Nitrogen oxides (mainly NO and NO2, named as NOx) emitted during the combustion of fossil fuels cause serious environmental problems such as acid rain, photochemical smog, and ozone depletion [1, 2]. Low-temperature selective catalytic reduction of NO with NH3 (NH3-SCR) was advocated to be a preferential method for the abatement of NOx emitted from the flue gas in steel, cement, glass, and other non-power-plant industries [3-6]. However, the formation of undesired by-products of N2O during the low-temperature NH3-SCR process is a great constraint because of their greenhouse effect and potential for the destruction of the stratospheric ozone layer [7]. Therefore, detailed knowledge of the mechanism of N2O formation during the NO abatement by low-temperature NH3-SCR is indispensable for the rational design of effective catalysts with good N2 selectivity. Several reaction mechanisms for the formation of N2O in the low-temperature NH3-SCR process over metal-oxide-supported noble metal catalysts have been proposed. For example, Otto et al. [8] and Takoudis et al. [9] proposed that N2O produced over Pt/Al2O3 is a combination of two nitroxyl species (HNO); however, the reaction intermediate HNO has never been observed experimentally. Pérez-Ramírez et al. [10] suggest that the coupling of ammonia intermediates (NHx) and nitric oxide is the exclusive route for N2O formation. The coupling of two adsorbed NO molecules, the recombination of an adsorbed NO molecule with an adsorbed NHx fragment, and the coupling of un-dissociated NO molecules with N• adspecies were considered to be possible pathways for N2O formation over a Pt−Ba/Al2O3 LNT catalyst, based on isotopic labeling experiments [11, 12]. Due to the complexity of this reaction system, the reaction mechanisms of the whole system are complicated, and more fundamental studies are expected to provide further insights into the pathways involved in N2O formation.
Cerium oxide (CeO2) has been extensively investigated in the field of heterogeneous catalysis due to its unique redox properties and high oxygen storage capacity [13]. Although pure CeO2 generally exhibits poor NH3-SCR activity as an active component [14], the use of CeO2 as a support can significantly enhance the catalytic activity via the increased dispersion and stabilization of the active species as well as the activation of the reaction gas [15-19]. Earlier studies revealed that Pd metal is capable of activating NO and enables the reduction of NO with H2 [20], CO [21], and hydrocarbons [22]. However, research on the Pd-containing catalysts applied in the NO abatement by NH3 is limited.
In this study, Pd nanoparticles were selected as the active components loaded on a CeO2 support, and the activity for low-temperature NH3-SCR of NO was studied; special attention was paid to N2O formation and its related mechanisms. Our results indicated that the formation of N2O was temperature-dependent; below 250 ℃, N2O was produced through the decomposition of the HON intermediate, and above 250 ℃, N2O resulted from the reaction of N• and NO in the absence of O2.
CeO2 nanorods were prepared via a hydrothermal method, according to the literature [23]. Typically, 20 ml Ce(NO3)3 solution (0.4 mol/L) was dropped into 140 ml NaOH solution (9 mol/L) and stirred for 30 min at room temperature. Then, the mixed solution was transferred into a 200 ml Teflon-lined stainless steel autoclave and heated at 100 ℃ for 24 h. After hydrothermal treatment, the precipitates were collected by centrifugation, washed with deionized water, and dried at 80 ℃ overnight. Finally, the product was calcined at 600 ℃ for 2 h under flowing air.
A Pd/CeO2 catalyst was prepared using the solid grinding method [24]. Pd(acac)2 powder was mixed with the CeO2 nanorods and ground in a mortar pestle for 30 min. Thereafter, the sample was calcined at 280 ℃ for 4 h under flowing air and then at 400 ℃ for 2 h under high-purity nitrogen.
The amount of Pd loaded on the CeO2 nanorods was measured to be 3%, via inductively coupled plasma analysis. The morphology, structure, and composition of the catalyst were identified using transmission electron microscopy (TEM; FEI Titan 80-300KV FEG TEM), X-ray diffractometry (XRD; Philips XD-98 X-ray diffractometer with Cu Kα radiation, λ = 0.15406 nm), and X-ray photoelectron spectroscopy (XPS; Thermo Scientific ESCALAB 250Xi system with Al Kα radiation of 1486.6 eV, using the C 1s binding energy of 284.8 eV as a reference).
NH3-temperature-programmed desorption (NH3-TPD) tests were carried out in a Micromeritics Autochem Ⅱ 2920 apparatus with a thermal conductivity detector (TCD). A 100 mg sample was pretreated at 300 ℃ under helium atmosphere (40 ml/min) for 1 h, and subsequently cooled to 60 ℃ under the same airflow. Thereafter, the sample was exposed to 10% vol NH3/He flow (20 ml/min) for 2 h, and the physisorbed NH3 was purged with helium before the TPD test. TCD curves were recorded from 50 to 800 ℃ at a heating rate of 10 ℃/min.
An NO-TPD experiment was carried out in a fixed-bed flow reactor. Prior to the experiment, the sample was treated at 400 ℃ for 1 h and then cooled to 100 ℃. 800 ppm NO balanced with N2 at a flow rate of 500 ml/min was passed through the sample for 2 h. After it reached a saturation stage, the sample was flushed with pure N2 (500 ml/min) until the NOx signal stabilized. Finally, the sample was heated up to 450 ℃ (ramp of 2 ℃/min), and the concentrations of NO/NO2 and N2O desorbed from the catalyst surface were detected, with a temperature interval of 10 ℃, using the Testo AG testo 335 and Medi-Gas G200 instruments, respectively.
In-situ transmission infrared (IR) spectra were recorded on a CRCP-7070 micro transient reaction integrated platform, which was built by the Tianjin Xianquan Company. The platform was equipped with an in-situ IR reaction cell containing a gas flow system and a liquid-nitrogen-cooled MCT detector. The catalyst was pretreated at 300 ℃ under flowing N2 for 30 min before spectrum acquisition. The background spectrum recorded under flowing N2 was subtracted from the sample spectrum.
Catalytic activity measurements were conducted in a fixed-bed flow reactor. 1.0 g catalyst powder was pasted on seven 4 cm × 10 cm aluminum plates. The plates were placed in a reactive tank, at 5 mm intervals. The reactant gas was composed of 800 ppm NO and 800 ppm NH3 balanced with N2 at a total flow rate of 500 ml/min, corresponding to a gas hourly space velocity (GHSV) of 36000 h−1. In order to simplify the system, O2 was removed intentionally in this study, and the reaction between NO and NH3 could be studied without interference. The Testo AG testo 335 instrument was applied to monitor the inlet and outlet concentrations of NO/NO2, whereas the N2O concentration was monitored using the Medi-Gas G200 instrument. The conversion of NO was calculated using the following equation:
Where [NOx]in and [NOx]out represent the sums of the concentrations of NO and NO2 at the inlet and outlet, respectively.
The structure and morphology of the Pd/CeO2 catalyst was characterized using TEM. As shown in Fig. 1(a), the CeO2 support exhibited a nanorod morphology with an average diameter of about 20 nm. Fig. 1(b) is an enlarged image of the area marked by a white dashed rectangle in Fig. 1(a). The d-spacing of approximately 3.13 Å measured on the nanorod was indexed as CeO2 {111}, and the supported nanoparticles on the catalyst surface, with a d-spacing of approximately 1.92 Å , could be attributed to the {200} planes of Pd. It is notable that there is no obvious interface between the Pd nanoparticles and CeO2, which can result from the formation of a solid solution of PdxCe1-xO2 at the interface [25]. Moreover, many subnanometer Pd clusters (marked by white circles) are present in Fig. 1(a) and (b), indicating a good dispersion of Pd on CeO2. The size distribution of Pd is shown in Fig. 1(c); almost all the clusters have a diameter less than 2 nm. The XRD pattern of the Pd/CeO2 catalyst is displayed in Fig. 1(d), wherein all the diffraction peaks are consistent with the characteristic peaks of cubic phase CeO2 (JCPDF 34-0394) and no diffraction peak can be identified for Pd, in good agreement with the TEM observation.
The atomic concentrations and element chemical states of the catalyst surface were investigated using XPS, and the results are shown in Fig. 1(e) and (f). By performing peak-fitting deconvolutions, the Ce 3d spectrum can be divided into eight characteristic peaks: v (882.3 eV), v' (884.4 eV), v'' (888.5 eV), v''' (898.2 eV), u (900.8 eV), u' (903.0 eV), u'' (907.4 eV), and u''' (916.6 eV). The bands labeled as v' and u' were ascribed to Ce3+, and the other six bands were assigned to Ce4+ [26]. The relative surface content of Ce3+/(Ce3++Ce4+) was measured to be about 12.6%; Ce3+ defects and oxygen vacancies in the CeO2 nanorods are reported to play an important role in their catalytic performance [27]. The spectrum of Pd 3d was divided into four peaks, as shown in Fig. 1(e). The peaks at 335.5 and 340.7 eV can be assigned to metallic Pd0, and the other two peaks (337.5 and 342.6 eV) can be attributed to ionic Pd2+ located in the structure of the PdxCe1-xO2 solid solution at the surface and subsurface layers of the ceria lattice [25, 28, 29].
It is well-recognized that the adsorption of reaction gases on the catalyst surface always plays a crucial role in the catalyst performance. Therefore, NH3-TPD and NO-TPD were performed to evaluate their adsorption (Fig. 2). The NH3-TPD profile shows distinct desorption peaks located at 50–400 ℃, belonging to NH3 desorption from weak (60–180 ℃) and medium (180–400 ℃) acid sites distributed on the catalyst surface [30]. The small peak located at 488 ℃ belongs to the strong acid sites [31, 32]. According to the NO-TPD profile shown in Fig. 2(b), NO started to desorb at 120 ℃ and showed a maximum concentration at 275 ℃; even above 400 ℃, NO desorption was detectable, suggesting that NO strongly adsorbed on the catalyst during the reaction. Moreover, N2O was also found to be desorbed in the temperature range of 250–370 ℃, with a maximum concentration of 18 ppm at 310 ℃, which could be attributed to the dissociation of NO to N• and O• on the catalyst surface. The curve for N2O formation in Fig. 2(b) also indicated that the direct dissociation of NO only occurred at temperatures above 250 ℃. It is notable that the appearance of N2O is accompanied by a decrease in NO desorption, which further indicates that the detected N2O is attributable to the NO dissociation.
The in-situ transmission IR spectra for the reaction between NO and the adsorbed NH3 species over the Pd/CeO2 catalyst at 250 ℃ are depicted in Fig. 3(a). The catalyst was firstly treated with NH3 for 30 min and then purged with N2. The bands at 1124 and 1570 cm–1 originated from coordinated NH3 on Lewis acid sites, whereas the other two bands at 1466 and 1635 cm−1 were ascribed to δas(NH4+) and δs(NH4+) bound to Br nsted acid sites, respectively [30, 33]. Activated NH3 species (–NH2) were also indicated by the bands at 1315 and 1370 cm–1 [26]. Earlier studies suggested that both CeO2 and Pd were the active sites for NH3 adsorption and activation [34, 35]. When NO was introduced into the system, all the adsorbed NH3 species, including the coordinated NH3 on Lewis acid sites, NH4+, and –NH2, were rapidly consumed within 5 min, implying that the adsorbed NH3, NH4+, and –NH2 species on the surface of the Pd/CeO2 catalyst at 250 ℃ reacted with NO quickly. Then, adsorbed NOx species, including chelating nitrites (1210 cm–1), NO3– (1326 and 1405 cm–1)[12], and bidentate nitrate (1498 and 1530 cm–1) [36], rapidly formed on the surface of the catalyst. It is worth noting that in addition to the aforementioned NOx species formed with the aid of CeO2, NO adsorbed on Pd, including linear NO [Pd0-NO] (1751 cm–1) and bent NO [Pd-NO–] (1664 cm–1), was also observed [21].
To investigate the reactivity of adsorbed NOx species with NH3, an in-situ IR spectrum was also recorded as a function of time over Pd/CeO2 catalyst at 250 ℃, as shown in Fig. 3(b). After the introduction of NO for 30 min and purging with N2, bands attributable to chelating nitrites (1210 cm–1), NO3– (1326 and 1405 cm–1) [12], bidentate nitrate (1498 and 1530 cm–1)[36], and Pd-NO– (1664 cm–1) [21] appeared. The band at 1751 cm–1 for Pd0-NO was not detected; this may have been because of the oxidation of metallic Pd upon NO introduction. The follow-up introduction of NH3 resulted in the disappearance of adsorbed chelating nitrites, bidentate nitrate, and Pd-NO–, which was attributed to their reaction with NH3 at 250 ℃; however, there was hardly a decrease in the intensity of the peaks attributed to NO3– (1326 and 1405 cm–1), suggesting that its reaction with NH3 was slow at 250 ℃. Furthermore, new bands at 1556, 1635, and 1370 cm–1, corresponding to coordinated NH3 on Lewis acid sites, NH4+, and –NH2 species over Pd[26, 30, 33], respectively, were detected. Based on Fig. 3(b), it can be concluded that nitrites, nitrate species, as well as adsorbed NO on Pd participate in the reaction with adsorbed NH3 on the surface of the Pd/CeO2 catalyst at 250 ℃.
The catalyst was further investigated under a flow of NO + NH3 with the temperature rising from 100 to 300 ℃, to gain an insight into the possible reaction routes versus temperature. As presented in Fig. 3(c), both NO and NH3 adsorbed on the catalyst over the entire test temperature range. At 100 ℃ adsorbed NOx species, including NO3– (1326 and 1405 cm–1), bidentate nitrate (1235 and 1550 cm–1) [12, 36], NH3 species bound to Br nsted acid sites, with peaks at 1466 and 1635 cm–1 [30, 33], and intermediates (NxHyNO3) with an extremely weak peak at 1357 cm–1 [37], were simultaneously detectable. This indicated that the reaction between NH3 and NO at 100 ℃ was slow, and only the reactions between –NH2 + NH3-L and nitrides + Pd-NO– occurred, accompanied by the formation of N2O and N2. The other reactants still existed as adsorption species.
With an increase in temperature to 150 ℃, the intensity of adsorbed bidentate nitrate at 1550 cm–1 decreased obviously, indicating that the reaction between bidentate nitrate and –NH2 + NH3-L species was promoted. At 175 ℃, the intensity of adsorbed bidentate nitrate (1550 cm–1) further decreased and the peaks of bidentate nitrate at 1235 cm–1 started to get activated. Moreover, the peaks at 1326 and 1405 cm–1 weakened slightly, implying that NO3– ions were also activated and consumed by NH3 species at this temperature. When the temperature reached 300 ℃, the adsorbed nitrate species, both NO3– and bidentate nitrate, disappeared, which implied that all the reaction species were activated and contributed to the NO reduction. Moreover, the IR peak located at 1466 cm–1, which was ascribed to NH4+, was found to shift to 1484 cm–1, and its intensity increased as the temperature increased from 200 to 300 ℃; this could be attributed the transformation of Lewis acid sites to Br nsted acid sites after contact with the OH groups from the H2O produced during the reaction[38]. Accordingly, it could be concluded from Fig. 3(c) that NO3– and bidentate nitrates only get activated above 175 ℃, and are fully activated at 200 ℃.
The catalytic performance of the Pd/CeO2 catalyst for NO reduction by NH3 is shown in Fig. 4. The catalyst exhibited a high performance for NO reduction with NH3 in the absence of O2. The NO conversion increased rapidly with an increase in the temperature, and the light-off temperature (with 50% conversion of NO) was about 125 ℃. A conversion of 100% was achieved throughout the temperature range of 175–400 ℃. Interestingly, the concentration of N2O firstly increased with the temperature and reached 160 ppm at 125 ℃; it then decreased to 0 when the temperature increased beyond 175 ℃, indicating that the reaction between a 1:1 molar ratio of NO and NH3 had a perfect N2 selectivity above 175 ℃. When the molar ratio of NO:NH3 changed, the N2O selectivity also changed, as shown in Fig. 5. With a decrease in the amount of NH3 introduced, the NO conversion decreased obviously. In detail, for NO:NH3 = 0.4:1, the NO reduction behavior was very similar to that of NO:NH3 = 1:1, but the maximum concentration of N2O (115 ppm) was produced at 120 ℃. For NO:NH3 = 1.5:1, 100% N2 conversion could still be achieved; however, at temperatures below 200 ℃, a relatively low NO conversion, compared with the result shown in Fig. 4, was observed. A further decrease in the NH3 amount resulted in an evident decrease in NO conversion, as shown in Fig. 5(a). It should be pointed out that the theoretical NO conversion is 60% and 37.5% when NO:NH3 equals 2.5:1 and 4:1, respectively. The measured NO conversion in these two cases was higher than the corresponding theoretical data (Fig. 5(a)), indicating a decrease in N2 selectivity, which was in good agreement with the results shown in Fig. 5(b), according to which a high concentration of N2O was detected at temperatures above 250 ℃. Furthermore, even after the catalyst was tested for more than 50 h, no evident activity loss was found, indicating its good stability.
Based on the above results, the route of NO reduction by NH3 over Pd/CeO2 can be approximately described as follows. It should be pointed out that in the absence of O2, the reaction routes for nitrides and nitrates were similar because all the possible reducible [O] species came from NO, and the lattice [O2–] was not involved in the reactions.
The possible reaction routes for N2O can be summarized as follows. (1) At low temperatures below 250 ℃, Reactions (1)–(7) are the reaction routes for NO removal. HON from Reaction (6) is the intermediate for N2O formation, which is in agreement with earlier reports [7, 8]. (2) At high temperatures above 250 ℃, Reactions (5) and (8)–(15) are the reaction routes for NO abatement, and Reaction (15) is the only route for N2O formation.
In detail, the reaction can be discussed based on three temperature windows. (1) At temperature window Ⅰ below 200 ℃, the yield of HON is related to the H• species. When sufficient NH3 is used, the H• species originating from NH3 activation would preferentially react with adsorbed NO to form HON (Reaction (6)), and Reaction (10) is prohibited due to the energy barrier. The combination of two formed HON species through Reaction (7) leads to the formation of N2O; simultaneously, HON can also react with H• to generate N2, as described in Reaction (8). Thus, excessive H• species promote Reaction (8), reducing the formation of N2O from Reaction (7); this is in good agreement with the catalytic performance tests under molar ratios of NO/NH3 below 1.5:1, as presented in Figs. 4 and 5. When NH3 is insufficient, H• is still produced, but Reaction (8) is suppressed; as a result, N2O formation increases with a decease in the NH3/NO molar ratio (Fig. 5(b)). (2) At temperature window Ⅱ between 200 and 250 ℃, Reaction (10) gets activated and suppresses Reactions (6)–(8); hence, the quantitative analysis of the formation of N2O was complicated, and no evident tendency could be observed. (3) At temperature window Ⅲ above 250 ℃, Reaction (6) is completely prevented because the H• species are completely trapped by Reaction (10); therefore, the N2O formation through Reaction (7) is prohibited. However, NO dissociation through Reaction (13) gets activated. In this case, when sufficient NH3 exists, absorbed NO species preferentially react with activated –NH2, which implies that reaction (13) is still prohibited, and the N2O formation through reaction (15) is still suppressed (in the cases of NO:NH3 = 0.4:1 and 1:1). However, when NH3 is insufficient, direct NO dissociation through Reaction (13) occurs, activating Reaction (15) between N• and –NO, and N2O formation is favored again, as shown in Fig. 5(b). The direct dissociation of NO, occurring at high temperatures, is in accordance with the NO-TPD results shown in Fig. 2(b). Evidently, the dissociation of NO is the rate-determining step for N2O formation at temperatures above 250 ℃, and the trapping of O• (Reaction (14)) by surface NH4+ further favors the formation of N2O; the existence of NH4+ at high temperatures was evidenced by the detection of the transformation of L-NH3 to NH4+ at these temperatures (Fig. 3(c)).
In summary, we investigated the mechanism of N2O formation over Pd/CeO2 catalyst during the catalytic reduction of NO using NH3 as the reductant in the absence of O2. The temperature-dependent formation mechanism of N2O was confirmed. When NH3 is sufficient for NO abatement, N2O formation only takes place at low temperatures, through the dissociation of HON intermediates. However, when NH3 is insufficient for NO removal, direct dissociation of NO into N• and O• takes place at high temperatures, and the combination of N• and NO- induces N2O emission. Our results provide detailed knowledge on the mechanism of the N2O formation, which is good reference for the suppression of N2O emission in the low-temperature SCR of NO with NH3.
2019,
Vol. 40 Issue (7): 1070-1077 DOI: 10.1016/S1872-2067(19)63328-0 
PDF
2019,
Vol. 40 Issue (7): 1070-1077 DOI: 10.1016/S1872-2067(19)63328-0 PDF