Selective catalytic reduction (SCR) of NOx is a promising de-NOx technique in excess oxygen [1]. NH3-SCR is a well- known process for NOx emission control in stationary sources [2]. However, many problems have been encountered [3, 4], such as NH3 slip, fouling of air heaters, and catalyst deterioration caused by ammonium sulfate deposition at low reaction temperatures [5]. Extensive research has been performed on SCR by hydrocarbons (HC) in recent decades [6, 7, 8], but no significant practical applications have been developed. The major problem is the considerable decrease in catalytic activity caused by the presence of H2O vapor in the reaction system.
H2 has also been investigated as a reductant for SCR of NOx in excess oxygen (H2-SCR) [9, 10, 11, 12]. Recently, because of the lower working temperature (< 200 °C) and significant progress in achieving N2 selectivity, H2-SCR has attracted much attention [13, 14, 15, 16, 17, 18]. Pt is the most active component for this reaction, and its most active state is Pt0 [19, 20, 21, 22]. Although the effect of Pt loading on the support has been widely investigated, the effects of the surface properties of the support on the catalytic performance, especially the effect on N2 selectivity, remain unclear because of the complex catalysts and reaction conditions used in studies. Na2O was used to modify Pt/γ-Al2O3 and Pt/SiO2 catalysts for the H2-SCR reaction [23]. A similar investigation was performed using Na-modified Pt-ZSM-5 [14]. The authors suggested that NO-type adsorption species led to N2O formation, whereas NO2− species were possible intermediates in N2 production, and an acidic-basic bifunctional catalyst was required for the reaction. It has also been proposed that acidic supports are favorable for the target reaction [24, 25]. Recently, Pt loaded on Al-MCM-41 with Brönsted acidic sites was reported to be active and selective for the reaction [26].
High temperatures increase N2 selectivity. Burch et al. [27] found that the primary product of H2-SCR over Pt/SiO2 is N2 formed from N2O at 150 °C. Yang et al. [28] reported that the temperature at which the primary product from N2O became N2 over a Pt/γ-Al2O3 monolithic catalyst was 80 °C. The process NO(ad) + N(ad) → N2O represents the N2O formation route on a Pt surface [26, 29, 30]. Burch et al. [29] suggested that four neighboring Pt sites were required for N2 formation, whereas N2O formation needed only three Pt sites. Machida et al. [21] proposed that the reaction product is determined by the number of H2 molecules taking part in nitrate species reduction over a Pt/TiO2-ZrO2 catalyst. Recently, Li et al. [31] proposed that N2O was also produced via the Eley-Rideal mechanism, i.e., NO + N(ad) → N2O, over a Pd/TiO2 catalyst.
Although the rational design of better catalysts for this reaction is still difficult, reports suggest that some surface properties of the support strongly influence the catalytic activity and N2 selectivity of the catalyst. We investigated this in a study using HZSM-5, ZrO2, γ-Al2O3, and MgO, which have different surfaces from each other in terms of acidity/basicity, as the supports for Pt catalysts in the H2-SCR. It was found that the ability of Pt to activate the H2 and NOx adsorption capacities of the support greatly influenced the catalytic activity and N2 selectivity of the catalyst, and these factors are closely associated with the acidity-basicity of the support.
A commercial HZSM-5 zeolite (SiO2/Al2O3 = 25) was purchased from Nankai University. MgO was purchased from the Dunhuang Chemical Plant. Commercial spherical γ-Al2O3 was purchased from the Shandong Ocean Chemical Plant and was ground to a powder before use as a support. ZrO2 was prepared by a precipitation method using ZrOCl2·8H2O and ammonia as precursors.
All the Pt catalysts were prepared using an impregnation method as follows. The supports were impregnated with an appropriate amount of H2PtCl6 aqueous solution at room temperature for 24 h, dried at 120 °C for 5 h, and calcined at 500 °C in air for 5 h. The Pt loading on each support was 0.5 wt%.
X-ray photoelectron spectroscopy (XPS) was performed using an ESCALAB 250Xi (Thermo Fisher Scientific Inc.) spectrometer with an Al Kα source. All spectra were calibrated using the C 1s binding energy, 284.6 eV.
High-resolution transmission electron microscopy (HRTEM; TECNAI G2 F30 instrument operated at an accelerating voltage of 300 kV) was performed after reduction of the catalysts using 1.3% H2/N2.
CO2 temperature-programmed desorption (TPD) was performed using a CHEMBETTPD/TPR apparatus (Quantachrome Company). For the experiments, each catalyst (0.100 g) was pretreated online at 600 °C for 30 min in high-purity He in a U-shaped quartz tube, exposed to pure CO2 at 50 °C for 60 min, and purged with He for 40 min at this temperature. The CO2-TPD profile was recorded using a thermal conductivity detector (TCD) from 50 to 950 °C at 10 °C/min.
Pyridine-adsorption Fourier transform infrared (FTIR) spectroscopy was used to study the support surface acidity (Bruker TENSOR 27 FTIR instrument with quartz IR cell with a CaF2 window). Self-supporting wafers of the samples, of weight 30 ± 0.5 mg, were used for the measurements. The sample was treated in a N2 flow at 500 °C for 30 min in the cell, and then the cell was evacuated for 30 s, cooled to room temperature, and exposed to a pyridine atmosphere for 30 min. The pyridine-adsorption FTIR spectrum was recorded at 200 °C under evacuation.
In situ FTIR spectra of the nitrous species on the catalysts were obtained at a resolution of 4 cm−1. The averaged spectrum from 32 scans was used. Before the measurements, the sample wafer was treated in 1.3% H2/N2 at 500 °C for 30 min in the IR cell. The spectrum of the solid catalyst at evacuation and that of the gas mixture without the wafer were obtained at an appropriate temperature for FTIR single-beam background subtraction.
The H2-SCR reaction was carried out in a quartz reactor (i.d. 4 mm) at atmospheric pressure. The granular catalyst (20-40 mesh, 0.200 g) was fixed in the reactor. To determine the chemical states of Pt in the catalysts, the catalysts were pretreated with air and with 1.3% H2/N2 at 500 °C, and their catalytic activities were examined. The reactant gas mixture, consisting of 1000 ppm NOx, 5000 ppm H2, and 10% O2 in N2, was passed over the catalyst at a total flow rate of 200 mL/min (W/F = 0.06 g s/mL) at the desired reaction temperatures from 280 to 60 °C. The catalytic data were recorded after 30 min at each temperature. H2 and N2O were analyzed online using a gas chromatograph (GC 7890) equipped with a 13X molecular sieve column (i.d. 2.5 mm, length 3 m) and a carbon unibead column (i.d. 2 mm, length 35 cm) with a TCD. NO and NO2 were analyzed using a chemiluminescence NOx analyzer (Eco Physics, CLD62). The catalytic performance of the catalysts in the H2-SCR reaction was assessed based on the NOx conversion, H2 conversion, and N2 selectivity, which were calculated using the following equations:
where c represents concentration, and the subscripts “in” and “out” denote inlet and outlet, respectively. Because NO is partially oxidized to NO2 in the gas phase by O2 at room temperature both before reaching and after leaving the catalyst bed, the conversion of NOx rather than that of NO was chosen to evaluate the H2-SCR activities of the catalysts.
Fig. 1 shows the NOx conversion, H2 conversion, and N2 selectivity at each reaction temperature over the Pt/γ-Al2O3, Pt/MgO, Pt/ZrO2, and Pt/HZSM-5 catalysts pretreated with air. The maximum NOx conversion (71.4%) was obtained over Pt/HZSM-5 at 120 °C, but almost no NOx conversion was observed over Pt/MgO at 130-250 °C. The activity of Pt/ZrO2 was slightly better than that of Pt/γ-Al2O3, but both had inferior activities to that of Pt/HZSM-5. The order of H2 conversions over the four Pt catalysts was Pt/HZSM-5 ≫ Pt/γ-Al2O3 > Pt/ZrO2 > Pt/MgO. Over Pt/MgO, almost no H2 consumption was observed below 200 °C. The NOx and H2 conversions achieved using Pt/MgO, Pt/γ-Al2O3, and Pt/ZrO2 pretreated with 1.3% H2/N2 were much larger than those obtained using the corresponding air-pretreated catalysts (Fig. 2). For Pt/HZSM-5, the pretreating with 1.3% H2/N2 gave quite a limited change in the catalytic performance. As the supports do not, in principle, undergo reduction during the reducing treatment, the increased catalytic activity caused by the reducing treatment must arise from the reduction of Pt4+ to Pt2+ and/or Pt0. The catalytic activity changes caused by the reducing treatment should therefore indicate the amount of high-valent Pt in the catalysts before reduction. This means that Pt is primarily present as Pt4+ in the catalysts that showed increased activities after the reducing treatment. For the four Pt catalysts before reduction, the order of the amounts of high-valent Pt is Pt/MgO > Pt/γ-Al2O3 > Pt/ZrO2 > Pt/HZSM-5, based on the activities of each catalyst before and after the reducing treatment (Figs. 1 and 2). This order is supported by the XPS results (Fig. 3), which show that Pt is primarily presents as PtIVO2 (with binding energy peaks 4f7/2 = 74.5 eV and 4f5/2 = 77.1 eV) in the Pt/MgO catalyst, whereas it is mainly present as PtIIO (with binding energy peaks 4f7/2 = 72.1 eV and 4f5/2 = 75.0 eV) in the Pt/ZrO2 catalyst [32]. Because the Pt 4f5/2 line overlaps with the Al 2p (74.3 eV) line, no clear result regarding the Pt valence could be obtained for Pt/γ-Al2O3 and Pt/HZSM-5 at Pt loadings of 0.5 wt%.
In pyridine-adsorption studies (Fig. 4), HZSM-5 had a much stronger band at 1543 cm−1 than the other three supports because pyridine was adsorbed on the Brönsted acidic sites. In contrast, pyridine adsorbed on Lewis acidic (band at 1450 cm−1) or Brönsted acidic sites was not observed for the MgO support at the FTIR-detectable level. For the ZrO2 and Al2O3 supports, only weak bands assigned to Lewis acidic sites were detected. In contrast, in CO2-TPD studies (Fig. 5), for the MgO and γ-Al2O3 supports, particularly MgO, large CO2 desorption peaks appeared at 100-980 °C, indicating a large number of basic sites on the support surface. These results show that the HZSM-5 support has an acidic surface, whereas the MgO support has a basic surface. Although the γ-Al2O3 support has some Lewis acidic sites, its surface is essentially basic. In the case of the ZrO2 support, there are only limited numbers of both acidic and basic sites on its surface.
It has been reported [25, 33, 34] that Pt loaded on acidic supports is less oxidized than that on basic ones. Based on this information and the surface properties of the supports indicated by the characterization results, it can be deduced that the ratios of Pt metal to oxides in the four as-prepared catalysts follow the order Pt/HZSM-5 > Pt/ZrO2 > Pt/γ-Al2O3 >Pt/MgO. This agrees well with the order deduced from the H2-SCR activity changes of the catalysts after reducing treatment.
Fig. 6 shows the HRTEM images of the catalysts pretreated with 1.3% H2/N2. The average Pt particle sizes were estimated to be 1.7, 2.2, and 3.3 nm, respectively, for the Pt/MgO, Pt/γ-Al2O3, and Pt/HZSM-5 catalysts. Pt particle agglomeration occurred in the Pt/ZrO2 catalyst, and the Pt particles were about 5 nm in diameter. The smaller average Pt particle sizes for Pt/MgO and Pt/γ-Al2O3 compared with Pt/HZSM-5 and Pt/ZrO2 indicate that Pt dispersions on the MgO and the γ-Al2O3 supports were better than those on the HZSM-5 and ZrO2 supports. This is ascribed to the strong interactions between the acidic Pt precursor, i.e., H2PtCl6, and the basic sites on the supports during catalyst preparation. The small size of the Pt particles on the basic support, resulting from stronger interactions with the support surface, may be one of the reasons for there being hardly any Pt metal under oxidizing conditions. This explains the inferior activity of the Pt/MgO catalyst in H2-SCR well. Although its activity was enhanced by reducing treatment, the activity of the Pt/MgO catalyst clearly deteriorated with time on stream, as shown in Fig. 2(a) (the second run).
As shown in Fig. 1, the N2 selectivity over Pt/HZSM-5 was much higher than that over Pt/ZrO2. No reliable N2 selectivity data were obtained using the as-prepared Pt/γ-Al2O3 and Pt/MgO catalysts because the NO conversions were too low to calculate. After reducing treatment with 1.3% H2/N2, the N2 selectivities achieved with these two catalysts were much lower than those with Pt/ZrO2 and Pt/HZSM-5 (Fig. 2). It is worth noting that the N2 selectivity changed with reaction temperature according to a “volcano” shape, as did the NO conversion. The Tm, at which the volcano peak appeared, was close to that at which H2 was almost completely consumed for each catalyst. This means that if the amount of active H is insufficient, lower N2 selectivity is achieved by the catalysts, whether the H insufficiency results from a low ability of the catalyst to activate H2, or from lack of H2 for the H2-SCR. The maximum N2 selectivity is therefore obtained when the largest amount of active H with respect to nitrous species covers the catalyst. This explains why the maximum N2 selectivity and the maximum NOx conversion appeared at the same temperature. The ability of the catalyst to activate H2 is therefore an important factor determining N2 selectivity in the H2-SCR. As discussed in Section 3.1, the support acidity affects the dispersion and chemical state of Pt on the support, and this in turn influences the ability of the catalyst to activate H2. The support acidity therefore also influences the N2 selectivity via its impact on the ability of the catalyst to activate H2. This could be why the order of the maximum N2 selectivities is Pt/HZSM-5 > Pt/ZrO2 > Pt/γ-Al2O3 ~ Pt/MgO.
A basic support surface favors the adsorption of acidic nitrogen oxides, therefore more nitrate and nitrite species may be produced under the same reaction conditions on MgO and γ-Al2O3 than on ZrO2 and HZSM-5. This is confirmed by Fig. 7, which shows the steady surface FTIR spectra of the four supports exposed to gas mixtures of 1000 ppm NOx + 10% O2 in N2 at 100 °C. Clearly, the IR absorption bands attributed to nitrous species [8, 35, 36, 37] were much stronger for MgO and γ-Al2O3 than for ZrO2 and HZSM-5. These results are consistent with the order of the times, i.e., 105, 60, 30, and 15 min on MgO, γ-Al2O3, ZrO2, and HZSM-5 wafers, respectively, required to reach adsorption balance in the FTIR study. All the results show that the NOx adsorption capacities of the MgO and γ-Al2O3 supports were much larger than those of ZrO2 and HZSM-5. A combination of these results with those for N2 selectivity over the corresponding catalysts (Fig. 2) suggests that a higher NOx adsorption capacity of the support is unfavorable for high N2 selectivity; this implies that the nitrous species on the supports were involved in the H2-SCR.
Given that the supports themselves are inert in H2 activation at low temperatures, as shown by the result that almost no H2 was consumed over the supports below 150 °C under H2-SCR conditions, the active H required to reduce the nitrous species on the supports must be provided by Pt particles via H spillover. This means that only some of the nitrous species on the supports take part in reduction, as the amount of active H supplied in this way is limited. Costa et al. [22] proposed that in H2-SCR over Pt/MgO-CeO2, only the nitrous species at the metal-support interface contributed to the reaction. To estimate the amount of nitrous species on the supports that take part in the H2-SCR for our catalysts, steady in situ surface FTIR spectra of the catalysts (pretreated with 1.3% H2/N2) were obtained by exposing the catalysts to a gas mixture containing NOx (1000 ppm) + H2 (0 or 5000 ppm) + O2 (10%) + N2 at the temperatures at which the respective catalysts gave their largest NO conversion in H2-SCR (Fig. 8). It is assumed that the decrease in the amount of nitrous species on each catalyst caused by H2 introduction into the gas mixture corresponds to the amount reduced at the Pt-support interface. This should be well reflected by the difference between the FTIR surface spectra with and without H2 in the gas mixture, because the in situ surface FTIR spectra (Fig. 8) of the catalysts were essentially the same as those of their supports (Fig. 7). Figure 8 shows that the amount of nitrous species involved in the H2-SCR changes in direct proportion with the NOx adsorption capacity of the support.
Catalysts with weak H2-activating abilities and large NOx adsorption capacities provide smaller amounts of active H, so few nitrous species are reduced to N2 at the Pt-support interface. This leads to more N2O and less N2 formation in the H2-SCR. In our study, extreme situations occur over Pt/MgO and Pt/γ-Al2O3 catalysts. This explains why the N2 selectivities were so low for Pt/MgO and Pt/γ-Al2O3 compared with Pt/ZrO2 and Pt/HZSM-5. Conversely, because of the stronger ability of the Pt/HZSM-5 support to activate H2 and the lower NOx adsorption capacity of the support, the Pt/HZSM-5 catalyst reduced fewer nitrous species with larger amounts of active H at the Pt-support interface. It must therefore produce more N2 and less N2O there, leading to higher N2 selectivity over the catalyst.
For the Pt/HZSM-5 catalyst, for the same reasons, significant amounts of NH3 were produced by reduction of nitrous species; this is supported by the strong band at 1454 cm−1 assigned to NH4+ ions [25, 26] on the catalyst (Fig. 8). The reduction of NO + O2 by NH4+ ions at Brönsted acidic sites exclusively produces N2 [25, 26], which is another reason for the high N2 selectivity obtained over the acidic Pt/HZSM-5 support.
The dispersion and chemical state of Pt is strongly influenced by the acidity/basicity of the support. The H2-SCR activity and N2 selectivity of a supported Pt catalyst are therefore greatly influenced by the acidity/basicity of the support. On HZSM-5, a representative acidic support, Pt primarily exists in the metallic state, making the supported Pt catalyst much less active at low reaction temperatures before reducing treatment. Moreover, the acidic properties of HSZM-5 result in little NOx being adsorbed on the support surface. The strong ability of the Pt/HSZM-5 catalyst to activate H2, together with the lower NOx adsorption capacity of the HSZM-5 support, leads to reduction of nitrous species at the Pt-support interface and the production of more N2 and less N2O; this is the main reason why Pt/HSZM-5 gives much higher N2 selectivity in H2-SCR. The high H2-SCR activity and N2 selectivity of Pt/HZSM-5, resulting from the strong acidity of the support, are also associated with favorable NH3 adsorption, enabling NO + O2 reduction by NH4+ at Brönsted acidic sites to take place. Conversely, MgO and γ-Al2O3, representative basic supports, do not favor the presence of Pt in the active state. Their large NOx adsorption capacities and low H2-activationg abilities are the primary reasons for the low N2 selectivities of the Pt/MgO and Pt/γ-Al2O3 catalysts in SCR.
2015, Vol. 36 Issue (2): 197-203
PDF (6373 KB)
2015, Vol. 36 Issue (2): 197-203
PDF (6373 KB)