Bi-reforming with a ratio of CH₄/CO₂/H₂O = 3/1/2 by gliding arc plasma catalysis for power to fuels Open Access

In this way, a syngas with H₂/CO = 2 has been achieved with excess CO₂ and H₂O feeding, as reported previously.^4–7 There have been no reports of direct production of a syngas with H₂/CO = 2 at the ideal stoichiometric ratio of CH₄/CO₂/H₂O = 3/1/2.

In nonthermal plasmas, electrons can be accelerated selectively in such a way that almost all the plasma energy goes into the formation of a thermal nonequilibrium state.⁸ Impacts between energetic electrons and molecules produces reactive species that can induce reactions. Plasmas are commonly generated by applying electrical power to a gas, which can be quickly switched on/off. Hence, it is feasible to power a plasma with fluctuating renewable electricity and use this as a form of fuel storage.⁹ To date, there have been no published reports about bi-reforming of CH₄ in plasmas, but various plasma types have been studied for dry, steam, and oxidative reforming of CH₄ to syngas.^10–14 Of these, warm plasmas (WPs), with intermediate gas and electron temperature between those of cold and thermal plasmas and formed by a gliding arc (GA), are believed to be among the most promising types,^13,15 especially for the conversion of CH₄ and CO₂, because they operate at atmospheric pressure and exhibits a high energy efficiency.^9,13,15 In addition, WPs have electron temperatures of 1–2 eV, and therefore provide a favorable environment for CO₂ conversion via efficient vibrational excitation.^15,16 However, previous works have concluded that plasma reforming suffers from low energy efficiency and poor conversion.^17,18 For practical application of plasma-based power to fuels, it would be of great interest to combine plasmas with catalysts to improve conversion, product selectivity, and energy efficiency.^19,20 WP reactors have already been used in combination with catalysts for CH₄ reforming, but the reactor designs were not optimized to supply the required activation temperature for the catalysts (∼800 °C), and thus low energy efficiencies (<50%) were obtained.^20,21 In our previous work, we have reported an efficient WP catalytic (WPC) reforming reactor combining a WP and a Ni/CeO₂/Al₂O₃ catalyst. This WPC reactor has shown excellent performance (high energy at high conversion) for oxidative and dry reforming of CH₄.^22–24 In addition, a novel route for power to fuel synthesis in this WPC reactor has been proposed.²² In the present work, we use the same reactor to investigate bi-reforming of CH₄ and demonstrate the production of a high-quality syngas with H₂/CO = 2 at the ideal stoichiometric ratio of CH₄/CO₂/H₂O = 3/1/2.

The WPC reactor is the same as that used in our previous work.^22,24 Briefly, the reactor is composed of a vortex gas flow GA reactor and a downstream tubular catalytic reactor, as illustrated in Fig. 1. To avoid carbon deposition in the plasma reactor, the reactants are fed from two inlets (F1 and F2). One flow (F1), which is a mixture of CH₄/CO₂/H₂O = 1.5/x/3 − x (x = 0–3), is used as a discharge gas to form a vortex flow in the GA reactor with a flow rate of 2.2 standard liters per minute (SLM). The other flow (F2), which is pure CH₄ with a flow rate of 0.7 SLM, is fed after the GA reactor but before the catalyst bed to adjust the total molar ratio of CH₄/CO₂/H₂O = 3/x/3 − x. The GA plasma is powered with a direct current power supply with a maximum voltage of 10 kV and a current of 100 mA. The discharge voltage, current, and power P_d are recorded by an oscilloscope (Tektronix DPO 4104B) via a voltage probe (Tektronix P6015A) and a 51 Ω sampling resistor, respectively.

The temperature of the catalytic reactor is kept by a homemade furnace at 1123 K, whose power P_h is monitored by a power meter. The Ni/CeO₂/Al₂O₃ catalyst pellets (∼2 mm diameter) with weight percentages of 11% nickel and 8% cerium, prepared by sequential incipient wetness impregnation as described in detail in Ref. 23, are packed into the catalytic reactor. This Ni-based catalyst is auto-reduced by the off gas from the GA plasma, and thus there is no pre-reduction in this process. The reactor can run in three modes: with a WP only (henceforth referred to as the WP case), with conventional catalysis (the CC case), and with a WP plus catalyst (the WPC case). The gas hourly space velocity (GHSV) in the CC and WPC cases is changed by varying the packing amount of catalyst.

In plasma reforming, especially in the case of endothermic reactions, the reactions are induced by electrons and reactive species produced by electron impact. Hence, the reactant conversion and product selectivity depend strongly on the plasma and electrical parameters. We measured the waveforms of discharge voltage, current and power at various SEI values (80–110 kJ/mol), gas flow rates, and CH₄/CO₂/H₂O ratios. The results show that similar trends are obtained under all the experimental conditions. Hence only the waveforms at an F1 molar ratio of CH₄/CO₂/H₂O = 1.5/1/2, a flow rate of 2.2 SLM, and an SEI of 110 kJ/mol is illustrated in Fig. 2. It can be seen that the GA discharge possesses the classic “ignition–gliding–extinguishment” pattern,^25,26 with an average cycle time of ∼1.6 ms. The arc channel is ignited at a voltage of ∼3.5 kV. Once the arc ignites, the discharge voltage drops to ∼2.6 kV at a current of ∼69 mA, and then the voltage and current respectively increase and decrease gradually to extinguishment/re-ignition with arc gliding. Balanced by the opposite trends of voltage and current, the discharge power fluctuates around 180 ± 6 W. Under the assumption that the largest arc length is equal to the distance from the high-voltage electrode to the edge of the GA reactor outlet (∼29 mm), the intensity of the electric field E along the arc channel is estimated to lie in the range of 90–120 V/mm during the arc evolution. This implies that the plasma generated in this GA reactor always remains in a nonthermal condition, but not undergo a transition from thermal to nonthermal.²⁷ In addition, the cycle time of this GA discharge (∼1.6 ms) is much shorter than that of the traditional 2D GA reactor (tens of ms), which can significantly enhance the interaction between plasma and reactants to increase reactant conversions.^27,28

In a plasma, dissociation of reactants through direct electron impact and stepwise vibrational excitation is the major contribution to the initiation of reactions.^29–31 This means that in a gas mixture of CH₄/CO₂/H₂O, plasma reforming should mainly start with the three reactions e + CH₄ → e + CH₃ + H, e + CO₂ → e + CO + O, and e + H₂O → e + H + OH. The H, O, and OH radicals thereby produced then induce chain reactions to form the final products. As mentioned above, the plasma parameters are insensitive to SEI, flow rate, and gas composition in this GA-based warm plasma, and thus the reaction pathway is only weakly dependent on SEI and flow rate at a given CH₄/CO₂/H₂O molar ratio. Changes in SEI and flow rates of reactants have strong effects on conversions, but only weak effects on product selectivities. This trend has been found previously,^32–34 and therefore we only present the effects of CO₂/(CO₂ + H₂O) molar ratio on bi-reforming of CH₄ in the WP case. As illustrated in Fig. 3, with increasing CO₂/(CO₂ + H₂O) molar ratio from 0 (steam reforming) to 1 (dry reforming) at an F1 flow rate of 2.2 SLM, an F2 flow rate of 0.7 SLM, and an SEI of 110 kJ/mol, X_CO2 and X_CH4 remain around 36% and 26%, respectively. However, $XH2O$ significantly decreases with increasing CO₂/(CO₂ + H₂O) molar ratio, even falling to negative values when the CO₂/(CO₂ + H₂O) molar ratio exceeds 0.5. X_CO2 is the highest among those three reactants, which can be attributed to the plasma parameters determined above. This GA-based warm plasma can achieve efficient vibrational excitation and provide favorable conditions for CO₂ activation.^15,35 In addition, CO₂ can also be reduced by H atoms produced as a result of CH₄ dissociation in a WP.^26,36 $XH2O$ is the lowest among the reactants because the cross section of H₂O for electron impact dissociation is lower than those of the other two molecules.^37,38 Moreover, the produced radicals are easily converted back to H₂O via their reactions. Figure 3(b) presents the effect of CO₂/(CO₂ + H₂O) molar ratio on selectivities of CO, C₂, and H₂. CO and H₂ are the main gaseous products, which indicates that steam and dry reforming are the main reaction routes in a WP. The by-products of C₂ hydrocarbons are generated by dehydrogenation of CH₄. As mentioned above, plasma parameters are insensitive to CO₂/(CO₂ + H₂O) molar ratio, which implies that the contributions of electrons to the conversion of CH₄, CO₂, and H₂O should be similar at various CO₂/(CO₂ + H₂O) molar ratios. Hence, variations in conversions and product selectivities are induced by changes in the radical distributions at various CO₂/(CO₂ + H₂O) molar ratios. When the initial reactions induced by electrons are considered, an increased CO₂/(CO₂ + H₂O) molar ratio results in more O atoms and CO being produced in the plasma. O atoms easily react with CH₄ and H to form OH radicals, and reduction of CO₂ by H atoms also produces OH radicals. These two pathways are more efficient than dissociation of H₂O. Obviously, this reduces H₂ selectivity and finally produces more H₂O to decrease its conversion.

As shown in Fig. 4, the H₂/CO molar ratio decreases monotonically with CO₂/(CO₂ + H₂O) molar ratio owing to the competition between dry and steam reforming of CH₄ in a WP. The ideal ratio H₂/CO = 2 is obtained at CO₂/(CO₂ + H₂O) = 1/3. The energy efficiency remains around (61 ± 1)% in the test range of CO₂/(CO₂ + H₂O) molar ratio.

Although H₂/CO = 2 is achieved at the ideal stoichiometric ratio of CH₄/CO₂/H₂O = 3/1/2 in a WP, the conversions and energy efficiency in a WP are still too low for practical application. To produce a high-quality syngas at a high conversion and a high energy efficiency, 11%Ni/8%CeO₂/Al₂O₃ catalyst is packed downstream of the GA-based warm plasma in the same way as in our previous work.²² The average gas temperature post GA reactor is measured by a thermocouple as ∼850 K at a SEI of 110 kJ/mol. This temperature is much lower than the required catalyst temperature for bi-reforming, and thus almost no reaction occurs over the catalyst bed if we simply pack the catalyst after the plasma, and the same reactant conversions and product selectivities are obtained as in the WP case. To activate the catalyst, a tubular furnace is used to maintain the catalyst bed temperature at 1123 K. By varying the amount of packed catalyst from 10 to 30 g, the effect of gas hourly space velocity (GHSV) on conversions and product selectivities was studied for the CC and WPC cases at the ideal CO₂/(CO₂ + H₂O) = 1/3 in the GHSV range of 5800–17 400 ml h⁻¹ g⁻¹. As illustrated in Fig. 5, all the conversions decrease with increasing GHSV. In the CC case, $XH2O$ becomes the highest and X_CO2 the lowest with increasing GHSV, because of the water gas shift reaction.⁴ These opposite trends in the CC and WP cases imply that CO₂ conversion is kinetically favored in the WP case and H₂O conversion in the CC case. In the WPC case, similar values of X_CH4, X_CO2, and $XH2O$ are achieved in the test range of GHSV, because WPC provides favorable conversion environments for both CO₂ and H₂O. In addition, the WPC case gives higher conversions than both the CC and WP cases, because it is a combination of these cases. In particular, the conversions of CO₂ (66%) and H₂O (64%) in the WPC case are higher than the sums of their respective values in the WP (X_CO2 = 37% and $XH2O=11%$⁠) and CC (X_CO2 = 26% and $XH2O=47%$⁠) cases at a GHSV of 17 400 ml h⁻¹ g⁻¹. By combining WP and catalyst, portions of the reactants, especially CO₂, are converted and produce active species in the plasma, and it is then possible for energy and reactive species with long lifetime carried by the plasma to reach the catalyst and enhance reactions. This higher reactant conversion also leads to the high syngas concentration in WPC case, which reaches a range of 85%–96% under the test GHSV conditions. By contrast, the syngas concentration in the CC case is only in the range of 67%–90%. This trend is also revealed by the temperature distribution of the catalyst bed in the CC and WPC cases. Figure 6 illustrates the temperature profiles of the catalytic reactor wall and catalyst bed at a GHSV of 17 400 ml h⁻¹ g⁻¹. Since the wall temperature of the catalytic reactor is controlled by the furnace in both the CC and WPC cases, it is similar in the two cases and is maintained at ∼1123 K along the catalyst bed. The temperature profiles measured along the central axis in the CC and WPC cases are much lower than the wall temperature, owing to the strong endothermic reaction over the catalyst. In addition, the temperature of the catalyst in the WPC case is higher than that in the CC case, which can be attributed to the WP supplying heat and active species in the WPC case.

The reactions of CH₄, CO₂, and H₂O are fed with the ideal stoichiometric ratio of bi-reforming (CH₄/CO₂/H₂O = 3/1/2). Theoretically, if all the reactants convert in equivalent amounts, the consumption rates of CH₄, CO₂, and H₂O should meet the values of R_CO2/R_CH4 = 1/3 and $RH2O/RCH4=2/3$⁠. With the WPC case, it is possible to overcome the respective disadvantages of the WP and CC cases. Both R_CO2/R_CH4 = 1/3 and $RH2O/RCH4=2/3$ are achieved in the test range of GHSV. This means that all the reactants are reformed with the ideal stoichiometric ratio for bi-reforming of CH₄ in the WPC case. This is a result of coupling the advantages of WP and CC for conversions of CO₂ and H₂O, respectively. In addition, with catalyst packing, only CO and H₂ are found with selectivities close to 100%, and almost all the C₂ hydrocarbons produced in the WP are converted to syngas. Therefore, the desired H₂/CO ratio of 2 is also obtained in the WPC case (see Fig. 7). With monitoring of the heating power of the furnace, the energy efficiencies for the CC and WPC cases can be calculated using Eq. (7). As the GHSV increases from 5800 to 17 400 ml h⁻¹ g⁻¹, the energy efficiency of bi-reforming of CH₄ in the WPC case decreases from 72% to 65%. In addition, to further compare CC and WPC cases, we characterized the used catalyst in the WPC case by x-ray diffraction and transmission electron microscopy. The results shows that the catalyst characteristics are similar to those found in previous work.²³ The nickel particle size is distributed in the range of 15–30 nm, with an average of ∼22 nm.

Bi-reforming in a warm plasma catalytic reactor using a warm plasma (WP) and a Ni/CeO₂/Al₂O₃ catalyst with the ideal stoichiometric ratio of CH₄/CO₂/H₂O = 3/1/2 has been investigated. A high-quality syngas with a H₂/CO = 2 is obtained at similar conversions of (89 ± 1)% for all of CH₄, CO₂, and H₂O. Electrical diagnostics show that the gliding arc plasma in a CH₄/CO₂/H₂O gas mixture is a warm one and can provide favorable conditions for CO₂ activation, as a consequence of which the highest conversion is achieved with a WP alone. A comparison of the results obtained with the WP alone (the WP case), a conventional catalyst alone (the CC case), and the WP plus catalyst (the WPC case) reveals that the WPC case can overcome the disadvantages of both the WP and CC cases. In the WPC case, CH₄, CO₂, and H₂O react with the ideal stoichiometric ratio of CH₄/CO₂/H₂O = 3/1/2 in the tested GHSV range. In addition, higher conversions and energy efficiencies are obtained in the WPC case than in the WP case.

See the supplementary material for the detailed products analytical method.

This work was supported by the National Natural Science Foundation of China (Grant No. 12375247).

The authors have no conflicts to disclose.

Jing-Lin Liu: Formal analysis (equal); Investigation (equal); Methodology (equal); Resources (equal); Writing – original draft (equal). Ai-Min Zhu: Conceptualization (equal); Writing – review & editing (equal).

The data that support the findings of this study are available within the article and its supplementary material.