Synthesis, structural properties, and catalytic behavior of Cu-BTC and mixed-linker Cu-BTC-PyDC in the oxidation of benzene derivatives

Article abstract of DOI:10.1016/j.jcat.2011.04.004

Mixed-linker metal-organic frameworks based on the Cu-BTC structure have been synthesized in which the benzene-1,3,5-tricarboxylate (BTC) linkers have been partially replaced by pyridine-3,5-dicarboxylate (PyDC). X-ray-based techniques (powder XRD and XAS), thermal analysis, and infrared spectroscopy proved that a desired amount of PyDC (up to 50%) can be incorporated without changing significantly the crystal structure. The pyridine unit can be seen as a defect site in the local coordination environment of the dimeric copper units, which is significantly altering their electronic structure and the catalytic properties. Both Cu-BTC and the mixed Cu-BTC-PyDCs catalyze the demanding direct hydroxylation of toluene both in acetonitrile and in neat substrate. Different selectivity toward the desired ortho- and para-cresol and other oxidation products (benzaldehyde, benzyl alcohol, methylbenzoquinone) was observed for Cu-BTC and the Cu-BTC-PyDCs, respectively. Leaching tests and comparison with homogeneously dissolved Cu catalysts indicate mainly a heterogeneous reaction pathway.

Full text of DOI:10.1016/j.jcat.2011.04.004

Journal of Catalysis 281 (2011) 76–87
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Synthesis, structural properties, and catalytic behavior of Cu-BTC and mixed-linker
Cu-BTC-PyDC in the oxidation of benzene derivatives
Stefan Marx a, Wolfgang Kleist a,b, Alfons Baiker a,
⇑
a Institute for Chemical and Bioengineering, ETH Zurich, Wolfgang-Pauli-Str. 10, CH-8093 Zurich, Switzerland
b Institute of Catalysis Research and Technology, Karlsruhe Institute of Technology (KIT), Herrmann-von-Helmholtz-Platz 1, D-76344 Eggenstein-Leopoldshafen, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Mixed-linker metal-organic frameworks based on the Cu-BTC structure have been synthesized in which
the benzene-1,3,5-tricarboxylate (BTC) linkers have been partially replaced by pyridine-3,5-dicarboxyl-
ate (PyDC). X-ray-based techniques (powder XRD and XAS), thermal analysis, and infrared spectroscopy
proved that a desired amount of PyDC (up to 50%) can be incorporated without changing significantly the
crystal structure. The pyridine unit can be seen as a defect site in the local coordination environment of
the dimeric copper units, which is significantly altering their electronic structure and the catalytic prop-
erties. Both Cu-BTC and the mixed Cu-BTC-PyDCs catalyze the demanding direct hydroxylation of toluene
both in acetonitrile and in neat substrate. Different selectivity toward the desired ortho- and para-cresol
and other oxidation products (benzaldehyde, benzyl alcohol, methylbenzoquinone) was observed for Cu-
BTC and the Cu-BTC-PyDCs, respectively. Leaching tests and comparison with homogeneously dissolved
Cu catalysts indicate mainly a heterogeneous reaction pathway.
Received 21 January 2011
Revised 2 April 2011
Accepted 4 April 2011
Available online 11 May 2011
Keywords:
Metal-organic frameworks
Cu-BTC
Mixed-linker Cu-BTC-PyDC
Benzene derivatives
Oxidation
Hydrogen peroxide
Hydroxylation of aromatics
Ó 2011 Elsevier Inc. All rights reserved.
1. Introduction
be found in the literature where substrate activation in catalytic
reactions takes place at coordinatively unsaturated framework
In addition to their ‘‘classic’’ applications in gas storage and sep-
aration, metal-organic frameworks (MOFs) have attracted the
interest of researchers in the field of catalysis in recent years [1].
The crystalline, porous structure in combination with huge specific
surface areas and (micro)pore volumes results in unique properties
of these materials. Up to now, several different strategies have
been presented in the literature to introduce catalytically active
sites in metal-organic frameworks [2].
In the first approach, the MOF plays simply the role of a sup-
port material, which is used for immobilization of the active
component in the form of nanoparticles or clusters inside the
pores of the material [3–8]. In contrast to that, also the MOF
structure itself can contain catalytically active centers at the or-
ganic linker molecules of the framework. In this case, organic
molecules are utilized that feature functional side groups at
the aromatic units which can act as active (e.g., basic) sites
themselves [9,10] or which can be further transformed by
post-synthetic modification. This allows for the introduction of
additional organic functional groups [11–13] as well as the
immobilization of transition metal complexes via covalent bond-
ing (or covalent tethering) [14–17]. Finally, some examples can
metals [18–20]. Such active sites are, in general, occupied by sol-
vent molecules which can either be replaced by the reactants or
which have to be removed using activation procedures before
reaction (e.g., vacuum drying).
In addition to the presence of accessible sites at the metal
centers, MOF-based catalysts have to provide a sufficient stabil-
ity under the applied reaction conditions. Besides thermal stabil-
ity, chemical resistance (e.g., against dissolution in a certain
solvent or at a given pH) of the material is required. Conse-
quently, many reports on MOF catalysis focus on the use of
well-studied frameworks where these factors have been thor-
oughly investigated. Cu-BTC (general formula: Cu3(BTC)2(H2O)3,
also known as HKUST-1 [21] or MOF-199 [22], represents one
of the most famous examples of a metal-organic framework pos-
sessing all the required conditions. While the synthesis of phase-
pure Cu-BTC was difficult in the beginning (metallic and oxidic
Cu species were found as the main side products in early solvo-
thermal procedures) [21], several optimized routes have been
published recently [23] that opened possibilities for large-scale
production of Cu-BTC, which is an important pre-requisite for
catalytic applications. These include electrochemical methods
[1] and high-throughput screening [24] as well as ultrasonic
[25] and microwave [26] assisted procedures. In addition, the
growth of Cu-BTC nanocrystals on support materials has been
presented [27–29].
⇑
Corresponding author. Fax: +41 44 63 21163.
E-mail address: baiker@chem.ethz.ch (A. Baiker).
0021-9517/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2011.04.004

S. Marx et al. / Journal of Catalysis 281 (2011) 76–87
77
The SBU (secondary building unit) of Cu-BTC consists of a
dimeric copper unit that is bridged by four carboxylate functional-
ities of four BTC molecules. The resulting paddle-wheel moiety
is a characteristic motif for copper carboxylates. Each copper
atom is octahedrally coordinated by four oxygen and one copper
atom. The remaining, free coordination site, which is in the as-pre-
pared form occupied by a solvent molecule like water, represents
the possible active site for catalytic transformations. In Fig. 1
(top), the crystal structure of Cu-BTC is depicted. The secondary
building unit (SBU) and the processes occurring during dehydra-
tion and substrate coordination are schematically shown in Fig. 1
(bottom).
The accessibility of these sites has been proven both experi-
mentally [30] and based on simulations [31–33]. Consequently,
Cu-BTC has been used in some catalytic applications up to now.
Kaskel et al. reported on the cyanosilylation of carbonyl com-
pounds, namely benzaldehyde and acetone, at low temperatures
(40–80 °C) in the liquid phase [34]. The catalytic activity of the Le-
wis-acidic Cu centers in the water-free form Cu3(BTC)2 has been
has been demonstrated that this mixed-linker metal-organic
framework (MIXMOF) concept can be used to introduce catalyti-
cally active groups as side groups of the organic-linkers, which
are well distributed over the whole framework [8,10]. In contrast
to that, our target was to substitute a certain amount of ben-
zene-1,3,5-tricarboxylate (BTC) linker molecules by pyridine-3,5-
dicarboxylate (PyDC) in order to check whether these molecules
can be incorporated in the Cu-BTC structure without changing its
crystal structure. The pyridine group could then be seen as a defect
site in the crystal lattice, which leads to a paddle-wheel unit where
one of the bridging carboxylate groups is missing and replaced by
the nitrogen of the pyridine ring (Scheme 1). This should affect
electronic (e.g., the bond strength and distance of the coordinated
water molecules or the Cu–Cu distance) and consequently also the
catalytic properties of the Cu centers in the framework. We applied
X-ray-based techniques (X-ray diffraction and X-ray absorption
spectroscopy) as well as thermal analysis, infrared spectroscopy
and electron microscopy for structural investigations and used
the direct hydroxylation of benzene and its derivatives (both using
acetonitrile as the solvent and in neat substrate) as the target reac-
tion to demonstrate the catalytic properties of the resulting Cu-
BTC-PyDC MIXMOFs.
tested in the isomerization of
a-pinene oxide and the cyclization
of citronellal by De Vos and coworkers [18]. Also other Cu-based
metal-organic frameworks have been reported as catalysts in oxi-
dation reactions of, e.g., xanthene [20] and hydrochinone [35] or
the epoxidation of olefins [36]. While the synthesis of phenol itself
is straightforward via the cumene process, the hydroxylation of
substituted benzene derivatives is possible by multi-level proce-
dures, e.g., by the synthesis of the corresponding diazonium com-
pound and its hydrolysis. Therefore, a direct hydroxylation of
benzene or related aromatic compounds under mild conditions
and with a green oxidation reagent would be desirable. In one of
the rare examples of direct hydroxylation of phenol (to produce
hydrochinone, catechol, and traces of benzoquinone), a copper
complex immobilized in molecular sieves was utilized [37]. Since
this transformation is a quite demanding reaction, only very small
turnover frequencies in the range of 0.15–1.04 hÀ1 have been re-
ported in that work. For an overview on the activities in the field
of oxidation of aromatics, the reader is referred to the review by
Lücke et al. [38].
2. Experimental
2.1. Material synthesis
About 1.126 g (4.84 mmol) of Cu(NO3)2 Ã 2.5H2O was dissolved
in 25 ml of distilled water, and 0.491 g of benzene-1,3,5-tricarbox-
ylic acid (H3BTC; 2.33 mmol) was dissolved in 25 ml of N,N-
dimethylformamide (DMF). The two solutions were combined in
a 100-ml round-bottom flask that was put in a pre-heated oil bath
at a temperature of 100 °C for 4 h. The blue precipitate was then fil-
tered off, washed thoroughly with DMF and water to remove resid-
ual precursor species, and dried at 200 °C in order to remove the
DMF solvent. The Cu-BTC-PyDC catalysts have been synthesized
accordingly, with the only difference that 10, 20, 30, 40, 50, and
90 mol% of H3BTC, respectively, have been replaced by H2PyDC.
While pure Cu-BTC is turquoise in its as-synthesized form, it
turned to dark blue after drying at 200 °C. The substituted Cu-
BTC-PyDC samples have the same color like the original one with
the only difference that with increasing degree of substitution
the color becomes a bit less intense.
The main objective of the present work was to answer the ques-
tion whether the local coordination sphere of the copper centers in
the structure, and thus the electronic and catalytic properties of
Cu-BTC, could be modified by introducing a desired amount of
modified linker molecules in the structure. In previous works, it
Fig. 1. Elementary cell of Cu-BTC (top) and structural changes at the Cu centers occurring during dehydration and substrate coordination.

78
S. Marx et al. / Journal of Catalysis 281 (2011) 76–87
Scheme 1. Illustration of replacing BTC by PyDC.
2.2. Atomic absorption spectroscopy AAS
Data analysis was performed using scattering paths calculated
with FEFF 7.0 [39]. The feff input file was generated using crystal
structures of Cu-BTC. The calculated amplitudes and phase shifts
have been fitted to the experimentally obtained data. In a first step,
the reference spectra of Cu-BTC were fitted taking four next neigh-
bors into consideration.
Leaching of Cu was determined by AAS (atomic absorption
spectroscopy). The measurement was performed with a Varian
SpectrAA 220 FS spectrometer. The reaction mixture was filtered,
the organic and aqueous phases were separated, and the copper
content was determined by a calibration curve recorded before.
2.6. Thermogravimetry
2.3. IR investigations
Thermoanalytical (TA) experiments were carried out isother-
mally or non-isothermally (heating rates were generally in the
range of 5–10 K minÀ1) on a Netzsch STA 409 simultaneous ther-
mal analyzer. The determination of the thermal stability of the
MOFs was made in He atmosphere, using a total gas flow of
IR-investigations were performed using a Vertex 77 FT-IR spec-
trometer (Bruker Optics) in ATR mode. The powder was milled and
stored at atmospheric conditions for 30 min to warrant similar
conditions and comparability of the spectra. The spectrometer
was equipped with a liquid nitrogen cooled MCT detector and a
Platinum ATR unit. The resolution was set to 4 cmÀ1, and the spec-
tra were recorded in the range between 7500 and 500 cmÀ1. About
200 scans were averaged to one spectrum.
50 mL minÀ1
.
2.7. Catalytic tests in acetonitrile
The catalytic test reactions were performed using an eightfold
parallel batch reactor system (Endeavor Catalysts Screening
System). Each reactor was equipped with a glass inlet and a stir-
rer. The temperature and the pressure could be programmed for
each reactor separately. The reaction mixture was purged with
nitrogen, and a pressure of 2 bar of nitrogen was added to the sys-
tem. About 440 mg (4.78 mmol) of toluene (VWR, 99.9%), 12.4 mg
(0.097 mmol) of nonane (Fluka 99.0%) as internal standard, and
4.5 ml acetonitrile (Sigma–Aldrich, 99.5%) were mixed in a glass in-
let. Twenty milligram (0.033 mmol) of the catalyst was added to
the solution as well as 1.06 g of an aqueous solution of H2O2
(Merck, 30%), which corresponds to 2 eq. of peroxide based on
toluene.
2.4. Electron microscopy
Cu-BTC and Cu-BTC0.5PyDC0.5 have been placed on a carbon
foil. The samples were measured under high vacuum on a 1530
Gemini (Zeiss; field emission gun) operated at 1 kV.
2.5. X-ray absorption measurements
The experiments were performed at the Hamburger Synchro-
tronstrahlungs-Labor (HASYLAB) at Deutsches Elektronen-Syn-
chrotron (DESY) at DORIS III (4.45 GeV, 120 mA current) using
beam line X1 (energy range: 7–100 keV). The monochromator
was a Si(1 1 1) crystal for the measurement at the Cu K-edge
(8.979 keV). To prevent higher harmonics, the crystal was detuned
to 65%. All samples were measured in transmission mode. To mon-
itor the dehydration of the Cu-BTC, an in situ flow cell was filled
with the as-prepared catalyst (admixed with Al2O3 in order to ob-
tain the optimum Cu concentration for the XAS measurements). He
was passed through the cell which was heated to 162 °C while
XANES spectra were taken. The final temperature was kept until
no changes in two subsequent XANES spectra were discernible.
After cooling to room temperature, EXAFS spectra of the samples
were taken.
2.8. Catalytic tests under solvent-free conditions
The catalytic test reactions were performed in the same parallel
batch reactor system. In a typical experiment, 20 mg (0.033 mmol)
of the catalyst, 46.9 mmol of the substrate, and nonane (Fluka
99.0%) as internal standard were placed in the glass inlet. The ratio
of the amount of copper atoms to the amount of substrate was kept
at around 0.21 mol%. The experiment was started with the addition
of 30% H2O2 aqueous solution to the mixture. The amount of the
H2O2 solution was adjusted that the molar ratio of H2O2 to sub-
strate was 0.1, 0.2, 0.4, or 0.8 (corresponding to 0.476, 0.938,
1.877, and 3.754 ml of hydrogen peroxide solution). The reaction
mixture was purged with nitrogen, and a pressure of 2 bar of nitro-
gen was added to the system. Stirring at 250 rpm guaranteed an
intimate mixing of the phases. The reaction mixture was analyzed
The raw data were energy calibrated toward copper foil, back-
ground corrected using a two polynomial fit, and normalized. The
v
(k) function was extracted from the EXAFS data in the range be-
tween 0.5 and 16.0 ÅÀ1 and Fourier transformation was performed
on the k3 weighted data in the interval k = 2.6–14.9 ÅÀ1
.

S. Marx et al. / Journal of Catalysis 281 (2011) 76–87
79
after separation of the phases and drying of the organic phase over
MgSO4 with a HP 6890 GC system on a HP-5 column (30 m)
equipped with a FID. A temperature program has been applied
starting at 40 °C increasing to 250 °C after 5 min.
3. Results and discussion
3.1. Structural investigations of Cu-BTC and the Cu-BTC-PyDC catalysts
3.1.1. Cu-BTC
The Cu-BTC catalyst used in this work was synthesized using a
modified experimental procedure at low temperature (100 °C)
using water/N,N-dimethylformamide (DMF) mixtures as the sol-
vent instead of water/ethanol. In contrast to the common solvo-
thermal routes, this low-temperature synthesis could be done via
precipitation in a round-bottom flask, because of the good solubil-
ity of the organic precursor in DMF (in contrast to ethanol). This
method has been chosen to allow for a controlled precipitation of
the catalyst in order to avoid the formation of undesired phases.
Furthermore, the applied soft precipitation method should be ben-
eficial for the synthesis of the mixed BTC-PyDC catalysts. The as-
synthesized material was subsequently dried at 200 °C in order
to remove the DMF solvent, which might be bound more strongly
than water. The measured BET surface area of the dried material
(1700 m2/g) was comparable to the values reported in the litera-
ture. To prove that solvent molecules could be removed from the
Cu centers by thermal treatment, in situ X-ray absorption studies
were performed to investigate the local environment of the Cu cen-
ters in the structure. To give further evidence to this claim, the sub-
sequent dehydration process of the material was monitored by
in situ XANES spectroscopy.
In Fig. 2 three exemplary XANES spectra at 33 °C, 109 °C, and
162 °C are depicted. The white line for the hydrated sample is
found at 8.9975 keV and shifts to 8.9982 keV for the dehydrated
one. Additionally, the intensity of the white line decreases upon
heating. The blue shift and the decrease in the white line intensity
indicate the decrease in the coordination number of the copper
atoms due to the loss of coordinated water. A second indication
for the removal of water is given by the increase and red shift of
the pre-edge feature at around 8.985 keV (inset in Fig. 3). This
feature corresponds to a 1s to 4p transition combined with a
charge transfer from the ligand to the metal. Both effects together
give a so-called shakedown transition [1s(2). . .3d(9)4p(0)L0] to
[1s(1). . .3d(10)4p(1)L1+]. Upon the loss of a coordinated water
molecule, the copper atom becomes more electrophilic, resulting
Fig. 3. Fourier-transformed EXAFS spectra (k3 weighted) of the as-prepared and the
dehydrated Cu-BTC.
in a strengthening of the Cu–Ocarboxylate bond. The stronger binding
of the ligand to the copper facilitates the shakedown transition,
resulting in the observed increase and red shift of this pre-edge
feature from 8.9848 keV to 8.9841 keV [40].
After the completion of the drying process, an EXAFS measure-
ment of the dried sample was performed (Fig. 3), which revealed
that the Fourier-transformed spectra of the as-prepared and the
dried sample differ slightly. For the activated material, a backscat-
ter contribution occurs as a shoulder at low R values, indicating a
shortening of the Cu–Ocarboxylat bonds after water removal. Addi-
tionally, the amplitude of the dried sample is less intense than
the one of the as-prepared sample hinting at the presence of less
next neighbors around the copper center.
Both methods indicate the removal of water and the generation
of a free coordination site (Fig. 1) in the material synthesized in
this work. This means that the copper atoms do not only serve as
SBU (secondary building unit) but could act as active centers in a
catalytic reaction.
3.1.2. Cu-BTC-PyDC
The influence of the substitution of benzene-1,3,5-tricarboxy-
late (BTC) by pyridine-3,5-dicarboxylate (PyDC) on the structure
has been studied by XRD and XAS techniques (see Scheme 1). By
exchanging one BTC molecule by PyDC, one carboxylate group
(out of three) of one linker molecule is replaced by one nitrogen
atom in the aromatic ring. Thus, the paddle-wheel structure should
become distorted, and a gap should occur between the copper di-
mer and the linker molecule. Up to 50% of the BTC linkers could
be introduced into the framework without changing the structure
of the original Cu-BTC as indicated by XRD (Fig. 4). Since higher ra-
tios of PyDC (e.g., 90%) resulted in the formation of a completely
different crystal structure, we used only the materials containing
10%, 20%, 30%, and 50% of PyDC for the further studies.
To gain a deeper understanding of the geometric and electronic
changes in the modified framework, EXAFS and XANES investiga-
tions have been performed. Comparison of Cu-BTC and Cu-
BTC0.5PyDC0.5 XANES spectra reveals several differences (Fig. 5).
The white line intensity of the substituted MOF is smaller com-
pared to the pure MOF indicating less next neighbors, similar to
observation made for dehydration of the MOF. In difference to
dehydration, substitution leads to a red shift of the white line by
about 0.0009 keV. Additionally, the pre-edge feature found for
Cu-BTC at 8.985 keV (see inset in Fig. 5) is not visible for the substi-
tuted MOF. Thus, the change in the coordination surrounding leads
to an increase in the electron density at the copper atoms [30].
Fig. 2. In situ XANES spectra of the dehydration process of Cu-BTC. The inset
magnifies the pre-edge feature attributed to a 1s ? 4p dipole shakedown transition.

80
S. Marx et al. / Journal of Catalysis 281 (2011) 76–87
Fig. 6. Fourier-transformed EXAFS spectra (k3 weighted) of Cu-BTC and substituted
Cu-BTCxPyDC1Àx samples.
Fig. 4. Powder X-ray diffraction pattern of pure Cu-BTC and the Cu-BTC-PyDC
materials; the percentage of BTC linkers that were replaced by PyDC is marked on
the curves.
sample with 50% PyDC, since half of the BTC ligands are replaced
by PyDC, meaning that 1/6 of all carboxylate ligands are missing.
Therefore, the average coordination numbers NCu–O1 and NCu–C1
should be 3.3, which is close to the value of 3.4 that is found by
simulation (Table 1). The amount of water bonded to the Cu atoms
stays constant for up to 30% PyDC, and a slightly higher coordina-
tion number of 1.2 was found for the last sample. We assume that
these additional water molecules might coordinate at the distorted
Cu centers, which would otherwise have only a coordination num-
ber of five. Consequently, the framework structure should be less
rigid compared to the pure Cu-BTC, which should also alter the
possibilities of substrate coordination during catalytic processes.
As mentioned above, a higher degree of substitution (e.g., 90% of
PyDC) leads to restructuring and the formation of a different crys-
tal phase, whereas the Cu-BTC structure stays intact when up to
50% of the linker is substituted by PyDC. In addition, the distance
between the two Cu atoms in the dimer becomes smaller by
0.3 Å, while the distance between Cu and C1 increases by 0.6 Å with
increasing PyDC content. Thus, a strain on the SBU as well as elec-
tronic effects upon substitution could be proven by XAS analysis.
Nevertheless, the question for the balance of the remaining po-
sitive charge is rather demanding. Upon substitution of BTC by
PyDC, one carboxylate group is missing and a positive charge re-
mains in that structure. Thus, residual nitrate ions or hydroxide
species might compensate the additional charge. On the other
hand, a certain amount of Cu2+ ions could in principle be reduced
to Cu+ to balance the charge. The change of the valence state seems
unlikely, since the XANES investigations rule out the existence of
Cu+ ions in the material. In the case of Cu-BTC0.5PyDC0.5, 1/6 of
all Cu ions should then be Cu+, which should be visible in the
XANES spectra (I) as a shift of the edge energy by 0.0035 keV to
lower values, (II) by an increase in the pre-edge feature at
8.984 keV, and (III) by a significant decrease in the white line
intensity [41]. Therefore, the presence of nitrate or hydroxide ions
as source of the additionally needed negative charges is likely
although it was not unambiguously possible to prove their incor-
poration with EXAFS and to determine their exact location.
Thermogravimetric results are shown in Fig. 7 and Fig. S1 in
Supporting information (SI), which clearly indicate that pure Cu-
BTC is the most stable compound and that the thermal stability
of the substituted materials decreased with increasing PyDC con-
tent (Fig. 7a). Thus, the onset of decomposition of Cu-BTC was
around 300 °C, while Cu-BTC0.5PyDC0.5 started to decompose al-
ready at 270 °C (Fig. 7b). Water could be removed between 50 °C
and 150 °C being in good agreement with the XANES investigation,
where the water was removed after reaching a temperature of
160 °C.
Fig. 5. Comparison of the XANES region of Cu-BTC (black line, —) and selected Cu-
BTC-PyDC materials. The inset magnifies the pre-edge feature attributed to
a
1s ? 4p dipole shakedown transition.
In Fig. 6, the Fourier-transformed EXAFS spectra of pure Cu-BTC
and the modified Cu-BTC-PyDC materials are depicted. With
increasing PyDC content, the amplitude becomes smaller, indicat-
ing a decrease in the coordination number of the copper atoms.
At larger distances, the scattering pattern shows also slight differ-
ences especially for the transition from the unmodified MOF to the
substituted ones. To quantify the changes visible in the EXAFS pat-
tern, the spectra have been fitted (Table 1).
Two assumptions have been made for the fitting procedure.
First, for a Cu-BTC-PyDC material, the coordination numbers
NCu–O1 and NCu–C1 should be identical, since the removal of a car-
boxylate ligand reduces the number of O1 and C1 atoms in the same
manner. Second, the coordination number NCu–Cu should be one in
all cases. The results of fitting for the pure Cu-BTC and the mixed
Cu-BTC-PyDC MOFs with 10%, 30%, and 50% of PyDC reveal several
changes in the local structure around the Cu atoms. As expected,
the coordination numbers NCu–O1 and NCu–C1 were decreasing upon
increasing the amount of PyDC. While a coordination number of
4.0 (which is expected for the unmodified Cu-BTC) is found for
the samples containing 0% and 10% of PyDC, respectively, this value
changes to 3.6 and 3.4 for the samples with 30% and 50% of PyDC.
The decrease in NCu–O1 and NCu–C1 should be around 0.66 for the

S. Marx et al. / Journal of Catalysis 281 (2011) 76–87
81
Table 1
Results of the EXAFS fit parameters for Cu-BTC and Cu-BTCxPyDC1Àx at the Cu K-edge; N = coordination number, R = distance (Å),
D
r
2 = Debey-Waller factor (Å2),
D
E0 = inner core
correction (keV Ã 103).
Sample
Cu-BTC
Edge
Cu K
Pair
N
R
D
r2
D
E0
Residual
4.9
Cu–O1
Cu–O2
Cu–Cu
Cu–C1
4.0
1.0
1.0
4.0
1.96
2.01
2.61
2.80
0.0043
0.0050
0.0068
0.0073
9.0
À4.8
9.9
8.8
10% PyDC
30% PyDC
50% PyDC
Cu K
Cu K
Cu K
Cu–O1
Cu–O2
Cu–Cu
Cu–C1
4.0
1.0
1.0
4.0
1.96
2.08
2.60
2.80
0.0044
0.0080
0.0080
0.0050
9.4
9.6
3.4
2.8
3.7
À0.2
À2.9
Cu–O1
Cu–O2
Cu–Cu
Cu–C1
3.6
1.0
1.0
3.6
1.95
2.04
2.59
2.84
0.0048
0.0083
0.0108
0.0063
8.4
9.7
À9.9
À4.0
Cu–O1
Cu–O2
Cu–Cu
Cu–C1
3.4
1.2
1.0
3.4
1.95
2.00
2.58
2.86
0.0054
0.0070
0.0111
0.0076
8.3
8.3
À8.1
1.2
the signal assigned to the BTC linkers) is in perfect accordance with
the higher degree of substitution, proving the incorporation of both
types of linker into the structure.
IR spectra of the series of MOFs are depicted in Fig. 8 and Fig. S2.
In the range between 1700 cmÀ1 and 1500 cmÀ1, the signals for the
antisymmetric stretch vibrations of the carboxylate group [
m
as(O-
CO)] of BTC and PyDC are discernible. For the BTC ligand,
mas(OCO)
is located at around 1640 cmÀ1 [42]. All six samples exhibited the
signal for BTC as shown in Fig. 8a, and only a small red shift of 20
wave numbers was observed upon substitution with PyDC. The sig-
nal for mas(OCO) of PyDC is normally found at lower wave number
[43]. Thus, the signal at 1560 cmÀ1 was assigned to the carboxylate
group of PyDC. This signal is increasing with increasing PyDC con-
tent. A second characteristic signal is shown in Fig. 8b. The C-H in
plane bending mode of the aromatic ring in Cu-BTC is observed at
1105 cmÀ1, which is in good agreement with literature values [42].
With increasing PyDC content, a signal at 1096 cmÀ1 appeared,
indicating the presence of a second linker with a different substitu-
tion pattern of the aromatic ring.
In summary, IR studies, TG analysis as well as EXAFS investiga-
tions corroborated the incorporation of both linker molecules into
the framework structure. Powder XRD indicated that the crystal
structure was sustained up to a substitution degree of 50%. We also
investigated whether the exchange of BTC by PyDC had an influ-
ence on the morphology of the crystals using scanning electron
microscopy. SEM images of Cu-BTC and Cu-BTC0.5PyDC0.5 com-
pared in Fig. 9 and Fig. S3 do not indicate a significant change in
the morphology up to the maximum possible substitution degree.
3.2. Catalytic tests of Cu-BTC and the modified Cu-BTC-PyDC materials
Fig. 7. (A) Thermogravimetric measurements of MOFs with 0, 10, 30, and 50 mol%
PyDC; (B) differential thermogravimetric measurements of MOFs with 0, 10, 30, and
50 mol% PyDC.
We have chosen the demanding direct hydroxylation of ben-
zene derivatives, e.g., toluene, in the presence of hydrogen perox-
ide (Scheme 2) as the target reaction to check the potential of
the modified Cu-BTC materials in catalysis and the influence of
the structural ‘‘defects’’ that have been introduced by the substitu-
tion of BTC by PyDC. The chosen test reaction is of importance
since phenol and its derivatives are versatile building blocks for
the synthesis of pharmaceuticals, polymers, and natural products
[44].
Evaluation of the DTG data (Fig. 7b) revealed that the thermal
decomposition of the mixed Cu-BTC-PyDCs occurred in two steps.
While only one maximum of the decomposition rate (at ca. 330 °C)
was found for the pure Cu-BTC, a second feature at 280–290 °C ap-
peared for the mixed-linker MOFs. Assuming that the missing car-
boxylate group of the PyDC results in a weaker interaction of the
pyridine function with the Cu centers (compared to the BTC linker),
it is reasonable that the structure starts to break at these defect
sites already at lower temperatures. The increasing intensity of this
low temperature signal (combined with the decreasing intensity of
3.2.1. Oxidation of toluene in acetonitrile
First, the oxidation of toluene using H2O2 as the oxidant
(Scheme 2) was studied in acetonitrile, which was chosen as the

82
S. Marx et al. / Journal of Catalysis 281 (2011) 76–87
Fig. 9. Comparison of morphologies of Cu-BTC (A) and Cu-BTC0.5PyDC0.5 (B), as
emerging from SEM investigations (white bar indicates 200 nm).
Fig. 8. A) IR spectra of MOFs with a molar PyDC content between 0% and 50%;
frequency range for
framework (gray area indicates the signal for
of the C–H in plane bending mode of BTC and PyDC.
m
as(OCO) of the BTC and PyDC ligand in the metal-organic
as(OCO) of PyDC); B) frequency range
toluene could not be observed under the chosen conditions. In
the presence of the dissolved Cu salts, the reaction mixture chan-
ged its color to deep red even at low reaction temperatures (40
or 50 °C) and a huge variety of side products could be found in
GC analysis. From these observations, we conclude that the heter-
ogeneous MOF catalyst represents the catalytically active species
facilitating a direct hydroxylation of toluene with high selectivity
toward the cresols. However, a possible contribution of leached
species cannot be ruled out completely. By means of AAS, 3% of
the total amount of copper was detected in the aqueous phase,
though these dissolved Cu species will probably favor the unde-
sired side reactions toward benzaldehyde or methylbenzoquinone.
Despite the observed leaching, catalyst separation and copper con-
tamination of the product do not represent a major problem, since
soluble copper species were only found in the aqueous phase. The
organic phase did not contain Cu in concentrations that could be
detected within the limits of AAS.
The product selectivity for the oxidation of toluene in the pres-
ence of Cu-BTC using acetonitrile as the solvent is visualized in
Fig. 10. Under these conditions, significant amounts of the side
products methylquinone, benzaldehyde, and benzyl alcohol were
observed in addition to the desired ortho- and para-cresol. Never-
theless, with increasing temperature (from 40 °C to 80 °C), the for-
mation of ortho- and para-cresol became favored again on the cost
of benzaldehyde formation. In contrast to the solvent-free studies
(vide infra), where the ratio of ortho- and para-cresol depended
on temperature (with ortho-cresol being favored at higher temper-
atures), the ortho-cresol was favored over the whole temperature
range. Obviously, the reaction temperature is strongly influencing
the chemoselectivity, whereas the regioselectivity is only affected
m
solvent because of the solubility of both reactants (toluene and
hydrogen peroxide) in this medium.
In Table 2, the conversion of toluene is listed under different
reaction conditions. As expected, higher temperatures and longer
reaction times are beneficial for the conversion. Almost no conver-
sion was observed at 40 °C (0.14%, Table 2, entry 1) and 50 °C
(0.21%, Table 2, entry 5) after one hour of reaction. Increasing the
temperature to 80 °C resulted, however, in a slightly higher con-
version of ca. 1.27% (Table 2, entry 13). Furthermore, conversion
could also be more than doubled by doubling the reaction time
(e.g., from 0.62% to 1.51% at 70 °C, Table 2, entries 11 and 12).
Much longer reaction times (18 h) turned, however, out to be not
very efficient, which can simply be explained by the fact that the
H2O2 oxidant is decomposed thermally after a certain time. Finally,
we discovered that a further dilution of the substrate concentra-
tion (from 10 to 5 vol.%) had a beneficial influence on the conver-
sion, which could be increased from 1.27% to almost 5% after one
hour at 80 °C (Table 2, entries 13 and 14). Note that this latter value
corresponds to a TOF of 2.37, which is more than twice the value
that has been reported in the work of Raja and Ratnasamy [37].
The catalytic results obtained from the reaction with Cu-BTC
were also compared with homogeneously dissolved Cu salts in or-
der to classify the catalytic performance of the MOF catalyst.
Therefore, Cu(NO3)2 and Cu(OAc)2 have been applied under identi-
cal reaction conditions, using the same amount of Cu that is pres-
ent in the Cu-BTC catalyst. However, a selective hydroxylation of

S. Marx et al. / Journal of Catalysis 281 (2011) 76–87
83
Scheme 2. Reaction scheme for the oxidation of toluene.
Table 2
Conversion of toluene in acetonitrile as solvent under different reaction conditions
(catalyst: Cu-BTC, eq. H2O2 = 2).
Entry
Temperature (°C)
Time (h)
Substrate (vol.%)
Conversion (%)
1
2
3
4
5
6
7
8
9
10
11
12
13
14a
40
40
40
40
50
50
50
60
60
60
70
70
80
80
1
2
3
18
1
3
18
1
2
3
1
10
10
10
10
10
10
10
10
10
10
10
10
10
5
0.14
0.22
0.39
0.46
0.21
1.0
1.66
0.2
0.46
3.7
0.9
1.5
1.27
4.91
2
1
1
a
The amount of hydrogen peroxide was kept constant; thus 4 equivalents were
used.
Fig. 11. Influence of the MOF composition on the product distribution. Substrate:
toluene, solvent: acetonitrile, vol.% substrate = 10%, T = 60 °C, reaction time = 1 h,
2 eq. H2O2. (j) = benzaldehyde, (d) = o-cresol, (H) = p-cresol, (Ã) = benzyl alcohol,
(s) = methylquinone. The error bars give an indication of the reproducibility of the
experiments.
product (29%). Therefore, the following studies were performed
under solvent-free conditions (in neat substrate) in order to obtain
a high selectivity toward the desired cresols.
The substituted Cu-BTC-PyDC materials were also tested in the
hydroxylation of toluene. Interestingly, the conversion for the
substituted MOFs was much higher and reached more than 3% at
60 °C compared to 0.3% for the unmodified Cu-BTC (see Table 2, en-
try 8 for comparison).
In addition to the substrate conversion, also the product selec-
tivity was significantly influenced by the introduction of PyDC to
the framework (Fig. 11). Pyridine did not only influence the elec-
tronic properties of the copper ions, but might also have changed
some other properties of the material, e.g., its basicity by the pres-
ence of the pyridine moiety. Thus, the observed selectivity for the
pure Cu-BTC was significantly different from that found in case of
the modified Cu-BTC-PyDCs. Noticeably, the degree of substitution
did not have a major effect, as more or less the same product dis-
tribution was achieved for all MIXMOFs. For all Cu-BTC-PyDC
materials, a distinct increase in the methylbenzoquinone selectiv-
ity was observed, while the concentration of the two cresols
dropped (from 28% and 22%, respectively, for Cu-BTC to ca. 22%
and 15% for the materials substituted with 20% or more of PyDC).
Apparently, under these conditions, further oxidation of the aro-
matic ring was favored over its hydroxylation. Independent of
these trends, benzaldehyde (30–38%) was the main product with
all catalysts.
Fig. 10. Influence of the temperature on the selectivity. Substrate: toluene, catalyst:
Cu-BTC, solvent: acetonitrile, vol.% substrate = 10%, time: 1 h, 2 eq. H2O2.
(j) = benzaldehyde, (d) = o-cresol, (H) = p-cresol, (Ã) = benzyl alcohol, (s) = meth-
ylquinone. The values given on the right side of the diagram indicate the product
distribution of a diluted sample (T = 80 °C, vol.% substrate = 5%, catalyst = Cu-BTC,
time = 1 h, 4 eq. H2O2). The error bars give an indication of the reproducibility of the
experiments.
to a minor extent. The aldehyde (side chain oxidation) was found
as the main product at low temperatures, while at higher temper-
ature the quinone (further oxidation of the ring) was favored. A
further dilution of the substrate from 10 vol.% to 5 vol.% led to an
increased conversion (1.27% vs. 4.91%, Table 2, entries 13 and 14)
but had a negative influence on the selectivity (Fig. 10). While
ortho- (32%) and para-cresol (26%) were favored over benzalde-
hyde (24%) for the high substrate concentration, dilution of the
substrate caused a decreased selectivity toward the cresols (27%
and 22%, respectively), and benzaldehyde was found as the main
In addition to the parameter studies, recycling experiments
have been performed with Cu-BTC and Cu-BTC0.6PyDC0.4 at 60 °C.
For pure Cu-BTC, an improvement of the conversion from 0.2%

84
S. Marx et al. / Journal of Catalysis 281 (2011) 76–87
( 0.1) for the first run up to 1.42% ( 0.002) was observed. Only
slight changes in the selectivity have been observed for the recycle
experiments (SI, Fig. S4). The powder XRD pattern of the used Cu-
BTC material (SI, Fig. S5) indicated no structural change in the
material after four experiments.
Cu-BTC0.6PyDC0.4 did not exhibit a significant change in conver-
sion after 3 runs (average conversion of three runs = 2.7%, standard
deviation of the three runs = 0.3%) or selectivity upon reuse (SI,
Fig. S6).
temperature. At low temperatures of 40 or 50 °C, the highest selec-
tivity after one hour was obtained for para-cresol (47%, compared
to 34% of the ortho-product; Fig. 13). With increasing temperature
(starting at 50–60 °C), a decreased selectivity to para-cresol was
observed, and the ortho-product became the dominant one. At a
reaction temperature of 80 °C, para- and ortho-cresol were formed
with a selectivity of 34% and 41%, respectively. No conversion was
observed in the absence of the catalyst (SI, Fig. S8 top and bottom).
3.2.2.2. Variation of Substrates. In addition to toluene, a variety of
other aromatic substrates were used in the test reaction to check
scope and limitations. The non-substituted benzene should yield
only phenol as the main product and benzoquinone as side prod-
uct. Compared to toluene, higher conversions could be obtained
for benzene (vide infra). In accordance with the toluene reaction,
the total conversion increased at higher temperatures (SI, Fig. S9
top), while the selectivity toward phenol decreased. A similar trend
was found with increased H2O2 amount (SI, Fig. S9 bottom). How-
ever, using an optimized reaction temperature of 40 °C phenol
could be detected as the only product of the reaction after one
hour. No benzene conversion was observed in the absence of the
catalyst.
In accordance with the findings made for toluene, also xylenes
could be converted in the test reaction, albeit at very little conver-
sion. 2,5-dimethylphenol was the only product observed in the
reaction of para-xylene, whereas the reaction of ortho-xylene gave
surprising results concerning the regioselectivity. While exclu-
sively 2,3-dimethylphenol was obtained at 40 °C, the selectivity to-
ward this isomer (41%) was reduced significantly at a reaction
temperature of 50 °C, and the other regioisomer 3,4-dimethylphe-
nol (59%) became the main product. The decrease of conversion in
the order benzene > toluene > p-xylene > o-xylene indicates that
the size and steric hindrance of the reactants may play a role and
suggests that diffusion processes might be a limiting factor for
the efficiency of this catalytic reaction (Table 3).
3.2.2. Solvent-free oxidation of aromatic substrates
3.2.2.1. Influence of reaction temperature and oxidant concentra-
tion. Based on the previous findings, Cu-BTC was tested in the sol-
vent-free oxidation reaction of aromatic compounds (e.g., neat
toluene) using hydrogen peroxide as the oxidant. Note that the
overall conversion of toluene and the other substrates under these
conditions was <1%, caused by the huge excess of the substrate. As
already indicated by the dilution experiments (see Section 3.2.1),
the catalyst did not significantly promote the oxidation of the side
chain to obtain benzyl alcohol, benzaldehyde, or benzoic acid, but a
direct hydroxylation of the aromatic ring could be observed with a
strong selectivity toward the para- and ortho-substituted products
in the absence of a solvent. In the reaction of toluene at 80 °C, only
small quantities of benzaldehyde and methylquinone were found
in addition to the main products ortho- and para-cresol. The che-
mo- and regioselectivity of the test reaction were investigated un-
der several reaction conditions.
In Fig. 12, the selectivity toward ortho-, para-cresol, benzalde-
hyde and methylbenzoquinone at 80 °C is depicted for different
amounts of hydrogen peroxide and after different reaction times.
After one hour of reaction time, ortho-cresol (41% selectivity) was
the main product, followed by para-cresol (38%), methylbenzoqui-
none (11%), and benzaldehyde (10%) using the lowest oxidant con-
centration. Increasing the H2O2 quantity led to an increased
production of the over-oxidized product methylbenzoquinone
(21%), while the selectivity of the two cresols decreased (37% and
28%). The same trend was observed after reaction times of two
and three hours, respectively. (SI Fig. S7 top and bottom). There-
fore, 0.2 equivalents of H2O2 (with respect to the substrate) were
used in the further tests.
In addition to toluene and benzene, also other substrates have
been screened in the reaction to investigate possible electronic
and steric influences. The results are listed in Table 3. A compari-
son of the three aryl halides (fluoro-, chloro-, and bromobenzene)
showed that also for these substrates, the para-isomer was ob-
tained with highest selectivity at 40 °C (53–61%, see Table 3). In
contrast to the observations made for toluene, increasing the tem-
perature to 50 °C did not significantly change the selectivity for
chloro- and bromobenzene. Interestingly, higher reaction temper-
ature resulted in higher para-selectivity (69% at 50 °C compared
Besides the oxidant concentration, also temperature affected
the product distribution significantly. In contrast to the former,
over-oxidation to quinone was not a major problem here (e.g., 4%
at 40 °C and 11% at 80 °C). However, the regioselectivity of the cre-
sols could be influenced significantly by variation in the reaction
Fig. 12. Influence of the oxidant concentration on the product selectivity after one
hour reaction time. Substrate: toluene, catalyst: Cu-BTC, no solvent, T = 80 °C, vol.%
substrate = 10%.
Fig. 13. Influence of the reaction temperature on the product selectivity after one
hour reaction time. Substrate: toluene, conditions: 0.2 equivalents of H2O2,
catalyst = Cu-BTC.

S. Marx et al. / Journal of Catalysis 281 (2011) 76–87
85
Table 3
Solvent-free oxidation of various aromatic substrates. (Catalyst: Cu-BTC, reaction time = 1 h; eq. H2O2 = 0.2).
Entry
1
Substrate
Toluene
T (°C)
Conversion (%)
0.04
Products
Selectivity (%)
40
Benzaldehyde
Methylbenzoquinone
o-Cresol
12
4
36
47
p-Cresol
2
50
0.07
Benzaldehyde
Methylbenzoquinone
o-Cresol
10
5
39
46
p-Cresol
3
4
Benzene
40
50
0.04
0.11
Phenol
Benzoquinone
>99.9
<0.1
Phenol
Benzoquinone
90
10
5
Para-xylene
Ortho-xylene
40
50
0.003
0.006
2,5-Dimethylphenol
2,5-Dimethylphenol
>99.9
>99.9
6
40
50
40
0.001
0.004
0.003
0.009
0.016
0.06
2,3-Dimethylphenol
3,4-Dimethylphenol
>0.1
<99.9
7
2,3-Dimethylphenol
3,4-Dimethylphenol
41
59
8
Fluorobenzene
Chlorobenzene
Bromobenzene
Para-fluorophenol
Ortho-fluorophenol
61
39
9
Para-fluorophenol
Ortho-fluorophenol
69
31
10
11
12
13
Para-chlorophenol
Ortho-chlorophenol
53
47
50
40
50
Para-chlorophenol
Ortho-chlorophenol
53
47
0.014
0.03
Para-bromophenol
Ortho-bromophenol
57
43
Para-bromophenol
Ortho-bromophenol
56
44
to 61% at 40 °C) for fluorobenzene. No quinone species could be de-
tected for all three aryl halides under the reaction conditions used.
Since the conversion of the halogenated aromatic compounds was
much smaller compared to toluene, with fluorobenzene having the
smallest conversion, it is assumed that the electron-withdrawing
properties of the halide substituents deactivate the aromatic com-
pound for that reaction.
3.2.2.3. Mechanistic considerations. In order to gain some insight
whether the reaction proceeds via a radical mechanism or not,
Scheme 3. Feasible mechanism for the hydroxylation of aromatic compounds (exemplarily shown for benzene).

86
S. Marx et al. / Journal of Catalysis 281 (2011) 76–87
hydrochinone was added to the reaction mixture as a scavenger.
Under solvent-less conditions, the presence of hydrochinone did
not stop the hydroxylation of toluene, indicating a non-radical
reaction mechanism.
(Electron Microscopy ETH Zurich) are acknowledged for micro-
scope time and recording of SEM pictures and Dr. Marek Maciejew-
ski (ETH Zurich) for performing the TG analysis.
This assumption is supported by the absence of meta-cresol,
which could be a possible product in a radical reaction, and the
low selectivity toward oxidation of the methyl group of toluene
that should be favored due to stabilization of the radical in benzylic
position. Further evidence of an ionic mechanism is given by the
fact that the conversion of fluorobenzene is much smaller com-
pared to that of chlorobenzene or bromobenzene. Fluorine is
known to be a substituent that deactivates aromatic compounds
for the electrophilic, aromatic substitution. Taking the aforemen-
tioned arguments into account, we favor an ionic pathway similar
to an electrophilic, aromatic substitution. Hydrogen peroxide is
activated by the copper ions, and a species like OH+ is transferred
to the aromatic system [45,46]. After release of a proton, the
hydroxylated aromatic product is generated, while water is re-
leased as by-product (Scheme 3) Nevertheless, an activation of a
hydrogen peroxide molecule at a Cu dimer can be virtually ex-
cluded, since the water molecules are axially located at the two
copper centers (Fig. 1, bottom) and thus antipodal (180° angle)
and pointing toward different pores. Therefore, the calculation of
TOFs is done with the assumption that a single Cu ion activates
one hydrogen peroxide molecule.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jcat.2011.04.004.
References
[1] U. Mueller, M. Schubert, F. Teich, H. Puetter, K. Schierle-Arndt, J. Pastré, J.
Mater. Chem. 16 (2006) 626–636.
[2] J.Y. Lee, O.K. Farha, J. Roberts, K.A. Scheidt, S.B.T. Nguyen, J.T. Hupp, Chem. Soc.
Rev. 38 (2009) 1450–1459.
[3] M. Sabo, A. Henschel, H. Fröde, E. Klemm, S. Kaskel, J. Mater. Chem. 17 (2007)
3827–3832.
[4] S. Hermes, M.K. Schröter, R. Schmid, L. Khodeir, M. Muhler, A. Tissler, R.W.
Fischer, R.A. Fischer, Angew. Chem. Int. Ed. 44 (2005) 6237–6241.
[5] S. Hermes, F. Schröder, S. Amirjalayer, R. Schmid, R.A. Fischer, J. Mater. Chem.
16 (2006) 2464–2472.
[6] S. Opelt, S. Türk, E. Dietzsch, A. Henschel, S. Kaskel, E. Klemm, Catal. Commun.
9 (2008) 1286–1290.
[7] F. Schröder, S. Henke, X. Zhang, R.A. Fischer, Eur. J. Inorg. Chem. (2009) 3131–
3140.
[8] W. Kleist, M. Maciejewski, A. Baiker, Thermochim. Acta 499 (2010) 71–78.
[9] J. Gascon, U. Aktay, M.D. Hernandez-Alonso, G.P.M. van Klink, F. Kapteijn, J.
Catal. 261 (2009) 75–87.
[10] W. Kleist, F. Jutz, M. Maciejewski, A. Baiker, Eur. J. Inorg. Chem. (2009) 3552–
3561.
[11] A.D. Burrows, C.G. Frost, M.F. Mahon, C. Richardson, Angew. Chem. Int. Ed. 47
(2008) 8482–8486.
4. Conclusions
[12] Z. Wang, S.M. Cohen, Angew. Chem. Int. Ed. 47 (2008) 4699–4702.
[13] Z. Wang, S.M. Cohen, J. Am. Chem. Soc. 131 (2009) 16675–16677.
[14] X. Zhang, F.X. Llabrés i Xamena, A. Corma, J. Catal. 265 (2009) 155–160.
[15] C.D. Wu, A. Hu, L. Zhang, W. Lin, J. Am. Chem. Soc. 127 (2005) 8940–8941.
[16] C.D. Wu, W. Lin, Angew. Chem. Int. Ed. 46 (2007) 1075–1078.
[17] C.D. Wu, L. Zhang, W. Lin, Inorg. Chem. 45 (2006) 7278–7285.
[18] L. Alaerts, E. Séguin, H. Poelman, F. Thibault-Starzyk, P.A. Jacobs, D.E. De Vos,
Chem. Eur. J. 12 (2006) 7353–7363.
[19] F.X. Llabrés i Xamena, A. Abad, A. Corma, H. Garcia, J. Catal. 250 (2007) 294–
298.
[20] A. Dhakshinamoorthy, M. Alvaro, H. Garcia, J. Catal. 267 (2009) 1–4.
[21] S.S.Y. Chui, S.M.F. Lo, J.P.H. Charmant, A.G. Orpen, I.D. Williams, Science 283
(1999) 1148–1150.
[22] D.J. Tranchemontagne, J.R. Hunt, O.M. Yaghi, Tetrahedron 64 (2008) 8553–
8557.
[23] A. Vishnyakov, P.I. Ravikovitch, A.V. Neimark, M. Bülow, Q.M. Wang, Nano Lett.
3 (2003) 713–718.
[24] E. Biemmi, S. Christian, N. Stock, T. Bein, Micropor. Mesopor. Mater. 117 (2009)
111–117.
We have synthesized new MIXMOF materials based on the
structure of the metal-organic framework Cu-BTC. Up to 50% of
the benzene-1,3,5-tricarboxylate (BTC) linkers could be replaced
by pyridine-3,5-dicarboxylate (PyDC) without changing the crystal
structure of the resulting materials, as confirmed by X-ray powder
diffraction. X-ray absorption spectroscopy (XAS) revealed that the
incorporation of the pyridine moiety into the crystal lattice leads to
a local defect site at some of the dimeric Cu units, where normally
a bridging carboxylate group would be coordinated. These struc-
tural defects cause a modified chemical and electronic environ-
ment of the metal centers, which consequently also affects their
accessibility and activity in catalytic applications.
Cu-BTC and the substituted Cu-BTC-PyDC materials were found
to catalyze the direct hydroxylation of aromatic compounds at low
temperatures. In contrast to homogeneously dissolved Cu salts, a
surprisingly high selectivity toward ortho- and para-cresol was ob-
served in the reaction of toluene and hydrogen peroxide. As ex-
pected, the product selectivity was significantly different for pure
Cu-BTC and the mixed-linker materials. Using acetonitrile as the
solvent, a maximum turnover frequency of up to 2.37 hÀ1 could
be reached which is twice as high as TOFs that have been reported
in literature before. The toluene concentration had a dramatic and
opposed influence on the conversion and the product selectivity.
Lower substrate concentrations led to an increased conversion,
but also to an increased selectivity toward the undesired side prod-
ucts benzaldehyde and methylbenzoquinone. Consequently, high
selectivity toward the cresol products was observed under sol-
vent-free conditions. Using optimized reaction conditions, the di-
rect hydroxylation of other aromatic substrates including
benzene, xylenes, and aryl halides became feasible.
[25] Z.Q. Li, L.G. Qiu, T. Xu, Y. Wu, W. Wang, Z.Y. Wu, X. Jiang, Mater. Lett. 63 (2009)
78–80.
[26] Y.K. Seo, G. Hundal, I.T. Jang, Y.K. Hwang, C.H. Jun, J.S. Chang, Micropor.
Mesopor. Mater. 119 (2009) 331–337.
[27] C. Shekhah, H. Wang, S. Kowarik, F. Schreiber, M. Paulus, M. Tolan, C.
Sternemann, F. Evers, D. Zacher, R.A. Fischer, C. Wöll, J. Am. Chem. Soc. 129
(2007) 15118–15119.
[28] J. Gascon, S. Aguado, F. Kapteijn, Micropor. Mesopor. Mater. 113 (2008) 132–
138.
[29] D. Zacher, O. Shekhah, C. Wöll, R.A. Fischer, Chem. Soc. Rev. 38 (2009) 1418–
1429.
[30] J.G. Vitillo, L. Regli, S. Chavan, G. Ricchiardi, G. Spoto, P.D.C. Dietzel, S. Bordiga,
A. Zecchina, J. Am. Chem. Soc. 130 (2008) 8386–8396.
[31] A. Martin-Calvo, E. Garcia-Pérez, J.M. Castillo, S. Calero, Phys. Chem. Chem.
Phys. 10 (2008) 7085–7091.
[32] T. Düren, Y.-S. Bae, R.Q. Snurr, Chem. Soc. Rev. 38 (2009) 1237–1247.
[33] R. Babarao, J. Jiang, S.I. Sandler, Langmuir 25 (2009) 5239–5247.
[34] K. Schlichte, T. Kratzke, S. Kaskel, Micropor. Mesopor. Mater. 73 (2004) 81–88.
[35] Y. Wu, L.-G. Qiu, W. Wang, Z.-Q. Li, T. Xu, Z.-Y. Wu, X. Jiang, Trans. Metal Chem.
34 (2009) 263–268.
[36] K. Brown, S. Zolezzi, P. Aguirre, D. Venegas-Yazigi, V. Paredes-Garcia, R. Baggio,
M.A. Novak, E. Spodine, Dalton Trans. (2009) 1422–1427.
[37] R. Raja, P. Ratnasamy, Appl. Catal. A: Gen. 143 (1996) 145–158.
[38] B. Lücke, K.V. Narayana, A. Martin, K. Jähnisch, Adv. Synth. Catal. 346 (2004)
1407–1424.
Acknowledgments
[39] J.J. Rehr, C.H. Booth, F. Bridges, S.I. Zabinsky, Phys. Rev. B 49 (1994) 12347–
12350.
[40] C. Prestipino, L. Regli, J.G. Vitillo, F. Bonino, A. Damin, C. Lamberti, A. Zecchina,
P.L. Solari, K.O. Kongshaug, S. Bordiga, Chem. Mater. 18 (2006) 1337–1346.
[41] J. Stötzel, D. Lützenkirchen-Hecht, R. Frahm, B. Kimmerle, A. Baiker, M.
Nachtegaal, M.J. Beier, J.D. Grunwaldt, 14th International Conference on X-Ray
Absorption Fine Structure (Xafs14), Proceedings, vol. 190, 2009.
The authors thank Hasylab at DESY for providing beamtime at
beamline X1 and Prof. Jan-Dierk Grunwaldt (KIT) for valuable dis-
cussions as well as the European Union for financial support (con-
tract number I-2009 0282 EC). Dr. Frank Krumeich and the EMEZ

S. Marx et al. / Journal of Catalysis 281 (2011) 76–87
87
[42] N. Drenchev, E. Ivanova, M. Mihaylov, K. Hadjiivanov, Phys. Chem. Chem. Phys.
12 (2010) 6423–6427.
[45] C.N. Kato, M. Hasegawa, T. Sato, A. Yoshizawa, T. Inoue, W. Mori, J. Catal. 230
(2005) 226–236.
[43] K.L. Gurunatha, T.K. Maji, Eur. J. Inorg. Chem. 12 (2009) 1592–1599.
[44] D. Zhao, N. Wu, S. Zhang, P. Xi, X. Su, J. Lan, J. You, Angew. Chem. Int. Ed. 48
(2009) 8729–8732.
[46] A.V. Ramaswamy, S. Sivasanker, P. Ratnasamy, Micropor. Mater. 2 (1994) 451–
458.