Nitrogen oxides (NOx) emitted from fossil fuel combustion are major air pollutants that cause many environmental and human health hazards, including acid rain, photochemical smog, ozone depletion, and the greenhouse effect [1]. Currently, the most widely used method for NOx removal in flue gas is selective catalytic reduction (SCR) with a V-W-Ti catalyst. This technology has high efficiency and stability for NOx removal, but there still exist many problems, e.g., high reaction temperature (300-400 °C) needed has limited its application. The requirement to add ammonia or urea increased system complexity and operation cost, and there is the possibility of secondary pollution caused by ammonia slip [2].
Extensive research has been conducted to solve these problems. Recently, NO catalytic oxidation (SCO) has begun to attract more attention. Over 90% of NOx formed in combustion is NO, and SCO is a process to convert most of the NO in flue gas to NO2, which is chemically more active and easier to capture in a later stage. This technology has two main applications: (1) as an intermediate step for the SCR reaction to improve the reaction rate and utilization of ammonia [3, 4, 5]; (2) because NO2 is more water soluble than NO, it is possible to vary out the simultaneous adsorption of SOx and NOx in the flue gas desulfurization plant [6, 7].
In early research, Pt-based catalysts were shown to have good performance for the SCO [8, 9, 10], but its high cost and easy poisoning by SO2 and H2O have limited its commercial application. Recently, it was demonstrated that transition metal oxides possess good catalytic activity in NO oxidation [11, 12, 13, 14]. Among these, Mn-based catalysts are receiving more attention because of their high catalytic activities, low prices, and environmental friendliness. Their multivalent chemical states and porous physical structures make MnOx a good candidate catalyst for NO oxidation. The performance of both unsupported and supported Mn-based catalysts have been investigated, including Mn-Co-Ce-Ox [15, 16], Mn-Ce [17] Mn/Ti [14, 18, 19, 20], Mn-Co/Ti [21], Mn/AC, Mn/C [22], and Fe-Mn/Ti [18], as well as the effects of different experimental conditions. Compared to unsupported catalysts, the supported one can improve its thermal stability and the dispersion of the active sites [23]. The nature of the support has a significant impact on the performance of the catalyst.
Existing research on supported Mn-based catalysts for NO oxidation has mainly focused on TiO2 supported MnOx, which was demonstrated to give high catalytic performance. It is well known that the crystal phase of the TiO2 affects catalyst activity [24], but its role in the NO catalytic oxidation has not been reported yet. In this paper, Mn catalysts supported on TiO2 with three different crystalline phases were prepared to investigate the influence of the TiO2 crystalline phase on the activity of MnOx/TiO2 for NO catalytic oxidation. The catalysts were characterized by N2 adsorption-desorption measurement, field emission scanning electron microscopy (FE-SEM), X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), H2 temperature-programmed reduction (H2-TPR), and O2 temperature-programmed desorption (O2-TPD) to understand the interactions between MnOx and TiO2.
All chemicals used were analytical reagent. Commercial nanosized TiO2 was used in this study. A 25 nm-anatase TiO2 and 25 nm-rutile TiO2 were supplied by Aladdin Industrial Inc. of China, and the 25 nm mixed crystal Type P25 TiO2 (20% rutile and 80% anatase) was provided by Evonik Degussa GmbH of Germany.
TiO2-supported 10 wt% Mn catalysts were prepared by the impregnation method. An appropriate amount of Mn(CH3COO)2·4H2O (Aladdin Industrial Inc.) was dissolved in 150 ml deionized water and heated to and kept at 80 °C. After TiO2 of the correct mass was added into the solution, water was slowly evaporated with electromagnetic stirring. The paste obtained was dried at 110 °C overnight, then crushed and sieved to obtain 40-60 mesh particles. Finally, the sample was calcined in air at 400 °C for 6 h. The Mn catalysts supported on P25 TiO2, anatase TiO2, and rutile TiO2 were labeled MnOx/ TiO2(P), MnOx/TiO2(A), and MnOx/TiO2(R), respectively.
NO catalytic oxidation measurement was carried out in a fixed-bed reactor (quartz tube, inner diameter 33 mm). The gas flow was controlled by a mass flow meter. The reaction temperature was controlled by a programmed electric furnace. The reactor inlet gas composition was 0.05% NO, 6% O2, and N2 balance. The total gas flow rate was controlled at 1.5 L/min. Catalyst activities were tested at 100-400 °C and GHSV of 20000 h-1. Catalyst (5.0 g) was used in a typical test. The concentrations of NO and NO2 at the reactor outlet were monitored by a Fourier transform infrared spectroscopy (FT-IR) gas analyzer (Nexus 670). Because the inlet and outlet concentrations of NOx in all experiments were recorded after equilibrium reached, NO conversion was calculated from the ratio of outlet concentration of NO2 and the sum of outlet concentrations of NO and NO.
A variety of characterization methods were used to study the physiochemical properties of the MnOx/TiO2 catalysts. N2 adsorption-desorption measurement were carried out on a Micromeritics ASAP2010 automated N2 adsorption analyzer. The specific surface area of samples was calculated by the BET method, and the pore volume and pore diameter were measured by the BJH method. A JEOL S-4800 FESEM was used to investigate the surface morphology of catalysts. XRD was used to characterize the crystalline phase of the sample using a German Bruker D8 Advance diffractometer with Cu radiation source, scanning range of 10° to 90° (2θ), and a step size of 0.05°. XPS was used to analyze the atomic concentrations and Mn valence state on the catalyst surface on a British EscaLab 250Xi X-ray diffractometer using Al Kα (1486.6 eV, pass energy 30.0 eV) radiation as the excitation source. H2-TPR and O2-TPD were carried out on a CHEMISORB 2720 chemisorption analyzer with a thermal conductivity detector (TCD). Catalyst (100 mg) was used in the test. For the H2-TPR analysis, the sample was pretreated at 250 °C for 30 min under N2 (30 mL/min), and then cooled to room temperature before reduction by 5% H2/Ar (20 mL/min), and the temperature was increased from room temperature to 800 °C (10 °C/min). For the O2-TPD analysis, the sample was pretreated first at 250 °C for 30 min under N2 (30 mL/min). After cooling to room temperature, N2 was switch to high purity O2 until adsorbed to saturation, and then the gas was switched to He (20 mL/min). When the TCD signal was stable, the temperature was increased from room temperature to 1000 °C (10 °C/min).
Figure 1 shows the NO conversion on the series MnOx/TiO2. For reference, the activity of unsupported MnOx calcined from Mn(CH3COO)2·4H2O under air at 400 °C for 6 h is also presented. The sequence of catalyst activities for NO catalytic oxidation was MnOx/TiO2(P) > MnOx/TiO2 (A) > MnOx > MnOx/TiO2(R). MnOx/TiO2(P) gave the highest catalytic activity with the NO conversion of 83%. MnOx/TiO2(P) and MnOx/TiO2(A) had the same maximum activity temperature of 300 °C. For MnOx and MnOx/TiO2(R), the maximum temperature was 350 °C. This indicates that P25 TiO2 and anatase TiO2 significantly enhanced the low temperature NO catalytic oxidation activity of MnOx, while rutile TiO2 seemed to have no catalytic activity promotion at all. The lower catalytic activity of MnOx/TiO2(R) compared to MnOx may be due to the relatively low MnOx loading. Thus, it could be concluded that, compared to rutile TiO2, anatase TiO2 can better interact with MnOx on the catalyst surface, while a small amount of rutile TiO2 in P25 TiO2 further promoted this interaction.
The specific surface area, pore volume, and pore diameter of the MnOx/TiO2 catalysts and supports are shown in Table 1. The data showed that anatase TiO2 had the largest surface area and pore volume. The larger surface area and pore volume of the support provided more gas-catalyst contact and more active sites, thus improving the catalytic activity. Compared to the pristine supports, after loading MnOx on the surface of the three types of TiO2, the surface area of MnOx/TiO2(A) and MnOx/TiO2(R) were reduced by 16.3% and 26.3%, respectively. The pore volumes were also much reduced. In contrast, the specific surface area of MnOx/TiO2(P) was only reduced slightly from 46.3 to 45.3 m2/g, while the pore volume remained almost the same. This indicates that MnOx was better dispersed on P25 TiO2. The pore sizes of the three catalysts did not change much compared to those of the pristine supports, indicating that the change in specific surface area was mainly due to blockage of the pores.
The morphology of the catalyst surface was observed by SEM. As shown in Fig. 2, the surface of MnOx/TiO2(P) was rough and catalyst particles were distributed evenly without agglomeration, indicating that MnOx existed in a highly dispersed phase on the TiO2(P) surface after calcination, as implied in Table 1. The catalyst particles of MnOx/TiO2(A) appeared to be much agglomerated, resulting in a smaller specific surface area and pore volume. For the MnOx/TiO2(R) catalyst, no obvious agglomeration was found on the surface, but more bulk adhesion could be seen. The uneven distribution of MnOx reduced the specific surface area and pore volume of the TiO2(R) supported catalyst. Based on BET and SEM results, it can be concluded that the small amount of rutile TiO2 contained in P25 TiO2 prevented agglomeration of catalyst particles and improved the dispersion of MnOx on the TiO2 surface.
The XRD patterns of the MnOx/TiO2 catalysts and their supports are illustrated in Fig. 3. The peaks corresponding to the crystalline phases of all three TiO2 supports were clearly observed. After MnOx was loaded, their peak position and intensity did not change. No peak corresponding to MnOx was detected over any MnOx/TiO2 sample. The monolayer surface coverage is defined as the maximum amount of MnOx on the TiO2 support [25]. From the bond length of M-O in Mn2O3, the nominal quantity of Mn needed to cover the TiO2 surface as a monolayer was calculated to 0.0532 wt% per m2 of the support [26]. It was estimated that the Mn loading at monolayer coverage is 6.6 wt% for all the samples. As no MnOx was identified on the TiO2 surfaces with 10% Mn loading, MnOx was distributed homogeneously on the support surface as an amorphous phase or low crystallinity grains, or there was a relatively high proportion of Mn2O3 in the MnOx loaded on the TiO2 surface because Mn2O3 is highly amorphous in nature [27].
The Mn atomic concentration and Mn/Ti ratio measured by XPS are shown in Table 2. The ratio of Mn/Ti is a measurement of the dispersion of MnOx on the surface of TiO2 [25]. Although the nominal Mn loadings in all the catalysts were the same, their Mn/Ti ratios on the catalyst surface were different. MnOx/ TiO2(P) has the lowest Mn/Ti ratio of 0.20, MnOx/TiO2(A) has a Mn/Ti ratio of 0.22 and MnOx/TiO2(R) has the highest ratio of 0.83. This revealed again that MnOx on the last catalyst existed in a highly dispersed phase and was intercalated with TiO2 lattice, which agreed well with the BET, SEM, and XRD results. P25 TiO2 has a lower specific surface area and Mn/Ti ratio compared to TiO2(A), signifying that MnOx interacted stronger with P25 TiO2. The high Mn atomic concentration and Mn/Ti ratio on TiO2(R) surface may be due to two reasons: one is the low specific surface area and pore volume of TiO2(R), and the other is the weak interaction between TiO2(R) and MnOx, which only allowed MnOx to cover the TiO2(R) surface as isolated species that do not interact with the support. According to the XRD results, no MnOx crystalline phase was present on the TiO2(R) surface, indicating that its MnOx existed in the form of polymeric or microcrystalline crystallites, rather than in monolayer form. This observation is consistent with the results reported in the literature [25].
Figure 4 shows the Mn 2p XPS spectra of the surface of the MnOx/TiO2 catalysts. It can be deduced from the asymmetry of the Mn 2p3/2 and Mn 2p1/2 profiles that MnOx existed in two states, MnO2 and Mn2O3. The peak at lower binding energy (641.3-641.7 eV) was attributed to Mn3+ [28], and the peak at higher binding energy (642.2-643 eV) was attributed to Mn4+ [29]. The proportion of Mn3+ to Mn is shown in Table 2. The relatively high proportion of Mn2O3 verified the XRD results. The confirmation of the two different oxidation states of MnOx by XPS indicated a possible mechanism of NO oxidation on the MnOx/TiO2 catalysts. Previous studies have shown that the activity of MnOx is significantly influenced by the oxidation states and crystalline phases [30], and for the oxidation of NO, Mn2O3 has a higher catalytic activity than MnO2 [20, 31]. In the reaction of CO oxidation, Cimino et al. [32] attributed the higher catalytic activity of Mn3+ to the relatively weak Mn3+-O bond. Similarly, it can be speculated here that the easy-to-break Mn3+-O bond promoted the formation and desorption of NO2. It was reported that the species of MnOx present strongly depended on the properties of the support, and TiO2 contributed to the formation of Mn2O3 and MnO2 mixed oxides [33]. Among the three TiO2 supports, Mn3+ was present as a higher proportion on the surface of P25 TiO2, suggesting that P25 TiO2 was more conducive to the formation of Mn2O3, which thus promoted the catalytic activity of MnOx for NO oxidation.
For transitional metal oxides, how the oxygen combines with the metal ions affects much the catalyst activity. The O 1s XPS spectra of the three catalysts are shown in Fig. 5. All the O 1s profiles have three overlapping peaks. The first peak at 529.56-529.82 eV corresponded to lattice O or chemisorbed atoms O2- in the metal oxides [34, 35], the second peak at 530.59-531.05 eV, corresponded to hydroxyl [35], and the third peak at 532.0-532.54 eV corresponded to oxygen in MnOx [36]. Similar to the findings in the literature [34], lattice oxygen was the dominant oxygen species on the surface of P25 TiO2. The strong interaction between P25 TiO2 and MnOx may be the most important reason for improved MnOx catalytic activity in NO oxidation. The hydroxyl oxygen content on the catalyst surface seemed to play a minor role in NO oxidation as no obvious relationship with catalyst activity was found.
H2-TPR was used to investigate the reducibility of MnOx on the TiO2 surface. H2 consumption was compared using the areas of the TPR profiles (Fig. 6). MnOx/TiO2(A) has a larger H2 consumption than MnOx/TiO2(R) and MnOx/TiO2(P), indicating that the average valence state of Mn on anatase TiO2 was the highest, which was consistent with the XPS results. All the three catalysts showed a two-step reduction process. The first peak was located between 300 and 400 °C, and the second peak was located at 400-600 °C, which agreed with the results reported in the literature [25, 37]. MnOx was reduced in the following sequence: MnO2→Mn2O3→Mn3O4→MnO [38], and MnO reduction would not be seen under our experimental conditions [39]. Therefore, the low temperature reduction peak corresponded to the reduction of highly dispersed and easy reducible MnO2, and the high temperature peak corresponded to the reduction of Mn2O3 or bulk MnOx phase [33]. MnO2 loaded on the TiO2(A) and TiO2(P) had the same reduction temperature, which was lower than that of MnO2 loaded on TiO2(R). This indicates that MnO2 on TiO2(A) and TiO2(P) was more easily reduced. The reduction temperature of Mn2O3 correlated with the catalytic activity, i.e., the lowest Mn2O3 reduction temperature corresponded to the highest NO oxidation activity, implying that the crystalline phase of TiO2 affected the reducibility of Mn2O3 significantly. It was concluded from the reducibility of MnOx that TiO2(A) and TiO2(P) had a stronger interaction with MnOx than TiO2(R) did. In the interaction with MnO2, TiO2(A) and TiO2(P) were similar, but with Mn2O3, the interaction was in the order of TiO2(P) > TiO2(A) > TiO2(R). The decrease in reduction temperature means the promotion of oxygen mobility [20], as discussed previously that Mn2O3 played a more important role in NO oxidation. Therefore, the better reducibility of MnOx, especially Mn2O3, on the surface of P25 TiO2 explained why P25 TiO2 gave the highest performance of the MnOx catalyst for NO oxidation.
O2-TPD was used to investigate the oxygen species on the catalysts surface. As shown in Fig. 7, all the samples had a large capacity for the adsorption of oxygen species. A desorption peak appeared before 200 °C for all catalysts. As the temperature increased, all three catalysts showed oxygen desorption without obvious peaks. A second desorption peak appeared at 510 °C for MnOx/TiO2(R) and 590 °C for MnOx/TiO2(P) and MnOx/TiO2(A). Above 700 °C, all the catalysts showed a third desorption peak at similar temperature. The first peak at low temperature corresponded to physically adsorbed or weakly chemisorbed O2- species. The peak areas were consistent with the specific surface areas of three catalysts. The high surface areas of MnOx/TiO2(A) and MnOx/TiO2(R) explained the slow desorption of weakly chemisorbed oxygen species. The second peak corresponded to strongly chemosorbed O2- species. The third peak corresponded to the chemically stable lattice oxygen in MnOx and TiO2. Combining the maximum activity temperature and desorption temperature of the catalysts, it can be deduced that the strongly chemisorbed O2- as indicated by the second peak played an important role in the reaction. Compared to MnOx/TiO2(A) and MnOx/TiO2(P), the second peak of MnOx/TiO2(R) was broader and weaker, indicating that chemisorbed O2- was different from those on the surfaces of MnOx/TiO2(A) and MnOx/TiO2(P). The former was generated from the interaction between MnOx and O2, while the latter were generated from the interaction between MnOx, TiO2, and O2. The relatively sharp peak also indicated that chemisorbed O2- species on the surface of MnOx/TiO2(A) and MnOx/TiO2(P) was more easily desorbed. Although MnOx/TiO2(A) possessed a larger specific surface area than MnOx/TiO2(P), the larger second peak of MnOx/TiO2(P) showed that P25 TiO2 better enhanced the formation of chemisorbed O2-. During the reaction, easier desorption of O2- and a larger oxygen storage capacity were beneficial to the performance of the catalyst [21]. Combining the XRD, XPS, and H2-TPR results, it can be speculated that the oxygen corresponding to the second peak came from the relatively weak Mn3+-O bond. TiO2(A) and TiO2(P) were capable of enhancing the activity of the Mn3+-O bond, thus promoting the desorption of chemisorbed O2-. Obviously, the promotion by P25 TiO2 was much stronger. The strong interaction between P25 TiO2 and MnOx promoted the formation of Mn2O3. Furthermore, P25 TiO2 improved the dispersion of MnOx on its surface, which improved oxygen mobility accordingly [12].
Xu et al. [21] studied the reaction mechanism of NO oxidation on Mn-Co/Ti catalyst, which was confirmed by Wu et al. [20] on MnOx/TiO2. Xu et al. [21] found the formation of bridging nitrates and bidentate nitrates on the catalyst surface during the reaction, and they inferred the following reaction mechanism:
O2 + e- → O2- (1)
NO + O2- → NO3- (2)
NO3- + NO → 2NO2 + e- (3)
The generation of adsorbed NO2 was faster than that of the nitrate species, and the NO2 generated could be easily desorbed from the catalyst surface [20, 21]. Thus, reaction (2) is the rate limiting step of the overall reaction. P25 TiO2 favored the formation of easily desorbed O2-, and consequently increased the reaction rate and NO conversion.
The activity of Mn catalysts impregnated on TiO2 with different phases was in the order MnOx/TiO2(P) > MnOx/TiO2(A) > MnOx/TiO2(R). The decrease in the specific surface area and pore volume of TiO2 after impregnation were due to blockage of pores, and MnOx/TiO2(P) showed the smallest changes in specific surface area and pore volume. Compared to TiO2(A) and TiO2(R), TiO2(P) gave a higher dispersion of MnOx on the catalysts surface, reduced the agglomeration and adhesion of catalyst particles and produced more Mn2O3 species, which was more active for the oxidation of NO. In addition, TiO2(P) enhanced the reduction of MnOx, especially Mn2O3, and the formation of easily desorbed O2- generated from the Mn3+-O bond. The chemisorbed O2- promoted the formation of intermediate nitrates, thus increasing NO conversion.
2014, Vol. 35 Issue (1): 120-126
PDF (797KB)
2014, Vol. 35 Issue (1): 120-126
PDF (797KB)