Experimental study of CO2 dissociation in an expanding dc arc discharge at atmospheric pressure with multiple pin-to-pin electrodes

In the following study, we examine the performance of a new design of the classic gliding arc discharge (GAD) with diverging electrodes between dielectric walls at atmospheric pressure. In the present design, a tungsten pin-to-pin electrode pair ladder replaces the standard curved diverging electrodes. A major problem with the classic GAD design is the surface erosion of the electrodes, which leads to bad repeatability and issues with long-term usage. The new construction provides controlled electrode wear at well-defined points of arc attachment. This ensures stable operation while retaining the arc expansion effect. This new configuration of the discharge is being developed for gas treatment. In the current work, the device was applied for CO2 dissociation. The most significant quantities for this application of the GAD, the CO2 conversion rate and energy efficiency, are measured at different gas flow rates (2–12 Ln/min) and arc currents (50–210 mA). The results are analyzed and compared with previous measurements using the classic GAD at the same conditions.


Introduction
Today, the emission of huge amounts of greenhouse gases into the environment is a global concern.Therefore, people are working on methods to reduce the amount of CO2 in the atmosphere.One of them is to convert CO2 gas into valuable products, by combining them with other gases.The latter usually requires the dissociation of CO2 into CO and O2, which are much more reactive substances.There are several chemical and physical methods to do this [1,2], with one being plasma utilization.Various gas discharges and configurations are being researched with the aim of achieving higher values for both conversion and energy efficiency of the process [35].This is where the GAD, working at atmospheric pressure, shows potentiality with its low-cost realization and capability for repeatability [4,6,7].
Gliding and expanding arc discharges are non-stationary gas discharges.The arc is ignited in the region with the highest value of the applied electric field, namely in the narrow region between two vertically diverging electrodes in a gas stream.In the classic design, the electrodes are flat plates cut to the appropriate curvature, usually made of aluminum, copper or stainless steel sheets [1,4,6,7].However, the high electric fields and temperatures at which the arc operates cause the electrodes to heat and erode.This is inevitable, particularly at the cathode and anode spots [1,6].Reducing these effects would result in a more stable and durable device.
This work aims to explore the effect of replacing solid electrodes with a cascade of pin-to-pin stages made of thin cylindrical electrodes with increasing distance between each pair.This would bring several advantages, which are particularly significant for industrial applications.It can provide controlled electrode wear at precise positions and stable operation by attachment of the arc on the same anode and cathode spots.
In the current work we investigate the configuration of the discharge with pin-to-pin electrodes under the same conditions as our previous measurements of the GAD [8,9].Cascades using either 2 or 3 pairs of cylindrical electrodes are investigated.

Experimental setup
The experimental setup is presented in figure 1 (a).The system works at atmospheric pressure.The CO2 gas flow passes through a system of pipes and is controlled using a mass-flow controller (MFC Bronkhorst EL-FLOW F-201CM).The gas is then introduced to the discharge device by a nozzle at the bottom of the discharge construction.The discharge device is presented in figure 1 (b) and will be described in more detail below.The discharge is connected to a high-voltage power supply.The voltage drop and the arc current are measured with a high-voltage differential probe (Pintek DP-30K) and а clamp probe (Pintek PA-699) respectively.Both probes are connected to an oscilloscope.The gas analysis is done using Fourier transform infrared absorption spectroscopy (FTIR Perkin Elmer Frontier NIR/MIR).To determine the CO2 dissociation, we use the intensity of a CO single line at 2209 cm -1 , which belongs to the molecule band.The intensity of the line multiplied by a coefficient gives us the CO2 conversion.The experiments were carried out with CO2 gas flows in the range 2 -12 Ln/min and discharge currents from 50 to 210 mA.
The discharge is a modification of the classic GAD, but the diverging electrodes are replaced by a cascade of tungsten pin-to-pin electrode pairs as shown in figure 1 (b).The horizontal distance between the electrodes is increased for each subsequent pair from bottom to the top.The distances between electrodes in our 3 pairs device is about 3,7, 11 mm.The arc ignites at the shortest distance between the electrodes, which in this case is between the bottom pair.As the gas rises, the arc moves to the next pair of electrodes.When the arc reaches its maximum length, it extinguishes and ignites again at the bottom.The cycle repeats itself.The difference between this configuration and the one used in the previous studies [8,9] is that the arc attaches to the same electrode spots every time, unlike the classic GAD where the arc moves on the electrode's surface and attaches to different spots with every ignition.
The electrodes are placed between two quartz glasses which additionally constrict the arc and the gas flow.The tungsten rods are covered with ceramic tubes which help in channeling the gas, so it cannot enter between the stages of the cascade.The electrodes used are 1.0 mm WR2 Tungsten rods with mixed oxides.

Experimental results
The main parameters that we examine are the conversion of CO2 and the energy efficiency.Conversion shows the percentage of the converted CO2 gas as the change of the CO2 concentration over the initial concentration of CO2 gas: The energy efficiency is the percentage of energy used for CO2 dissociation to the input energy:  = (  2 ∆  /( × where ∆  = 279.8/ is the reaction enthalpy for the CO2 splitting reaction and SEI (J/Ln) is the specific energy input.A more detailed explanation of these calculations can be found in [9].For the sake of proper comparison of the current results with the classic solid-electrode GAD, the measurements of the conversion rate and the energy efficiency are performed for the same conditions and setup as in [9].The comparison is shown in figures 2 and 3.The classic GAD results are presented as trend lines (denoted as NSGAD in [9]), obtained by polynomial fitting of multiple data points.
The conversion of CO2 as a function of the gas flow for the three arc current values 50, 100 and 210 mA is presented in figure 2 (a).The data shows similar conversion in both discharge configurations and the same trend.The conversion increases with the increase of the current and decreases with the increase of the flow.
In figure 2 (b) one can see the conversion as a function of the Specific energy input (SEI).The graph shows that the conversion reaches a saturation region at around 2.5 SEI.The lines are a polynomial fit of data points from [9].
The other important quantity, the energy efficiency, is presented in figure 3 (a) as a function of the gas flow rate and in figure 3 (b) as a function of the SEI.The dotted lines again show the polynomial approximations of the data from [9].The behaviour of the efficiency is the same as classic GAD -it increases with the gas flow and decreases with current increase.Using the pin-to-pin electrodes the energy efficiency seems to be about 7 -10 % higher compared to classic GAD.This result is probably due to the device's design which favors an arc regime operation, instead of glow regime operation even at low currents of 50 mA.As it was shown in [9], at current in to the order of 100 mA the discharge is preferably in glow regime (large cathode spot controlled by secondary electron emission), despite the fact that the positive column is contracted.In the current design, the longtime attachment of the arc to a single spot leads to its heating and a transition to an arc regime controlled by thermo-field emission.In that case the cathode fall is much smaller and energy losses are lower than in the monolith electrodes in glow regime.The cathode fall in arc regime is expected to be lower than 10 V (not measured) while in glow regime could be a few hundred volts.
Figure 3 (b) shows the energy efficiency as a function of SEI.There is a tendency for energy efficiency to decrease as SEI increases, which was also observed previously [1,2,5,9].The lines are a polynomial fit of data points from [9].
A comparison was also made between the 2-stage and 3-stage cascades.In figure 2 (a) the empty dots represent the 2-stage cascade, and the solid ones represent the 3-stage cascade.There are slight differences in the conversion rates and the energy efficiencies that can be seen at flow rates of 10 and 12 Ln/min.The conversion rate for 2 pairs of electrodes seems to be slightly higher.At lower flow rates the difference between the 2 and 3-stage cascades is within the uncertainty range, which was previously calculated to be 10 -12% [9].The input power for the two configurations (at the same flow rate) is the same, thus, the deviation in the energy efficiency is due to the conversion rate.

Conclusions
A new design based on the classic gliding arc discharge was developed.The solid diverging electrodes were replaced by cascade arrangements using either 2 or 3 stages (pairs) of cylindrical tungsten electrodes with increasing interelectrode distance for each pair.The general behaviour of both the conversion rate and the energy efficiency is similar compared to this of classic GAD -the conversion decreases with the flow rate and increases with the SEI, while the energy efficiency increases with the flow rate and decreases with the SEI.The comparison of the new design with the classic GAD shows some advantages of the discharge with pin-to-pin electrodes.A higher energy efficiency at nearly the same CO2 conversion rate was observed, attributed to the different operation regimes between the setups.For the studied discharge current range, GAD favours glow regime, while the pin-to-pin setup operates in arc regime.
A comparison between 2 and 3-stage cascades was also made.No significant differences for the conversion rates and energy efficiencies were observed at lower flow rates.Both parameters are slightly different at higher flow rates.The higher conversion rate cannot be explained by a higher power input.It may be due to some other effects, but the available data is not sufficient to draw any firm conclusions.
To summarize, the preliminary experiment shows that the pin-to-pin electrodes setup holds promise, however, further in-depth research is required for more conclusive results.

Figure 1 .
Figure 1.Experimental setup in (a) and discharge device in (b).

Figure 2 .
Figure 2. The conversion XCO2 as a function of the gas flow rate (a) and the specific energy input (b).The lines are a polynomial fit of data points from[9].

Figure 3 .
Figure 3.The energy efficiency η as a function of the gas flow rate (a) and specific energy input (b).The lines are a polynomial fit of data points from[9].