1. Introduction
NH
3-SCR (selective catalytic reduction of NO
x with NH
3) with V
2O
5-WO
3/TiO
2 as the catalyst is used successfully in stationary plants [
1,
2,
3,
4,
5]. The support of choice is TiO
2 (anatase form) due to its higher surface area relative to rutile phase and the fact that SO
3 does not deteriorate the TiO
2 support. The commercial catalyst exhibits high selectivity and activity in the NH
3-SCR of NO at 300–400 °C [
3,
4,
5]. In order to operate at this temperature, the SCR unit is usually installed at a high dust position.
However, by placing the SCR unit at the high dust position, the catalyst’s life is shortened due to the high content of ash with alkali metals in the flue gas [
6,
7,
8]. Therefore, the tail-end position, which is located behind the SO
2/SO
3-removing unit is attractive. Decreased erosion and fouling at the low dust level also increases the catalyst’s lifetime [
9]. In order to avoid costly reheating of the flue gas to around 350 °C, tail-end placement necessitates the SCR catalyst to be significantly more active than the vanadia-tungsta based one.
In recent years, a large number of research articles on NH
3-SCR of NO at low-temperature have been published. Among the reported catalysts, Mn/TiO
2 based formulations are the most promising [
10,
11]. Furthermore, bimetallic Mn catalysts showed higher activities and selectivities. Hence, Mn-Fe/TiO
2 [
12,
13,
14,
15], MnOx-CeO
2 [
16], and Mn-Ce/TiO2 [
17,
18] have been reported to be highly active bimetallic catalysts for NH
3-SCR of NO at low temperatures. Recently, we reported highly active low temperature Mn-Fe/TiO
2 catalysts prepared by deposition-precipitation using ammonia carbamate as a precipitating agent [
19].
The low-temperature SCR activity of the MnO
x catalysts depends on the precursor, preparation method, and metal loading. Kapteijn et al. [
20] reported that a Mn/Al
2O
3 catalyst was more active when prepared with Mn-acetate than with Mn-nitrate. Likewise, Li et al. [
21] concluded that a Mn-acetate derived Mn/TiO
2 catalyst had better activity than its Mn-nitrate based version. However, Peña et al. [
22] showed that a Mn/TiO
2 catalyst prepared from manganese nitrate and calcined at 400 °C had better activity at lower temperatures than a catalyst obtained from manganese acetate. Detailed investigation of the precursor effect on more active formulations like Mn-Fe/TiO
2 and Mn-Fe-Ce/TiO
2 catalysts has not been reported.
The optimum Mn loading of low temperature Mn/TiO
2 catalysts was reported to be 20 wt.% [
22] while the optimum loadings of Mn and Fe in Mn-Fe/TiO
2 catalysts synthesized by impregnation were both 10 wt.% [
14]. In our previous article [
19], we reported that it was possible to further reduce the total metal loading on Mn/TiO
2 catalysts from 20 wt.% to 5 wt.% with a change in the method of synthesis from conventional impregnation to deposition, while the total metal loading of the Mn-Fe/TiO
2 catalysts could be reduced from 35 wt.% to 25 wt.%. Catalysts based on Mn-Fe/TiO
2 contain high amounts of active metals (about 20–25 wt.%) compared to the traditional V
2O
5-WO
3/TiO
2 system (about 7–10 wt.%). Additionally, unsupported manganese oxide in hollandite form [
23] and MnOx-CeO
2 [
16] exhibited high NH
3-SCR activity at low temperatures.
The present article deals with the preparation of Mn/TiO2, Mn-Fe/TiO2, and Mn-Fe-Ce/TiO2 using several metal precursors. Various methods of characterization were employed to understand the differences in catalyst properties and activities.
2. Results and Discussion
The SCR NO conversion profiles of the 5Mn
Nit and 5Mn
Ace supported catalysts are shown in
Figure 1. Among the catalysts studied, Mn deposited on TiO
2 showed superior catalytic activity followed by ZrO
2 and Al
2O
3. In particular, Mn
Ace/Ti was more active compared to the Mn
Nit/Ti. At 250 °C, the Mn
Ace/Ti and Mn
Nit/Ti catalysts displayed a NO conversion of 81 and 65%, respectively. The low temperature activity of the Mn/TiO
2 catalysts were compared with silica and alumina by Simirniotis et al. [
24] and they concluded that Lewis acid sites, a high surface concentration of MnO
2, and good redox properties were important in achieving low temperature SCR activity. For further experiments, TiO
2 was chosen as the unique support.
Mn/TiO
2 doped with transition metals (e.g., Ni, Cu, and Fe) had high resistance to sintering and more favorable Mn dispersion [
12]. It is also reported that Mn/TiO
2 catalysts promoted with transition metals showed activity for NO oxidation to NO
2 [
15]. In our previous publication, we reported the promotional effect of Fe on Mn/TiO
2 catalysts and the optimum formulation was found to be 25 Mn
0.75Fe
0.25/Ti using deposition-precipitation [
19].
Figure 2a shows the NO conversion profiles of the 25 wt.% Mn
0.75Fe
0.25/Ti catalysts with Mn and Fe precursor combinations as a function of reaction temperature. All the catalysts, except for the 25Mn
0.75NitFe
0.25Ace/Ti catalyst, showed full conversion above 225 °C. The 25Mn
0.75AceFe
0.25Nit/Ti, 25Mn
0.75NitFe
0.25Nit/Ti, 25Mn
0.75AceFe
0.25Ace/Ti, and 25Mn
0.75NitFe
0.25Ace/Ti catalysts exhibited NO conversion of 69.6, 55.6, 47.0, and 43.1% at 150 °C, respectively, illustrating the importance of precursors on catalyst activity.
It is also well known that the presence of Ce can enhance the SCR performance and selectivity to N
2 [
17]. It is also known that the presence of Ce further overcomes the SO
2 deactivation and water inhibition effects [
18].
Figure 2b shows the effect of Ce precursor on the optimum 25Mn
0.75AceFe
0.25Nit/Ti catalyst. The presence of Ce in the 25Mn
0.75AceFe
0.20NitCe
0.05Ace/Ti catalyst showed slightly better performance while 25Mn
0.75AceFe
0.20NitCe
0.05Nit/Ti showed less performance than the previously optimized 25Mn
0.75AceFe
0.25Nit/Ti catalyst. This further confirms the sensitivity of SCR catalysts to the choice of precursors. The 25Mn
0.75AceFe
0.25Nit/Ti, 25Mn
0.75AceFe
0.20NitCe
0.05Ace/Ti, and 25Mn
0.75AceFe
0.20NitCe
0.05Nit/Ti catalysts displayed NO conversions of 69.6, 73.0, and 54.4% at 150 °C, respectively.
Table 1 summarizes the N
2O formation data obtained at 150 °C over the 25Mn
0.75AceFe
0.25Nit/Ti, 25Mn
0.75AceFe
0.20NitCe
0.05Ace/Ti, and 25Mn
0.75AceFe
0.20NitCe
0.05Nit/Ti catalysts, which respectively produced 35, 15, and 20 ppm of N
2O under wet conditions (2.3 vol% H
2O). Thus, the presence of Ce can increase the selectivity to N
2. A moderate Ce content in Mn-Fe/TiO
2 catalysts contributed to decreased N
2O formation by hindering the over oxidation of NH
3, the dominant step in N
2O formation. N
2O formation is controlled by selective reaction of NO with NH
3, limiting the oxidation of NH
3. However, at a higher concentration of water (≈10 vol%), no N
2O was formed for all catalysts.
Figure 3a shows the NO conversion profiles of 20–30 wt.% Mn
0.75AceFe
0.20NitCe
0.05Ace/Ti catalysts vs. the reaction temperature. The NO conversion was in the following order: 25 Mn
0.75AceFe
0.20NitCe
0.05Ace/Ti > 20 Mn
0.75AceFe
0.20NitCe
0.05Ace/Ti > 30 Mn
0.75AceFe
0.20NitCe
0.05Ace/Ti between 150–200 °C. Above 200 °C, the catalysts showed almost 100% NO conversion and it was not possible to discriminate between them. The 20, 25, and 30 wt.% Mn
0.75AceFe
0.20NitCe
0.05Ace/Ti catalysts displayed a NO conversion of 59.8, 73.0, and 55.8% at 150 °C, respectively.
Figure 3b shows the SCR activity of the 25 wt.% Mn
AceFe
NitCe
Ace/Ti catalyst with different Mn-Fe-Ce mole fractions. The highest NO conversion was attained at a Mn mole fraction of 0.75 and the lowest activity at a mole fraction of 0.725, indicating that the minimum Mn content should be 0.75. Maximum NO conversion was obtained at a Fe mole fraction of 0.20 followed by the mole fractions 0.175 and 0.225. Maximum NO conversion was obtained at a Ce mole fraction of 0.05 followed by 0.075 and 0.025. The 25Mn
0.75AceFe
0.20NitCe
0.05Ace/Ti, 25Mn
0.75AceFe
0.175NitCe
0.075Ace/Ti, 25Mn
0.75AceFe
0.225NitCe
0.025Ace/Ti, and 25Mn
0.725AceFe
0.225NitCe
0.05Ace/Ti catalysts displayed NO conversions of 73.0, 65.2, 48.4, and 41.4 at 150 °C, respectively.
The effect of space velocity (mLg
−1h
−1) on the most active 25Mn
0.75AceFe
0.20NitCe
0.05Ace/Ti catalyst is shown in
Figure 4. Space velocity is an important factor to be considered in the catalyst design as well as to compare to catalysts in the open literature. The catalyst displayed a NO conversion of 98, 88 and 79%, respectively, at space velocities of 360,000, 450,000, and 600,000 mL g
−1 h
−1 at 200 °C. The fact that at the lowest space velocity, NO conversions of above 90% can be attained at temperatures above 200 C indicates that unselective oxidation of NH
3 is not a major side reaction. The temperature window between 125 and 175 °C was used to determine an apparent activation energy of 38.6 ± 4.0 kJ/mol. This value clearly indicates that the present activity measurements on powdered samples were not strongly influenced by transport limitations in the pursued temperature window of the tail-end operation (125–175 °C).
The BET (Brunauer–Emmett–Teller) surface area, H
2-TPR, and NO oxidation results of the MnFe/Ti and MnFeCe/Ti catalysts are summarized in
Table 2. The BET surface area of DT51-TiO
2 was 83 m
2/g, while those of the MnFe/Ti and MnFeCe/Ti catalysts showed an increased surface area even with 25 wt.% active metal content. Thus, pore blockage of TiO
2 is unlikely and the active metal oxides are probably highly dispersed on the TiO
2 support. The increased surface area was due to increased microporosity compared to TiO
2 (see
Figures S1 and S2, Supplementary Materials).
Ease of reduction of metals oxides is known to be an indicator for favorable low temperature SCR activity [
19]. The H
2 consumption profiles of the 25MnFe/Ti and 25MnFeCe/Ti catalysts are shown in
Figure 5 and the integrated values (µmol/g) are summarized in
Table 2. All 25MnFe/Ti catalysts showed almost similar reduction patterns. To distinguish the bimetallic reduction patterns of the MnFe/Ti catalysts, the reduction patterns of Fe/TiO
2 and Mn/TiO
2 were reported [
19]. The Fe/TiO
2 catalyst reduced from Fe
2O
3 to Fe at 338 °C. The Mn/TiO
2 catalyst showed three peaks corresponding to step-wise reduction of MnO
2 to Mn
2O
3, Mn
2O
3 to Mn
3O
4, and Mn
3O
4 to MnO [
19]. The 25MnFe/Ti materials exhibited only two peaks with the first (maximum at ≈255–270 °C) corresponding to the MnO
2 reduction, and the second one (maximum at ≈350–390 °C) could be due to the reduction of subsequent manganese oxide phases mixed with iron oxide. The 25Mn
0.75AceFe
0.25Ace/Ti and 25Mn
0.75NitFe
0.25Ace/Ti catalysts showed visible shoulder peaks at around 230 and 340 °C, and that of the 25Mn
0.75AceFe
0.25Nit/Ti and 25Mn
0.75NitFe
0.25Nit/Ti catalysts did not display visible shoulder peaks because of the broad nature of the reduction profiles. The origin of the shoulder peak toward lower temperature is unclear, but might be because of the presence of smaller, more easily reducible manganese oxide particles.
The 25Mn0.75AceFe0.25Nit/Ti, 25Mn0.75NitFe0.25Nit/Ti, 25Mn0.75AceFe0.25Ace/Ti, and 25Mn0.75NitFe0.25Ace/Ti catalysts consumed 4120, 4004, 3996, and 3972 µmol/g of H2, respectively. The difference in H2 consumption between the catalysts was small, but the 25Mn0.75AceFe0.25Nit/Ti catalyst was reduced at a relatively low temperature. Thus, the ease of reduction pattern and the dominating MnO2 phase (first peak) for the 25Mn0.75AceFe0.25Nit/Ti catalyst seem to be the main contributors to the superior low temperature SCR activity.
The 25MnFeCe/Ti catalysts exhibited three reduction peaks, where the first two peaks can be assigned as similar to those of the MnFe/Ti catalysts and then the third reduction peak about 550 °C is due to the reduction of CeO
2 [
25]. Most importantly, the 25Mn
0.75AceFe
0.20NitCe
0.05Ace/Ti and 25Mn
0.75AceFe
0.20NitCe
0.05Nit/Ti catalysts displayed a H
2 consumption of 5040 and 4907 µmol/g, respectively. This H
2 consumption, which is higher than for the 25MnFe/Ti catalysts, could be due to better dispersion of Mn, Fe, and Ce. The 25Mn
0.75AceFe
0.20NitCe
0.05Ace/Ti catalyst was reduced at lower temperatures (≈10–15 °C) compared to the 25Mn
0.75AceFe
0.20NitCe
0.05Nit/Ti catalyst. Thus, also in this case, the ease of reduction and the dominating MnO
2 phase (first reduction peak) in the 25Mn
0.75AceFe
0.20NitCe
0.05Ace/Ti catalyst were the main reasons for the superior SCR activity at low temperature.
The X-ray powder diffraction (XRPD) patterns of the 25MnFe/Ti and 25MnFeCe/Ti catalysts are shown in
Figure 6. The TiO
2 anatase phase was dominant in all catalysts and manganese oxide, iron oxide, and cerium oxides or other mixed phases of Mn, Fe, or Ce were not observed. This is a clear indication that active metal oxides are highly dispersed and/or in an amorphous state. To understand the amorphous state of the active metal oxides on the surface of the catalysts, thermal treatment at 400, 500, and 600 °C for 2 h was performed. It is anticipated that amorphous to crystalline phase transformation can happen by thermal treatment.
The XRPD patterns of the 5MnAce/Ti, 25Mn
0.75AceFe
0.25Nit/Ti, and 25Mn
0.75AceFe
0.20NitCe
0.05Ace/Ti samples calcined at several temperatures are shown in
Figure 7. The 5MnAce/Ti catalyst showed similar diffraction patterns of anatase-TiO
2 at 400 °C and there was a small deviation compared to the TiO
2 patterns at 500 °C. With a further increase in temperature to 600 °C, MnO
2, Mn
2O
3, and anatase phases were observed. Similarly, the 25Mn
0.75AceFe
0.25Nit/Ti catalyst showed only anatase phases of TiO
2 at 400 and 500 °C, and crystalline Mn
2O
3 and anatase TiO
2 phases were observed at 600 °C. No diffraction patterns of iron oxide were observed. Thus, we can clearly see that the presence of Fe can cause the transformation of the mixed manganese oxide phase into particles big enough for XRD detection to consist only of the Mn
2O
3 phase. The 25Mn
0.75AceFe
0.20NitCe
0.05Ace/Ti catalyst showed only anatase TiO
2 phases even at 600 °C, thus the combined presence of Fe and Ce on Mn/Ti can hamper the transformation of the amorphous to the crystalline phase.
Figure 8 shows the NO conversion profiles of 5MnAce/Ti, 25Mn
0.75AceFe
0.25Nit/Ti, and 25Mn
0.75AceFe
0.20NitCe
0.05Ace/Ti catalysts calcined at 400, 500, and 600 °C. The catalysts were most active when calcined at 400 °C where the catalysts calcined at 500 and 600 °C showed lower activity. When calcined at 400 °C, the catalysts were rich in the amorphous metal oxide phases (manganese, iron or cerium oxides), which are known to be SCR active. Further increase in calcination temperature resulted in the partial transformation of amorphous manganese oxide to crystalline manganese oxides (MnO
2 or Mn
2O
3). Overall, the SCR activity of the catalysts was in parallel to the amorphous to crystalline transformation of the catalysts as also reported by Kang and Tang et al. [
26,
27].
The impact of the calcination temperature and transformation of active oxides can also be studied in combination with H
2-TPR. The redox properties of the 5MnAce/Ti, 25Mn
0.75AceFe
0.25Nit/Ti, and 25Mn
0.75AceFe
0.20NitCe
0.05Ace/Ti catalysts calcined at 400, 500, and 600 °C are shown in
Figure 9. The 5MnAce/Ti catalyst showed three different reduction peaks, which correspond to stepwise reduction of MnO
2 to Mn
2O
3 (≈260 °C), Mn
2O
3 to Mn
3O
4 (≈360 °C), and Mn
3O
4 to MnO (≈460 °C) [
19]. Increasing the calcination temperature from 400 to 600 °C, the 5MnAce/Ti catalyst shifted the first reduction peak to higher temperatures due to strong metal–support interactions [
22], and the intensity of the second reduction peak was increased, which further indicates that the SCR active MnO
2 phase decreases and the Mn
2O
3 phase increases. The shifting of both reduction peaks to higher temperatures might be due to particle growth (sintering).
The 25Mn0.75AceFe0.25Nit/Ti and 25Mn0.75AceFe0.20NitCe0.05Ace/Ti catalysts showed almost similar reduction patterns at 400 and 500 °C of calcination temperature, further indicating that Fe and Ce are inhibiting the phase transformation of MnO2 to Mn2O3 and possibly particle growth (sintering). At 600 °C, the catalyst displayed a shift in the MnO2 reduction peak to high temperatures and the intensity of the second reduction peak was increased. Thus, the combined presence of Fe and Ce on Mn/TiO2 can increase the thermal stability.
The observed low temperature activity of Mn catalysts can also be explained by the NO to NO
2 oxidation ability as reported previously [
15,
18].
Table 2 shows the oxidation of NO to NO
2 on the 25Mn
0.75Fe
0.25Ti-DP and 25Mn
0.75Fe
0.25Ce
0.05Ti-DP catalysts at 300 °C under wet conditions. All the catalysts displayed high NO to NO
2 conversion of 41.6 to 66%. The 25Mn
0.75AceFe
0.20NitCe
0.05Ace/Ti and 25Mn
0.75AceFe
0.25Nit/Ti catalysts displayed the highest NO to NO
2 conversion. The observed NO oxidation is consistent with the increased SCR activity of the catalysts, since partial conversion of NO into NO
2 is helpful to promote the fast SCR reaction, which is also known to go on at very low temperatures [
28].
The surface composition as obtained by XPS (X-ray photoelectron spectroscopy) characterization is shown in
Table 3. The 25Mn
0.75AceFe
0.25Nit/Ti and 25Mn
0.75AceFe
0.20NitCe
0.05Ace/Ti showed a surface Mn/Fe molar ratio of 2.08 and 1.30, respectively. Thus, it appears that the precursor/promoter combination has an influence on forming MnFe oxide species on the surface of the support. Importantly, the 25Mn
0.75AceFe
0.25Nit/Ti catalyst showed an O
α concentration of 50.1% of the total oxygen while the Ce promoted 25Mn
0.75AceFe
0.20NitCe
0.05Ace/Ti catalyst yielded 83.8%, respectively. High concentrations of chemisorbed oxygen have been reported to have a beneficial influence on the low-temperature SCR reaction [
12,
29] and is explained by an increased rate of NO to NO
2 oxidation [
29]. Our XPS results showed a significantly higher concentration of more reactive surface oxygen (O
α) in the 25Mn
0.75AceFe
0.20NitCe
0.05Ace/Ti catalyst than in the 25Mn
0.75AceFe
0.25Nit/Ti catalyst. This is reflected in the higher NO to NO
2 oxidation activity (66% vs. 60.4% conversion).