Non-Thermal Plasma Incorporated with Cu-Mn/γ-Al₂O₃ for Mixed Benzene Series VOCs’ Degradation

Abstract

In this work, a coaxial dielectric barrier discharge (DBD) reactor was constructed to degrade the mixture of toluene and o-xylene, two typical benzene series. The Cu-MnO₂/γ-Al₂O₃ series catalysts prepared by redox and impregnation methods were filled into the plasma device to degrade VOCs synergistically and explore the degradation effect. The experimental results showed that the introduction of a Cu-doped MnO₂ catalyst significantly improved the pollutants’ removal efficiency and CO₂ selectivity, and greatly inhibited the formation of by-products. Among them, Cu_0.15Mn/γ-Al₂O₃ showed the highest removal efficiency (toluene was 100% and o-xylene was 100%), and the best CO₂ selectivity (92.73%). The XRD, BET, XPS and SEM results confirmed that the synergistic effect between Cu and Mn in the Cu-Mn solid solution could promote the amount and reducibility of the surface active oxygen species, which improved the catalytic performance. Finally, the toluene and o-xylene decomposition pathways in the NTP catalytic system were speculated according to the detected organic matter. This work provides a theoretical and experimental basis for the application of DBD-catalyzed hybrid benzene series VOCs.

1. Introduction

The emission of volatile organic compounds (VOCs) not only harms human health, but also causes environmental problems such as haze, photochemical smog, and ozone depletion [1]. Among VOCs, the benzene series are a typical kind of VOC, which are usually used as organic solvents in industrial production due to their good stability and indissolubility with most organics. However, the benzene series can cause air pollution and be toxic to living organisms [2]. Many benzene series have pungent odors, causing discomfort to the human body. They also cause photochemical smog in cities, producing secondary pollution. Technologies such as adsorption, absorption, condensation, membrane separation, catalytic oxidation/combustion, biological treatment, photocatalytic oxidation, and non-thermal plasma have been applied to the degradation of VOCs [3]. Photocatalytic oxidation technology and NTP technology are two promising technologies. Among them, non-thermal plasma (NTP) technology has attracted considerable attention due to its advantages of simple operation, high oxidation degradation activity, and low cost [4,5]. However, the formation of by-products and low energy efficiency relatively limit the application of NTP [6]. Jiang et al. [7] detected intermediate products, including HCOOH, CO, CO₂, H₂O, N₂O, and O₃, in the experimental process of studying the pulsed-modulated dielectric barrier discharge (DBD) degradation of toluene. Chang et al. proved that O₃ was an inevitable by-product when gas containing O₂ was used as the carrier gas, and that the O₃ concentration increased with the increase in the specific input energy (SIE) [8]. Ye et al. achieved a removal rate of 100% when degrading benzene containing gas, but the mineralization rate was not 100% [9]. It was found that aerosols were generated during the discharge process, and phenolic substances were deposited on the inner wall of the DBD tubes to form brown solids. Fan et al. [10] found that large amounts of O₃ and trace amounts of NO₂ were produced when dealing with low concentrations of p-xylene. Compared with other traditional treatment technologies, the non-thermal plasma degradation of VOCs has certain advantages, but it has disadvantages such as a low energy efficiency and undesirable by-products, such as NO_X and O₃, which inevitably remain in the exhaust gas. To overcome these shortcomings, researchers have found that combining plasma with catalytic technology can not only significantly improve energy efficiency and selectivity, but also greatly reduce the production of by-products. Plasma co-catalytic technology combines the advantages of plasma and heterogeneous catalytic technology, and has the advantages of a high VOC emission reduction efficiency, good product selectivity, less by-product generation, and high energy efficiency. It is a low-cost and energy-saving method for VOC removal.

The addition of the catalyst can effectively improve the efficiency of the system, especially for the degradation of lower concentrations of VOCs [11,12,13]. On the one hand, compared with the plasma alone, the catalyst can help the system produce active species (free radicals and various ions) with a lower input power. In turn, plasma can provide the thermal energy required for catalysis. On the other hand, catalysts can provide abundant oxygen vacancies and further generate oxygen-active species by utilizing O₃ and gaseous O₂ in the system. In addition, it also improves the selectivity of CO₂ and inhibits the generation of O₃ [14]. The catalyst is generally composed of support and active components. Common supports include Al₂O₃, SiO₂, molecular sieve, diatomite, activated carbon, etc. These catalysts are usually characterized by a large specific surface area, high mechanical strength, high temperature resistance, and low fluid resistance, and their shapes are generally honeycomb, spherical, lumpy, and other shapes. The support can increase the interaction time between VOCs and catalysts, thus facilitating the degradation of VOCs [15]. γ-Al₂O₃ is widely used as a support due to its low price, convenient use, high reactivity, and chemical stability in VOCs’ degradation [15]. Wu et al. [16] load NiO on Al₂O₃ support to remove toluene, which has a higher dispersion than SBA-15, TiO₂ support. Therefore, the catalytic effect was better. Dang et al. [17] used a DBD reactor filled with catalysts M/13X-γ-Al₂O₃ (M: Ag, Ce, Mn, and Co) to evaluate the removal effect of toluene, and the results showed that Ag/13X-γ-Al₂O₃ had the highest penetration and the highest mineralization rate (about 68%) at 20 kV. Jiang et al. [18] loaded the transition metals of Mn, Co, Fe, and Cu onto SiO₂, ZSM-5-300, ZSM-5-38, and γ-Al₂O₃ and filled them into the DBD reactor to degrade toluene. The catalyst supported by γ-Al₂O₃ showed the best performance in toluene decomposition and the highest toluene decomposition was 97% when the MnO₂ loaded on γ-Al₂O₃ was 1 wt%.

Transition metal oxides have been widely used as catalysts for VOCs’ degradation in plasma catalytic systems because of their low price and effective catalytic decomposition of gaseous pollutants [19,20]. MnO₂ has oxygen vacancies and a valence state, which can promote the degradation of VOCs and improve CO_X selectivity. Meanwhile, MnO₂ was effective for ozone removal [21,22]. Ge et al. [23] used a combination of DBD and MnO₂ catalyst to degrade benzene and were able to significantly increase the CO_X selectivity from 38% to 80% with almost no O₃ generated. Wang et al. [24] found that the MnO_X catalysts prepared with different precursors had different o-xylene removal performances in the plasma discharge region, and this difference was mainly caused by the presence of Mn⁴⁺ and lattice oxygen on the catalyst. More Mn⁴⁺ and lattice oxygen were conducive to the degradation of o-xylene. Yang [25] et al. prepared catalyst three-dimensional hollow urchin α-MnO₂ and combined it with the PPC (post-plasma catalysis) system to degrade toluene. The maximum toluene degradation rate can reach 100%, and the energy efficiency was improved by 64% compared with NTP alone. Zheng [26] et al. combined solar irradiation and PPC to degrade toluene over a MnO₂/GFF (bifunctional graphene fin foam) catalyst. When the SIE was 350 J/L, the degradation rate of toluene reached 93% and the selectivity of CO₂ reached 83%, which significantly reduced the energy consumption of the plasma catalytic gas purification process. Trinh et al. [27] reported that ceramics loaded with MnO₂ exhibited a better catalytic performance for acetone decomposition than ZnO no matter if in the plasma discharge area or after the discharge. However, pure MnO₂ catalysts usually require more energy in the degradation process of VOCs and may lead to the formation of other by-products, which makes it difficult to achieve the ideal catalytic degradation effect. Therefore, it is necessary to enhance the interaction of supports [28]. Roberto [29] reviewed the application of bimetallic catalysts in VOC oxidation in terms of the catalytic activity and physical and chemical properties. It was found that the bimetallic catalyst had a great improvement in properties compared with a nonmetallic catalyst due to the synergistic effect. In recent years, Cu-doped MnO₂ catalysts have shown broad application prospects with high surface areas and redox performances [21,30]. A study proved that a Cu-O-Mn interface could be formed by introducing copper into manganese oxide, which improves the reducibility of Cu-OMS [31]. Therefore, a Cu-doped MnO₂ catalyst can be considered to deal with benzene series VOCs.

Most of the current studies focus on a single target and ignore the interaction between the different components in the degradation process. In addition, toluene and o-xylene are two typical benzene VOCs, usually from the petrochemical, coating, automobile manufacturing, packaging, and printing industries. In this work, a coaxial DBD reactor was built to degrade mixed gases of toluene and o-xylene. A series of Cu-Mn/γ-Al₂O₃ catalysts with various Cu/Mn molar ratios were prepared and filled in the NTP reactor to co-degrade benzene series. The removal efficiency, CO_X selectivity, catalyst characterization, and energy efficiency were studied and discussed deeply. Organic matter during the degradation of toluene and o-xylene in the NTP catalytic system was detected by GC-MS and the possible decomposition pathways were speculated.

2. Results and Discussion

2.1. Catalysts Characterization

The XRD patterns of MnO₂, Cu_0.05Mn, Cu_0.15Mn, Cu_0.25Mn, and Cu_0.5Mn are shown in Figure 1. In total, 5 main diffraction peaks can be seen in the XRD diffraction patterns of the pure MnO₂ in the figure, centered at 2θ = 28.7°, 37.5°, 42.9°, 46.2°, and 67.4°, respectively, corresponding to standard MnO₂ (PDF-72-1984). According to the standard card, they correspond to the (110), (101), (111), (210), and (310) crystal planes, respectively. After Cu was gradually doped, the primary crystal structure of MnO₂ was unchanged and no visible diffraction peaks of CuO appeared, indicating the high dispersion of Cu species in MnO₂ [14]. As the amount of Cu increased, a slight peak shift (shaded parts) can be seen until the molar ratio of Cu-to-Mn reached 0.15. Since the ionic radius of octahedral manganese ion in MnO₂ was much smaller than that of Cu²⁺,the latter can easily replace Mn⁴⁺ in MnO₂ [32]. However, only a small amount of Cu²⁺ can enter the MnO₂ lattice to form a solid solution and lead to lattice expansion.

The SEM photos of Cu-doped MnO₂ catalysts can be seen in Figure 2. As can be seen from the scan image magnified 5000 times, the pure MnO₂ and Cu-doped MnO₂ both showed similar morphologies, with rough and uneven surfaces, and a part of the larger particle loaded, which could easily fall off due to collision and friction in the later stage [33]. After comparative analysis, small particulate matter appeared on the original MnO₂/γ-Al₂O₃ pattern after copper doping. As can be seen from the scanning figure magnified 20,000 times, in the pore structure of the γ-Al₂O₃ particles, Cu entered the pore in the form of small pellets and honeycomb and was loaded well inside. Because MnO₂ is doped into the γ-Al₂O₃ support, it is judged that the gray area is alumina and the white area is MnO₂. Most of the MnO₂ (white) was embedded in γ-Al₂O₃ (gray) as particles. The surface structure of the support did not change greatly after loading.

Figure 3 shows the N₂ adsorption/desorption isotherms of the catalysts. It can be seen that each catalyst had H₁ hysteresis loops [34]. According to the results listed in

Table 1. Comparison of specific surface area, average pore diameter, and pore volume of different catalysts.

Catalyst Type	SBET (m2·g−1)	Average Pore Size (nm)	Pore Volume (cm3·g−1)
γ-Al2O3	322.560	4.777	0.314
MnO2/γ-Al2O3	295.580	5.093	0.297
Cu0.05Mn/γ-Al2O3	304.894	4.760	0.289
Cu0.15Mn/γ-Al2O3	303.457	4.614	0.316
Cu0.25Mn/γ-Al2O3	294.615	4.625	0.277
Cu0.5Mn/γ-Al2O3	234.726	4.955	0.233

, the catalysts prepared by a redox precipitation method were typical mesoporous material in the range of 2–50 nm. The N₂ physisorption curve belongs to the type IV isotherm. The γ-Al₂O₃ support pellets selected in this study had the largest specific surface area of 322.560 m²·g⁻¹. When MnO₂ and Cu were loaded onto the catalyst, the specific surface area decreased, indicating that the catalyst was better coated on the surface of the support. Compared with other catalysts, Cu_0.15Mn had the largest pore volume of 0.316 cm³·g⁻¹ and the smallest pore size of 4.614 nm. The small pore size and large pore volume meant that there was more space for micro reactions, which was more conducive to the production of reactive species by micro discharge. Although the main mesoporous structure of MnO₂ would not be affected by a small amount of Cu introduction, the pore volume can be changed effectively.

The XPS scanning spectra of O1s, Mn2p, and Cu2p are shown in Figure 4a–c, respectively. According to the deconvoluted O1s XPS spectra in Figure 4a, all catalysts showed 2 distinct sub-bands at around 532.18 eV and 529.88 eV, belonging to the surface labile oxygen (O_ads) and lattice oxygen (O_lat), respectively [35,36]. The redox cycle between Mn⁴⁺/Mn³⁺ and Cu²⁺/Cu⁺ consumes lattice oxygen and generates oxygen vacancies. The existence of oxygen vacancies is conducive to the formation of O_ads on the catalyst surface, which has a higher mobility than O_lat and is conducive to the catalytic oxidation reaction. The depleted oxygen vacancy can be filled by the decomposition of oxygen and O₃ in the plasma gas phase [37].

Figure 4b shows the deconvoluted Mn2p XPS results of the Cu-doped MnO₂ catalysts with a range of 635~660 eV. There were two peaks at 643.38 eV and 642.38 eV, which belong to the Mn⁴⁺ and Mn³⁺ species, respectively [38]. Similarly, there were also two sub-peaks of the Cu2p spectrum, as shown in Figure 4c. The region at 933.58 eV can correspond to the surface Cu²⁺, and the region near 931.78 eV was assigned to the Cu⁺ species, respectively [39].

2.2. The Degradation Efficiency of Toluene and o-Xylene

The removal efficiency of toluene and o-xylene in NTP alone and the NTP-catalysis system have been studied as a function of SIE under the same conditions. It can be seen from Figure 5 that the degradation efficiency of toluene and o-xylene increased with the increase in SIE, regardless of whether the catalyst was used. This was because, with the increase in SIE, the number of high-energy electrons and active free radicals generated in the reactor increased, and the probability of each VOC molecule sharing high-energy electrons and active free radicals increased, thus leading to the improvement of VOCs’ degradation efficiency. In addition, the degradation efficiency of o-xylene was higher than that of toluene at the same energy density. For example, when SIE was 5.72 kJ·L⁻¹, the degradation efficiency of 700 mg·m⁻³ o-xylene was as high as 82.05%, while for toluene with the same concentration, the degradation efficiency was only 53.17%. This meant that the degradation efficiency of VOCs was related to VOC types and their chemical structure. On the one hand, the reason for the degradation was that the chemical bonds within VOCs were broken by the energetic electrons and active particles. In the case of toluene and o-xylene, the main fractures were C-H (methyl), C-H (benzene ring), C-C (methyl with benzene ring), and C-C (benzene ring). The chemical bond energy and molecular structure of the VOC molecules determined their degradation efficiency during NTP degradation [40,41]. The C-H bond energy on methyl (3.7 eV) is smaller than that on the benzene ring (4.3 eV), and the C-C bond energy between the methyl and benzene ring (4.4 eV) is also smaller than that on the benzene ring (5.0~5.3 eV). There are two methyl groups in o-xylene; therefore, when the number of energetic electrons and active radicals in the discharge area was constant, C-H (methyl) and C-C (methyl and benzene ring) on o-xylene molecules were more likely to collide with these active particles. Therefore, the o-xylene was more easily degraded compared with toluene. In toluene, on the other hand, the superchoke effect between the methyl and benzene rings [42] brought the electrons of the C-H bond on methyl closer to H, so that methyl was easily oxidized. In addition, the electron cloud density of benzene ring increased under the influence of the methyl group, and it was also prone to substitution reaction [43]. O-xylene contains two methyl groups, and the superconjugate effect between the methyl group and the benzene ring was more obvious. The electron cloud density of the benzene ring increased further, and the substitution reaction was more likely to occur, which caused the o-xylene to be degraded more easily.

In order to investigate the interaction between toluene and o-xylene in the degradation process, a mixture of 900 mg·m⁻³ toluene and 900 mg·m⁻³ o-xylene was prepared to study the degradation efficiency. It can be seen from the Figure 5c, in the whole range of energy density, that the degradation efficiency of toluene and o-xylene after mixing is higher than that of a single component with the same concentration. The degradation of single component VOCs can produce reactive intermediates, such as aldehydes and peroxides, which can improve the degradation efficiency of VOCs to a certain extent. More of this reactive substance is produced in the mixture degradation than in the degradation of a single component, and it is easier to form reaction intermediates/free radicals. Therefore, the degradation of the toluene and o-xylene mixture can improve the utilization of the active substance produced in the DBD reactor. Some scholars have also observed that VOCs are easily activated in mixtures [44,45]. Therefore, the degradation efficiency of toluene and o-xylene in the mixture is higher.

As shown in Figure 5a, when toluene was degraded under plasma alone, the conversion rate was 53.17% at the SIE of 5.72 kJ·L⁻¹. After adding Cu-doped MnO₂ catalysts to the plasma system, the removal efficiency of toluene was obviously improved, following the order of Cu_0.15Mn > Cu_0.05Mn > Cu_0.25Mn > Cu_0.5Mn > MnO₂. The results indicated that the catalysts could improve the removal rate of toluene, and the catalytic effect of catalyst was related to the amount of the Cu load in the catalysts. Among these catalysts, Cu_0.15Mn had the best performance, achieving 100% conversion at 5.72 kJ·L⁻¹ SIE, which was about 2 times higher than that of separated plasma. As shown in Figure 5b, the Cu-doped MnO₂ catalysts in the plasma reactor also greatly promoted the degradation of o-xylene. Similar to toluene, Cu_0.15Mn still showed the highest removal efficiency (~100%). MnO₂ had a special pore structure, high thermal stability, and various in valence states of Mn^x+ (x = 2, 3 and 4). These unique properties gave MnO₂ a good catalytic activity [46]. The solid solution effect between Cu and Mn in the catalyst may be the reason why the degradation efficiency was obviously improved with different amounts of Cu doping. According to the magnitude of the peak shift in the XRD pattern, the MnO₂ structure may be affected by Cu, which may cause a charge imbalance and unsaturated chemical bonds on the catalyst, leading to the formation of more oxygen vacancies. Gaseous oxygen could fill the oxygen vacancy on the catalyst surface and generate reactive oxygen species [39,47]. Surface reactive oxygen species could be regarded as the main active substances in the plasma catalytic oxidation of toluene and o-xylene. A comparison of the degradation efficiency for this and other studies is shown in Table S2.

It also can be seen from the Figure that the degradation efficiency does not improve with the increase in Cu doping, which may be due to the following two reasons. First, the performance of catalyst was related to the dispersion state of the active component Cu on the support. As the load increased, the number of active centers of catalytic reaction exposed in the plasma space increased, and the oxidation rate of toluene gradually accelerated. When the loading exceeded a certain level, the Cu aggregated into larger particles, and the active components could be distributed on the support in the form of multilayer accumulation or clusters. The number of catalytic reaction centers did not increase with the excessive increase in the loading, but the number of reaction centers exposed to the plasma space decreased, which caused a decrease in the catalytic activity. Therefore, the catalytic effect of Cu_0.05Mn and Cu_0.5Mn was slightly worse than that of the Cu_0.15Mn catalyst. Second, the pore size affects the residence time of the reaction; a suitable pore size was conducive to reactants into the active site of the catalyst for a reaction. A small pore size and large pore volume cause more space for small reactions, which was more conducive for micro discharges to produce reactive species and participate in the oxidation reaction of VOCs [48]. According to the BET characterization results, compared with other catalysts, Cu_0.15Mn had the smallest pore size of 4.614 nm and the largest pore volume of 0.316 cm³·g⁻¹. In conclusion, the catalyst Cu_0.15Mn had the best degradation performance.

2.3. The CO_X Selectivity

As can be seen from the Figure 6, with the increase in SIE, the selectivity of CO₂ increased and CO decreased in both the NTP alone system and the NTP-catalytic system. High specific input energy may lead to more electrons with high energy. Furthermore, high-energy electrons are also conducive to the production of active substances that completely oxidize toluene and o-xylene into CO₂ [47], and part of CO would be oxidized into CO₂, so the selectivity of CO is reduced while the CO₂ selectivity is increased. Among the projects studied, Cu_0.15Mn produced the highest CO₂ selectivity (~92.73%) and the lowest CO selectivity (~3.57%) at 5.72 kJ·L⁻¹. In contrast, a single plasma produced the lowest CO₂ selectivity (~52.08%) and the highest CO (~12.20%). This is because the addition of a catalyst in the discharge gap increased the average field strength and the corresponding average electron energy, resulting in a surface discharge on the catalyst particles, which induced the effective interaction between plasma and catalyst to activate CO₂, leading to the enhancement of the CO₂ conversion. The Cu_0.15Mn catalyst had more space for micro reactions, which was conducive to micro discharge, generated more active species, and made more intermediate products converted to CO₂ and H₂O.

2.4. Inhibition Effect of the NO₂ and O₃ Generation

NO₂ and O₃ are the main by-products in the plasma degradation process. The concentrations of NO₂ and O₃ produced by the degradation of VOCs in NTP alone and NTP-catalyst systems were determined by gas chromatography and indigo sodium disulfonate spectrophotometry from 0.51 to 5.72 kJ·L⁻¹. Since NO was easily oxidized to NO₂ by a large amount of ozone after discharge, the main NO_X species was NO₂ [15]. It can be seen from Figure 7a that the NO₂ concentration increased with the increase in SIE. The main reason was that with the increase in SIE, the number of electrons, ions, excited molecules, and free radicals formed increased, and electrons with energy exceeding 5.1 eV or 9.18 eV increased, which led to the ionization and excitation of the N₂, O₂, and H₂O molecules in the carrier gas, resulting in the generation of some primary active substances, such as O·, OH·, H·, N₂*, and O₃. As shown in Equations (1)–(4) [49,50]:

$$\mathrm{e}+{\mathrm{O}}_2\to \mathrm{e}+\mathrm{O}\left({}_{ }{}^1\mathrm{D}\right)+\mathrm{O}\left({}_{ }{}^3\mathrm{P}\right)$$

(1)

$$\mathrm{e}+{\mathrm{N}}_2\to \mathrm{N}·+\mathrm{N}·+\mathrm{e}$$

(2)

$$\mathrm{e}+{\mathrm{H}}_2\mathrm{O}\to \mathrm{H}·+\mathrm{OH}·+\mathrm{e}$$

(3)

$$\mathrm{O}·+{\mathrm{O}}_2+\mathrm{M}\to {\mathrm{O}}_3+\mathrm{M}$$

(4)

Among them, the N· and O· generated NO₂ reacted with O₂ and H₂O [51,52], as shown in Equations (5)–(9):

$$\mathrm{N}·+{\mathrm{O}}_2\to \mathrm{NO}+\mathrm{O}·$$

(5)

$$\mathrm{N}·+{\mathrm{H}}_2\mathrm{O}\to \mathrm{NO}+{\mathrm{H}}_2$$

(6)

$$\mathrm{NO}+\mathrm{O}·\to {\mathrm{NO}}_2$$

(7)

$$\mathrm{O}·+{\mathrm{NO}}_2\to {\mathrm{NO}}_3$$

(8)

$$\mathrm{O}·+{\mathrm{NO}}_3\to {\mathrm{O}}_2+{\mathrm{NO}}_2$$

(9)

OH· reacted with NO and NO₂ to form HNO₂ and HNO₃, which in turn reacted with OH· to form NO₂:

$$\mathrm{OH}·+\mathrm{NO}\to {\mathrm{HNO}}_2$$

(10)

$$\mathrm{OH}·+{\mathrm{NO}}_2\to {\mathrm{HNO}}_3$$

(11)

$$\mathrm{OH}·+{\mathrm{HNO}}_2\to {\mathrm{H}}_2\mathrm{O}+{\mathrm{NO}}_2$$

(12)

$$\mathrm{OH}·+{\mathrm{HNO}}_3\to {\mathrm{H}}_2\mathrm{O}+{\mathrm{NO}}_3$$

(13)

O₃ was formed by the reaction of O₂ and O·, and participated in the formation reaction of NO₂, as shown in Equation (14):

$${\mathrm{O}}_3+\mathrm{NO}\to {\mathrm{NO}}_2+{\mathrm{O}}_2$$

(14)

Therefore, the formation rate of NO₂ increased with the increase in SIE. The combination of NTP and the catalyst could obviously inhibit the formation of NO₂, and Cu_0.5Mn had the best effect on the inhibition of NO₂ emission. When SIE increased from 0.51 kJ·L⁻¹ to 5.72 kJ·L⁻¹, the NO₂ concentration increased from 211 ppm to 394 ppm mg·m⁻³, compared with NTP alone, which reduced the concentration of NO₂ by 36%, indicating that the Cu-doped MnO₂ was also beneficial to inhibit NO_X formation, and the inhibition effect was related to the amount of Cu doping. It was preliminarily speculated that the reducibility of Cu⁺ can directly reduce the original NO₂ to N₂. Compared with other catalysts, the effect of Cu_0.5Mn was the best.

As can be seen from Figure 7b, in the whole energy density region, with the increase of SIE, the O₃ concentration first increased and then decreased. The reason for this was that when the SIE was within 4.42 kJ·L⁻¹, the number of free radicals and high-energy electrons increased with the electric field strength, and these substances collided or reacted with the air molecules to produce more atomic oxygen, leading to a greater chance of colliding with oxygen, which caused the concentration of O₃ to increase [53]. Furthermore, when the SIE increased to more than 4.42 kJ·L⁻¹, the temperature increased in the discharge area and the ultraviolet light emitted by a large number of excited particles when they transitioned to ground state particles, which caused O₃ to break bonds or degrade.

Compared with NTP alone, the Cu-doped MnO₂ catalyst could effectively reduce O₃ emission. When SIE was 5.72 kJ·L⁻¹, the O₃ concentration decreased from 125.6 mg·m⁻³ to 55.8 mg·m⁻³. Different catalysts inhibited O₃ generation in the order of Cu_0.15Mn > Cu_0.25Mn > Cu_0.5Mn > Cu_0.05Mn > MnO₂. Cu_0.15Mn had the highest O₃ inhibition, and the concentration of O₃ was about 56.6%, lower than that with NTP alone. This suggests that the catalyst Cu_0.15Mn had more space for micro reactions. The active species generated by micro discharge decomposed O₃ on the surface of Cu_0.15Mn and generated reactive oxygen atoms, which participated in the oxidation process of VOCs.

2.5. Intermediate Product and Mechanism

2.5.1. Intermediate Product Analysis

Some by-products are produced in the non-thermal plasma catalysis system. The experimental results are shown in Figure S1, and the by-products are listed in Table S1. In this study, the plasma device input SIE was 5.72 kJ·L⁻¹, the initial concentration of toluene and o-xylene was 700 mg·m⁻³, and the flow rate was 1 L·min⁻¹. The exhaust gas produced after degradation was studied and analyzed by gas chromatography mass spectrometry (GC-MS). In the plasma reactor filled with five different Cu-doped MnO₂ catalysts, the by-products were different. The detected 8 open benzene ring by-products and 15 benzene ring containing by-products are shown in Table 2.

2.5.2. Mechanism Analysis

The process of VOC treatment by non-thermal plasma co-catalysis was mainly composed of two parts, one was the reaction in the gas phase and the other was the reaction on the surface of the catalyst.

The reaction in the gas phase mainly had three processes: high-energy electrons collided with gas molecules to produce active substances, energetic electrons and active particles reacted with VOCs molecules, and unstable active particles such as free radicals further reacted with the small molecules of organic matter after decomposition.

(1) High-energy electrons collided with gas molecules to produce active substances: High-energy electrons collided with N₂, O₂, CO₂, and H₂O in the carrier gas at room temperature and a large number of O·, OH·, H·, N·, ·HO₂, O₃ and other active particles were generated. These active particles and free radicals promoted the oxidative decomposition reaction of VOC molecules, as shown in Equations (1)–(3).

(2) Energetic electrons and active particles reacted with VOCs molecules: In the process of non-thermal plasma discharge, a large number of high-energy electrons would be generated [53]. The high-energy electrons generated would react with the VOC molecules. The excitation and dissociation of VOC molecules depended on the energy of the high-energy electrons and the bond energy of chemical bonds in the molecule. When the energy of high-energy electrons was greater than the bond energy of the chemical bonds in VOC molecules, the chemical bonds broke, accompanied by unstable intermediates. The unstable intermediates either collided with the energetic electrons again or combined with other free radicals to form other stable substances [54].

The bond energies of different chemical bonds in VOCs and the reactions that would occur when they collided directly with high-energy electrons are shown in Table 2. According to Table 2, the lower the bond energy, the more likely the bond was to break in direct collision with high-energy electrons. The C-H bond energy on methyl was relatively small, and high-energy electrons could easily destroy this bond to form H· and phenylmethyl [55,56]. Secondly, the C-H bond on the benzene ring was broken [55]. Next, the C-C bond between methyl and benzene ring was broken to form phenyl and methyl [56]. Finally, the C-C bond on the benzene ring was broken and two small molecule hydrocarbons were generated [54]. Jiang [18] found that the degradation process of toluene in the plasma system loaded with 1 wt%Mn/γ-Al₂O₃ was as follows: toluene → benzyl → benzyl alcohol → benzaldehyde → benzyl acid → carbonates and bicarbonates → CO₂ and H₂O. However, because the structure of benzene ring was relatively stable, carbon atoms were bonded with the π bonds, which made it difficult to break or the energy required to break the C-C (benzene ring) bonds was high. Therefore, it was difficult to make a ring-opening reaction happen by pure electron collision as the reaction depended more on the oxidation decomposition of O₃ and various active particle free radicals in the reaction system.

The reaction rate constants of electrons with toluene are greater than those of electrons with OH·, O·, and O₃. The contribution of electrons to toluene degradation is very low. Electrons react with toluene by ionization, dissociation, and excitation of O₂ and N₂ to produce reactive oxygen species and reactive nitrogen species [37]. The excited nitrogen plays a very important role. Robby [57]’s results suggested that the N₂(A³${{\displaystyle \sum }}_U^{+}$) species play the most important role in the destruction of ethylene. The direct destruction of ethylene by electron impact was negligible. In a previous study, Zhang [58] simulated the removal of formaldehyde under the N₂ atmosphere and air atmosphere, respectively. It was found that N₂(A³${{\displaystyle \sum }}_U^{+}$) can obviously remove formaldehyde in the N₂ atmosphere, and it was removed by the collision of N₂(A³${{\displaystyle \sum }}_U^{+}$) with formaldehyde. However, the removal of HCHO in the air atmosphere was determined by O· and OH· radicals, and N₂(A³${{\displaystyle \sum }}_U^{+}$) mainly acted to increase the concentration of O· and OH· radicals in the system.

The unstable intermediates (such as C₆H₄·CH₃ and C₆H₃·CH₃CH₃) generated by the high-energy electrons reacted with electrons again are shown in Equations (15)–(18):

$${\mathrm{C}}_6{\mathrm{H}}_4·{\mathrm{CH}}_3+\mathrm{e}\to {\mathrm{C}}_3{\mathrm{H}}_2·{\mathrm{C}}_3{\mathrm{H}}_2·+{\mathrm{CH}}_3·/{\mathrm{C}}_6{\mathrm{H}}_4·{\mathrm{CH}}_2·+\mathrm{H}·$$

(15)

$${\mathrm{C}}_6{\mathrm{H}}_3·{({\mathrm{CH}}_3)}_2+\mathrm{e}\to {\mathrm{C}}_3{\mathrm{H}}_2·{\mathrm{C}}_3\mathrm{H}·{\mathrm{C}}_3\mathrm{H}+{\mathrm{CH}}_3·$$

(16)

$${\mathrm{C}}_6{\mathrm{H}}_3·{({\mathrm{CH}}_3)}_2+\mathrm{e}\to {\mathrm{C}}_6{\mathrm{H}}_3·{\mathrm{CH}}_2·{\mathrm{CH}}_3+\mathrm{H}·$$

(17)

$${\mathrm{C}}_6{\mathrm{H}}_3·{({\mathrm{CH}}_3)}_2+\mathrm{e}\to {\mathrm{C}}_6{\mathrm{H}}_3·{\mathrm{CH}}_2·{\mathrm{CH}}_2+2\mathrm{H}·$$

(18)

(3) The unstable active particles continued to react with the decomposed intermediates C₆H₅CH₂·, C₆H₄CH₂·CH₂·, C₆H₄·CH₃, C₆H₃·CH₃CH₃, C₆H₅·, and CH₃·, as shown in Figure 8. In other words, VOCs were ultimately decomposed into CO₂, H₂O, and other environmentally pollution-free substances through various pathways under the action of free radicals O·, OH·, H·, N·, and ·HO₂.

Table 2. The reaction of breaking different chemical bonds in toluene and o-xylene [56].

Chemical Bonds	Energy/eV	Bond Breaking Reactions That Occurred after Collision
C-H (Methyl)	3.7	C7H8 + e → C6H5CH2· + H·
		C8H10 + e → C7H7CH2· + H·/C8H10 + e → C6H4CH2·CH2· + 2H·
C-H (Benzene ring)	4.3	C7H8 + e → C6H4·CH3 + H·
		C8H10 + e → C6H3·CH3CH3 + H·
C-C (Between methyl and benzene)	4.4	C7H8 + e → C6H5· + CH3·
		C8H10 + e → C6H4·CH3 + CH3·/C8H10 + e → C3H2·C3H2· + 2CH3·
C-C (Benzene ring)	5.0–5.3	C7H8 + e → C3H4· + C4H4·
		C8H10 + e → C4H6· + C4H4·

Table 2. The reaction of breaking different chemical bonds in toluene and o-xylene [56].

Chemical Bonds	Energy/eV	Bond Breaking Reactions That Occurred after Collision
C-H (Methyl)	3.7	C7H8 + e → C6H5CH2· + H·
		C8H10 + e → C7H7CH2· + H·/C8H10 + e → C6H4CH2·CH2· + 2H·
C-H (Benzene ring)	4.3	C7H8 + e → C6H4·CH3 + H·
		C8H10 + e → C6H3·CH3CH3 + H·
C-C (Between methyl and benzene)	4.4	C7H8 + e → C6H5· + CH3·
		C8H10 + e → C6H4·CH3 + CH3·/C8H10 + e → C3H2·C3H2· + 2CH3·
C-C (Benzene ring)	5.0–5.3	C7H8 + e → C3H4· + C4H4·
		C8H10 + e → C4H6· + C4H4·

O·, OH·, and other free radicals determined the degree of oxidation in the formation of VOCs and small molecular organics. However, considering the limited number of active free radicals, some small molecules of organic matter would not be completely oxidized into CO₂ and H₂O, thus inhibiting the degradation of VOCs. Therefore, the removal efficiency of VOCs could be improved by introducing a Cu-doped MnO₂ catalyst.

After the catalyst was introduced, toluene, o-xylene, and organic intermediates in the gas phase could be adsorbed on the surface of the catalyst. The gas phase products after the degradation of toluene and o-xylene were analyzed. Figure S1 shows that the types of by-products were reduced compared with the types of intermediate by-products generated by plasma degradation alone. This showed that the introduction of a Cu-doped MnO₂ catalyst could transform more intermediate by-products into small molecule inorganic substances and improve the removal efficiency of VOCs. The synergistic effect between non-thermal plasma and a Cu-doped MnO₂ catalyst was mainly reflected in the following aspects:

(4) Under the action of the applied electric field, a large number of active particles produced by the discharge, such as high-energy electrons, ions, excited atoms, and free radicals, impacted the surface of the catalyst, polarized the catalyst particles, and caused the secondary emission of electrons. The enhanced field region was formed on the surface of the catalyst, and the unreacted molecules in the discharge region were further excited into reactive free radicals to participate in the reaction [59].

(5) The oxidation of toluene and o-xylene molecules on mangan-based catalysts was carried out by the Mars-van Krevelen mechanism [60]. In the catalyst filling zone, toluene and o-xylene were adsorbed on the catalyst surface. There are two types of sources of surface reactive oxygen species. One is the decomposition of O₃ and the collision between the plasma and the oxygen adsorbed by the catalyst [61]. The other is supplemented by the REDOX cycles of the Cu and Mn cations (Mn⁴⁺/Mn³⁺ and Cu²⁺/Cu⁺) in the Cu-Mn/γ-Al₂O₃ catalysts. The REDOX cycle between Mn⁴⁺/Mn³⁺ and Cu²⁺/Cu⁺ consumes lattice oxygen to create oxygen vacancies that can be filled by the decomposition of oxygen and O₃ in the plasma gas phase [62].

(6) A large number of active particles generated by the discharge had abundant energies, which could reduce the activation energy required by part of the catalyst reaction and speed up the oxidation reaction rate.

(7) The catalyst selectively reacted with some of the by-products produced in the reaction process, further converted them into simple small molecule safety substances, and then inhibited the production of toxic by-products.

The mechanism of the co-catalytic treatment of toluene and o-xylene by non-thermal plasma was shown in Figure 9. Figure 9a depicted the main reactions of toluene and o-xylene in the gas phase, and Figure 9b depicted the main reactions of toluene and o-xylene on the surface of the catalyst.

3. Experimental Instruments and Methods

3.1. Catalysts Preparation

The Cu-doped MnO₂ catalysts were prepared by the KMnO₄ redox precipitation method [63]. γ-Al₂O₃, with a diameter of 2–3 mm, was used as the support of the catalyst. Typically, 17.82 g of MnCl₂·4H₂O and the desired amount of CuSO₄·5H₂O (1.875 g, 5.625 g, 9.375 g, 18.75 g) were first dissolved in the 100 mL of DI H₂O. Then, 100 mL of the KMnO₄ solution (9.48 g) was added dropwise into the above solution with vigorous stirring. After being stirred for 24 h, the pre-dried γ-Al₂O₃ support was immersed in the solution and stirred for 6 h, and then washed and filtered alternately by DI water and ethanol. The obtained precipitates were dried in an oven at 120 °C for 12 h. The dried samples were further calcined in air at 300 °C for 2 h with the heating rate of 5 °C·min⁻¹. The catalysts were abbreviated as Cu_XMn for ease of reference, where x represents the molar ratio of Cu to Mn. In this experiment, the catalysts with different molar ratios were labeled with Cu_0.05Mn, Cu_0.15Mn, Cu_0.25Mn, and Cu_0.5Mn, respectively. In addition, the pure MnO₂ catalyst was also synthesized with a similar method for comparison.

3.2. Catalysts Characterization

The specific surface area, pore size, and pore volume of the catalysts were measured using N₂ adsorption-desorption isotherm (Mike Tristar 2020 surface area analyzer, Atlanta, GA, USA). The X-ray diffraction (XRD) pattern of the catalyst samples was performed on Brook D8 Advance with Cu-Kα radiation in the scan range of 10°~90° and scan rate of 2°·min⁻¹. The surface microtopography of the samples were characterized with a Hitachi S4800 JSM-7500F scanning electron microscope (Tokyo, Japan). The X-ray photoelectron spectroscopy (XPS) was measured with the Thermo ESCALAB 250Xi equipment (Waltham, MA, USA) with Al Ka X-ray radiation. All binding energies were corrected by standard carbon (C1s, 284.8 eV).

3.3. Experimental System

As shown in Figure 10, the experimental system was mainly composed of VOC generator systems, a DBD reactor, and an exhaust gas detection and analysis system. The total gas flow was controlled to 1 L·min⁻¹ by setting the flow rates of the two syringe pumps to 0.5 L·min⁻¹, respectively. The concentration of toluene and o-xylene was maintained at 700 mg·m⁻³ by controlling the temperature and air flow rate of the vaporization chamber.

The structure of the coaxial dielectric barrier discharge plasma reactor filled with catalyst is shown in Figure 11. The inner electrode is a stainless-steel cylinder in the middle of the reactor, and the blocking medium is a quartz tube coaxial with an inner electrode (inner diameter: 20 mm). The outer side of the quartz tube is covered with a layer of stainless-steel net as the outer electrode (inner diameter: 25 mm) [15]. The discharge area of the reactor is the gap between the outer surface of the inner electrode and the inner surface of the quartz. The discharge gap is 3 mm and the length of discharge zone is 150 mm [6]. The DBD reactor is powered by an AC high voltage power supply (CTP-2000 K, Nanjing Suman Plasma Technology, Nanjing, China). We adjusted the voltage and current by adjusting the knob of the high voltage supply. The digital storage oscilloscope (TBS1052B) was used to measure the input power and applied voltage [15].

3.4. Test and Analysis Methods

(1) The toluene and o-xylene concentrations were adsorbed on an activated carbon adsorption tube, desorbed by CS₂, and then they were measured by gas chromatography (Agilent Technologies 7890B, Santa Clara, CA, USA), which used a hydrogen flame ionization detector (FID) and a DB-FFAP capillary column. The conditions were as follows: the oven temperature was 40 °C, the detector temperature was 250 °C, and the inlet temperature was 150 °C [6].

The degradation efficiencies η₁ and η₂ were used to evaluate the treatment effect of toluene and o-xylene, respectively:

$${\mathsf{\eta }}_1=\frac{{[{\mathrm{C}}_7{\mathrm{H}}_8]}_0-{[{\mathrm{C}}_7{\mathrm{H}}_8]}_1}{{[{\mathrm{C}}_7{\mathrm{H}}_8]}_0}\times 100\%$$

(19)

$${\mathsf{\eta }}_2=\frac{{[{\mathrm{C}}_8{\mathrm{H}}_{10}]}_0-{[{\mathrm{C}}_8{\mathrm{H}}_{10}]}_1}{{[{\mathrm{C}}_8{\mathrm{H}}_{10}]}_0}\times 100\%$$

(20)

where ${[{\mathrm{C}}_7{\mathrm{H}}_8]}_0, {[{\mathrm{C}}_7{\mathrm{H}}_8]}_1$ and ${[{\mathrm{C}}_8{\mathrm{H}}_{10}]}_0, {[{\mathrm{C}}_8{\mathrm{H}}_{10}]}_1$ represented the mass concentrations (mg·m⁻³) of toluene and o-xylene before and after the reaction, respectively.

(2) The concentrations of CO₂ and CO at the outlet of the reactor were determined by gas chromatograph equipped with a thermal conductivity detector (TCD). The chromatographic column was an HQ-packed column, the detection temperature was 80 °C in the column box, and the front detector was 150 °C. The CO₂ and CO concentrations were calibrated by standard gas.

The selectivity of CO₂, CO, and CO_X were:

$${\mathrm{S}}_{{\mathrm{CO}}_2}=\frac{2[{\mathrm{CO}}_2]}{8\{{[{\mathrm{C}}_8{\mathrm{H}}_{10}]}_0-{[{\mathrm{C}}_8{\mathrm{H}}_{10}]}_1\}+7\{{[{\mathrm{C}}_7{\mathrm{H}}_8]}_0-{[{\mathrm{C}}_7{\mathrm{H}}_8]}_1\}}\times 100\%$$

(21)

$${\mathrm{S}}_{\mathrm{CO}}=\frac{2\left[\mathrm{CO}\right]}{8\{{[{\mathrm{C}}_8{\mathrm{H}}_{10}]}_0-{[{\mathrm{C}}_8{\mathrm{H}}_{10}]}_1\}+7\{{[{\mathrm{C}}_7{\mathrm{H}}_8]}_0-{[{\mathrm{C}}_7{\mathrm{H}}_8]}_1\}}\times 100\%$$

(22)

$${\mathrm{S}}_{{\mathrm{CO}}_{\mathrm{x}}}={\mathrm{S}}_{{\mathrm{CO}}_2}+{\mathrm{S}}_{\mathrm{CO}}$$

(23)

where $\left[{\mathrm{CO}}_2\right], \left[\mathrm{CO}\right]$ were the concentration of CO₂ and CO in the outlet gas, respectively.

(3) The Lissajous graph method was used to calculate the discharge power of the plasma. Figure 12 shows the Lissajous graph. The discharge power, P, of the DBD reactor in one discharge cycle can be obtained by integrating the area of the Lissajous graph [64]:

$$P_{consume}=f\times C\times S$$

(24)

where f was the frequency, C was the capacitance at 0.47 µF, and S was the area of the Lissajous figure.

(4) The specific input energy (SIE, kJ·L⁻¹) was used to indicate the power consumed by the plasma when processing a unit flow of gas:

$$\mathrm{SIE}=\frac{P_{consume}}{\mathrm{Q}}\times 60\times {10}^{-3} \left(\mathrm{KJ}·{\mathrm{L}}^{-1}\right)$$

(25)

where Q was the flow rate of the total gas (L/min).

(5) The comprehensive energy efficiency performance of DBD was evaluated by the energy efficiency (η_E, g·kWh⁻¹), which represented the removal quality of pollutants per unit energy:

$${\mathsf{\eta }}_{\mathrm{E}1}=3.6\times \frac{{[{\mathrm{C}}_7{\mathrm{H}}_8]}_0-{[{\mathrm{C}}_7{\mathrm{H}}_8]}_1}{\mathrm{SED}\times 1000} \left(\mathrm{g}·{\mathrm{kWh}}^{-1}\right)$$

(26)

$${\mathsf{\eta }}_{\mathrm{E}2}=3.6\times \frac{{[{\mathrm{C}}_8{\mathrm{H}}_{10}]}_0-{[{\mathrm{C}}_8{\mathrm{H}}_{10}]}_1}{\mathrm{SED}\times 1000} \left(\mathrm{g}·{\mathrm{kWh}}^{-1}\right)$$

(27)

(6) The content of O₃ was determined by the spectrophotometric method of indigo disulfonate; the concentration of NO₂ at the outlet of the reactor was determined by gas chromatograph equipped with a thermal conductivity detector (TCD) under the same conditions as (2).

(7) The intermediate products were analyzed by GC-MS (Agilent Technologies 7890B GC/5977B MSD). The column was an HP-5 capillary column (30.00 m in length, 0.32 mm in diameter and 0.25 µm in thickness). The detection conditions were as follows: the oven temperature was 30 °C, kept for 5 min. The subsequent program was raised to 120 °C, kept for 2 min at the heating rate of 3 °C·min⁻¹. The sample inlet temperature was 200 °C; the MS conditions were as follows: ion source temperature was 230 °C, quadrupole temperature was 150 °C, and mass-to-charge ratio scanning range was 35~270 amu.

4. Conclusions

In this paper, the effects of the Cu-doped MnO₂ catalyst on the toluene and o-xylene catalytic reactions in non-thermal plasma were investigated, and the VOCs’ removal efficiency, CO_X selectivity, and by-product formation were compared. It was found that the removal efficiency of VOCs and the CO₂ selectivity were significantly improved by introducing the catalyst into the plasma treatment process. When the catalyst Cu_0.15Mn was packed into the discharge region, this plasma-catalytic system showed the highest removal efficiency (toluene was 100% and o-xylene was 100%), the best CO₂ selectivity (92.73%), and the strongest by-product inhibition rate (36% for NO₂ and 56.6% for O₃) at 5.72 kJ·L⁻¹. It was concluded that the excellent performance of the Cu-doped MnO₂ catalyst was related to the formation of the Cu-Mn solid solution and surface reactive oxygen species. Finally, the possible degradation pathways of VOCs in plasma and plasma catalysis were proposed and discussed with the help of GC-MS. The results can provide a theoretical basis for the study on the performance of VOC degradation by NTP and have certain guiding significance for the application and mechanism research of VOC degradation by NTP.