(L1)
2A
N
material of [Zn2(LX)
2A(dabco)ACTHUNGTRENNUNG
room temperature the material was washed twice with fresh DMF. The
as-synthesised materials were characterised with PXRD and IR spectros-
[12] C. Volkringer, S. M. Cohen, Angew. Chem. 2010, 122, 4748–4752;
Angew. Chem. Int. Ed. 2010, 49, 4644–4648.
copy. The crystals of [Zn2(L1)
2ACHTUNRGTENNUG(dabco)ACHTUNTGERN(NUGN DMF)m]n were further analysed by
single-crystal X-ray diffraction. For activation the materials were stirred
in chloroform (50 mL) for 3 d, followed by filtration and drying under re-
duced pressure (ꢀ10À3 mbar) at 400 K for 48 h. Characterisation of the
[13] Z. Wang, S. M. Cohen, J. Am. Chem. Soc. 2009, 131, 16675—16677.
[14] J. Seo, R. Matsuda, H. Sakamoto, C. Bonneau, S. Kitagawa, J. Am.
Chem. Soc. 2009, 131, 12792–12800.
[15] D. Dybtsev, H. Chun, K. Kim, Angew. Chem. 2004, 116, 5143–5146;
Angew. Chem. Int. Ed. 2004, 43, 5033–5036.
dried materials [Zn2(LX)
(dabco)]n (X=1, 2, 3 and 4) was performed with
elemental analysis, PXRD, thermogravimetric analyses, IR and 13C-MAS-
Chem. Eur. J. 2010, 16, 14296 – 14306
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14305

R. A. Fischer et al.
[16] a) O. Mitsunobu, Synthesis 1981, 1; b) S. D. Lepore, Y. He, J. Org.
Chem. 2003, 68, 8261–8263.
[22] V. Zelenꢆk, Z. Vargovꢆ, K. Gyçryovꢆ, Spectrochem. Acta Part A
2007, 66, 262–272.
[23] a) A. D. Burrows, C. G. Frost, M. F. Mahon, C. Richardson, Angew.
Chem. 2008, 120, 8610–8614; Angew. Chem. Int. Ed. 2008, 47, 8482–
8486; b) A. P. Nelson, O. K. Farha, K. L. Mulfort, J. T. Hupp, J. Am.
Chem. Soc. 2009, 131, 458–460.
[17] Q. Li, W. Zhang, O. Miljani, C.-H. Sue, Y.-L. Zhao, L. Liu, C. B.
Knobler, J. F. Stoddart, O. M. Yaghi, Science 2009, 325, 855–859.
[18] D. J. Tranchemontagne, J. R. Hunt, O. M. Yaghi, Tetrahedron 2008,
64, 8553–8557.
[24] K. Uemura, Y. Yamasaki, Y. Komagawa, K. Tanaka, H. Kita,
Angew. Chem. 2007, 119, 6782–6785; Angew. Chem. Int. Ed. 2007,
46, 6662–6665.
[25] a) K. Seki, Phys. Chem. Chem. Phys. 2002, 4, 1968–1971; b) R. Ki-
taura, K. Seki, G. Akiyama, S. Kitagawa, Angew. Chem. 2003, 115,
444–447; Angew. Chem. Int. Ed. 2003, 42, 428–431.
[26] T. Ressler, J. Synchrotron Radiat. 1998, 5, 118–122.
[27] Y. Miura, E. Torres, C. A. Panetta, R. M. Metzger, J. Org. Chem.
1988, 53, 439–440.
[19] a) S. Kitagawa, K. Uemura, Chem. Soc. Rev. 2005, 34, 109–119, and
references therein; b) N. Yanai, W. Kaneko, K. Yoneda, M. Ohba, S.
Kitagawa, J. Am. Chem. Soc. 2007, 129, 3496–3497; c) C.-D. Wu, W.
Lin, Angew. Chem. 2005, 117, 1994–1997; Angew. Chem. Int. Ed.
2005, 44, 1958–1961.
[20] a) Z. Wang, S. M. Cohen, Chem. Soc. Rev. 2009, 38, 1315–1329;
b) H. Deng, C. J. Doonan, H. Furukawa, R. B. Ferreira, J. Towne,
C. B. Knobler, B. Wang, O. M. Yaghi, Science 2010, 327, 846–850.
[21] M. Eddaoudi, D. B. Moler, H. Li, B. Chen, T. M. Reineke, M.
OꢃKeeffe, O. M. Yaghi, Acc. Chem. Res. 2001, 34, 319–330.
Received: August 15, 2010
Published online: December 7, 2010
14306
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 14296 – 14306