The experiments were conducted in an ED-DBD reactor, consisting of two coaxial quartz tubes with a wall thickness of 1 mm and three electrodes, as shown schematically in Figure 1. The reactor was filled with SiO₂ beads or SiC blocks, and the discharge gap between the tubes was 3.5 mm. The inner tube was filled with a copper rod with a diameter of 5 mm as a high-voltage electrode. The ED-DBD reactor was tightly wrapped around a copper sheet with a length of 6 cm, as the first ground electrode. Four stainless steel electrodes (2 mm in diameter) were added to the discharge gap as the second ground electrodes, which were closely fitted to the inner wall of the outer quartz tube and paralleled to the high-voltage electrode. The packing material (i.e., SiO₂ beads and SiC blocks) was used in the size fraction of 1–2 mm filled the entire discharge gap.

The experimental system, as shown in Figure 2, consisted of the gas supply, discharge and detection systems.

The gas supply system used high-purity argon (99.999%, 13.5 MPa, Shandong Honda Biotechnology Co., Ltd.), benzene (0.06%, the rest was argon, 5.0 MPa, Shandong Hongda Biotechnology Co., Ltd.), and CO₂ (1.0%, the rest was argon, 10.0 MPa, Shandong Hongda Biotechnology Co., Ltd.) as background gas, model molecule, and oxidizing agent, respectively. CO₂ and benzene were mixed in a gas mixing bottle to obtain an initial concentration of 3,000 ppm CO₂ and 300 ppm benzene. The total gas flow rate was 250 mL/min.

The discharge system utilized high-voltage AC power supply (CTP-2000K, Nanjing Suman Plasma Technology Co., Ltd.) to drive the electric field local enhanced DBD reactor, with a discharge voltage range of 4–12 kV and discharge frequency range of 6–10 kHz. The discharge voltage and current waveforms were detected using a high-voltage probe (Tektronix, P6015A) and current probe (Tektronix, TCP0030A), and the oscilloscope (Tektronix, MDO3054) recorded the voltage, frequency, current, and other discharge data of the reactor discharge process.

The product detection system consisted of a gas chromatograph (Agilent GC 7890B with FID and TCD detectors) and an emission spectrometer (Princeton Instruments, SP-2750). The benzene concentration was quantitatively detected using a DB-HeavyWax (J&W 123–7163) column with a high-purity nitrogen carrier gas (99.999%, 13.5 MPa, Shandong Hongda Biotechnology Co., Ltd.). The vaporization chamber temperature was set at 250°C, the column box temperature was initially set at 40°C for 5 min, then raised to 100°C at a rate of 15°C/min for 2 min, and the FID detector operating temperature was 300°C with hydrogen flow rate of 32 mL/min (99.999%, 13.5 MPa, Shandong Hongda Biotechnology Co., Ltd.) and an air flow rate of 400 mL/min (99.999%, 13.5 MPa, Shandong Hongda Biotechnology Co., Ltd.). The CO₂ concentration was detected using a GS-CarbonPLOT (Agilent 113–3133) column with a vaporization chamber temperature of 180°C, a 30:1 shunt, a column box temperature initially set at 40°C for 10 min, and TCD detector temperature of 200°C. The CO and H₂ concentrations were detected using HP-PlotQ (Agilent 19095P-Q04) and HP-Plot Molesieve (Agilent19091P-MS4) columns in series by chromatographic center cutting technology. The carrier gas was high-purity argon (the specifications were the same as above), with a flow rate of 10.917 mL/min, a vaporization chamber temperature of 180°C, and a column box temperature initially set at 35 °C for 11.5 min. The programmed temperature was then increased to 120°C at a rate of 15°C/min for 10 min, and the TCD detector temperature was 250°C. The tail blowing gas was high-purity argon (the specifications were the same as above), with a reference flow rate of 10 mL/min. The emission spectrum generated an emission spectrum for diagnosing the plasma reaction process, and the characteristic peak wavelength data was used to analyze the possible species in the plasma. The signal in the range of 200–800 nm was collected and processed, with working parameters including an exposure time of 80 m, a light intake slit of 500 μm, and a grating scale of 600 g/mm.

Before each discharge, the reactor was injected with the necessary gas source for 90 min. An infrared camera (FLIR T660) was used to measure the temperature variation of the reactor during the 20-min continuous operation. The temperature of the quartz glass tube on the right side of the first grounding electrode was specifically selected as the focal point, which is not representative for the plasma (gas) itself.

The discharge power was calculated using Equation 1, where P (kW) represents the plasma discharge power, C _M denotes the measurement capacitance, U represents the voltage measured by the high voltage probe, U _M represents the voltage of the measurement capacitance and f represents the discharge frequency. P = 1 T ∫ 0 T U I d t = f ∫ 0 T U   ×   C M   ×   d U M d t d t = f C M ∫ 0 T U d U M (1)

The benzene and CO₂ conversion rate Υ B and Υ C O 2 were calculated using Equations 2, 3, respectively. γ B = φ C 6 H 6 in − φ C 6 H 6 out   φ C 6 H 6 in × 100 % (2) γ C O 2 = φ C O 2 in − φ C O 2 out φ C O 2 in × 100 % (3)

The molar yield of CO and H₂ were calculated using Equations 4, 5, where Y B represents the molar yield of benzene degradation and Y C O represents the molar yield of CO₂ conversion. Y H 2 =   φ H 2 out 3 × φ C 6 H 6 in × 100 % (4) Y C O =   φ CO out φ CO 2 in + 6 × φ C 6 H 6 in × 100 % (5)

The energy efficiency of CO₂ reforming was the ratio of the chemical energy of CO₂ reforming of benzene harvested to the electrical energy spent, η C O 2 =   V 2 Δ H C O ⋅ Y C O + 0.5 Δ H H 2 ⋅ Y H 2 × φ CO 2 in φ CO 2 in + φ C 6 H 6 in 22.4 × 60 × P × φ CO 2 in × γ C O 2 × 100 %

In the above equation, V (L min^-1) represents the volumetric flow rate, P (kW) represents the plasma discharge power, ΔH _CO represents the fuel value of CO (283 kJ mol⁻¹), ΔH _H2 represents the fuel value of H₂ (285.8 kJ mol⁻¹), and Y _CO or Y _H2(mol min⁻¹) represents the CO or H₂ yield, Υ C O 2 o r   Υ B represents CO₂ or benzene conversion rate. The factor of 22.4 (L mol⁻¹) represents gas molar volume constant, and 60 is used to convert seconds into minutes.

The input power in dielectric barrier discharges (DBDs) plays a crucial role in determining the number of electrons generated, which subsequently affects the ensuing chemical processes. Figure 3 presented the discharge voltage and current waveforms for both the ED-DBD and D-DBD reactors, as well as the cases with packing materials. Comparing the blank D-DBD and ED-DBD reactors, as illustrated in Figures 3A, B, it was evident that the current pulse of the ED-DBD reactor was six times higher than that of the blank D-DBD reactor. However, this pulse occurred only after the current peak in the D-DBD reactor. Furthermore, the discharge intensity of the ED-DBD reactor exhibited an increase, with the pulse amplitude attaining its maximum at the current peak. This was likely due to the formation of a new electric field caused by the electrons adhering to the stainless-steel rods.When SiO₂ beads were introduced into the gap, the discharge intensity was further augmented as surface discharge occurs on the SiO₂ bead surface (Mujahid and Hala, 2018). As shown in Figure 3C, the amplitude of the current pulse increased, and a secondary current pulse occurred before the main current pulse due to surface discharge. On the other hand, in the case of SiC blocks filled, the amplitude of the main current pulse was maintained, while the amplitude of the secondary current pulses increased, and their number was greater than that in the case of SiO₂ beads, as depicted in Figure 3D. This was probably because the rougher surface of SiC could withstand a higher maximum current density than glass (Wang et al., 2015), resulting in more current pulses under the same discharge voltage.

The variation in discharge power with discharge voltage under different conditions was shown in Figure 4. For ED-DBD reactor, the discharge power could range from 59.8 W to 130.3 W for voltages ranging from 8 kV to 12.2 kV, while the range was 47.2–108.0 W for the D-DBD reactor. The partially narrowed gap caused by the second ground electrode in the ED-DBD reactor strengthened the local electric field and improved the discharge intensity, as shown in Figure 3. The discharge power was further enhanced by filling the reactor with insulating materials beads. For SiO₂ beads and SiC blocks, the discharge powers ranged from 63.4 W to 132.3 W and 65.8 W to 133.5 W, respectively. The maximum discharge power increased by 1.5% and 2.5%, respectively, compared to that without packing. Based on the above analysis, the ED-DBD reactor had the advantage of higher discharge power and more discharge current pulses compared to the D-DBD reactor. Therefore, subsequent experiments only focused on ED-DBD reactor under different filling conditions.

The temperature of the reactor is a critical parameter for assessing the conversion of electrical energy during the discharge process. An elevated reactor temperature signified a greater proportion of energy being converted into heat, thereby diminishing energy efficiency (Duan et al., 2021). Figure 5 demonstrated the variation in reactor temperature with discharge time under different conditions. As time advanced, heat energy accumulated within the discharge gap, leading to a gradual increase in reactor temperature. After 20 min of discharge, the temperature stabilized, following the order: T _{(SiC blocks in ED-DBD)} = T _{(SiO2 beads in ED-DBD)} > T _(ED-DBD) under different conditions. In comparison to the case without packing material, the maintaining temperature of packing with SiO₂ beads and SiC blocks increased by about 9.1%. When the reactor was filled with SiC or SiO₂, the generated heat could be transferred to the packing medium, which had a higher specific heat capacity than air under normal temperature. Therefore, under the same discharge conditions, the temperature of the ED-DBD reactor containing SiO₂ and SiC-filled packing was higher than that of the ED-DBD reactor without packing material.

The conversion of benzene and CO₂ varied with discharge voltage given initial concentrations of 300 ppm for benzene and 3,000 ppm for CO₂, as illustrated in Figure 6. An increase in discharge voltage resulted in benzene conversion efficiencies ranging from 40% to 50% (Figure 6A). The highest conversion efficiencies were observed at 49.0% with SiC block filling, 47.8% with SiO₂ bead filling, and 45.2% without filling. The improved benzene conversion in the reactor containing packing materials could be attributed to the synergistic effects of surface discharge and prolonged residence time, which impeded the gas flow within the discharge region. The benzene degradation reached its maximum when the SiC blocks were utilized, due to the largest number of discharge current pulses. On the other hand, the maximum conversion of CO₂ exhibited a notable increase, reaching 72.7% with SiC blocks, 71.0% with SiO₂ beads, and 69.5% in the absence of any filing material (Figure 6B). The incorporation of packing materials within the reactor was advantageous for plasma generation, resulting in enhanced discharge intensity in the ED-DBD reactor and higher CO₂ conversion rates. Interestingly, the CO₂ conversion in the ED-DBD reactor filled with SiO₂ beads and SiC blocks exhibited minimal dependence on the driving voltage. This was likely because benzene conversion was primarily influenced by reactive species generated from CO₂ in the plasma, which were not directly associated with the applied voltages (Alliati et al., 2018).

To investigate the influence of CO₂ concentration on benzene reforming, the reaction system was subjected to CO₂ concentrations of 0.6% and 1.0%, alongside 300 ppm of benzene. The results indicated a slight increase in the maximum conversion of benzene, reaching 51.0% with 1.0% CO₂ and 49.7% with 0.6% CO₂, compared to 49.0% with 0.3% CO₂, while the maximum conversion of CO₂ presented a slight decline, decreasing from 72.7% at 0.3% CO₂ to 71.8% at 0.6% CO₂ and 71.7% at 1.0% CO₂ (Figure 7). The presence of benzene facilitated interactions between reactive species derived from CO₂ and benzene or its intermediates, resulting in the consumption of active species and enhancing the conversion of benzene and CO₂. However, CO₂ also competed with benzene for active particles (Xu et al., 2021; Zhu et al., 2020). When the concentration of CO₂ reached a sufficiently high level, this competitive interaction was intensified, thereby diminishing the facilitative effect of CO₂ on benzene degradation.

To further understand the impact of benzene concentration, we continuously introduced 600 ppm of benzene with 0.3% CO₂ into the reaction system. The performances of benzene and CO₂ conversion were depicted in Figure 8. The maximum conversion of benzene slightly decreased from 49.0% at 300 ppm benzene to 46.5% at 600 ppm benzene, while the maximum conversion of CO₂ showed a modest increase from 72.7% at 300 ppm benzene to 75.0% at 600 ppm benzene upon the adding more benzene to the reaction system. It suggested that the increased presence of hydrogen atom donors from benzene facilitated the conversion of CO₂, albeit at the expense of a reduced benzene conversion. Notably, CO₂ conversion exceeded 60% across all ED-DBD reactors with SiC blocks, indicating potential suitability for industrial applications (Olivier et al., 2023).

Further investigation was conducted to assess the yield of gas-phase inorganic products in the reaction system characterized by an initial concentration of 300 ppm benzene and 3,000 ppm CO₂, as shown in Figure 9. The analysis utilizing infrared (IR) spectroscopy and gas chromatography (GC) indicated that the primary products of benzene CO₂ reforming included CO, H₂, CH₄ and H₂O. The maximum CO yields obtaind in the ED-DBD reactor filled with SiC blocks recorded at 26.5%, representing 3% and 10% increase compared to the yields in the SiO₂-filled ED-DBD reactor and blank ED-DBD reactor, respectively. Similarly, the maximum H₂ yields in the ED-DBD reactor filled with SiC blocks were 19.2% and about 3% and 5% higher than the ED-DBD reactor with SiO₂ beads filled and the ED-DBD reactor, respectively. When the discharge power gradually increased from 60.0 W to 130 W in the ED-DBD reactor with SiO₂ beads filled, the CO yield increased slowly from 18.9% to 23.6% and the H₂ yield slightly increased from 10.0% to 17.8%. These findings suggested that enhancing discharge power might not be the primary factor in improving the yields of H₂ and CO, as evidenced by the nearly identical growth rates of CO and H₂.

Furthermore, the introduction of 0.6% and 1.0% CO2, along with 300 ppm benzene, into the reaction system resulted in the maximum CO yield of 29.4% at 1.0% CO₂, while a maximum H₂ yield of 19.1% at 0.6% CO₂, as shown in Figure 10. These results indicated that an increase in the initial concentration of CO₂ could substantially enhance the molar yields of CO and H₂. However, an excessive increase in the initial concentration of CO₂ might lead to a reduction in the molar yield of H₂. Additionally, the maximum molar ratio of H₂/CO decreased from 0.135 to 0.045 as the initial concentration of CO₂ increases from 0.3% to 1.0%, as shown in Figure 12A. It suggested that the initial concentration of CO₂ was excessively high, resulting in a more rapid growth rate of CO compared to H₂.

Figures 10B, D presented the yields of CO and H₂ within a reaction system containing 600 ppm benzene and 0.3% CO₂. In this system, the maximum yields of CO and H₂ were observed at 29.0% and 24.0%, respectively, suggesting that an increase in the initial concentration of benzene could substantially enhance the molar yields of both CO and H₂. Furthermore, as depicted in Figures 10B, D, the maximum molar ratio of H₂/CO increased from 0.135 to 0.226 as the initial benzene concentration raise from 300 ppm to 600 ppm. It was indicated that the generation rate of H₂ could surpass that of CO with the increased initial concentration of benzene.

Finally, the energy efficiency of 32%–75% was achieved in all the ED-DBD reactor system which filled SiC blocks (Figure 11), exceeding the thresholds of 60% for future development of CO₂ utilization (Olivier et al., 2023).

To verify the role of SiC, X-ray photoelectron spectroscopy (XPS) measurements were conducted to examinte the chemical states of Si and C in the SiC blocks, both prior to and following plasma exposure. Figures 12A, D showed the XPS spectra of Si 2p orbital before and after utilization, indicating the presence of Si-C bonds at approximately 99.5 eV. There was also a common Si-O peak at about 103.2 eV, indicating the presence of an oxide layer (SiO_x) on the SiC surface due to oxidation (Charpentier et al., 2012; Tamayo et al., 2018). Previous studies have documented a significant peak at about 105 eV for pre-use SiC samples, attributed to processes. (Hollering et al., 1999; Önneby and Pantano, 1997). However, this peak at 105 eV disappears after utilization, as the plasma bombardment facilitates the exposure of a new SiC layer. The C1s resulted in Figures 12B, E further confirm this, showing a significant increase in the proportion of -C bonds with plasma reaction. Combining with findings in Figure 6, it could be inferred that a small amount of C elements from the SiC blocks in the plasma region were involved in the reaction process on the SiC surface. Meanwhile, the O 1s XPS spectra depicted in Figures 12C, F indicated a substantial increase in the proportion of Si-O bonds following plasma treatment. The above observation suggested that Si within the packing materials played a crucial role in the CO₂ reforming of benzene. The detailed mechanisms were still under investigation.

To gain the insights into the discharge process in the SiC-packed ED-DBD reactor, optical emission spectroscopy was performed at a voltage amplitude of 12 kV and atmospheric pressure. As depicted in Figure 13, the emission spectral lines were mainly observed in two regions: 200–400 nm and 700–850 nm. The OH (A²∑⁺→X²∏) emission band (306.3–309.0 nm) was clearly identified, along with weaker CO₂ ⁺(A²→X²) emission bands at 325 nm, 337 nm, 355 nm, and 370 nm. Additionally, prominent Ar (2P→1S) emission lines were observed at wavelengths of 696.6 nm, 706.7 nm, 720.2nm, 738.4 nm, 751.2 nm, 763.3 nm, 772.7 nm, 794.9 nm, 811.2 nm and 842.6 nm.

By utilizing emission spectroscopy, the relative intensities of the emission lines emitted by the active OH species were evaluated with applied voltages (Figure 14). The results indicated that the intensities of OH emission lines in the SiC and SiO₂ packed ED-DBD reactor were higher than those in the blank ED-DBD reactor, with a slight increase in intensity as the applied power increased from 60 W to 120 W. Furthermore, the OH emission intensity in the SiC-packed ED-DBD reactor was higher than in the SiO₂ beads filled reactor. This might serve as indirect evidence suggesting that the SiC-packed ED-DBD reactor performed better in the CO₂ reforming of benzene compared to the SiO₂ beads.