Microstructure characters of Cu/ZnO catalyst precipitated inside microchannel reactor

Article abstract of DOI:10.1016/j.molcata.2016.07.046

Cu/ZnO catalyst was prepared by co-precipitation method inside microchannel reactor and characterized by X-Ray diffraction (XRD), thermo gravimetric analysis (TGA), X-ray photoelectron spectroscopy (XPS), and high-resolution transmission electron microscopy (HRTEM). The XRD analysis of precursors demonstrates that, compared with the sample prepared by conventional batch reactor, more Zn²⁺ are incorporated into malachite structure, which is attributed to the relatively uniform distribution of Cu, Zn elements in initial precipitates caused by the excellent mixing performance of the microchannel reactor. Higher decomposition temperature of carbonate species trapped in the interfaces between CuO and ZnO and higher binding energy of Cu2p_3/2 indicate that sample prepared by the new reactor possess a stronger interface interaction, which derives from the more intimate contact between oxide components. This supposition is confirmed by the HRTEM images and the stronger interface interaction in the final reduced catalyst can improve catalytic performance on methanol synthesis.

Full text of DOI:10.1016/j.molcata.2016.07.046

Journal of Molecular Catalysis A: Chemical 423 (2016) 457–462
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Microstructure characters of Cu/ZnO catalyst precipitated inside
microchannel reactor
a,∗
a
a
b
Xin Jiang , Lu Zheng , Zhiyong Wang , Jiangang Lu
a
College of Chemical and Biological Engineering, Zhejiang University, Hangzhou 310027, PR China
State Key Laboratory of Industrial Control Technology, College of Control Science and Engineering, Zhejiang University, Hangzhou 310027, PR China
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 1 June 2016
Received in revised form 20 July 2016
Accepted 25 July 2016
Available online 26 July 2016
Cu/ZnO catalyst was prepared by co-precipitation method inside microchannel reactor and character-
ized by X-Ray diffraction (XRD), thermo gravimetric analysis (TGA), X-ray photoelectron spectroscopy
(
XPS), and high-resolution transmission electron microscopy (HRTEM). The XRD analysis of precursors
2+
demonstrates that, compared with the sample prepared by conventional batch reactor, more Zn are
incorporated into malachite structure, which is attributed to the relatively uniform distribution of Cu, Zn
elements in initial precipitates caused by the excellent mixing performance of the microchannel reac-
tor. Higher decomposition temperature of carbonate species trapped in the interfaces between CuO and
ZnO and higher binding energy of Cu2p3/2 indicate that sample prepared by the new reactor possess a
stronger interface interaction, which derives from the more intimate contact between oxide components.
This supposition is confirmed by the HRTEM images and the stronger interface interaction in the final
reduced catalyst can improve catalytic performance on methanol synthesis.
Keywords:
Cu/ZnO catalyst
Microchannel reactor
Interaction between components
Intimate contact
Methanol synthesis
©
2016 Elsevier B.V. All rights reserved.
As well known, the precipitation process of Cu2+, Zn2+ has a close
relationship with the preparation parameters and styles [7]. In the
normal concentration range of reagents, the precipitation reaction
is a very fast process which is more rapid than the mixing process
of reagents, resulting that the precipitation process occurs under
inhomogenous mixing conditions [8]. Combined with the fact that
Cu, Zn cations have inconsistent precipitation conditions and rates,
the distribution of Cu, Zn elements in initial precipitate is inevitably
inhomogeneous. It seems the redissolution and recrystallization
[9] processes in subsequent ageing step could smooth and even
eliminate the inhomogeneity, however, the pronounced “chemical
memory” phenomenon indicates the initial uneven character has
been taken into the microstructure of final catalyst.
1
. Introduction
The well-known Cu/ZnO/Al O3 catalyst has been widely
2
employed in methanol synthesis from syngas for more than 40
years [1]. The binary Cu/ZnO catalyst system with a Cu:Zn ratio of
7
:3 serves as a model for investigating the characters of the indus-
trially applied ternary catalysts [2]. The synthesis process is started
by co-precipitation reaction between aqueous solutions of metal
nitrates and sodium carbonate. During the following ageing pro-
cess, the initial products are transformed into crystalline precursor
2+
and correspondingly Cu in malachite are gradually substituted
2
+
by Zn , forming the most active precursor, e.g. zincian malachite
3]. Active catalysts are obtained after the subsequent drying, cal-
[
cination and reduction steps. The exposed Cu surface area and its
intrinsic activity [4–6] are regarded as crucial factors with respect to
the catalyst’s performance. According to the previous literature [5],
these two factors are deeply affected by the microstructure of final
Cu/ZnO catalyst which is predetermined by the various synthesis
parameters during the preparation process, and this phenomenon
is termed as “chemical memory” of the system.
The conventional preparation of Cu/ZnO catalyst is carried out
in the batch reactors, in which even under “constant” precipitating
conditions the chemical potential of reagents are neither spatially
nor temporally homogenous [10]. Thus, the inconsistency on mix-
ing and reaction conditions at different time and position inevitably
bring about the uneven distribution of Cu, Zn elements and effects
on the final microstructure of catalysts. Accordingly, to some extent
the Cu surface area and its intrinsic activity are influenced.
In the past years, microchannel reactors have aroused extensive
attention. They are widely used in preparing nanoparticles and can
be regarded as continuous and steady state flow devices in which
the reaction conditions do not change with the time. Furthermore,
∗ Corresponding author.
E-mail address: jiangx@zju.edu.cn (X. Jiang).
http://dx.doi.org/10.1016/j.molcata.2016.07.046
1
381-1169/© 2016 Elsevier B.V. All rights reserved.

4
58
X. Jiang et al. / Journal of Molecular Catalysis A: Chemical 423 (2016) 457–462
Fig. 1. (a) the scheme of preparation process by microchannel reactor (b) the photograph of inner flow channel of the microchannel reactor.
Table 1
their tiny space and turbulence-inducing inner structure and con-
figuration can significantly enhance and homogenize the mixing
degree [11] in the channel. In this article we discuss the influ-
ences of these new features on the microstructure of precursors
and catalysts.
the lattice plane reflection characters of two precursors.
1¯ 20 (2␪)/ C
◦
121 (2␪)/ C
¯
◦
d (201)/Å
¯
BR-Precursor
MR-Precursor
32.15
32.40
33.15
33.29
2.761
2.744
2.3. Catalytic reaction
2
. Experiment methods
Catalysis activity measurement was performed in an up-flow
2
.1. Catalyst preparation
tubular reactor (inner diameter 4 mm) made of stainless steel,
loaded typically with 0.1 g of powder catalyst diluted with 1.0 g
of quartz sand in 40–60 mesh. The reduced catalyst was main-
The Cu, Zn (Cu/Zn molar ratio = 7/3) nitrate solution (0.4 M) and
the aqueous Na CO solution (0.4 M) were preheated to reaction
2
3
◦
tained at 220 C for 2 h in N2 flowing at 40 ml/min. Then, syngas
◦
temperature (70 C), and then were simultaneously pumped into
the microchannel reactor which has internal bas-relief structures
to enhance the mixing degree of reagents (the scheme of prepara-
tion process and the image of inner flow channel of microchannel
reactor are shown in Fig. 1). The mixed materials flowed through a
(
H : CO = 2:1, with Ar as the internal standard, the specific pro-
2
portions respectively are 64.2%, 32.1% and 3.7%) with a total flow
−
1
rate at 60 ml/min (GHSV = 10276 h ) was switched in. The cat-
alytic activity was determined under the following conditions:
◦
◦
pressure 2 MPa, temperature range 220 C–270 C. After leaving
the reactor, the organic products were analyzed by on-line GC
◦
coiled tube (20 m length, 2 mm ID) submerged in 70 C water bath,
and the pH value at outlet was regulated to set value by control-
ling the flow rate. After a 2 h ageing process, the obtained products
were collected by filtration, and washed 6 times with deionized
water. Subsequently, the samples were dried at 110 C for 16 h fol-
lowed by calcination at 350 C in static air for 4 h. Finally the oxide
(
model 1102) equipped with a FID detector and the inorganic prod-
ucts were analyzed by on-line GC (model 9790) equipped with a
TCD detector. The CO conversion was calculated from the formula:
◦
X(CO) = (Mol(CO)in/min −Mol(CO)out/min)/Mol(CO) min × 100%,
◦
in
Mol(CO)in/min and Mol(CO)out/min were respectively denoted the
molar flow-rate of CO in inlet and outlet. Methanol production on
per unit copper surface area was used to evaluate the catalytic
performance.
◦
components were reduced by 20 vol% H2 in N2 at 300 C for 2 h.
For comparison, the conventional co-precipitation catalyst was
prepared under the same conditions expect that precipitation pro-
cess was conducted in a batch reactor with stir bar. Samples
prepared by microchannel reactor and conventional batch reactor
are respectively labeled as MR-X, BR-X, X represents for precursor,
oxide, or catalyst.
3. Results and discussion
3
.1. Microstructure characterization of precursors
The XRD patterns of the precursors prepared by two methods
2
.2. Characterization method
were shown in Fig. 2. All the marked peaks belong to the reflections
of zincian malachite (JCPDS-PDF 41-1390). As shown in the figure,
X-ray powder diffraction (XRD) was performed on a X’Pert Pro
¯
¯
the reflections of 201, 211 lattice planes of zincian malachite in the
MR-Precursor shift towards higher angles and the corresponding
MPD X-ray diffractometer using Cu-K␣ radiation (␭ = 0.15406 nm).
Thermogravimetric analyses were performed on a Netzsch STA449
thermobalance (10 K/min, synthetic air). XPS measurements were
performed using a Physical Electronics PHI-5400 ESCA work station
equipped with a magnesium anode (Mg K␣ = 1254 eV) at a power
of 240 W (12 kV and 20 mA). A Tecnai G2 F20 microscope operated
at 200 KV and equipped with a field emission gun was used for
¯
d-spacing of 201 reflection has a lower value (see Table 1). Pointed
in the literatures, the observed phenomenon is indicative of more
2+
9
Cu (d , Jahn-Teller distorted coordination preferred) substituted
2+
10
by Zn [13] (d , no Jahn-Teller distortion), which would exert a
larger contraction on the Jahn-Teller distortions of MO octahedral
6
¯
in the malachite structure and thus lower the d-spacing value of 201
HRTEM. Cu surface area was measured by applying N O reactive
frontal chromatography (RFC) [12].
2
reflection. Estimated by the established function [9] between the
2+
amount of incorporated Zn and d-spacing value, the Zn content in

X. Jiang et al. / Journal of Molecular Catalysis A: Chemical 423 (2016) 457–462
459
Fig. 2. the XRD patterns of precursors prepared by two methods.
the zincian malachite prepared by microchannel reactor is 28.4%,
while for the conventional co-precipitation method the value is
2
3.8%. As the Zn content incorporated into the malachite depends
heavily on the preparation parameters and styles [9,14], we deduce
that the gap comes from different courses of precipitation process
in the two reactors.
As a very fast process, the precipitation process of Cu2+ and Zn2+
proceeds before the homogeneity of the reagents has been estab-
lished [15]. Thus, concentration and pH gradients of reactants at
the micro-level cannot be avoided before the precipitation reac-
tion process ends. Due to the fact that the precipitation behaviors
of Cu and Zn are inconsistent [16] (e.g. the starting pH of Cu2+ is
2
+
2+
2
+
about 3 and Zn near pH 5) and seriously affected by pH, the exist-
ing gradient of pH causes an inconsistent precipitation and uneven
distribution of Cu, Zn elements in initial precipitation product. For
the microchannel reactor, the tiny space and turbulence-inducing
channel structures can relatively enhance and homogenize the
mixing degree inside the reactor. While in the conventional batch
reactor, there is a remarkable difference in the mixing performance
from point to point and weaker turbulent degree inevitably appears
in the regions far away from the stirring center. The lower tur-
bulence aggravates the inhomogeneous distribution of Cu and Zn
2
+
elements and then suppresses the incorporation process of Zn
in the ageing step. Therefore, the content of Zn2+ incorporating
into zincian malachite in the BR-Precursor is lower than that in
the MR-Precursor.
Fig. 4. the X-ray photoelectron spectra of the Oxide samples (a) the XPS spectra of
Cu2p, (b) the XPS spectra of Zn2p, (c) the XPS spectra of O1s.
Fig. 3 is the TGA profiles of precursors. Both samples show
the typical thermal behaviors of Cu, Zn mixed precursors. BR-
Precursor exhibits two weight loss steps at 327 and 427 C, while
◦
◦
MR-Precursor at 331 and 472 C. According to the literature [17],
the first weight loss step is owned to the simultaneous dehydra-
tion and decarbonation process. During this step, the interfaces and
grain boundaries between CuO and ZnO are formed. With respect to
the higher temperature weight loss step, it exclusively results from
the decomposition of carbonate species (HT-CO32 ) trapped in the
−
interfaces between CuO and ZnO. As claimed in literatures [10,17],
the higher temperature of HT-CO32 is an indicator of stronger
interaction across interfaces between CuO and ZnO. Accordingly,
a stronger interface contact exists in the MR-Oxide microstructure.
Estimated from Fig. 3, the fraction of carbonate species decom-
posed at higher temperature in the total weight loss is 12.3% for
BR-Precursor and 22.1% for MR-Precursor. Confirmed by many
−
experiments, the abundance of HT-CO32 has a positive correlation
−
with the amount of interfaces and grain boundaries [18] which is
2+
determined by the Zn content in the zinc malachite. Hence, more
amount of HT-CO32 in the MR-Precursor malachite structure is
the function of more incorporated Zn2+, which is correspond with
the description above.
−
Fig. 3. the TGA profiles of precursors prepared by two methods.

4
60
X. Jiang et al. / Journal of Molecular Catalysis A: Chemical 423 (2016) 457–462
Fig. 5. HRTEM images of two samples (a) BR-Oxide, (b) FFT pattern of area I, (c) FFT pattern of area II, (d) MR-Oxide, (e) FFT pattern of area III, (f) FFT pattern of area IV.
3.2. Microstructure characters of oxide samples
Fig. 4 shows XPS spectra of core level Cu2p, Zn2p, O1 s and
these data are fitted with Gaussian function. Proved by the peak
positions of Cu2p3/2 binding energy (Fig. 4(a), 934.4 eV, 934.1 eV)
as well as by the presence of intensive shake-up satellites in the
range of 940–945 eV [19,20], both calcined samples only con-
2
+
tained Cu at the surface. Compared with the binding energy of
Cu2p3/2 (933.6 eV) in the pure CuO, both samples have a signifi-
cant shift towards high values, which is caused by the interaction
between CuO and ZnO [21]. As can be seen from Fig. 4(b), the bind-
ing energy of Zn2p3/2 of MR-Oxide and BR-oxide are respectively
1
021.9 eV and 1021.8 eV. Compared with the binding energy of
Zn2p3/2 (1022.2 eV) in the pure ZnO, both samples have a shift
towards lower values which is also aroused by the interaction
between CuO and ZnO. It was reported that higher value of the
Fig. 6. XRD patterns of the two reduced catalysts.
Cu binding energy and lower value of Zn2+ binding energy corre-
sponds to stronger interaction between CuO and ZnO [21,22]. The
obtained results further confirm the so-called stronger interface in
the MR-Oxide microstructure.
2+
are hardly detected, which could be the impact of homogeneous
distribution of Cu, Zn elements. Thus, we speculate the more inti-
mate interface contact and better inter-dispersion between oxide
components may be the direct reason for enhanced interface inter-
action in the MR-Oxide microstructure.
As shown in Fig. 4(c), the O1s-spectra of both samples are con-
sistedof two peaks. Describedin the literature[20], the peak located
at 529.5 eV is assigned to the lattice oxygen in CuO and ZnO and
the other peak located at 531.6 eV is assigned to the residual car-
3.3. Reduced catalysts and catalytic performance
2−
bonate species in oxide samples (HT-CO3 ). Calculated from the
corresponding peak area, the amount of residual carbonate species
in MR-Oxide is 1.8 times than that in BR-Oxide, which is in good
agreement with characterization results (the value is 1.9) in Fig. 3.
Fig. 5 is the HRTEM images of two samples accompanied with
diffraction patterns of the selected areas via FFT method. Similar
to what has been presented by other researchers [5], the micro-
morphology of BR-Oxide shows a feature that CuO particles are
separated by some individual ZnO particles, forming a porous
framework of particles. Whereas, in the MR-Oxide, the diffraction
patterns of lattice planes of CuO and ZnO can be simultaneously
found in a small enough area, individual separated oxide particles
Fig. 6 shows the XRD patterns of the final reduced Cu/ZnO cat-
alysts. All diffraction peaks can be attributed to Cu (JCPDS-PDF
4-836) and ZnO (JCPDS-PDF 36-1451) phases. The sizes of Cu and
ZnO crystallites are calculated by Scherrer Formula according to
the corresponding reflection line (Cu (111), 2␪ = 43.297 , ZnO (101),
2␪ = 36.252 ) and the values don’t show obvious changes. For the
◦
◦
convenience of the following calculating process, the sizes of Cu
crystallites are listed in Table 2. The corresponding Cu surface area
(SACu) is measured by applying N O reactive frontal chromatogra-
2
phy. The interface ratios[17] (IFR) (i.e. the fraction of Cu particles
not exposed to the gas atmosphere but present as interface to ZnO
phase) equal to the ratio of measured SACu to the theoretically avail-

X. Jiang et al. / Journal of Molecular Catalysis A: Chemical 423 (2016) 457–462
461
Table 2
the performance parameters of two reduced catalysts.
0
d(Cu ) nm SACu
IFR % Lattice strain Catalytic activities
2
−1
−2
−1
m
g
mmol m Cu h
BR-Reduced 13.4
MR-Reduced 13.2
6.5
4.0
80.1 0.181%
88.1 0.241%
126
260
Fig. 8. Williamson-Hall plots of two reduced catalysts.
plement evidence for the described higher intrinsic activity in the
MR-Catalyst.
4. Conclusion
Fig. 7. the Catalytic activities of two reduced catalysts at different reaction temper-
ature.
Compared with the conventional batch reactor, the characters
of flow and mixing in the microchannel reactor can bring a more
homogenous Cu, Zn distribution in the initial precipitation prod-
ucts. As a result, more Zn2+ ions are incorporated into the malachite
structure confirmed by higher reflection angels and lower d(20 1¯ )
able Cu surface calculated from the XRD determined average Cu
crystallites sizes. As can be seen from the calculated results listed in
the Table 2, the MR-Catalyst sample has a lower Cu surface area but
a higher IFR which manifests that the Cu particles in MR-catalyst
have a stronger interface interaction with the ZnO particles [5]. This
is in agreement with the stronger interface interaction in the MR-
Oxide microstructure, demonstrating that the enhanced interface
interaction survives after the reduction process.
spacing of zincian malachite. In the following calcination step, a
more intimate contact between the oxide components improves
the interface interaction, which is supported by DTG, XPS and
HRTEM characterization methods. The stronger interface interac-
tion survives in the reduced catalyst and then enhances the intrinsic
activity for methanol synthesis.
In order to compare their catalytic performance both reduced
catalysts were tested in the methanol synthesis reaction, the results
are shown in Fig. 7. The catalytic activities of both samples initially
increase with the reaction temperature and the maximum catalytic
activities of per unit copper surface area (the values are given in
Acknowledgements
Financial support from National Science Foundation of China
Grants (Contract 21276223), the Chinese Key Technology R&D Pro-
gram (2010AA064905) and the Program for Zhejiang Leading Team
of Science & Technology Innovation (Contract 2009R50020) are
gratefully acknowledged.
◦
Table 2) are obtained at 270 C. As the equilibrium conversion for
◦
methanol synthesis at 280 C (2 MPa, H :CO = 2:1) is only 7.4% (the
2
conversions of two samples are respectively 6.3% and 5.1% which
are not very far from equilibrium value), the catalytic activities
measured at this temperature show a downward trend.
References
Obviously, the activity of MR-catalyst is about twice as large as
the value of BR-catalyst. Considering of the characterizations men-
tioned above, it is reasonable to assume that the intrinsic activity
of exposed Cu surface in MR-catalyst has been improved by the
presented stronger Cu-ZnO interface interaction which is useful
for the stabilization of the strained and defective form of Cu par-
ticles [23]. Thus, high value of Cu lattice strain and abundance of
Cu defects identified as symbols of high intrinsic activity [23,24]
can be expected. It is worth mentioning that the CO conversion
[
1] J.-L. Li, T. Inui, Appl. Catal. A 137 (1996) 105–117.
[2] Malte Behrens, Stefan Kißner, Frank Girsgdies, Igor Kasatkin, Felix
Hermerschmidt, Katharina Mette, Holger Ruland, Martin Muhler, Robert
Schlögl, Chem. Commun. 47 (2011) 1701–1703.
[
3] M. Behrens, J. Catal. 267 (2009) 24–27.
[4] P. Munnik, M. Wolters, A. Gabrielsson, S.D. Pollington, G. Headdock, J.H. Bitter,
P.E. de Jongh, K.P. de Jong, J. Phys. Chem. C 115 (2011) 14698–14706.
[
5] M. Behrens, A. Furche, I. Kasatkin, A. Trunschke, W. Busser, M. Muhler, B.
Kniep, R. Fischer, R. Schlögl, ChemCatChem 2 (2010) 816–818.
6] Melanie Kurtz, Natalia Bauer, Christoph Büscher, Hagen Wilmer, Olaf
Hinrichsen, Ralf Becker, Stefan Rabe, Klaus Merz, Matthias Driess, Roland A.
Fischer, Martin Muhler, Catal. Lett. 92 (2004) 49–52.
[
of MR-catalyst is not very far from the equilibrium conversion
[
7] M. Behrens, R. Schlögl, Z. Anorg, Allg. Chem. 639 (2013) 2683–2695.
◦
(
10.3%) under our reaction conditions (2 MPa, 270 C, H : CO = 2:1),
2
[8] Martain A.J. Harting, Nikolas Jacobsen, Wolfgang Peukert, Chem. Eng. Sci. 109
(2014) 158–170.
[
10] M. Schur, B. Bems, A. Dassenoy, I. Kassatkine, J. Urban, H. Wilmes, O.
Hinrichsen, M. Muhler, R. Schlögl, Angew. Chem. Int. Ed. 42 (2003)
3815–3817.
demonstrating that the intrinsic activity of MR-catalyst has been
substantially enhanced.
9] S. Zander, B. Seidlhofer, M. Behrens, Dalton Trans. 41 (2012) 13413–13422.
[
Williamson-Hall method was used to calculate the values of Cu
lattice strain (Fig. 8) in two samples [25,26]. The method assumes
that if the XRD patterns can be satisfactorily fitted by Lorentz
function, the total physical broadening of XRD pattern is a lin-
ear addition of ‘size’ and ‘strain’ effects. For isotropic effects, the
plot of ␤ cos(␪)/␭ versus 2sin (␪)/␭ should results in a straight line.
The mean value of lattice strain (See Table 2) equals the half of
the slope in Fig. 8. As expected, the value of Cu lattice strain has
improved a lot in the MR-Catalyst, which can be regarded as a sup-
[
[
11] X. Yao, Y. Zhang, L. Du, J. Liu, J. Yao, Renew. Sustain. Energy Rev. 47 (2015)
519–539.
12] G.C. Chinchen, C.M. Hay, H.D. Vandervell, K.C. Waugh, J. Catal. 103 (1987)
79–87.
13] M. Behrens, F. Girgsdies, Z. Anorg, Allg. Chem. 636 (2010) 919–927.
14] B. Bems, M. Schur, A. Dassenoy, H. Junkes, D. Herein, R. Schlögl, Chem. Eur. J. 9
[
[
(
2003) 2039–2052.
[15] A. Nagaki, M. Togai, S. Suga, N. Aoki, K. Mae, J. Yoshida, J. Am. Chem. Soc. 127
(2005) 11666–11675.

4
62
X. Jiang et al. / Journal of Molecular Catalysis A: Chemical 423 (2016) 457–462
[
16] M. Behrens, D. Brennecke, F. Girgsdies, S. Kißner, A. Trunschke, N. Nasrudin, S.
Zakaria, N.F. Idris, S.B.A. Hamid, B. Kniep, R. Fischer, W. Busser, M. Muhler, R.
Schlögl, Appl. Catal. A 392 (2011) 93–102.
[21] D. Ji, W. Zhu, Z. Wang, G. Wang, Catal. Commun. 8 (2007) 1891–1895.
[22] G. Fabio, G. Petrini, J. Catal. 90 (1984) 106–112.
[23] S. Zander, E.L. Kunkes, M.E. Schuster, J. Schumann, G. Weinberg, D. Teschner,
N. Jacobsen, R. Schlögl, M. Behrens, Angew. Chem. Int. Ed. 52 (2013)
6536–6540.
[24] I. Kasatkin, P. Kurr, B. Kniep, A. Trunschke, R. Schlögl, Angew. Chem. Int. Ed. 46
(2007) 7324–7327.
[25] T. Kandemir, I. Kasatkin, F. Girgsdies, S. Zander, S. Kühl, M. Tovar, R. Schlögl,
M. Behrens, Top. Catal. 57 (2014) 188–206.
[
[
[
[
17] S. Kühl, M. Friedrich, M. Armbrüster, M. Behrens, J. Mater. Chem. 22 (2012)
9
632–9638.
18] M. Behrens, F. Girgsdies, A. Trunschke, R. Schlögl, Eur. J. Inorg. Chem. 7 (2009)
10–921.
19] Jong Wook Bae, Suk-Hwan Kang, Yun-Jo Lee, Ki-Won Jun, App. Catal. B 90
2009) 426–435.
9
(
20] D.A. Svintsiskiy, A.P. Chupakhin, E.M. Slavinskaya, O.A. Stonkus, A.I.
Stadnichenko, J. Mol. Catal. A: chem. 368–369 (2013) 95–106.
[26] B.L. Kniep, T. Ressler, A. Rabis, F. Girgsdies, M. Baenitz, F. Steglich, R. Schlögl,
Angew. Chem. Int. Ed. 43 (2004) 112–115.