Magdalena Jabłońska*ab,
Miren Agote Aránc,
Andrew M. Bealecd,
Kinga Góra-Mareke,
Gérard Delahayf,
Carolina Petittof,
Kateřina Pacultovág and
Regina Palkovits*ab
aChair of Heterogeneous Catalysis and Chemical Technology, RWTH Aachen University, Worringerweg 2, 52074 Aachen, Germany. E-mail: Palkovits@itmc.rwth-aachen.de; Jablonska@itmc.rwth-aachen.de; Fax: +49 241 80 22177; Tel: +49 241 80 26497
bCenter for Automotive Catalytic Systems Aachen – ACA, RWTH Aachen University, Schinkelstr. 8, 52062 Aachen, Germany
cDepartment of Chemistry, University College London, 20 Gordon Street, London, WC1H 0AJ, UK
dUK Catalysis Hub, Research Complex at Harwell, Rutherford Appleton Laboratories, Didcot, Oxon OX11 0FA, UK
eFaculty of Chemistry, Jagiellonian University in Kraków, Gronostajowa 2, 30-387 Kraków, Poland
fInstitut Charles Gerhardt de Montpellier, 240 Avenue du Professeur Emile Jeanbrau, 34296 Montpellier Cedex 5, France
gVŠB-Technical University of Ostrava, 17. listopadu 15, 708 33 Ostrava, Czech Republic
First published on 29th January 2019
Cu–Al–Ox mixed metal oxides with intended molar ratios of Cu/Al = 85/15, 78/22, 75/25, 60/30, were prepared by thermal decomposition of precursors at 600 °C and tested for the decomposition of nitrous oxide (deN2O). Techniques such as XRD, ICP-MS, N2 physisorption, O2-TPD, H2-TPR, in situ FT-IR and XAFS were used to characterize the obtained materials. Physico-chemical characterization revealed the formation of mixed metal oxides characterized by different specific surface area and thus, different surface oxygen default sites. The O2-TPD results gained for Cu–Al–Ox mixed metal oxides conform closely to the catalytic reaction data. In situ FT-IR studies allowed detecting the form of Cu+⋯N2 complexes due to the adsorption of nitrogen, i.e. the product in the reaction between N2O and copper lattice oxygen. On the other hand, mostly nitrate species and NO were detected but those species were attributed to the residue from catalyst synthesis.
The chemical composition of mixed metal oxides was determined by ICP-MS using an Agilent Technologies 8800 Triple Quad spectrometer. Prior to the measurement, the sample (50 mg) was dissolved in 6 cm3 mixture of concentrated acids (HCl:HNO3, 1:1), and afterwards, the resulting mixture was diluted with 64 cm3 deionized water before warming up to 40 °C for 24 h.
The specific surface area (SBET) of the mixed metal oxides was determined by low-temperature (−196 °C) N2 sorption using a Quantachrome Quadrasorb SI. Prior to nitrogen adsorption, the samples were outgassed at 250 °C for 12 h using a Quantachrome Flovac degasser. The specific surface area (SBET) was calculated using the Brunauer–Emmett–Teller (BET) multiple point method in the p/p0 range from 0.05 to 0.3.
The temperature-programmed desorption of O2 (O2-TPD) was performed to investigate oxygen desorption behaviour using a Micromeritics AUTOCHEM 2910. 49–58 mg of sample was loaded into a quartz tube and pretreated in reconstituted air (20.0 vol% O2/80.0 vol% N2) at 600 °C (30 cm3 min−1). After cooling down to 80 °C, the pretreated sample was heated up to 600 °C with a linear heating rate of 10 °C min−1 in a carrier gas of He (15 cm3 min−1) and kept for 10 min at 600 °C and cooled down up to 80 °C in the same flow (He). Then this material was heated, under a flow of 1.0 vol% N2O/He (30 cm3 min−1), at 600 °C (plateau 10 min) and cooled to 80 °C. Afterward the second run of O2-TPD was performed under He flow (15 cm3 min−1). O2-TPD was also carried out (15 cm3 min−1 of He, 10 K min−1, 80–600 °C, 10 min at 600 °C) over 52–56 mg of sample after heating (up to 600 °C, plateau 10 min) and cooling down (80 °C) in the presence of 1.0 vol% N2O/He (30 cm3 min−1).
The redox properties of the mixed metal oxides were studied by the temperature-programmed reduction (H2-TPR) using Quantachrome ChemBET Pulsar TPR/TPD. H2-TPR runs for the samples (50 mg) were carried out starting from room temperature to 1000 °C, with a linear heating rate of 10 °C min−1 and in a flow of 5.0 vol% H2/Ar (25 cm3 min−1). Water vapour was removed from the effluent gas by the means of a cold trap placed in an ice-water bath. The H2 consumption was detected and recorded by a TCD detector.
IR spectra were recorded with a Tensor 27 Bruker spectrometer equipped with an MCT detector. Prior to FTIR studies selected mixed metal oxides were pressed into the form of self-supporting wafers (ca. 5–10 mg cm−2) and pretreated in situ in a homemade quartz IR cell at 400 °C under vacuum conditions for 1 h. The spectral resolution was 2 cm−1. Sorption of N2O was performed at room temperature. Next, the N2O contacted sample was heated to 330, 390, and 450 °C, kept at this temperature for 2 min and cooled down to room temperature, while collecting the spectrum.
Cu–Al–Ox | CuO particle sizea/nm | Cu/Al molar ratio | Cu/wt% | Na content/wt% | SBET/m2 g−1 | H2 uptakec/mmol g−1 | O2(des)d/μmol g−1 (e/f/g) | |
---|---|---|---|---|---|---|---|---|
Theoretical | Determinedb | Determinedb | ||||||
a Estimated by the Scherrer's formula for (111) reflection.b Determined with ICP-MS analysis.c Calculated by the equation: Y = 9 × 10−9X + 2 × 10−7, R2 = 0.9996, and X, Y referred to the area of each reduction peak and the H2 consumption, respectively.d Estimated from direct O2-TPD of known amount of as-stored Ag2O (Stream Chemicals). O2 desorbed during O2-TPD for pretreated materials under reconstituted air (20.0 vol% O2/80.0 vol% N2), (e), and subsequently reoxidised by 1.0 vol% N2O/He (f), or for pretreated materials under 1.0 vol% N2O/He (g). | ||||||||
60/30 | 15 | 2.00 | 2.29 | 48.4 | 0.6 | 84 | 7.8 | 46.9/27.3/36.3 |
75/25 | 14 | 3.00 | 3.06 | 55.0 | 1.0 | 68 | 8.4 | 32.1/24.9/25.1 |
78/22 | 17 | 3.55 | 3.11 | 41.8 | 1.1 | 70 | 9.4 | 30.1/17.5/17.5 |
85/15 | 21 | 5.67 | 4.65 | 51.3 | 0.5 | 34 | 9.4 | 26.2/17.5/14.3 |
Cu–Al–Ox mixed metal oxides show specific surface area (SBET) in the range of 34–84 m2 g−1. SBET varied between materials with different molar ratios. While no correlation existed between copper content (wt%) and specific surface area, the specific surface area increased with decreasing Cu/Al molar ratio. Fig. 2 presents O2 desorption rates evaluated for the pretreated materials under reconstituted air and subsequently reoxidized by 1.0 vol% N2O/He, or for pretreated materials under 1.0 vol% N2O/He. The O2-TPD profiles of Cu–Al–Ox mixed metal oxides showed peaks related to the desorption of surface oxygen species around 150–400 °C, whilst the peak above 400 °C is attributed to the desorption of lattice oxygen.15 The O2-TPD profiles are dominated by high-temperature peaks (Table 1), indicating that the increasing Cu/Al molar ratios decreased the molar amount of desorbed lattice oxygen. Also, the presence of copper in CuxCo3−xO4 led to a lower amount of desorbed O2 compared to Co3O4.16 Nevertheless, the quantity of oxygen desorbed seems to increase with the specific surface area. It should be emphasized that the amount of oxygen desorbed per gram of catalyst is negligible compared to the amount of oxygen consumed by hydrogen during the TPR (see below H2-TPR results). This point seems to indicate that the desorbed oxygen comes rather from the surface of the material.
The H2-TPR profiles of Cu–Al–Ox mixed metal oxides revealed one main broad peak between 200 and 400 °C corresponding to the reduction of bulk copper oxide species to metallic copper, as shown in Fig. 3. The shape of peak maxima and H2 uptake (Table 1) obtained for mixed metal oxides matched to that of pure CuO (maximum at about 350 °C, H2 uptake of 10.7 mmol g−1). According to the XRD analysis, the peaks associated with the CuO were the main peaks observed in the mixed metal oxides. Otherwise, for Cu75Al25Ox and Cu60Al30Ox the reduction of copper in CuAl2O4 cannot be excluded. A quantitative analysis of H2 consumption based on integrating the H2-TPR curves confirmed that H2 uptake did not change significantly over mixed metal oxides (7.8–9.4 mmol g−1).
Fig. 3 H2-TPR profiles of Cu–Al–Ox mixed metal oxides; experimental conditions: mass of the catalysts = 30 mg; [H2] = 5.0 vol%, Ar balance, flow rate = 25 cm3 min−1, linear heating of 10 °C min−1. |
Fig. 4 shows the results of N2O decomposition over Cu–Al–Ox mixed metal oxides with varying Cu/Al molar ratios. The effect of sample composition on the activity was studied to find the optimum materials for maximum conversion. The highest activity among the tested catalysts reached material with a molar ratio Cu/Al = 60/30 (conversion of 25% at 450 °C). The other catalysts reached significantly lower activities (below 20% at 450 °C). Comparable results were obtained by Kannan,9 who found a conversion of 48% at 450 °C over Cu–Al–Ox (Cu/Al = 3/1 mol. ratio, 0.1 g catalyst, 0.0985 vol% N2O/He, 100 cm3 min−1). The most active material was also tested in simulated waste gas conditions – in the presence of 5.0 vol% O2 and 2.0 vol% H2O. O2 present in the feed caused a steep drop in conversion; what is important, the inhibiting effect was fully reversible. The stepwise addition of H2O (results not shown) had a detrimental effect on N2O conversion probably due to the strong adsorption of the water molecule on the surface, since the initial activity in inert gas was not fully recovered after removal of O2 and H2O. Table 2 lists examples of Cu-containing catalyst tested for N2O decomposition. It can be found that either (supported) copper oxide or hydrotalcite derived mixed metal oxides are inherently not active in deN2O. Further modification of Cu-containing material with noble/rare earth metals can significantly improve their catalytic activity [e.g. ref. 1 and 9].
Fig. 4 Results of catalytic tests performed for Cu–Al–Ox mixed metal oxides; experimental conditions: [N2O] = 0.1 vol%, ([O2] = 5.0 vol%), N2 balance, SV = 60 l g−1 h−1. |
Catalyst code/Composition | Reaction conditions | Temp. for N2O conversion | Ref. |
---|---|---|---|
Copper oxide, supported copper oxide | |||
CuO commercial | 0.5 vol% N2O/He, SV = 6 l g−1 h−1 | 400 °C/<5% | 3 |
CuO mesoporous | 400 °C/30% | ||
Cu–Zn/Al2O3, 35 wt% of Cu–Zn | 0.7 vol% N2O/2 vol% O2/N2, GHSV = 7200 h−1 | 480 °C/50% | 4 |
Hydrotalcite derived mixed metal oxides | |||
Cu–Al–Ox, Cu/Al = 60/30, mol. ratio | 0.1 vol% N2O/N2, SV = 60 l g−1 h−1 | 450 °C/25% | This study |
Cu–Al–Ox, Cu/Al = 3/1, mol. ratio | 0.0985 vol% N2O/He, SV = 60 l g−1 h−1 | 450 °C/48% | 9 |
Cu–Mg–Al–Ox, Cu/Mg/Al = 10/61/29, mol. ratio | 0.5 vol% N2O/4.5vol% O2/He, SV = 30 l g−1 h−1 | 600 °C/100% | 8 |
The most active catalyst (Cu/Al = 60/30, mol. ratio) was studied by in situ EXAFS during N2O decomposition. The Cu K-edge absorption spectra were collected: at room temperature, during the temperature ramp (100 to 450 °C) under N2O/He feed, and finally at room temperature after deN2O. The Cu K-edge EXAFS spectra for the catalyst collected at different stages of the experiment (Fig. 5A) presents similar phases as the CuO reference; in line with the XRD results, suggesting CuO as the main species discernible in the samples. The reduced EXAFS oscillation amplitudes in Cu60Al30Ox can be attributed to differences in sample measuring conditions: CuO was measured in pellet form while the catalyst was measured as granulated powder in a microreactor, inhomogeneities and pin-hole effects in the latter case resulted in EXAFS amplitude reduction. No significant changes in the Cu60Al30Ox spectra were observed during activation or N2O exposure indicating the CuO local structure remains constant.
The Fourier transform of the X-ray absorption fine structure (FT-EXAFS) spectra of the catalyst in Fig. 5B gives insight into the distance of neighboring atoms around the absorber atom. The peaks with maxima around 0.19 and 0.29 nm correspond to neighbor O and Cu atoms, respectively. A decrease in peak intensity was observed for the data collected at increasing temperatures. This decrease is attributed to the increasing atom vibration due to thermal effects which smear out the EXAFS oscillations affecting the signal intensity in the Fourier transform. The position of the peaks did not vary significantly throughout the experiment up to 250 °C evidencing no changes in bond distance or local Cu geometry in the sample measured at 450 °C. When the catalyst is active, there seems to be a slight shift to shorter Cu–O distance which could suggest the formation of Cu+. Nonetheless, a definite conclusion cannot be drawn here as no changes were discernible in the position of the rising absorption edge in the XANES to account for Cu2+ reduction to Cu+.
For FT-IR, the selected materials – Cu60Al30Ox and Cu85Al15Ox (the most active and less active sample in deN2O, respectively) studies were contacted with the N2O dose (60 Tr of N2O per 10 mg of the sample) at room temperature, then the IR spectrum was collected (Fig. 6). Next, the catalysts were heated to 330 °C, kept for 10 min and then cooled down to RT to collect IR spectrum of all the reaction products. The procedure was repeated for 390 and 450 °C. An intense absorption band around 2220–2230 cm−1 recorded at room temperature appeared due to adsorbed N2O. In the spectra recorded at higher temperatures, bands appeared below 1700 cm−1 due to the appearance of nitrites, nitrates and nitro compounds. Also, the gaseous NO can be easily identified in the IR spectra due to the vibration-rotation bands at 1950–1850 cm−1. Nitric oxide was found to play a crucial role in the kinetic oscillations caused by a complex interaction of different reactions. Both NO and molecular oxygen can be formed by the decomposition of the residual nitrate species from the catalyst synthesis.17 It is generally accepted that nitrate species on Cu-ZSM-5 and other zeolite catalysts modified with copper are stable even at high temperatures. Furthermore, nitrate moieties have been proposed to be important intermediates in the decomposition of NO,18 the selective catalytic reduction of NO by hydrocarbons,19 and the SCR of NO by ammonia.20 The spectra of Cu–Al–Ox mixed metal oxides were consistent with the formation of bridged nitrates at 1674 cm−1, surface nitrates at 1614 and 1576 cm−1, monodentate nitrates at 1416 cm−1, nitro groups at 1353 cm−1, as well as bidentate nitrates at 1312 cm−1.21,22
For the most active Cu60Al30Ox, the existence of N-species implied a reaction between N2O and copper lattice oxygen. What is more, the formation of end-product of the reaction, i.e. N2 molecule is strongly supported by the presence of 2301 cm−1 band attributed to Cu+⋯N2 complexes. N2 molecule interacting with Cu+ ions inside zeolites leads to a ν(NN) band in the 2300–2290 cm−1 region, which is significantly downshifted with respect to gas-phase value of 2321 cm−1. This bathochromic shift could be explained in terms of chemical interactions involving molecular orbitals of the probe molecule and a suitable d-orbital of the metal cation.23 Basing on the wavenumber and half-bandwidth of 2301 cm−1 band identical to the band of Cu+⋯N2 adducts in Cu-zeolites we advocated on the formation of this adducts also in Cu–Al–Ox mixed metal oxides. The presence of stable Cu+⋯N2 complexes indicated the co-presence of both Cu+ and Cu2+ cations on the catalyst surface, as expected for a redox mechanism requiring a balance between these two sites. It is also in line with the CO sorption results on spent material Cu60Al30Ox (spectra not shown) which indicated that the surface Cu is poorer in the Cu2+ cationic species. Such observation allows for concluding on the efficient catalyst which should be characterized by the high redox ability to reduce Cu2+/Cu+ redox pair. Furthermore, in this particular case, Cu+ cations are not able to efficiently transform back to a Cu2+ state. The poorest activity of Cu85Al15Ox was evidenced as no completed decomposition of N2O (the 2224 cm−1 band) could be achieved even at a temperature as high as 450 °C (Fig. 6).
The activity varied among the tested materials, however, no clear trend related either to the Cu or Na content became evident. While alkali metals can act as basic centres and significantly influence catalytic activity.24 For example, Obalová et al.25 pointed out that 1.15 wt% of Na introduced by impregnation already slightly enhanced activity of Co4MnAlOx. In our case, Na residual remained after preparation procedure, and actually with lower values than 1.15 wt% (Table 1). At this stage, it is not possible to precisely justify the influence of residual Na on materials catalytic activity in deN2O.
N2O decomposition on Cu–Al–Ox yielded N2 and O2, and the reoxidation of copper side took place with gas phase N2O (eqn (1)):
Cu* + N2O → Cu*O + N2. | (1) |
2Cu*O → 2Cu* + O2. | (2) |
This journal is © The Royal Society of Chemistry 2019 |