Toluene Decomposition in Plasma–Catalytic Systems with Nickel Catalysts on CaO-Al 2 O 3 Carrier

: The decomposition of toluene as a tar imitator in a gas composition similar to the gas after biomass pyrolysis was studied in a plasma–catalytic system. Nickel catalysts and the plasma from gliding arc discharge under atmospheric pressure were used. The effect of the catalyst bed, discharge power, initial toluene, and hydrogen concentration on C 7 H 8 decomposition, caloriﬁc value, and unit energy consumption were studied. The gas ﬂow rate was 1000 NL/h, while the inlet gas composition (molar ratio) was CO (0.13), CO 2 (0.15), H 2 (0.28–0.38), and N 2 (0.34–0.44). The study was conducted using an initial toluene concentration in the range of 2000–4500 ppm and a discharge power of 1500–2000 W. In plasma–catalytic systems, the following catalysts were compared: NiO/Al 2 O 3 , NiO/(CaO-Al 2 O 3 ), and Ni/(CaO-Al 2 O 3 ). The decomposition of toluene increased with its initial concentration. An increase in hydrogen concentration resulted in higher activity of the Ni/(CaO-Al 2 O 3 ) catalysts. The gas composition did not change by more than 10% during the process. Trace amounts of C2 hydrocarbons were observed. The conversion of C 7 H 8 was up to 85% when NiO/(CaO-Al 2 O 3 ) was used. The products of the toluene decomposition reactions were not adsorbed onto its surface. The caloriﬁc value was not changed during the process and was higher than required for turbines and engines in every system studied.


Introduction
The gas produced during the gasification and pyrolysis of biomass contains a significant amount of pollutants, mainly aromatic hydrocarbons [1]. Depending on the type of biomass, the amount of tars varies from 5 to 100 g/Nm 3 [2,3]. As a result, up to 15% of the effective energy of biomass is lost, which is against the principles of a clean energy source [4]. In addition, unpurified biogas cannot be used as a fuel for engines (limit 50-100 mg/Nm 3 of tar) or turbines (5 mg/Nm 3 ) because the tar concentration is too high [5][6][7].
Many methods have been developed to remove tars using catalysts, plasma, adsorption, or filtration, depending on the requirements that the gas after pyrolysis or gasification must meet [8][9][10][11][12]. In catalytical methods, nickel has been shown to be the most efficient active phase while set on oxide carriers, such as Al 2 O 3 . The main disadvantage of using nickel catalysts has been deactivation of the catalyst, mostly due to the formation of carbon deposits on its surface [13][14][15][16].
Plasma and plasma-catalytic systems for toluene decomposition have been successfully conducted in many scientific groups [17][18][19]. The use of both plasma and catalysts deserves special attention similar to the tar decomposition method due to being more efficient compared to the separate use of catalyst and plasma. However, most studies have been conducted in small-scale reactors with low gas flow ratio and discharge power.
A new Ni 3 Al catalyst in the form of a honeycomb as well as commercial nickel catalysts designed for carbon oxide methanation (RANG-19PR) and methane water shift (G-0117) have been successfully used in plasma-catalytic systems for the decomposition of C 7 H 8 , using a flow rate of 1 Nm 3 /h and an inlet gas composition similar to that produced in biomass gasification [20][21][22][23]. However, these catalysts could be deactivated by carbon deposits formed during the process. Reducing the amount of active phase on the catalyst's bed could result in less soot formation. It was found that a high toluene conversion rate could be achieved with 10% nickel on Al 2 O 3 , but soot formation remained a problem; therefore, further studies were needed [24]. The addition of trace amounts of calcium to a catalyst used for catalytic reforming resulted in an increase in its resistance and activity by reducing the formation of carbon deposits [21,[25][26][27]. New nickel catalysts were prepared on a CaO-Al 2 O 3 bed to study the effect of calcium oxide addition in the decomposition of toluene in the plasma-catalytic system. Furthermore, catalysts' activity, the reaction products deposited on the catalyst bed, the calorific value of the outlet gas, the unit energy consumption, and the initial concentrations of C 7 H 8 and hydrogen in plasma-catalytic systems were studied.

Results and Discussion
In the plasma-catalytic system, higher toluene conversion (x C 7 H 8 ) was observed than in systems without catalysts (plasma only). The highest x C 7 H 8 was observed for the coupled system with the NiO/(CaO-Al 2 O 3 ) catalyst-85%. In the plasma-catalytic system with Ni/(CaO-Al 2 O 3 ) and NiO/Al 2 O 3 catalysts [24], the highest results were 77% and 82%, respectively. Without a catalyst, up to 68% of toluene was decomposed [21].
The toluene conversions obtained were lower than those obtained with a commercial nickel catalyst (G-0117, manufactured by INS Pulawy, Poland) used in the plasma-catalytic system (conversion-99%). It could be a consequence of lower temperature during the process and lower concentration of active phase on the catalyst [22,23,28].
The trace addition of calcium oxide onto the catalyst surface prevented the sintering of the catalyst and changed the interaction between the active phase and the catalyst carrier. This resulted in an increase in toluene conversion [26,29]. Studies on the effect of Ca addition on the steam reforming of ethanol showed that the addition of a trace amount of calcium to the nickel catalyst on Al 2 O 3 support increased coke formation on its surface. However, due to the rapid formation, it was less stable and amorphous, making it easier to oxidize than the graphitic carbon formed on the nickel catalyst on Al 2 O 3 without Ca addition [30].
A methanation reaction of carbon oxide occurred in all systems studied. A higher methane concentration was observed in a gas composition with a higher hydrogen concentration (series B) because the reaction rate at the catalyst was higher. More H 2 in the initial gas lowered the toluene conversion in plasma-catalytic systems with NiO/(CaO-Al 2 O 3 ) and NiO/Al 2 O 3 [24] but increased in the system with the Ni/(CaO-Al 2 O 3 ) catalyst. In series B, more hydrogen radicals were present during the process, which could be used not only for the methanation reaction but also for the hydrogenation of toluene radical, resulting in lower C 7 H 8 conversion rates in the system with NiO/(CaO-Al 2 O 3 ). The amount of CH 4 produced in the plasma-catalytic system with the use of Ni/(CaO-Al 2 O 3 ) catalyst reached 600 ppm (Figure 1), which was significantly higher than that produced with NiO/(CaO-Al 2 O 3 )-70 ppm-or Ni/Al 2 O 3 -50 ppm [24]. The addition of CaO to the catalyst carrier resulted in an increased methane concentration in the outlet gas. The highest conversions were observed with the highest initial toluene concentration (4500 ppm) in each system studied, as this increased the reaction rate on the catalyst surface. In contrast to previous studies, a slightly higher decomposition of toluene was observed when catalysts with oxidized nickel were used [22,23]. This could be related to the fact that byproducts of toluene decomposition were adsorbed on Ni/(CaO-Al2O3), which limited the activity of the catalyst due to blocked access to the active sites.
Toluene conversion rates were stable for NiO/(CaO-Al2O3) and NiO/Al2O3 catalysts and an increase in power did not increase because the catalytical process affected toluene decomposition to a greater extent than the plasma one ( Figure 2). While the Ni/(CaO-Al2O3) catalyst was used, conversion rates increased with increasing discharge power because the process was more dependent on plasma, and more active radicals could react with toluene.
In previous studies, a reduction of NiO to Ni was observed when the G-0117 [23] catalyst was used. This had a positive effect since Ni was more active than NiO in the decomposition of tar. However, in the case of this study, this might have had a negative influence. The reason was the different Ni/Ca ratio in the catalyst bed. A very low addition of calcium (Ca/Ni below 0.2) increased the resistance of the Ni/CaO-Al2O3 catalyst to sintering and the formation of carbon deposits, while a higher ratio resulted in greater soot formation. This was due to the growth of Ni crystallites and the coverage of the catalyst surface with Ca, which hindered the interaction of Ni with the catalyst bed [26,30]. The optimal Ca/Ni ratio for the NiO/CaO-Al2O3 catalyst might be different from that of the Ni/CaO-Al2O3 catalyst, which would explain why the addition of calcium did not decrease the toluene conversion rate in the plasma-catalytic system with NiO/CaO-Al2O3 [29].
In previous studies with plasma-catalytic systems and nickel catalysts, it was observed that turning off the plasma discharge resulted in a rapid decrease in toluene conversion [21]. Therefore, the process of toluene decomposition required both the presence of plasma and temperature to achieve high C7H8 conversion rates. The highest conversions were observed with the highest initial toluene concentration (4500 ppm) in each system studied, as this increased the reaction rate on the catalyst surface. In contrast to previous studies, a slightly higher decomposition of toluene was observed when catalysts with oxidized nickel were used [22,23]. This could be related to the fact that byproducts of toluene decomposition were adsorbed on Ni/(CaO-Al 2 O 3 ), which limited the activity of the catalyst due to blocked access to the active sites.
Toluene conversion rates were stable for NiO/(CaO-Al 2 O 3 ) and NiO/Al 2 O 3 catalysts and an increase in power did not increase x C 7 H 8 because the catalytical process affected toluene decomposition to a greater extent than the plasma one ( Figure 2). While the Ni/(CaO-Al 2 O 3 ) catalyst was used, conversion rates increased with increasing discharge power because the process was more dependent on plasma, and more active radicals could react with toluene. Energy efficiency (EE) calculations were used to compare the results obtained with those of other studies that used gliding arc discharge (GA) and a nickel catalyst in a coupled plasma-catalytic system for tar decomposition. At an initial toluene concentration of 4500 ppm (18.5 g/m 3 ), EE values for NiO/Al2O3, NiO/(CaO-Al2O3), and  In previous studies, a reduction of NiO to Ni was observed when the G-0117 [23] catalyst was used. This had a positive effect since Ni was more active than NiO in the decomposition of tar. However, in the case of this study, this might have had a negative influence. The reason was the different Ni/Ca ratio in the catalyst bed. A very low addition of calcium (Ca/Ni below 0.2) increased the resistance of the Ni/CaO-Al 2 O 3 catalyst to sintering and the formation of carbon deposits, while a higher ratio resulted in greater soot formation. This was due to the growth of Ni crystallites and the coverage of the catalyst surface with Ca, which hindered the interaction of Ni with the catalyst bed [26,30]. The optimal Ca/Ni ratio for the NiO/CaO-Al 2 O 3 catalyst might be different from that of the Ni/CaO-Al 2 O 3 catalyst, which would explain why the addition of calcium did not decrease the toluene conversion rate in the plasma-catalytic system with NiO/CaO-Al 2 O 3 [29].
In previous studies with plasma-catalytic systems and nickel catalysts, it was observed that turning off the plasma discharge resulted in a rapid decrease in toluene conversion [21]. Therefore, the process of toluene decomposition required both the presence of plasma and temperature to achieve high C 7 H 8 conversion rates.
Energy efficiency (EE) calculations were used to compare the results obtained with those of other studies that used gliding arc discharge (GA) and a nickel catalyst in a coupled plasma-catalytic system for tar decomposition. At an initial toluene concentration of 4500 ppm (18.5 g/m 3 ), EE values for NiO/Al 2 O 3 , NiO/(CaO-Al 2 O 3 ), and Ni/(CaO-Al 2 O 3 ) catalysts in plasma-catalytic systems were up to 7.5 g/kWh, 9.0 g/kWh and 8.9 g/kWh, respectively. It decreased with the increase in SEI, while the conversion rates of toluene remained stable (Figure 3).  The calorific values of the gas after the process were similar to the initial calorific values. They were as high as 5.8 MJ/m 3 and 7.3 MJ/m 3 for series A and B, respectively, in the system with the NiO/(CaO-Al2O3) catalyst (Figure 4). Tar imitator conversion rates obtained in this study were lower than those reported in other groups in which as much as even 95.7% of toluene and 83.4% of naphthalene were decomposed when the Ni-Co/γ-Al 2 O 3 catalyst was used in the plasma-catalytic with gliding arc plasma system [31]. Other studies also reported a very high conversion of toluene up to 95.2% when the Ni/γ-Al 2 O 3 catalyst with rotating gliding arc discharge was used [32]. The energy efficiency values obtained in the studies were in a similar range to those reported in the literature, i.e., 6.7 g/kWh [20] and 3.6 g/kWh [33]. However, when GA discharge plasma was used, much higher EE values were obtained, up to 46.3 g/kWh [34]. The difference between EE and toluene conversion rates reported in the literature and in this study could be due to the lower gas flow rate (up to 0.54 Nm 3 /h) [32], which allowed for longer residence time and contact of excited radicals with toluene particles.
The calorific values of the gas after the process were similar to the initial calorific values. They were as high as 5.8 MJ/m 3 and 7.3 MJ/m 3 for series A and B, respectively, in the system with the NiO/(CaO-Al 2 O 3 ) catalyst ( Figure 4). Figure 3. Effect of SEI on toluene conversion and energy efficiency for 4500 ppm initial C7H8 concentration in the plasma-catalytic system with NiO/(CaO-Al2O3) catalyst. Series A.
The calorific values of the gas after the process were similar to the initial calorific values. They were as high as 5.8 MJ/m 3 and 7.3 MJ/m 3 for series A and B, respectively, in the system with the NiO/(CaO-Al2O3) catalyst ( Figure 4).

Catalysts Preparation
Catalysts NiO/Al2O3, NiO/(CaO-Al2O3), and Ni/(CaO-Al2O3) with 10 wt% active phase on Al2O3 and on the commercial catalysts support G-2117-7H/C (CaO-Al2O3), manufactured by INS Pulawy, Poland, had a specific surface area not exceeding 10 m 2 /g ( Table 1). In previous studies, catalysts on the G-117 bed were successfully used and the optimal concentration of the active phase was identified to be 10 wt% [21][22][23][24]. Catalysts were prepared by the impregnation method using Ni(NO3)2•6H2O, which was deposited   Table 1). In previous studies, catalysts on the G-117 bed were successfully used and the optimal concentration of the active phase was identified to be 10 wt% [21][22][23][24]. Catalysts were prepared by the impregnation method using Ni(NO 3 ) 2 ·6H 2 O, which was deposited on the catalyst support and dried at 90 • C for 3 h. Afterward, they were calcinated at 500 • C for 5 h. A catalyst with metallic active phase was reduced at 400 • C for 14 h under 4 Nl/h hydrogen flow rate. Table 1. The specific surface area before and after the process [24].

Methods
Two gas compositions similar to the gas composition after biomass gasification were used (series A and B) in the plasma-catalytic systems. The CO and CO 2 amounts were the same for both series ( Table 2). The total inlet gas flow was 1000 Nl/h, and the initial toluene concentrations were 2000, 3000, and 4500 ppm. The gliding arc discharge power was in the range of 1500-2000 W and was measured with an energy meter, Schrack MGDIZ065. The catalyst bed was placed over the end of the electrodes. A detailed description of the reactor was given in a previous article [23]. Gas samples were taken for three increasing discharge power; afterwards, the discharge power was reduced to the initial value (1500 W) to study the effect of the catalyst temperature on toluene decomposition. The results obtained were compared to those of plasma-only and plasma-catalytic systems with commercial nickel catalysts [21][22][23].
After the process, the catalysts were taken out from the reactor and rinsed with acetone. The solutions obtained were analyzed by Thermo-Scientific ISQ mass spectrometer. The gas composition before and after the process was analyzed using gas chromatograph Agilent 6890 N with ShinCarbon column and TCD and FID detectors.
Toluene conversion rate was calculated using the following equation [23]: x-toluene conversion, Co-initial toluene concentration (g/m 3 ), and C-toluene concentration on outlet gas (g/m 3 ). For the calculation of the calorific value, the following equation was used [21,23]: W-calorific value (MJ/m 3 ); Q p -heat of combustion (kJ/m 3 ); and n-mole fractions CO, CH 4 , C 2 H 2 , C 2 H 4 , and C 2 H 6 . For the calculation of specific energy input (SEI) and energy efficiency (EE), the following equations were used [34]: SEI-specific energy input (kwh/m 3 ), P-discharge power (kW), and Q-total flow rate (m 3 /h).

Reaction Mechanism
The catalysts were rinsed with acetone after the process; then, the solution obtained was analyzed with MS to identify toluene decomposition products ( Figure 5). Retention times (RTs) between 0.48 and 0.54 min were methanol, acetone, water, and C2-C4 hydrocarbons. 2-Pentanone was identified (RT = 2.75 min) as a product of the acetone condensation reaction. The proposed reaction mechanism was based on previous [20,21,24] and current studies' toluene decomposition intermediate products as well as decomposition mechanisms proposed in the literature [27,[31][32][33][34][35][36]. The process was inducted by excited electrons, which could cause both radical and electron dissociation of the particles. Electron attachment to toluene particles could lead to the formation of CH 3 or H radials. Since the gas consisted of nitrogen, carbon oxides, and hydrogen, those could also be dissociated by the impact of electrons [21]. Water was formed during the methanation reaction. The presence of humidity in nitrogen-rich gas was a source of HO • and NO • radicals, but a higher density of hydrogen oxide radicals was observed [37]. The reason could be that the source of nitrogen radicals, the N≡N bond, required more energy to break (9.8 eV) than the H-OH bond in the water molecule (5.11 eV) [35].
hydrocarbons. 2-Pentanone was identified (RT = 2.75 min) as a product of the acetone condensation reaction. The proposed reaction mechanism was based on previous [20,21,24] and current studies' toluene decomposition intermediate products as well as decomposition mechanisms proposed in the literature [27,[31][32][33][34][35][36]. The process was inducted by excited electrons, which could cause both radical and electron dissociation of the particles. Electron attachment to toluene particles could lead to the formation of CH3 or H radials. Since the gas consisted of nitrogen, carbon oxides, and hydrogen, those could also be dissociated by the impact of electrons [21]. Water was formed during the methanation reaction. The presence of humidity in nitrogen-rich gas was a source of HO • and NO • radicals, but a higher density of hydrogen oxide radicals was observed [37]. The reason could be that the source of nitrogen radicals, the N≡N bond, required more energy to break (9.8 eV) than the H-OH bond in the water molecule (5.11 eV) [35]. No toluene decomposition reaction products were present on the surface of NiO/(CaO-Al2O3) catalysts, and only a trace amount of C7H8 (RT = 1.53 min) was adsorbed (Table 3). No nitrogen oxides in gas after reaction or nitrogen-containing intermediate products such as nitrotoluene on catalysts' surfaces were identified for all of the systems studied.  (Table 3). No nitrogen oxides in gas after reaction or nitrogen-containing intermediate products such as nitrotoluene on catalysts' surfaces were identified for all of the systems studied. Table 3. Compounds adsorbed on catalysts surface after the process [24].

Substance NiO/(CaO-Al 2 O 3 ) Ni/(CaO-Al 2 O 3 ) NiO/Al 2 O 3 [24]
Toluene On the nickel oxide catalyst without calcium addition, only 3-hexen-2-one was identified as a product of the toluene ring-opening reaction (5) (Figure 6). On the metallic nickel catalyst, tar formation was observed despite CaO addition in the catalyst carrier. A higher hydrogen concentration in the initial gas and an increase in discharge power resulted in a higher amount of hydrogen radicals, which could then react with other particles during the process. In the current study, the deactivation of Ni/CaO-Al 2 O 3 could be caused by too high of a Ni/Ca ratio on the catalyst bed, which led to a decrease in the number of active sites on the catalyst. A higher amount of hydrogen radicals did not lead to a higher conversion rate or resistance to deactivation in previous studies [24]. Instead, the radicals were used in the hydrogenation of toluene decomposition intermediates due to a lack of the active sites on the catalyst surface, which lowered the C 7 H 8 conversion rate.
Intermediate products of toluene decomposition further reacted with each other [36] to form polycyclic hydrocarbons (8 and 9), which were identified during the studies. Radical and electrons reacted with toluene decomposition products to form simple particles such as hydrogen, water, carbon oxide, and dioxide [24]. A trace amount of methanol identified at RT 0.48 min ( Figure 5) on the catalyst carrier was formed during methane oxidation [38]. For each catalyst, a similar gas composition was observed after the process. In the outlet gas, trace amounts of C2 hydrocarbons were formed. The concentration of hydrogen was lowered mainly due to the methanation process and hydrogenation of toluene decomposition products. In the presence of plasma or excited molecules, CO2 could decompose to CO and oxygen or oxide radicals [21], which then reacted with  Figure 6. Possible reactions of toluene in plasma-catalytic system.
In this study, polycyclic hydrocarbons such as diphenylmethane (RT = 11.93 min) were identified on the catalyst surface. The formation of benzyl alcohol identified in a previous study [24] was a product of oxidation reactions (6a and 6b) with HO • or oxide (7) radicals, which could also lead to the formation of other intermediates such as benzoic acid, benzene [20,21], or phenol [34] as a result of further oxidation reactions [35]. Intermediate products of toluene decomposition further reacted with each other [36] to form polycyclic hydrocarbons (8 and 9), which were identified during the studies. Radical and electrons reacted with toluene decomposition products to form simple particles such as hydrogen, water, carbon oxide, and dioxide [24]. A trace amount of methanol identified at RT 0.48 min ( Figure 5) on the catalyst carrier was formed during methane oxidation [38].
For each catalyst, a similar gas composition was observed after the process. In the outlet gas, trace amounts of C2 hydrocarbons were formed. The concentration of hydrogen was lowered mainly due to the methanation process and hydrogenation of toluene decomposition products. In the presence of plasma or excited molecules, CO 2 could decompose to CO and oxygen or oxide radicals [21], which then reacted with hydrogen to form water. CO 2 could not undergo the methanation process because its reaction is strongly inhibited by CO methanation. Moreover, the reaction of hydrogen with CO was twice as fast as with CO 2 [39]. During the process, the concentration of CO increased, although it was used for methane production. This also led to the conclusion that it was a product of the decomposition of carbon dioxide. This reaction had a favorable effect on the calorific value of the outlet gas.
Plasma-catalyst modeling, focused on the effects of both the catalysts surface and plasma effects, has not been extensively studied due to the complexity of the subject matter. However, several limited models have been reported. The mechanism of the chemical reactions involved in the removal of NF 3 in a hybrid plasma BaTiO 3 packed-bed with a CaCO 3 absorbent was modeled. Theoretically predicted byproducts were successfully identified by the authors during the experiments. The removal efficiency of NF 3 was improved with a BaTiO 3 dielectric constant enhancing electric field but decreased with the increase in gas flow [40]. A negative influence of gas flow on the conversion rate was also found when comparing the results presented in this paper with other studies [31][32][33]. Another group used fluid modeling and Particle-in-cell/Monte Carlo collision (PIC/MCC) to study the electric field enhancement and micro discharge formation in catalysts' pores on the catalyst-DBD system. Plasma formation in the pores occurred more easily in common catalysts supports such as Al 2 O 3 with low dielectric constant than in ferroelectric catalysts. This resulted in a larger contact area for plasma and catalysts and could improve the plasma-catalyst process [41]. On the contrary, the effect of electric field enhancement was more prominent in materials with higher dielectric constant, which could lead to stronger oxidative power of plasma discharge [42]. It became clear that further studies on the plasma-catalytic mechanisms are needed as this area has not yet been studied in depth.

Conclusions
In plasma-catalytic systems with nickel catalysts deposited on Al 2 O 3 and CaO-Al 2 O 3 , an efficient tar decomposition process was possible. High conversion of C 7 H 8 was observed-up to 85%, which exceeded the results obtained without the catalyst.
An increase in hydrogen concentration resulted in a higher conversion of toluene when the Ni/(CaO-Al 2 O 3 ) catalyst was used. Hydrogen was consumed in the methanation reaction and water formation. These reactions did not significantly affect the calorific value of the gas. With the addition of CaO, the catalyst bed decreased the amount of toluene decomposition products adsorbed on NiO/(CaO-Al 2 O 3 ) catalyst surface.

Conflicts of Interest:
The authors declare no conflict of interest.