Introduction

N2O is a stratospheric ozone depleting greenhouse gas that has a long lifetime of approximately 114 years and a global warming potential that is 310 times higher than that of CO21,2,3,4,5,6,7,8,9,10,11. In addition, recently, anthropogenic N2O emissions, from fossil fuel-using industries and biomass combustion, as well as from chemical plants that produce adipic and nitric acid, have been annually increasing8. If every country and/or international organisation in the world does not put in place mitigation strategies, N2O emissions are forecast to approximately double by 20508,9,10,11. To overcome these issues, catalytic decomposition of N2O to N2 and O2 is one of the most promising and economical techniques for emissions abatement, because the N2O emissions from combustion and chemical plants can be controlled using only end-of-pipe technologies employing exhaust heat. Over the past few years, supported metal oxides, noble metals and/or metal oxides and composite oxide (perovskites, hydrotalcites, spinels and hexa-aluminates) catalysts for N2O decomposition have been studied8,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34. In terms of supported catalysts, it has been shown that their activities for N2O decomposition follow the order of Ru, Rh, Ir > Pd > Cu > Fe > Pt > Ni > Mn8.

In the case of active Ru catalysts, as Ru species are efficient catalysts for the dissociation of the N–O bond in N2O, their catalysts are thus promising candidates for N2O decomposition26,27,28,29,30,31,32,33,34. However, the catalytic activities of Ru species are strongly related to their sizes, distributions and degrees of agglomeration. Zheng et al. studied the effect of support materials (MgO, SiO2, CeO2, Al2O3, TiO2, active carbon and SiC) on the catalytic properties of supported Ru for the N2O decomposition reaction31. They concluded that Ru/TiO2 shows high activity for N2O decomposition, where the activity was found to be related to the reducibility of the catalyst. Lin et al. also suggested that Ru supported on rutile-type TiO2 exhibited higher N2O decomposition activity compared with Ru/anatase-TiO2, Ru/Al2O3, Ru/SiO2 and other such systems, because its catalytic properties are induced by Ru metal dispersion and its monolayer structure30. Komvokis et al. prepared highly dispersed Ru/γ-Al2O3 via a conventional impregnation and ethylene glycol (EG) method and reported that Ru/γ-Al2O3 prepared using the EG method featured metallic Ru nanoparticles with a size of ca. 1–3 nm and high activity for N2O decomposition with H2O, SO2 and NO28. Recently, active Ru supported on perovskite-like La1.6Sr0.4NiO4 was prepared by Sui et al.33, who indicated that the activity arose due to the desorption of a large amount of oxygen from the active sites at low temperature and the ability of the oxygen vacancies to regenerate. Therefore, as stated above, Ru species are expected to be possible candidates for N2O decomposition catalysts due to their specific catalytic properties.

For N2O emissions from industrial combustion processes (i.e. fluidised bed combustion), it is considered that these emissions contain, amongst other gases, ca. 50–200 ppm of N2O, excess O2 (2–10%) and 10% H2O (water vapour)8,35,36. Therefore, as any catalysts that are developed may potentially be used in the decomposition of N2O from industrial combustion processes as a practical application, the evaluation of their catalytic properties under steady gas emissions conditions is required. Therefore, in this study, the N2O decomposition reaction properties of Ru supported on various oxide materials under excess O2 conditions were focused on, and the identities of the N2O adsorption species on these catalysts were confirmed. As the novel Ru/SnO2 catalysts in this work showed high activity for N2O decomposition, their local structures, optimal Ru loading and N2O adsorption species were determined. Finally, the reproducibility of the effects that H2O (water vapour) have on the N2O decomposition properties was evaluated.

Results and discussion

Ru supported on various metal oxide materials

Figure 1 shows the powder X-ray diffraction (PXRD) patterns of Ru supported on various metal oxide materials. The diffraction peaks for the Ru of all of the supported catalysts could be assigned to RuO2, and the Al2O3 of Ru/Al2O3 and TiO2 of Ru/TiO2 could be assigned to γ phase and brookite structures, respectively. As the other support materials also had compositional formulas, the solid-state reaction of RuO2 and the support materials was not observed.

Figure 1
figure 1

PXRD patterns of 5.0 wt% Ru supported on various different metal oxide materials prepared by impregnation, followed by drying and calcination at 600 °C for 3 h in air.

Full size image

Figure 2 shows a comparison of the temperature dependence of N2O conversion for the various different supported Ru catalysts. As none of the catalysts showed NO production, as detected by non-dispersive infrared (NDIR) spectroscopy, it was presumed that N2O decomposed into N2 and O2. For Ru/SnO2 and Ru/ZrO2, the light-off curves of N2O were obtained at approximately 200 °C, although the light-off temperature at which 90% conversion of N2O occurred was not reached for all catalysts at a reaction temperature of 600 °C. In the case of Ru/Nb2O5, it showed almost no activity.

Figure 2
figure 2

Catalytic activity for N2O decomposition reaction over 5.0 wt% Ru supported on various different metal oxide materials. Reaction conditions: 200 ppm of N2O, 10% O2 and N2 balance at 100 cm3 min−1 (W/F = 5.0 × 10−4 g·min cm−3).

Full size image

Table 1 summarises the properties of the different catalysts, in which the activity is expressed in terms of the light-off temperature at which 50% conversion of N2O was achieved (T50). The T50 values increased in the order of SnO2 < ZrO2 < Al2O3 < CeO2 < Ta2O5 < TiO2 ≈ WO3 ≈ Nb2O5, which bears no relation to the Brunauer − Emmett − Teller surface area (SBET) values of their catalysts. The order of the T50 values is almost consistent with the order of the reduction temperatures observed from the H2-temperature-programmed reduction (TPR) experiments (see Table 1 and Supplementary Figure S1): Al2O3 < SnO2 < ZrO2 < CeO2 < Ta2O5 < TiO2 < WO3 < Nb2O5, suggesting that the redox properties for Ru (RuO2) at low reaction temperature are closely related to the N2O decomposition activity. Because, in addition, the supported Ru catalysts with lower reduction temperatures tended to exhibit higher metal dispersion (smaller particle size), it is considered that highly dispersed Ru (RuO2) particles act as an active catalyst for N2O decomposition reaction. Ru/CeO2 exhibited high metal dispersion (18%), but its activity showed medium. In case of CeO2-supported catalysts, the overestimation for the metal dispersion was previously reported, because of CO adsorbed on CeO2 support as carbonate species37. Therefore, it is suggested that the overestimation for the dispersion for Ru/CeO2 was also caused by CO adsorption on CeO2, and its relationship between the dispersion and activity for Ru/CeO2 are low. Ru3d XPS spectra for 5.0 wt% Ru catalysts supported on various oxide materials were obtained (Supplementary Figure S2). According to the previous report for Ru3d XPS spectra analysis38, the binding energies of Ru3d5/2 peaks for all catalysts could be assigned to the oxidation state of Ru4+ (approximately 280.4 eV), which is consistent with the assignment for RuO2 of XRD patterns. In addition, Ru3d5/2 peak area and intensity for Ru/SnO2 showed slightly higher in comparison with the other catalysts, which indicates that Ru surface concentration for Ru/SnO2 is higher in agreement with high Ru metal dispersion estimated by CO adsorption.

Table 1 Catalytic properties of 5.0 wt% Ru supported on various different oxide materials.
Full size table

To study the acid and base properties of the supported Ru catalysts, the amount of desorbed gas per weight on the catalyst was estimated using NH3- and NO-temperature-programmed desorption (TPD) and was found to be in the range 100–500 °C (see Supplementary Figure S3 and Figure S4 for more details), the data of which are also summarised in Table 1. Ru/Al2O3 and Ru/TiO2 showed high amounts of desorbed NH3, whereas Ru/CeO2 exhibited a low amount despite it having a relatively high SBET value. The other catalysts also showed lower acidity than Ru/Al2O3. However, in terms of NO desorbability, there was no significant difference as well as no relation to the SBET value. However, it was also implied that the supported Ru with a higher amount of NO desorbability tended to approximately exhibit higher N2O decomposition activity, which therefore indicated that there is almost a correlation between the base properties and the catalytic N2O decomposition activity.

To confirm the identities of the N2O adsorption species on the various different supported Ru catalysts and the reasons for their different activities, in situ Fourier-transform infrared (FTIR) spectra were recorded at 200 °C, which is the approximate initiation temperature for N2O decomposition (Fig. 3). Based on previous reports32,39,40,41,42, two bands at 2238 and 2008 cm−1, attributed to adsorbed N2O, were observed for all of the catalysts. Several other bands were also observed in the range of 1000–1700 cm−1, which were attributed to nitrites, nitrates and nitro compounds that were adsorbed on the catalysts. These bands determinably appeared for Ru/CeO2 and Ru/ZrO2, which is consistent with the high NO adsorbability (basicity) in the range of 300–600 °C (see Supplementary Figure S4 for more details). In addition, the band at approximately 1870 cm−1 attributed to adsorbed NO was observed for Ru/Al2O3, Ru/CeO2, Ru/ZrO2 and Ru/SnO2, which showed that these systems have relatively high activity for N2O decomposition. Therefore, it was concluded that this band attributed to NO species can be considered as belonging to activated N2O.

Figure 3
figure 3

In situ FTIR spectra of N2O adsorbed on 5.0 wt% Ru supported on various different metal oxide materials, measured at 200 °C in gas feeds of 200 ppm of N2O and N2 balance.

Full size image

Effects of Ru loading on catalytic N2O decomposition

As the Ru/SnO2 catalyst exhibited high N2O decomposition activity, the optimal amount of Ru loading, its local structure and catalytic properties were comprehensively studied. In the PXRD patterns of the catalysts with different loading amounts of Ru (see Supplementary Figure S5 for more details), the diffraction peaks for SnO2 can be observed for all of the catalysts, whereas the diffraction peaks for RuO2 can be observed in the patterns for the catalysts with a Ru loading of higher than 5.0 wt%. In addition, the intensities of the diffraction peaks of RuO2 increased upon an increase in the Ru loading, which probably suggests an increase in the crystallinity and/or particle size upon the increase in RuO2. In terms of the SBET values of the catalysts with different amounts of Ru loading (see Supplementary Table S1 for more detail), the values were observed to decrease upon an increase in the Ru loading, probably because of the lower proportion of SnO2 to a higher proportion of RuO2. The local structure of the 5.0 wt% Ru/SnO2 sample was characterised using scanning transmission electron microscopy and energy-dispersive X-ray mapping (STEM-EDX) mapping (Fig. 4). The bright-field STEM image revealed SnO2 particles present in the sample with sizes of approximately 10–100 nm. From the overlay image of the EDX mapping, Sn − L (blue) and Ru − K (green) fluorescence lines can be observed. Therefore, based on these and the PXRD results, RuO2 particles with a size of approximately 50 nm, as shown by the arrows in Fig. 4, were dispersed on the SnO2 support.

Figure 4
figure 4

(left) TEM image and (right) EDX mapping analysis of 5.0 wt% Ru/SnO2. The blue and green dots correspond to the Sn − L and Ru − K fluorescence lines, respectively.

Full size image

In Fig. 5 and Supplementary Table S1 the temperature dependence of N2O conversion are compared for 0.5–20 wt% Ru/SnO2. It was found that 0.5 and 1.0 wt% Ru/SnO2 showed a lower activity of catalytic N2O decomposition, whereas 5.0–20 wt% Ru/SnO2 exhibited almost the same light-off profile for N2O and T50. Therefore, it was assumed that the optimal Ru loading supported on SnO2 was approximately 5.0 wt%. Supplementary Figure S6 also shows N2O conversion, NO selectivity and mass signal for N2O decomposition reaction over 5.0 wt% Ru/SnO2. NO selectivity was not observed. For the production of NO2, the mass signal for m/z value of 44 for N2O decreased, which is consistent with N2O conversion, whereas the mass signal for m/z value of 46 for NO2 was constant. In addition, we estimated the energy difference to generate NO2 from NO and O by density functional theory calculations at the level of BP8643,44/def-SV(P)45 under the resolution of identity approximation46 using TURBOMOLE47. Although NO2 can be formed barrierlessly from NO and O, the preparation of O from O2 requires large energy of 6.24 eV, which is larger than the NO2 formation energy from NO and O, 4.23 eV. Therefore, it is considered that not only NO but also NO2 were not produced during the N2O decomposition reaction over 5.0 wt% Ru/SnO2.

Figure 5
figure 5

Catalytic activity for the N2O decomposition reaction over 0.5–20 wt% Ru/SnO2. Reaction conditions: 200 ppm of N2O, 10% O2 and N2 balance at 100 cm3·min−1 (W/F = 5.0 × 10−4 g·min·cm−3).

Full size image

To also confirm the relationship between the N2O adsorption species and decomposition activities, in situ FTIR spectra were recorded for 0.5–20 wt% Ru/SnO2 at 200 °C (Fig. 6). For all of the Ru/SnO2 samples, two bands were observed for adsorbed N2O, at 2238 and 2008 cm−1. However, bands for adsorbed NO (1870 cm−1) and NOx compounds (1000–1700 cm−1) were present for the 5.0–20 wt% Ru/SnO2 that have high activities, which is consistent with the results that is Ru supported on the different metal oxide materials. Therefore, it is expected that this NO band can be considered as belonging to activated N2O. In situ FTIR spectra of N2O adsorbed on 5.0 wt% Ru/SnO2 at 300 °C and 400 °C were also confirmed (Supplementary Figure S7). The two bands attributed to adsorbed N2O (2008 and 2238 cm−1) were observed at 300 °C, whereas the bands were not observed at 400 °C. These bihaviors probably imply that catalytic N2O decomposition reaction was proceeded.

Figure 6
figure 6

In situ FTIR spectra of N2O adsorbed on 0.5–20 wt% Ru/SnO2. The spectra were measured at 200 °C in gas feeds of 200 ppm of N2O and N2 balance.

Full size image

Stability, reproducibility and the effects of water vapour

To evaluate the stability and reproducibility of the 5.0 wt% Ru/SnO2 catalyst, the N2O decomposition reaction was repeated five times. Figure 7, Supplementary Figure S8 and Supplementary Table S2 give summaries of the temperature dependence of N2O conversion for repeated cycles using 5.0 wt% Ru/SnO2. Upon repetition of the catalytic cycle, the light-off temperature and T50 tended to slightly increase. Because there is concern for the vaporisation as RuO3 and/or RuO448, it is considered that the stability of RuO2 is low at high temperature. XRD patterns and Ru3d XPS spectra for 5.0 wt% Ru/SnO2 after N2O decomposition reaction were obtained (Supplementary Figure S9). For XRD pattens, the diffraction peaks for Ru/SnO2 after the reaction could be assigned to RuO2 and SnO2. In comparison with as-prepared catalysts (before reaction), not only the diffraction peaks but also Ru3d XPS spectra for 5.0 wt% Ru/SnO2 after reaction are almost no change. The Ru metal dispersion for 5.0 wt% Ru/SnO2 after N2O decomposition reaction at 600 °C was also estimated, and it showed 6%. In comparison with the dispersion for as-prepared 5.0 wt% Ru/SnO2 (7%), it is slight low. Therefore, it is suggested that the sintering of Ru (RuO2) was slightly induced by the N2O decomposition reaction at 600 °C. However, the light-off temperature at which 90% conversion was achieved was almost reached for all of the catalysts at 600 °C. Therefore, the stability and reproducibility of 5.0 wt% Ru/SnO2 were confirmed. In addition, time-on-stream stability of catalytic activity for the N2O decomposition reaction over 5.0 wt% Ru/SnO2 at 400 °C was also studied (Supplementary Figure S10). Because the stable N2O conversition (ca. 55%) was confirmed in approximately 2 h, it is considered that the the catalytic stability for 5.0 wt% Ru/SnO2 was verified.

Figure 7
figure 7

Catalytic activity (T50) for the N2O decomposition reaction over 5.0 wt% Ru/SnO2 as a function of the repetition number, and the effects of water vapour (10% H2O). Reaction conditions: 200 ppm of N2O, 10% O2 and N2 balance at 100 cm3·min−1 (W/F = 5.0 × 10−4 g min cm−3).

Full size image

The effects of water vapour (10% H2O) on the N2O decomposition reaction over as-prepared 5.0 wt% Ru/SnO2 were also evaluated (see the plots in Supplementary Figure S8). The N2O conversion profile for the reaction with H2O shifted to a higher temperature than that without H2O. In comparison with the T50 values for the N2O decomposition reaction carried out in the absence of H2O, the T50 values of that carried out in the presence of H2O was higher, at 50 °C (see Supplementary Table S2 for more details). Therefore, it could be deduced that the deactivation of 5.0 wt% Ru/SnO2 and/or inhibition of N2O adsorption on 5.0wt % Ru/SnO2 was induced by the presence of H2O.

Conclusion

Ru catalysts supported on various different metal oxides were prepared by impregnation to evaluate their decomposition properties for N2O, which is a powerful greenhouse gas that is present in industrial emissions. From the PXRD patterns, the diffraction peaks for Ru of all of the supported catalysts could be assigned to RuO2. The T50 values were found to increase in the order of: SnO2 < ZrO2 < Al2O3 < CeO2 < Ta2O5 < TiO2 ≈ WO3 ≈ Nb2O5, which is almost consistent with the order of the reduction temperatures observed from the H2-TPR measurements. Therefore, it is considered that the redox properties for Ru (RuO2) at low reaction temperature are closely associated with N2O decomposition activity. In addition, according to the NO-TPD profiles and in situ FTIR data, a correlation can also be observed between the NO adsorption properties as well as the NO species considered to be activated N2O and the catalytic N2O decomposition activity. Among the supported Ru catalysts, Ru/SnO2 showed a high catalytic performance for the N2O decomposition reaction. SnO2 support materials induced the physicochemical properties, high reducibility (redox property), Ru (RuO2) dispersion and basicity for the catalysts, which are required for high N2O decomposition activity. As 5.0–20 wt% Ru/SnO2 exhibited almost the same light-off profiles for N2O and T50, the optimal amount of Ru loaded on the SnO2 support was found to be approximately 5.0 wt%. Although the deactivation of Ru/SnO2 was induced by H2O, the stability and reproducibility for N2O decomposition activity of Ru/SnO2 were confirmed.

Methods

Catalyst preparation

A wide variety of commercially available metal oxides, Al2O3 (JRC-ALO-8, Catalysis Society of Japan), CeO2 (JRC-CEO-5, Catalysis Society of Japan), TiO2, Nb2O5, ZrO2, Ta2O5, WO3 and SnO2 (Kojundo Chemical Lab.), were used as support materials for Ru. Supported Ru (5.0 wt% loading as Ru) samples were prepared via the impregnation of an aqueous solution of RuCl3 (Fujifilm Wako Pure Chemical Corporation), followed by drying and calcination at 600 °C for 3 h in air. To study the effects of Ru loading, 0.5–20 wt% of Ru/SnO2 catalysts that show a high performance in N2O decomposition were prepared in a similar manner for comparison.

Characterisation

PXRD measurements were performed using mono-chromated Cu Kα radiation (40 kV, 15 mA, MiniFlex600, Rigaku). The chemical compositions of the samples were determined from X-ray fluorescence measurements (XRF, MESA-500 W, Horiba). Spectra from X-ray photoelectron spectroscopy (XPS) were obtained using Al Kα radiation (PHI 5000-VersaProbe, Ulvac-Phi). The C1s signal at 284.8 eV that was derived from adventitious carbon was used as a reference to correct for the effect of surface charge. STEM-EDX mapping were performed using a JEM-ARM200CF microscope (Jeol). SBET calculations were performed using N2 adsorption isotherms, which were obtained at − 196 °C (ASAP2020, Micromeritics). H2-TPR measurements were performed in a flow system (5% H2/Ar) at a constant rate of 10 °C min−1 (Bel-cat, Microtrac-Bel). The NH3 and/or NO adsorbability of the catalysts were also studied through TPD. Prior to the measurements, the catalysts were treated at 500 °C for 1 h under an Ar flow and were subsequently cooled at 100 °C for 30 min in 5% NH3/Ar and 1% NO/Ar (50 cm3 min−1). After pre-treatment, the catalysts were heated to 500 °C under a He flow at a constant rate of 10 °C·min−1. The concentrations of the desorbed NH3 and/or NO in the effluent gas were analysed using an online thermal conductivity detector (TCD) signal (Bel-cat, Microtrac-Bel). The Ru metal dispersion was determined by pulsed CO chemisorption at 50 °C (Bel-metal, Microtrac-Bel) after the catalyst was oxidised using O2 and subsequently reduced using H2 at 400 °C. The metal dispersion was calculated from the molar ratio of chemisorbed CO to loaded Ru by assuming that the chemisorption stoichiometry of Ru:CO was 1:1.

Catalytic N2O decomposition tests

The catalytic decomposition of N2O was performed in a flow reactor at atmospheric pressure. Catalysts (10–20 mesh, < 0.3 mm thickness, 50 mg) were fixed in a quartz tube (outside diameter: 6 mm, inside diameter: 4 mm) with quartz wool at both ends of the catalyst bed. The temperature dependence of the catalytic activity was evaluated by heating the catalyst bed from room temperature to 600 °C at a constant rate of 10 °C min−1 while a gas mixture containing 200 ppm of N2O, 10% O2 and N2 balance was supplied at 100 cm3 min−1 (W/F = 5.0 × 10−4 g min cm−3). For the catalysts that exhibited high performance, the reactions were repeated to evaluate the stability of the catalysts and the reproducibility of the results. Effects of water vapour (10% H2O) on the N2O decomposition reaction were also evaluated. The N2O, NO and gas concentrations were analysed using an NDIR analyser (VA-3011, Horiba) and quadrupole mass spectrometer (PrismaPlus, Pfeiffer).

In situ FTIR spectra were recorded on a FT/IR-6600 spectrometer (Jasco) using a diffuse-reflectance reaction cell with a BaF2 window connected to a gas supply and a heating system to enable measurements to be conducted under atmospheric pressure. First, the catalysts were preheated in situ in a flow of Ar at 400 °C for 30 min prior to each experiment. After pre-treatment, the temperature of the catalyst was decreased to 200 °C, followed by the subsequent purging of the cell with Ar and then filling with 200 ppm of N2O/N2 gas. Finally, FTIR spectra were recorded while the catalysts were maintained under a stream of N2O/N2.