1 Introduction

NOx (nitrogen oxides, mainly NO and NO2) is one of the main pollutants in the atmosphere, however, over 95 % of the NOx in the atmosphere is NO. NO2 is found in the air only in very small proportions. About 90 % of NOx in the flue gas is NO, which mainly comes from vehicle exhaust, fossil fuel combustion, and emissions from stationary sources. NO can cause acid rain, photochemical smog, greenhouse effects, and damage to human health [1–3]. Selective catalytic reduction (SCR) methods aim to reduce NOx in flue gas to N2 using a reducing agent such as NH3; this reaction must be controlled at a certain temperature with a certain amount of oxygen content, and using specific catalysts. The most mature commercial catalyst for this process is V2O5–WO3/TiO2, however, the temperature window of the catalyst is between 350 and 400 °C. Thus, SCR devices must be placed before the air pre-heater, particularly because poor N2 selectivity and high SO3 conversion rates are evident at this temperature range [4]. If the SCR reactor is placed after the electrostatic precipitator and desulfurization devices, the effects of SO2 and dust alkaline earth metal on the SCR catalysts can be reduced significantly. The outlet temperature of flue gas during sintering after semi-desulfurization is generally about 80–150 °C. If flue gas reheating cannot be introduced, the cost of de-nitration can be significantly reduced [5]. Hence, developing efficient and economic catalysts at lower temperatures, specifically at 80–180 °C, is a popular focus of research on de-nitrification of sintering flue gas. Research on the effects of complex cryogenic stage as sintering flue gas to low temperature catalyst is still at the blank stage.

Researchers have carried out many in-depth studies on SCR catalyst supports at low temperature as well as their active components. Findings indicate that Mn-based catalysts exhibit good low temperature catalytic properties [6, 7]. In addition, anatase TiO2 is often used as a support of SCR catalysts because of its good mechanical and strong resistance to sulfuration properties [8]. However, the small specific surface area and poor thermal stability of the material have restricted its catalytic properties to some extent [9, 10]. Carbon-based materials (e.g., activated carbon, carbon fibers, carbon nanotubes, etc.) have larger specific surface areas and better chemical stability; thus, these materials have been used as a supports in SCR DeNOx experiments [11–14]. Mn-based catalysts with carbon-based materials as supports, such as MnOX/ACF [11], MnOX/AC/ceramic, [12] and MnOX/MWCNTs (manganese oxides supported on multi-walled carbon nanotubes), have been widely used in research and can yield good SCR performance at temperatures below 250 °C. Zhao Xueying and other researchers [15] have prepared TiO2–CNT composite powder materials of different mass ratios by the sol–gel method. Results show that this method allows anatase nano-TiO2 to be evenly distributed on the CNT surface so that agglomeration of TiO2 solution molecules is effectively inhibited during hydrolysis. Graphene oxide (GO) and graphene (GE) are new nanoscale carbon materials bearing the basic features of not only carbon-based materials but also surface active groups with electron transfer ability, especially during GO preparation to promote oxidation. SO4 2− is formed on the catalyst surface by addition of a large quantity of concentrated sulfuric acid, which plays a role of pre-sulfuration to the support and this feature will necessarily improve the resistance to poisoning of catalyst to SO2. Currently, composite supports of TiO2–GO and TiO2GE are only used in photo catalysis, although their characteristics may also be applicable in denitration by SCR at low temperatures. Yang et al. [16] found that –OH, –COOH, and other oxygen-containing functional groups can improve CNT dispersion in water and enhance particle interactions with the substrate to improve the catalytic activity of the SCR reaction. GO surfaces also have these functional groups.

In this paper, pertinent to the outlet temperature of desulfurization process by semi-dry flue gas sintering, the temperature range of 80–180 °C is selected. MnOX/TiO2–GO catalysts were prepared using the sol–gel method and ultrasonic impregnation. The TiO2 in the TiO2–GO supports were anatase in structure, which is conducive to enhancing SCR catalytic activity. XRD characterization confirms that introduction of GO does not affect polymorphs of TiO2. Using field emission scanning electron microscopy (FESEM), EDS spectrum analysis, X-ray diffraction (XRD), Brunauer–Emmett–Teller analysis (BET), transmission electron microscopy (TEM), X-ray photoelectron spectroscopy (XPS), and Fourier transform infrared (FT-IR) spectroscopy, interactions among the GO support, TiO2, and MnOX were further explored to determine why SCR reactivity is improved under specific conditions. The effects of H2O and SO2 on the catalysts and the resistance to poisoning mechanism were investigated to provide theoretical guidance for future research on low temperature SCR catalysts for sintering flue gas and their industrial applications.

2 Materials and Methods

2.1 Catalyst Preparation

2.1.1 Preparation of TiO2–GO Supports

In the experiments, an improved Hummers method was applied to prepare GO [17, 18]. Firstly, 23 ml of concentrated H2SO4 was placed in a 1000 ml beaker. Graphite flakes weighing 1 g and NaNO3 weighing 0.5 g (Kim Electronic Materials Plant, Lacey City of Qingdao Municipality) were slowly added to concentrated H2SO4 at moderate stirring speed and in an ice bath to maintain the temperature and prevent the bump splash of reaction. When the chemicals had been sufficiently dissolved and uniformly mixed, 3 g of KMnO4 was slowly added to the mixture at vigorous stirring speed. A thermometer was used to monitor and maintain the temperature below 20 °C.

Upon addition of KMnO4, a violent reaction took place and the color of the liquid in the beaker changed from black to dark green. Upon complete dissolution of KMnO4, the ice bath was removed and a water bath at a temperature of 35 ± 3 °C was used instead. The liquid remained incubated for 30 min. The contents of the beaker began foaming and turned into a black paste-like substance. Following this development, 46 ml of de-ionized water was added to the solution under stirring. The paste-like material became thinner and was accompanied by strong bubbling. The beaker was moved to a water bath at 90 °C. Incubation was performed for another 15 min, and the beaker was transferred to an ice bath. The liquid in the beaker was diluted to 140 ml using warm de-ionized water and stirred continuously for 5 min. Finally, H2O2 was gradually added to the solution until a color change to bright yellow, which signifies completion of oxidation.

The resulting solid filter cake was obtained after suction filtration. About 500 ml of 10 % hydrochloric acid was used to clean the cake, followed by three cycles of final washing. Solid oxidized graphite was obtained. To remove impurities from strong oxidants introduced to the reaction, the resulting solid material was centrifuged for 30 min at a speed of 4000 r min−1 for three cycles. Finally, the solid material was dried in an 80 °C air-blowing drying chamber for 12 h to obtain flaky black solid GO. After dissolving in deionized water and sonicating for 2 h, the layers of the GO flakes were stripped to prepare the GO solution. A certain volume of tetra-n-butyl titanate and ethanol were vigorously stirred to obtain solution A. Glacial acetic acid, the GO solution (different concentrations based on mass ratio; 30 min of ultrasound was applied to release the flake structure of GO prior to use), and anhydrous ethanol were dissolved to obtain solution B. Solution B was slowly dropped into solution A, and 90 min of continuous stirring was performed to obtain a yellow–brown homogeneous titanium–GO solution. Brown crystals were obtained after aging of this sol at room temperature for 24 h and then drying at a constant temperature of 80 °C for 12 h to remove the organic solution. Finally, the support was ground, calcined in a tubular furnace in a nitrogen atmosphere, and kept warm at 450 °C for 6 h to obtain the TiO2–GO power. In experiments based on aqueous solutions of various concentrations, TiO2–GO supports with different mass percentages were prepared as follows: TiO2–2 % GO, TiO2–0.8 % GO, TiO2–0.4 % GO, and TiO2–0.2 % GO.

2.1.2 Preparation of MnOX/TiO2–GO Catalysts by Ultrasonic Impregnation

Using manganese acetate (C4H6MnO4·4H2O) as a precursor linked to the active component on the support, both liquor and support were mixed and treated by an ultrasonic device at 60 °C for 90 min to ensure full contact with the support and that the active components are uniformly attached to the surface of the support. The catalysts were placed in an oven set to 80 °C for drying. The materials were burned at 450 °C for activation in a nitrogen atmosphere and incubated for 3 h to prepare supported catalysts MnOX/TiO2–GO: X % MnOX/TiO2–GO where X % stands for the mass percentage of Mn elements in MnOX/TiO2–GO.

2.2 Catalyst Characterization

Microscopic characterization of the morphology of catalyst was performed using a Hitachi S4800 field emission scanning electronic microscope.

An X Flash Detector Type 5010 (BRUKER) to analyze elemental distributions on the surface of the catalysts.

Phase analysis of sample materials was done using a Rigaku DMAX-RB 12 kW rotating anode X-ray diffractometer. A copper target (wavelength, 0.15406 nm), 40 kV acceleration voltage, 150 mA current intensity, 10° min−1 scanning speed, and scanning range 2θ = 10°–100° were applied.

A QuadraSorb SI type analyzer was applied to determine the specific surface areas and pore structure parameters of the catalyst via the static nitrogen physical adsorption method.

By means of NEXUS 670 type FT-IR infrared instrument made by Thermo Nicolet samples were tested for Fourier transformation infrared spectroscopy. Samples were prepared by use of potassium bromide KBr Tablet. Before the test N2 was applied to blow off the impurities produced during the reaction process followed by sampling and analysis. The wave number used for the test was 400–4000 cm−1.

A Kratos Axis Ultra DLD multifunctional electronic spectrometer was used to analyze the XPS spectra of the catalysts. This instrument was coupled to a monochromatic Al Kα anode target and multi-channel delay detector. The whole spectrum scanning of 160 eV and narrow spectrum scanning of 40 eV. The analytical results were calibrated using the C 1s binding energy.

A JEOL JEM-2100HR high resolution transmission electronic microscope was used to observe the microstructural features of grouped samples loaded with active components.

2.3 Catalyst Activity Testing

Steady-state SCR reaction experiments were performed in a quartz tube fixed-bed continuous flow reactor using 500 mg catalyst of 60–100 mesh. The inner diameter of the quartz reaction tube was 9 mm, the reactor was placed in an electrically heated furnace with a programmable controller. The typical reactant gas composition included 0.1 vol% NO, 0.1 vol% NH3, 7 vol% O2, 10 vol% H2O (when used), 200 ppm SO2 (when used) with a balance of Ar. The Ar flow gas was divided into two branches. One branch converged with NO, NH3, O2, and SO2 to form the main gas flow, while the other one passed through a heated gas-wash bottle containing deionized water to introduce water vapor into the system when required. The feed flow rate was fixed at 600 ml min−1, which were controlled by mass flow controllers and corresponded to a gas hourly space velocity (GHSV) of 67,000 h−1, and the reaction temperature ranged from 80 to 180 °C. To test water resistance, H2O was added using a controlled evaporator mixer. In resistance to poisoning experiments, the SO2 concentration was 200 ppm. In selectivity tests, the content of N2 in the vent served as a reference. A gas chromatograph (GC 7890A; Agilent Company; United States) was applied with Ar as the support gas.

Using a Thermo Scientific 42i-HL flue gas analyzer, the concentration of NOx at the inlet and the outlet was determined to calculate the removal rate of NOx. Taking into account the effect of gas adsorption on the catalyst, NO was fed into the reactor at the initial stage of the experiment. When the concentration at the outlet had stabilized, NOX was fed into the reactor until the concentration of NH3 at the outlet was stable upon reading the data. The removal rate of NOx and selectivity of N2 were obtained using Formulas (1) and (2), respectively. Here, \(C_{{{\text{NO}}_{\text{x}} }}^{\text{in}}\), \(C_{{{\text{NO}}_{ 2} }}^{\text{in}}\), and \(C_{{{\text{NH}}_{ 3} }}^{\text{in}}\) respectively represent the concentrations of NOX, NO, and NH3 at the inlet. In the experimental system, only NO gas was fed into the reactor. However, because NO inevitably reacts with O2 in the air and generates NO2, we indicate the final results as NOX. \(C_{{{\text{NO}}_{\text{x}} }}^{\text{out}}\) and \(C_{{{\text{NO}}_{ 2} }}^{\text{out}}\) represent the concentrations of NOX and NO2, respectively, and \(C_{{{\text{N}}_{ 2} {\text{O}}}}^{\text{out}}\) represents the concentration of N2O at the outlet.

$${\text{NO}}_{\text{X}} \;{\text{conversion }}\left( \% \right) = 100 \times \frac{{\mathop {\text{C}}\nolimits_{{{\text{NO}}_{\text{X}} }}^{\text{in}} - \mathop {\text{C}}\nolimits_{{{\text{NO}}_{\text{X}} }}^{\text{out}} }}{{\mathop {\text{C}}\nolimits_{{{\text{NO}}_{\text{X}} }}^{\text{in}} }}$$
(1)
$${\text{N}}_{ 2} \;{\text{selectivity}}\;\left( \% \right) = 100 \times \frac{{\mathop {\text{C}}\nolimits_{\text{NO}}^{\text{in}} + \mathop {\text{C}}\nolimits_{{{\text{NH}}_{ 3} }}^{\text{in}} - \mathop {\text{C}}\nolimits_{{{\text{NO}}_{ 2} }}^{\text{out}} - 2\mathop {\text{C}}\nolimits_{{{\text{N}}_{ 2} {\text{O}}}}^{{}} }}{{\mathop {\text{C}}\nolimits_{{{\text{NO}}_{{}} }}^{\text{in}} + \mathop {\text{C}}\nolimits_{{{\text{NH}}_{ 3} }}^{\text{in}} }}$$
(2)

3 Results and Discussion

3.1 SCR Activity of MnOX/TiO2–GO Catalyst

3.1.1 Effect of Different Mass Ratios of GO

Figure 1 shows the NH3-SCR activity of 5 % Mn/TiO2–GO catalysts with different mass ratios of GO.

Fig. 1
figure 1

Effect of different mass ratios of GO of Mn/TiO2–GO catalysts on catalytic activity

Full size image

Comparing with sample of only Mn/TiO2 or Mn/GO support, the rising trend of the curve in Fig. 1 shows that increasing the amount of GO in the support increases the SCR activity of the catalyst and reached the maximum when Mn/TiO2 mixed with 0.8 % GO, activity up to 93.38 % at 180 °C, especially, the catalytic activity of 5 % Mn/TiO2–0.8 % GO was obviously better than the other samples at the low temperature range of 80–120 °C. When the amount of doped GO reached 2 %, the catalyst activity of 5 % Mn/TiO2–2 % GO began to decrease, possibly because of clogging of the pore structures. This observation shall be discussed further in terms of BET analysis of the supports. As the GO content increased, the activity of the 5 % Mn/TiO2–GO catalyst significantly improved compared with those of the 5 % Mn/TiO2 and 5 % Mn/GO single-component catalysts in the low temperature range of 80–180 °C. As the reaction temperature increased, the curves of the activity of all samples converged. Taken together, results indicate that TiO2–0.8 % GO is the best catalyst support for this experiment; as such, this catalyst was used for subsequent experiments and research.

3.1.2 Effect of Different Mn Loading and N2 Selectivity

Figure 2a, b indicate the relationships between different Mn loadings in the Mn/TiO2–0.8 % GO catalysts respectively, and the removal rate of NOX and N2 selectivity. Mn loadings of 0, 3, 5, 7, 9, and 11 % were applied in this experiment.

Fig. 2
figure 2

a Effect of different Mn loadings on the removal rate of NOX. b N2 selectivity and N2O formation over the various MnOX/TiO2–GO catalysts. Reaction conditions 0.1 vol% NO, 0.1 vol% NH3, 7 vol% O2, Ar to balance, GHSV 67,000 h−1, 500 mg of catalysts

Full size image

Figure 2a shows that increasing the Mn loading increases initially and then decreases the SCR activity of the Mn/TiO2–0.8 % GO catalysts. When the Mn loading of the catalysts reached to 9 %, the maximum removal rate of NOX up to 95 % at 180 °C. As the Mn loading increased to 11 %, the activity began to decline due to agglomeration of excess MnOX particles on the surface of the TiO2–0.8 % GO support. Nevertheless, the results were further elaborated by the following TEM and BET analysis.

Figure 2b shows the N2 product selectivity and N2O formation for these catalysts. All of the samples show high N2 selectivity over the entire reaction temperature range (all greater than 96 %). With increasing loading of the active component and reaction temperature, N2 selectivity in the catalysts showed a downward trend, although N2O levels generated remain low (<5 ppm). This result is due to N2O generation and NH3 oxidation reactions occurring at higher temperatures. Singoredjo [19] and other researchers found that, in MnOX/Al2O3 catalysts, increases in Mn loading and reaction temperature generates greater amounts of N2O and reductions in N2 selectivity. These results resemble findings in this experiment.

3.2 Catalyst Characterization Results

3.2.1 FESEM Morphology of the Support and Scanning Analysis of the EDS Surface

Figure 3 shows the FESEM patterns of GO and the TiO2–0.8 % GO support as well as scanning analysis results of the EDS mapping of the TiO2–0.8 % GO support; these results are used to observe the morphology of the support and the distribution of each element in the support surface.

Fig. 3
figure 3

FESEM patterns of the support and EDS energy pattern a GO, b, c TiO2–0.8 % GO, and d, f EDS mappings of the TiO2–0.8 % GO support

Full size image

The monolayer morphology of the prepared GO is shown in Fig. 3a. Figure 3b, c show the lamellar morphology and the cross-sectional topography of the TiO2–0.8 % GO support. The figures show that tiny white TiO2 particles are uniformly distributed on the surface layer of flaky GO. In the cross-sectional diagram, several stacked GO layers can be observed. Further, EDS mapping analysis proved that the TiO2 particles in TiO2–0.8 % GO were distributed on the surface of GO. Figure 3d–f represent the respective distribution statuses of C, O, and Ti on the surface of the support. The three elements were integrated in an orderly and tight manner, which confirms that the TiO2–GO supports prepared in this experiment present a uniform composition, which helps increase the specific surface area of the support, improves the pore structure characteristics of the support, and promotes uniform loading of the active component.

3.2.2 XRD Analysis Results of the Support

Figures 4a–c show the XRD spectra of as-received flaky graphite, GO, and TiO2–GO supports with different mass ratios respectively.

Fig. 4
figure 4

XRD spectra of a as-received graphite, b GO, and c TiO2–GO of different mass ratios

Full size image

Figure 4a demonstrates that the flaky graphite sample acquired for this experiment has high purity, and the graphite peaks appear only at around 2θ = 26°. No impurity peaks were observed for this material. Characteristic GO peaks appeared at a 2θ value of 9° in Fig. 4b. This small angular offset of the characteristic peaks shows excellent oxidation of GO in the present experiment. The absence of characteristic peaks at 2θ = 26° indicates that the flaky graphite was completely oxidized to GO. The XRD spectra of TiO2–GO with different mass ratios in Fig. 4c illustrates that all of the samples have characteristic GO peaks present at 2θ = 9°, which confirms successful combination of GO and TiO2. Anatase TiO2 had rich active sites, which enhanced the SCR activity of the catalyst, and the peaks at 2θ value of 25.3°, 37.8°, 48.1°, 53.9°, 55.0°, 62.7°, 68.7°, 70.3° and 75.1° were indexed to (101), (004), (200), (105), (211), (204), (116), (220) and (215) crystal planes (JCPDS 21-1272). Here, increases in mass ratio of TiO2–GO, slightly decreases and broadens the intensity of diffraction peaks. These findings indicate that introduction of GO does not alter the crystal structure of TiO2 and that the support completely maintains its anatase structure. This outcome is consistent with the research of previous scholars [20, 21].

3.2.3 TEM Analysis of MnOX/TiO2–GO Catalyst

Figure 5 shows TEM images of the Mn/TiO2–0.8 % GO catalyst with different Mn loadings. The distribution of the active component particles were dispersed well on the lamellar GO.

Fig. 5
figure 5

TEM images of different Mn loading catalysts a 3 % MnOX/TiO2–0.8 % GO, b 5 % MnOX/TiO2–0.8 % GO, c 7 % MnOX/TiO2–0.8 % GO, d 9 % MnOX/TiO2–0.8 % GO, and e 11 % MnOX/TiO2–0.8 % GO

Full size image

As shown in Fig. 5a–c, only a small amount of MnOX particles are supported on the support surface. As the Mn loading continues to increase up to 9 wt%, the surface of the GO sheets is covered by a uniform layer of MnOX particles with sizes ranging from 10 to 40 nm. They are all mesoporous materials. However, manganese particles appeared severe agglomeration when the manganese loading was increased to 11 wt%. Mutual coverage by MnOX particles masks some of the active sites and reduces opportunities for contact with NH3, thereby decreasing the activity of the catalyst. These characterization results verify our previous inference on changes in catalyst activity with varying Mn loadings.

3.2.4 BET Surface Areas and Pore Size Distributions

The detailed textural data of the specific surface area and pore structure of catalysts are shown in Table 1.

Table 1 Structural parameters of TiO2, TiO2–GO, and MnOX/TiO2–GO
Full size table

Analysis of the results of pore structure in Table 1 indicated that the specific surface area and pore volume of TiO2–GO increase first by added GO, and average pore size showed the opposite trend, reaching a maximum when the mass ratios of GO reached 0.8 %. The distribution of pore sizes in all of the samples fell within 2–10 nm, which indicated that all of the samples are mesoporous materials.

However, when the mass ratios of GO reached 2 %, the specific surface area appeared to decrease. This was due to the excess GO caused agglomeration which had blocked some mesopores and micropores, resulting in the slight decline in the specific surface area. This result also verifies the cause of the change in catalytic activity with respect to mass ratio. Analysis of the results in the given table shows that as the MnOx loading increases, the specific surface area and pore volume of Mn/TiO2–GO catalyst slightly decrease before reaching the catalytic activity expected at optimal loading. Beyond a certain loading, however, significant decreases in activity were observed. This tendency agrees with the trend previously reported by scholars. The specific surface area and pore volume of carbon-based supports, such as AC [22], ACF [23], and MWCNTs [24], decline after loading with metal oxides. In this experiment, the same trend was observed in the TiO2–GO support after MnOX loading. Upon analysis, incorporation of GO appeared to provide TiO2 with the characteristics of carbon-based materials. When the MnOX loading exceeded a critical value, as shown by the TEM morphology in the previous section, agglomerated was an important reason for the significantly reduced of the catalyst activity. Therefore, the catalytic activity of carbon-based titanium in this experiment is dependent not only on the specific surface area but also pore structure of the catalyst. Further analysis of the effect of additional GO on the redox capacity of catalyst must be performed through XPS.

3.2.5 XPS Analysis of the MnOX/TiO2–GO Catalysts

The concentrations of elements of MnOX/TiO2–0.8 % GO on the catalyst surface in Table 2. Figure 6 shows XPS analysis results of the 9 % Mn/TiO2–0.8 % GO catalyst; Fig. 6a shows XPS spectrum of C 1s, while Fig. 6b–d illustrate the XPS spectra of Ti 2p, Mn 2p, and O 1s, respectively.

Table 2 Atomic surface compositions of 9 % Mn/TiO2 and 9 % Mn/TiO2–0.8 % GO on the surface of the catalyst
Full size table
Fig. 6
figure 6

XPS spectrum of 9 % Mn/TiO2–0.8 % GO catalyst

Full size image

The surface element concentrations of the 9 % Mn/TiO2–0.8 % GO catalyst as well as its valence states were further analyzed by XPS with reference to the NIST XPS database to compare peaks obtained under different energy bindings. Figure 6a shows two main peaks at 285.0 and 286.6 eV, respectively, representing amorphous carbon and graphite carbon from GO, may also be observed [25, 26]. The weak peak appearing at 289.2 eV is attributed to O=C–O [27]. Oxygen-containing groups present in GO provide an abundance of active sites for connecting GO to TiO2 particles. This result suggests that, –OH on the surface of TiO2 particle may reacts with –COOH and C–OH functional group, the surface is formed GO O=C–O–Ti, and C–O–Ti bond respectively by esterification and dehydration reactions, to achieve a closer integration [28]. XPS analysis detected the atomic surface compositions of fresh 9 % Mn/TiO2–0.8 % GO catalysts.

From the XPS Ti 2p spectrum in Fig. 6b, two distinct curves which are Ti 2p1/2 and the Ti 2p3/2 spin–orbit splitting peaks at the binding energy of 464.3 and 458.6 eV are assigned to Ti4+. De-convolution peaks of the Ti 2p spectrum also appeared at binding energies of 465.8 eV (Ti 2p1/2) and 460.2 eV (Ti 2p3/2). By referring to the XPS database, the peaks were identified as two characteristic spectra line of Ti–C. The results show a combination of chemical bonds between TiO2 and GO. The characteristic XPS spectra line of Mn 2p and O 1s can be seen in Figs. 6c, d, wherein Mn exists in multiple forms, such as MnO, MnO2, Mn2O3, and MnOX/Mn, in the catalyst. Fitting of the de-convolution peaks showed three peaks of Mn 2p3/2 at binding energies of 640.8 eV (MnO), 642.0 eV (Mn2O3), and 643.4 eV (MnO2). Two peaks, 652.9 eV (MnOX/Mn) and 653.4 eV (MnO), belong to Mn 2p1/2. Results showed that the MnO and MnOX/Mn contents of the fresh 9 % Mn/TiO2–0.8 % GO catalyst obviously increased relative to those in the catalyst without GO, the results were consistent with the research achievements of Lu [29] before. The variety of MnOX species with different valences present in the catalyst, especially non-stoichiometric MnOX/Mn, activated electron mobility in the catalyst, which favors the redox reaction of the catalyst and further increases the removal rate of NOX in SCR reaction [28]. In Fig. 6d, XPS fitting of separated peaks of O 1s demonstrates two types of O in the catalyst: lattice oxygen on the surface of the catalyst (529.4–530.1 eV, denoted as Oβ) and chemisorbed oxygen (531.3–532.3 eV, denoted as Oα). Here, chemical adsorption of oxygen as O2 2− or O− occurs in the form of OH− and carbonate (CO3 2−) [30]. A comparison of the catalyst surface element concentrations of 9 % Mn/TiO2 and 9 % Mn/TiO2–0.8 % GO in Table 2 shows that introduction of GO not only improves the effective loading capacity of Mn but also enriches the MnOX species in the catalyst. Introduction of GO also increases the catalytic surface oxygen content, improves the oxygen storage capacity of the catalyst and the redox performance.

3.3 Impacts of H2O and SO2

To investigate the effect of H2O and SO2 on NOX conversion of 9 % Mn/TiO2–0.8 % GO catalyst, the volume fraction of 10 % water vapor and 200 ppm of SO2 for separate resistance to H2O and SO2 experiments were introduced to simulated flue gas for the transient response test. At the same time, 10 % water vapor and 200 ppm SO2 were fed into the system to carry out water resistance and resistance to sulfur experiments at 180 °C. The effects of H2O and SO2 on the catalytic activity of 9 % Mn/TiO2–0.8 % GO are shown in Fig. 7.

Fig. 7
figure 7

Effects of H2O and SO2 on the catalytic activity of 9 % Mn/TiO2–0.8 % GO and 9 % Mn/TiO2

Full size image

Figure 7 shows that the removal rate of NOX from the system becomes stable 0.5 h into the reaction. After 2 h, the reaction stabilized when water vapor and SO2 were added to the system to study the influence of poisoning on catalyst activity. Results show that the resistance to H2O ability of the 9 % Mn/TiO2–0.8 % GO catalyst is not significantly improved compared with a conventional low temperature SCR catalyst, which is attributed to the GO added to support. The surface of this support has many hydrophilic groups, such as carboxyl, hydroxyl, and epoxy groups. The water vapor can be easily adsorbed onto the catalyst surface and occupy active sites required by NH3 adsorption, which results in competitive adsorption and a decline in SCR activity. The independent resistance to SO2 poisoning of the catalyst was significantly improved. The catalytic activity after poisoning stabilized at around 78 %. After SO2 feeding was removed, the activity of this catalyst was restored to 83 %. The sulfur/water resistance increased to about 73 % and was restored to 79 % when feeding was removed.

The FT-IR spectra of the 9 % Mn/TiO2–0.8 % GO catalyst before and after poisoning are shown in Fig. 8 for explanation of this phenomenon. Peaks at 2925 and 3425 cm−1 are attributed to –OH vibrations, while the peak at 1407 cm−1 is attributed to C–OH bond vibrations. The peak at 1618 cm−1 is attributed to stretching vibrations of the C–O–C bond, that at 1709 cm−1 corresponds to the stretching vibrations of C=O bond, and that at 1226 cm−1 is attributed to C–O–C bonds of the epoxy base. At 1102 cm−1 of the fresh and poisoned catalyst SO4 2− could be observed, which may result from the large amount of concentrated sulfuric acid during the preparation of GO and is equivalent to the pre-sulfuring of catalyst, thereby inhibiting the sulfation of the active component when SO2 was fed and substantially increasing resistance to SO2 ability of the catalyst. The resistance mechanism to SO2 poisoning of the catalyst needs further research. Taking the results together, a new low temperature SCR catalyst with excellent resistance to sulfur and water was developed in this work. Our work provides a theoretical basis for adding SCR de-nitration devices for sintering flue gas in highly humid, rich in oxygen conditions with a certain amount of sulfur.

Fig. 8
figure 8

FT-IR patterns of fresh and poisoned 9 % Mn/TiO2–0.8 % GO

Full size image

4 Conclusion

In summary, TiO2–GO supports of different Ti/C mass ratios were successfully prepared by using an in situ sol–gel and ultrasonic impregnation methods, which features the largest specific surface area (160 m2 g−1), showed the best catalytic activity. Through XPS analysis, the high catalytic activity of 9 % Mn/TiO2–0.8 % GO catalyst was determined to come from introduction of the catalyst, which improves the transfer capability of electrons on the surface. MnOX of multiple valence was found to exist, and the non-stoichiometric MnOX/Mn maximize the Redox capacity of catalyst and significantly enhanced the removal rate of NOX. Among the catalysts prepared, the 9 % Mn/TiO2–0.8 % GO catalyst exhibited the highest SCR activity (up to 95 %) at 180 °C, and GHSV of 67,000 h−1. All of the samples showed excellent N2 selectivity of over 96 %. During investigation of the effects of feeding 10 % water vapor and 200 ppm SO2 separately into the flue gas at 180 °C on the catalyst activity of 9 % Mn/TiO2–0.8 % GO, the removal rate of NOX decreased to 91 % from 95 % before feeding. When feeding of H2O was stopped for 2 h, the removal rate of NOX gradually increased to the initial level. When SO2 was fed separately, the removal rate of NOX fell to 78 % and returned to 83 % when feeding is stopped. The removal rate of NOX dropped to 73 % when 10 % H2O and 200 ppm SO2 were fed at the same time and was restored to 79 % within 2 h when feeding was stopped. FT-IR analysis showed that introduction of GO induces pre-sulfuration effects on the catalyst itself and inhibits sulfation of the active component. Thus, resistance to H2O and –SO2 poisoning in the catalysts is significantly improved. This research provides a theoretical basis for adding SCR de-nitration devices for sintering flue gas in conditions of humidity, rich in oxygen, and with a certain amount of sulfur in the future.