Nitrogen oxides (NOx), one of the major classes of air pollutants, can cause a series of environmental problems such as acid rain and photochemical smog [1]. Selective catalytic reduction (SCR) has been demonstrated to be one effective method for removing NOx in the presence of oxygen, and different reducing agents have been employed in this process [2]. Currently, SCR of NOx by ammonia (NH3-SCR) has been widely adopted for the control of NOx [3-5]. However, NH3 slip and vanadia emissions associated with the NH3-SCR process have restricted its application [6, 7]. The SCR of NOx by hydrocarbons (HC-SCR) also has some disadvantages such as poor activity at low temperatures [8]. As an alternative technology, selective catalytic reduction of NOx by H2 (H2-SCR) has recently attracted increasing attention [9-12]. When H2 is used as the reducing agent, NOx can be effectively removed at relatively low temperatures (100–300 ℃) [13]. Moreover, the combustion product of H2 is H2O, which is environmentally benign, and corresponds with the goal of environmental protection.
In the process of H2-SCR, noble metals, especially Pt and Pd, were found to be the active components [14-17]. Many studies have focused on the development of Pd-based catalysts for the H2-SCR, due to the lower cost of Pd than that of Pt [18-22]. Lee et al. [23] reported that Pd/Al2O3 prepared using PdCl2 as the precursor was more active than that prepared using Pd(NO3)2 as the precursor. Ueda et al. [24] found that Pd/TiO2 was much more active than a Pd/Al2O3 catalyst. The activity of Pd-based catalysts in the H2-SCR of NOx is dependent upon the chemical nature of Pd (PdO, metallic Pd0, and Pd alloy). Wen [25] proposed that Pd0 plays a significant role in the H2-SCR reaction. Xu et al. [26] investigated the activities of a series of Pd-doped aluminate spinel catalysts for the H2-SCR reaction, and found that the activity decreased in the following sequence: Pd/CoAl > Pd/ZnAl > Pd/CuAl. Based on an XPS analysis, they proposed that the most-reduced state of Pd in Pd/CoAl was responsible for the highest activity.
The activity of a catalyst is closely related to the preparation method [8]. Chiarello et al. [21] found that Pd/LaCoO3 prepared by the flame-spray pyrolysis method was more active than that prepared by the impregnation method. Pd/MFI prepared by the sublimation method also exhibited higher activity in the reduction of NOx by H2 than that prepared by the impregnation method [25]. However, little research has been reported on the effect of preparation method on the activity of Pd/TiO2 catalysts for the H2-SCR of NOx. In the present study, Pd/TiO2 catalysts were prepared by different preparation methods, and it was found that the catalyst prepared by the polyethylene glycol reduction method exhibited the highest activity in the reduction of NOx. On the basis of the catalyst characterization, the reason for this difference in activity resulting from the different preparation methods has been revealed.
0.5% Pd/TiO2 catalysts were prepared by the impregnation (abbreviated as Pd/TiO2 (IM)), deposition-precipitation (abbreviated as Pd/TiO2 (DP)), and polyethylene glycol reduction (abbreviated as Pd/TiO2 (PR)) methods. Degussa AEROSIL TiO2 P25 was used as the support.
For Pd/TiO2 (IM), TiO2 was impregnated with an appropriate amount of PdCl2 solution, stirred for 4 h, dried at 120 ℃, and then calcined in air at 500 ℃ for 4 h.
For Pd/TiO2 (DP), TiO2 was dispersed into an appropriate amount of PdCl2 solution and mixed for 1 h, and then NaOH (10 mol/L) was added dropwise into the homogeneous mixture until the pH of the suspension reached 10. The suspension was filtered and washed thoroughly with deionized water, dried in an oven for 12 h at 120 ℃, and finally calcined in air at 500 ℃ for 4 h.
For Pd/TiO2 (PR), an appropriate amount of PdCl2, polyethylene glycol (PEG), and polyvinylpyrrolidone (PVP) were combined in a beaker and mixed well, and then the mixture was transferred to a three-neck flask placed in an oil bath. An appropriate amount of TiO2 was dispersed into the mixed solution in the three-neck flask. The mixed solution was heated at 150 ℃ under N2 flow for 2 h, then allowed to cool to room temperature. The suspension was then centrifuged and washed thoroughly with ethanol. The obtained solid was dried under vacuum at 60 ℃ for 12 h.
The H2-SCR measurements were carried out in a fixed-bed quartz reactor using 0.2 g of 40–60 mesh catalyst. The typical reactant consisted of 2000 ppm NO, 8000 ppm H2, 5% O2, and the balance was He. The total flow rate was 200 cm3 min–1 (GHSV = 35000 h–1). The reaction temperature was increased from 100 ℃ to 400 ℃. The compositions of the inlet and outlet gases were monitored by a NO/NO2 chemiluminescence analyzer (Thermo Scientific, model 42i-HL) and an FTIR spectrometer (Gasmet FTIR DX4000). The NOx conversion and N2 selectivity were calculated using equations reported previously [19]. At each temperature, the catalytic activity data were collected after the catalytic reaction reached a steady-state condition.
N2 adsorption and desorption isotherms were measured on a Quantachrome Autosorb AS-1 system. The specific surface area and the pore size distribution were calculated by BET plots and the BJH method, respectively. The Pd loading of the prepared catalyst was analyzed by inductively coupled plasma atomic emission spectroscopy (ICP-AES) on a Plasma-Spec-I spectrometer. The particle size of Pd was measured with a JEOL JEM-2100 transmission electron microscope (TEM). The crystal structures were determined by X-ray diffraction (XRD; Brucker D8 ADVANCE system) with a Cu Kα radiation source (45 kV, 200 mA). XPS analysis was performed on an ESCALab220i-XL electron spectrometer (VG Scientific), using 300 W Mg Kα radiation. The binding energies were referenced against C 1s (284.8 eV).
A Thermo Scientific Nicolet 6700 spectrometer, which was equipped with a high-temperature environmental cell fitted with a ZnSe window as well as an MCT detector cooled with liquid N2, was employed to explore the mechanism of the H2-SCR. The catalyst was first pre-treated at 400 ℃ under helium at a total flow rate of 100 cm3 min–1 for 1 h in order to remove impurities on the surface of the catalyst, and then cooled to the desired temperature. A background spectrum was collected under a helium atmosphere and was subtracted from the sample spectra. DRIFT spectra were recorded by accumulating 100 scans with a resolution of 4 cm–1.
The effects of different catalyst preparation methods on the performance of a Pd/TiO2 catalyst in the H2-SCR are illustrated in Fig. 1. As shown in Fig. 1(a), both Pd/TiO2 (IM) and Pd/TiO2 (DP) were inactive at 100 ℃. For the Pd/TiO2 (IM) catalyst, the maximum NOx conversion was achieved at 125 ℃, with NOx conversation decreasing as the temperature was increased further. Above 150 ℃, Pd/TiO2 (DP) was more active than Pd/TiO2 (IM), and the maximum NOx conversion was obtained at 250 ℃. Compared to the Pd/TiO2 (IM) and Pd/TiO2 (DP) catalysts, Pd/TiO2 (PR) exhibited the highest activity over the widest temperature window. From Fig. 1(b), it can be seen that Pd/TiO2 (PR) exhibited much better N2 selectivity than Pd/TiO2 (IM) and Pd/TiO2 (DP) in the low temperature range (100–200 ℃). As shown in Fig. S1 (see the Supporting Information), over the Pd/TiO2 (PR) catalyst, the main by-product was N2O and only a small amount of NH3 formed between 150 and 250 ℃. In contrast, the amounts of NH3 formed over Pd/TiO2 (IM) and Pd/TiO2 (DP) catalysts were higher than that over Pd/TiO2 (PR). Therefore, the preparation method exerted a significant effect on the H2-SCR performance, and the catalyst prepared by polyethylene glycol reduction method was found to be the most active.
The physicochemical properties of TiO2, Pd/TiO2 (IM), Pd/TiO2 (DP), and Pd/TiO2 (PR) catalysts are summarized in Table 1. Compared to the bare TiO2 support, the BET surface areas of the three Pd/TiO2 catalysts were lower due to the introduction of Pd. The BET surface areas of Pd/TiO2 (DP) and Pd/TiO2 (PR) were similar, and larger than that of Pd/TiO2 (IM). The Pd loading of Pd/TiO2 (PR) was slightly lower than those of Pd/TiO2 (IM) and Pd/TiO2 (DP).
XRD patterns of the different Pd/TiO2 catalysts are shown in Fig. 2. Strong characteristic diffraction peaks corresponding to the anatase phase and weak diffraction peaks ascribable to the rutile phase [5] were observed in all three catalysts. However, no diffraction peaks attributed to Pd species were observed, which suggests that Pd was highly dispersed on the TiO2 support, or that the loading of Pd was so low that it could not be detected.
The TEM images and the particle size distributions of the Pd/TiO2 (IM), Pd/TiO2 (DP), and Pd/TiO2 (PR) catalysts are displayed in Fig. 3. It can be seen that Pd was highly dispersed on the surface of TiO2 in all the three catalysts. The average particle sizes of Pd on the Pd/TiO2 (IM), Pd/TiO2 (DP), and Pd/TiO2 (PR) catalysts were 1.28, 1.50, and 1.02 nm, respectively. The smallest Pd particle size, observed in Pd/TiO2 (PR), may have contributed to its high H2-SCR activity.
The chemical nature of each of Pd/TiO2 (IM), Pd/TiO2 (DP), and Pd/TiO2 (PR) was analyzed by XPS. Fig. 4(a) illustrates the Pd 3d XPS spectra of the Pd/TiO2 (IM), Pd/TiO2 (DP), and Pd/TiO2 (PR) catalysts. For the Pd/TiO2 (IM) catalyst, the Pd 3d3/2 and 3d5/2 peaks are observed at approximately 341.7 and 336.6 eV, respectively, which are the classical XPS peaks of Pd2+ [13, 18, 27]. Similar XPS spectra were also observed for the Pd/TiO2 (DP) catalyst. In the case of the Pd/TiO2 (PR) catalyst, the Pd 3d3/2 and 3d5/2 peaks were shifted to lower bonding energies by approximately 1.3 eV. The peaks at 340.2 and 335.1 eV were assigned to metallic Pd [18, 27, 28]. As shown in Fig. 4(b), two Ti 2p peaks at around 458.5 eV (Ti 2p3/2) and 464.3 eV (Ti 2p1/2), which were assigned to Ti4+ [29, 30], appeared in the spectra of all three catalysts.
O 1s XPS spectra of Pd/TiO2 (IM), Pd/TiO2 (DP), and Pd/TiO2 (PR) are exhibited in Fig. 4(c). Two remarkable characteristic peaks are present in the spectra of all three catalysts: the peak at 529.5–529.7 eV was attributed to lattice oxygen (denoted as Oβ), and the other at 531.5–531.6 eV was ascribed to surface chemisorbed oxygen (denoted as Oα). It is worthwhile to note that the content of Oα in Pd/TiO2 (PR) was 23.39%, which is much higher than those in Pd/TiO2 (DP) and Pd/TiO2 (IM) (11.45% and 14.86%, respectively). It has been reported that Oα contributed to the adsorption and oxidation of NO [19], thus promoting the reduction of NOx by H2 [31].
Fig. 5 shows the in situ DRIFTS spectra of NO+O2 co-adsorption at different temperatures on Pd/TiO2 (IM) and Pd/TiO2 (PR). As can be seen in Fig. 5(a), a series of bands at 1851, 1815, 1749, 1612, 1581, 1498, 1290, and 1246 cm–1 were observed for the Pd/TiO2 (IM) catalyst, which were respectively assigned to physisorbed NO (1851 and 1815 cm–1) [32], trans-(NO)2 (1749 cm–1) [5, 14], NO2 (1612 cm–1) [33], bidentate nitrate (1581 cm–1) [34, 35], monodentate nitrate (1498 cm–1) [10, 14], chelating nitrite (1290 cm–1) [33, 36], and monodentate nitrite (1246 cm-1) species [14]. The peak ascribed to NO adsorption disappeared as the temperature reached 100 ℃. On the contrary, the intensities of the peaks assigned to trans-(NO)2 increased slightly in intensity with increasing temperature. With increasing temperature, new bands also appeared at 1529 and 1450 cm–1, which correspond to bidentate nitrate [37] and monodentate nitrate [14], respectively.
Fig. 5(b) illustrates the in situ DRIFT spectra of NO+O2 co-adsorption on the Pd/TiO2 (PR) catalyst at different temperatures. Compared with the spectra observed for Pd/TiO2 (IM), the intensities of the peaks ascribed to trans-(NO)2 (1759 cm–1), chelating nitrite (1286 cm–1), and monodentate nitrite (1246 cm–1) species were stronger. Moreover, at 30 ℃, a new band was observed due to a bent mono-nitrosyl adsorbed on metallic Pd0 (1681 cm−1) [19, 22], which further demonstrates the existence of metallic Pd on the Pd/TiO2 (PR) catalyst.
Fig. 6 shows the dynamic changes of the in situ DRIFTS spectra of adsorbed species on Pd/TiO2 (PR) at 150 ℃. As shown in Fig. 6(a), after exposure of Pd/TiO2 (PR) to NO+O2 at 150 ℃, some bands assignable to NO2 (1607 cm–1), bidentate nitrate (1583 cm–1), bidentate nitrate (1531 cm–1), chelating nitrite (1283 cm–1), and monodentate nitrite (1245 cm–1) species appeared. After changing the gas to H2, the bands assigned to adsorbed NOx species disappeared in 20 min, during which time the band at 1200 cm–1 appeared, which was attributed to NH3 on Lewis acid sites [33, 38]. This indicates that the adsorbed NOx species can react with H2 to form NH3 species. When the gas was changed back to NO+O2, the peaks assigned to the NH3 species disappeared and the adsorbed NOx species appeared again. This observation indicates that the NH3 species formed were reactive intermediates in the H2-SCR of NOx over the Pd/TiO2 (PR) catalyst. From Fig. 6(b), it can be seen that if the gas was switched to H2+O2 instead of H2, the adsorbed NOx species could also react with H2+O2. Meanwhile, the band assigned to NH3 on Lewis acid sites also appeared, meaning that it is also reactive. Therefore, NH3 species played an important role in the reduction of NOx by H2 over the Pd/TiO2 (PR) catalyst.
In the H2-SCR of NOx, the catalytic activity is dependent on the nature of the noble metal, and the metallic state can contribute to improving the activity in the reduction of NOx [25, 39]. In the case of Pd/TiO2-Al2O3, the added Sn led to the formation of Pd0, thus exerting a promoting effect on the activity of the catalyst in the H2-SCR [40]. For a Pd-Au/TiO2 catalyst, the interaction between Pd and Au resulted in the formation of Pd0, leading to improved activity in the reduction of NOx [18]. The activity of Pd/TiO2 (IM) pretreated by H2 at 400 ℃ for 1 h (denoted herein as Pd/TiO2 (IM-H2)) was also evaluated in this study, and the result is shown in Fig. S2. In comparison to Pd/TiO2 (IM), the activity of Pd/TiO2 (IM-H2) at 100 ℃ was significantly improved. This confirms that Pd0 is active in the reduction of NOx, which is consistent with previous reports [24, 25]. The highly dispersed metallic Pd on the Pd/TiO2 (PR) catalyst played a significant role in the high activity. With increased temperature, the NOx conversion over Pd/TiO2 (IM-H2) decreased, even falling lower than that over Pd/TiO2 (IM). TiO2 is a reduced oxide, and during the 400 ℃ pretreatment of Pd/TiO2 (IM-H2), TiO2 could be reduced along with Pd, thus leading to an improved oxidation activity as reported by Li et al. [41]. The improved oxidation activity would cause the combustion of H2 with O2 to become predominant at elevated temperatures over the Pd/TiO2 (IM-H2) catalyst, and thus less H2 would remain for the reaction with NOx. This is a possible reason for the decreased NOx conversion over Pd/TiO2 (IM-H2) at high temperatures. In contrast to the pretreatment by H2, in the polyethylene glycol reduction process, Pd0 was formed while TiO2 was seldom reduced. Therefore, Pd/TiO2 (PR) exhibited high activity in the H2-SCR of NOx.
In comparison to Pd/TiO2 (IM) and Pd/TiO2 (DP), Pd/TiO2 (PR) exhibited superior catalytic activity in the H2-SCR of NOx over a wide temperature window. Characterization results showed that on Pd/TiO2 (PR), Pd existed entirely in the form of Pd0, which is active in the reduction of NOx. In situ DRIFTS results revealed that more chelating nitrite and monodentate nitrite species formed over the Pd/TiO2 (PR) catalyst, both of which are reactive in the reduction of NOx. DRIFTS studies also indicated that NH3 species are reactive intermediates in the H2-SCR of NOx. The presence of metallic Pd and more reactive chelating nitrite and monodentate nitrite species on the Pd/TiO2 (PR) catalyst was responsible for its high activity in the H2-SCR of NOx.
The authors acknowledge the financial support from the National Key R & D Program of China (2017YFC0210700), the National Natural Science Foundation of China (21876009, 21611130170), the Beijing Municipal Natural Science Foundation (8162030), and the Fundamental Research Funds for the Central Universities (XK1802–1).
2019,
Vol. 40 Issue (6): 849-855 DOI: S1872-2067(19)63297-3 
PDF
2019,
Vol. 40 Issue (6): 849-855 DOI: S1872-2067(19)63297-3 PDF