2.1. Experimental Method

In the plasma treatment of biomass, the input voltage of 225 V and the frequency of 10.5 kHz were used. The discharge power and output voltage under different reaction conditions is shown in Table 1. About 0.2 g of the sample was placed on top of the sieve plate, which was in the middle of the quartz reactor. Before the experiment, the carrier gas, including Ar, N₂, and CO₂, respectively, with a flow rate of 40 ml/min was used to purge the sample for 10 min to remove air in the reactor, and then the plasma power supply was turned on. During the experiment, the plasma reactor was kept at room temperature and atmosphere pressure. Gas production from plasma treated biomass was detected using an Agilent 7890B gas chromatograph equipped with two TCD detectors as well as an FID detector. As GC does not allow for real-time analysis of gas components, the exhaust gas was first collected in a gas bag, followed by a detailed analysis of its composition. Gas samples were collected and analyzed every 5 min during the first 30 min, and subsequently every 15 min after the experiment 60 min.

In order to further investigate the physical and chemical properties of the carbon materials after the reaction, a series of characterization methods were used. Scanning electron microscopy (SEM; ZEISS Sigma HD) was used to study the surface morphology and pore structure of the carbon materials produced after pyrolysis. Raman spectroscopy (LabRAM HR Evolution) was used to investigate the degree of graphitization and the degree of surface defects of carbon materials produced after pyrolysis. Raman spectra were used with an Ar laser, and the wave number ranging between 500 and 3,500 cm⁻¹.

2.2. Calculation Method

DFT is a quantum mechanical modeling method used to investigate the electronic structure (principally, the ground state and transition state) of many-body systems, particularly atoms and molecules. Quantum chemistry calculations using the DFT were performed using the Gaussian 09 software. All the atoms were described by B3LYP/6-31G (d) base set [20, 21, 22], the structures of reaction products, transition states (ts), intermediates, and products were optimized. The accuracy of each optimized structure is verified by frequency calculation, and no virtual frequency is determined. The TS method was employed to calculate the transition state in the pyrolysis reaction process to make sure that the transition state had the unique imaginary frequency [23, 24, 25]. The intrinsic reaction coordinates (IRC) were used to track the reaction path of the searched transition state, which ensures the rationality of the structural connection before and after the reaction process. By comparing the energy barrier to be overcome under different reactions, the path with the minimum energy barrier is the possible path.

3.1. Gasification Characterization

The gas production characteristics of millet and soybeans under different ambientes were studied. At the same time, the gasification characteristics of millet and soybean model compounds (glucose and aspartic acid) under the same experimental conditions were also studied. Figure 2 shows the gas distribution of millet and soybeans treated by plasma under different ambientes. The gas distribution of millet during plasma gasification under N₂, CO₂, and Ar ambientes, respectively, are shown in Figures 2(a), 2(b), and 2(c). It can be found that the concentration of each gas component showed a similar trend with time. With the progress of the reaction, the gas concentration showed a reaction trend of first increasing and then decreasing. Under N₂ and Ar ambient, when the experiment was carried out for about 15–20 min, the concentration of these four gases reached the peak and gradually decreased, and the largest gas production was H₂, followed by CO gas. The total concentration of CO₂ gas was not much different from that of CO gas, but the concentration of CH₄ was always lower than that of the other three gases. Since the carrier gas ambient is CO₂, the CO₂ concentration is not marked in the gas production distribution in Figure 2(b). Because CO₂ will decompose into CO under the action of plasma, CO content is the highest in the gas products of millet after plasma treatment under CO₂ ambient [26, 27]. For the gasification under N₂ and Ar ambientes, CO and CO₂ gases still existed in the gasification product although there was no oxidizing gas in the reaction ambient, which was caused by the reaction between the oxygen-containing functional group existing in the structure of millet itself and the O₂ adsorbed on the surface. The gas production of millet in Ar ambient was much smaller than that in N₂ ambient, because the metastable state energy of the Ar is lower than that of the N₂, less of the energetic metastable (or excited states) through the Ar are transferred to the sample than N₂ [28, 29, 30], which on the one hand hindered the propagation of high-energy electrons, and also reduced the interaction between high-energy electrons and plasma generated after carrier ionization. It is noted that the temperature of the reactor was measured using an infrared thermometer during the reaction, and the core temperature was about 250°C, which did not cause tar to be produced.

Figure 3 shows the gas production distribution of soybean gasification process under N₂, CO_2, and Ar ambient, respectively. The gas production of soybean under N₂ and CO₂ ambient was similar to that of millet. When soybean was gasification by plasma under N₂ and CO₂ ambient, the gas production rate of H₂ was still the highest, and the gas production rate of CH₄ was the lowest. After the reaction exceeded 50 min, the gas concentration of each component basically reached a stable state. Figure 3(c) shows the gas production distribution of soybeans during plasma treatment under Ar ambient. It can be found that the time when the gas concentration of each component reaches the highest value was about 47 min, which was later than that of soybean in the other two gases (N₂ and CO₂) and millet under the same ambient. On the one hand, due to the weak ionization energy of Ar itself and fewer high-energy electrons, the gasification rate is reduced compared with N₂ and CO₂, so the gas production rate is weakened, resulting in the peak gas production concentration extending to the middle and later stages of the reaction. On the other hand, it shows that different grain substances have different gas production rates under the same reaction conditions, which may be caused by the composition of the grain itself and the shape of the grain.

By analyzing the data in Table 1 and the gas production characteristics of biomass gasification above, it can be found that there is little difference in plasma discharge power under N₂ and CO₂ atmosphere, but the gas production concentration of millet and soybeans after plasma treatment under N₂ atmosphere is much higher than that under CO₂ atmosphere, indicating that plasma is not directly interacting with biomass and gasification. At the same time, when CO₂ is used as carrier gas, there is a large amount of CO in the tail gas, which indicates that the high-energy electrons produced by the plasma fully interact with the carrier gas and produce a large number of plasma gas species. It can be inferred that during biomass gasification under plasma influence, gases are transformed into plasma gas species through plasma action. Subsequently, these plasma gas species interact with the biomass to facilitate gasification. Du et al. [31] also found that after the introduction of water vapor in the experiment, under the action of high-energy electrons, the water vapor in the plasma discharge region will produce OH·, H·, O·, and other active free radicals.

In order to further explore the gas production mechanism of food waste under different ambientes by plasma action, glucose and aspartic acid were selected as the model compounds of food waste, and their gas production characteristics by plasma were studied in experiment, and their decomposition paths were explored by DFT simulation.

Figure 4 shows the gas distribution of glucose during plasma gasification under N₂, CO_2, and Ar ambientes, respectively. Figure 4(a) shows the gas production of glucose during gasification under N₂ ambient. It can be found that the peak gas production of glucose during plasma treatment gasification under N₂ ambient occurred between 20 and 25 min, and the overall gas production concentration was in the order of H₂ > CO > CO₂ > CH₄ from high to low. The time of the maximum gas production was later than that of millet and soybean, but the overall gas production law was similar to that of millet and soybean. Figure 4(b) shows the gas distribution of glucose during plasma treatment under CO₂ ambient. Under the action of CO₂ ambient, when glucose is treated by plasma, the maximum concentration of CO output can reach 10%, and part of CO comes from the ionization of CO₂. The maximum concentration of hydrogen production can reach 1.0%. Figure 4(c) shows the distribution of gas production during plasma treatment of glucose under Ar ambient. The peak time of gas production concentration of glucose under Ar ambient is earlier than that of N₂ and CO₂, but the gas production is significantly lower than that of the other two ambientes, and the output of H₂ is only 1.3%.

Figure 5 shows the gas production distribution of aspartic acid in N₂, CO₂, and Ar ambientes under plasma treatment gasification process. It can be found that the gas concentration of aspartic acid produced by plasma treatment under the three ambientes is lower than that of glucose. Figure 5(a) shows the gas production of aspartic acid during gasification under N₂ ambient. It can be found that compared with glucose, the gas production time of aspartic acid is longer, and H₂ gas production starts after plasma treatment and lasts for about 80 min. The gas production of CO, CO₂, and CH₄ lasted for about 30 min after the reaction. Figure 5(b) shows the gas distribution of aspartic acid during plasma treatment under CO₂ ambient. Aspartic acid gasification basically only produces CO, H₂, and CH₄ concentrations less than 0.25%. Figure 5(c) shows the gas production distribution of aspartate in Ar ambient by plasma treatment. Similar to the gas production rule under the other two ambientes, the gas production time is also extended, and the gasification process basically ends after about 80 min of reaction. The gas production law of the two model compounds is basically consistent with that of millet and soybean. Therefore, the gas production mechanism of millet and soybean under the action of plasma can be explored by studying the gas production mechanism of the model compounds.

3.2. Mechanism of Gas Production

In order to further understand the mechanism of gasification of kitchen waste model compounds, DFT simulation was used to study the decomposition pathways of glucose and aspartic acid. According to where the glucose molecule breaks the bond, three different pathways (G-P1, G-P2, and G-P3) are considered. Figure 6 shows the reaction path and Gibbs free energy of glucose decomposing. The molecular structure formula corresponding to the glucose decomposition path is shown in Figure 7. For G-P1 pathway shown in Figure 6(a), the glucose molecule with ring structure first underwent the ring-opening reaction to produce glucose with chain structure (G-P1-IM2). This process needs to cross an energy barrier of 182.7 kJ/mol and absorb 18.6 kJ/mol of energy in the reaction. G-P1-IM2→G-P1-IM3 had C–C bond breaking and hydrogen transfer. The process of G-P1-IM3→G-P1-IM4 was elimination reaction, and H₂O was formed at the same time. This process needs to cross an energy barrier of 280.6 kJ/mol and absorb 725 kJ/mol of energy in the reaction. G-P1-IM4 has an enol-like structure, which is extremely unstable and will produce products with a stable ketone structure. Then, it underwent decarbonylation and elimination to form CO and H₂O. It can be found that the speed control step of the G-P1 path is the breaking of C─C bond and the transition of H (G-P1-IM2→G-P1-IM3).

As shown in Figure 6(b), in the G-P2 pathway, the glucose molecule underwent ring-opening reaction, elimination reaction, hydrogen transfer reaction, C─C bond breaking, and other processes to finally decompose. After the ring-opening reaction, the G-P2-IM3→G-P2-IM4 process was the elimination reaction, which removes one molecule of water from glucose. The G-P2-IM4→G-P2-IM5 process was a hydrogen transfer reaction, and the energy barrier that need to be overcome for the G-P2-IM4→G-P2-IM5 process was 235.0 kJ/mol. Finally, C─C bond breakage and hydrogen transfer occurred, and the decomposition of glucose molecules was completed.

Figure 6(c) depicts the reaction path and Gibbs free energy of G-P3, the glucose molecule first underwent the ring-opening reaction, and then decarburization reaction formed G-P3-IM3 and formaldehyde. The activation energy of this process is 362.2 kJ/mol. G-P3-IM3 underwent elimination again and produced H₂O. G-P3-IM4 was decomposed by carbon─carbon bond breaking, decarbonization reaction, hydrogen transfer, and so on. Finally, the glucose molecule was broken down. By comparing the rate-controlling steps of the three paths, it can be found that the G-P3 path is the most likely path, and the activation energy of G-P3 to be overcome in the rate-controlling step is 363.4 kJ/mol. The decomposition process of glucose is shown in Figure 7, mainly includes four parts, which are ring opening reaction, decarbonylation reaction, dehydration reaction, and carbon─carbon bond breaking. And the rate-controlling step is the formation of H₂O.

The reaction pathway and Gibbs free energy of aspartic acid decomposing are shown in Figure 8. Also, three different pathways (A-P1, A-P2 and A-P3) of aspartic acid decomposing are considered. The molecular structure formula corresponding to the aspartic acid decomposition path is shown in Figure 9. Figure 8(a) shows the reaction path and Gibbs free energy of A-P1 path, aspartic acid was decarbonized to produce A-P1-IM2 and CO₂, which released 50.1 kJ/mol of energy. A-P1-IM2 was decomposed into A-P1-IM3 and NH₃, and the activation energy of this process was 208.7 kJ/mol. A-P1-P and CO₂ are the products of the decarbonylation of A-P1-IM, which overcomes the energy barrier of 268.3 kJ/mol and releases 34.7 kJ/mol of energy. It can be found that the speed control step of A-P1 path is A-P1-IM1→A-P1-IM2+CO₂, and the energy barrier overcome by this process is 308.2 kJ/mol.

For A-P2 shown in Figure 8(b), A-P2-IM1 is first decomposed into ammonia and A-P3-IM2, which requires an energy barrier of 199.1 kJ/mol to overcome, and 26.7 kJ/mol is absorbed in the reaction. The decarbonylation of A-P2-IM2 could form A-P2-IM3 and CO₂, which the activation energy is 289.6 kJ/mol. Finally, A-P3-IM3 is decarbonized to produce CO₂ and ethylene with an energy barrier of 268.2 kJ/mol. It can be found that the speed control step of the P2 path is A-P2-IM2→A-P2-IM3 process, and the activation energy of the reaction is 289.6 kJ/mol.

In the Figure 8(c), A-P3-IM1 first underwent decarbonylation to form A-P3-IM2. The energy barrier of this process is 296.4 kJ/mol, and the energy released by the reaction is 64.9 kJ/mol. The decomposition of A-P3-IM2 to form A-P3-IM3 and CO₂, which is the rate-controlled step of the reaction, and the activation energy is 306.3 kJ/mol. At least, the A-P3-IM3→A-P4-P reaction could form NH₃ and ethylene with the energy barrier value of 294.3 kJ/mol. By comparing the activation energy of three different path control steps, it is found that A-P2 path is the best path for the decomposition of aspartic acid. The decomposition process of aspartic acid mainly includes decarbonylation and deammonia reaction. And the decarbonylation process controlled the decompose of aspartic acid.

3.3. Characterization of Gasification Solid Product

In order to use the solid products produced by gasification with DBD treatment efficiently, the physicochemical properties of the solid product carbon material after kitchen waste gasification were also studied by some characterization methods. Figure 10 shows the Raman spectra of carbon materials of millet and soybean treated by plasma under different reaction ambientes, and Table 2 presents the value of I_D/I_G of the carbon materials prepared by millet and soybean. The Raman spectra of the carbon material prepared by millet is shown in Figure 10(a). It can be found that the typical G and D peaks of graphite materials appear at 1,582 cm⁻¹ and 1,350 cm⁻¹, respectively, indicating that biomass materials are transformed into carbon materials after plasma treatment under N₂ and CO₂ ambientes [32, 33]. Among them, peak D represents structural defects or impurities present in the carbon material. Peak G represents the lattice vibration or pure isomer of the carbon material, resulting from the surface vibration of the sp² carbon atom [34]. From Table 2, the graphitization degree of the carbon material produced by the plasma treatment of millet in N₂ ambient is smaller than that of the carbon material produced in CO₂ ambient. It can be seen that the biomass carbon material treated in the CO₂ ambient has a higher degree of chaos, which is caused by the continuous reaction of the carbon material generated in the reaction process with the CO₂ ambient and the active oxygen generated by ionization.

The Raman spectra of carbon materials of soybean treated by plasma under different reaction ambientes are shown in Figure 10(b). It can be found that there are still two characteristic peaks, peak D and peak G, but the intensity of peak D was much higher than that of peak G. The results indicated that the graphitization degree of soybean after plasma treatment was higher than that of millet [35, 36]. Compared with the carbon materials produced by plasma treatment of soybean in N₂ ambient and CO₂ ambient, the graphitization degree of the samples of millet and soybean in N₂ ambient is greater than that of the samples in CO₂ ambient. However, although the carbon material produced by plasma treatment of soybean has the highest degree of graphitization, its gas production under N₂ and CO₂ ambient is lower than that of millet, which may be due to the influence of the mechanism of plasma treatment of biomass related to the particle size of biomass.

Figure 11 shows the SEM images of the raw materials of millet and the carbon materials treated by plasma under different reaction ambientes. As shown in Figure 11(a), it is the original material of, the surface of the original millet was complete and there were many bumpy small particles. Figure 11(b) shows the SEM image of the millet treated by plasma under N₂ ambient. Compared with the original material surface, convex particles disappeared, and a small number of pore structures appeared. The pore structure was mainly large, with a diameter of about 40 μm. Figure 11(c) shows the SEM diagram of the millet treated by plasma in CO₂ ambient. Similar to the carbon material treated by N₂, a small number of pore structures appeared on the surface, but the pore structure was more developed, which may be due to the oxidizing ambient of CO₂ itself caused a more developed pore structure was produced. Figure 11(d) shows the SEM image of the millet treated by plasma under Ar ambient, which was quite different from the morphology of the products under the other two ambientes. There were still a small number of particles on the surface of the millet treated under Ar ambient, and the traces of particles eroded on the surface of the carbon material can also be obviously observed. This is due to the weak ionization energy of Ar ambient, the production of less active ions, resulting in weak biomass gasification performance. It can also further reveal the reason for the low gas production of millet after plasma treatment in Ar ambient. At the same time, through the above microscopic appearance, it can also be inferred that the mechanism of plasma treatment of biomass is layer by layer erosion. First, high-energy electrons react with particles present on the surface of the biomass. When the biomass surface is flat, high-energy electrons further interact with it, producing syngas and forming pore structures on the solid surface.

Figure 12 shows the SEM images of soybean raw materials and carbon materials treated by plasma under N₂, CO₂, and Ar ambientes. Figure 12(a) shows the morphology of soybean surface, the surface structure shows a scaly structure, and the surface is relatively flat as a whole. As shown in Figure 12(b), the surface of the sample treated by plasma under N₂ ambient is not smooth and a pit-like structure appears. This may be due to the high energy electrons produced by the plasma action and the gasification of the reactants layer by layer, the soybean particle size is large and the effect is weak under the same action time. Therefore, only pit-like structure appeared on the surface of soybean and no pore structure appeared. It is again proved that the mechanism of plasma action is the interaction of high energy electrons with reactants layer by layer. Figure 12(c) shows the SEM image of soybean products treated by plasma under CO₂ ambient. It can be found that the morphology of soybean products under N₂ conditions is similar, and only pit-like structure is produced on the surface, which is similar to that of another biomass. The pit structure of carbon materials produced by soybean under CO₂ is more obvious than that under N₂ ambient. Through the physical and chemical properties of the above carbon materials, it can be found that the high-energy electrons generated by the plasma will react layer by layer after contact with the biomass surface. Figure 12(d) depicts the SEM images soybean products treated by plasma under Ar ambient. It can be found that the surface of soybean after Ar treatment is not much different from the surface of the original soybean.