Non-Thermal Plasma-Assisted Catalytic Dry Reforming of Methane and Carbon Dioxide Over G-C₃N₄-Based Catalyst

Abstract

In the present study, pure plasma and plasma-assisted catalytic activation of reforming of methane and carbon dioxide into syngas production were performed in a coaxial dielectric barrier discharge (DBD) plasma reactor at low temperature. For pure plasma, higher input power was favorable for the conversions of CH₄ and CO₂ but led to lower selectivities of syngas, while selectivities of syngas increased under lower flow rate of feed gases. A high CH₄/CO₂ ratio was more selective for C₂ hydrocarbon formation. Novel g-C₃N₄-based catalysts were prepared and filled in the discharge gap at low temperature. The cooperation effect of DBD plasma with g-C₃N₄ catalyst was achieved that contributed more to the activation of carbon dioxide than that of methane as well as the yield of CO for its low reduction potential. TiO₂/g-C₃N₄ with heterogeneous structure as packing the catalyst bulks into the discharge zone was capable of further improving the reforming of methane and carbon dioxide. TiO₂ mass ratio exhibited significant effect on dry reforming reactions and only TiO₂/g-C₃N₄ with 1% TiO₂ mass ratio presented enhanced hybrid effect than that of g-C₃N₄ catalyst.

1 Introduction

Greenhouse gases, such as carbon dioxide (CO₂) and methane (CH₄), have been widely investigated through dry reforming of CH₄ and CO₂, producing syngas (CO and H₂) for source utilization and environmental protection [1–3]. The reforming of CO₂ and CH₄ has been reported via the following reaction [4–6].

$${\text{C}}{{\text{H}}_4}+{\text{C}}{{\text{O}}_2} \to 2{{\text{H}}_2}+2{\text{CO,}}\;{{\text{G}}_{298{\text{K}}}}=69\;{{{\text{kJ}}} \mathord{\left/ {\vphantom {{{\text{kJ}}} {{\text{mol}}}}} \right. \kern-\nulldelimiterspace} {{\text{mol}}}}$$

However, it is a thermodynamically unfavorable reaction that high operating temperature is required for the production of syngas from methane and carbon dioxide with low H₂/CO molar ratio. The traditional catalytic processes have been widely studied for the reforming of methane with carbon dioxide, while the problems of catalytic deactivation causing by coke deposition and active metals sintering at high operating temperature (above 937 K) usually restrain catalytic technology application [7]. Plasma technology can provide a large number of reactive species generated in the plasma field, including electrons, radicals, etc., which help to make the dry reforming reaction of methane. Non-thermal plasma has attracted increasing attention in recent years because of its low energy requirement [8–13]. Under non-thermal plasma, the reactant molecules such as CH₄ and CO₂ are readily captured by the high-energy electrons to give excited species and free radicals, which facilitate for further chemical reactions. A series of ionization, dissociation and excitation chemical process are initiated and the temperature of the reactants remains only at several 100 K [8].

Dielectric barrier discharge (DBD) plasma, as typical non-thermal plasma, possesses advantages of high electron temperature (1–10 eV) and electron density (10¹⁸–10²¹ m⁻³) [14–18], and has been widely investigated to be effective for the dry reforming reaction at low temperature [19–23]. On the other hand, the dry reforming of methane under DBD plasma alone usually suffers from low conversion of reactants, poor selectivity of desired products and undesirable energy efficiency [12, 24–26]. The combination of a heterogeneous catalyst with DBD plasma has been reported to exhibit synergetic effect on improving the reaction performance both in conversion and selectivity [27–30]. Ni-based catalysts, Ni-based bimetallic oxides and La₂O₃/γ-Al₂O₃ catalyst have been extensively applied as catalysts in DBD reactor for the reforming of methane under atmosphere pressure [18, 29–34]. The catalyst placing in the plasma zone can affect discharge mode of DBD plasma that transfers from micro discharge to the combination of surface discharge and micro discharge [35, 36], and further the discharge stability can be destroyed in some cases [22, 30, 37]. The activation of catalyst in the discharge is found to be after the activation of the reactants [38]. However, effect of coke formation on catalysts in discharge zone is a problem facing by combination of DBD plasma and catalyst that leads to deactivation of the catalysts, and moreover the sintering of active metals is inevitable using metal-based catalyst, both of which cause reduction of carbon balance of plasma-catalysis much lower than 100% [25, 35, 39]. Furthermore, some studies have reported that combination of plasma with catalyst doesn’t play a positive role in the reforming reaction of methane [14, 40]. Therefore, it is necessary to develop novel catalyst with improving catalytic performance in discharge zone for enhancing the dry reforming of methane.

Graphitic carbon nitride (g-C₃N₄) was firstly reported as a novel stable photocatalyst by Wang et al. in 2009 because of its metal-free composition, non-toxicity and long-term stability [41], which is composed of carbon and nitrogen elements, possessing layered structure. Because of the band gap of approximately 2.7 eV [8, 42], g-C₃N₄ has been found to possess superior properties and exhibit efficient photocatalytic performance under visible light. Goettmann et al. reported that CO₂ could be activated by g-C₃N₄ [43] and Dong et al. also found that CO₂ could be reduced into CO [44]. Accordingly, g-C₃N₄ exhibits good application prospect in catalytic reduction of CO₂. Nevertheless, the catalytic efficiency of pure g-C₃N₄ is far from satisfaction due to its small surface area and short lifetimes of generated electrons and holes [45]. Coupling g-C₃N₄ and TiO₂ materials (g-C₃N₄ /TiO₂) is a promising strategy to improve the catalytic activity because of matched band gaps between g-C₃N₄ and TiO₂, which facilitates the separation of generated electrons from valence band (VB) of g-C₃N₄ (−1.12 eV vs. NHE) to conduction band (CB) of TiO₂ (−0.29 eV vs. NHE), as the VB edge potential of g-C₃N₄ is more negative than that of TiO₂. Under the powerful driving force causing by the differences in the CB edge potentials between g-C₃N₄ and TiO₂, the generated electron–hole pairs on the g-C₃N₄ surface are efficiently separated. TiO₂ acts as the acceptors of the generated electrons due to the matched band gaps between g-C₃N₄ and TiO₂. In DBD discharge zone, the high electron temperature is satisfactory to activate the g-C₃N₄/TiO₂ catalyst and it will be of important significance to exploring the feasibility of applying this novel catalyst for reforming of methane and carbon dioxide.

In this paper, CO₂ reforming of CH₄ to syngas by combination of DBD plasma with g-C₃N₄-based catalyst was studied under ambient conditions without extra heat to mitigate greenhouse emissions and reuse the carbon-hydrogen source. G-C₃N₄-based catalyst was firstly used as catalyst to convert methane and carbon dioxide because of its metal-free composition and good catalytic activity. The experimental conditions in the performed experiment and the effect of catalyst on the CO₂ reforming of CH₄ for syngas production are well investigated.

2 Experimental

2.1 Experimental System

The experimental system is shown in Fig. 1a. The non-thermal plasma-assisted catalytic experiments are performed in a coaxial DBD reactor that consists of a quartz tube with an inner diameter and thickness of 12 and 1.5 mm, respectively, as shown in Fig. 1b. A stainless tube with an inner diameter of 8 mm on the inside of the tube serves as high voltage electrode and the outer electrode is aluminum foil wrapping around the outer tube. The discharge length is 13 cm. DBD plasma is generated by an AC high-voltage generator (CTP-2000K) with a variable frequency of 4–12 kHz. A gas mixture of methane and carbon dioxide, controlled by mass flow controllers (MFC), flows to the DBD reactor. After reaction, on-line detection of the reaction mixture is performed on a gas chromatography (Shanghai Tianmei, GC7900) equipped with a thermal conductive detector (TCD) and a flame ionization detector (FID).

Fig. 1

Experimental system (a) and schematic diagram of reactor set-up (b)

2.2 Preparation of G-C₃N₄-Based Catalysts

G-C₃N₄ was prepared by heating melamine powder (heating rate 5 ^oC/min) at 500 °C in a muffle furnace for 2 h and then heated at 520 °C for 2 h. After naturally cooling to room temperature, the yellow product was collected and milled into powder. TiO₂/g-C₃N₄ catalyst was prepared by mixing a certain amount of TiO₂ (P25) and melamine powder in absolute ethyl alcohol, stirring for 3 h, treating by ultrasound for 1 h, dried for 8 h, and finally heated at the same conditions of preparing g-C₃N₄. The product was milled into small pieces with average diameter of approximately 0.4–1.0 mm. TiO₂/g-C₃N₄ with different contents of TiO₂ was obtained. In the plasma-assisted catalytic experiments, the catalyst packed in the plasma zone was shown in Fig. 1b.

2.3 Characterization of Catalysts

The crystallinity characteristic of catalyst was determined using a X-ray diffractometer (Shimadzu, XD-3A) with Cu kα radiation at a scanning speed of 5 ^oC/min. The morphology of catalyst was observed on a NOVA Nano SEM 450 field-emission scanning electron microscope (FESEM). UV–Vis diffuse reflection spectra (UV-DRS) were recorded on UV–Vis spectrometer (UV-2450, Shimadzu) in the range of 200–800 nm using BaSO₄ as the reflectance standard. The X-ray photoelectron spectroscopy (XPS) measurement was performed using ESCALAB™ 250Xi XPS spectrometer equipped with Al Kα radiation in the fixed analyzer transmission mode. All peaks in XPS spectra have already been calibrated with C 1s peak at 284.6 eV. The Brunauer–Emmett–Teller (BET) of the catalyst was analyzed by nitrogen adsorption–desorption at 77 K in a Quantachrome AS-1 Fully automatic physical nitrogen adsorption instrument. The BET surface area was determined using adsorption data in a relative pressure (P/P₀) range of 0.05–0.3, and the Barrett–Joyner–Halenda (BJH) method was used to determine the pore-size distribution.

2.4 Analytical Method

In the present work, the reactants and products were analyzed by an on-line gas chromatography. The CH₄ and CO₂ conversions, products selectivities, yields of products, and carbon balance are defined as follows:

$${C_{{\text{C}}{{\text{O}}_2}}}\left( \% \right)=\frac{{{\text{ C}}{{\text{O}}_2}\;{\text{converted}}\;({\text{mol}}/{\text{s}})}}{{{\text{ C}}{{\text{O}}_2}\;{\text{input}}\;({\text{mol}}/{\text{s}})}} \times 100$$

$${C_{{\text{C}}{{\text{H}}_{\text{4}}}}}(\% )=\frac{{{\text{C}}{{\text{H}}_{\text{4}}}\;{\text{converted}}\;{\text{(mol/s)}}}}{{{\text{C}}{{\text{H}}_{\text{4}}}\;{\text{input}}\;{\text{(mol/s)}}}} \times 100$$

$${S_{{\text{CO}}}}(\% )=\frac{{{\text{CO}}\;{\text{produced}}\;({\text{mol}}/{\text{s}})}}{{{\text{C}}{{\text{O}}_2}\;{\text{converted}}\;({\text{mol}}/{\text{s}})+{\text{C}}{{\text{H}}_4}\;{\text{converted}}\;({\text{mol}}/{\text{s}})}} \times 100$$

$${S_{{{\text{H}}_{2{\text{ }}}}}}(\% )=\frac{{{{\text{H}}_2}\;{\text{produced}}\;({\text{mol}}/{\text{s}})}}{{2 \times {\text{C}}{{\text{H}}_4}\;{\text{converted}}\;({\text{mol}}/{\text{s}})}} \times 100$$

$${S_{{{\text{C}}_2}{{\text{H}}_6}}}(\% )=\frac{{{{\text{C}}_2}{{\text{H}}_6}\;{\text{produced}} \times 2\;({\text{mol}}/{\text{s}})}}{{{\text{C}}{{\text{O}}_2}\;{\text{converted}}\;({\text{mol}}/{\text{s}})+{\text{C}}{{\text{H}}_4}\;{\text{converted}}\;({\text{mol}}/{\text{s}}){\text{ }}}} \times 100$$

$${Y_{{\text{CO}}}}(\% )=\frac{{{\text{CO}}\;{\text{produced}}\;({\text{mol}}/{\text{s}})}}{{{\text{C}}{{\text{O}}_2}\;{\text{input}}\;({\text{mol}}/{\text{s}})+{\text{C}}{{\text{H}}_4}\;{\text{input}}\;({\text{mol}}/{\text{s}})}} \times 100$$

$${Y_{{{\text{H}}_2}}} {(\%)}=\frac{{{{\text{H}}_2}\;{\text{produced}}\;({\text{mol}}/{\text{s}})}}{{2 \times {\text{C}}{{\text{H}}_4}\;{\text{input}}\;({\text{mol}}/{\text{s}}){\text{ }}}} \times 100$$

$$Y_{\!{{\text{C}}_2}{{\text{H}}_6}}\!(\% )=\frac{{{{\text{C}}_2}{{\text{H}}_6}\;{\text{produced}}\;({\text{mol}}/{\text{s}}) \times 2}}{{{\text{C}}{{\text{O}}_2}\;{\text{input}}\;({\text{mol}}/{\text{s}})+{\text{C}}{{\text{H}}_4}\;{\text{input}}\;({\text{mol}}/{\text{s}}){\text{ }}}} \times 100$$

$${S_{{\text{C}}{{\text{H}}_4}\;{\text{to}}\;{\text{C}}{{\text{O}}_2}}}\left( \% \right)=\frac{{{\text{CO}}\;{\text{produced}}\;({\text{mol}}/{\text{s}}){\text{ }} - {\text{C}}{{\text{O}}_2}\;{\text{converted}}\;({\text{mol}}/{\text{s}})}}{{{\text{C}}{{\text{H}}_4}\;{\text{converted}}\;({\text{mol}}/{\text{s}}){\text{ }}}} \times 100$$

$${B_{{\text{carbon}}}}\;\left( \% \right)=\frac{{{\text{CO}}\;{\text{produced}}\;({\text{mol}}/{\text{s}})+{\text{C}}{{\text{O}}_2}\;{\text{produced}}\;({\text{mol}}/{\text{s}})+{{\text{C}}_{\text{x}}}{{\text{H}}_{\text{y}}}\;{\text{produced}}\;({\text{mol}}/{\text{s}}) \times x}}{{{\text{CO}}\;{\text{input}}\;({\text{mol}}/{\text{s}})+{\text{CH}}\;{\text{input}}\;({\text{mol}}/{\text{s}}){\text{ }}}} \times 100$$

Specific energy density (SED) and energy utilization efficiency (E) are defined as:

$$SED\;\left( {{\text{kJ}}/{\text{L}}} \right)=\frac{{{\text{Discharge}}\;{\text{power}}\;({\text{kW}})}}{{{\text{C}}{{\text{O}}_2}\;{\text{flow}}\;{\text{rate}}\;({\text{L}}/{{\text{s}}})+{\text{C}}{{\text{H}}_4}\;{\text{flow}}\;{\text{rate}}\;({\text{L}}/{{\text{s}}})}}$$

$$E\left( {{\text{mol}}/{\text{kJ}}} \right)=\frac{{{\text{C}}{{\text{O}}_2}\;{\text{converted}}\;({\text{mol}}/{\text{s}})+{\text{C}}{{\text{H}}_4}\;{\text{converted}}\;({\text{mol}}/{\text{s}}){\text{ }}}}{{{\text{Power}}\;{\text{input}}\;({\text{kW}})}}$$

3 Results and Discussion

3.1 Characterization of Catalysts

The phase structures of the prepared catalysts are determined by XRD analysis. XRD patterns of pure g-C₃N₄, and TiO₂/ g-C₃N₄ prepared at different mass rations of g-C₃N₄ and TiO₂ are illustrated in Fig. 2. The pure g-C₃N₄ shows two characteristic diffraction peaks at 13.1° and 27.4° (JCPDS No. 87-1526). The minor peak of (100) at 13.1° and the stronger peak of (002) at 27.4° correspond to the inter-layer structural packing [46] and the interplanar graphitic stacking of aromatic system [47], respectively. The phases of TiO₂ are observed in the TiO₂/g-C₃N₄ samples. With the increase of the mass ratio of TiO₂ in the TiO₂/g-C₃N₄, the intensity of the diffraction peak of (101) at 25.2°, assigning to anatase phase (JCPDS No. 21-1272) of TiO₂, becomes stronger, while the diffraction peak intensity at 27.6° of g-C₃N₄ becomes weaker. In addition, the peak positions for all TiO₂/g-C₃N₄ samples are observed to keep unchanged, indicating that TiO₂ probably deposits on the surface of g-C₃N₄.

Fig. 2

XRD patterns of pure g-C₃N₄ and TiO₂/ g-C₃N₄

SEM imaging is performed to observe the morphologies of g-C₃N₄ and TiO₂/g-C₃N₄. From the point of the entire field of vision (Fig. 3a), it fills with large number of continuous lamellar g-C₃N₄ with 2D lamellar structure. As shown in Fig. 3b, after mixing TiO₂ (1%) with g-C₃N₄, TiO₂ is loaded on the surface of g-C₃N₄, which is benefited for the formation of heterojunction between g-C₃N₄ and TiO₂. The surface of TiO₂ is rough and porous, possibly being caused by the agglomeration of TiO₂ to some extent due to the calcination treatment.

Fig. 3

SEM images of a g-C₃N₄ and b 1% TiO₂/g-C₃N₄

The XPS survey spectrum of TiO₂/g-C₃N₄ sample is shown in Fig. 4a, in which C 1s, N 1s, Ti 2p and O 1s can be observed. The weak peaks of Ti and O are the results of low amount of TiO₂ in the TiO₂/g-C₃N₄ composite. Figure 4b shows the spectrum of C 1s for TiO₂/g-C₃N₄ sample. The peaks at about 284.96, 288.41 and 293.89 eV correspond to C–C, C–N–C and C–(N)₃ groups in the g-C₃N₄, respectively [48]. Figure 4c shows the spectrum of N 1s for TiO₂/g-C₃N₄ that displays four peaks, ascribable to sp₂ hybridized nitrogen (C=N–C) at about 398.8 eV, tertiary N in N–(C)₃ at about 399.65 eV, N atoms in amino moieties at about 401.35 eV [49], and weak peak located at 404.2 eV which is due to the charging effects or positive charge localization in the heterocycles [50]. Figure 4d presents the Ti 2p spectra of TiO₂/g-C₃N₄ and pure TiO₂. Two peaks for Ti 2p_3/2 and Ti 2p_1/2 of pure TiO₂ located at 459.1 and 464.8 eV, and the peaks at 458.7 and 464.0 eV correspond to Ti 2p_3/2 and Ti 2p_1/2 of TiO₂/g-C₃N₄, respectively. Comparison of O 1s spectra of TiO₂/g-C₃N₄ and pure TiO₂ is shown in Fig. 4e. The blinding energies of 530.6 eV and 532.7 eV for pure TiO₂ can be ascribed to Ti-O bond and O-H bond, respectively [48]. Compared with pure TiO₂, O 1s spectrum of TiO₂/g-C₃N₄ shows the negative shift with an order of 0.8 eV, and the same shift of Ti 2p in Fig. 4b is also observed, which indicates that chemically bound interfaces may be formed between g-C₃N₄ and TiO₂ phase.

Fig. 4

XPS of a survey spectrum, b C 1s and c N 1s for 1% TiO₂/g-C₃N₄, d Ti 2p and e O 1s for 1% TiO₂/g-C₃N₄ and pure TiO₂

The nitrogen adsorption–desorption isotherms of g-C₃N₄ and 1% TiO₂/g-C₃N₄ were present in Fig. 5. It can be seen that both samples have the isotherms type of H4 according to the IUPAC (International Union of Pure and Applied Chemistry) classification, which indicates irregular pore structure and micropore and mesopore both exist in the samples [51]. The graph insert in Fig. 5 shows the pore-size distributions of g-C₃N₄ and 1% TiO₂/g-C₃N₄, which are broad and the samples are mainly composed by mesopore. Specific surface area (S_BET) of the as-prepared samples with different amount of TiO₂ in g-C₃N₄ was shown in Table 1. The S_BET increases with the increasing amount of TiO₂ except 20%TiO₂/g-C₃N₄, which is possibly attributed to the sintering caused by calcine process.

Fig. 5

Nitrogen adsorption–desorption isotherms and corresponding pore size distribution (inset) of the g-C₃N₄ and 1% TiO₂/g-C₃N₄

Table 1 Specific surface area of the as-prepared samples with the different amount of TiO₂ in TiO₂/g-C₃N₄ composite

The optical absorption properties of TiO₂, g-C₃N₄ and as-prepared TiO₂/g-C₃N₄ samples are investigated by UV-DRS analysis, as shown in Fig. 6. The absorption edge for TiO₂ is under 400 nm, only having a response to UV light. The g-C₃N₄ exhibits a wider photo-absorption till visible light, the main absorption edge of which occurs at approximately 500 nm due to its wide band gap of 2.57 eV. For TiO₂/g-C₃N₄ samples, a certain extent red-shift of the absorption edge to visible region and a certain extent blue-shift of the absorption edge to UV region are observed as compared to pure TiO₂ and pure g-C₃N₄, respectively, indicating the band gap of g-C₃N₄ becomes narrower when combing it with TiO₂. In the plasma zone, the catalyst is expected to be excited by the active species in the plasma with energy higher than 2.7 eV, exhibiting catalytic performance on reforming of methane efficiently.

Fig. 6

UV–vis absorption spectra of pure g-C₃N₄, pure TiO₂, and TiO₂/g-C₃N₄ composite with different ratios

3.2 Effect of Discharge Power on the Plasma Dry Reforming Reaction

Under non-thermal plasma, the discharge power determines the strength of the internal electric field and the electron energy, which is important for inelastic collisions between energetic electrons and reactant molecules. In the present work, the conversions of CO₂ and CH₄ were found to increase from 6% and 11% to 18% and 34%, respectively, with the discharge power increasing from 20 to 60 W. High level of discharge power provides more energy for the formation of active species, such as electrons, ^·OH, and O^·, which are likely to attack CH₄ and CO₂, leading to enhanced conversions of reactants. It is also found that the conversion of CH₄ is always higher than that of CO₂ as shown in Fig. 7a. Similar result is also reported in [30]. With the increase of discharge power, the balance between the dehydrogenation rates of CH_x (x = 1–4) radicals and the dissociation rate of active CO₂ species can be deteriorated causing by the excess energy. The result of the breaking of the balance favors for the reaction between H and O but is disadvantageous for the generation of CO from C and O atoms [52]. The selectivities of CO and H₂ decreased with the increase of discharge power (Fig. 7b), indicating the coke formation at higher energy level, the results of which consist with the conversions of CO₂ and CH₄. Gaseous hydrocarbon, C₂H₆, existing in the products, is possibly formed by the complex reaction between two CH₃ radicals. In Fig. 5b, the selectivity of C₂H₆ hardly changed with the increasing discharge power, suggesting that C₂H₆ formation was not affected by the discharge power. The carbon balance was found to decrease with the increase of discharge power (Fig. 7c), further supporting the facts of coke formation at higher energy level. As shown in Fig. 7d, discharge power exhibited less effect on the H₂/CO ratio, which indicated the consistent changes of H₂ formation and CO formation was not affected with the increase of discharge power.

Fig. 7

Effect of discharge power on the performance of dry reforming reaction: a conversion of reactants; b selectivity of main products; c carbon balance; d ratio of H₂/CO (CH₄/CO₂ = 1:1, feed flow rate 50 mL/min)

3.3 Effect of Total Flow Rate on the Plasma Dry Reforming Reaction

The total flow rate of feed determines the residence time of the reactants in the discharge zone and its effect on the plasma dry reforming reaction is shown in Fig. 8. For a constant input power, the conversions of CH₄ and CO₂ decreased from 35.3% and 19.7% to 14.4% and 7.5% (Fig. 8a), respectively, and the yields of CO and H₂ both decreased by approximately 56% (Fig. 8b), but the selectivities of CO and H₂ changed slightly (Fig. 8c), with the increase of total flow rate from 35 to 100 mL/min. The total flow rate also influenced the formation of the hydrocarbon byproduct, C₂H₆, the yield of which decreased by 42%, while the selectivity of it increased obviously. It is apparent that increasing the total flow rate benefits the conversion of CH₄ to hydrocarbon byproduct. As shown in Fig. 6d, the carbon balance as well as conversion of CH₄ to CO increased with the total flow rate, because increase of the flow rate would reduce the collision chance among CH_x active species, and thus more CH₄ was converted to low-carbon hydrocarbon compounds such as C₂H₆ and CO instead of long chain hydrocarbons. Energy efficiency is found to decrease when increasing the total flow rate, which is caused by the reduction of residence time of reactants and active species in plasma zone. On the consideration of the effect of the total flow rate on the dry reforming reaction, the optimal flow rate of 50 mL/min is selected in the following experiments.

Fig. 8

Effect of total flow rate on the performance of plasma dry reforming reaction: a conversion of reactants and energy efficiency; b yield of main products; c selectivity of main products; d carbon balance (CH₄/CO₂ = 1:1, discharge power 50 W)

3.4 Effect of CO₂/CH₄ Ratio on the Plasma Dry Reforming Reaction

Effect of reagent composition on the plasma dry reforming reaction is performed by varying the CO₂/CH₄ ratios from 6/1 to 1/6 at a constant total flow rate (50 mL/min) and a constant discharge power (50 W). The results are shown in Fig. 9. With the increase of molar ratio of CH₄ to CO₂, the conversion of CH₄ was always higher than that of CO₂. However, the CO₂/CH₄ ratio had greater influence on the conversion of CH₄ which decreased from 53% to 17% obviously but the conversion of CO₂ changed slightly. As shown in Fig. 9b, the yields of H₂ and CO both decreased with the increase of molar ratio of CH₄ to CO₂ while the yield of C₂H₆ increased. Figure 9c further exhibited the fact that increasing the amount of CH₄ in the feed gas would restrain its conversion to CO. These experimental results can be explained by the possible reactions involving active methane species, the amount of which increase with the CH₄ molar, and thus more active methane species involve in recombination reaction than methane oxidation to form CO, generating more hydrocarbons [18, 30]. In the present studied range of CO₂/CH₄ ratios, the carbon balance kept at approximately 95% while the ratio of H₂/CO increased significantly with the increasing CH₄/CO₂ ratio (Fig. 9d). This result further confirms that increasing CH₄/CO₂ ratio leads to less oxygen, favoring for the production of hydrocarbon accompanying by carbon deposition.

Fig. 9

Effect of CO₂/CH₄ ratio on the performance of dry reforming reaction: a conversion of reactants and energy efficiency; b carbon balance and the ratio of CH₄ convert to CO; c yield of main products; d ratio of H₂/CO (feed flow rate 50 m L/min, discharge power 50 W)

3.5 Plasma-Catalyst Association

Combination of non-thermal plasma with catalysts has been widely investigated in dry reforming reactions aiming at improving the conversion and selectivity. For non-thermal plasma, most of the electrical energies provided by the input power are consumed for energetic electrons production, which play key role in plasma-alone assisted dry reforming by way of reacting with reactants molecules to generate excited species. With the presence of catalyst bulks in the discharge gap, surface discharge may occur on the surface of the catalyst and the catalyst can trap energetic electrons in the active particles. Accordingly, effect of catalyst on the discharge process other than catalytic activity should be under consideration to clarify the plasma-catalyst association for dry reforming reaction.

In the present work, samples of g-C₃N₄ and γ-Al₂O₃ with the same particle size are placed in the discharge zone, respectively. The γ-Al₂O₃ has no catalytic activity and here helps to explore how the drying reforming reaction is influenced by the sample packing that might cause the discharge process change. Effect of input power on the conversions of reactants and the yields of CO and H₂ over the catalyst samples placed at the DBD reactor is shown in Fig. 10. As γ-Al₂O₃ was packed in the discharge zone, the conversion of CH₄ as well as yield of H₂ increased to some extent, while the conversion of CO₂ changed slightly and the increase in yield of CO was lower than that of H₂. The results exhibit the fact that the change of discharge mode causing by γ-Al₂O₃ packing would affect the dry reforming reaction. After γ-Al₂O₃ packing, the discharge mode transfers from filament discharge to the combination of filament discharge and surface discharge. Polarization field is formed on the surface of γ-Al₂O₃ particles, making the electric field intensity in local area strengthen. The reason of the enhancements of the conversion of CH₄ and yield of H₂ is due to lower dissociation energy of CH₄ than that of CO₂. The followings are the main reactions happened in the discharge zone.

Fig. 10

Effect of packing g-C₃N₄ catalyst and γ-Al₂O₃ in plasma discharge area on the performance of plasma dry reforming reactions: a conversion of CH₄; b conversion of CO₂; c yield of CO; d yield of H₂. (CH₄/CO₂ = 1:1, feed flow rate 50 mL/min)

$$e+{\text{C}}{{\text{O}}_2} \to {\text{CO}}+{\text{O}}+e$$

(1)

$$e+{\text{C}}{{\text{H}}_4} \to {\text{C}}{{\text{H}}_3}+{\text{H}}+e$$

(2)

$$e+{\text{C}}{{\text{H}}_3} \to {\text{C}}{{\text{H}}_2}+{\text{H}}+e$$

(3)

$$e+{\text{C}}{{\text{H}}_2} \to {\text{CH}}+{\text{H}}+e$$

(4)

$$e+{\text{CH}} \to {\text{C}}+{\text{H}}+e$$

(5)

$${\text{C}}+{\text{O}} \to {\text{CO}}$$

(6)

In case of g-C₃N₄, the conversions of CH₄ and CO₂ increased and the yields of H₂ and CO were all higher than those obtained in plasma-alone system. As a novel metal-free catalyst, it can be excited by DBD plasma because its narrow band gap (2.7 eV), and moreover, the generated electrons have a strong reducing ability because of extremely negative position (−1.35 vs. NHE) of the g-C₃N₄ conduction band, which is expected to favor the conversion of CO₂ in the dry reforming reaction. Furthermore, the results are also higher than those using γ-Al₂O₃ particles in plasma zone. The cooperation effect of DBD plasma with g-C₃N₄ catalyst is verified. Compared with plasma-alone, stronger externally applied electric field as well as stronger micro- and surface discharges are formed as packing g-C₃N₄ catalyst into DBD plasma zone. More excited species are adsorbed on the surface of the catalyst and favors the conversion of the reactants comparing, achieving desired products [34].

The stability of g-C₃N₄ prepared in this study was tested under the conditions of discharge power 30 W, flow rate of feed gas 50 mL/min and molar ratio of CH₄/CO₂ 1:1. The product was detected every 15 min by the on-line detection and the results was shown in Fig. 11. After 120 min treatment, the as-prepared catalyst performed a favorable stability in the conversion of CH₄ and CO₂.The yield and selectivity of main products of H₂ and CO slightly decrease after 45 min reaction which may be attributed to the carbon formation during the collision of high energy electrons and CH₄.

Fig. 11

Stability of g-C₃N₄ catalyst with the increase of reaction time: a conversion, b yield and c selectivity

Pure g-C₃N₄ usually suffers from small surface area and short lifetimes of generated electrons and holes. Approach of constructing heterogeneous structure by coupling g-C₃N₄ and TiO₂ materials is a promising strategy to improve the charge separation efficiency. Accordingly, TiO₂/g-C₃N₄ catalysts with different TiO₂ mass ratio were prepared in the present work and the catalytic activity towards dry reforming reaction is explored. As shown in Fig. 12, the TiO₂ mass ratio exhibited significant effect on dry reforming reactions. Only TiO₂/g-C₃N₄ with 1% TiO₂ mass ratio presented a hybrid effect as packing the catalyst bulks into the discharge zone. The efficient separation of electrons and holes occurred when mixing TiO₂ (1%) with g-C₃N₄. However, further increasing mass ratio of TiO₂ restrained the conversions of feed gases and also the yields of CO and H₂. The reasons can be explained by the higher dispersion of TiO₂ on the surface of g-C₃N₄ with further increasing the mass ratio of TiO₂, which could adsorb more high energy electrons and thus reduce the reaction probability between high energy electrons and g-C₃N₄. In the present work, the electron energy might not be higher enough to excite TiO₂, whose band gap is 3.1 eV, and thus excitation of g-C₃N₄ is expected to play significant role in the plasma-assisted catalytic reaction and 1% of TiO₂ is believed to contribute to improve separation of electrons and holes generated on g-C₃N₄. Control experiment was further carried out with 1% TiO₂/γ-Al₂O₃, which prepared in same synthesis procedure of 1% TiO₂/g-C₃N₄, and the results are shown in Fig. 13. As the discharge area packing with 1% TiO₂/g-C₃N₄ and 1% TiO₂/γ-Al₂O₃ catalysts, respectively, with the increasing of discharge power, conversions of CO₂ and CH₄ increase while the selectivities of main products decrease, which is in accordance with aforementioned result. Comparing the catalytic activity of TiO₂/g-C₃N₄ with TiO₂/γ-Al₂O₃ combining with DBD plasma, TiO₂/g-C₃N₄ exhibits better performance in both CO₂ and CH₄ conversion and the selectivity of CO, while the selectivity of H₂ is slightly lower than that using TiO₂/γ-Al₂O₃ catalyst. The experimental result further proves the good catalytic activity of TiO₂/g-C₃N₄ composite in conversion of CO₂ and CH₄ and enhancing CO selectivity. However, the detailed mechanism of DBD plasma- g-C₃N₄-based catalysis has not been fully understood yet. The appropriate optimization of the DBD plasma-assisted g-C₃N₄-based catalytic system is needed to be further studied.

Fig. 12

Effect of TiO₂/g-C₃N₄ composite packing in discharge area with different TiO₂ mass ratios on the performance of dry reforming reaction: a, b conversions of reactants; c, d yield of products. (CH₄/CO₂ = 1:1, feed flow rate 50 mL/min)

Fig. 13

Conversions of CH₄ and CO₂, and selectivities of CO and H₂ as the DBD reactor packing with 1% TiO₂/g-C₃N₄ and 1% TiO₂/γ-Al₂O₃ catalysts, respectively

4 Conclusion

Dry reforming of methane and carbon dioxide under pure plasma and synthetic action of plasma-assisted g-C₃N₄-based catalysis was investigated in DBD reactor. In pure plasma process, input power, total flow rate of feed gas and CH₄/CO₂ ratio affected the reaction. High input power and low total flow rate favored for increasing the conversions of reactants. The selectivities of CO and H₂ decreased with the input power, indicating the coke formation at higher energy level, and only the ratio of CH₄/CO₂ had a significant influence on the products ratio of H₂/CO. Combining DBD plasma with g-C₃N₄ catalyst, the conversions of CH₄ and CO₂ increased and the yields of H₂ and CO were all higher than those obtained in plasma-alone system, which was attributed to the synergistic effect of plasma and g-C₃N₄ catalyst. In the DBD plasma assisted with 1% TiO₂/g-C₃N₄ composite system, the maximum conversion rate of reactants and yield of product were obtained. With further increasing the mass ratio of TiO₂, more high energy electrons could be adsorbed and thus reduced the reaction probability between high energy electrons and catalyst. The present study makes it clear that a synergy between g-C₃N₄-based catalyst and DBD plasma is achieved.