Soot Oxidation over γ-Al₂O₃-Supported Manganese-Based Binary Catalyst in a Dielectric Barrier Discharge Reactor

Abstract

In this work, soot oxidation was conducted over a series of Mn-X/γ-Al₂O₃ (M = Ce, Co and Cu) binary catalysts in a dielectric barrier discharge reactor. The soot conversion in the plasma–catalytic system was in the order of Mn/γ-Al₂O₃ (57.7%) > Mn-Co/γ-Al₂O₃ (53.9%) > Mn-Ce/γ-Al₂O₃ (51.6%) > Mn-Cu/γ-Al₂O₃ (47.7%) during the 30 min soot oxidation process at 14 W and 150 °C. Meanwhile, the doping of Ce, Co and Cu slightly improved the CO₂ selectivity of the process by 4.7% to 10.3% compared to soot oxidation over Mn/γ-Al₂O₃.It is worth to note that the order of CO₂ selectivity was in the opposite order with soot oxidation rate. The effects of discharge power, oxygen content in the carrier gas and reaction temperature on plasma–catalytic soot oxidation was systematically analyzed. The catalyst characterizations, including N₂ adsorption–desorption, X-ray diffraction, X-ray photoelectron spectroscopy, temperature-programmed reduction by H₂ and temperature-programmed desorption of O_2, were conducted to illustrate the reaction mechanisms of plasma–catalytic soot oxidation and reaction pathways.

1. Introduction

The diesel engine is widely used as a modern power source due to its advantages in power, fuel economy and high reliability compared to gasoline engines. However, particulate matter (mainly soot, namely elemental carbon) emissions from diesel engines could result in an acute impact on both human health and the environment [1]. In practice, diesel engine filters (DPFs) are used to capture soot and meet the stringent emission legislations for diesel engines. The captured soot is further oxidized for DPF regeneration at the exhaust temperature via many methods, including catalytic regeneration, fuel burners, external heating and recirculation of exhaust gas [2,3]. Among these methods, catalytic DPF regeneration is promising since the ignition temperature could be reduced to 400–500 °C compared with the temperature of above 600 °C without a catalyst [4,5]. However, the ignition temperature is still significantly higher than the exhaust temperature of engines, especially for the idling working conditions [6,7]. Thus, the catalytic regeneration of DPFs could not work properly under these conditions.

To overcome the drawbacks of catalytic DPF regeneration at low temperatures, plasma–catalysis has emerged as a promising method due to its unique characteristics of a compact system, quick response and mild reaction conditions [8,9,10]. In typical air plasma, highly energetic electrons (1 eV to 10 eV) could be generated. The electrons could then collide with background gas molecules to generate short-lived chemically reactive radicals (e. g., O· and O₃) for soot oxidation. However, the relatively low surface area of soot particles makes it difficult to achieve both a high soot oxidation rate and CO₂ selectivity of the process in cases of using plasma alone [9,11]. The integration of plasma and a heterogeneous catalyst, also known as “plasma catalysis”, could result in a synergistic effect since the short-lived radicals play a crucial role in activating effective collisions between oxidative species and soot on catalyst surfaces [12].

The composition of a catalyst imposes significant effects on the reaction performance of plasma–catalysis, including gas-phase pollutant removal, NH₃ synthesis, water purification, steam reforming, etc. [13,14]. As for soot oxidation, Ranji-Burachaloo reported the soot oxidation rate in a pin-to-plate plasma reactor over Mn, Fe and Co oxides. The CO₂ selectivity of the process reached the maximum values of 90%, 85% and 79% at the discharge power of 7.4 W and 350 °C, while energy yields of soot conversion were 7.0, 6.0 and 5.6 g kWh⁻¹, respectively [10]. Liu et al. found that the synergistic effect between plasma and LaAgMnCoO₃ perovskite could greatly improve the soot oxidation below 200 °C in a dielectric barrier discharge (DBD) plasma reactor. The soot conversion and CO₂ selectivity were 63% and 90% in a plasma–LaAg_0.2MnCo_0.2O₃ system, respectively, compared to 14% and 75% in the plasma–LaMnO₃ system [15]. Manganese has shown potential as the active metal of catalytic oxidation of volatile organic compound oxidation and soot [16,17]. Chen et al. prepared a series of MnO_x/wire-mesh monoliths for soot oxidation. The sample calcined at 700 °C showed a T₉₀ value of 421 °C and CO₂ selectivity of 93.2%, which were significantly improved compared with 618 °C and 46.5% for pure soot oxidation [18]. Yu et al. reported that the T₉₀ value of the MnCe₁O_x catalyst (370 °C) was 237 °C lower than that of pure soot, while the CO₂ selectivity (99.6%) was 47.2% higher compared to that of pure soot [19]. However, there is very limited literature on plasma–catalytic soot oxidation over Mn-based catalysts to the best of the authors’ knowledge, so the effect of various reaction factors on the reaction performance and the underlying mechanisms are not fully understood [20].

In this work, three kinds of Mn-X/γ-Al₂O₃ (X = Ce, Co and Cu) catalysts were prepared using wet impregnation method for soot oxidation in a DBD plasma reactor, while the effects of various parameters on plasma–catalytic soot oxidation were analyzed. A series of catalyst characterizations, including N₂ adsorption–desorption, X-ray diffraction, X-ray photoelectron spectrum, temperature-programmed reduction by H₂ (H₂-TPR) and temperature-programmed desorption of O₂ (O₂-TPD), were conducted to illustrate the textural and chemical properties of the Mn-X/γ-Al₂O₃ catalysts. The correlations between plasma-induced soot oxidation performance and the catalyst properties were discussed together with the underlying mechanisms.

2. Results and Discussions

2.1. Textural Properties of Mn-X/γ-Al₂O₃ Catalysts

Table 1. Physicochemical properties of the Mn-X/γ-Al₂O₃ catalysts.

Sample	SBET (m2·g−1)	Total Pore Volume (cm3·g−1)	Average Pore Diameter (Å)	Oads/(Oads+Olatt) (%)	H2 Consumption (mmol g−1)	O2 Desorption (mmol g−1)
						<350 °C	>350 °C
MnOx/ γ-Al2O3	208	0.38	7.4	77.8	0.36	0.69	2.50
Mn-Ce/ γ-Al2O3	190	0.37	7.9	74.9	0.35	0.71	1.66
Mn-Co/ γ-Al2O3	193	0.37	7.7	75.8	0.37	1.05	1.83
Mn-Cu/ γ-Al2O3	171	0.34	7.8	72.0	0.35	0.93	2.14

shows the specific surface area, pore volumes and average pore diameter of the Mn-X/γ-Al₂O₃ (X = Ce, Co and Cu) catalysts. All catalysts possessed type V isotherms and H4-type hysteresis loops, indicating the existence of narrow slit-like pores in the Mn-X/γ-Al₂O₃ catalyst, and the loading of active metal did not significantly affect the textural properties of the γ-Al₂O₃ support [21,22]. The specific surface area (S_BET) of pure γ-Al₂O₃ was 224.6 m²·g⁻¹, while its pore volume and average pore diameter were 0.39 cm³·g⁻¹ and 6.9 Å. The doping of Mn species slightly decreased the S_BET value to 207.6 m²·g⁻¹, and the S_BET value further decreased to between 171.4 m²·g⁻¹ and 192.5 m²·g⁻¹ after promoter (Ce, Co or Cu species) doping. The pore volume of the Mn-X/γ-Al₂O₃ catalysts (0.34–0.37 cm³·g⁻¹) was slightly lower than that of Mn/γ-Al₂O₃ (0.38 cm³·g⁻¹) and the pure γ-Al₂O₃ support (0.39 cm³·g⁻¹). The doping of active metal on the γ-Al₂O₃ support may not have covered the γ-Al₂O₃ surface and partly blocked the pores within the γ-Al₂O₃ support, leading to the decrease in the S_BET value and pore volume of the Mn-X/γ-Al₂O₃ catalysts. As for the average pore diameter, the values for Mn-X/γ-Al₂O₃ catalysts (7.4–7.9 Å) were slightly higher than that of γ-Al₂O₃ (6.9 Å). This phenomenon could be attributed to the clogging of pores with a small diameter within γ-Al₂O₃ after active metal loading. Similar results were reported in our previous work and by Zakaria et al. [23].

The XRD patterns of the Mn-X/γ-Al₂O₃ catalysts are given in Figure 1. The main diffraction peaks of all Mn-X/γ-Al₂O₃ and Mn/γ-Al₂O₃ catalysts at the 2θ of 37.6°, 45.9° and 67.0° were in good agreement with the cubic structure of pure γ-Al₂O₃ (JCPDS No. 00-010-0425), indicating that the crystalline structure of γ-Al₂O₃ was maintained after active metal doping. The intensities of diffraction peaks for Mn-X/γ-Al₂O₃ catalysts were slightly lower than that of pure γ-Al₂O₃, which could be ascribed to the dispersion of dopants on the surface of the support, and part of the dopants may have entered the crystal form of the support and formed a solid solution [24,25]. The diffraction peaks at 42.8° belonged to the MnO₂ phase (JCPDS No. 03-065-2821), which were observed for the four catalysts. In addition, weak diffraction peaks of CeO₂ (JCPDS No. 03-065-5923), Co₃O₄ (JCPDS No. 00-042-1467) and CuO (JCPDS No. 01-089-5899) could be found in the corresponding XRD pattern, indicating that active metal oxides existed in the form of crystallites on the γ-Al₂O₃ support. Moreover, Mn-Ce/γ-Al₂O₃ showed strong diffraction peaks of CeO₂ species. The crystalline sizes of the Mn, Ce, Co and Cu oxides were 5.2 nm, 4.3 nm, 3.7 nm and 7.6 nm, respectively, as calculated using the Scherrer equation. The small particle sizes would benefit the heterogeneous catalytic reactions [26].

2.2. Chemical Properties of Mn-X/γ-Al₂O₃ Catalysts

XPS was performed to investigate the valence and relative content of each element on the catalyst surface. The O 1s XPS spectra of the Mn-X/γ-Al₂O₃ catalysts are presented in Figure 2. The O 1s spectra could be deconvoluted into two major peaks. The peaks around 531.3 eV corresponded to the surface-adsorbed oxygen (O_ads), while the peaks around 529.6 eV could be ascribed to lattice oxygen (O_latt) species [27]. The relative concentrations of O_ads/(O_ads+O_latt) over the four catalysts are listed in Table 1. The O_ads/(O_ads+O_latt) value slightly decreased by 2.0–5.8% after the doping of Ce, Co and Cu on the Mn/γ-Al₂O₃ catalyst compared with Mn/γ-Al₂O₃ (77.8%), while the highest O_ads/(O_ads+O_latt) value of 75.8% was observed over Mn-Co/γ-Al₂O₃ among the Mn-X/γ-Al₂O₃ catalysts. It is widely reported that the O_ads species are closely related to the oxygen vacancies on the catalyst surface. As a result, the catalyst with a higher O_ads/(O_ads+O_latt) value would effectively promote the performance of heterogeneous catalytic reaction [28].

The reducibility of the Mn-X/γ-Al₂O₃ catalysts was analyzed through a H₂-TPR experiment (Figure 3). The H₂-TPR profiles of Mn/γ-Al₂O₃ showed two major reduction peaks at 297 °C and 410 °C. As the pure γ-Al₂O₃ support had no reduction peak between 100 °C and 600 °C [29], the first reduction peak could be attributed to the reduction of Mn⁴⁺ to Mn³⁺, while the second peak belonged to the reduction of Mn³⁺ to Mn²⁺ species [30]. The doping of Ce, Co and Cu significantly changed the reduction behavior of Mn-X/γ-Al₂O₃ catalysts. For Mn-Ce/γ-Al₂O₃, the two reduction peaks shifted to lower temperatures of 290 °C and 357 °C, respectively, indicating the mutual effect between Mn and Ce species on the catalyst surface [31]. In the case of Mn-Cu/γ-Al₂O₃, the predominant reduction peaks were observed at 276 °C and 362 °C, which could be ascribed to the reduction of Cu²⁺ and Mn⁴⁺ species [32]. The H₂-TPR profile of Mn-Co/γ-Al₂O₃ showed a broad peak between 272 °C and 379 °C, which belonged to the reduction of Mn⁴⁺ and Co³⁺ [24]. The lower reduction temperatures implied better oxygen mobility of the Mn-Co/γ-Al₂O₃ and Mn-Cu/γ-Al₂O₃ catalysts compared with Mn-Ce/γ-Al₂O₃ and Mn/γ-Al₂O₃ catalysts. To gain a better insight into reducibility, H₂ consumption of the catalysts was calculated (Table 1). It could be observed that all four catalyst samples exhibited a similar H₂ consumption amount due to the fixed loading amount of metal dopants. It could be deduced that the reducibility of Mn-Co/γ-Al₂O₃ and Mn-Cu/γ-Al₂O₃ was much better than that of Mn-Ce/γ-Al₂O₃ and Mn/γ-Al₂O₃ catalysts, taking the reduction temperature and H₂ consumption amount into consideration.

O₂-TPD experiments were conducted to study the nature of surface oxygen species, as shown in Figure 4. The O₂-TPD profiles for the four catalysts were quite similar. The desorption peaks between 100 °C and 350 °C were attributed to the adsorbed surface oxygen species (α-O₂) on oxygen vacancies, while the peaks between 350 °C and 700 °C belonged to chemically adsorbed oxygen species (β-O₂) [33]. The surface oxygen species would go through the transformation process of O₂(ad) → O₂⁻(ad) → O⁻(ad) → O₂⁻(ad/latt) via electron gaining [34]. The O₂ desorption amount is presented in Table 1, while the peaks above 750 °C, which could be assigned to lattice oxygen, were not observed in this work. Compared with pure γ-Al₂O₃, the peak areas of α-O₂ for Mn-X/γ-Al₂O₃ catalysts were in the range of 0.71 mmol·g⁻¹ to 1.05 mmol·g⁻¹, which were much higher than that of Mn/γ-Al₂O₃ (0.69 mmol·g⁻¹). The highest α-O₂ value was obtained over the Mn-Co/γ-Al₂O₃ catalyst. On the other hand, the peak areas of β-O₂ for Mn-X/γ-Al₂O₃ catalysts were lower than that of Mn/γ-Al₂O₃. These results are in accordance with the XPS spectra and H₂-TPR profiles. The adsorbed oxygen species (O_ads) are closely associated with the oxygen vacancies and oxygen mobility on catalyst surfaces, which could in turn contribute to the process of soot oxidation on the catalyst surfaces in the presence of plasma discharge [35].

2.3. Effect of Catalyst Composition

Figure 5 shows the effect of catalyst composition on plasma–catalytic soot oxidation in terms of the oxidation rate and CO₂ selectivity as a function of reaction time. The soot oxidation increased over time for all plasma reactors with Mn-X/γ-Al₂O₃ catalyst-packing, while the CO₂ selectivity of the process showed only a 0.5% to 1.0% improvement within the 30 min soot oxidation process. The soot oxidation rate of the undoped Mn/γ-Al₂O₃ catalyst was higher than that of the Mn-X/γ-Al₂O₃ catalysts. The relative low soot oxidation rate could be ascribed to the duration of the reaction period, which takes about 60 min to achieve 100% soot oxidation as reported in our previous work [36]. For example, the soot oxidation rate reached 53.9% for the Mn-Co/γ-Al₂O₃-packed plasma reactor, followed by Mn-Ce/γ-Al₂O₃ (51.6%) and Mn-Cu/γ-Al₂O₃ (47.7%) after the 30 min reaction. In contrast with the soot oxidation rate, the CO₂ selectivity of the plasma–catalytic process followed the order of Mn/γ-Al₂O₃ (59.8%) < Mn-Co/γ-Al₂O₃ (64.5%) < Mn-Ce/γ-Al₂O₃ (64.8%) < Mn-Cu/γ-Al₂O₃ (70.1%). It is worthy to note that the soot oxidation rate and CO₂ selectivity of Mn/γ-Al₂O₃ were close to those of the Mn-Co/γ-Al₂O₃ catalyst.

In plasma region, the soot oxidation was initiated by the plasma-generated energetic electrons and reactive radicals, including O·, O₃, etc., while the generation of these species was directly associated with the energy deposited into the DBD plasma reactor [8,10]. In the presence of catalyst pellets and/or soot particles, the discharge mode of DBD plasma would be the combination of filamentary micro-discharge in the gas phase and surface discharge along the surfaces of packed catalysts and soot particles [37]. Moreover, the plasma discharge region would be extended and intensified within the packed-bed layer, which could contribute to the generation of reactive species. Once generated, these species would be transported onto the external surfaces of soot particles and further oxidize the soot particles to CO and CO₂ [38]. It is worth noting that the activation energy of soot oxidation to CO (ΔH = −110 kJ·mol⁻¹) (Reaction (1)) is only 27.9% that of soot oxidation to CO₂ (ΔH = −394 kJ·mol⁻¹) (Reaction (2)) [39]. Consequently, the possibilities of CO formation were much higher than for CO₂ in plasma-induced soot oxidation, which led to the low CO₂ selectivity of the process. In addition, the direct oxidation of CO to CO₂ was also expected both in the gas phase and on the surfaces of the catalyst or soot.

$${\mathrm{C}}_{\mathrm{s}\mathrm{o}\mathrm{o}\mathrm{t}}+\mathrm{O}\to \mathrm{C}\mathrm{O}$$

(1)

$${\mathrm{C}}_{\mathrm{s}\mathrm{o}\mathrm{o}\mathrm{t}}+{\mathrm{O}}_3\to \mathrm{C}\mathrm{O}+{\mathrm{O}}_2$$

(2)

$${\mathrm{C}}_{\mathrm{s}\mathrm{o}\mathrm{o}\mathrm{t}}+\mathrm{O}\to \mathrm{C}{\mathrm{O}}_2$$

(3)

$${\mathrm{C}}_{\mathrm{s}\mathrm{o}\mathrm{o}\mathrm{t}}+{\mathrm{O}}_3\to \mathrm{C}{\mathrm{O}}_2+{\mathrm{O}}_2$$

(4)

$$\mathrm{C}\mathrm{O}+\mathrm{O}\to \mathrm{C}{\mathrm{O}}_2$$

(5)

$$\mathrm{C}\mathrm{O}+{\mathrm{O}}_3\to \mathrm{C}{\mathrm{O}}_2+{\mathrm{O}}_2$$

(6)

In the packed-bed DBD reactor, besides soot oxidation by O· radicals and O₃ from the gas phase, the soot oxidation could be promoted by the active sites on the surfaces of Mn-X/γ-Al₂O₃ via the soot–catalyst contact points. Moreover, the oxidation of CO to CO₂ may also be improved via the soot–catalyst contact points. It is well-known that the properties of catalysts would significantly affect the reaction performance of plasma–catalytic soot oxidation. Previous work also confirmed that the activated oxygen species on the catalyst surfaces can participate in the oxidation of soot and CO to form CO₂ [40]. In addition, the transfer of oxygen species to the surfaces of soot via the physical contact points is also one of the major roles of Mn-X/γ-Al₂O₃ catalysts [41]. The N₂ adsorption–desorption results show that the S_BET, pore volume and pore diameters of Mn-X/γ-Al₂O₃ and Mn/γ-Al₂O₃ catalysts were slightly lower than those of the γ-Al₂O₃ support. In the loose-contact mode, the contact between the catalyst pellets and soot was at the scales of μm-size and mm-size rather than at the nanoscale. It could be deduced that the textural properties of the catalysts may not play a vital role in improving plasma-induced soot oxidation.

For the binary Mn-X/γ-Al₂O₃ catalysts, the catalytic active sites introduced by the dopants would contribute to the soot oxidation as confirmed by the improved redox properties of Mn-X/γ-Al₂O₃ [42,43]. The XPS spectra of O 1s showed improved O_ads/(O_ads+O_latt) values over Mn-X/γ-Al₂O₃ compared to those of Mn/γ-Al₂O₃. The presence of O_ads was correlated with the formation of oxygen vacancies on catalyst surfaces. The oxygen vacancies on Mn-X/γ-Al₂O₃ could act as the reservoir of O_ads species and O· radicals generated by plasma [44]. The O₂-TPD profiles evidenced that the O_ads was the major oxygen species on the surfaces of the Mn-X/γ-Al₂O₃ catalysts, and the order of O_ads amount was Mn-Co/γ-Al₂O₃ > Mn-Cu/γ-Al₂O₃ > Mn-Ce/γ-Al₂O₃ > Mn/γ-Al₂O₃. The complete soot oxidation strongly depends on the oxygen adsorption capacity of the supported metal oxide catalysts and the facilitated oxygen transfer from catalyst surfaces to soot particles [45]. In plasma region, the catalysts with higher O_ads/(O_ads+O_latt) values may be beneficial for soot oxidation since O_ads possessed better mobility compared to O_latt species, and it is much easier to release and transport O_ads species to the surface of soot particles near the interfaces between soot particles and catalyst pellets (i.e., via the contact points) [19]. The H₂-TPR profiles also showed the best reducibility of Mn-Co/γ-Al₂O₃ among the tested Mn-X/γ-Al₂O₃ catalysts, while the oxygen vacancies and O_ads species could be refilled during the plasma process by the generated O· radicals and O₃ species. However, it could be deduced that less active sites would be effectively utilized due to the “loose contact” mode between catalyst pellets and soot since the proximity of the catalyst and soot is of great significance in the “solid-solid reaction” [46]. In the “loose contact” mode, most of the plasma-generated reactive species would be diminished before reaching the interfaces between soot and the gas phase or between soot and catalysts [8,47], especially those decomposed to O_ads species on the surfaces of Mn-X/γ-Al₂O₃ catalysts. The oxidation of CO to CO₂ around the contact points between the catalyst and soot would be consequently improved, leading to the formation of more CO₂ molecules and higher CO₂ selectivity in the presence of Mn-X/γ-Al₂O₃ catalysts. Similarly, Ranji-Burachaloo et al. reported that the loose-contact mode would result in a lower soot oxidation rate and CO₂ selectivity compared with the tight-contact mode [10]. A very good correlation between the CO₂ selectivity and the redox properties of Mn-X/γ-Al₂O₃ catalysts was observed in this work, while the order oxidation rate contrasted with the order of redox properties.

2.4. Effect of Operation Parameters

Figure 6 shows the effect of discharge power on the plasma–catalytic soot oxidation rate and CO₂ selectivity at the 30th min over Mn-Co/γ-Al₂O₃ and Mn/γ-Al₂O₃ catalysts. The soot oxidation rate increased from 33.4% to 71.3% in the presence of the Mn-Co/γ-Al₂O₃ catalyst in the discharge power range of 10 W to 18 W, while the value for the Mn/γ-Al₂O₃ catalyst increased from 35.1% to 78.1% in the same range. It is well-accepted that the discharge power is essential for the initial plasma-induced reactions as various types of energetic electrons and chemically reactive species (e.g., O and O₃) can be generated with the help of input discharge power [48]. The species could be transported onto the surfaces of catalyst and soot particles, resulting in the oxidation of soot and formation of CO and CO₂. In the plasma region, more micro-discharge channels would be generated at a higher discharge power, which could result in the generation of more energetic electrons and chemically reactive species during the discharge process [13]. Consequently, the soot oxidation process could be greatly improved since the packed soot and catalyst particles could directly interact with the micro-discharge and reactive species. As with the CO₂ selectivity, the values for Mn-Co/γ-Al₂O₃ and Mn/γ-Al₂O₃ catalysts both decreased slightly by ~3.3% in the range of 10 W to 18 W. As discussed before, the oxidation of soot to CO is the predominant reaction pathway rather than the oxidation of soot to CO₂ due to the distinct difference in activation energy of these two reactions. The decrease in CO₂ selectivity could be ascribed to the formation of more CO in soot oxidation at a higher discharge power regardless of the catalyst type.

Figure 7 presents the effect of oxygen content on plasma–catalytic soot oxidation and the selectivity of CO₂ over Mn-Co/γ-Al₂O₃ and Mn/γ-Al₂O₃ catalysts. With the increase in oxygen content from 5% to 20%, the soot oxidation rate increased from 41.2% to 60.2% for Mn-Co/γ-Al₂O₃ and 43.9% to 66.5% for the Mn/γ-Al₂O₃-catalyst-packed reactor, respectively. However, the CO₂ selectivity of the process decreased very slightly from 66.5% and 61.7% at 5 vol.% oxygen content by 2.6% and 3.3% at the oxygen content of 20% for these two catalysts. Previous work reported that oxygen radicals, including O radicals and O₃, were the major oxidative species in N₂/O₂ plasma [49]. At a higher oxygen content, more oxidative species would be generated via electronic excitation for soot oxidation. As the oxidation of soot was carried out in a “gas-solid-solid” heterogeneous system, the oxidative species would participate in the oxidation of the external surface of the model soot particles and form the final products of CO and CO₂ [47]. The presence of the Mn-based catalyst would contribute to the decomposition of O₃ to O· radicals and surface-adsorbed oxygen species (O_ads) as reported by Zhang et al. [50]. These species could oxidize the soot particles to CO or convert the CO species in the gas phase to CO₂. Considering the low activation energy required for soot oxidation to CO compared to CO₂, it is reasonable that the CO₂ selectivity decreased with the increase in oxygen content.

The effect of the reaction temperature on plasma–catalytic soot oxidation and CO₂ selectivity at the 30th min over Mn-Co/γ-Al₂O₃ and Mn/γ-Al₂O₃ is presented in Figure 8. The soot oxidation rate decreased with the increase in reaction temperature from 25 °C to 150 °C. The soot oxidation rate decreased from 58.6% at 25 °C to 53.9 at 150 °C, while the value of the Mn/γ-Al₂O₃-catalyst-packed reactor decreased from 59.0% to 57.7% in the same temperature range. On the other hand, the CO₂ selectivity increased from 54.9% to 59.8% for the Mn/γ-Al₂O₃-packed plasma reactor, while the value for Mn-Co/γ-Al₂O₃ improved from 58.4% to 64.8%. At a higher reaction temperature, the diminish in plasma-generated reactive species would be improved due to the enhancement of the Brownian motion of these radicals via the collisions between them and the carrier gas molecules [51]. The decrease in the soot oxidation rate could be ascribed to the diminish of chemically reactive species in the gas phase as most of these species would fail to be transported to the surfaces of soot particles and catalyst pellets at higher temperature [52,53,54]. However, the oxygen species on the catalyst surfaces would be slightly activated at a higher temperature, leading to the further oxidation of -CO groups on soot surfaces via the contact points between catalyst pellets and soot particles. Consequently, the formation of more CO₂ could be expected, and the CO₂ selectivity increased with the increasing reaction temperature.

3. Materials and Methods

3.1. Catalyst Characterizations

N₂ adsorption–desorption experiments were conducted at −196 °C over an Autosorb-iQ instrument from Quantachrome Co. Ltd., Boynton Beach, FL, USA. The Brunauer–Emmet–Teller specific surface area (S_BET) value of the Mn-X/γ-Al₂O₃ catalysts was calculated using the equation, while the pore diameters and pore volumes were calculated using the Barrett–Joyner–Hallender (BJH) method.

X-ray diffraction (XRD) patterns of the catalysts were recorded using a D/max-2000 diffractometer (Rikagu Co. Ltd., Tokyo, Japan) at 100 mA and 40 kV using a Cu-Kα radiation source. The scanning 2θ range was between 10 and 80° with the step size of 0.02°.

X-ray photoelectron spectroscopy (XPS) was performed on an ESCALAB 250Xi spectrometer (ThermoFisher Co. Ltd., Waltham, MA, USA) fitted with Al Kα radiation (hν = 1486.6 eV) X-ray source. All binding energies were calibrated to the C 1s neutral peak at 284.6 eV.

Both hydrogen temperature-programmed reduction (H₂-TPR) and oxygen temperature-programmed desorption (O₂-TPD) were conducted using an AutoChemII 2920 chemical adsorption analyzer (Micromeritics Co. Ltd., Norcross, GA, USA). For H₂-TPR, a 50 mg sample was pretreated at 300 °C for 1 h in 30 mL·min⁻¹ Ar flow and then cooled down to 50 °C. After that, the gas flow was switched to 5% H₂/Ar (30 mL·min⁻¹), and the sample was heated from 50 °C to 850 °C at a heating rate of 10 °C·min⁻¹. As for O₂-TPD, a 50 mg catalyst was purged at 250 °C for 1 h at 5% O₂/He flow rate (30 mL·min⁻¹) and subsequently cooled down to 50 °C. Then, the gas flow was switched to He (30 mL·min⁻¹), and the temperature increased linearly to 850 °C at a heating rate of 10 °C·min⁻¹.

3.2. Catalyst Preparation

In this work, the Mn-X/γ-Al₂O₃ (X = Ce, Co and Cu) catalysts were prepared using the wet impregnation method. All reagents including metal nitrates and γ-Al₂O₃ were purchased from Aladdin Co. Ltd. and of analytic grade. Firstly, weighted metal nitrates were dissolved in deionized water to obtain a solution with the metal ion concentration of 0.1 mol·L⁻¹. Secondly, the desired amount of γ-Al₂O₃ was added to the metal ion solution and magnetically stirred for 4 h at 80 °C. Then, the samples were dried in an oven at 110 °C for 12 h and calcined at 500 °C for 5 h. Finally, the obtained catalysts were crashed and sieved to 40 to 60 meshes prior to use. The loading amount of metal ions was 5% in weight, while the weights of Mn and the dopants were both 2.5 %. In addition, the Mn/γ-Al₂O₃ catalyst with the same loading amount (5% in weight) was prepared with the same procedure for comparison.

3.3. Experimental Setup

The diagram of the experimental setup for plasma–catalytic soot oxidation is shown in Figure 9. In this work, all carrier gases (purity > 99.99%, N₂ and O₂) were supplied from gas cylinders purchased from Fangxin Co. Ltd., China. The gas streams were regulated by mass flow controllers (Sevenstars D07-B, China), while the total gas flow rate was typically 500 mL min⁻¹ with a N₂/O₂ molar ratio of 9:1. Before being introduced into the DBD plasma reactor, the gas streams were premixed.

The DBD plasma reactor consisted of a quartz tube, a stainless-steel rod, a stainless-steel net and two PTFE caps. The inner diameter of the quartz tube was 9 mm, while the wall thickness was 1.5 mm. The stainless-steel rod with a 4 mm diameter was fixed at the axis of the quartz tube by the PTFE seals. The stainless-steel rod was connected to a CTP-2000 K high-voltage power supply (Suman, Nanjing, China) as a high-voltage electrode. A stainless-steel mesh was wrapped outside the quartz tube and acted as the ground electrode. The resulting discharge gap and discharge length were 2.5 mm and 30 mm. Before the experiments, typical Printex-U soot (Degussa Co. Ltd., Frankfurt, Germany) was mixed with the Mn-X/γ-Al₂O₃ catalyst in a mortar for 10 min to obtain a loose contact between soot and catalyst particles. The weight ratio of soot to Mn-X/γ-Al₂O₃ catalyst was 1:9. In each test, 100 mg of soot–catalyst mixture was tightly packed between the discharge gap and held by quartz wool. For all experiments in this work, the DBD reactor was placed in an oven (Lichen Tech., Shanghai, China), and the heating temperature was regulated between 25 °C and 180 °C with an accuracy of ±1 °C. The plasma discharge was initiated after the reaction temperature reached steady state.

All the electrical signals were recorded by a Tektronix TDS2024C oscilloscope, and the discharge power was calculated using the Lissajous method. The whole process of plasma–catalytic soot oxidation was performed for 30 min, while the concentrations of major gaseous products of CO and CO₂ were measured by an on-line infrared gas analyzer (GXH-3010/3011AE, Huayun, China) with an accuracy of ±3%. All experiments were performed three times, and the average values are presented. The reaction performance was evaluated considering the soot oxidation rate and CO₂ selectivity of the process, which were calculated as follows:

$${\mathrm{m}}_{\mathrm{s}\mathrm{o}\mathrm{o}\mathrm{t}}=\frac{\mathrm{Q}\times {\int }_0^{\mathrm{t}}({\mathrm{c}}_{\mathrm{C}\mathrm{O}}+{\mathrm{c}}_{\mathrm{C}{\mathrm{O}}_2})\mathrm{d}\mathrm{t}}{{\mathrm{V}}_{\mathrm{m}}}\times {\mathrm{M}}_{\mathrm{c}}$$

(7)

$${\mathrm{m}}_{\mathrm{s}\mathrm{o}\mathrm{o}\mathrm{t}}=\frac{\mathrm{Q}\times {\int }_0^{\mathrm{t}}({\mathrm{c}}_{\mathrm{C}\mathrm{O}}+{\mathrm{c}}_{\mathrm{C}{\mathrm{O}}_2})\mathrm{d}\mathrm{t}}{{\mathrm{V}}_{\mathrm{m}}}\times {\mathrm{M}}_{\mathrm{c}}$$

(8)

$$\mathrm{C}{\mathrm{O}}_2\mathrm{s}\mathrm{e}\mathrm{l}\mathrm{e}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{i}\mathrm{t}\mathrm{y}(\%)=\frac{{\int }_0^{\mathrm{t}}{\mathrm{c}}_{\mathrm{C}{\mathrm{O}}_2}\mathrm{d}\mathrm{t}}{{\int }_0^{\mathrm{t}}({\mathrm{c}}_{\mathrm{C}\mathrm{O}}+{\mathrm{c}}_{\mathrm{C}{\mathrm{O}}_2})\mathrm{d}\mathrm{t}}\times 100\%$$

(9)

where m_soot is the weight of converted soot during the plasma–catalytic oxidation process, Q is the flow rate, V_m is the gas molar volume under experimental conditions, M_c is the molar mass of carbon, t is the reaction time, M_soot is the weight of initially packed soot, and c_co and c_co2 are CO and CO₂ concentrations in the effluent, respectively.

4. Conclusions

In the presented work, plasma–catalytic soot oxidation was conducted over a series Mn-X/γ-Al₂O₃ (X = Ce, Co and Cu) catalysts. The doping of Ce, Co and Cu species slightly decreased the soot oxidation rate compared with Mn/γ-Al₂O₃, while the order of soot oxidation was in the order of Mn/γ-Al₂O₃ (57.7%) > Mn-Co/γ-Al₂O₃ (53.9%) > Mn-Ce/γ-Al₂O₃ (51.6%) > Mn-Cu/γ-Al₂O₃ (47.7%) during the 30 min soot oxidation process at 150 °C. The CO₂ selectivity of the process was improved in the presence of dopants and followed the opposite order of oxidation rate.

Various catalyst characterizations, including N₂ adsorption–desorption, XRD, XPS, H₂-TPR and O₂-TPD, were performed to illustrate the structure–performance relationships of plasma–catalytic soot oxidation. High O_ads/(O_ads+O_latt) values of Mn-X/γ-Al₂O₃ catalysts indicated abundant oxygen vacancies on their surfaces, which was beneficial for the release of O_ads species during soot oxidation compared with Mn/γ-Al₂O₃. The lower reduction temperature of Mn-X/γ-Al₂O₃ also confirmed the feasibility of O_ads species activation over the catalysts, which could benefit the oxygen transfer from catalyst surfaces to soot particles. O₂-TPD showed that Mn-Co/γ-Al₂O₃ also possessed the most O_ads species on its surface. The order of reducibility of Mn-X/γ-Al₂O₃ catalysts was in line with the order of CO₂ selectivity of the process, and directly opposite that of the soot oxidation rate. The decrease in the soot oxidation rate could be attributed to the consumption of plasma-generated reactive species in the presence of a catalyst, while the improvement in CO₂ selectivity may be ascribed to the utilization of oxidative species, including O_ads, O and O₃, for the oxidation of -CO groups on soot surfaces via the interfaces between soot particles and catalyst pellets.

The effects of discharge power, oxygen content in the carrier gas and reaction temperature on plasma–catalytic soot oxidation were systematically analyzed. The results show that a higher discharge power and lower oxygen content and reaction temperature were beneficial for soot oxidation, while the CO₂ selectivity of the process under these working conditions would be inhibited.

The results obtained in this work could help us to further design and optimize the plasma-catalytic instrument for soot oxidation in terms of catalyst composition, contact type and the screening of proper working conditions. More importantly, it should be noted that the present work mainly focused on the feasibility of using Mn-based catalysts for plasma–catalytic soot oxidation at the lab scale. The future work should be designed to be able to cope with the real conditions of DPF regeneration, i.e., the contact type between soot and catalysts, the design of catalyst composition, the distribution of active metal catalysts on the support, loading method and number of catalysts on the real DPF should be considered. Besides the catalysts, the configuration of plasma reactor structure, types of discharge power and the working conditions are also very important.