Abstract

Removal of nitric oxides (NOx) from stationary and transportation sources has been desired for environmental benefits. Selective catalytic reduction (SCR) of NOx by NH3 is attractive for its cost effectiveness and high efficiency but still technically challenging in consideration of operable temperatures. In this research, MnOx-CeOx hybrid nanoparticles supported on graphene aerogel (MnOx-CeOx/GA) are fabricated as the monolithic catalysts for potential applications to low-temperature SCR. The impacts of the particle size along with the amount and valency of catalytic elements in the nanocomposite on the catalytic activities are studied with the help of scanning electron microscopy (SEM), transmission electron microscopy (TEM), and X-ray photoelectron spectroscopy (XPS). The catalyst crystallites are a few tens of nanometers and uniformly disperse on the surface of three-dimensional (3D) directionally aligned hierarchical porous graphene aerogel (GA) networks. The novel nanocomposite catalysts exhibit over 90% NOx conversion rate in a broad temperature range (200–300°C). Addition of CeOx into the MnOx-GA catalysts significantly reduces the operational temperature at the same conversion rate. In addition to Mn4+ ions in the catalysts, the adsorbed oxygen species which can be increased by the presence of low-valence cerium contribute to high catalytic activities in the MnOx-CeOx/GA catalysts.

1. Introduction

Nitric oxides (NOx) have raised global awareness for their serious impacts on the environment such as acid rain, photochemical smog, ozone depletion, and greenhouse effects. With the increasingly stringent legislation and regulation on NOx emission, both industrial and academic laboratories have dedicated extensive efforts to developing novel catalysts towards effective reduction of NOx. Among various technologies, selective catalytic reduction (SCR) of NOx with NH3 as reductant is favorable for its low cost, high efficiency, and good stability [1, 2]. However, the present commercialized SCR catalysts applied in power plants and diesel vehicles, represented by V2O5-WO3/TiO2, still encounter some inevitable challenges. In particular, the high operating temperatures (300~400°C) in diesel engine emission result in a technical dilemma [36]. If the SCR system is installed before the dust collector and desulfurization tower, the catalyst is readily poisoned and deactivated. If the SCR system is located after desulfurization and particle removal process, the temperature of the flue gas can hardly meet the catalyst requirement. Therefore, developing high-performance SCR catalysts operable at temperatures lower than 300°C, preferable in the range of 100–250°C, is highly demanded.

Mn-based catalysts are attractive candidates because they exhibit good low-temperature SCR activities [1, 2, 711]. Ce is commonly added in theses catalysts for the benefits of improved catalytic efficiency. According to Qi et al. [7], MnOx-CeOx catalyst with the mole ratio Mn/(Mn + Ce) of 0.4 increased NO conversion starting at 150°C. Various supporting materials for the catalysts are reported, such as Al2O3, TiO2, and active carbon fibers. Appropriate support not only provides large surface area to anchor the catalysts but also supplies the space to facilitate catalyst homogeneous distribution. Cao et al. [12, 13] found out that in the Mn-Ce/γ-Al2O3 series the high surface area of γ-Al2O3 was an important factor to increase the active catalytic sites. Lee et al. [9] reported that MnOx-CeO2 catalyst supported on TiO2 exhibited high NO conversion of 30~90% from 120 to 180°C at a space velocity of 60,000 h−1. Although the MnOx/TiO2 catalysts exhibited much higher catalytic activity at temperature below 200°C, they have various deficiencies including low acidity, low surface area, and poor temperature endurance [10]. Better performances using carbon-based supports such as activated carbons, carbon nanotubes, and graphene have been demonstrated in SCR under excess oxygen and temperatures around 200°C [11].

We recently fabricated MnOx-CeOx nanoparticles supported on three-dimensional graphene aerogel (GA), i.e., MnOx-CeOx/GA, of different compositions and explored their SCR performances at low temperatures. Graphene aerogel inherits all the advantages of 2D graphene such as superior electrical conductivity, high chemical stability, and good mechanical properties [14]. In addition, graphene aerogel possesses the characteristics of 3D aerogel which has ultra-low density and porous network on micro-, meso-, and macroscales. The hierarchical pore structure in GA provides not only high surface area for anchoring nanoparticles but also good accessibility for gases to the active surface. These structure and property merits render GA an ideal catalyst support. In the study, the novel nanocomposite catalysts MnOx-CeOx/GA were subjected to systematic morphological, structural, compositional, and chemical analyses in order to gain insights of the key factors correlated with their catalytic activities.

2. Experimental

2.1. Synthesis of GA-MnCe Nanocomposite Catalysts

Firstly, graphene oxide (GO) was synthesized using the modified Hummers method [15, 16]. Typically, in the experiment, 1 gram graphite (300 mesh with 99.9% purity) and 0.5 gram sodium nitrite were mixed in the 70 mL concentrated sulfuric acid. Then, 3 grams of potassium permanganate was gradually added to the mixture and stirred in a water bath at 35°C. After 2 hours, 15 mL hydrogen peroxide was slowly added to the mixture until the color of the mixture turned into bright yellow. The resultant mixture was then rinsed thoroughly with deionized water and diluted hydrochloric acid (10 wt%) until the pH value reached around 7. The resulted graphene oxide powders were filtered out and completely dried in a vacuum furnace.

An appropriate amount of the as-prepared GO powders was added into an aqueous solution of manganese acetate with or without cerium nitrate at the predetermined composition. Gradually, ammonium hydroxide was added until the pH value of the solution reached 10. After the precipitation reaction finished, the concentration of GO in the solution was adjusted to 2 mg/mL and ultrasonicated for 2 hrs. Later, ethylenediamine was added as the reducing agent. The mixture was then sealed in a cylindrical vessel and heated at 95°C for 6 hrs. Subsequently, the obtained hydrogel was on dialysis for 24 hours, followed by the freeze-drying process. To remove all the volatile components, the aerogel complex was heated to 500°C at the ramping rate of 10°C/min in a tube furnace under the flowing Argon gas. The as-prepared MnOx-CeOx/GA catalysts were referred to as GA-Mn(X) or GA-Mn(X)Ce(Y), where represents the nominal weight percentage in the composite and represents the molar ratio of Mn/Ce. The specimens and their corresponding nominal compositions are listed in Table 1.

2.2. Material Characterizations and Analyses

The morphologies of the as-prepared MnOx-CeOx/GA nanocomposite catalysts were analyzed with the help of field emission scanning electron microscopy (FESEM, ZEISS-ΣIGMA HD) and transmission electron microscopy TEM (JEOL 2100).

The weight percentages of manganese and cerium in the catalysts were determined by using inductively coupled plasma (ICP) atomic emission spectroscopy (Thermo IRIS Intrepid II). For this analysis, the catalysts were dissolved in an acid solution.

The phases of the components in the catalysts were determined using X-ray diffraction (XRD) (PANalytical X’pert Powder Diffractometer,  nm). The operating current and voltage were 40 mA and 40 kV, respectively. The diffraction profiles were collected in the range of 10–80° at the step length of 0.013 s and the residence time of 5 s.

The elements’ bonding state and valency were analyzed using X-ray photoelectron spectroscopy (XPS) collected on ESCALAB250Xi (Thermo-Fisher). The XPS peaks were fitted to Voigt functions using the XPSpeak 4.1 software having 80% Gaussian and 20% Lorentzian character, after performing a Shirley background subtraction.

2.3. SCR Activity Tests

The SCR activity tests were performed on a fixed-bed reactor in the temperature range of 60 to 320°C. The reactor contains 2 mL catalyst with a gas space velocity of 30,000 h−1. The schematic reactor diagram is shown in Figure 1. A simulate gas consists of 500 ppm NO, 500 ppm NH3, 5 vol. % O2, and balanced with N2. Flue gas analyzer (ZR-3200, Qingdao) was used to analyze the concentrations of NO, NO2, and O2 at the gas outlet. The activity data were collected and recorded when the NH3-SCR reaction reached the steady state at each temperature point. NO conversion was calculated as follows: where and represent the inlet and outlet concentration of NO, respectively.

3. Results and Discussion

The ICP results of the as-prepared GA-MnCe catalysts are listed in Table 2. Apparently, the actual weight percentages of both Mn and Ce loadings in the composite catalyst products are in general consistent with the nominal compositional values, with error less than 1 wt%.

The morphology and microstructure of the as-prepared nanocomposite catalysts GA-Mn and GA-MnCe series are elucidated by FESEM and TEM. Figures 2(a)2(h) exhibit representative images illustrating the morphological characteristics of the GA-Mn and GA-MnCe catalysts. Clearly seen in Figures 2(a) and 2(b), the graphene aerosol has a three-dimensional network structure made up of directionally aligned macropores with the pore diameter in the range of 20–30 μm. On the pore wall, there also exist pores of submicrometers. High-resolution TEM reveals that the pore walls are self-assembled by the intertwisted graphene sheets around 5 nm thick (see Figure 2(g)). The large pores and thin porous walls of the GA network structure ensure the high gas flow and large amount of anchor sites for nanoparticle catalysts.

As seen in Figures 2(c) and 2(d), the MnOx catalyst nanoparticles disperse uniformly on the surface of graphene sheets. Upon increasing the catalyst loading, more particles are visualized (compared Figure 2(c) for GA-Mn5 with Figure 2(d) for GA-Mn10). When increasing MnOx loading to 15–20 wt%, some agglomerates in the micrometer size are also observed (see Figure 2(e) for GA-Mn15). Such agglomerates are anticipated to reduce the effective active sites and hence catalytic activity.

High-resolution SEM images reveal that the sizes of catalyst particles, whether MnOx or CeOx, are a few tens of nanometers. Figure 2(f) is a representative image. Under TEM, it is also seen that some catalysts are having crystallite size less than 10 nanometers (see Figure 2(g)), which may be hardly visible under SEM.

The XRD profiles of the as-prepared GA-Mn and GA-MnCe hybrid materials are shown in Figure 3. In both GA-Mn and GA-MnCe series, there are broad peaks located at 26° and 44°, which originate from graphene aerosol supports. These two peaks correspond to the diffractions from (002) plane and (100) plane in the hexagonal graphitic structure, respectively. In addition, there are peaks observed at 18.3°, 32.7°, 35.2°, 40.7°, and 59.1° in the GA-Mn catalysts (see Figure 3(a)), confirming the coexistence of crystalline Mn3O4 and MnO. As the Mn loading increased, these peaks are more prominent. Upon adding Ce and gradually increasing the Ce loading, it is seen that the peaks from MnOx become weaker and broader. For the samples with high content of Ce, like GA-Mn10Ce41, GA-Mn10Ce21, and GA-Mn10Ce11, the distinguishable peaks located at 28.7°, 47.6°, 56.8°, and 78° are all from CeO2. These observations indicate the potential formation of amorphous or nanocrystalline solid-solution of MnOx and CeOx. The occurrence of Mn4+ replacing Ce4+ lattice position in the fluorite structure is most likely due to their size and structural similarity [17].

Figure 4 presents the NO conversion efficiency as a function of operating temperature in the presence of GA-Mn or GA-MnCe catalysts with different manganese loading or different Mn/Ce molar ratio. The SCR catalytic activities of the catalysts are strongly dependent on the operating temperatures. For the GA-Mn catalyst series, the NO conversion efficiency increases exponentially at low-temperature regions. The conversion efficiency reaches maximum at 280°C. Further increasing the temperature to 310°C, the NO conversion starts to decrease due to the NH3 oxidation, which is a competitive reaction with SCR reaction [18, 19]. Among the four Mn loading, GA-Mn5 has the lowest NO conversion attributed to the insufficient Mn loading. The catalytic activity increases with Mn loading and appears to saturate when Mn loading is over 10 wt%. The slightly lower performances of GA-Mn15 and GA-Mn20 may be related to the agglomeration of the excessive amount of MnOx nanoparticles, manifested in SEM images. The enlarged agglomerates will lead to the reduced active sites and hence the reduced conversion efficiency.

As seen also in Figure 4, the GA-MnCe catalysts exhibit superior low-temperature activity compared to the GA-Mn catalysts. The GA-Mn10Ce catalysts show 80% NO conversion at 150°C, which is over 50°C lower than GA-Mn10. The novel nanocomposite catalysts exhibit over 90% NO conversion rate in a broad temperature range (200–300°C). The molar ratio of Mn/Ce has slight impacts on the SCR activities for the GA-MnCe catalysts. GA-MnCe41 is the best throughout the entire temperature range and reached the highest NO conversion of 98% at 250°C.

Further, the stability/durability of the GA-MnCe catalysts was assessed. Figure 5 shows NO conversion efficiency as a function of time obtained from the GA-MnCe series at 250°C. At the initial stage, the NO conversion efficiency reduced gradually with time on stream for all the four GA-MnCe catalysts, possibly resulting from the catalyst sintering. However, the conversion efficiency tends to reach a steady value after a few hours. The GA-Mn10Ce41 remains the highest SCR catalytic activity with efficiency of 89% after 14 h continuous conversion.

To gain insights of the key factors affecting the SCR catalytic activities, the elemental composition and valency of nanocomposite catalysts anchored on GA surface were semiquantified with the help of XPS analyses. The survey spectra confirm the existence of four elements, i.e., C, O, Mn, and Ce, in the GA-MnCe nanocomposite catalysts. Figures 6(a)6(g) present the typical deconvoluted spectra of C1s, O1s, Mn2p, and Ce3d spectra. For comparison, the spectra of GA-Mn (Figures 6(a), 6(c), and 6(e)) and GA-MnCe (Figures 6(b), 6(d), 6(f), and 6(g)) are presented side-by-side. The values of the computed composition are summarized in Table 3.

As seen in Figures 6(a) and 6(b), the C1s spectra can be deconvoluted into five peaks both in GA-Mn and in GA-MnCe series. Peak centered at 284.3 eV is well known corresponding to sp2 C-C bonded carbons. The peak at 284.7 eV can be assigned to sp3 component in the amorphous carbon residue. The peaks between 285 eV and 289 eV resulted from carbons bonded with oxygen, including single bond C-O (epoxy, carboxyl) and double bond C=O (carbonyl and quinzone) functional groups on the graphene surfaces. The weak broad peak around 290 eV is related to - shakeup [20, 21]. Based on the relative intensities, it is calculated that the amount of carbon bonded with oxygen is in average around 20% in all the specimens.

The O 1s spectra were fitted well with three peaks (see Figures 6(c) and 6(d)). The peak around 529.5 eV represents the lattice oxygen bonding with Mn or Ce (denoted as Olatt). It is noticeable that the binding energy of Olatt shifts to a low value with the addition of Ce. The amount of Olatt in GA-MnCe is lower than that in GA-Mn10. The peak at 532.7 eV is correlated with oxygen anchored on the graphene surface associated with C-O bonding, which is denoted as OC. Oxygen species bonded to carbon or metal usually have insignificant impacts on the SCR reactions [22, 23]. The peak centered at 531.5 eV is assigned to the chemisorbed oxygen like O- and on the catalytic metallic oxides and hence denoted as Oads [24, 25]. The chemisorbed oxygen Oads is known as the strong oxidizing agent which is involved in the activation of NH3 and oxidation of NO in the SCR reaction. It is noteworthy that this peak is higher in GA-MnCe than in GA-Mn (compare Figures 6(c) and 6(d)). Quantitatively, the amount of chemisorbed oxygen Oads in GA-Mn is consistently lower, i.e., over 5%, than that in GA-MnCe (see Table 3). The observed higher Oads in the GA-MnCe catalysts corroborated well with the better NO conversion efficiency at lower temperatures.

In the Mn 2p spectra, a spin-orbit doublet of Mn 2p 1/2 and Mn 2p 3/2 with a binding energy gap of 11.5 eV is observed. The deconvoluted Mn 2p 3/2 spectra for all samples display in four peaks (see Figures 6(e) and 6(f)). The first three peaks with binding energies of 640.5 eV, 641.8 eV, and 643.2 eV originate from the different valency of manganese, i.e., Mn (II), Mn (III), and Mn (IV), respectively [11, 26, 27]. The fourth peak at 644.5 eV is a shakeup from the charge transfer between the outer electron shell and an unoccupied orbit with higher energy during the photoelectron process. The Mn 2p binding energies in GA-MnCe, compared those in GA-Mn, slightly shift to lower values, confirming the existence of interactions between manganese and cerium and formation of a solid solution. This observation is consistent with Qi and Yang’s report [28] and corroborates with our XRD results. As seen in Table 3, GA-Mn5 has the lowest amount of Mn4+, i.e., 12%, while all others contain Mn4+ in the amount of 17–19%. Since GA-Mn5 has the lowest catalytic activity throughout the operating temperature, Mn4+ is believed to be another key factor affecting the SCR catalytic activity. Mn4+ readily promotes SCR reaction because its strong oxidizing state enhances the oxidation of NH3 [13].

Ce 3d XPS spectra are presented in Figure 6(g). In the binding energies range 880–890 eV, the first three characteristic peaks are attributed to Ce3+ and Ce4+ [29]. The assignment of the last peak is inconclusive, which varies from Ce3+ to Ce2+. In all the GA-MnCe catalysts, peaks from Ce4+ are prevailing but there exists a decent amount of Ce3+ species. As seen in Table 3, upon the addition of Ce into the catalyst, the percentage of Olatt is reduced from average 43% to 29%. In contrast, Oads and Oc increase to 40% and 30%, respectively. It is reported that the existence of Ce3+ causes the oxygen deficiencies and unsaturated chemical bonds, which cannot only promote the transformation of Olatt to Oads but also help to adsorb more gaseous oxygen [30, 31]. As a consequence, the transfer process of the active oxygen is accelerated and the SCR activity is enhanced, which is observed in this study.

4. Conclusions

In this study, a series of novel 3D monolithic MnOx-CeOx nanocatalysts supported on graphene aerogel, i.e., MnOx-CeOx/GA, was successfully fabricated through an in situ hydrothermal and freeze-drying method. The GA support has directionally aligned micropore networks and thin porous walls, ensuring the high gas flow and large amount of anchor sites for nanoparticle catalysts. The catalyst nanoparticles are a few tens of nanometers and uniformly disperse on the surface of GA. The catalytic activity increases with Mn loading and appears to saturate when Mn loading is over 10 wt%. The slightly reduced performances of GA-Mn series at higher Mn loadings are attributed to the agglomeration of the excessive amount of MnOx nanoparticles. Upon addition of cerium, the agglomeration is alleviated and the SCR operable temperature is lowered by 50°C at the same NO conversion rate. The novel nanocomposite GA-MnCe catalysts exhibit over 90% NOx conversion rate in a broad temperature range (200–300°C). GA-MN10Ce41 has over 98% conversion efficiency at 250°C with relatively good stability. Based on the XPS analyses, Mn4+ and Oads are believed to be the two key factors contributing the SCR catalytic activity in the as-prepared GA-Mn and GA-MnCe series. Increasing Mn loading and adding Ce can slightly increase the amount of Mn4+. The presence of Ce3+ in the GA-MnCe catalysts appears to promote the transformation of Olatt to Oads leading to the enhancement of the SCR activity.

Data Availability

Raw data were generated at corresponding facilities described in the paper. Upon the acceptance/publishing of the paper, the derived data supporting the findings of this study are available from the corresponding author [YZ] upon request. YZ’s contact info is [email protected]; +8613029368800.

Conflicts of Interest

The authors declare that there is no conflict of interest regarding the publication of this paper.

Acknowledgments

The financial support of National Science-Technology Support Plan Projects of China (no. 2014BAB02B03) is gratefully acknowledged.