The catalytic reduction of NO to N2 by CO has been widely applied in emission control for automotive exhaust [1]. In recent years, CeO2 has generally been used as an oxygen buffer in three-way catalytic converter systems owing to its oxygen storage/release capacity, undergoing redox processes involving the Ce4+/Ce3+ couple, CeO2 ⇌ CeO2-x + (x/2) O2, that can greatly promote CO oxidation[2]. Among transitional metals, rhodium shows the strongest ability for specifically enhancing NO reduction due to its ability to efficiently dissociate NO, which is the main reason for Rh addition to commercial three-way catalysts [3, 4]. Additionally, Rh/ceria interactions improve both resistance to thermally induced catalyst sintering and catalytic activity due to additional bifunctional sites at the metal-support interfaces [4, 5].
Few reports exists concerning the formation of NH3 during the catalytic reduction of NO by CO, but N2O selectivity has been discussed [2, 4, 6, 7], resulting in a misunderstanding of the reaction mechanisms [8-11]. N2 selectivity is defined as 'the percentage of NO conversion to N2 (total reduced NO concentration minus double N2O concentration) in total NO conversion' [7, 12], indicating that only N2, N2O, and CO2 products are taken into consideration during the catalytic reduction of NO by CO on Rh catalyst. Dropsch et al. [13] suggested that CO could react with hydroxyl groups on the support to release hydrogen gas. Until now, NH3 detection in the effluent gas has not been used to analyze N2 selectivity, resulting in an improper mechanistic and kinetic understanding of the NO/CO reaction on Rh catalyst [11, 14].
NH3 emission from automobiles is principally due to the reduction of nitrogen oxides (NOx) by H2 generated from water-gas shift and steam-reforming reactions [15-17]. Recent studies [18, 19] have emphasized that NH3 is among the main factors causing the formation of secondary inorganic aerosols (NH4+, SO42-, and NO3-). Therefore, an exact understanding of the NH3 formation mechanism in the catalytic reduction of NO to N2 by CO plays a significant role in effective NH3 emission control for gasoline-fueled vehicles under transient driving conditions. This work focuses on determining the behavior of NH3 generation and proposing a possible reaction mechanism for NH3 formation during the catalytic reduction of NO by CO, leading researchers to a more accurate analysis of light-off test results.
CeO2 was synthesized using a precipitation method. A slight excess of 25% NH4OH solution was added dropwise into the requisite amount of aqueous solution of Ce(NO3)3·5H2O under continuous stirring. The obtained gelatinous precipitate was aged at 80 ℃ for 3 h under vigorous stirring, filtered, and washed with distilled water several times until no pH change was observed. The precipitate was dried at 120 ℃ overnight and then calcined at 300 ℃ in static air for 1 h. The resultant powder was calcined at 550 ℃ for 4 h under an air atmosphere.
Rh-CeO2 catalyst was prepared by mixing the obtained gelatinous precipitate with Rh nanoparticles (NPs), followed by aging, drying, and calcinating using the same method outlined for CeO2 preparation. For Rh NP synthesis, polyvinylpyrrolidone (PVP) and citric acid (CA) were added into 85 mL of Rh(NO3)3 aqueous solution with a PVP monomer unit:CA:Rh molar ratio of 5:4:1, and then placed in a 250-mL three-necked flask. The mixture was preheated under a nitrogen atmosphere and magnetic stirring to 60 ℃ for 5 min and the pH was adjusted to 3.0 by the addition of 10% NH4OH solution. The requisite borane tert-butylamine (TBAB), which served as a reducing agent in the reaction, was dissolved in 15 mL of distilled water and added to the above mixture with a TBAB:Rh molar ratio of 5:1. The reaction proceeded at 60 ℃ for 10 min.
Fourier transform infrared (FT-IR) measurements of the solid samples were carried out in transmission mode on a Nicolet iS10 FT-IR spectrometer with a resolution of 2 cm-1 using the KBr wafer technique and scanning 36 times. For wafer preparation, catalyst sample (10 mg) was mixed with KBr (200 mg, Specpure) and pressed under a pressure of 20 MPa for several minutes. The actual Rh loading amount was analyzed on a PerkinElmer Optima 8000 ICP-OES instrument.
In-situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) was performed on a Nicolet 6700 FT-IR spectrometer equipped with a mercury cadmium telluride (MCT) detector. Each sample (100 mg) was pretreated at 500 ℃ for 60 min in a N2 stream and then cooled to room temperature. Before starting the experiment, background spectra were recorded at 100 ℃ intervals from 100 to 500 ℃ in a N2 stream at a heating rate of 10 ℃/min. Each corresponding spectrum was measured in a reaction gas stream containing 0.1% CO and/or 0.1% NO for transient experiments. The total flow rate of the mixed gas stream in these experiments was 100 mL/min.
The catalytic activity and selectivity of the catalyst samples were studied in the catalytic reduction of NO by CO in a continuous flow fixed-bed microreactor. Catalyst samples (500 mg) with an average diameter of 40-60 mesh were placed in a stainless-steel tube with an inner diameter of 5 mm to conduct activity tests, and CO temperature-programmed reduction (CO-TPR) and NO temperature-programmed oxidation (NO-TPO) measurements. The reaction mixture gas, as shown in Table 1, was supplied at a flow rate of 500 mL/min and the space velocity was 60000 mL/(g·h). The temperature dependence of the reaction was determined in temperature-programmed mode by heating from ambient temperature to 500 ℃ at a heating rate of 10 ℃/min. The composition of the effluent gas was analyzed online using a 2030DBG2EZKS13T MultiGas FT-IR Analyzer purchased from MKS Instruments, Inc.
The conversion rate of NO, X (NO), was calculated using equation (1):
where [NO]in and [NO]out are the concentrations of injected and effluent NO gas, respectively. The conversion rate of CO, X (CO), was calculated using the same method. The selectivities to N2O and NH3, S (N2O) and S (NH3), respectively, were calculated using equations (2) and (3), based on the concentrations of N2O and NH3 products, respectively.
Activation energies were estimated for CO conversion using the Arrhenius equation. About 200 mg of catalyst sample was used in each measurement. The samples were first activated at 400 ℃ in a CO/NO or CO/NO/H2O reaction mixture gas stream at a total flow rate of 500 mL/min and then cooled to room temperature. The composition of the effluent gas was analyzed online at 10 ℃ interval in the (5-20)% CO conversion temperature range. The specific reaction rate was defined as the total number of moles of CO2 produced per second per mass of catalyst at a specific temperature.
The catalytic reduction of NO by CO is an important model reaction for simulating three-way catalysis. In this work, model CO/NO and CO/NO/H2O reactions were carried out using a typical fixed-bed flow reactor under a CO/NO stoichiometric mixture. CO and NO conversion curves for the model reactions are shown in Fig. 1(a). T50 values for CO and NO conversions in the CO/NO reaction over Rh-CeO2 catalyst with an actual Rh loading amount of 0.296% were 152 and 156 ℃ (Table 2), respectively, which were significantly lower than those over the CeO2 catalyst. This was further confirmed by the lower apparent activation energy (Ea) of 36 kJ/mol for the Rh-CeO2 catalyst (Fig. 2), compared with 105 kJ/mol for the CeO2 catalyst, which explained the lack of intrinsic reactivity for base metals compared with precious metals [3, 20]. For the Rh-CeO2 catalyst, water-induced active inhibition of CO and NO conversions was explained by the corresponding increases in T90 and Ea (up to 21 kJ/mol) compared with results under dry conditions.
NO is partially reduced to N2O as the main product on the Ce site below 250 ℃, and N2O selectivity decreased with increasing temperature in the CO/NO reaction (Fig. 3). Notably, Fig. 1(b) shows that clear NH3 gas detection started at about 190 ℃ and a maximum NH3 selectivity of 14.8% was obtained on the CeO2 support at 268 ℃ in the H2O-free reaction mixture. Furthermore, CO and NO conversions of 17.1% and 12.6%, respectively, were achieved at the same temperature. By contrast, N2O selectivity strongly decreased when Rh nanoparticles were supported because N2O was more readily formed and binded less strongly. Therefore, N2O was more likely to desorb into the gas phase on CeO2 surfaces than on Rh sites [11]. The Rh-CeO2 catalyst also showed higher NH3 selectivity up to 236 ℃, but gave lower NH3 selectivity at higher temperatures due to a rhodium-mediated increase in reactivity.
Fig. 4(b) shows that the amount of NH3 product on the Rh-CeO2 catalyst increased in the following order: CO/NO/H2O (0.0158% NH3 at 203 ℃) > CO/NO (0.0094% NH3 at 222 ℃) > NO/H2 (0.004% NH3 at 240 ℃). There was no clear NH3 formation during the NO-TPO and NO/H2O reactions at low temperatures. The results indicated that CO and NO might react with hydroxyl groups on the surface of the CeO2 support to release NH3 gas via a water-gas shift reaction, comprising the chemisorption and reaction of CO with hydroxyl species, followed by NO reduction, as described in equations (4) and (5) [13, 21, 22].
Results concerning the influence of water on NH3 selectivity are shown in Fig. 1(b). Two peaks for NH3 formation, centered at 199 and 342 ℃ with NH3 selectivities of 25.2% and 20.9%, respectively, were observed in the water mixed feed stream. The maximum NH3 selectivity at low temperature was attributed to OH groups, while NH3 was formed at high temperatures due to the reduction of NOx by H2, where CO reacts with water to produce H2 and CO2 according to the water-gas shift reaction, as shown in equation (6). Compared with the CO-TPR results, CO conversion rates increased with increasing temperature above 230 ℃ due to the water-gas shift reaction, as shown in Fig. 5. Theoretically, 0.05% H2 would be released when a total conversion of 0.1% CO was achieved in the water-gas shift reaction. The competitive reaction of H2O and NO with CO during the CO/NO/H2O reaction might form H2 as a product.
As shown in Figs. 4 and 5, CO conversion was observed during the CO-TPR experiment, illustrating that CO indeed reacted with active species on the catalyst surface. There was little NO conversion and NH3 production under the NO/H2O stream, which indicated that H2O inhibited NO adsorption due to the competitive adsorption of H2O and NO at the same sites[23]. This inhibition effect would reduce the reactivity of NO reduction by NH3, leading to increased NH3 emission.
FT-IR spectroscopy was used to further confirm that adsorbed CO and NO reacted with hydroxyl groups on the surface of CeO2 to release NH3. The FT-IR spectrum of pristine Rh-CeO2 catalyst (Fig. 6) showed three main bands at 3693, 3650, and 3506 cm-1, which were assigned to Ce-OH terminal hydroxyl groups (type Ⅰ), Ce2-OH double-bridging groups (type Ⅱ), and cerium oxyhydroxide impurities, respectively [24, 25]. After the CO/NO/H2O reaction, three main stretching vibrations of OH groups were still detected, while the Ce-OH vibration showed a sharp decrease in intensity and shifted to higher wavenumber (3704 cm-1). However, no band was observed in the range 3800-3400 cm-1 after CO/NO cycling. This was explained by the interaction of adsorbed H2O with basic OH groups. Demoulin et al. [22] suggested that hydroxylated surfaces resulted from water vapor dissociatively chemisorbing to become hydroxyl ions. Surface water molecules can also react with chemisorbed CO species to form H+ and formate ions [26].
Surface species evolutions of CO and/or NO on pretreated Rh-CeO2 catalyst were also investigated using in-situ DRIFTS measurements. The stretching vibrations of hydroxyl groups exhibited a significant wavenumber shift after CO adsorption at 100 ℃, as shown in Fig. 7(a). This was mainly due to the presence of surface hydroxyls on the CeO2 support, which (1) allows the migration of adsorbed CO on Rh across the CeO2 surface to isolated OH sites to form carbonates (1583, 1453, and 1402 cm-1), and (2) alters the coordination of chemisorbed CO at the oxygen end, weakening the C-O bond followed by facilitating CO dissociation during CO oxidation [24, 25, 27, 28]. At this temperature, the features at 2168, 2073, 2010, and 1760 cm-1 were attributed to Ce4+-carbonyls, gem-dicarbonyls, linear carbonyls, and bridged carbonyls on Rh, respectively [27]. The vibrational IR spectra of carbonyl species, and type Ⅰ and Ⅱ hydroxyls, showed a decrease in intensity, which indicated the consumption of adsorbed CO. The band at 3616 cm-1 was assigned to O-H stretching vibrations of hydrogen carbonate species formed by the interaction of CO with basic OH- ions [10, 28]. In summary, the reaction of type Ⅰ and Ⅱ hydroxyls with adsorbed CO led to a weakening of their vibrations and hydrogen carbonate formation.
Figs. 7(b) and (c) show in-situ DRIFTS results for temperature-programmed experiments in NO and CO/NO streams, respectively. Under NO, the obvious consumption of three types of hydroxyl groups was accomplished by the reaction with NO, with a peak apparition appearing at 3622 cm-1 (when the temperature was raised to 300 ℃). This was associated with N-H stretching in coordinated ammonia from NH3 product adsorbed on the oxide surface [29]. The consumption of adsorbed NO on Rh (band at 1922 cm-1) and the formation of nitrates on CeO2 (1594, 1568, and 1530 cm-1) were observed with increasing temperature [23, 30]. These results suggested that NO reacted with OH groups to form nitrates and release NH3, explaining NH3 formation and clear NO conversion during the NO-TPO experiment. In contrast, the peak at 3622 cm-1 started to appear at a lower temperature (200 ℃) and showed a higher intensity at the same temperature when CO and NO were introduced simultaneously, illustrating that CO played an important role in NH3 formation during the NO reaction. These results were in good agreement with those in Fig. 4(b).
NH3 production in the catalytic reduction of NO to N2 by CO under dry conditions was ascribed to NO reduction by molecular hydrogen present in hydroxyl groups on the CeO2 surface. In-situ DRIFTS results suggested that the consumption of cerium oxyhydroxide impurities was directly accomplished by NO reduction, whereas type Ⅰ and Ⅱ hydroxyls were consumed via a water-gas shift reaction followed by NOx reduction, which led to NH3 generation. In the presence of water, water-induced hydroxylation could promote NH3 formation, whereas the competitive adsorption of H2O and NO on the same sites would inhibit the reactivity of NO reduction by NH3, leading to increased NH3 emission.
This work was supported by Xiaodong Wu (School of Materials Science and Engineering, Tsinghua University) for in-situ DRIFTS characterization.
2017,
Vol. 38 Issue (8): 1399-1405 
PDF
2017,
Vol. 38 Issue (8): 1399-1405 PDF