Nickel catalyst in coupled plasma-catalytic system for tar removal


 Tar formation is a significant issue during biomass gasification. Catalytic removal of tars with the use of nickel catalyst allows to obtain high conversion rate but coke formation on catalysts surface lead to its deactivation. Toluene decomposition as a tar imitator was studied in gliding discharge plasma-catalytic system with the use of 5%, 10% and 15% by weight Ni and NiO catalyst on Al2O3 (α-Al2O3) and Peshiney (γ-Al2O3) carrier in gas composition similar to the gas after biomass pyrolysis. The optimal concentration of nickel was identified to be 10% by weight on Al2O3. It was stable in all studied initial toluene concentrations, discharge power while C7H8 conversion rate remained high – up to 82%. During the process, nickel catalysts were deactivated by sooth formation on the surface. On catalysts surface, toluene decomposition products were identified including benzyl alcohol and 3-hexen-2-one.


INTRODUCTION
Obtaining energy from solid biomass is one of the ways of reducing fossil fuels usage 1 . Gasifi cation of biomass converts it into CO, CO 2 , H 2 and CH 4 , which can be used as fuel for engines, turbines or as an ingredient for chemical synthesis 2-6 . One of the most signifi cant problems during gasifi cation of biomass is tar formation, which concentration may reach up to 100 g/Nm 3 . It is mostly containing mono and polycyclic hydrocarbons 7, 8 . Tars presence in the gas, may result in plugging and damaging the equipment, so it has to be cleaned before it can be used in turbines (tar limit 5 mg/ Nm 3 ) or engines (50-100 mg/Nm 3 ) 9, 10 . Various purifying techniques are being used, including physical, non-catalytical and catalytical methods. Compared to the other methods, use of catalysts decreases temperatures required for tar elimination and increases selectivity. The most widely studied catalysts are nickel based due to the high activity and lower costs compared to noble metals (Pd, Pt). However, they can be easily deactivated due to coke formation on the surface 11- 13 .
Plasma has been proven to be a promising solution for hydrocarbons reforming (i.e. toluene, naphthalene) but lacks catalyst's selectivity. Combining plasma and catalysts in hybrid system for tars removal results in higher conversion rate and selectivity to syngas than using those methods separately 14- 18 .
In previous studies, commercial catalysts were used in plasma-catalytic system. High toluene conversion was obtained (up to 99%) but large amount of soot was formed on catalysts surface 19, 20 . Amount of coke formed during tar cracking process increases with active phase concentration 21 . Lowering nickel concentration on catalyst bed may lead to lower soot formation and result in higher activity. For previously studied reduced nickel catalyst RANG-19PR 19 , once the discharge power was completely shut off, toluene conversion dropped signifi cantly, therefore toluene decomposition reaction might be depending on both temperature and discharge power. The purpose of this work was to identify the most suitable active phase concentration for toluene decomposition as tar model compound. Effect of initial C 7 H 8 and hydrogen concentration, discharge power, catalysts surface area on unit energy consumption, calorifi c value and toluene conversion were studied.

EXPERIMENTAL
Two gas compositions were used during these studies (series A and B), total inlet volumetric gas fl ow coeffi cient was 1000Nl/h. CO and CO 2 amounts were constant in each series (Table 1). Gliding discharge power in plasma--catalytic system was in range of 1500-2000 W. Toluene concentration in inlet gas was 2000, 3000 and 4500 ppm. Gliding discharge reactor was described in the previous work 19 . The catalyst bed was located above the end of the electrodes where the plasma discharge ended.
Polish Journal of Chemical Technology, 23,3,[24][25][26][27][28][29]10.2478/pjct-2021-0026 Table 1. Inlet gas composition The research was conducted using 1500, 1750 and 2000 W of discharge power. Power range of the reactor was limited due to plasma discharge being unstable below 1500 W and 2000 W being its highest achievable value. After the process conditions stabilized, samples were taken. During the studies, temperature of the gases increased with the increase of discharge power. To investigate temperature infl uence on toluene conversion and gas composition, additional samples were taken for low power and high temperature. Catalyst activity in time was also studied. Toluene conversion was calculated using below equation 20 : x -toluene conversion, c o -inlet toluene concentration, c -outlet toluene concentration.
To identify the composition of process products adsorbed on catalyst bed, its surface was risen with the acetone. Obtained samples were analyzed on mass spectrometer Thermo-Scientifi c ISQ. Gas composition both before and after the process was analyzed on gas chromatograph -Agilent 6890N with TCD and FID detectors on ShinCarbon column. Catalysts' specifi c surface area was determined using a BET absorption isotherm. ASAP2020 (Micromeritics) apparatus was used for nitrogen physisorption measurements. Calorifi c values were calculated to evaluate outlet gas utility as a fuel for engines using below relations 20 : W -calorifi c value, MJ/m 3 , Q p -heat of combustion, kJ/m 3 , a -molar fraction of CO, CH 4 , C 2 H 2 , C 2 H 4 and C 2 H 6 .
To choose the most effi cient system and active phase concentration, unit energy consumption was also calculated using below equation 22 :

Catalysts preparation
Catalyst NiO/Al 2 O 3 and Ni/Al 2 O 3 containing 5%, 10% and 15% by nickel weight on commercial catalysts bed G-117 (produced by INS Pulawy, Poland) had specifi c surface area below 15 m 2 /g. Nickel and nickel oxide catalysts on Peshiney bed (γ -Al 2 O 3 ) had surface 200 m 2 /g ( Table 2). As a precursor of active phase, (NO 3 ) 2 · 6H 2 O was used. It was placed on catalyst bed by impregnation method and dried at 90 o C for 3 hours, then it was calcinated at 500 o C for 5 hours. To obtain metallic active phase, nickel oxide was reduced under hydrogen volumetric fl ow coeffi cient equal to 14 Nl/h, at 400 o C for 14 hours.
Discharge power, nickel form and toluene concentration infl uence were not observed when using 10% Ni and NiO catalysts, which were nearly equally active in the most studied discharge power and initial toluene concentration. The highest conversion of toluene (82%) was obtained while using 4500 ppm of C 7 H 8 concentration. Toluene conversion dependency on initial concentration was not visible for 5% and 10% catalysts, as the radicals termination reactions were dominating over radicals reactions with toluene 23 . Only at 15% dependency on toluene initial concentration occurs.
The difference in toluene conversion rate for active phase concentration between 10% and 15% was significant, while between 5% and 10% was smaller. However, it was decided that 10% Ni concentration is the most benefi cial as conversion rates were stable in different discharge power and initial toluene concentrations. The uncertainty of toluene conversion was estimated at 5% and it was related to the measurement error.
The infl uence of higher concentration of hydrogen in inlet gas was studied (series B). For the most systems, higher conversions were obtained while using the gas with lower concentration of H 2 (series A) (Fig. 3). Low conversion rate for series A while using 3000 ppm toluene concentration and NiO as active phase might be the result of taking the sample before the system was stabilized. Methanation reaction was catalyzed in all studied systems, similar to previously studied nickel catalysts 19, 20 . Initial hydrogen and toluene concentration had a slight impact on the amount of methane formed during the process, it did not exceed 50 ppm. Molar fractions of inlet gas components did not change by more than 10% percent and remained stable during the process for both studied series (Fig. 4). Slight decrease in hydrogen concentration might be caused by methane formation and reactions with oxygen radical that could be formed by CO 2 dissociation 19 . Trace amount of oxygen was also detected in outlet gas.
In p reviously studied systems with the use of nickel catalyst (RANG-19PR, G-117) 19, 20, 24 toluene decomposition was rising with both temperature and discharge power. To check the infl uence of each of those separately, by the end of experiments when the temperature was the highest (250-300 o C), power was lowered to 1500 W. For both, reduced and oxidized nickel catalysts in past studies, conversion of toluene was higher while using high temperature at low power 1500 W. The next step with the use of RANG-19PR catalyst was to completely shut of the discharge power. It resulted in rapid decrease in toluene conversion rate 19 . Those steps have shown, that the studied process is not only catalytical or thermal decomposition but the plasma discharge is also important.
Dependence similar to previous studies was observed while using 5% and 10% active phase concentration (Fig. 5a). Conversion increased with both power and temperature. When power was reduced to 1500 W, the conversion was the highest due to higher temperature by 50 o C than using 2000 W. While using 10% Ni/Al 2 O 3 catalyst, conversion of toluene was constant with an increase of power and temperature (Fig. 5b). The high temperature generated in gliding discharge increased the reaction rate and produced radicals, which reacted with toluene and its intermediate products. Table 2. Specifi c surface area before and after process

RESULTS AND DISCUSSION
The measurements of toluene decomposition were performed in plasma-catalytic systems with the use of nickel catalysts. Inlet gas composition was similar to that obtained after biomass gasifi cation and contained hydrogen, carbon monoxide and carbon dioxide. Nitrogen was added to increase volumetric gas fl ow to 1000 Nl/h. Unit energy consumption (UEC) depended highly on concentration of C 7 H 8 . The lowest UEC was obtained while using 4500 ppm toluene concentration for each studied system. Nickel form and its concentration on catalyst surface did not have a signifi cant infl uence on UEC. Calorifi c value did not vary more than 10% from the initial value in most studied systems.
High toluene conversion, up to 80%, was obtained for systems using 5% to 15% by weight of nickel active phase (Fig. 1). Ni as active phase was more active than NiO. Comparison of fresh and used (after all studied series) 15% Ni/Al 2 O 3 catalyst, showed decreased activity of toluene decomposition and methanation reactions (Fig. 2). Samp les were taken using the same discharge power after the conditions have stabilized. In the systems with 15% NiO/Al 2 O 3 and Ni/Al 2 O 3 catalysts, higher amounts of methane were produced. It resulted in reaching higher temperatures during the process than while using the catalyst with lower active phase concentration.
The infl uence of the specifi c surface area of the catalyst bed was also studied in this work. Two catalysts with very different surface areas were compared. The specifi c area of α-Al 2 O 3 was not exceeding 10 m 2 /g while Peshiney's (γ-Al 2 O 3 ) area was over 200 m 2 /g. Surface areas were comparable both before and after the process, the difference might be coming from carbon deposits that formed during the process on the catalyst surface. Catalysts on Peshiney absorbed toluene below its boiling point, therefore samples were not taken while using low power and temperatures below 110 o C. Toluene conversions were lower than those obtained during this study for catalysts with a small surface area -up to 60% for NiO/Peshiney and 80% for Ni/Peshiney. Unit energy consumption was higher than in other studied in this work systems due to lower conversion, which was another disadvantage of using catalysts on Peshiney bed.
One of the problems during previous studies was mechanical damages of catalyst and carbon deposits formed on its surface, which led to catalysts deactivation. What is more, intermediate products of toluene decomposition could also be adsorbed on the surface of the catalyst. It resulted in blocking of the active sites on the surface and lowered its activity. Toluene decomposition products were identifi ed by mass spectrometry. In outlet gas benzyl alcohol, 3-hexen-2-one, methanol and trace amount of C2 were present as products of toluene decomposition. A possible reaction mechanism was proposed for identifi ed products (Fig. 6). Water, which was a product of carbon monoxides methanation was important source of excited HO • radical. The fi rst step of the decomposition process was H atom removal from methyl group. This C-H bond had the lowest energy in toluene particle so it was the most vulnerable for high energetic radicals infl uence. Reactive methyl radical was formed and might be oxidized to benzoic alcohol, aldehyde and acid. Further steps might lead to creation of benzene radical and phenol. High energy radicals  were able to break toluene ring to form stable particles like 3-hexen-2-one identifi ed in this work. Radicals and electrons were able to initialize further reactions that could lead to formation of simple particles like CO, CO 2 , hydrogen and water. Identifi ed methanol might be formed in catalytic methane oxidation 25-28 . Figure 6. Possible reaction mechanism of toluene decomposition

CONCLUSIONS
Toluene decomposition as a tar model was possible in coupled plasma-catalytic system using nickel catalysts in NiO and Ni forms. The optimal concentration of the active phase was identifi ed to be 10% of nickel by weight on Al 2 O 3 catalyst bed. Obtained high conversion -up to 82% was stable in all studied toluene concentrations and used power. Nickel catalysts were deactivated during the process due to soot formation. Initial gas similar in composition to the gas after biomass gasifi cation was benefi cial for tar elimination, additional hydrogen was not enhancing toluene decomposition process. Calorifi c value was stable in most systems.