Al2O3-supported Pt/Rh catalysts for NOx removal under lean conditions

Al2O3-supported Pt/Rh catalysts for NOx removal under lean conditions L. Castoldi1*, R. Matarrese1, M. Daturi2, J. Llorca3, L. Lietti1* 1Politecnico di Milano, Laboratory of Catalysis and Catalytic Processes, Dipartimento di Energia, Via La Masa, 34, 20156 Milano (Italy) 2Normandie Université, ENSICAEN, UNICAEN, CNRS, Laboratoire Catalyse et Spectrochimie, 14000 Caen, France 3Institute of Energy Technologies and Centre for Research in Nanoengineering, Universitat Politècnica de Catalunya, Barcelona, Spain *corresponding author: [email protected]; [email protected]   Abstract In this work the reactivity of Pt-Rh NOx storage-reduction (NSR) catalysts in the reduction of NOx under lean conditions is investigated. It is found that significant amounts of NOx are stored on both Rh- and Pt-based samples at all the investigated temperatures (in the range 150–350¿°C). Mostly chelating nitrites are adsorbed at the lowest investigated temperature (150¿°C), while nitrates (both bidentate and ionic) at higher temperatures. However, at all temperatures nitrites prevail at the beginning of the storage phase, while nitrates represent the most abundant adsorbed species after prolonged contact. Pt-containing catalysts (either monometallic Pt or bimetallic Pt/Rh) show higher NOx storage capacity than the Rh monometallic sample, possibly due to the higher dispersion of Pt vs. Rh and/or to the higher oxidizing capability of Pt vs. Rh. The stored NOx species show relevant thermal stability, and decompose to NOx and O2 upon heating. In particular, nitrites disproportionate to gaseous NO and nitrates; these latter then decompose to NOx and O2. On the Rh-Ba/Al2O3 catalyst the disproportionation reaction is observed with a higher temperature onset if compared to the Pt-based samples. The analysis of the reactivity of the stored NOx species (probed by isotopic labeling experiments and reduction with H2 and NH3) showed the lower reactivity of the Rh-Ba/Al2O3 sample; however Rh shows activity in the ammonia decomposition reaction to N2 and H2, unlike Pt. The lower reactivity of the Rh-Ba/Al2O3 sample is also pointed out by experiments under cyclic lean-rich conditions. However, the presence of Rh increases the reactivity of the catalyst in the steam reforming of hydrocarbons, especially at high temperature, and accordingly the reactivity of the bimetallic Pt/Rh sample at high temperatures is higher than that of the Pt and Rh monometallic catalysts. 1. Introduction The push for better fuel economy and lower vehicular CO2 emissions has led to increased deployment of lean-burn engines, being widely used in heavy duty diesel and light duty gasoline engine. However, exhaust gases from these engines contain NOx and excess O2, which renders NOx reduction into N2 impractical over conventional three-way catalysts (TWCs) [1]. Consequently, two main technologies have been developed for mobile lean NOx removal, i.e. selective catalytic reduction (SCR) of NOx using urea as a reductant for heavy duty diesel applications [2,3], and lean NOx trap (LNT) catalysts (otherwise known as NOx adsorber or NOx storage-reduction catalysts) for light duty applications [4]. NSR operation is cyclic: during the lean phase, NOx is trapped on the catalyst; intermittent rich excursions are used to reduce the NOx to N2 and restore the original catalyst surface, after which lean operation resumes [5]. Platinum and barium deposited on γ-Al2O3 (Pt-Ba/Al2O3 catalyst) is the most commonly studied model composition for NOx storage-reduction catalysis. Unfortunately, this catalytic system exhibits good catalytic performance at temperatures higher than 300°C; consequently, the low temperature activity of LNT catalysts becomes important, being for light duty diesel applications typical engine-out temperatures in the EOC (Efficiency Optimization Control) and FTP-75 (Regulatory Vehicle Mode) about 180-350°C [6]. Therefore, the low-temperature performance of LNT catalysts is becoming increasingly important in order to facilitate the implementation of highly fuel efficient lean-burn engines. A number of efforts have been made to improve the low temperature activity of LNT catalysts; between them, different dopant species have been considered in the catalyst formulation, for example Mn, Fe [7], or Ce [8]. Kim et al. studied the effect of Co and Rh promoter [9]. From these studies, it is evident that doping Pt-Ba/Al catalysts with redox active metal oxides can improve their NOx storage capacity at low temperature. In this work, we have considered the effect of Rh promoter in the classical Pt-Ba/alumina formulation focusing the attention of the fundamental aspects of the reactions occurring during the typical leanrich cycling of the catalyst. Pt-Rh-Ba/Al2O3 bimetallic catalyst, and the monometallic reference ones, has been prepared and deeply characterized by means of BET, XRD, XPS and TEM analysis. FT-IR has been used to characterized the nature of the stored NOx species, while TPD and TPIE for the thermal stability and the reactivity. Finally, isothermal lean-rich cycles using a complex reducing mixture (i.e. H2 + CO + C3H6) have been useful for the reactivity in the NOx removal and selectivity of this process. 2. Materials and Methods Model Pt-Rh-BaO/Al2O3 (1/0.5/20/100 w/w/w) bimetallic sample has been prepared by incipient wetness impregnation of a commercial -alumina (Versal 250 UOP) support with Pt(NO2)2(NH3)2 (Strem Chemicals, 5% w/w) aqueous solutions at first, and Rh(NO3)3 (Sigma Aldrich, 10% w/w) solution later on, followed by impregnation with Ba(CH3COO)2 (Sigma Aldrich, 99% w/w) aqueous solution. After each impregnation step the samples have been dried at 80°C overnight and calcined at 500°C for 5 h. Model Pt-BaO/Al2O3 (1/20/100 w/w/w) and Rh-BaO/Al2O3 (0.5/20/100 w/w/w) monometallic catalysts have been prepared in the same way. The obtained catalysts have been characterized by BET analysis for specific surface area and pore size distribution by N2 adsorption–desorption at 77 K (BET method, Micromeritics TriStar 3000 Instrument). The metal dispersion has been measured by H2 chemisorption at 40°C over a pre-reduced catalyst with a TPD/R/O ThermoFisher Instrument. X-ray powder diffraction (XRD) analysis has been performed with a Bruker D8 instrument using graphite monochromated CuKα radiation; the diffraction patterns were collected in the 2θ range of 10 - 70° with a step of 0.05° and a counting time of 12.5 s per step. High resolution transmission electron microscopy (HRTEM) was carried out using a JEOL 2010F electron microscope equipped with a field emission source at an accelerating voltage of 200 kV. Samples were deposited on holey Cu grids. The point-to-point resolution achieved was 0.19 nm and the resolution between lines was 0.14 nm. X-ray photoelectron spectroscopy (XPS) was performed on a SPECS system equipped with an Al anode XR50 source operating at 150 mW and a Phoibos MCD-9 detector. The pass energy of the hemispherical analyzer was set at 25 eV and the energy step was set at 0.1 eV. The pressure in the analysis chamber was kept below 10−7 Pa. The area analysed was about 2 mm × 2 mm. Data processing was performed with the CasaXPS program (Casa Software Ltd., UK). The interaction of NO/O2 with the catalytic surface has been investigated by IR spectroscopy and temperature-programmed desorption. NO/O2 mixture (1000 ppm NO + 3% O2 in He) has been fed to the IR reactor for the analysis of the surface species. It consists in a “Sandwich” IR cell containing the catalyst (15 mg) in form of selfsupported wafer; the gases leaving the cell were analyzed by mass spectrometer (ThermoStar TM GSD 301) and chemiluminescence analyzer (42i-HL MEGATEC). Total flow was 25 cm3min−1 (at 1 atm and 0°C). Surface FT-IR spectra have been collected with a FT-IR Nicolet Nexus spectrometer with 4 cm-1 spectral resolution and accumulation of 64 scans using DTGS detector. The thermal stability/reactivity of the stored species has been characterized by Temperature Programmed Desorption (TPD, up to 400°C @ 10°C/min in He) and by isotopic labelling experiments with 15NO (Temperature Programmed Isotopic Exchange, TPIE). Further details on the experimental procedure and setup can be found elsewhere [10]. The reactivity of the prepared catalysts in the NOx removal has been studied by means of lean-rich cycles at different temperature in the 150-350°C range. During the 15 min of the lean phase NO / O2 mixture (1000 ppm NO+ 3% v/v O2) has been fed to the reactor in a stepwise way; in the subsequent 15 min rich phase the reductant mixture, constituted by H2 (500 ppm) + CO (1500 ppm) + C3H6 (450 ppm), has been fed. During all the experiments CO2 (1000 ppm) and H2O (2.5% v/v) are present in the flow. Prior to any catalytic run the catalysts have been conditioned through NOx lean-rich cycles (typically 3-4 cycles) at 350°C starting from NO/O2 and H2 mixtures. Notably, due to the low reductant concentration, experiments have been carried out under nearly isothermal conditions, i.e. in the absence of significant temperature effects during the experiments. 3. Results and discussion 3.1 Morphological and structural characterization BET analysis – The composition of the prepared catalysts, the specific surface area, pore volume and pore sizes are reported in Table 1, along with the metal dispersion. Pt-Ba/Al2O3 catalyst shows a specific surface area of 140 m2 g-1 with a final metal dispersion near 60 %; Rh-Ba/Al2O3 exhibits a specific surface area of 150 m2g-1 with a Rh dispersion near 7 %. The bimetallic system shows the same morphological characteristics as the monometallic ones; it was not possible to measure the metal dispersion. XRD analysis – In Figure 1 are reported the XRD spectra of the calcined samples. From the XRD patterns it is possible to recognize the characteristic peaks of microcrystalline face-centered cubic - Al2O3 (JCPDS 10-425), BaCO3 both monoclinic and whiterite (JCPDS 78-2057 and 5-378, respectively) in all the catalysts. In the samples containing Pt, also the face-centered cubic metallic Pt (JCPDS 4-802) phase was recognized while in the Rh-containing samples, the Rh phase is not identified. It is worth t note that when the Pt is present, Ba is present mainly in the whiterite form, while in the Rh-Ba catalyst also the monoclinic form is detected. HRTEM analysis – Figures 2A-C show nice images recorded by STEM-HAADF on Pt-Rh-Ba/Al2O3 catalyst. This sample exhibits an excellent dispersion of noble metal nanoparticles. The bright spots correspond to the noble metal nanoparticles, which are not agglomerated; they are finely dispersed over the alumina support. The mean particle size of the noble metal nanoparticles is about 2-3 nm. A HRTEM image is shown in Figure 2D. Again in HRTEM mode is hard to distinguish the noble metal nanoparticles, although in some cases it is possible to infer their presence. Given the small size of the noble metal nanoparticles the EDX spectra does not allow to distinguish between Pt and Rh. Also, it is not possible to get lattice fringe images of the noble metal nanoparticles so no individual characterization is possible. In this regard, it is not possible to discuss if the nanoparticles are monometallic or bimetallic. High resolution TEM images of Pt-Ba/Al2O3 catalyst are reported in Figure 3. Figure 3A and its enlargement in Figure 3B show numerous Pt nanoparticles of similar size and extremely welldispersed. The mean particle size of Pt is about 1-2 nm, indicating a certainly remarkable dispersion of the Pt nanoparticles. In Figure 3C, the EDX spectra recorded in the areas labeled “a” and “b” are shown. Note that Cu signal comes from the TEM grid. The EDX spectrum of area “a does not contain any bright spot corresponding to Pt and shows the presence of Al and O, the last coming from the alumina support as well as Ba; this demonstrates that Ba is distributed all over the alumina support. The EDX spectrum of area “b”, in which it is possible to recognize several bright spot corresponding to Pt nanoparticles, shows in addition to Al, O and Ba, signals of Pt as expected. Finally, Figures 3D and 3E show HRTEM images, where the morphology and dimensions of the alumina flakes are seen. Also in this case, no Ba segregated is observed in any part of the sample. STEM-HAADF images of Rh-Ba/Al2O3 catalyst are displayed in Figures 4A-E. Both individual Rh nanoparticles as well as Rh aggregated nanoparticles are observed. Moreover, comparing with the TEM images of the binary Rh/Al2O3 catalyst (not shown), the presence of Ba does not seem to improve dispersion of the noble metal nanoparticles. Also, barium is not identified occurring as a separate phase and it is likely dispersed in all the sample. Figure 4F shows a HRTEM image where only the alumina support crystallites are clearly identified. XPS analysis – The results of the XPS analysis are reported in Table 2 in terms of surface atomic ratio. First of all, the atomic concentrations of the noble metals and Ba at the surface of the samples determined by XPS are always lower than those of the bulk. The dispersion of Ba is virtually identical in all samples containing this element (2.3-2.4 %). The dispersion of Rh is also very similar between the different samples containing this element. This could be apparently in contradiction to the TEM results, where the Rh/Al2O3 and Rh-Ba/Al2O3 samples, for instance, contained agglomerated Rh nanoparticles. However, given their small dimension, the area of the Rh nanoparticles analyzed by XPS includes all the Rh. In other words, it is not possible to distinguish between isolated Rh nanoparticles and agglomerated Rh since everything is indeed analyzed. 3.2 Activity tests 3.2.1 Characterization of the adsorbed species Effect of temperature – The interaction of NO/O2 mixture with the surface of Pt-Rh-Ba/Al2O3 catalyst has been performed after completely removal of Ba carbonates present over the freshly calcined catalysts and due the weakly adsorbed atmospheric CO2. After this, the NO/O2 mixture has been admitted to the IR reactor at different temperatures and the FT-IR spectra are recorded during this phase. Figure 5 shows the results as difference spectra at 150°C, 250°C and 350°C (Figure 5A, B, C, respectively). The FT-IR spectra recorded during the NOx storage at 150°C (Figure 5A) show the formation of chelating nitrites on the barium phasewith characteristic bands at 1356 and 1233 cm-1 related to νsym.(NO2) and νasym.(NO2) modes, respectively. The concentration of such species increases monotonically during storage. In addition, very minor amounts of bidentate nitrates (1565 cm−1, ν(N=O) mode) are formed after long exposure times, by very slow nitrite oxidation, [11;12;13]. By increasing the temperature up to 250°C (Figure 5B), the FT-IR spectra show that at low exposure times mainly nitrites (main band at 1220 cm−1) are formed at the catalyst surface. Nitrites continue to increase with time on stream and, in parallel, both ionic nitrates (νasym.(NO3) mode split at 1403 and 1350 cm−1 and νsym. (NO3) mode at 1040 cm−1) and bidentate nitrates at 1546 cm−1 start to form. After long exposures the band at 1220 cm−1 related to nitrites goes through a maximum and start to decrease in intensity due to the oxidation of nitrites into nitrates. On the other hands, the bands characteristic of ionic nitrates increase markedly. The parallel decrease of the band at 1546 cm−1 suggests the transformation of the bidentate nitrates into the ionic ones. The presence of two isosbestic points in the FTIR spectra confirms that both nitrites bidentate nitrates are transformed into ionic nitrates. Finally, the small band near 1755 cm-1 might be associated to the formation of N2O4 species [14]. Figure 5C shows the FT-IR spectra recorded as a function of time on stream during the NOx storage at 350°C. In this case, both nitrites and nitrates are almost simultaneously formed. Upon increasing the exposure time, the nitrite species decrease and eventually disappear while the nitrate bands continue to increase. Also in this case the band of bidentate nitrates shows a maximum while those of ionic nitrates grows monotonically with time. At saturation, only nitrates are present on the surface along with minor amounts of N2O4 species. The storage temperature has also a clear effect on the oxidation capacity of the catalyst (i.e. on the amount of NO2 produced at steady state) and on the NOx storage capacity. Indeed, at 150°C negligible amount of NO2 are observed at the end of the lean phase, being near 1.53 10-4 mol gcat-1 the amounts of NOx stored mainly as nitrites (see Figure 5A). By increasing the temperature, the concentration of NO2 at steady state increases as well, reaching near 200 ppm at 350°C; at the same way increases the storage capacity, resulting 3.89 10-4 mol gcat-1 at 350°C. Effect of noble metal – The same characterization of the adsorbed species in the 150-350°C temperature range have been carried out over monometallic Rh-Ba/Al2O3 catalyst and the results are reported in Figure 6A, B, C. The results of NOx adsorption for the model Pt-Ba/Al2O3 catalyst are not here reported since they have been already discussed in previous papers of some of us [Lietti et al. ChemCatChem 4 (2012), 55–58; S. Morandi et al. Catalysis Today 231 (2014) 116–124; Castoldi et al. Applied Catalysis B: Environmental 224 (2018) 249–263]. The surface FT-IR analysis performed at 150 °C over Rh-Ba/Al2O3 (Figure 6A) shows that, as in the case of Pt-Rh-Ba/Al2O3, NOx adsorption results mainly in the formation of nitrites (bands at 1360 and 1234 cm−1) that grow during storage. In addition, bidentate nitrates (1559 cm−1) are seen after long exposure times. However, at variance to what observed over Pt-Rh-Ba/Al2O3, also ionic nitrates (1394 and 1368 cm−1) and monodentate nitrates [15] (band at 1442 cm−1) are likely formed at this temperature. Surface FT-IR analysis recorded at 250°C (Figure 6B) shows the formation of both nitrites and nitrates (i.e. ionic and bidentate nitrates). Of note, differently from the case of Pt-Rh- Ba/Al2O3 both nitrites and nitrates increase monotonically with time on stream. As a matter of fact, no transformation of nitrites and bidentate nitrates into ionic nitrates can be observed. Besides, at 350°C (Figure 6C) the behaviour of Rh-Ba/Al2O3 is similar to Pt-Rh-Ba/Al2O3: nitrite species are observed only at a very short exposure time which rapidly evolve to ionic and bidentate nitrates. The transformation of bidentate nitrates into the ionic ones is also apparent. The FTIR experiments performed at low temperature (e.g. 150◦C) over the model Pt-Ba/Al2O3 showed the formation of surface nitrites only. Besides, at higher temperatures (e.g. 250°C) the formation of nitrates (i.e. both ionic and bidentate) was observed, along with nitrites. In particular nitrites were the prevalent species at the initial NOx uptake and their surface concentration showed a maximum with time. Instead, nitrates bands were found to grow monotonically with time and represented the most abundant adsorbed species after prolonged contact. Finally, when the NOx As already discussed, the NOx release is a chemical reaction occurring at the Pt/Ba interface and driven by metallic Pt/Rh centers. Therefore, it is expected that the release of stored NOx is favored in the presence of a reductant, due to the formation of reactive metallic centers. The reactivity of stored nitrates with different reducing agents (H2, NH3) has been previously studied by TPSR experiments[21,22,23]. The study has been enriched here by the reduction of nitrites. TPSR of nitrates - The reactivity of nitrate ad-species stored over bimetallic Pt-Rh-Ba/Al2O3 catalysts has been investigated under temperature programming using H2 or NH3 as reductant (H2-TPSR or NH3-TPSR); the results of the gas phase analysis are reported in Figure 9A and 9B, respectively [24]. When H2 is employed as reducing agent (Figure 9A), the temperature onset for its consumption is observed near 150°C. At 200°C H2 is totally consumed and both NH3 and N2 are observed as reduction products; neither NO nor N2O were detected during the reduction. The temperature onset for NH3 formation is slightly lower than that for N2, which matches with the temperature of total hydrogen consumption. Initially, H2 is consumed in the reduction of stored nitrates and only ammonia is formed (reaction (5)). Then, as the reduction front moves along the reactor axis, H2 is completely consumed and so-formed ammonia becomes reactive in the reduction of stored species (reaction (6)). Ba(NO3)2 + 8 H2  2 NH3 + BaO + 5 H2O (5) 3 Ba(NO3)2 + 10 NH3  8 N2 + 3 BaO + 15 H2O (6) The ammonia profile, indeed, shows a minimum in correspondence to the maximum of nitrogen production. Finally, H2 concentration raises due to the depletion of stored nitrates, nitrogen concentration quickly decays to zero while the ammonia decrease is much slower. At 500°C the hydrogen concentration is very close to the inlet value (2000 ppm). The data in Figure 9A show that the reduction of stored NOx is active at temperatures well below that of NOx thermal decomposition and accordingly the NOx reduction is a metal-catalyzed process that does not require the thermal release of stored NOx as the preliminary step [NS LAVORI]. For comparative purposes, Figure 9B-C shows also the results obtained in the case of the H2-TPSR of nitrates over the Pt- and Rh-monometallic samples, respectively [18,24,25,]. The reduction profile in the case of Pt-monometallic catalyst (Figure 9B) is very similar to that of bimetallic one, showing i) the consumption of H2 and formation of ammonia at first; ii) the complete H2 consumption in correspondence to the consumption of NH3 and the maximum N2 production; iii) the increase in H2 concertation and the decrease in the reduction products. It is notice that the onset of NH3 and N2 formation are closer and the H2 consumption sharper than over the bimetallic system. Over Rhmonometallic system (Figure 9C), only NH3 is observed between the products with a higher onset with respect to Pt-containing catalysts (200°C vs 170°C). Note that starting from 400°C a small production of N2 is observed, probably due to the decomposition of ammonia formed by reaction (6). The reactivity of stored nitrates has been studied also using ammonia as reducing agent and the results are reported in Figure 9D-F. Note that in this case isotopically labeled nitrates are present at catalyst surface, being obtained by 15NO/O2 adsorption. Over all the catalytic systems the reduction of nitrates by ammonia is slower since the onset of this reaction is observed at higher temperature than the one with hydrogen, and is fully selective to nitrogen, being this the only detected species. As already observed, the Rh-monometallic system results the less active showing the lower amount of reduction products. As already reported [26], the dynamics of the evolution of the three dinitrogen isotopes is the same; the unlabeled 14N2 and the single-labeled 15N14N species are most abundant, while 15N2 accounts for 19% of total N2 in the Pt-Ba/Al2O3 catalyst and lower amount in the other cases. Finally, over the Rh-containing catalysts, both mono- and bi-metallic catalysts, H2 is observed between the products starting from 300°C together with an increase in the unlabeled 14N2. Since the stored NOx are isotopically labeled (i.e. 15NOx), this species originates only from 14NH3; hence, at this temperature the decomposition of ammonia occurs according to reaction (7): 2 14NH3  14N2 + 3 H2 (7) TPSR of nitrites - The reduction of stored nitrites has been performed both by H2-TPSR or NH3- TPSR over all the catalytic systems. In general, with H2 (Figure 10A-C) occurs at even lower temperatures than nitrates; indeed, in these case H2 is consumed below 100°C with formation of NH3 and traces of N2. Moreover, Pt-containing catalysts are more reactive that Rh-monometallic one, both as onset reduction temperature and reduction extent (i.e. low H2 consumption and NH3 formation). The consumption of H2 and the NH3/N2 production follow the stoichiometry of the global reactions (8) and (9): Ba(NO2)2 + 6 H2  2 NH3 + BaO + 3 H2O (8) Ba(NO2)2 + 3 H2  N2 + BaO + 3 H2O (9) Also in the case of nitrites, when H2 is used as reductant no NO or N2O are observed among the products. In the case of the NH3-TPSR experiment (Figure 11D-F), the desorption of ammonia is observed at first; indeed, ammonia is being stored on the catalyst upon admission at the beginning of the experiment (not shown in the Figures). From nearly 120°C when Pt is present (and slightly higher temperature in the Rh-monometallic catalyst), a net consumption of ammonia is seen together with the formation of N2O and N2 due to the reaction of NH3 with stored labeled nitrites [27,28]. Nitrogen is by far the most abundant product, with different isotopic composition. The single-labeled isotope (i.e., 15N14N) is initially observed in greater amounts, whereas the unlabeled (14N2) and double-labeled (15N2) species are seen simultaneously and with a short delay respect to the single-labeled nitrogen reported [26]. The evolution of very small amounts of single-labeled nitrous oxide (15N14NO), whose concentration is multiplied by a factor of 10, is also observed at the onset of the reaction, near 100°C in the Pt-containing catalysts and near 180°C on Rh-monometallic sample. The consumption of NH3 and the N2 (and N2O) production agree with the stoichiometry of the global reaction (10): Ba(NO2)2 + 2 NH3  2 N2 + BaO + 3 H2O (10) These data clearly indicate that ammonia is an effective reductant for the stored NOx, even the reduction by ammonia is slower than with H2 being the onset of this reaction observed at higher temperature. As in the case of nitrates, at high temperature over Rh-containing catalysts ammonia is decomposed to N2 and H2; accordingly, the concentration of unlabeled 14N2 increases. The order observed for the onset temperature of reduction of stored NOx (i.e. T(H 2) < T(NH3)) compares fairly well with the efficiency of the reductants in scavenging the O-adatoms from Pt(Rh)- O and then in the activation of stored NOx by metal reduced centers, that seems to represent the rate determining step of the reduction of stored NOx [10]. In the presence of efficient reductants like H2 (and NH3), it is likely that NO formed upon the activation of stored NOx does not desorb from the metal sites, but is readily decomposed to N- and O-adatoms over them. Once formed, O-adatoms are scavenged by H-adatoms formed upon decomposition of H2 and of NH3 with formation of H2O whereas N-adatoms combine to form N2 (or are hydrogenated to give NH3 in the case of H2), which are indeed observed among the reaction products. Nevertheless, it is not possible to exclude that NO might be released in the gas phase and subsequently re-adsorbed and decomposed at metal sites to give N2 and NH3. When the reactivity of the reductant is lower, like in the case of NH3, oxidized metal sites are less efficiently reduced to metal form. Accordingly, released NO is not readily decomposed into N- and O-adatoms over partially O-covered metal site, so the detection of traces of N2O is also possible. Upon increasing the temperature, the evolution of N2O is no more observed, because ammonia efficiently scavenge O-adatoms from metal, and this favors NO decomposition to N- and O-adatoms and subsequent coupling of N-adatoms to give N2. Note that N2O is observed in greater amounts only in the presence of Pt, while over Rh-monometallic catalysts its concentration is lower. This could be related to the poor reactivity of Rh with respect to Pt at low temperature. Moreover, Rh becomes active at higher temperature where ammonia is already efficient in the O-adatoms removal. 3.2.4 Reactivity of the stored species under realistic conditions In a previous work by some of us [29], Pt- and Rh-based catalysts have been tested under lean/rich conditions at different temperatures. It has been found that the Rh-based samples exhibit a superior ability to release O2 from the surface at lower temperatures with respect to Pt, suggesting the presence of a promoting effect on the spillover process of NOx to the precious metal, controlling the subsequent release and reduction of NOx. Aiming to a better understanding of the different catalytic behavior of Rh- and Pt-based catalysts, we have here analyzed the behavior of the bimetallic Pt-Rh catalyst and monometallic references in the lean-rich cycles using a complex reducing mixture, maintaining a continuous flow of CO2 and water. Reactivity of NOx Species Stored at 250°C - Figure 11A-C show the gas-phase results obtained during the lean-rich cycles performed at 250°C over Pt-Rh-Ba, Pt-Ba and Rh-Ba catalysts, respectively. Note that in the Figure is reported a representative cycle, i.e. the catalysts are fully conditioned at this temperature. In the case of the Pt-Rh-Ba sample (Figure 11A), the breakthrough of NO was detected after 30 s. Then, the nitrogen oxide concentration increases with time, until a steady state is reached (near 800 ppm). The NOx storage is accompanied by evolution of NO2 (near 60 ppm) due to oxidation of NO over Pt (Rh) sites. After 15 min, at the end of the lean phase 2.66 10-4 mol/gcat result stored onto the catalyst surface. Then, the NO concentration is decreased to zero in a stepwise manner and the reducing mixture (H2 + CO + C 3H6) is fed. At the lean-to-rich switch a tailing in the NOx concentration is observed due to the desorption of weakly NOx adsorbed (near 2.68 10-5 mol/gcat); an instantaneous production of N2 and of N2O is observed, along with NH3 which is however detected with a delay. At the rich-to lean transition a new N2O peak is observed. Similar behavior is observed with the Pt- and Rh-monometallic catalysts (Figure 11B and 11C, respectively). Over Pt-Ba/Al2O3 the NOx breakthrough is observed after 60 s; after 15 min of lean phase NO reaches near 700 ppm while NO2 accounts for 160 ppm, i.e. the oxidation capacity is higher than the bimetallic catalyst. In the case of Rh-Ba/Al2O3 catalyst, upon NO addition (Figure 11C) its concentration increases monotonically with time showing an almost nil dead time, while NO2 is not observed being the NO/NO2 oxidation not effective at this temperature. The total amount of stored NOx result near 3.44 10-4 mol/gcat over Pt-monometallic catalyst and near 1.7 10-4 mol/gcat over Rhmonometallic one. Both the rich phases show the same dynamics described for the bimetallic system, being the lowest N2O production observed over Rh-monometallic catalyst. Only over Ptmonometallic sample, N2O is observed also at the rich-to lean transition. It is worth to note that over all the systems, H2 and CO concentration are observed to increases later than propylene, when the concentration of nitrogen decreases. It would be notice that a complex set of reactions, involving the reductant species, takes place during the reduction: i) steam reforming reaction of C3H6 with water (reaction (11)): C3H6 + 3 H2O  3 CO + 6 H2 (11) ii) water gas shift reaction (12): CO + H2  CO2 + H2O (12) iii) reverse water gas shift reaction (13): CO2 + H2O  CO + H2 (13) The consumption of propylene could be related to the reduction of stored NOx, but also to the occurrence of the steam reforming reaction (11). However, the H2 concentration at steady state remain lower than the inlet value (i.e. 500 ppm) suggesting that or reaction (11) is not active or it is balanced by the occurrence of reaction (12) and/or reaction (13). Reactivity of NOx Species Stored at 350°C – At higher temperature, i.e. 350°C, the bimetallic catalyst shows a dead time in NO breakthrough of 110 s; the dead time is longer for Pt-catalyst and shorter for Rh-catalyst (near 170 s and 50 s, respectively). Also the NO/NO2 oxidation follows the same trend, being higher for Pt-Ba/Al2O3, lower for Rh-Ba/Al2O3 and intermediate for the bimetallic system. At the end of the lean phase, the bimetallic Pt-Rh catalyst exhibits the best storage capacity (6.04 10-4 mol/gcat); lower mounts are calculated for the monometallic systems, being higher in the case of Pt (4.59 10-4 mol/gcat over Pt vs 2.74 10-4 mol/gcat over Rh). Then, upon the rich switch, the reductant mixture is admitted to the reactor. The NOx concentration rapidly decreases and an instantaneous production of N2 and of N2O is observed, along with NH3 which is however detected with a delay. The reduction of the stored NOx is more selective to N2 at this temperature since minor amounts of NH3 and of N2O are detected. Finally, at this temperature the concentration of hydrogen at steady state results higher than the inlet value (i.e. 500 ppm) suggesting that the steam reforming and the reverse water gas shift reactions (11- 13) occur to a higher extent than reaction (12). Moreover, the low CO concentration suggests that the water gas shift reaction (12) also occurs to some extent. As already discussed and demonstrated in previous work [30], the reduction of the stored NOx is based on a metal-catalyzed process that does not involve the thermal release of the stored NOx as the preliminary step. Instead, the process is initiated by the reduction of the noble metal by the reductant; this leads to the release of NO from the NOx species stored nearby the Me sites (Pt, Rh), and to the migration toward Pt of the species adsorbed far-away from the noble metal. This process is followed by decomposition of the released NO into N- and O-adatoms over the metal sites (reactions 14 and 15). NO dissociation as the initial step for NOx reduction has been pointed out by transient TAP reactor experiments31, and by other evidences as well [32]. NO + Me  Me-NO (14) Me-NO + Me  Me-N + Me-O (15) N-containing products are then formed upon reaction of adsorbed N-adatoms with undissociated NO, with other N-adatoms or with H-adatoms, leading to N2O (reaction 16), N2 (reaction 17) and NH3 (reaction 18), respectively, while the role of the reducing agent is to remove the O-adspecies from the metal that is hence kept in a reduced state. Me-NO + Me-N  N2O + 2 Me (16) 2 Me-N  2 Me + N2 (17) Me-N + 3 Me-H  4 Me + NH3 (18) Me-O species formed upon dissociation of NO (reaction 15) are scavenged by the reductant restoring Me sites, while Me-H species are formed by the H-activation on Me sites (H-atom deriving from both H2 and C3H6). At the active-site level, the selectivity of the reduction process depends on the operating conditions (temperature, gas-phase NO and H2 concentration), i.e. on the degree of reduction of the noble metal and on the local concentration of N-, H-adatoms: at low temperatures and at the beginning of the rich phase the noble metal is not fully reduced and this favors N2O formation, whereas high temperatures and/or high H-atoms concentration favor NH3 formation. Accordingly, high selectivity to N2 formation is observed when a high concentration of N-adspecies is attained. Inspection of the literature [33,34,35,36,37,38,39] reveals that during lean-rich operations over fully formulated LNT catalysts N2O is formed. Indeed, N2O evolution is observed at both lean-to-rich and rich-to-lean transitions (primary and secondary N2O emissions). It has been suggested that primary N2O formation (lean-to-rich transition) is formed at the regeneration front, when the reductant reaches oxidized/not fully reduced Platinum-Group-Metal (PGM) sites, which are in a close proximity to NOx ad-species. The secondary N2O formation (rich-to-lean transition) originates from reaction between residual surface NOx with reductive species (like NCO, CO or NH3) in an adsorbed state. In the lean-rich cycles of Figure 11A-C and 12A-C, both primary and secondary N2O peaks are detected with different concentration depending mainly on temperature and on the noble metal present in the catalysts. Primary N2O is occurring during lean/rich cycles according to the lines previously depicted (reactions (14) – (16)). At high temperature primary N2O formation is reduced because the metal sites are easier reduced by the reducing agent, and this enhances the NO dissociation thus favoring the coupling of N-adatoms forming N2 (reaction (17)) and/or the hydrogenation to NH3 (reaction (18)). Even if the lower N2O production is observed for Rh-Ba/Al2O3 catalysts, it would seem that this is more the effect of the lower storage capacity (that means low reduction products in general) than of the specific metal activity. At variance, secondary N2O may originate from reaction of NO/O2 with minor quantities of reductive species (like NCO, CO or NH3) left adsorbed on the surface during the rich phase. This leads to the N2O evolution observed during the rich-to-lean switch. As in the case of primary N2O, also secondary N2O formation is decreased as increasing the temperature being the metal sites more easily reduced. Moreover, in this case over Rh-containing catalysts, both bi- and mono-metallic, the secondary N2O results lower than over Pt-one at each temperature. This could be related to the different poisoning effect that CO have over Rh and Pt metal sites. References 1 S.i. Matsumoto, Catalytic Reduction of Nitrogen Oxides in Automotive Exhaust Containing Excess Oxygen by NOx Storage-Reduction Catalyst, CATTECH, 4 (2000) 102-109 2 M. Koebel, M. Elsener, M. Kleemann, Urea-SCR: a promising technique to reduce NOx emissions from automotive diesel engines, Catalysis Today, 59 (2000) 335-345. 3 S. Roy, A. Baiker, NOx Storage−Reduction Catalysis: From Mechanism and Materials Properties to Storage−Reduction Performance, Chemical Reviews, 109 (2009) 4054-4091 4 N. Takahashi, H. Shinjoh, T. Iijima, T. Suzuki, K. Yamazaki, K. Yokota, H. Suzuki, N. Miyoshi, S.-i. Matsumoto, T. Tanizawa, T. Tanaka, S.-s. Tateishi, K. Kasahara, The new concept 3-way   catalyst for automotive lean-burn engine: NOx storage and reduction catalyst, Catalysis Today, 27 (1996) 63-69 5 LNT BOOK 6 T.W. Chan, E. Meloche, J. Kubsh, R. Brezny, Black Carbon Emissions in Gasoline Exhaust and a Reduction Alternative with a Gasoline Particulate Filter, Environmental Science & Technology, 48 (2014) 6027-6034 7 N. Le Phuc, X. Courtois, F. Can, S. Royer, P. Marecot, D. Duprez, NOx removal efficiency and ammonia selectivity during the NOx storage-reduction process over Pt/BaO(Fe, Mn, Ce)/Al2O3 model catalysts. Part I: Influence of Feand Mn addition, Applied Catalysis B-Environmental, 102 (2011) 353-361 8 C. Shi, Y. Ji, U.M. Graham, G. Jacobs, M. Crocker, Z. Zhang, Y. Wang, T.J. Toops, NOx storage and reduction properties of model ceria-based lean NOx trap catalysts, Applied Catalysis B- Environmental, 119-120 (2012) 183-196 9 J.-G. Kim, H.-M. Lee, M.-J. Lee, J.-H. Lee, J.-G. Kim, J.-Y. Jeon, S.-K. Jeong, S.-J. Yoo, S.-S. Kim, Effect of Co and Rh promoter on NOx storage and reduction over Pt/BaO/Al2O3 catalyst, Journal of Industrial and Engineering Chemistry, 14 (2008) 841-846. 10 L. Castoldi, L. Righini, R. Matarrese, L. Lietti, P. Forzatti, J. Catal. 328 (2015) 270 11 Lietti et al. ChemCatChem 4 (2012) 55–58 12 Morandi et al. Catalysis Today 231 (2014) 116–124 13 Castoldi et al. Applied Catalysis B 224 (2018) 249–263 14 Yi C-W, Kwak JH, Szanyi J (2007) J Phys Chem C 111:15299–15305 15 Daturi et al, Oil & Gas Science and Technology – Rev. IFP Energies nouvelles, Vol. 66 (2011), No. 5, pp. 845-853 16 39 in ANTONIA PAPER 17 40 in ANTONIA PAPER 18 ANTONIA PAPER Figure 2 Figure 3 Figure 4 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 2000 1800 1600 1400 1200 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Absorbance [a.u.] 1233 1356 1556 A Absorbance [a.u.] 1755 1546 1403 1350 1220 1040 B Wavenumbers [cm -1 ] Absorbance [a.u.] 1752 1551 1412 1320 1032 1207 C Figure 5 0.00 0.05 0.10 0.15 0.20 0.25 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 2000 1800 1600 1400 1200 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 Absorbance [a.u.] 1559 1442 1394 1368 1234 A B Absorbance [a.u.] 1758 1551 1391 1359 1231 1044 Absorbance [a.u.] Wavenumbers [cm -1 ] 1754 1552 1400 1340 1221 1038 C Figure 6 100 200 300 400 500 600 700 800 0 100 200 300 0 100 200 300 0 100 200 300 100 200 300 400 500 600 700 800 100 200 300 400 500 600 700 800 Temperature [°C] Temperature [°C] Concentration [ppm] Concentration [ppm] Concentration [ppm] Temperature [°C] C O2 NO NO 2 hold B O2 NO NO 2 A O2 NO NO 2 O2 NO NO2 D E O2 NO2 NO F NO2 O2 NO hold hold NO NO 2 O 2 G H NO 2 NO O 2 L O2 NO2 NO Figure 7 0 200 400 600 800 1000 50 100 150 200 250 300 350 400 450 500 550 600 0 200 400 600 800 1000 0 200 400 600 800 1000 50 100 150 200 250 300 350 400 450 500 550 600 B 14 NO 2 NO x 15 NO 14 NO hold Concentration [ppm] Concentration [ppm] Concentration [ppm] Temperature [°C] NO x 14 NO 2 14 NO 15 NO Temperature [°C] C hold 14 NO 2 A NO x 15 NO 14 NO 14 NO 2 NOx 15 NO 14 NO E F 14 NO 2 NOx 15 NO 14 NO D 14 NO 2 NOx 15 NO 14 NO Figure 8 0 200 400 600 800 1000 1200 100 200 300 400 500 600 0 200 400 600 800 1000 1200 0 200 400 600 800 1000 1200 0 200 400 600 800 1000 1200 1400 1600 100 200 300 400 500 600 0 200 400 600 800 1000 1200 1400 1600 0 200 400 600 800 1000 1200 1400 1600 A N 2 NH 3 H 2 /2 Concentration [ppm] Concentration [ppm] Concentration [ppm] C NH3 H2/2 N2 B N2 NH3 H2/2 15 N 14 N E H 2 NH 3 15 N 2 14 N 2 Temperature [°C] Temperature [°C] F H 2 NH 3 14 N 2 15 N 14 N 15 N 2 D 15 N 14 N 15 N 2 H 2 NH 3 14 N 2 Figure 9 0 200 400 600 800 1000 1200 1400 100 200 300 400 500 600 0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400 100 200 300 400 500 600 Temperature [°C] Concentration [ppm] Concentration [ppm] Temperature [°C] NH 3 N 2 H 2 /2 F BE AD Concentration [ppm] C NH 3 *2 H 2 /2 N 2 N 2 NH 3 H 2/2 15 N 14 NO*10 H 2 /2 NH 3 15 N 14 N 14 N 2 15 N 2 15 N 2 H 2 NH3 14 N 2 15 N 14 N 15 N 14 NO*10 NH 3 H 2 15 N 2 14 N 2 15 N 14 N 15 N 14 NO*10 Figure 10 66000 66500 67000 67500 0 200 400 600 800 1000 1200 1400 0 100 200 300 400 500 C 3 H 6 CO NH 3 Concentration [ppm] Time [s] N 2 NO 2 NO x N 2 O NO CO 2 /2 H 2 N 2 O concentration [ppm] A lean phase rich phase 4th cycle 48750 49000 49250 49500 49750 50000 50250 50500 0 200 400 600 800 1000 1200 1400 0 100 200 300 400 500 CO2/2 C3H6 CO NH 3 N 2 NO 2 NO x N 2 O NO CO2/2 H2 B rich phase lean phase 5th cycle Concentration [ppm] Time [s] N 2 O concentration [ppm] 64000 64500 65000 65500 0 200 400 600 800 1000 1200 1400 0 100 200 300 400 500 N 2 O concentration [ppm] CO 2 /2 C 3 H 6 CO/2 NH 3 N 2 NO x N 2 O NO CO 2 /2 H 2 C rich phase lean phase 4th cycle Concentration [ppm] Time [s] Figure 11