4.1 Structural Characterization of x%LaMnO3/Fe2O3
In order to study the physical state and morphology of the catalyst, XRD analysis was conducted, and the results were shown in Fig. 6a. 30%LaMnO3/Fe2O3 shows strong diffraction peaks at 2θ = 30°, 35°, 43°, 53°, 57°, 64° and 75°. 40%LaMnO3/Fe2O3 shows strong diffraction peaks at 2θ = 30°, 32°, 35°, 43°, 45°, 53°, 57°, 64°, 68° and 75°. By comparing the diffraction curve with the standard PDF, it is found that after doping a certain amount of LaMnO3, the diffraction peaks of LaMnO3 and Fe2O3 appear at the same time. Compared with 30%LaMnO3/Fe2O3, the diffraction peak of Fe2O3 at 2θ = 30° and 64° becomes weaker and narrower, while the diffraction peak of LaMnO3 appears at 2θ = 32° and 45°, indicating that LaMnO3 in crystal form is formed on the catalyst surface. 30%LaMnO3/Fe2O3 only has the diffraction peak of Fe2O3 but no other peaks, indicating that the active substance is distributed evenly on the catalyst surface and may exist in an amorphous form, which has a certain promotion effect on the denitrification activity of low-temperature SCR. Compared with 40%LaMnO3/Fe2O3, the LaMnO3 diffraction peak is stronger and the peak width is narrower when 2θ = 32°, 45° and 68°, which indicates that the distribution uniformity of the active substance on the catalyst surface becomes worse, this may be the reason for the poor catalytic denitration efficiency.
The XRD patterns of 40%LaMnO3/Fe2O3 catalysts with different Ce doping amounts are shown in Fig. 6b. It can be seen that 40%La0.9Ce0.1MnO3/Fe2O3 presents strong diffraction peaks at 2θ = 30°, 32°, 35°, 43°, 45°, 53°, 57°, 64° and 75°. Compared with 40%LaMnO3/Fe2O3, the intensity of partial diffraction peaks gradually becomes weaker and the peak width becomes wider. The results showed that LaMnO3 was evenly distributed in the catalyst with Ce element.
In samples 40%La0.8Ce0.2MnO3/Fe2O3 and 40%La0.7Ce0.3MnO3/Fe2O3, diffraction peaks of composite oxides are almost not detected, indicating that the lattice of LaMnO3 composite oxides is damaged after the addition of Ce element. In addition, the LaMnO3 diffraction peak of 40%La0.6Ce0.4MnO3/Fe2O3 curve disappears completely, which indicates that with the increase of Ce doping amount, the active component gradually changes from crystal form to amorphous form, which is conducive to the highly uniform dispersion of active substance on the carrier. It is also explained that Ce doping can improve the catalytic activity of catalysts.
The specific surface area and pore structure of the catalyst were analyzed as shown in Table 1. With the increase of LaMnO3 loading, the specific surface area increased from 33.601m2/g to 54.720m2/g, pore volume increased from 0.110m3/g to 0.145m3/g, and pore size increased from 100.692 nm to 158.273nm. The increase of specific surface area is conducive to the increase of contact area between catalyst and reaction gas, which is conducive to the progress of SCR reaction. In addition, with the increase of pore volume, the internal diffusion resistance of the reaction gas in the catalytic reaction process is relatively reduced, which is conducive to the adsorption of the reaction gas on the catalyst surface, thus improving the SCR reaction rate.
The specific surface area and pore structure of the catalyst were analyzed as shown in Table 2. With the increase of Ce doping amount, the specific surface area of the catalyst increased from 48.452m2/g to 53.510m2/g, pore volume increased from 0.114m3/g to 0.144m3/g, and pore size increased from 136.201nm to 136.580nm. The increase of specific surface area, pore volume and pore diameter is conducive to the adsorption and diffusion of gas within the catalyst, which also explains the doping of Ce element is conducive to the improvement of denitrification efficiency of the catalyst to a certain extent.
Table 1
BET data of x%LaMnO3/Fe2O3
x%
|
Specific surface area
|
Pore volume
|
Hole diameter
|
0%
|
33.601
|
0.110
|
100.692
|
10%
|
39.157
|
0.120
|
106.459
|
20%
|
41.146
|
0.138
|
103.475
|
30%
|
43.482
|
0.140
|
129.025
|
40%
|
48.452
|
0.144
|
136.201
|
50%
|
54.720
|
0.145
|
138.273
|
Table 2
BET data of 40%La1 − xCexMnO3/Fe2O3
40%La1 − xCexMnO3/Fe2O3
|
Specific surface area
|
Pore volume
|
Hole diameter
|
0
|
48.452
|
0.114
|
136.201
|
0.1
|
48.406
|
0.114
|
136.347
|
0.2
|
49.551
|
0.117
|
137.509
|
0.3
|
51.151
|
0.132
|
137.330
|
0.4
|
53.510
|
0.144
|
136.580
|
Figure 7a shows the surface morphology of Fe2O3 magnified 10K times. It can be seen that there are more spherical particles on the surface of Fe2O3, some particles are agglomerated, and there are fewer voids. Figure 7b shows the 40%LaMnO3/Fe2O3 catalyst. It can be seen that after the catalyst is loaded with LaMnO3, the voidage increases, the surface is relatively loose, and the particle uniformity improves, but the particle agglomeration phenomenon still exists locally. Figure 7c shows 40%La0.6Ce0.4MnO3/Fe2O3. Compared with Fig. 7b, the surface diameter is significantly larger, and the catalyst surface is loose and porous. The increase of specific surface area and the uniform distribution of particles are conducive to the contact, adsorption and activation of the catalyst surface with the reaction gas, which is favorable for the catalytic reaction process of NH3-SCR. Although 40%La1 − xCexMnO3/Fe2O3 has a high efficiency in catalytic denitrification experiments, it can be found that the local surface particle agglomeration and voidage are less, and the active substances cannot be uniformly loaded on the catalyst surface, which also indicates that the impregnation method for catalyst needs to be improved.
The citric acid impregnation method has the advantages of high utilization rate of active components, low cost and simple production method, but the active components are easy to migrate. Therefore, it is necessary to investigate the load of active substances. As can be seen from the Fig. 8, when the load is 10%, 20% and 30%, the actual load is 9.408%, 19.370% and 28.608%. When the load was 40%, the actual load was 37.029%, while when the load was 50%, the actual load was only 42.821%. The results indicated that citric acid impregnation method was not suitable for the preparation of catalyst with large loading capacity. The oxides of other trace elements, such as ZrO2, SO3 and P2O5, were also detected in the prepared catalyst, and the sum of oxide content was less than 3%.
4.2 Catalytic Performance of 40%La1-xCexMnO3/Fe2O3 When Reducing NO
The H2-TPR distribution of 40%La1 − xCexMnO3/Fe2O3 is shown in Fig. 9. Three reduction peaks appear in all three catalysts, which are related to the continuous reduction of Fe2O3→Fe3O4→FeO→Fe.44–46 The reduction behavior of Fe2O3 changed obviously after loading active substance. The reduction peak area increased at 402℃, and a new reduction peak appeared at 370℃. This is because the addition of Ce provides a new active site on the catalyst surface. According to the integration, as the doping amount of Ce increases from 0.3 to 0.4, the area of the first reduction peak and the second reduction peak of the catalyst gradually increases and moves towards low temperature (the first reduction peak :402℃→398℃→391℃, the second reduction peak:540℃→538℃→512℃). In addition, 40%La0.6Ce0.4MnO3/Fe2O3 reduction peak corresponds to the lowest desorption temperature. This indicates that 40%La0.6Ce0.4MnO3/Fe2O3 has the strongest reducing ability at low temperature. The above results show that LaMnO3 can improve the catalytic denitrification capacity of the catalyst, and the addition of Ce can improve the catalytic capacity of catalyst at low temperature. This further explains why the catalytic effect of 40%La0.6Ce0.4MnO3/Fe2O3 is higher than other catalysts.
The NH3-TPD of La1 − xCexMnO3/Fe2O3 is shown in Fig. 10. As we all know, the thermal stability of NH3 adsorbed at Brønsted acid site is lower than that of NH3 coordinated with LAS. Thus, desorption peaks below 300 ℃ are attributed to the desorption of NH4+ from the Brønsted acid site(BAS), while peaks above 300 ℃ are attributed to the desorption of NH3 coordinated with Lewis acid(LAS). It can be found from the spectrum that LaMnO3/Fe2O3 has three desorption peaks at 248℃, 415℃ and 440℃. La0.8Ce0.3MnO3/Fe2O3 and La0.7Ce0.4MnO3/Fe2O3 have two desorption peaks, and the acid distribution is typical bimodal ammonia desorption curve. The acid distribution is typical bimodal ammonia desorption curve. The desorption peak of LaMnO3/Fe2O3 at 248℃ belongs to the weak acid position, while the desorption peak at 415℃ and 440℃ belongs to the strong acid position. When the Ce doping amount was 0.3, two strong acid sites became one strong acid site (445℃). When the doping amount continued to increase (x = 0.4), the desorption peak attributed to the strong acid site shifted significantly (450℃), and the peak area of both weak acid site and strong acid site increased. This indicates that Ce doping can change the intensity and quantity of Brønsted acid sites and Lewis acid sites of catalysts. The La0.6Ce0.4MnO3/Fe2O3 Lewis acid site is stronger, which also explains why the catalytic denitration efficiency of La0.6Ce0.4MnO3/Fe2O3 is the highest among the three catalysts.
4.3 Chemisorption Properties of 40%La0.6Ce0.4MnO3/Fe2O3
In situ DRIFT spectra of NH3 adsorption and activation on the catalyst surface changing with time are shown in Fig. 11a. At 50℃, the adsorption peaks of NH3 on the catalyst surface mainly appeared at 3684, 1760, 1540, 1419, 1355 and 1250cm− 1. 45min after NH3 was introduced, the vibration peaks at 1540 cm− 1 and 1250 cm− 1 were weakened. The vibration peaks at 1540 cm− 1 and 1250 cm− 1 are attributed to the antisymmetric variable Angle vibrations (δas(NH3)) and symmetric variable Angle vibrations (δs(NH3)) of the coordination NH3 adsorbed at the Lewis acid site (LAS). 26 According to the vibration peak area of the spectrum, the surface of catalyst is mainly LAS. The peak at 3684 cm− 1 was attributed to the hydroxyl group produced at the Brønsted acid site (BAS) on the catalyst surface. Peaks at 1760 cm− 1 and 1463 cm− 1 were attributed to NH4+ species bound to BAS. 47–49 Peaks at 1540, 1362 and 1250 cm− 1 were attributed to NH3 species adsorbed to LAS. According to the integration, the peak area at 1760 cm− 1 and 1463 cm− 1 gradually increased after NH3 began to flow into the area, while the vibration peaks showed a weakening trend after N2 purge. Among them, the vibration peak at 1760 cm− 1 showed a larger weakening trend, while the peak at 1463 cm− 1 showed little weakening trend. This indicates that the BAS at 1463 cm− 1 is stronger than that at 1760 cm− 1. The adsorption peak at 1362 cm− 1 is attributed to NH2, the active intermediate product produced by adsorption and dehydrogenation of NH3. 50–52 The adsorption peak at 3684 cm− 1 is due to the formation of NH4+ by the consumption of surface acidic hydroxyl by NH3, and the resulting adsorption peak at 1463 cm− 1 belongs to the binding of NH4+ at the Brønsted acid site. Infrared spectra of quasi-steady-state adsorption of NH3 on the catalyst surface at 50℃-400℃ is presented in Fig. 11b. It can be seen that as the temperature goes from 50℃ to 400℃, the infrared adsorption peak at 1362 cm− 1 which is attributed to the LAS starts at 150℃, and the peak area is basically unchanged, while the adsorption peak at 3684, 1760, 1540 and 1463 cm− 1 decreases at different rates. This indicates that the adsorption of acid sites at 1362 cm− 1 is relatively stable, while the strength of other sites is not strong. The coverage of NH3 molecules at acid sites on the surface deteriorates with the increase of temperature.
It has been reported that the adsorption capacity of NO and O2 together on Fe2O3 at low temperature is stronger than that of NO alone.26 In situ DRIFTS spectra of transient adsorption of NO and O2 on the catalyst surface at 50℃ is presented in Fig. 12a. At 50℃, the main infrared adsorption peaks on the catalyst surface are 1772, 1548, 1424 and 1350 cm− 1. It can be seen that when NO and O2 were first introduced into the catalyst surface, obvious infrared adsorption peaks appeared at 1424 cm− 1 and 1350 cm− 1, indicating that the acid site was adsorbed at a faster rate. The adsorption capacity of NO and O2 at 1548, 1424 and 1350 cm− 1 increased with the increase of catalyst adsorption time. According to reports, the band 1622 − 1548 cm− 1 is mainly produced by NO and lattice oxygen in the catalyst, which belongs to the typical N-O antisymmetric stretching vibration region, belonging to nitrate or adsorption state NO2.53–57 1548 cm− 1 was attributed to the vas(NH3) vibration mode of the v3 cleavage of the undented nitrate, and the infrared peaks at 1424 and 1350 cm− 1 were attributed to the undented nitrate and nitro compounds.58,59 It can be seen that the main adsorption products of NO and O2 on the surface are single-tooth nitrate and nitro compounds. A small infrared peak appeared at 1770 cm− 1 about 30min after gas injection, which was attributed to Fe2+-N = O species. After N2 purge, strong adsorption peaks were still found at 1424 and 1350 cm− 1, which may be because when N2 purge the catalyst surface, NO was oxidized by Fe3+ to generate NO2, which returned to the catalyst surface to generate nitro, resulting in enhanced peak at 1424 cm− 1. In situ DRIFTS spectra of quasi-steady-state adsorption at programmed temperature on the catalyst surface is presented in Fig. 12b. The main products of catalyst surface adsorption were nitrate (1424 cm− 1) and nitro compound (1350 cm− 1). It can be seen that the characteristic peaks of nitrate (1424 cm− 1) and nitro compounds (1350 cm− 1) showed an increasing trend with the increase of temperature, indicating that more nitrate substances were generated after further oxidation of nitrite and nitro. The characteristic peak area at 1548 cm− 1 decreases significantly at 100℃, indicating that the weakly adsorbed single-tooth nitrate and adsorbed NO2 produce desorption on the catalyst surface. This may be due to the involvement of NO2 in the rapid SCR reaction, resulting in a significant reduction. There is literature indicating that when NO2 is present, nitric acid will be generated to further generate NH4NO3. Under the action of surface acidic sites, NH4NO3 will be reduced by NO to NH4NO2, which is extremely unstable and can be decomposed into N2 and H2O at 100 ℃, which is a rapid SCR reaction60. Therefore, for this catalyst, strong oxidizability can lead to the production of NO2 and promote the SCR reaction process.
The reaction of NH3 with NO and O2 at 240℃ is shown in Fig. 13. The main infrared adsorption peaks on the catalyst surface at 270℃ are 3265, 1781, 1540, 1412 and 1355 cm− 1. It can be seen that, with the influx of NO and O2, ammonia species are consumed. Meanwhile, the peak value at 1540 cm− 1 increases at 40min of reaction, while the peak value at 1412 cm− 1 decreases. The peak at 1540 cm− 1 was monodontic nitrate, which belonged to nitrate species. The peak at 1412 cm− 1 was attributed to NH4+. The peak at 3265 cm− 1 was attributed to NH3 adsorbed at the L acid site. Although there was the presence of monondonate nitrate in the spectrum, no characteristic peak of bridging nitrate was detected, indicating that the pre-adsorbed NH3 was not completely consumed and NH3 still participated in the reaction on the catalyst surface. It was found from the spectrum that the thermal stability of the Lewis acid site was still stronger than that of the Brønsted acid site under the two different pre-adsorbed SCR reactions, so the enhancement of the Brønsted acid site played a key role in improving the catalytic activity at low temperature.61
4.4 SO2-tolerance Mechanism of 40%La0.7Ce0.4MnO3/Fe2O3
In order to understand the formation of ammonium sulfate salts on the catalyst surface in SO2 and H2O environment, TG characterization was performed on fresh and 40%La0.6Ce0.4MnO3/Fe2O3 catalysts after sulfur toxicity resistance at 240℃, and the results were shown in the Fig. 14. HAADF image and EDS mappings of used Fe2O3 and used 40%La0.6Ce0.4MnO3/Fe2O3 are as show in Fig. 15. It can be seen that compared with the fresh catalyst, the mass of the poisoned catalyst gradually loses during the heating process, indicating the presence of sulfate on the surface of the catalyst. Previous studies have found that there are two ways for SO2 to cause catalyst poisoning. On the one hand, NH4HSO4 generated by SO2 reacts with NH3 in the environment of H2O is attached to the catalyst surface, resulting in the active site being covered; on the other hand, SO2 can directly generate relatively stable metal sulfate with metal oxides.61,62
Weightlessness of catalyst can be divided into three steps. It is attributed to the loss of H2O on the catalyst surface at 50–200℃.63 In the second step, 200–500℃ is attributed to the decomposition of (NH4)2SO4 and NH4HSO4.64 In the third step, 500–800℃ is attributed to the decomposition of Fe2 (SO4)3. It can be seen that the mass loss of used catalysts is significantly increased compared with that of fresh catalysts. In the second stage, the weight loss of 40%La0.6Ce0.4MnO3/Fe2O3 catalyst before and after the experiment is ~ 0.62%, while the weight loss of Fe2O3 catalyst before and after the experiment is ~ 0.68%, indicating that compared with pure Fe2O3, the surface of 40%La0.6Ce0.4MnO3/Fe2O3 catalyst has less sulfate. At 500 ~ 800℃, the weight loss difference of Fe2O3 catalyst before and after use decreases gradually, indicating that Fe2 (SO4)3 decomposition occurs at this temperature. The weight loss degree of used 40%La0.6Ce0.4MnO3/Fe2O3 catalyst is lower than that of used Fe2O3 catalyst, indicating that there is more Fe2 (SO4)3 on the surface of supported catalyst. On the other hand, in the third stage, the weight loss rate of 40%LaMnO3/ Fe2O3 catalyst before and after the experiment is lower than 40%La0.6Ce0.4MnO3/ Fe2O3 catalyst, indicating that the doping of Ce element improves the formation of Fe2 (SO4)3, this is consistent with simulation results from previous studies.65