Exploring Simultaneous Upgrading and Purification of Biomass−Gasified Gases Using Plasma Catalysis

Abstract

Tar and substantial CH₄ and CO₂ are contained in gasified fuels, which pose an obstacle to direct chemical synthesis, and this is a predominant challenge for biomass gasification technology. Herein, a packed−bed dielectric barrier discharge (DBD) reactor was built for simultaneous CH₄ dry reforming and tar removal with a La−Ni/γ−Al₂O₃ catalyst. The interaction between CH₄ dry reforming and tar removal in plasma catalysis was investigated. The results indicated that plasma catalysis can achieve high−efficiency simultaneous tar removal and CH₄ dry reforming, as indicated by the reactants’ conversion (14% increase for C_CH4 and C_CO2 at 450 °C in the presence of tar and a 37% increase for the tar removal rate at 360 °C when CH₄ and CO₂ were introduced), and the mechanism for mutual promotion of CH₄ dry reforming and tar removal was elucidated through catalyst characterization results. In addition, a possible reaction mechanism for tar removal via plasma catalysis was proposed. These findings provide valuable insights for simultaneous upgrading and purification of gases generated by biomass gasification.

1. Introduction

The continuously−increasing depletion of fossil fuels and the growing threat of global warming are driving the development of carbon−neutral biomass energy utilization technology [1,2,3]. Gasification, which involves partial oxidation of the carbon contained in biomass by controlling the oxidant amount and gasifying agent, is considered a promising and sustainable thermochemical conversion procedure for biomass utilization. Gasification enables high−efficiency conversion of biomass into valuable syngas, mainly containing H₂, CO and CH₄, and it provides great flexibility for the generation of heat and electricity or syntheses of chemical products and value−added byproducts [4,5]. However, gasification syngas inevitably contains non−directly utilized gases, such as CH₄ and CO₂, both of which contribute significantly to global warming [6]. In addition, the presence of tars produced from the gasification process is also a major challenge for most syngas applications. The tars in the producer gas cause scaling, blocking and corrosion of downstream equipment, limiting the use of syngas for industrial applications [7,8]. Thus, simultaneous effective CH₄–CO₂ conversion/reforming and tar removal from raw fuel gas are of great importance for the gasification process.

CH₄–CO₂ reforming, described as dry reforming of methane (DRM) (Equation (1)), is an effective channel for the utilization of greenhouse gases (GHGs) as carbon sources [9,10]. In addition, DRM has been recommended as a promising way to generate valuable syngas with a H₂/CO molar ratio equal to unity; syngas with this composition can be converted into value−added chemicals or used in chemical processes such as Fischer–Tropsch synthesis [11] and methanol synthesis [12].

CH₄ + CO₂ → 2CO + 2H₂ △H⁰_298K= + 247 kJ/mol

(1)

Plasma technologies have been regarded as attractive research topics due to their superior performance in terms of significant activation ability and purification ability and they have shown potential for application in CH₄–CO₂ reforming and tar processing [13,14,15,16,17]. Plasma technologies are classified into thermal and non−thermal plasmas, of which non−thermal plasmas (NTPs) produced through dielectric barrier discharge (DBD) are the more favourable due to their simple geometrical configuration, stability, reproducibility, steady operation and easy management [18,19]. A DBD reactor generates energetic and high−density electrons, which are accelerated by an applied electric field towards the positive electrode and collide with gas molecules in the discharge zone, inducing ionisation, excitation and dissociation [20,21]. Although DBD offers distinct advantages for use in unfavourable reactions at low temperatures [22], the high energy consumption and low selectivity may restrict its application [23,24]. In recent years, there has been an increasing amount of literature on the introduction of heterogeneous catalysts to DBD reactors, which has led to excellent synergistic effects in DRM and tar removal processes. The combination of DBD and catalysts can achieve high conversion and selectivity for the targeted products [25]. Heterogeneous catalysts, such as Ni/γ−Al₂O₃ [26], BaTiO₃ [27], Ni−Co/γ−Al₂O₃ [28], M/γ−Al₂O₃ (M=Ni, Co, Cu and Mn) [29], Ni/(sic) −Al₂O₃−MgO [30] and La−Ni/γ−Al₂O₃ [31], coupled with DBD for DRM have been widely studied and related experimental results have demonstrated that DBD catalysis exhibits a synergistic effect through catalyst surface discharge and micro−discharge [32]. Several investigations of tar removal via DBD have shown its superior performance, significant activation ability and purification ability, high resistance to carbon deposition and good product distribution [33]. Liu et al. [34] found that coupling DBD and Ni/γ−Al₂O₃ significantly enhanced the toluene conversion, hydrogen yield and energy efficiency in a hybrid plasma process. In the plasma treatment process, different carrier compositions have a significant impact on the tar removal [35]. Nair et al. [36] researched the effects of different gas compounds on naphthalene removal in a pulsed corona discharge reactor. They found that the main removal mechanism under a biogas atmosphere involved CO₂ decomposition that produced CO and O radicals. Wnukowski and Jamroz [37] found that the addition of CO₂ or H₂ with N₂ provided the optimal removal efficiency of 97–98%, while the addition of CH₄ decreased the efficiency to 62% due to benzene reformation. Although extensive research has been carried out on the dry reforming of CH₄ and tar removal through plasma–catalytic processes, simultaneous CH₄–CO₂ reforming and tar removal in a DBD–catalytic process has not been studied to date.

Herein, this work carries out a DBD–catalytic process for simultaneous effective CH₄–CO₂ reforming and tar removal. La−Ni/γ−Al₂O₃ catalysts possessing high activity and selectivity for both DRM and tar removal have been developed [31]. CO₂, CH₄ and a tar model compound (benzene) were added simultaneously to a plasma–catalytic reactor to better understand their interactions; experiments on tar removal under N₂ atmosphere and CH₄ dry reforming without tar addition were also studied. The physicochemical properties of the fresh and spent catalysts were characterized by XRD, N₂ physisorption, FTIR, XPS, SEM, TEM, Raman spectroscopy and TGA. Additionally, the liquid products collected after the reaction were analysed by GC–MS. All these experimental studies led to a deeper understanding of the interactions between DRM and tar removal. We aim to develop a strategy to achieve simultaneous effective CH₄–CO₂ reforming and tar removal in a DBD catalytic process, providing basic data and guidance for the simultaneous upgrading and purification of the gases derived from biomass gasification.

2. Results

2.1. Characterization of the Fresh Catalysts

The La−Ni/γ−Al₂O₃ catalyst, calcined at 600 °C, was characterized by XRD and the results of the XRD analysis are presented in Figure S1. The La−Ni/γ−Al₂O₃ catalyst exhibited obvious diffraction patterns with peaks at 2$\theta$ = 37.4, 39.7, 45.8, 60.4 and 67.3°, which corresponded to γ−Al₂O₃ (PDF#04−0880) with NiO (PDF#47−1049) peaks at 2$\theta$ = 37.2, 43.3, 62.9, 75.4 and 79.4°. However, no diffraction peaks related to La were detected, which may be attributed to highly dispersed nanosized La₂O₃ promoters with crystal sizes smaller than the XRD detection limit, as confirmed by other studies [38].

Microscopic analysis is widely used for the characterization of Ni−based catalysts and provides useful information regarding the morphological structures of the surfaces and interiors of solid samples. Figure 1a shows the HRTEM image of the fresh La−Ni/γ−Al₂O₃ catalyst. La₂O₃ (100), NiO (111), Ni (111) and γ−Al₂O₃ (311) planes are distinguished by the corresponding lattice fringes at distances of 0.348, 0.241, 0.203 and 0.24 nm, respectively. In addition, energy−dispersive X−ray spectroscopy (EDX) elemental maps of Al, Ni and La for the La−Ni/γ−Al₂O₃ catalyst are shown in Figure 1b. The above elements can be observed in the catalyst and Ni−based particles are fully dispersed with uniform distribution of La−based particles over the bulk of the catalyst, although aggregation occurs in some areas around the edges, demonstrating the successful preparation of the catalysts.

2.2. Interaction between Dry Reforming of CH₄ and Benzene Removal

In the plasma–catalytic process, dry reforming of CH₄ and benzene removal were studied simultaneously because interactions between the two processes were hypothesized to exist. To further understand the effect of tar addition on DRM and of CH₄ and CO₂ introduction on tar removal, CH₄–CO₂ reforming reactions with and without tar introduction and tar removal were carried out under N₂ or a gas mixture of CH₄, CO₂ and N₂ at various temperatures (MRT: CH₄ dry reforming with tar addition, MRO: CH₄ dry reforming only, TRO: tar removal only), with the results shown in Figure 2. For CH₄–CO₂ reforming in the presence of tar (MRT), C_CH4 and C_CO2 decreased first with increasing temperature, reached the lowest value at 330 °C and then increased significantly; CH₄–CO₂ reforming in the absence of tar (MRO) showed that C_CH4 and C_CO2 monotonically and gradually increased with increasing temperature. The former behaviour suggested that CH₄–CO₂ reforming and tar removal in the low−temperature zone (250–330 °C) mainly depended on gas−phase reactions excited by plasma and the discharge intensity may have dropped with increasing temperature [2,39]. In the high−temperature regime (330–450 °C), CH₄–CO₂ reforming is a strongly endothermic reaction, which could be enhanced with increasing reaction temperature. Moreover, the catalyst interacted with plasma to form a synergy significantly enhanced by high temperature, thus greatly improving the reforming performance [40,41]. C_CH4 reached maxima of 68% and 54% at 450 °C for two different cases: MRT and MRO, respectively. Significantly, the tar removal results were similar to the CH₄–CO₂ reforming results achieved under MRT; that is, the tar removal rate decreased at low temperatures and significantly improved at high temperatures.

For the interaction between CH₄–CO₂ reforming and tar removal, the results showed that the addition of tar significantly changed the CH₄–CO₂ reforming process, which was divided into two stages. At 288–400 °C, the C_CH4 and C_CO2 values for MRT were lower than the conversion achieved with MRO. However, upon increasing the temperature to 400–450 °C, the addition of benzene effectively improved the CH₄–CO₂ reforming performance: C_CH4 increased from 48.2% to 61.8% (~14% increase) and C_CO2 increased from 54.5% to 67.7% (~13% increase). Interestingly, a similar situation was observed for tar removal. The R_tar for MRT was lower than the value obtained by TRO in the low−temperature range (250–320 °C); when the temperature reached 320–450 °C, the presence of CH₄ and CO₂ significantly improved the tar removal effect and the maximum value was approximately 37% at 360 °C. There was a mutual promotion effect between CH₄ dry reforming and tar removal at high temperatures.

Although CH₄ and CO₂ exhibited high conversion rates at high temperatures with tar addition, the yield and selectivity for H₂ and CO were significantly reduced by tar addition (Figure 2d); this may indicate that the pollution of the catalyst was mainly caused by two factors. First, there was a competition between aromatic compounds and CH₄/CO₂ gases for active sites, which led to fewer catalytic conversion opportunities for DRM. Free radicals and the active components were generated from CH₄ and CO₂ activated by plasma, which may recombine with aromatic compounds to give low yield and selectivity for H₂ and CO; additionally, the generation of H₂ involved benzene decomposition, thus leading to the low yield for H₂. Second, deposition of carbon generated by aromatic ring cracking on the catalyst active sites affected the performance of the catalyst and its ability to selectively catalyse CH₄–CO₂ reforming [42]. From the above experimental results, there was a phenomenon worth exploring under MRT experimental conditions, as shown in Figure 2e. When the temperature was increased from 250 °C to 330 °C, the reactants’ (CH₄, CO₂, tar) conversion decreased, while the CO and H₂ selectivity increased from ~27% to ~42% and ~32% to ~45%, respectively; as the temperature was further increased to 450 °C, the reactants’ conversion increased significantly, while H₂ selectivity showed a decrease from ~45% to ~38%. Wang et al. [2] similarly observed that the reactants’ conversion first decreased and then increased with increasing temperature. Their further studies found that the mean electron energy of plasma decreased with increasing temperature and low−energy electrons were unable to break the molecular bonds of the reactants, resulting in lower products yield. They concluded that the synergistic effect of plasma catalysis was dominant at low temperature (250 °C), while the catalyst played a leading role at high temperature (550 °C) rather than the plasma. Wang et al. [39] observed that the increase in temperature negatively affected the dielectric coefficient of the medium, leading to a reduction in discharge strength. In the plasma–catalytic process at 330 °C, the lowest catalytic activity revealed an unsatisfactory plasma–catalyst synergy at 330 °C. High products selectivity implied that 330 °C was the trigger temperature for catalyst activity. However, the selectivity decreased when the temperature was increased to 450 °C. To further understand the catalytic behaviour of the reaction process, the fresh and spent catalysts were characterised. Figure S2 shows the TEM diagrams and the corresponding particle size distributions of the catalysts. The results showed that the particle size distribution at 330 °C remained almost constant compared to the fresh catalyst, but there was a small amount of particle sintering; a large amount of particle sintering occurred at 450 °C and the catalyst particles tended to agglomerate into larger particles, resulting in a decrease in the distribution of active sites, which was not conducive to mass transfer [43]. Figure S3b−e shows high−resolution C 1s XPS data for the spent catalysts and the results of peak deconvolution, with the relative contents of all the peaks of the spent catalysts listed in Table S1. The broad peak was deconvoluted into the peaks for graphitic carbon (C−C peak, at 284.8 eV), carbon bound to oxygen (C−O from hydroxyl or phenol at 286.54 eV and C=O from carboxyl at 288.86 eV) and the shakeup transition of aromatic carbons (π–π* transition, at 290.92 eV) [44,45]. For the MRT experiment conducted at 330 °C, the relative content of aromatic carbon (290.92 eV) reached 10.1%, considerably higher than the value of 2.8% achieved at 450 °C; this corresponded to the experimental results presented in Figure 2c, indicating that there were more aromatic compounds adsorbing to the catalyst surface. Moreover, the sample acquired under MRT at 330 °C exhibited a remarkable O relative content (66.5%), which was higher than the value obtained at 450 °C, which may be attributed to the high selectivity of the catalyst at 330 °C.

According to the results shown in Figure 2f, the H₂/CO molar ratios achieved by MRO were lower than 1 over the entire temperature range, which may be attributed to the reverse water gas shift reaction (RWGS, Equation (2)). It is noteworthy that the ratios obtained at low temperatures were higher than those obtained at high temperatures, mainly because the CH₄ decomposition reaction, Equation (3), was favoured at low temperatures, while the carbon gasification reaction, Equation (4), was accelerated at high temperatures. Compared with MRO, the H₂/CO molar ratios for MRT increased significantly at low temperatures while they were nearly constant at high temperatures. In the DRM reaction, the adsorption and dissociation of CH₄ typically occurred on metal active sites, while CO₂ showed a preference for adsorbing at the metal–support interfacial sites [46]. In addition, Table S2 shows the H₂ molar flow from different reactions (MRT, MRO, TRO) at 250~330 °C. It can be seen that the ratios of H₂ molar flow (MRO’s/TRO’s) at different temperatures were close to 10:1, indirectly indicating that H₂ production under MRT was mainly from the CH₄ cracking reaction, while the tar cracking reaction contributed little. Thus, higher H₂/CO ratios at low temperatures may reveal that the addition of benzene aggravates the competitive adsorption through the occupation of the metal–support interfacial sites, which may result in a lower CO yield.

CO₂ + H₂ → CO + H₂O △H⁰_298K= + 41.2 kJ/mol

(2)

CH₄ → C + 2H₂ △H⁰_298K= + 74.6 kJ/mol

(3)

C + CO₂ → 2CO △H⁰_298K= + 172.5 kJ/mol

(4)

Dry reforming of CH₄ and tar removal at 450 °C were significantly mutually promoting over the plasma–catalytic process: C_CH4 and C_CO2 were increased by ~13% with the presence of tar and the tar was completely removed when CH₄ and CO₂ were introduced. To further understand the mutual promotion effect, the physicochemical properties of the spent catalysts after 2 h of the MRO, TRO and MRT reactions at 450 °C were characterized by TGA, Raman, FTIR, XPS, N₂−physisorption, SEM and TEM techniques. This section discusses the effect of tar addition on CH₄–CO₂ reforming and the effect of CH₄ and CO₂ introduction on tar removal.

For the effect of tar addition on dry reforming of CH₄, the TGA profiles for the spent catalysts are shown in Figure 3a. After the introduction of benzene, the carbon content of the catalyst increased from 1.21% to 4.29%, which may significantly increase the risk of excessive carbon deposits on the active sites, hindering the adsorption of CH₄ and CO₂ on activated Ni/La metal particles. Figure 3b shows the Raman spectra for the spent catalysts. Two peaks located at 1340 and 1580 cm⁻¹ are observed in the Raman spectrum, corresponding to the D and G bands of carbon, respectively [47]. The D and G bands are due to the vibrations of sp³ disordered carbon and the vibrations of the completely graphitized sp² structure, respectively [48]. The ratio of G−band intensity and D−band intensity (I_G/I_D) indicated graphitization of the carbon species. A higher I_G/I_D value implies a greater degree of graphitization on the catalyst surface [49]. Figure 3b indicates that both disordered carbon and graphitized carbon were formed during the DRM reaction. The catalyst obtained from MRO exhibited the minimum I_G/I_D ratio, indicating that more defects were generated in the amorphous carbon, which may be beneficial for the adsorption and activation of reactants.

The FTIR spectra of the spent catalysts are shown in Figure S3a. According to the literature [50,51], the bands between 3429 and 3422 cm⁻¹ are attributed to stretching vibrations of hydroxyl (O−H) groups. The weak peaks located at 2919 cm⁻¹ represented C−H stretching vibrations for the cycloalkanes. A peak for C=C stretching vibrations appeared at 1600 cm⁻¹, which was due to the presence of olefins and aromatic rings [52]. Notably, the hydroxyl (O−H) vibrations were evident for the catalysts obtained from MRT and MRO, mainly due to the combination of H radicals and O radicals obtained from CO₂ and CO. A weak hydroxyl (O−H) vibrational peak was found for the TRO catalyst, mainly because of the adsorption of moisture from the air. Further illustration from the side showed that the hydroxyl (O−H) vibrational peaks for MRO and MRT were not from moisture.

As observed from the relative contents of the functional groups (Table S1), the C−C peak percentage for MRO was 27.4% and increased to 48.5% with tar addition, which was consistent with the results of the TGA. Notably, the relative contents of O (C−O and C=O) for MRO dropped from 70.3% to 48.6% (~21.7% decrease) when tar was added, further suggesting that the presence of benzene weakened the adsorption of CO₂ or oxygen−reactive species on the catalyst surface. More oxygen−active species were present in the gas phase and involved in the reaction of benzene derivatives. Therefore, it could be concluded that there was a bi−channel reaction path for CH₄−CO₂ reforming. First, CH₄ and CO₂ adsorbed on the catalyst surface were dissociated and excited by the plasma, inducing the formation of activated molecules and active oxygen species, which may react with carbon deposits or benzene derivatives. Second, CH₄ and CO₂ were activated directly in the gas phase by plasma, generating radicals that collide with benzene ring macromolecules or recombine with other reactant fragments. Although this was a faster reaction rate than the single−channel reaction pathway of MRO to improve CH₄ and CO₂ conversion, it came at the expense of H₂ and CO yield.

For the effect of CH₄ and CO₂ introduction on tar removal, as shown in the TGA profiles, the ~4.29% mass loss measured for the MRT process was lower than that for the TRO, suggesting that the introduction of CH₄ and CO₂ inhibited carbon deposition and released the active sites of the catalyst. In addition, from the Raman spectra results, the highest I_G/I_D ratio was exhibited by TRO, revealing that the presence of CH₄ and CO₂ mitigated graphitization and prevented catalyst deactivation due to severe graphitization. The hydroxyl (O−H) vibrational peak seen in the FTIR spectra was hardly found for the TRO catalyst but was remarkable after CH₄ and CO₂ were introduced; this indicated enhanced oxidation reactions in carbonaceous species, a claim that was supported by the TGA data. Additionally, the relative contents of O (48.6% for MRT > 30.4% for TRO) were consistent with the FTIR data, further confirming the beneficial effects obtained by the introduction of CH₄ and CO₂. It should be noted that the existence of O in TRO was due to the inevitable oxidation of the sample upon exposure to air during preparation. The general trend (68.3% for TRO > 48.5% for MRT) for the C−C peaks from graphitic carbon was consistent with the results of the Raman analysis shown in Figure 3a.

As shown in Table S3, the La−Ni/γ−Al₂O₃ catalyst exhibited a smaller specific surface area (97.445 m²·g⁻¹) than its unpromoted counterpart (99.114 m²·g−¹). The slightly decreased surface area indicated that the La nanoparticles were most likely well−dispersed on Al₂O₃ [53]. The BET specific surface area, total pore volume and average pore size of the spent catalysts were slightly lower compared with those of the fresh catalysts, indicating some partial blockage of the pores caused by the carbon deposits. Figure S4a shows that the fresh Ni/γ−Al₂O₃ and the La−Ni/γ−Al₂O₃ catalysts exhibited similar Type IV N₂ adsorption–desorption isotherms with typical H1 hysteresis loops, indicating that the addition of La did not change the mesoporous structure of the Al₂O₃ support. Ma et al. [53] found a similar result. In addition, the spent catalysts obtained under different reaction conditions were also analysed, exhibiting similar results (Figure S4b). This result demonstrated well−developed mesopores, despite such severe reaction conditions (simultaneous reactions with ~35 W DBD power at ~450 °C for 2 h). Figure S4c,d shows that the BJH pore width distributions for fresh and spent catalysts were concentrated within the range of 2–20 nm and the largest mesopore size was ~9 nm, i.e., in the low mesoporous range. Notably, Figure S4c shows that the number of centrally distributed pores for the La−Ni/γ−Al₂O₃ catalyst decreased slightly compared with those of the Ni/γ−Al₂O₃ catalyst. This was probably due to La particle modifications occurring during the impregnation step.

The SEM images of the fresh and spent catalysts (Figure 4a–h) show that the fresh catalyst exhibited spherical nanoparticles and that the spent catalyst surfaces were covered with filamentous carbon nanotubes. Together with the Raman analysis (Figure 3a), the presence of this filamentous carbon may indicate the high degree of graphitization [54]. Zhang et al. [55] investigated the performance of pure lanthanum nickelate (LNO) under plasma–DRM reaction. They found that the nanofibrous carbon deposition observed in the SEM image can be ascribed to the temperature rise and improved conversion for the catalyst [56]. Significantly, compared with the catalysts from MRT and MRO, abundant filamentous carbon was observed from the TRO process. Combined with Raman and TGA characterization data, this revealed that the filamentous carbon was mainly due to benzene decomposition; additionally, it may result in catalyst breakdown and blocking of the active sites via growth in the channels and pores on the catalyst surface, as observed in Figure 4e–h [57]. Figure 4i–k shows the TEM and HAADF images and EDX elemental maps of the spent catalyst for TRO. The Ni metal sites were covered by filamentous carbon nanotubes, causing the separation of Ni from the support and suggesting that carbon was formed at the Ni–support/promoter interface [58]; this confirmed that the addition of benzene aggravated competitive adsorption through occupation of the metal–support interfacial sites. It is important to note that a few hollow filaments were observed in Figure 4i. Snoeck et al. [59] reported similar observations and suggested that hollow filaments were formed with a high carbon nucleation rate. A similar phenomenon was found in other studies, suggesting the occurrence of the tip−growth mechanism on the Ni sites [60,61].

Figure 4d shows that the catalyst obtained under MRO exhibits strong nanoparticle aggregation and encapsulated graphitic carbon on the particle surface. In fact, relative to filamentous carbon, encapsulated graphitic carbon was the major cause of catalyst deactivation [62], because it could encapsulate the Ni active sites, making Ni inaccessible to the reactants. Wu et al. [63] found that the activation of CH₄ preferentially reacted on the exposed active Ni metal surface. Thus, the encapsulated graphitic carbon most likely was generated by the activation and decomposition of CH₄. Furthermore, Figure 4l,m shows the FFT–HRTEM images and phase identification for the spent catalyst obtained from the MRO process. From the FFT–HRTEM images, Ni and La₂O₃ were identified with the interplanar distances of 0.203 nm and 0.202 nm for the (111) and (110) lattice planes, respectively. It is worth noting that the presence of La in the catalysts mainly existed in the form of La₂O₃ before and after the reaction. Xu et al. [64] concluded that the La−O bond is very stable and unable to be reduced below 900 °C. In addition, the HRTEM image indicated the presence of a phase derived from graphitic carbon with an interplanar distance of 0.36 nm for the (002) lattice plane, indicating that the formation of encapsulated graphitic carbon occurred to the same extent for both La₂O₃ and Ni, which revealed that La₂O₃ was active for reactants in DRM and promoted the carbon generation reactions (CH₄ cleavage or CO disproportionation). Some studies have shown that the basic nature of La₂O₃ significantly facilitated the adsorption of CO₂ and the breaking of C=O bonds during DRM [65]. Figure S3f,g shows La 3d spectra and Ni 2p spectra for the spent catalyst obtained under MRO, from which it was found that the valence state of La was +3 and its main forms were La₂O₃ and La₂(CO₃)₃ at 834.8 [66] and 835.5 eV, respectively. The satellite peak of La₂(CO₃)₃ was at 839.11 eV and the ∆E between the satellite peak and La 3d5/2 peak was ~3.5 eV, which was the characteristic feature of the La³⁺ in the state of La₂(CO₃)₃ [67]. The formation of La₂(CO₃)₃ was attributed to the interaction between La₂O₃ and CO₂ [68]. In addition, ~852.3, ~853.6, ~855.3 and ~856.2 eV correspond to Ni⁰ [53], Ni²⁺ (NiO) [53], Ni²⁺ (Ni(OH)₂) [69] and Ni²⁺ (NiAl₂O₄) [67], respectively, of which the spinel peak was the strongest. A sharp spinel peak indicated the strong Ni−Al interaction formation; most of the Ni⁰ converted to Ni²⁺ due to the presence of CO₂ in the reaction atmosphere. There was a peak at ~850.22 eV related to La 3d3/2 region, which was at an overlap between La and Ni peaks. For the catalyst obtained from MRT (Figure 4b), no encapsulated graphitic carbon was observed on spherical nanoparticles and slightly filamentous carbon scattered on the catalyst surface was observed. This phenomenon can be attributed to a mutual promotion effect between CH₄–CO₂ reforming and benzene removal. First, the O involved in the CO₂ decomposition reduced the extent of filamentous carbon formation. Second, the benzene fragments were dissociated and converted into viable molecules through plasma catalysis and reacted with CH_X, efficiently inhibiting the formation of encapsulated graphitic carbon. The relevant reaction mechanism is shown in Figure 5.

2.3. Reaction Mechanism Analysis

To reveal the possible reaction pathways and mechanisms for benzene destruction in the plasma catalytic process, the collected liquid byproducts and their structures are listed in Figure S5 and Table S4.

Studies of simultaneous CH₄ dry reforming and benzene removal in the plasma reactor are complex due to the formation of free radicals, active molecules and ions [70]. A proposed reaction mechanism was based on the detection of liquid byproducts and the most thermodynamically favourable reactions [71]. The GC–MS analysis results shown in Figure S5 and Table S4 reveal the effects of pure N₂ and the gas mixture (N₂, CH₄ and CO₂) on the liquid phase byproduct distribution produced during tar removal. The distribution of byproducts from the gas mixture was similar to that of the pure N₂ atmosphere and mainly included aromatic hydrocarbons, aliphatic hydrocarbons and oxygen−containing compounds. The presence of aliphatic hydrocarbons proved the operation of a ring−opening reaction involving benzene and an intermediate of benzene. After the electron impact−induced dissociation of N₂ in the reaction system, excited nitrogen molecules N₂* were generated initially (Equation (5)). The destruction of benzene was caused by direct attack of energetic electrons and N₂*. The chemical binding energies determine the order in which different chemical bonds are destroyed. The C−H binding energy of the methyl group on the benzene ring is 3.7 eV, the C−H binding energy on the aromatic ring is 4.3 eV, the C−C binding energy between the aromatic ring and methyl group is 4.4 eV, the C−C binding energy on the aromatic ring is 5.0–5.3 eV, the binding energy of C=C is 5.5 eV and the average energy of electrons generated by ionization of macromolecules after energetic electron bombardment by plasma ranges from 1–10 eV, which provides a good dissociation and activation environment for the whole reaction system; thus, it becomes easier to realize chemical reactions that are difficult to achieve under normal conditions. In the plasma reaction, there are two main destructive ways to remove benzene: (a) benzene is directly attacked by energetic electrons and undergoes ring opening or (b) the benzene ring is bombarded by energetic electrons and N₂* to form phenyl, which reacts with active species to form new aromatic compounds.

N₂ + e → N₂* + e

(5)

Possible reaction pathways for benzene reforming under a N₂ atmosphere are shown in Figure 6a. After bombardment by energetic electrons and N₂*, benzene is first converted to phenyl because the dissociation energy of the C−H bond is the smallest in the benzene molecule [72]. Therefore, cracking of benzene can occur through this path to generate phenyl and then phenyl recombines with a methyl radical to form toluene. Toluene is the most abundant liquid−phase byproduct after benzene, which indicates that it is on the main path of forming liquid−phase byproducts. The generated toluene can also be used as a raw material for the formation of other liquid−phase byproducts, such as ethylbenzene and trans decahydronaphthalene. In addition, the C−C bond on the benzene ring can also be dissociated by energetic electrons, resulting in the formation of the HC=CH biradical and cyclobutadiene. The biradical HC=CH finally forms ethylene and cyclobutadiene can be used as a raw material for the production of naphthalene. However, no naphthalene peak was detected via GC–MS, indicating that in this reaction system full of energetic electrons, naphthalene finally generated trans−decahydronaphthalene through hydrogen reduction. The biradical HC=CH, produced by the dissociation of a C−C bond from the benzene ring, can be used as the raw material for the production of styrene. In addition, energetic electrons may also promote complete decomposition of benzene rings and intermediate compounds of benzene to produce methyl radicals and linear hydrocarbons, which was confirmed by the detection of n−heptane and n−decane via GC–MS.

A possible reaction pathway for reforming benzene under a gas mixture is shown in Figure 6b. CH₄ and CO₂ gas molecules collide with energetic electrons and N₂* to form CH_X, H, O, CO and H₂ free radicals or active molecules (Equations (6)–(11)), which increases the concentration of active substances in the reaction system and greatly improves the chance of reactions between active components and other reactants. CO₂ dissociates under the action of energetic electrons and introduces O active species [73]. According to the GC–MS results, unlike benzene reforming under a N₂ atmosphere, benzene removal under a gas mixture produces the oxygen−containing compound 2−butyl acrylate, which indicates that after ionization in the discharge area, O radicals participated in the reaction and the oxygen active particles obtained from collisions of energetic electrons with CO₂ molecules were recombined with the cracked fragments from benzene. In addition, the concentrations of toluene, ethylbenzene and styrene in the liquid phase byproducts obtained under the gas mixture were higher than those seen for the pure N₂ atmosphere, which can be explained by the increased concentrations of methyl and H radicals produced under the gas mixture. The concentration of benzene was lower than that seen in the pure N₂ atmosphere because benzene is more likely to be destroyed or reconstituted under the gas mixture than under the pure N₂ atmosphere.

CH₄ + e/ N₂* → CH₃ + H + e/N₂

(6)

CH₄ + e/ N₂* → CH₂ + H₂ + e/N₂

(7)

CH₄ + e/ N₂* → CH + H + H₂ + e/N₂

(8)

CH₄ + e/ N₂* → C + 2H₂ + e/N₂

(9)

CO₂ + e/ N₂* → CO + O + e/N₂

(10)

CO₂ + CH₄ → 2CO + 2H₂

(11)

3. Materials and Methods

3.1. Catalyst Preparation

The 2%La−10%Ni/γ−Al₂O₃ catalyst was prepared by an equal−volume impregnation method. The appropriate weight of γ−Al₂O₃ (160–200 items) was added to the metal precursor solution (Ni(NO₃)₂·6H₂O and La(NO₃)₂·6H₂O) and impregnated for 24 h. The above solution was dried at 105 °C until most of the water evaporated. Then, the obtained samples were calcined at 600 °C for 4 h in air atmosphere.

3.2. Experimental Setup

The experiments were carried out in a coaxial DBD reactor (Figure S6). The dielectric barrier discharge (DBD) plasma reactor was composed of a cylindrical aluminium tube (i.d.: 20 mm, o.d.: 25.8 mm) wrapped with stainless steel mesh in the middle of the tube. A stainless−steel bar (diameter 14 mm), which served as the inner electrode, was placed along the axis of the tube and a layer of steel mesh (thickness 2 mm, diameter 20 mm) was fixed at the bottom of the bar as a reaction bed support. The plasma was generated in the annular space between the coaxial cylindrical tubes. The corresponding discharge distance was 3 mm and the discharge volume was 16.02 mL. During the experiments, the La−Ni/γ−Al₂O₃ catalyst was placed in the discharge space.

The reaction temperature was controlled by a temperature controller and measured with a K−type thermocouple during the experiments. Before the reactions, the catalyst was reduced by pure hydrogen at 600 °C for 2 h. The DBD reactor was powered by an AC high−voltage power supply (Nanjing Suman Plasma Technology Co., CTP−2000K, Nanjing, China) with a peak voltage of 30 kV and a frequency of 5–20 kHz. In this research, the average input power and the frequency were 35 W and 7 kHz, respectively. The high voltage applied to the DBD was measured with a high voltage probe (Tektronix, p6015a, Beaverton, OR, USA). A capacitor (0.1 μF) was connected between the outer electrode of the reactor and the grounding electrode and all electrical signals were recorded with a digital oscilloscope (Tektronix, DPO2024B, Beaverton, OR, USA). The V−Q Lissajous method was used to determine the ability to discharge the reactor and then calculate the discharge power by multiplying the area of Lissajous by the frequency.

In the experiments, the gases (99.999% purity) were provided by high−pressure cylinders with the flow rates controlled by a mass flow controller (Beijing Seven Star Huachuang Flowmeter Co., D07, Beijing, China). Benzene vapour (9000 ppmv) was introduced into the reactor by using a nitrogen carrier gas that had flowed through a benzene−containing bubbler, which was placed in a water bath at a temperature of 40 °C. The heating belt was wrapped along the gas pipeline to ensure vaporisation of the benzene. The benzene, carried by the nitrogen stream and carrier gas (N₂:CO₂:CH₄ = 2:1:1), was fully mixed with the gas stream in the mixing chamber before feeding into the DBD reactor. The total flow rate and the concentration of benzene were kept at 300 mL/min (based on room temperature and atmospheric pressure) and 9000 ppmv (29.6 g/m³), respectively. At the outlet, the gas stream passed through an absorption bottle placed in an ice water bath and the flow rate was determined by a soap film flow meter. Unless otherwise stated, the tar discussed in this work is the tar model compound benzene.

3.3. Method of Analysis and Definition of Parameters

The composition of the produced gas was analysed with a gas chromatograph (GC, Shimadzu Corporation, GC−2014, Kyoto, Japan) equipped with a flame ionization detector (FID) for analyses of C1–C4 hydrocarbons and a thermal conductivity detector (TCD) used to analyse H₂, CO, CO₂ and CH₄. During the experimental process, the exhaust gas was introduced to an absorption bottle filled with 75 mL of hexane, which was placed in an ice water bath to collect the liquid products. The liquid products were analysed with an off−line gas chromatography–mass spectrometry system (GC–MS, Thermo Fisher Scientific (China) Co., Trace 1300−ISQ, Shanghai, China) equipped with a DB−5 capillary column. The exhaust gas was collected after running the DBD for 20 min when the reaction reached a stable situation.

For the CH₄ dry reforming reaction, the conversion rates for CH₄ and CO₂ were defined as

$$C_{C{H}_4}\left(\%\right)=\frac{ \mathit{moles} \mathit{of} C{H}_4 \mathit{converted} \left(mmol/s\right)}{ \mathit{moles} \mathit{of} C{H}_4 \mathit{input} \left(mmol/s\right)}\times 100\%,$$

(12)

$$C_{C{O}_2}\left(\%\right)=\frac{ \mathit{moles} \mathit{of} C{O}_2 \mathit{converted} \left(mmol/s\right)}{ \mathit{moles} \mathit{of} C{O}_2 \mathit{input} \left(mmol/s\right)}\times 100\% ,$$

(13)

The selectivity, S, and yield, Y, of the products can be calculated as

$$S_{H_2}\left(\%\right)=\frac{ \mathit{moles} \mathit{of} H_2 \mathit{produced} \left(mmol/s\right)}{2\times \mathit{moles} \mathit{of} C{H}_4 \mathit{converted} \left(mmol/s\right)+3\times \mathit{moles} \mathit{of} C_6{H}_6 \mathit{converted} \left(mmol/s\right)}\times 100\%,$$

(14)

$$S_{CO}\left(\%\right)=\frac{ \mathit{moles} \mathit{of} CO \mathit{produced} \left(mmol/s\right)}{ \mathit{moles} \mathit{of} C{O}_2 \mathit{converted} + \mathit{moles} \mathit{of} C{H}_4 \mathit{converted} +6\times \mathit{moles} \mathit{of} C_6{H}_6 \mathit{converted} }\times 100\%,$$

(15)

$$Y_{H_2}\left(\%\right)=\frac{ \mathit{moles} \mathit{of} H_2 \mathit{produced} \left(mmol/s\right)}{2\times \mathit{moles} \mathit{of} C{H}_4 \mathit{input} \left(mmol/s\right)+3\times moles \mathit{of} C_6{H}_6 \mathit{input} \left(mmol/s\right)}\times 100\% ,$$

(16)

$$Y_{CO}\left(\%\right)=\frac{ \mathit{moles} \mathit{of} CO \mathit{produced} \left(mmol/s\right)}{ \mathit{moles} \mathit{of} C{O}_2 \mathit{input} + \mathit{moles} \mathit{of} C{H}_4 \mathit{input} +6\times \mathit{moles} \mathit{of} C_6{H}_6 \mathit{input} }\times 100\%,$$

(17)

The H₂/CO molar ratio was defined as

$$\frac{H_2}{CO}=\frac{ \mathit{moles} \mathit{of} H_2 \mathit{produced} \left(mmol/s\right)}{ \mathit{moles} \mathit{of} CO \mathit{produced} \left(mmol/s\right)},$$

(18)

The tar removal efficiency was calculated as follows:

$$R_{tar}\left(\%\right)=\frac{ \mathit{amount} \mathrm{of} \mathit{benzene} \mathit{converted} \left(g/min\right)}{ \mathit{amount} \mathit{of} \mathit{benzene} \mathit{input} \left(g/min\right)}\times 100\%.$$

(19)

4. Conclusions

There was a certain mutual promotion effect operating between CH₄–CO₂ reforming and tar removal in the high−temperature range, as indicated by reactant conversion (~13% increase for C_CH4 and C_CO2 at 450 °C in the presence of tar and a ~37% increase for the tar removal rate at 360 °C when CH₄ and CO₂ were introduced). Strictly speaking, the reinforcing effect of CH₄ dry reforming was only understood to be the effectiveness of CH₄ and CO₂ conversion. The mutual promotion effect between CH₄ dry reforming and tar removal can be classified into two aspects:

• For CH₄–CO₂ reforming, the addition of benzene affected the physicochemical properties of the catalyst surface. The excessive graphitic carbon caused encapsulation of the active sites or coverage of the defective sites, which weakened the adsorption of CO₂ and CH₄. In addition, the addition of benzene increased the gas−phase reaction density by generating ring−opening products and benzene radicals via the plasma action, which reacted or recombined with the active components (CH_X, H, H₂, CO, O) that originated from CH₄ and CO₂.
• For tar removal, the introduction of CH₄ and CO₂ improved the catalyst surface activity by increasing the concentration of reactive oxygen species on the catalyst surface, which in turn increased the resistance of the catalyst to graphitization and carbon deposits. This further enhanced the oxidation of carbonaceous species on the catalyst surface and facilitated the cracking of benzene to gas phase products (CO, CO₂) via indirect or direct reaction pathways. On the other hand, the introduction of CH₄ and CO₂ may increase the amount of H, O, H₂ and CO_X in the gas phase by plasma action, facilitating benzene reforming and ring−opening reactions; for example, benzene radicals recombine with CH_X to form other aromatic compounds (e.g., toluene, ethylbenzene and styrene).

Finally, although several reasons for the mutual promotion of tar removal and CH₄ dry reforming have been summarized, there is still some work that needs to be further investigated. For example, the effects of benzene, CH₄ and CO₂ on the plasma system, such as changes in the discharge morphology or discharge patterns due to changes in the electrical conductivity, have not yet been investigated and will be carried out in future work.