Effect of High Voltage Electrode Material on Methanol Synthesis in a Pulsed Dielectric Barrier Discharge Plasma Reactor

Abstract

Plasma methanol synthesis from captured CO₂ and renewable H₂ is one of the most promising technologies that can drastically lower the carbon footprint in methanol production, but the associated high energy costs make it less competitive. Herein, we investigated the impact of the high-voltage electrode configuration on methanol formation. The effect of electrode materials Cu, Al, and stainless steel (SS) SUS304 on CO₂ hydrogenation to methanol using a temperature-controlled pulsed dielectric barrier discharge (DBD) plasma reactor was examined. The electrode surface area (ESA) was varied from 157 mm² to 628 mm² to determine the effect on discharge characteristics and the overall influence of plasma surface reactions on methanol production. The Cu electrode showed superior methanol synthesis performance (0.14 mmol/kWh) which was attributed to its catalytic activity function, while the Al electrode had the least production (0.08 mmol/kWh) ascribed to the excessive oxide coating on its surface, passivating its ability to promote methanol synthesis chemical reactions. In all electrode materials, the highest methanol production was achieved at 157 mm² ESA at a constant applied voltage. Lastly, the plasma charge concentration per discharge volume was determined to be an important parameter in fine-tuning the DBD reactor to enhance methanol synthesis.

1. Introduction

Historically, methanol has been widely used as a valuable chemical feedstock for the manufacture of various consumer goods ranging from building materials and textiles to pharmaceuticals [1]. Recently, its emerging application as a hydrogen energy carrier and a low-carbon substitute fuel, especially in the maritime sector, is increasingly gaining popularity [2,3]. In one of the reports by the International Energy Agency (IEA), methanol production is expected to exceed that of ammonia and other high-value chemicals when combined by 2030 [4]. However, despite the exceptional qualities, conventional methanol production is heavily reliant on fossil fuels (99%) and thus negatively contributes towards climate change [5]. In addition, in the concluded United Nations Conference of Parties (COP 28) more than 180 countries agreed on the need to transition away from fossil fuels in order to maintain the overarching goal of the Paris Accord of limiting the temperature increase beyond 1.5 °C [6]. Therefore, as a matter of urgency, a more sustainable methanol synthesis process that is capable of overcoming this challenge together with difficult production conditions of very high pressure and temperatures is required.

To substitute this complex thermal process, nonthermal plasma technology (NTP) has been considered alongside other novel solutions for CO₂ hydrogenation to liquid methanol. The uniqueness of NTP is that highly energetic electrons (mean electron energy 1–10 eV) generated are able to dissociate stable molecules such as CO_2, allowing the creation of an avalanche of chemically reactive species that could lead to the formation of new materials [7]. Moreover, this could be attained at a favorable environment of atmospheric pressure and temperature. NTP powered by renewable electricity coupled with carbon capture and utilization (CCU) for hydrogenation of CO₂ can substantially lower the carbon footprint of methanol production.

In the past, studies involving CO₂ hydrogenation with the dielectric barrier discharge (DBD) plasma have been popular among many researchers [8,9,10,11,12,13], with considerable effort dedicated to plasma catalysis. This unidirectional focus has contributed to a significant knowledge gap on the fundamental effects of plasma reactor design and parameters on the methanol synthesis mechanism. From these studies, Wang et al. [13] successfully investigated reactor design modifications in addition to using various catalyst packings. In their findings, the methanol concentration and selectivity increased 37 and 54 times, respectively, (from 0.1 to 3.7 mmol/L and 1 to 54.2%) when they changed the ground electrode structure from aluminum foil to circulating water. Their effective reactor design propelled the use of water-ground electrodes in subsequent studies on methanol synthesis with DBD plasma but still, the problematic challenge of low energy efficiency remains. From this, we deduce that the electrode configuration type, electrode/dielectric material, and other components of the plasma reactor can be very instrumental in enhancing methanol synthesis with DBD.

Table 1. Catalyst-free methanol synthesis performance under various DBD plasma configurations.

DBD Plasma System Electrode Configuration	Conditions	Feedstock Gasses	Selectivity CH3OH	Energy Consumption/Efficiency	Ref.
Stainless steel (high voltage), Water (ground)	9.2 kHz, 24 W, 30 °C	CO2 + H2	7.5%	387.5 kJ/mmol	[8]
Stainless steel (high voltage), Water (ground)	9 kHz, 18 W, 60 °C	CO2 + H2	17%	181.7 kJ/mmol	[11]
Stainless steel (high voltage), Water (ground)	9.5 kHz, 26 W, 60 °C	CO2 + H2	3.7%	855.9 kJ/mmol	[9]
Stainless steel (high voltage), Water (ground)	9 kHz, 10 W, 30 °C	CO2 + H2	54.2%		[13]
Copper (high voltage), Water (ground)	40 kHz, 30 W, 40 °C	CO + CO2 + H2 + (N2 additive)	86%	4512 kJ/mmol	[14]
Stainless steel (high voltage), Water (ground)	9.2 kHz, 20 W, 5 °C	CH4 + CO2	43%	>0.8 mol/kWh	[15]
Stainless steel (high voltage), Water (ground)	30 W, 85 °C	CH4 + O2	42%	0.76 mol/kWh	[16]

shows a summary of various DBD plasma system configurations for methanol synthesis.

Other researchers have shown that different central electrode materials have had tremendous effects on plasma chemical reactions in, ozone generation [17,18], CO₂ decomposition [19,20,21], and plasma-aided methanol decomposition experiments [22,23], among others. Fitria et al. [18] found that hexagonal stainless steel electrodes produced high ozone concentration compared to perforated aluminum electrodes due to a high electric field from the sharp edges in the spacings. Dey and Kamble [19], reported that Cu resulted in a superior CO₂ conversion than SS and Al, while Emeraldi et al.’s [20] results showed that SUS630 and Cu had superior and almost similar energy efficiency performance on CO₂ decomposition at higher applied voltages (18–20 kV_p-p) compared with SUS400 and carbon. Also, Iwamoto et al. [24] investigated the performance of eleven different electrode materials on plasma ammonia synthesis. They observed that the catalytic activity on ammonia production rate (μmol/min) of Pt, Pd, Ag, Cu, and Ni wools increased as the experiment was repeated, while that of Au, Fe, Mo, Ti, W, and Al remained constant.

Fewer studies have focused on the combined effect of electrode material and the electrode surface area in various gas mixture plasmas. For instance, Ma et al. [25] conducted a study on the plasma synthesis of ammonia using different inner electrode materials. In their findings, Cu outperformed SS, titanium, and NiFe alloy electrodes with its ability to hinder reverse reaction on formed ammonia on the surface. In addition, the expanded surface area of the tangled wire (63 cm²) compared to the rod electrode (45 cm²) showed a clear transformation of the DBD charge characteristics by enhancing the charge transfer leading to more efficient ammonia synthesis. Similarly, Suraidin et al. [17], in their study on ozone production using various inner electrodes, reported that SS material was most preferred over Cu and galvalume due to its high corrosion resistance. It is important to note that when the discharge length increased five times at a constant applied voltage and input flow rate, the ozone concentration nearly doubled.

Despite this insightful evidence, the high-voltage electrode material effect on multiple gas mixture plasmas remains underexplored, particularly in methanol synthesis reactions. In this study, we aim to explore and maximize the inherent properties of various electrode materials such as charge conductivity to promote chemical reactions enhancing methanol production. We perceive that to be able to substantially reduce the energy cost for methanol synthesis in DBD plasma systems, developing a foundational understanding of the effects of inner electrode materials and their surface configuration is of critical importance.

Herein, we investigated the effect of high-voltage electrode material in the DBD plasma using three commonly used elements, Cu, Al, and SS. The methanol synthesis experiment using a DBD plasma reactor was carried out using three primary reactant gasses, H₂, CO₂, and CO, at a fixed stoichiometric number. The concentrations, conversion, and selectivity of the gaseous and liquid products were analyzed. In addition, the electrical and discharge characteristics of the different electrode materials were examined. The electrode surface area was varied (by changing the electrode length) at a constant flow rate and power to determine the effect on charge characteristics and the overall influence on plasma surface reactions on methanol production. Also, a comprehensive analysis of the electrode surface morphological changes due to plasma exposure was conducted.

2. Results and Discussion

2.1. Effect of the Electrode Material on the Discharge Characteristics

We characterized the electron release on the surface of the electrode by evaluating the behavior of the electrical current and voltage signals during plasma operation (Figure 1). The data were captured 30 min after plasma ignition when the discharge was visibly stable. From the current waveform results (Figure 1a), both the Cu and Al electrodes had higher current amplitude peaks ranging from 0.09 A to 0.35 A and 0.11 A to 0.41 A, respectively, in the positive half cycle unlike the SS electrode (0.12 A to 0.3 A) which was slightly lower. These stronger current peaks translate to more high-energy electrons with greater acceleration capable of enhancing the electron density for robust chemical reactions. Furthermore, in all electrode materials, there was almost a similar intensity of the micro-discharges displayed both in the positive half-cycle and the negative half-cycle. The applied voltage was fixed at 8.5 kV (as seen in Figure 1b), corresponding to a constant discharge power of 9.3 W except for the Al electrode which was slightly higher by about 7.2%.

To obtain important information about the discharge in the DBD plasma reactor, a simple Lissajous-figure method proposed by T.C. Manley in 1943 was adopted [26]. The Lissajous figure of the plasma discharge for the water ground electrode reactor (Figure 2a) was an elliptical shape indicative of significant residual electron density in the discharge gap surviving through the capacitive phase period between sequential discharge cycles [27] or the presence of residual ions at all times [26]. The ellipse limits accurate estimates of the effective capacitance for a fully bridged discharge gap (C_diel) and the integral capacitance of the DBD reactor (C_cell) from the slopes. In Figure 2a, we can infer that the discharge characteristics of the DBD reactor are almost alike in all three electrodes at ESA of 157 mm² due to the uniform Lissajous shapes. Moreover, the minimum external voltage (U_min) where plasma ignition takes place is almost constant in the Cu, Al, and SS electrodes.

Several Lissajous plots were obtained during the methanol synthesis plasma reaction using different electrode materials at various ESA as summarized in Figure 2b. Notable transformations in the Lissajous plots were observed (in Supporting Information Figure S1a–c) when the ESA changed from 157 mm² to 628 mm² for Cu, Al, and SS electrodes, respectively. This behavior shows that the change in ESA impacts the discharge characteristics of the plasma. As the electrode surface area increased, the peak-to-peak charge (Q_p-p) and the applied voltage decreased in a similar trend across all electrode materials at fixed power (Figure 2b). The Q_p-p and applied voltages were greatest in all electrodes at the smallest ESA (157 mm²) and least at the highest electrode surface area (628 mm²). The area of the Lissajous figures increased with the decrease in the electrode surface area due to an increase in plasma charge density (herein defined as the peak-to-peak charge concentration per discharge volume).

2.2. Effect of the High-Voltage Electrode Material on the Plasma Methanol Synthesis Process

Figure 3 highlights the effects of electrode material on the methanol synthesis process in relation to the reactant gas conversion and selectivity of products. The Cu and SS electrodes displayed minor differences in the conversion of CO₂, unlike the Al electrode which was much lower as seen in Figure 3a. SS had the highest CO₂ conversion with 8.2% and 46% more than Cu and Al, respectively. In all electrode materials, the CO quantity at the reactor exit was higher than at the entry point implying more generation of CO gas than what was consumed for the formation of other products. Thus, the apparent CO conversion was zero though the injected CO might have been converted to other hydrocarbons.

The selectivity of methanol, CH₄, CO, and C₂H₆ were determined and highlighted in Figure 3b. The Cu and Al electrodes had the highest and lowest methanol selectivity, respectively. Whereas the methanol synthesis reaction in the Cu electrode was the most dominant (contributing to 50.5% methanol selectivity), the reverse water gas shift (RWGS) reaction was seen to be the main reaction for the Al electrode leading to a high CO selectivity of 53.2%. In addition, the methanation reaction which is another competitive chemical reaction taking place in the plasma discharge, resulted in a CH₄ selectivity of 19.4%, 28.6%, and 24.9% in the Cu, Al, and SS electrodes, respectively. The Cu electrode also showed better selectivity towards C₂H₆ formation at 50% more than the Al and SS electrodes.

In Figure 3c, we observe that methanol concentration had a relatively matching trend to the methanol selectivity trend seen earlier in Figure 3b. The Cu electrode had the highest methanol concentration at 0.057 mol/L followed by SS (0.0475 mol/L) and then the Al electrode (0.036 mol/L). From the electrical diagnostics, Cu, Al, and SS displayed an almost identical discharge mode (Figure 2a) and very minor differences in the streamer-filaments behavior (which indicate various chemical reaction channels and reactive species [15]). This implies that at a high applied voltage, the electrode material has a weaker influence on the plasma discharge since it is easier for electrons to break the potential barrier on the electrode surface [28]. Therefore, the superior methanol synthesis performance of the Cu electrode over Al and SS was attributed to the catalytic properties of metallic Cu on CO₂ hydrogenation.

Our case is supported by both Yang et al. [29] and Liu et al. [30] extensive studies that combine DFT calculations and thermal catalytic experiments. They reported that methanol is synthesized exclusively from CO₂ on metallic Cu surfaces. Similarly, some previous studies propose that metallic Cu surfaces play a key role in methanol synthesis via formate pathway (Route I) [31,32], while the oxide supports help in the dispersion of Cu particles and stabilizing metallic Cu sites [33]. Also, Wang et al. [13] in their study on plasma catalytic hydrogenation of CO_2, presented that the Cu-based catalyst (Cu/γ-Al₂O₃) had much higher methanol concentration (25.6 mmol/L) than the Pt-based catalyst (Pt/γ-Al₂O₃) at 7.5 mmol/L though they had comparable selectivity results. They suggest that elemental Cu’s ability to moderately bind with key intermediates in the hydrogenation process could be the reason for better performance.

Route I. Methanol synthesis progression on metallic Cu surfaces

At the end of the experiments, we performed SEM-EDS analysis on the electrodes to obtain a deeper insight into the surface morphological changes after exposure to plasma. SEM images of the electrodes before the experiment (Figure 4a–c) clearly contrast with the images after exposure to plasma. The Cu electrode surface (Figure 4(a*)) had unevenly distributed spots of confirmed carbon deposits while Al electrode in Figure 4(b*) had more visible pattern imprints on the surface. Similarly to both the Cu and Al surfaces, carbon particles were observed on the SS electrode surface (Figure 4(c*)).

Figure 4(a1–c1 and a2–c2) further illustrates the surface distribution of carbon (red dots) and oxygen (light blue dots) elements before the electrodes were exposed to plasma and after the methanol synthesis experiment, respectively. The mapping image was obtained to visualize the distribution of deposited elements, without which it would be difficult to know how vast the invisible oxygen covered the electrode surface. Furthermore, the SEM images alone limit the carbon distribution to the visible specks unlike the mapping image that highlights the carbon micro-particles scattered across the surface. Both oxygen and carbon were distributed at varying quantities along the whole electrode surface with a reasonable matching carbon concentration on the mapping image compared to the spots on the SEM images.

On the Cu electrode surface, the oxygen distribution before the experiment (Figure 4(a1)) compared to after exposure to the plasma (Figure 4(a2)) appeared to have faded slightly while the carbon distribution seemed to have increased. On the other hand, the Al electrode surface before exposure (Figure 4(b1)) had a limited oxygen distribution compared to after the plasma experiment (Figure 4(b2)). This shows that it was highly oxidized in contrast to Cu and SS electrodes after the experiment. Regarding the SS electrode at the start (Figure 4(c1)) and the end (Figure 4(c2)) of the experiment, the mapping results displayed a fair increase in both the oxygen and carbon elements on the surface. This result was further summarized graphically in Figure 5 to distinguish the approximate percentage elemental weight distribution on the electrode surface.

As evident in the EDS mapping images (Figure 4), the concentration of oxygen reduced by about 18.5% after the plasma experiment on the Cu electrode, whereas it increased by 16% on the SS surface and about five times on the Al electrode (Figure 5). We suspect that reduction reactions of the formed oxides on the Cu electrode surface by reducing agents CO (Equations (1) and (2)) [19,32,34] and H₂ (Equations (3) and (4)) [35,36] contributed to this result.

$$CuO+CO \to Cu+{CO}_2$$

(1)

$$2CuO+CO \to {Cu}_{2}O+{CO}_2$$

(2)

$$CuO+H_2 \to Cu+H_{2}O$$

(3)

$${Cu}_{2}O+H_2 \to 2Cu+H_{2}O$$

(4)

The Al electrode had the highest oxygen distribution on its surface at 18% followed by SS (2.5%), then Cu (0.8%) after exposure to plasma. Carbon deposits ($C{O}_2+e^{-}\to C•+•2O+e^{-}$) on the SS electrode surface after plasma exposure was the lowest, compared to the Cu and Al electrodes.

From the CO₂ conversion results, the SS electrode displayed the best CO₂ conversion compared to the Cu and Al electrodes though the current amplitude peaks were lower comparatively at a fixed applied voltage of 8.5 kV (as seen in Figure 1). Past reports have shown that higher and stronger current pulses imply greater electron density in the DBD plasma discharge thus promoting vigorous collision among electrons and reactant species leading to higher CO₂ conversion since decomposition takes place via electron impact dissociation reactions (Equations (5) and (6)) [20]. However, this was not the case for both the Cu and Al electrodes in our scenario.

$$C{O}_2+e^{-}\to C{O_2}^{*}+e^{-}$$

(5)

$$C{O_2}^{*}+e^{-}\to CO•+•O+e^{-}$$

(6)

We suspect that for the Cu electrode, the net CO₂ conversion was compromised by the reduction in formed oxides (CuO and Cu₂O) on the surface of the electrode to metallic Cu or Cu₂O by CO (Equations (1) and (2)). This was further confirmed by the decrease in oxygen concentration on the Cu surface as seen in the EDS electrode surface analysis (Figure 4 and Figure 5).

In the case of the low CO₂ conversion with the Al electrode, it was observed that the surface oxygen changed significantly from 3.4% to 18% compared with 2.2% to 2.5% in the SS electrode (Figure 4). This implies that the Al surface was extremely oxidized ($2 Al+3O \to {Al}_2{O}_3$) and as a result, passivated to become chemically inert from plasma surface reactions [37]. This oxidation could be taking place largely from the oxygen atoms when CO₂ is decomposed. Moreover, this reactive oxygen species has been linked to the poisoning of catalyst surfaces in other CO₂ hydrogenation studies leading to low methanol formation [38].

Whereas the Al₂O₃ coating formed on the Al electrode surface is non-reducible [39,40], the reducibility of the oxides formed on the Cu electrode seemed to have contributed to the enhancement of the methanol synthesis performance. Reducible metal oxides have been shown to play an important role in the selective production of oxygenates such as formaldehyde and methanol [41]. We thus speculate that the dynamic plasma-assisted oxidation-reduction surface reactions taking place on the Cu electrode could have been vital in achieving a proper balance in the migration of the oxygen radicals to renew the active sites for the continuous production of methanol intermediate species.

2.3. Effect of Varying the Electrode Surface Area (ESA) on Methanol Production

Previously, we illustrated how a change in the ESA affected the charge concentration and applied voltage at constant power (Figure 2b). This phenomenon was investigated further on how it influences the conversion of reactant gasses and product selectivity for methanol synthesis. In Figure 6a we see that a change in the ESA has a dramatic effect on the conversion of reactant gasses especially in the SS electrode material. In all the electrodes, the CO₂ conversion decreased by 49.1% (Cu), 50% (Al), and 56.2% (SS) when the ESA changed from 157 mm² to 628 mm². A steep decline in the CO₂ conversion was observed when the ESA changed from 157 mm² to 314 mm². On the contrary, there was a moderate decrease in the CO₂ conversion for Cu and Al electrodes when the ESA changed from 314 mm² to 628 mm² and a marginal increase of 1.5% for the SS electrode. We also observed a slight CO conversion of the originally injected amount in the feedstock mixture at 628 mm² for the Cu electrode.

These findings correlate well with the change in the plasma peak-to-peak charge concentration (Q_p-p) in the discharge zone when the ESA changed from 157 mm² to 628 mm². The decrease in Q_p-p was greatest between 157 mm² and 314 mm² than between 314 mm² and 471 mm² and 471 mm² and 628 mm². At the highest peak-to-peak charge concentration in the plasma discharge (smallest area 157 mm²), the environment was rich with dense reactive and energetic species which resulted in intense collisions leading to maximum CO₂ conversion. As the density of these actively charged species reduced with the increase in the ESA, the CO₂ conversion followed a similar downward trend.

The selectivity results have been summarized in Figure 6b. From these findings, we see that the impact of the ESA varies largely depending on the type of electrode material with optimal methanol selectivity attained at different ESA. In the Cu electrode, the methanol selectivity gradually increased as the ESA changed from 157 mm² to 628 mm² attaining its maximum selectivity at 471 mm² and then decreased slightly. For both the Al and SS electrodes, the methanol selectivity reached its peak at 314 mm² then decreased with further increase in the ESA. We noted that all three electrodes had very high CO selectivity at the least ESA, correlating well with the CO₂ conversion results in Figure 6a. This suggests that the CO₂ decomposition was much higher than both the methanol synthesis and methanation reactions, where in the latter two processes several hydrogenation steps are required.

On the other hand, the CH₄ selectivity was greatest at the highest ESA for all electrode materials implying that the methanation reaction was the most dominant compared to methanol synthesis reaction or RWGS reaction. Logically, the increased ESA provides favorable residence time for the chain hydrogenation process to be achieved for methane formation (Route II) [42,43,44]. The influence of the Cu electrode was further noted in the C₂H₆ selectivity results. The C₂H₆ selectivity was at its highest at around 0.9% and remained relatively stable between 314 mm² and 628 mm² but in both Al and SS electrodes, the C₂H₆ selectivity was lower by about 40%.

Route II. Stepwise methane hydrogenation process

The methanol production per kWh was determined and highlighted in Figure 6c. Here, the trend was almost identical to the methanol selectivity trend when the ESA changed from 157 mm² to 628 mm². In all electrode materials, the methanol synthesis was least at 157 mm² but increased when the ESA increased to 314 mm². The Cu electrode sustained an upward methanol production trend beyond 314 mm² to reach the highest amount at 471 mm² whereas for the Al and SS electrodes, we observed a minimal decline in production of 8.3% and 11.7%, respectively, between 314 mm² and 628 mm².

From our analysis, we can deduce that the plasma charge density has a reasonable influence on the plasma reactions leading to varying output products. Earlier we observed that a higher plasma charge density, i.e., 22.4 nC/mm³ (represented by a high concentration of charged species consisting of electrons and ions within the discharge zone) at 157 mm² amplified collision of reactant particles leading to high CO₂ decomposition (Figure 6a). However, this intensely charged environment was not conducive for the formation of liquid products (methanol) thus resulting in low methanol production across all three electrodes (Figure 6c) while enhancing gaseous products such as CH₄ and CO. As the Q_p-p decreased with the increase in ESA, the methanol production increased at 314 mm² and later decreased at 471 mm² for Al/SS electrodes and at 628 mm² for the Cu electrode.

The SEM-EDS analysis of the electrode surface was conducted on all three electrodes at 157 mm², 314 mm², 471 mm², and 628 mm² ESA. The quantified elemental data by weight composition across the electrode sample surface was summarized as seen in Figure 7. The distribution of oxygen and carbon elements on the electrode surface seemed not to be affected much by the change in ESA. The Cu electrode still maintained a mean decrease in the oxygen content, but the Al electrode showed >250% change across all ESA. The SS electrode had a moderate increase in the oxygen content while the carbon percentage change was in the order of Cu > Al > SS. It has been reported that CO dissociation (one of the carbon sources) rarely occurs on the Cu surfaces [38], this could possibly explain the very low carbon deposition observed in Figure 7 when compared to the Al and SS electrodes. As the ESA increased, the change in the distribution of oxygen and carbon averaged within a similar range on all electrode materials implying that the plasma surface reactions along the electrode length adjusted in a fairly proportional manner.

We investigated further how the methanol synthesis chemical reactions were affected by the changing plasma charge density when the ESA changed from 157 mm² to 628 mm² at a constant flow rate of 48 mL/min and fixed applied voltage of 8.5 kV_p-p. The Lissajous plots obtained were almost similar implying no major changes in the plasma discharge mode when the ESA increased from 157 mm² to 628 mm² (Figure S2). The plasma charge density was determined from the peak-to-peak charge concentration against the discharge volume and plotted in relation to the changing methanol production, discharge power, and gas residence time (Figure 8). In all three electrodes (Cu, Al, and SS), the methanol production decreased indiscriminately with the increase in the ESA from 157 mm² to 628 mm². This trend was nearly identical to the changing charge concentration per unit discharge volume at constant applied voltage when the electrode surface area changed, suggesting an interesting link between them. Uniform discharge characteristics (Lissajous plot) need to be maintained to achieve this kind of relationship.

As the ESA changed from 157 mm² to 628 mm², the power increased proportionately varying from 9.3 W to 37 W, respectively. Despite the low specific energy input (SEI, ratio of input power to flowrate) at ESA 157 mm², the CO₂ conversion was moderately high perhaps due to the high charge density observed (11.2 nC/mm³) that provided a conducive atmosphere for increased collision among reactive species. This in turn shifted the selectivity towards methanol leading to high production. When the charge density dropped to 4.9 nC/mm³ at 314 mm², the CO₂ conversion was at its lowest, but it eventually increased between 314 mm² and 628 mm². This turn-around could be a result of more energetic plasma species (from the increasing SEI) that accelerated the CO₂ decomposition. At this interval (314 mm² and 628 mm²), both the CO and CH₄ selectivity were enhanced even though the charge density declined. On the contrary, methanol selectivity assumed a downward trend with increasing ESA. The increase in CO and CH₄ selectivity with the increase in ESA suggests that high SEI enhanced RWGS and methanation reactions while low SEI promoted the methanol synthesis process. Moreover, at the least ESA (157 mm²) the utilization of the plasma discharge energy for methanol synthesis reactions rather than other plasma surface reactions such as oxidation was more efficient, enhancing the methanol yield. From this, we learn that the DBD plasma processing for methanol production can be tuned up by controlling both the energy and density of reactive species through the ESA to achieve the desired results.

Nonetheless, the individual electrode materials also play a critical role in influencing methanol production. Cu electrode had the highest production at ESA 157 mm² followed by SS and lastly Al electrode. As discussed earlier, the catalytic properties of Cu material cannot be underestimated. This made the Cu outperform the other electrode materials while the excessive oxidation on the Al electrode resulted in a poorer performance. With this perspective, more consideration can be laid on electrode material selection to improve the DBD reactor design and performance in future plasma methanol synthesis studies.

Herein, we have demonstrated that there is a likelihood of considerable plasma-assisted surface reactions along the central high-voltage electrode that promote (in the case of Cu) or inhibit (in the case of Al) methanol synthesis besides the gas phase chemical reactions. The Cu electrode seemed to undergo plasma cleaning (with oxygen removal) renewing the surface for more effective performance while the SS electrode remained resilient throughout the plasma operation with moderate surface modifications. Though the Al electrode showed poor performance comparatively due to over-oxidation of its surface, we draw an important insight on how the electrode can be exploited for oxygen trapping to prevent poisoning of the catalyst surface for greater methanol yield [38].

3. Materials and Methods

3.1. Experimental Setup

Figure 9 shows the basic experimental setup. The coaxial DBD reactor had an inner diameter of 4 mm, and a total volume of 35 mL between the outer and inner cylinders. 2 mm diameter high-voltage inner electrodes of different materials (copper, aluminum, and stainless steel SUS304) were used. The electrodes were standard rods available commercially and were used as received without any prior treatment. The purity of the materials was determined to be, Cu—98.6%, Al—95.8%, and SS (Fe—70.6%, Cr—19.9% Ni—6.2%, Mn—1%) using the X-ray fluorescence spectrometer (Bruker AXS S8 TIGER-MA 1 kW, Karlsruhe, Germany). The discharge gap was 1 mm and the electrode surface area (ESA) was varied from 157 mm², 314 mm², 471 mm², to 628 mm² corresponding to electrode length (EL) of 25 mm, 50 mm, 75 mm, and 100 mm, respectively. A self-cooling reactor operated with circulating water solution (30 g of sodium chloride per liter) as the ground electrode, was maintained at $40 °C$ using a cooling unit (Inverter cool bath, LTBi-550). The water ground electrode was selected to maintain low operating temperatures which are favorable for the synthesis of oxygenates like methanol [41]. The plasma was ignited by a high-voltage pulsed power supply with a maximum discharge power of 2 kW and a frequency range of 0 to 100 kHz (Hayden PHF-2K-2U). The frequency was fixed at 40 kHz, and the voltage was maintained at 8.5 kV_p-p. A four-channel digital oscilloscope (Tektronix, MD034, Beaverton, OR, USA) was used for electrical measurements. A high-voltage probe (Tektronix, P6015A 1000X) was used to measure the applied voltage, and a current probe (Tektronix, TCP0020) was used to monitor the current. The charge transferred in the plasma zone was determined using an external measuring capacitor (33.2 nF) placed between the high-voltage inner electrode and the ground electrode and was measured by another high-voltage probe (Tektronix, P6015A 1000X).

3.2. Product Analysis

Mass flowmeters (KOFLOC) were used to monitor the in-flow of feed gasses, H₂, CO₂, and CO while the soap-film flowmeter was used to check the change in gas volume after the reaction. The flow rate was maintained at 48 mL/min. The molar ratio was based on the stoichiometric number (in this case SN of 3.45), which is the ratio of the difference between H₂ and CO₂ moles and the sum of CO and CO₂ moles as in Equation (7) [45]. From the methanol synthesis equations ($C{O}_2+3H_2\rightleftharpoons C{H}_{3}OH+H_{2}O$ and $CO+{2H}_2\rightleftharpoons C{H}_{3}OH$), SN of 2 is most ideal [1,45], but based on our previous findings [14] the SN > 3 showed better methanol synthesis by minimizing competitive reverse water gas shift (RWGS) reactions.

$$SN=(\mathrm{m}\mathrm{o}\mathrm{l} H_2-\mathrm{m}\mathrm{o}\mathrm{l} C{O}_2)/(\mathrm{m}\mathrm{o}\mathrm{l} C{O}_2+\mathrm{m}\mathrm{o}\mathrm{l} CO)$$

(7)

The reaction product gasses exiting the reactor were cooled at $-12.5 °\mathrm{C}$ using a cooler (THOMAS, TRL 108H) to condense the oxygenates present. The condensed liquid sample collected in a glass tube was then analyzed using gas chromatography (Shimadzu, GC 2014B, Kyoto, Japan) equipped with a flame ionized detector (FID) and Porapak-Q 50-80 F-3444 column, after 1.25 h of the experiment. The exhaust gaseous products were directly analyzed online using a gas chromatograph (Shimadzu, Nexis^TM GC-2030) with a barrier discharge ionization detector and micro-packed st MP-01 column by injecting samples at time intervals of 25 min.

The concentration of oxygenate products condensed from the exhaust gasses was calculated using standard calibrated concentration curves to evaluate the reaction performance. The conversion and selectivity calculations were based on the reactant concentration at the inlet and outlet of the DBD plasma reactor.

CO₂ and CO conversion were calculated by Equation (8) and Equation (9), respectively.

$$X_{{CO}_2} \left(\mathrm{\%}\right)={\displaystyle \frac{C{O}_{2,in} \left[\mathrm{m}\mathrm{o}\mathrm{l}\right]-C{O}_{2,out} \left[\mathrm{m}\mathrm{o}\mathrm{l}\right]}{C{O}_{2,in} \left[\mathrm{m}\mathrm{o}\mathrm{l}\right]}}\times 100$$

(8)

$$X_{CO} \left(\mathrm{\%}\right)={\displaystyle \frac{C{O}_{in} \left[\mathrm{m}\mathrm{o}\mathrm{l}\right]-C{O}_{out} \left[\mathrm{m}\mathrm{o}\mathrm{l}\right]}{C{O}_{in} \left[\mathrm{m}\mathrm{o}\mathrm{l}\right]}}\times 100$$

(9)

The CO, CH₄, and C₂H₆ selectivity were calculated as per Equations (10), (11) and (12), respectively.

$$S_{CO} \left(\%\right)={\displaystyle \frac{C{O}_{out} \left[mol\right]}{(C{O}_{2, in} \left[mol\right]-C{O}_{2, out} \left[mol\right])+(C{O}_{in} \left[mol\right]-C{O}_{out} \left[mol\right])}}\times 100$$

(10)

$$S_{{CH}_4} \left(\mathrm{\%}\right)={\displaystyle \frac{C{H}_{4,out} \left[\mathrm{m}\mathrm{o}\mathrm{l}\right]}{(C{O}_{2,in} \left[\mathrm{m}\mathrm{o}\mathrm{l}\right]-C{O}_{2,out} \left[\mathrm{m}\mathrm{o}\mathrm{l}\right])+(C{O}_{in} \left[\mathrm{m}\mathrm{o}\mathrm{l}\right]-C{O}_{out} \left[\mathrm{m}\mathrm{o}\mathrm{l}\right])}}\times 100$$

(11)

$$S_{{C_{2}H}_6} \left(\mathrm{\%}\right)={\displaystyle \frac{2\times { C}_2{H}_{6,out} \left[\mathrm{m}\mathrm{o}\mathrm{l}\right]}{(C{O}_{2,in} \left[\mathrm{m}\mathrm{o}\mathrm{l}\right]-C{O}_{2,out} \left[\mathrm{m}\mathrm{o}\mathrm{l}\right])+(C{O}_{in} \left[\mathrm{m}\mathrm{o}\mathrm{l}\right]-C{O}_{out} \left[\mathrm{m}\mathrm{o}\mathrm{l}\right])}}\times 100$$

(12)

Also, the selectivity of liquid products was calculated as follows:

Total liquid selectivity:

$$S_{Liq} \left(\mathrm{\%}\right)=100\%-[S_{CO}+S_{{CH}_4}+S_{{C_{2}H}_6}]$$

(13)

The quantity of methanol in the liquid sample was determined using standard calibration concentration curves. The mole percent concentration was obtained and used to calculate the methanol selectivity below, Equation (14).

$$S_{C{H}_{3}OH} \left(\mathrm{\%}\right)=\mathrm{m}\mathrm{o}\mathrm{l} \% \mathrm{o}\mathrm{f} C{H}_{3}OH in liquid output\times S_{Liq}$$

(14)

Finally, the methanol production rate per kWh was calculated as per Equation (15).

$$\mathrm{M}\mathrm{e}\mathrm{t}\mathrm{h}\mathrm{a}\mathrm{n}\mathrm{o}\mathrm{l}\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}/\mathrm{e}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{g}\mathrm{y}\left(\mathrm{m}\mathrm{m}\mathrm{o}\mathrm{l}/\mathrm{k}\mathrm{W}\mathrm{h}\right)={\displaystyle \frac{C{H}_{3}O{H}_{\left[out\right]} \left(\mathrm{m}\mathrm{m}\mathrm{o}\mathrm{l}/\mathrm{m}\mathrm{i}\mathrm{n}\right)}{Discharge power \left(\mathrm{k}\mathrm{W}\right)}}\times 60$$

(15)

where the discharge power is the time-average dissipated power in the plasma discharge.

3.3. Electrode Surface Characterization

The surface features of the electrodes (before and after the experiment) were characterized by employing a scanning electron microscope with energy dispersive spectroscopy (SEM-EDS, S-4800 and S-4300 HITACHI High-Tech Co., Ltd. Tokyo, Japan). SEM images were captured with an accelerating voltage of 10 kV.

3.4. Electrical Parameters Evaluation

The discharge power was calculated using the voltage-current-time integration method in Equation (16), based on the measured applied voltage and current data obtained using the oscilloscope.

$$P\left(\mathrm{k}\mathrm{W}\right)=f{\int }_0^{T}V\left( t \right) I\left(t\right) dt$$

(16)

where f is the frequency of the discharge cycle and T is time.

Other electrical diagnostics were performed using the charge-voltage measurement technique, where the total charge (Q) was calculated from the voltage $U_C$ across the external measuring capacitor $C_{ext}$ (33.2 nF) as shown in Equation (17).

$$\mathrm{Q}=C_{ext}\times U_C$$

(17)

Moreover, a Q-V Lissajous plot was used to determine the discharge parameters of the DBD reactor such as the capacitance of the discharge gap $C_{gap}$ (Equation (18)), breakdown voltage $U_b$ (Equation (19)) and reactor charge transfer $Q_{trans}$ (as seen in Figure S3).

$$C_{gap}(\mathsf{\mu }\mathrm{F})={\displaystyle \frac{C_{diel} \left(\mathsf{\mu }\mathrm{F}\right)\times C_{cell}\left(\mathsf{\mu }\mathrm{F}\right)}{C_{diel}\left(\mathsf{\mu }\mathrm{F}\right)-C_{cell}\left(\mathsf{\mu }\mathrm{F}\right)}}$$

(18)

$$U_b(\mathrm{k}\mathrm{V})={\displaystyle \frac{1}{1+\left[C_{gap}\right(\mathsf{\mu }\mathrm{F})/C_{diel}(\mathsf{\mu }\mathrm{F}\left)\right]}}\times U_{min}(\mathrm{k}\mathrm{V})$$

(19)

The dielectric layer capacitance $C_{diel}$ and U_min (the minimum external voltage where micro discharges are observed [26]) were also obtained from the Lissajous figures.

4. Conclusions

The need to minimize inefficiencies and losses in the self-cooling DBD reactor contributing to high energy costs for methanol production led us to investigate the effect of central electrode material on plasma methanol synthesis reactions. In this work, we studied how different materials (Cu, Al, and SS) affected the conversion of reactant gasses and the selectivity of related products. The behavior of the plasma discharge characteristics of the different electrode materials on methanol synthesis was analyzed. Moreover, modification of the electrode surface after exposure to plasma that promoted/inhibited certain chemical reactions was studied. The electrode surface area of the high-voltage electrode was adjusted (from 157 mm² to 628 mm²) to understand the contribution of plasma surface reactions in the methanol synthesis process. An in-depth electrical and electrode surface material characterization was conducted by use of an oscilloscope and SEM-EDS equipment, respectively.

Overall, the methanol synthesis performance was in the order of Cu > SS > Al at 157 mm² ESA despite all electrode materials sharing similar discharge characteristics. The Cu electrode’s high methanol production (0.14 mmol/kWh) was attributed to the intrinsic catalytic properties of the material. Furthermore, the dynamic oxidation and reduction plasma surface reactions on the Cu electrode seemed to renew the active sites allowing for enhanced methanol synthesis. On the other hand, the poor methanol production by the Al electrode (0.08 mmol/kWh) was brought about by the excessive oxide coating on the surface passivating its ability to promote methanol formation chemical reactions. Also, the Cu electrode showed better C₂H₆ selectivity compared to SS and Al electrodes even at the minimum ESA (157 mm²).

Significant differences in the plasma discharge mode were observed when the electrode surface area changed. The peak-to-peak charge concentration was found to be inversely proportional to the ESA and at a charge density of 11.2 nC/mm³, the methanol production was at its highest for all electrode materials. The plasma charge concentration per discharge volume was determined as one of the important parameters in optimizing the conversion of reactant gasses, product selectivity, and methanol synthesis.