Flame-made Cu/ZnO/Al2O3 catalyst for dimethyl ether ...

Catalysis Communications 43 (2014) 52–56

Contents lists available at ScienceDirect

Catalysis Communications journal homepage: www.elsevier.com/locate/catcom

Short Communication

Flame-made Cu/ZnO/Al2O3 catalyst for dimethyl ether production Ruaa Ahmad a, Melanie Hellinger a,b, Maria Buchholz c, Hikmet Sezen c, Loubna Gharnati a,b, Christof Wöll c, Jörg Sauer a, Manfred Döring d, Jan-Dierk Grunwaldt a,b,⁎, Ulrich Arnold a,⁎⁎ a

Institute of Catalysis Research and Technology, Karlsruhe Institute of Technology, Hermann-von-Helmholtz-Platz 1, D-76344 Eggenstein-Leopoldshafen, Germany Institute for Chemical Technology and Polymer Chemistry, Karlsruhe Institute of Technology, Engesserstraße 20, D-76128 Karlsruhe, Germany Institute of Functional Interfaces, Karlsruhe Institute of Technology, Hermann-von-Helmholtz-Platz 1, D-76344 Eggenstein-Leopoldshafen, Germany d Fraunhofer Institute for Structural Durability and System Reliability LBF, Schlossgartenstraße 6, D-64289 Darmstadt, Germany b c

a r t i c l e

i n f o

Article history: Received 22 July 2013 Received in revised form 22 August 2013 Accepted 23 August 2013 Available online 30 August 2013 Keywords: Dimethyl ether Flame-spray pyrolysis Methanol Synthesis gas

a b s t r a c t For the single step synthesis of dimethyl ether (DME) from synthesis gas a Cu/ZnO/Al2O3-catalyst has been prepared using flame-spray pyrolysis. The resulting powder was co-mixed with γ-alumina to give an admixed system for DME production. The flame-made catalyst was analyzed using the BET method, in situ XRD, N2O decomposition, TPR and XPS. These studies unraveled that the catalyst exhibited a high Cu surface area including good contact with zinc oxide and alumina as well as small Cu particles resulting in high catalytic activity and product selectivity, also in comparison to a commercially available catalyst. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Dimethyl ether (DME) is one of the most promising substitutes for LPG- and Diesel-Fuels [1]. DME exhibits a high cetane number and its combustion is soot free with significantly reduced NOx- and SOx-emissions. Additionally, the ozone depletion potential of DME is zero and its global warming potential is significantly lower compared to other greenhouse gases [2]. Formerly, applications of DME were limited mainly to its use as propellant and for the preparation of dimethyl sulfate [3] but the markets have shifted and today DME is predominantly used as LPG-substitute [4]. The variety of applications caused an increasing interest in DME, which is reflected in the growth of the global DME market. Global DME capacity rose from 30 000 t/a in 2003 to 5 Mio t/a in 2009. As an example, Chinese capacities for DME production are estimated to reach 13 Mio t/a in 2018 [4]. DME is industrially produced by dehydrating methanol over an acidic catalyst. Another strategy is the direct synthesis of DME from synthesis gas combining CO hydrogenation to methanol and the following dehydration to DME in a single step. So far, the direct synthesis is not applied on industrial scale but has been tested in several pilot plants

⁎ Correspondence to: J.-D. Grunwaldt, Institute of Catalysis Research and Technology, Karlsruhe Institute of Technology, Hermann-von-Helmholtz-Platz 1, D-76344 EggensteinLeopoldshafen, Germany. Tel.: +49 721 608 42120. ⁎⁎ Corresponding author. Tel.: +49 721 608 23694. E-mail addresses: [email protected] (J.-D. Grunwaldt), [email protected] (U. Arnold). 1566-7367/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.catcom.2013.08.020

located in various countries [5]. Thus, there are no commercially available catalysts for the single step synthesis and specially designed catalysts for this reaction are still on a research level. Typically, admixed catalyst systems are employed. These contain a Cu/ZnO/Al2O3-catalyst (CZA) for methanol production and an acidic catalyst, e.g. γ-Al2O3 or zeolite, for methanol dehydration to DME [6]. Usually, catalysts for low-pressure methanol synthesis are synthesized by co-precipitation and the preparation procedure has been investigated and improved over decades [7]. The correlation between catalyst features and activity is not fully understood yet, but recent studies revealed that the crucial features are a high Cu0 surface area, strong metal support interaction of the active Cu species with ZnO which lead to dynamic changes of the copper particles [8] and the presence of Cu steps, that are stabilized by bulk defects [9]. Within this work an admixed catalyst system for the single step synthesis of DME was developed. The CZA-component has not been prepared by co-precipitation but via flame-spray pyrolysis (FSP). This method has the advantage that it leads to a strong interaction between the components being sprayed and that several components can be mixed in one step despite of the fact that the precursor materials may be more expensive [10,11]. Moreover, homogeneous mixing with microcrystalline structure can potentially be achieved and effects like ageing and pH-control during precipitation can be avoided. The method has not been described for the preparation of methanol or DME catalysts yet. However, FSP-synthesis of CZA-catalysts for the water-gas-shift (WGS) reaction was reported [12,13]. Catalysts for the WGS-reaction are similar to methanol catalysts, but their composition regarding Cu,

R. Ahmad et al. / Catalysis Communications 43 (2014) 52–56

ZnO and Al2O3 is quite different and hence, properties differ as well. Jensen et al. [14] prepared a CZA-catalyst for methanol synthesis using flame combustion synthesis, also known as vapour-fed aerosol flame synthesis (VAFS). VAFS uses volatile metal salts as precursors, which are ignited and combusted to give the metal oxides. Compared to VAFS, FSP-synthesis offers several advantages. Flame spray pyrolysis is not restricted to volatile precursors and higher yields can be reached [15]. Here we describe a simple and economic one-step FSP procedure for the preparation of a highly active catalyst for methanol production and its use in catalyst systems for the single step production of DME from CO-rich synthesis gas. A H2:CO ratio of 1:1 was employed, which is typical for the single step synthesis and which is typically received from gasification of agricultural residues and wastes [6]. 2. Experimental 2.1. Catalyst synthesis Copper (II) acetylacetonate, zinc acetate and aluminum triisopropoxide (molar ratio of 62:30:8) were dissolved in methanol, toluene and propionic acid (volumetric ratio of 1:2:1) to give a 0.125 M precursor solution, which was sprayed into a methane supported flame with a flow rate of 5 ml/min using a syringe pump with a needle injector similar to the studies reported before [16]. A mixture of 0.75 l/min methane and 1.6 l/min O2 was used as fuel for the supporting flame. O2 for dispersion was fed into the flame with a flow rate of 5 l/min. The obtained metal oxides were cooled down and collected on a glass-fiber filter with the help of a vacuum pump at the end of a container, distanced 40 cm away of the flame. A fine black powder was peeled off the filter to give CuO/ZnO/Al2O3 (CZA-FSP) with a composition of 65:28:7 wt%. For the single step synthesis of DME an admixed catalyst system (CZA-FSP-AA) was used comprising 1 g of CZA-FSP and 1 g of a commercially available γ-Al2O3 (Alfa Aesar). For comparison, a commercially available methanol catalyst (MComm) with a similar composition (62:28:10 wt%) was used. This catalyst was also mixed with γ-Al2O3 to give an admixed catalyst for DME production (MComm-AA). 2.2. Catalyst characterization Metal loadings were determined by scanning electron microscopy coupled with energy-dispersive X-ray spectroscopy (SEM-EDX). Analyses were carried out on several locations of the catalysts employing a FESEM DSM 982 Gemini microscope (ZEISS) and an INCAPentaFET-x3 unit from Oxford Instruments. Transmission electron microscopy (TEM), scanning electron microscopy with high angle annular detector (STEM-HAADF), high resolution transmission electron microscopy (HRTEM) and elemental analysis (EDX) were performed with an FEI Tecnai F20 ST TEM (operating voltage 200 kV) equipped with a field emission gun and an EDAX EDS X-ray spectrometer (Si(Li) detecting unit, super ultra thin window, active area 30 mm2, resolution 135 eV (at 5.9 keV)). For TEM analysis, the catalysts were initially suspended in isopropanol. A small droplet of the catalyst suspension was deposited onto holey carbon-filmed copper grids (for HRTEM imaging) and carbon-filmed nickel grids (for EDX analysis) and eventually air dried. The specific surface area (BET; p/p0 = 0.05–0.20) and total pore volume (p/p0 = 0.99) were determined by N2-physisorption using a Quantachrome Nova 2000e instrument. H2-TPR and N2O-pulse chemisorption were carried out in an AutoChem 2950 HP device from Micromeritics. The sample was dried at 200 °C while under Ar, cooled down and then exposed to a diluted H2 atmosphere (20% H2 in Ar) and heated up to 300 °C with 2 °C/min. The H2 uptake was recorded employing a thermal conductivity detector (TCD). For determination of the specific Cu surface area, N2O-pulse chemisorption was applied. The sample was reduced by H2-TPR as described above. At 60 °C pulses of N2O were given into the sample cell. Redundant N2O was collected in

53

a cooling trap and the evolved N2 was detected and recorded by the TCD-detector. The Cu surface area was calculated by assuming a surface density of 1.46 · 1019 Cu atoms/m2 [17]. An in situ study combining H2-TPR and XRD was also used to analyze CZA-FSP. The sample was subjected to a 5% H2/N2-atmosphere and heated up with 5 °C/min to 100 °C. Then a diffraction pattern was recorded while keeping the temperature constant. This procedure was repeated for each temperature (150, 200, 250, 300 °C). The diffractometer (Bruker D8 Advance) was equipped with a rotating sample holder (XRK 900 from Anton Paar), a copper anode X-ray source, a nickel filter and a graphite monochromator for Cu Kα (λ = 1.54 Å) radiation. The accelerating voltage and anode current were 40 kV and 35 mA, respectively. The diffraction patterns were measured in the scattering angle range from 20 to 80° (2 s per step, step: 0.016°). The average crystallite size of the reduced Cu-species was determined by the Scherrer-equation [18] with a correction for the instrumental broadening. Complementary X-ray absorption spectra were recorded at the Cu K-edge at the SuperXAS-beamline at SLS using a capillary as in situ cell and 5% H2/He as reduction gas. X-ray photoelectron spectroscopy (XPS) measurements have been carried out using an IR/XPS–UHV system (Prevac) equipped with a VG Scienta R4000 electron energy analyzer. Al Kα (nonmonochromatic, 1486.68 eV) radiation has been used as excitation source. The energy scale of the instrument has been calibrated to the reference peaks Au4f7/2 (84.00 eV) and Ag3d5/2 (368.26 eV). To eliminate the surface charging the binding energies were shifted to the C1s line at 285.00 eV. The surface composition was determined from the computed peak areas of the Cu2p3/2, the Zn2p3/2, and Al2p peaks after Shirleytype background subtraction and using the sensitivity factors for XPS (CasaXPS database, v2.3.16). 2.3. Catalyst testing The catalyst screening was carried out in a continuously operating laboratory plant [6]. The laboratory plant contained a fixed-bed reactor in which 2 g of the admixed catalyst system (CZA-FSP-AA or MCommAA) was placed. Reduction of CuO to the active species Cu0 took place in a reducing atmosphere (2% H2 in Ar) while heating up to 240 °C with a heating rate of 17 °C/h. Afterwards, the gas flow was switched to pure H2 and the temperature was increased to 250 °C within 1 h. After keeping the system for 2 h under these conditions it was purged with Ar and subsequently pressurized to 51 bar. A gas mixture (H2 + CO):(Ar + N2) with a ratio of 3:7 was employed and a WHSV of 4.19 h−1 was adjusted. The H2:CO-ratio was 1:1 to simulate biomass-derived synthesis gas and the Ar:N2-ratio was 6:1. All gas streams were monitored by online-GC (HP 6890, Agilent) using two columns (RT®-U-BOND and RT-Msieve 5A, RESTEK) and two detectors (TCD and FID). In brief, conversion is based on the consumption of CO and selectivity is based on the fraction of converted carbon in each product species. Details are described in [6]. 3. Results and discussion 3.1. Catalyst synthesis and characterization Catalyst CZA-FSP was prepared by feeding the respective metal salt precursor solution into a FSP-reactor. The obtained metal oxides were collected on a glass fiber filter which was located on top of the reactor container. During preparation a constant feed as well as a steady flame was assured. Hence, a fine black powder was obtained and structural features were investigated by recording TEM-images (Fig. 1). The sample is a fine, nanoparticulate powder consisting of small agglomerated particles (mean particle size estimated from TEM images: b10 nm; calculated mean particle size: 6.4+/−2.5 nm). Since the different CuO, ZnO and Al2O3 species cannot be clearly distinguished, an exact determination of the single particle sizes is not possible. The lattice spacings, as indicated by arrows in Fig. 1b, correspond to the

54


a

b

Fig. 1. TEM-images of CZA-FSP in as-prepared state.

measurements. However, N2O-pulse frontal chromatography may overestimate the Cu surface area if also underlying layers are oxidized. The oxidation states of the metal oxides on the catalyst surface were determined by XPS analysis and spectra including the binding energies are summarized in Fig. 4. The Cu2p3/2 peak at 934.20 eV and the respective shake-up satellite peaks confirm directly the presence of Cu2+ on the surface. The peak position of the Zn2p3/2 signal, which indicates the presence of Zn2+, can be seen at 1021.73 eV. The peak around 75 eV (Fig. 4c) can be attributed to Al2p. However, there is a contribution from Cu3p. Hence, the signal derives from the Cu3p peak at

CuO Cu 300 °C

250 °C Intensity (a.u.)

(110), (002), (−111) and (111) planes of the CuO tenorite phase and the (002) plane of the ZnO phase. STEM-HAADF images and EDXanalyses showed a homogeneous distribution of CuO, ZnO and Al2O3 within the catalyst. The total pore volume and the surface area were determined by N2-physisorption and calculated using the BET-method. CZA-FSP exhibits a total pore volume of 0.62 ml/g and a surface area of 68 m2/g. The composition of CZA-FSP was determined by XRD and the diffraction pattern shows the typical diffractions for CuO (2 θ = 35.38, 38.54, 48.44, 53.06, 58.12, 61.39, 66.01, 67.81, 74.96°), ZnO (2 θ = 30.80, 47.46, 56.24, 62.62°) and Al2O3 (53.06, 59.78°), which are overlapping. Hence, the presence of spinels and different oxidation states cannot be deduced from the XRD measurement but this does not prove their absence. The diffraction pattern displays relatively broad diffractions that indicate small particle sizes. Temperature dependent XRD measurements under reducing atmosphere were also carried out (Fig. 2). At temperatures above 150 °C, reduction of CuO to active Cu0 species takes place. Since the CuO-diffractions disappear almost completely, it can be assumed that the majority of CuO is reduced. The Cu0-peak at 2 θ ≈ 43° (250 °C) was used for calculating the particle size by the Scherrer-equation and a particle size of 10 nm was estimated. At 250 °C a shift in the diffraction patterns can be observed that disappears at 300 °C. This might be due to a change in lattice parameters caused by residual stress or elimination of defects. The reducibility of the catalyst was studied by H2-TPR (Fig. 3). CZA-FSP is mainly reduced at 210 °C but the reduction curve is not uniform and shows a shoulder at 176 °C. There are different interpretations by various authors regarding such reduction profiles with more than one peak. According to previous studies, the main peak at 210 °C can be attributed to the reduction of bulk CuO and the signal at 176 °C to the reduction of well-dispersed CuO that cannot be detected by XRD [19,20]. In contrast, it was reported that the two reduction steps are due to the consecutive reduction of Cu2+ first to Cu+ and then to Cu0 [21] or to the reduction of copper-aluminates [22]. XAS experiments were carried out and proved the reduction of two different CuO phases to metallic copper particles, since a first and a second temperature regime for the reduction to metallic copper was observed in 5% H2/He and no evidence for the presence of Cu+ was found (e.g. pre-edge feature corresponded from the start to Cu0 and not Cu+). The specific Cu surface area of CZA-FSP was determined by N2O-pulse frontal chromatography. From these data a high metal dispersion around 10% and a high Cu surface area around 34 m2/gSample were estimated. The particle size of the active Cu-species is around 8.5 nm and thus it is slightly lower than the particle size that was obtained via XRD-

200 °C

150 °C

100 °C

20

30

40

50

60

70

2 θ (°) Fig. 2. Diffraction pattern of CZA-FSP under reducing atmosphere (conditions: 5% H2 in N2, heating rate: 5 °C/min).


55

Intensity (a.u.)

surface area is determined compared to XPS analysis. Employing XPS, only outermost surface layers are determined.

3.2. Catalyst testing

100

150

200

250

300

Temperature (°C) Fig. 3. H2-TPR diagram of CZA-FSP (conditions: 20% H2 in Ar, heating rate: 2 °C/min).

76.85 eV and the Al2p peak at 74.75 eV. The surface stoichiometric ratio of Cu:Zn:Al is approximately 1:1.4:5.9 for the pristine sample, which differs strikingly from the overall composition (Cu:Zn:Al = 3.9:1.6:1). Hence, zinc oxide and especially alumina segregated on the sample surface of CZA-FSP. This phenomenon has also been described in the literature before [23–26]. Comparing XPS results with results from N2O-pulse frontal chromatography, the different surface compositions may be explained by the above-mentioned oxidation of subsurface layers during N2O-pulse frontal chromatography. Thus, a larger Cu

a

Cu2p3/2 934.2 eV

Satellite

Cu2p1/2

960

970

For the single step synthesis of DME, catalyst CZA-FSP and the commercially available methanol catalyst MComm, as a reference catalyst, were mixed with γ-Al2O3 in a weight-ratio of 1:1, respectively. The resulting admixed catalyst systems CZA-FSP-AA and MComm-AA were investigated in a laboratory plant at 250 °C, 51 bar and a WHSV of 4.19 h−1. Fig. 5 summarizes CO-conversion and DME-selectivity for both systems over a period of 72 h. Both systems do not show any deactivation and CO-conversion remains constant. Hence, the average CO-conversion of CZA-FSP-AA and MComm-AA was calculated by linear fitting and 56 and 54% were obtained, respectively. Regarding product selectivity, both catalysts accomplish almost complete dehydration of methanol to DME with ca. 67% selectivity. As expected, CO2 is formed with a selectivity of around 32% and the percentage of byproducts in the product mixture (methanol and hydrocarbons) does not exceed a total of 2%. The results of catalyst characterization and testing show that methanol/DME-catalysts obtained via FSP are highly active and stable systems that can compete with commercially available methanol catalysts. CZA-FSP exhibits a homogeneous distribution of CuO/ZnO, high

b

Zn2p3/2 1021.73 eV

Satellite

950

940

930

1025

1030

1020

1015

Binding Energy (eV)

Binding Energy (eV)

c

d

O1s 530.89 eV

Zn3p1/2, 3/2 89.07

91.87

Cu3p

Al2p 76.81

78.60

90

80

Binding Energy (eV)

70

537

534

531

528

525

522

Binding Energy (eV)

Fig. 4. a) XP spectrum of Cu2p lines, b) XP spectrum of Zn2p signal, c) Zn3p, Al2p and Cu3p lines, d) XP spectrum of O1s signal.


CO-conversion and DME-selectivity (%)

56

Acknowledgments

90 CZA-FSP-AA, CO-conversion CZA-FSP-AA, DME-selectivity MComm-AA, CO-conversion MComm-AA, DME-selectivity

80 70 60

R.A., M.H. and M.B. gratefully acknowledge financial support from the Helmholtz Research School “Energy-Related Catalysis”. J.-D.G. thanks the BMBF for financial support within the project “Materials in Action” and SLS for providing beamtime for XAS measurements. We thank Marina Tepluchin for in situ XRD measurements and Dr. Silke Behrens for TEM-images.

50

References 40 30 0

10

20

30

40

50

60

70

Time on stream (h) Fig. 5. CO-conversion and DME-selectivity for CZA-FSP-AA and MComm-AA.

metal dispersion and small Cu0 crystallites, properties that lead to high catalyst- and Cu-surface areas. It is well-known that the activity of methanol catalysts strongly depends on the Cu0 surface area, good interaction between copper and zinc and Cu0 crystallite size [27,28]. Catalysts with high Cu0 surface areas and small Cu0 crystallites are among the most efficient for methanol production from synthesis gas. Catalysts with an approximate CuO:ZnO:Al2O3 ratio of 60:30:10 wt% typically exhibit Cu0 surface areas around 35 m2/g and the crystallite size of Cu0 is around 10 nm [29–32]. CZA-FSP shows similar characteristics and thus, similar catalytic performance including high stability. In contrast to Cu/ZnO-based catalysts prepared by wet-chemical methods [33,34] preparation is carried out in one single step using economically affordable and common precursor salts. Hence, FSP offers the opportunity to reduce catalyst synthesis steps and combine them in a single and straightforward procedure that leads to efficient, microcrystalline and homogeneously distributed catalysts despite of the discussed “memory effect” [35]. 4. Conclusion and future work Flame-made Cu/ZnO/Al2O3-compounds are effective catalysts for methanol and DME synthesis. Compared to the vapour-fed aerosol flame synthesis method, higher yields and easier upscaling can be achieved. Their activity derives from the small particle size of the active Cu species, their close interaction with ZnO and the high surface area. These features add up to give catalysts with high metal dispersion, long-term stability and high activity as well as product selectivity. The current experiments indicate that further improvements in catalyst preparation and an extended catalyst testing offer a large optimization potential. Furthermore, an elaborate catalyst shaping could lead to catalysts with high long-term-stability and resistance to sintering and coking. The possibility to prepare Cu/ZnO/Al2O3-catalysts within one step opens a promising pathway, considering that preparation is simple, cost-efficient, up-scalable and reproducible. Future work will include the preparation of a flame-made bifunctional DME-catalyst that combines a Cu/ZnO/Al2O3-component and γ-Al2O3 in one single compound using the double-flame-spray pyrolysis.

[1] M. Marchionna, R. Patrini, D. Sanfilippo, G. Migliavacca, Fuel Process. Technol. 89 (2008) 1255–1261. [2] T.A. Semelsberger, R.L. Borup, H.L. Greene, J.Power Sources 156 (2006) 497–511. [3] M. Müller, U. Hübsch, Ullmann's Encyclopedia of Industrial Chemistry, Online Release, 11, Wiley-VCH, Weinheim, 2012, pp. 305–308. [4] J.A. Taupy, DME Industry and Association Overview, 4th International DME Conference, Stockholm, Sweden, 2010. [5] DME Handbook, in: Japan DME Forum, JDF (Ed.), Japan DME Forum, Tokyo, 2011. [6] M. Stiefel, R. Ahmad, U. Arnold, M. Döring, Fuel Process. Technol. 92 (2011) 1466–1474. [7] M.S. Spencer, Top. Catal. 8 (1999) 259–266. [8] J.-D. Grunwaldt, A.M. Molenbroek, N.-Y. Topsøe, H. Topsøe, B.S. Clausen, J. Catal. 194 (2000) 452–460. [9] M. Behrens, F. Studt, I. Kasatkin, S. Kühl, M. Hävecker, F. Abild-Pedersen, S. Zander, F. Girgsdies, P. Kurr, B.-L. Kniep, M. Tovar, R.W. Fischer, J.K. Nørskov, R. Schlögl, Science 336 (2012) 893–897. [10] W.Y. Teoh, R. Amal, L. Mädler, Nanoscale 2 (2010) 1324–1347. [11] S. Hannemann, J.-D. Grunwaldt, D. Günther, F. Krumeich, P. Lienemann, A. Baiker, Appl. Catal. A Gen. 316 (2007) 226–239. [12] (a) H. Meland, T. Johannessen, B. Arstad, H.J. Venvik, M. Rønning, A. Holmen, Stud. Surf. Sci. Catal. 162 (2006) 985–992; (b) F. Huber, H. Meland, M. Rønning, H. Venvik, A. Holmen, Top. Catal. 45 (2007) 101–104. [13] R. Kam, C. Selomulya, R. Amal, J. Scott, J. Catal. 273 (2010) 73–81. [14] J.R. Jensen, T. Johannessen, S. Wedel, H. Livbjerg, J. Catal. 218 (2003) 67–77. [15] B. Schimmoeller, S.E. Pratsinis, A. Baiker, ChemCatChem 3 (2011) 1234–1256. [16] M.J. Beier, B. Schimmoeller, T.W. Hansen, J.E.T. Andersen, S.E. Pratsinis, J.-D. Grunwaldt, J. Mol. Catal. A Chem. 331 (2010) 40–49. [17] J.W. Evans, M.S. Wainwright, A.J. Bridgewater, D.J. Young, Appl. Catal. 7 (1983) 75–83. [18] P. Scherrer, Göttinger Nachrichten Gesell. 2 (1918) 98–100. [19] E.R.A. Matulewicz, M.S. de Keijser, J.C. Mol, F. Kapteij, Thermochim. Acta 72 (1984) 111–116. [20] M. Shimokawabe, H. Asakawa, N. Takezawa, Appl. Catal. 59 (1990) 45–58. [21] B. Lindström, L.J. Pettersson, P.G. Menon, Appl. Catal. A Gen. 234 (2002) 111–125. [22] J.-D. Grunwaldt, B. Kimmerle, S. Hannemann, A. Baiker, P. Boye, C.G. Schroer, J. Mater. Chem. 17 (2007) 2603–2606. [23] Y. Okamoto, K. Fukino, T. Imanaka, S. Teranishi, J. Phys. Chem. 87 (1983) 3740–3747. [24] A.A.G. Lima, M. Nele, E.L. Moreno, H.M.C. Andrade, Appl. Catal. A Gen. 171 (1998) 31–43. [25] S.V. Merzlikin, N.N. Tolkachev, L.E. Briand, T. Strunskus, C. Wöll, I.E. Wachs, W. Grünert, Angew. Chem. Int. Ed. 49 (2010) 8037–8041. [26] H. Bluhm, M. Hävecker, E. Kleimenov, A. Knop-Gericke, A. Liskowski, R. Schlögl, D.S. Su, Top. Catal. 23 (2003) 99–107. [27] M. Kurtz, H. Wilmer, T. Genger, O. Hinrichsen, M. Muhler, Catal. Lett. 86 (2003) 77–80. [28] Y. Guo, W. Meyer-Zaika, M. Muhler, S. Vukojević, M. Epple, Eur. J. Inorg. Chem. (2006) 4774–4781. [29] G.C. Chinchen, K.C. Waugh, D.A. Whan, Appl. Catal. 25 (1986) 101–107. [30] W.X. Pan, R. Cao, D.L. Roberts, G.L. Griffin, J. Catal. 114 (1988) 440–446. [31] W.R.A.M. Robinson, J.C. Mol, Appl. Catal. 60 (1990) 73–86. [32] Q. Sun, Y.-L. Zhang, H.-Y. Chen, J.-F. Deng, D. Wu, S.Y. Chen, J. Catal. 167 (1997) 92–105. [33] C. Kiener, M. Kurtz, H. Wilmer, C. Hoffmann, H.-W. Schmidt, J.-D. Grunwaldt, M. Muhler, F. Schüth, J. Catal. 216 (2003) 110–119. [34] M. Behrens, J. Catal. 267 (2009) 24–29. [35] K. Nishida, I. Atake, D. Li, T. Shishido, Y. Oumi, T. Sano, K. Takehira, Appl. Catal. A Gen. 337 (2008) 48–57.