The Characterization and SCR Performance of Mn-Containing α -Fe 2 O 3 Derived from the Decomposition of Siderite

: In this work, a nano-structured iron-manganese oxide composite was prepared by calcining natural manganese-rich siderite at di ﬀ erent temperatures (450, 500, 550, 600 ◦ C, labeled as H450, H500, H550, H600, respectively), and their performances of selective catalytic reduction (SCR) of NO by NH 3 were investigated. XRD, XRF, BET, XPS, SEM, and TEM were used to investigate the morphology, composition, and surface characteristics of the catalyst. The results showed that the decomposition of siderite occurred from 450 ◦ C to around 550 ◦ C during the calcination in air atmosphere; moreover, the siderite could be converted into nano-structured α -Fe 2 O 3 . The speciﬁc surface area of the material increased, and Mn 2 + was transformed into Mn 4 + , which were beneﬁcial to the SCR. Among these catalysts, H550 had the best SCR performance, with NO removal of 98% at a temperature window from 200 to 250 ◦ C. The presence of water vapor and sulfur dioxide can inhibit the SCR performance of the catalysts, but this inhibition e ﬀ ect was not obvious for H550 at the optimum reaction temperature (250 ◦ C). The ﬁndings presented in this study are signiﬁcant toward the application of the Mn-rich siderite as a precursor in preparing the Fe-Mn oxides for catalytic de-NO x by SCR.


Introduction
Nitrogen oxides (NO x ) emitted from various sources are a major atmospheric pollutant, causing acid rain, photochemical smog, and endangering human health and the ecological environment [1][2][3][4]. The world energy infrastructure being based on fossil energy further aggravates NO x pollution to a certain extent. Many technical methods have been purposed for controlling nitrogen oxides, of which selective catalytic reduction (SCR) for NO removal with NH 3 has been the most investigated technology for several decades [5][6][7][8], especially low-temperature SCR [9]. For the catalysts, TiO 2 -supported vanadium-based catalysts such as V 2 O 5 −WO 3 /TiO 2 catalysts are widely used due to their advantages, including high NO conversion, N 2 selectivity, resistance to sulfur toxicity, and strong mechanical strength [10][11][12][13]. However, considering the high reaction temperature and environmental toxicity of vanadium-based catalysts [14][15][16], vanadium-free catalysts like metal oxides (Fe, Mn, Cu, Ce) have received more and more attention [17][18][19].
Iron oxides and manganese oxides are preferred as vanadium-free catalysts due to their good low-temperature activity and high NO conversion. In previous studies, α-Fe 2 O 3 was shown to have outstanding SCR performance in the temperature range 200-400 • C [20]. Moreover, studies have shown that calcined siderite can form nano-scale α-Fe 2 O 3 , accompanied during the transformation process by the phases FeCO 3 -Fe 3 O 4 -γ-Fe 2 O 3 -α-Fe 2 O 3 [21]. Meanwhile, manganese oxides are generally used as an active component supported on a carrier. Common examples including MnO x -CeO 2 /CNTs and Mn-Fe/TiO 2 have excellent SCR activity [22][23][24].
On the other hand, siderite is abundant in China with a low utilization rate [25,26]. Calcined siderite can form Fe 2 O 3 and Fe 3 O 4 [27]. The calcination process can increase the specific surface area still [28], and the calcined product is an environmentally friendly and highly active mineral material [27,29]. Additionally, the substitution of Mn for Fe in the structure of naturally occurring siderite is universal. Moreover, the presence of manganese oxides can improve the low temperature activity of a catalyst [30]. Thus, Fe-Mn oxides with abundant micropores and large specific surface areas should be obtained by the thermal decomposition of siderite.
In this present study, natural Mn-rich siderite was used to prepare Fe-Mn oxides with abundant micropores for SCR evaluation. The effects of calcination temperature on the material and SCR activity of the catalysts were investigated by the laboratory-scale catalytic system. The samples were characterized by X-ray Diffraction (XRD), X-ray Photoelectron Spectroscopy (XPS), Raman, Scanning Electron Microscope (SEM), Transmission Electron Microscope (TEM), and Brunauer-Emmett-Teller (BET). The objectives of the present work were (i) to obtain an iron-manganese oxide catalyst by calcining siderite, (ii) to investigate the effects of calcination temperature and various parameters (such as water vapor and sulfur) on the SCR reaction, and (iii) to examine the relationship between the structure of calcined products and SCR performance.

Catalyst Preparation
Natural siderite was collected from Rushan, Shandong Province, China. The sample was separated by a 40-60 mesh standard sieve after grinding and crushing. Afterwards, four groups of the catalysts were prepared by calcining in muffle furnace for 1 h at different temperatures (450, 500, 550, 600 • C) and labeled as H450, H500, H550, H600, respectively.

Catalyst Characterization
Crystal phases of the samples were determined by using a Dandong DS-2700 X-ray diffractometer (Cu Kα, with 2θ ranging between 5 • and 70 • and operated at 40 KV and 30 mA). Determination of the decomposition temperature was carried out using a thermogravimetric analyzer (EXSTAR S II TG/DTG7300, Hitachi, Tokyo, Japan). An X-ray photoelectron spectrometer (Thermo ESCALAB250Xi, ThermoFisher, Waltham, MA, USA) was used to analyze the elemental distribution and valence state of the catalysts. A scanning electron microscope (SEM, SU8020, Hitachi, Tokyo, Japan) and a transmission electron microscope (TEM, JEM-2100F, JEOL, Tokyo, Japan) were used to characterize the morphology of the samples. The molecular structures of samples were analyzed by Raman spectroscopy using an HR Evolution Raman spectrometer from HORIBA (Jobin Yvon, Paris, France). Chemical composition was measured on an X-ray fluorescence spectrometer (Shimadzu XRF-1800, Shimadzu, Kyoto, Japan) with Rh radiation. The specific surface area, pore volume, and pore size distribution of the samples were analyzed using a surface area and pore size analyzer (Quanta NOVA 3000e, Quantachrome, Shanghai, China).

Catalytic Activity Testing
The SCR evaluation of the as-prepared catalysts was carried out by fixing the catalyst volume in a quartz tube reactor with an inner diameter of 6 mm. The simulated flue gas was controlled by mass flow controllers at a flow rate of 150 mL·min −1 (including the carrier gas (Ar), 500 ppm of NH 3 , 500 ppm of NO, and 1.5% of O 2 ). The reaction temperature ranged from 50 to 400 • C with a gas hourly space velocity (GHSV) of 16000 h −1 . When the SCR reaction was stable, the concentrations of the outlet gases were detected by a flue gas analyzer (C500, ONUEE, Shenzhen, China). The experimental setup involving gas mixing system, reaction system, and detecting system is shown in Figure 1. NO conversion is calculated by the following Equation (1):

Results and Discussion
3.1. XRD, XRF, and TG of Siderite before and after Calcination Figure 2b shows the TG and DTG results of the natural siderite calcined in air atmosphere from 20 to 800 • C with a heating rate of 10 • C/min. According to previous studies, the main decomposition of siderite started at 478 • C and ended at 580 • C, with a maximum rate at 529 • C [31]. In this experiment, the natural siderite was almost in a stable state with a mass loss of roughly 1.13% before 400 • C. Then, a significant mass loss of approximately 30.94% occurred from 420 to 600 • C, which should be ascribed to the decomposition of siderite (FeCO 3 ) into Fe 2 O 3 [32]. According to the TG/DTG results, the calcination temperatures of 450, 500, 550, and 600 • C were selected for catalyst preparation.

XPS
XPS analysis was conducted to identify the species on the surface of the natural siderite and . The change to iron oxide was very obvious based on the XRD pattern, while the change to manganese oxide was not. This may be due to low content of manganese oxide or the substitution of Mn for Fe in the structure of the siderite [10].
The XRF results show that the sample was composed of Fe 2 O 3 (58.995 wt %), MnO (3.46 wt %), SiO 2 (1.73 wt %), and traces of CaO and MgO, with an ignition loss of about 32.5%. According to previous studies, the existence of manganese oxides had favorable effect for the SCR reactions [33,34]. Therefore, the confirmation of the valence of Mn will be carried out in the next section.

XPS
XPS analysis was conducted to identify the species on the surface of the natural siderite and thermally treated samples. According to the XRD patterns, the distinct peak at about 711.0 eV belonged to Fe 3+ [35,36], as shown in Figure 3a. The peaks for Fe 2+ and FeCO 3 were detected at around 709 and 713 eV, respectively [37,38]. The peaks of FeCO 3 experienced an obvious decrease as the calcination temperature increased to 450 • C and were then replaced by the characteristic peak of Fe 3+ as the temperature reached 500 • C [39]. This also proved that the increasing temperature favored the decomposition of siderite, especially for the temperature of 550 • C.
As displayed in Figure 3c, the peaks at the binding energy of 530 and 532 eV were assigned to the lattice oxygen O 2− (denoted as O β ) and the surface-adsorbed oxygen (denoted as O α ), respectively [40]. As the calcination temperature increased to 500 • C, the intensity of O β was gradually enhanced, while the intensity of O α became weaker. Interestingly, this variation exhibited an opposite trend when the calcination temperature further increased to 600 • C. It was reported that O α was more beneficial for the SCR reactions via a "fast SCR" process because of its higher mobility [41]. The O α /(O α + O β ) ratio was the lowest in the case of H550, but O α still occupied the main part, with 60.5%. Therefore, it is believed the thermal treatment products will have a good SCR performance.
To the best of our knowledge, Mn 4+ species were considered as the active sites for SCR reactions. Previous studies have shown that Mn 4+ , Mn 3+ , and Mn 2+ often existed in material [33,42]. Therefore, the distribution of the Mn valence was determined. As presented in Figure 3b, the peaks at 641.2 and 642.6 eV were attributed to Mn 3+ and Mn 4+ , respectively [36,43]. The peaks near 640.8 and 641.7 eV were attributed to Mn 2+ and Mn 3+ , respectively [7,23]. Other XPS peaks were attributed to Mn 4+ at a binding energy of around 642.8 eV [44,45]. It can be deduced from Figure 3 that with an increase in calcination temperature to 500 • C, the Mn 4+ on the surface increased gradually. When the calcination temperature attained 550 • C, a large amount of Mn 4+ is formed accompanied by a small amount of Mn 3+ . As the temperature increased to 600 • C, the content of Mn 4+ decreased, which is not conducive to SCR performance [7].

SEM and TEM
The SEM images of the natural siderite and thermally treated samples are displayed in Figure 4. The bulk siderite with a relatively even and smooth surface can be observed in Figure 4a [46]. Moreover, there were much fewer siderite aggregated particles with irregular shape on the surface of the bulk siderite [47]. With an increase in the calcination temperature, the bulk siderites were cleaved into small grains, and the surfaces of H500 (Figure 4b), H550 (Figure 4c), and H600 (Figure 4d) became loose. In addition, almost no pores can be observed in the structure of the natural siderite, as shown in Figure 4e. The EDX also confirmed the existence of Mn oxide in the siderite. However, numerous nanopores with diameters of 10 to 20 nm appeared in H550 (Figure 4f). Based on the XRD results, siderite was completely decomposed into hematite, and a large amount of carbon dioxide was produced during the calcination process at 550 • C. Thus, the formation of these nanopores was related to the decomposition of siderite. It was reported that the property of the nanoporous structure was beneficial for SCR reactions [7]. As shown in Figure 4e,f, the EDX elemental mapping images of Mn and Fe demonstrated that the distribution of Mn was partially overlapped with that of Fe. This phenomenon was probably due to the substitution of Mn for Fe in the structure of the siderite, since the isomorphous substitution of Mn for Fe is a universal occurrence in the natural siderite [48].

BET and Raman
The BET specific surface area (BET-SSA), pore volume, and pore size are displayed in Table 1. The order of BET-SSA was as follows: H550 (54.99 m 2 ·g −1 ) > H500 (52.04 m 2 ·g −1 ) > H600 (37.26 m 2 ·g −1 ) > H450 (14.1 m 2 ·g −1 ), indicating the thermal treatment significantly affected the specific surface area of the samples. As is well known, catalysts with a large specific surface area are beneficial to the catalytic reaction due to the greater number of active sites [5].
The N 2 adsorption-desorption isotherms and pore size distributions of H450, H500, H550, and H600 are shown in Figure 5a. In general, the isotherms of the samples exhibited a type IV isotherm, suggesting that the prepared samples were typical mesoporous materials [49]. The pore sizes of these samples were in the range 2-20 nm. These mesopores should provide more internal specific surface area and pore volume promoting the SCR efficiency [14]. According to the XRD, SEM, and TEM results, the decomposition of FeCO 3 and escape of CO 2 should be taken into consideration for the formation of nanoparticles and pore structures. A rich mesoporous structure allows the material to be more accessible to the reactant gases [50].
Raman spectroscopy was used to further analyse the molecular structure of the active element component in this study and the results are presented in Figure 5b. According to previous studies, the bands appearing at 182, 282, 734, and 1084 cm -1 were assigned to siderite [51,52]. After calcination, some new bands at 220, 288, 404, 608, and 652 cm -1 were found and identified as hematite (α-Fe 2 O 3 ) [53]. It was reported that the band around 650 cm -1 was not a perfect spectrum of hematite (α-Fe 2 O 3 ) [54]. Some studies have found that the bands near 580-650 cm −1 can also be assigned to MnO 2 [55,56]. Therefore, the band at 650 cm -1 could also be ascribed to MnO 2 [10]. This was consistent with the XPS results. Table 1. Specific surface area, pore volume, and average pore size of H450-H600.

SCR Performance and Resistance
The SCR activities of the as-prepared samples were examined in terms of the NO conversions under different reaction temperatures. It was noted that siderite exhibited very low SCR activity, with about 20% NO conversion (Figure 6a). However, the SCR performance was greatly enhanced after calcination at different temperatures. NO conversion of 64% was obtained for H450 at 250 • C, nearly 100% for H500 and H550 at 200-250 • C, and 95% for H600 at 250 • C, respectively. The relatively low SCR activities of H450 and H600 could be due to the low specific surface area and lattice oxygen content. In addition, the thermally treated Mn-rich siderite (H500 and H550) had a lower reaction temperature window compared with that of the common α-Fe 2 O 3 (250-300 • C) [40]. This phenomenon could be ascribed to the oxidization of MnCO 3 to MnO 2 during the calcination process, which greatly enhanced the low temperature activity of the material. However, the NO conversion gradually decreased when the reaction temperature was over 250 • C, which was likely caused by the NH 3 oxidation at high temperatures. In summary, large specific surface area and porous structure, the existence of MnO 2 , and the particular adsorption state of oxygen improved for the SCR performance of the newly formed α-Fe 2 O 3 [14,57].
According to the literature [34], the content of water vapor in the flue gas is approximately 6 to 12%. Therefore, it is necessary to evaluate the SCR performance of the samples under the existence water vapor. As shown in the Figure 6b, NO conversions of H550 under different contents of water vapor showed an obvious decline in the range 150-250 • C, indicating the addition of water vapor had a certain influence on the SCR activity. It was reported that the presence of water vapor would lead to the transformation of Lewis acid to Brønsted acid [58]. This is also the reason why many studies have shown that the effect on SCR is reversible when relieving water vapor [24,59]. Comparing different water vapor concentrations in H550 catalytic treatment of NO, the higher the concentration of water vapor, the more obvious the activity inhibition at low temperature. However, the effect of water vapor gradually decreases with the increase of reaction temperature, and when the reaction temperature exceed 300 • C, the presence of water vapor promotes the SCR performance, which is similar to previous studies [34,60]. This is due to the presence of little water at high reaction temperatures increased the Brønsted acid sites [61].
As is well known, the sulfation of the active species is the main reason for catalyst deactivation [62,63]. The sulfate species binding with the active sites would cut off the L-H reaction mechanism by the sulfation process [34,64]. As shown in Figure 6c, the SCR activity of H550 varied significantly in the range 100-250 • C under different concentrations of SO 2 . In particular, the inhibition effect was found to be greater under higher concentrations of SO 2 . This could be due to the formation of NH 4 HSO 4 and/or (NH 4 ) 2 SO 4 , which could accumulate on the surface of the catalyst and block the active sites [65]. When the reaction temperature exceeded 300 • C, the presence of SO 2 gave a promoting effect.

Conclusions
The calcination of natural siderite at different temperatures in air atmosphere was carried out to prepare the Mn-containing α-Fe 2 O 3 for NH 3 -SCR. The natural siderite was gradually decomposed at 450 • C and converted into Mn-containing α-Fe 2 O 3 . The escape of CO 2 in the process of decomposition of siderite improved the specific surface area. H550 had the best SCR performance with NO conversion of 98% at 200 • C, due to its larger specific surface area and nanoscale structure. In addition, with the increase of calcination temperature, the gradual conversion of Mn 2+ into Mn 4+ was responsible for the decrease of the reaction temperature window. The water and sulfur dioxide resistant experiments demonstrated that the materials have a certain resistance to water and sulfur, especially at 250 • C. Thus, this calcined siderite is an active material for NH 3 -SCR reaction. The results of this study provide a new way for the utilization of natural minerals in SCR.

Conflicts of Interest:
The authors declare no conflict of interest.