⇦Chinese Physics Letters, 2017, Vol. 34, No. 8, Article code 085203⇨ A Method of Using a Carbon Fiber Spiral-Contact Electrode to Achieve Atmospheric Pressure Glow Discharge in Air * Wen-Zheng Liu(刘文正)**, Shuai Zhao(赵帅), Mao-Lin Chai(柴茂林), Jiang-Qi Niu(牛江奇) Affiliations School of Electrical Engineering, Beijing Jiaotong University, Beijing 100044 Received 30 March 2017 ^*Supported by the National Natural Science Foundation of China under Grant No 51577011.
^**Corresponding author. Email: wzhliu@bjtu.edu.cn Citation Text: Liu W Z, Zhao S, Chai M L and Niu J Q 2017 Chin. Phys. Lett. 34 085203 Abstract During discharge, appropriately changing the development paths of electron avalanches and increasing the number of initial electrons can effectively inhibit the formation of filamentary discharge. Based on the aforementioned phenomenon, we propose a method of using microdischarge electrodes to produce a macroscopic discharge phenomenon. In the form of an asymmetric structure composed of a carbon fiber electrode, an electrode structure of carbon fiber spiral-contact type is designed to achieve an atmospheric pressure glow discharge in air, which is characterized by low discharge voltage, low energy consumption, good diffusion and less ozone generation. DOI:10.1088/0256-307X/34/8/085203 PACS:52.80.Hc, 52.80.Dy, 52.80.Tn © 2017 Chinese Physics Society Article Text Low-temperature plasma is increasingly employed in air purification, thin film deposition, surface modification, and other industrial applications.[1-4] Owing to the low efficiency of corona discharge, glow discharge plasma with moderate power density is the most promising among low-temperature plasma technologies. By inhibiting the production of secondary electrons using a double dielectric barrier discharge (DBD), it is relatively easy to generate atmospheric pressure glow discharge (APGD) plasma in rare gases such as helium.[5-8] However, the air has a special insulating property, resulting in easier transmission from a glow discharge to a filamentary discharge at high voltage, i.e., when the electric field strength is higher. It is generally reckoned that atmospheric air glow discharge can be formed only under the condition of a nanosecond-pulsed power or a nanometer-sized gap.[9,10] In our previous studies, utilizing the contact electrodes and non-uniform electric field distributions in gaps, we could achieve the steady generation of glow discharge plasma in atmospheric air.[11] In the existing methods, the uniform low-temperature plasma performed in atmospheric air exhibits poor diffusion. When the capillary jet is considered, glow discharge plasma can be generated outside the electrodes, but the generation area is small and difficult to expand.[12-14] However, a large amount of ozone will be produced during high-voltage discharges since the air contains oxygen, and excessive production of ozone would cause environmental pollution and would affect human health.[15] If the discharge voltage is lowered, i.e., the discharge is finished at a lower intensity of average electric field, it will contribute to reducing the decomposition rate of oxygen molecules, effectively suppressing the generation of ozone.[16] Therefore, to further explore the method of generating plasma with good diffusion performance, less ozone generation, and applicability to air purification, we present a spiral-contact electrode structure based on a carbon fiber material, wherein APGD is formed in air at a low voltage. The main aspect of the carbon fiber electrode structure is to utilize a microdischarge electrode to produce a macro glow discharge in atmospheric pressure air (microdischarge indicates that at least one dimension of the discharge space is below a millimeter scale). To further analyze the discharge characteristics of this electrode structure, we should discuss the electric field distribution around the discharge electrodes, which are composed of two fine electrodes placed in a cross-contact manner, as shown in Fig. 1(a). We obtain the distribution features of electric field lines in this structure utilizing Ansoft Maxwell 3D, as shown in Fig. 1(b). In the upper part of the figure, there is a carbon fiber electrode with an electrical conductivity, whose diameter is 0.2 mm (to conduct a better simulation study, the size of the carbon fiber electrode is increased appropriately). Further, in the lower part, there is a metal wire with an external diameter of 0.2 mm, wrapped in polytetrafluoroethylene (PTFE) dielectric in thickness 0.2 mm. In the absence of other external forces, we assume that the main electron avalanches are directed along the directions of the electric field lines. We have drawn the curves 1 and 2 along two of the electric field lines, representing the development paths of electron avalanches in the process of impact ionization, as shown in Fig. 1(c). It can be observed that the fiber electrode is surrounded by electric field lines, which is different from the case wherein the electric fields of macroscopic electrodes only exist between the two discharge electrodes, indicating that glow discharge may occur in the periphery and above the fiber electrode. Moreover, the discharge channels along the directions of electric field lines in the micro gaps become shorter (0–100 μm), limiting the lengths of electron avalanches. The density of plasma cannot increase continuously, causing the inhibition of filamentary discharge and contributing to the generation of glow. The discharge begins with a form of microdischarge in the minute gaps, and provides seed electrons for the open space, thus resulting in a discharge phenomenon in the millimeter scale.

Fig. 1. Characteristics of electric field lines for the carbon fiber electrode structure: (a) the electrode structure of the carbon fiber electrode, (b) the cross-section distribution of electric field lines around the electrode, and (c) the model of electron impact ionization paths when we draw curves along two of the electric field lines.

Fig. 2. Distribution of electric field strength along curve 1 and the development of an initial electron along this path.

When considering the path length along the electric field lines as an independent variable $d$ and the magnitude of the electric field as a dependent variable $E$, Fig. 2(a) shows the distribution of electric field along curve 1 (shown as curve a) and the result after nonlinear fitting (shown as curve a$'$). The function after fitting can be expressed as $$ E(d)=6.11\times 10^6e^{-d/93.99}+2.85\times 10^6\,{\rm V/m}.~~ \tag {1} $$ If only the process of impact ionization ($\alpha$ process) is considered, the number of electrons will increase according to $$\begin{align} n=n_0 e^{\alpha d'},~~ \tag {2} \end{align} $$ where $n_{0}$ is the initial number of electrons, $d'$ is the development distance of the electron avalanche, and $\alpha$ is the rate coefficient of electron impact ionization. Further, in air, $\alpha$ can be calculated as[17] $$\begin{align} \alpha = 6460e^{-1.9\times 10^7/E},~~ \tag {3} \end{align} $$ where $E$ is the magnitude of the applied electric field in the range 1.5$\times$10$^{6}$–1.1$\times$10$^{7}$ V/m. The electron mean free path in air is approximately 0.377 μm. It is assumed that electron impact ionization occurs in each mean free path. Subsequently, an iterative calculation is performed according to Eqs. (1)-(3). We could obtain the development of an initial electron in the process of impact ionization along path 1 from point A to point B, as shown in Fig. 2(b). We can observe that the impact ionization process is represented as a finite increase of electrons along the path of curve 1 (the final number is 107), and the process of increase is gradual. If an average intensity of electric field is chosen on curve 1 ($\bar {E}=5.90\times10^{6}$ V/m), it can be assumed that the impact ionization process occurs in this electric field strength and develops along the straight path from A to B. We can demonstrate that 2315 electrons are produced at point B, which is far greater than the number along the curve path. It is evident that this is a faster growth. This indicates that the special electric field distribution formed by the fiber electrode via the bending of electric field lines and formation of the non-uniform electric field with a gradient change effectively suppresses the development of electron avalanches in the process of impact ionization and provides theoretical support for glow discharge. The same method can be adopted when an electron starts the impact along the path of curve 2 from point C to point D. However, the number of electrons reaches a maximum at point P, and no longer increases because the electric field strength is too low to satisfy the ionization requirements after point P. Therefore, curve $l$ can be drawn by connecting all the points such as P. The plasma density outside curve $l$ remains constant, which is similar to the positive column in a low pressure dc glow discharge. Furthermore, the electric field strength exhibits not only a gradual decline along the directions of the electric field lines, but also a radial drop from the inside to the outside of the contact point. The number of electrons purely generated by the ionization is larger in the regions near the contact point owing to the stronger electric field intensity. However, the change of electric field from inside to outside and the open-type electrode structure enhance the diffusion probability and quantity of inner electrons diffusing outward, compensating for the number of outer electrons and guaranteeing the uniformity of discharge. However, the discharge space of the carbon fiber electrode is limited to a microscale, thus producing a microdischarge. The field emission at the cathode surface becomes non-negligible in the microdischarge under high electric field strength.[18,19] The threshold electric field required for field emission is a function of the cathode material and surface properties,[20] $$ D_{\rm FN} =(6.85\times 10^7)\frac{\phi ^{3/2}}{\beta}\,{\rm V/cm},~~ \tag {4} $$ where $D_{\rm FN}$ is the threshold electric field, $\phi$ is the work function of the cathode material, and $\beta$ is the geometric enhancement factor dependent on the geometry of the electrode and the distance between the two electrodes. The threshold electric field of the carbon fiber is usually lower, generally in the range of $1\times10^{6}$–$3\times10^{6}$ V/m.[21] The work function of the carbon fiber is 4.7 eV, and $\beta$ of the carbon fiber can be estimated using Eq. (4). It is approximately in the range of $2.33\times10^{4}$–$6.98\times10^{4}$, which is close to that of carbon nanotubes.[22] Compared with a metal electrode, even though the carbon fiber material does not have a relatively small work function, the enhancement factor is sufficiently large to achieve field emission at significantly lower field strength. A large number of initial electrons fill the discharge space owing to strong field emission, effectively reducing the initial discharge voltage and easily realizing glow discharge under a relatively low average field strength. Moreover, the larger value of $\beta$ further confirms that it is influenced by the shape of the carbon fiber electrode (sufficiently small radius of curvature) and the small gaps. By winding the insulated electrode (a metal wire covered with an insulating layer) with the carbon fiber electrode at a certain angle and bringing them in close contact to each other, a spiral-contact electrode structure is formed, as shown in Fig. 3(a).

Fig. 3. Distribution characteristics of the electric field strength and the electric field lines for the spiral-contact electrode structure: (a) the structure of the carbon fiber spiral-contact electrode, (b) the cross section of electric field strength distribution, and (c) the distribution of electric field lines.

The cross section of the electric field distribution is shown in Fig. 3(b). If the diameter of the metal wire, thickness of the dielectric, and diameter of the carbon fiber electrode are all set to 0.2 mm, and if the spiral pitch of carbon fiber is 5 mm, when the applied voltage is 1.8 kV, the whole electrode will be wrapped by a strong electric field region ($E>1\times10^{6}$ V/m) distributed as a wavy profile. The electric field strength in the vicinity of the contact points can be up to a magnitude of 10$^{7}$ V/m. It can be predicted that charged particles generated in the vicinity of the contact points, under the effect of the electric field, will drift to the periphery where the electric field is relatively weak, achieving a glow discharge phenomenon surrounding the entire electrode. Figure 3(c) shows a partial distribution of the electric field lines in the spiral-contact electrode structure. Each point on the dielectric surface and between the two spirals is affected by the carbon fiber spirals in all directions. As a result of the multi-directional superposition, the wrapping caused by the electric field distribution is helpful for the charged particles to drift and diffuse in a three-dimensional space, and ensures that the charged particles are limited to a certain space, which is very conducive to the generation of an APGD plasma in air.

Fig. 4. Experimental system.

To verify the analysis of the electric field of the carbon fiber spiral-contact electrode and the feasibility of glow discharge in atmospheric pressure air, discharge experiments have been carried out with the experimental system shown in Fig. 4. The experimental system includes three components: power supply, electrode structure, and measuring system. The power supply is a high-frequency high-voltage sine wave generator with a voltage range of 0–10 kV and a frequency range of 5–60 kHz. The frequency used in this study is approximately 20 kHz. The metal electrode is connected with the high-voltage terminal of the output, and the carbon fiber electrode is grounded. The outer diameter of the metal electrode and the thickness of PTFE dielectric both are 0.2 mm, and their lengths are 20 cm. A carbon fiber bundle (Tenax HTA40 1 K) for the spiral electrode is produced by Toho Co., with a resistivity of $1.6\times10^{-3}$ $\Omega$$\cdot$cm. The diameter of the monofilament is only 7 μm. The tufted carbon fiber spirally twines around the dielectric with a pitch of 5 mm. The discharge voltage $U$ is measured using a high-voltage probe (Tektronix P6015A), and the discharge current $I$ is obtained by measuring the voltage across a resistor $R$ in series with the electrode. The waveforms of discharge voltage and current are recorded using a Tektronix digital oscilloscope (TDS1012B-SC).

Fig. 5. Discharge phenomenon under a voltage of 1.8 kV (exposure time 0.083 s).

Figure 5 illustrates the uniform discharge phenomenon of the carbon fiber spiral-contact electrode in atmospheric pressure air. When the applied voltage is 960 V, the luminescence first appears in the vicinity of the carbon fiber electrode. When the applied voltage is up to 1.8 kV, a three-dimensional and homogeneous discharge emerges from the surface of the entire electrode, and is light blue in color. The soft glow diffuses along the radial direction of the wire electrode and completely wraps around it with almost no dark areas, and the carbon fiber is also completely buried in the plasma.

Fig. 6. Characteristics of steady discharge under a voltage of 1.8 kV.

Figure 6 indicates the characteristics of steady discharge under an applied voltage of 1.8 kV. Figure 6(a) shows that the values of discharge currents are in the magnitude of mA, demonstrating that there does not appear high concentration of plasma channels in the discharge space, which indicates that a typical filamentous discharge does not occur. Since the electrode structure of the carbon fiber spiral-contact type is in the form of a single DBD (an asymmetric electrode structure), the currents corresponding to the positive and negative half cycles are evidently unequal. During the discharge, the PTFE dielectric on the high-voltage electrode could absorb electrons and could inhibit the development of electron avalanches from the carbon fiber electrode. In contrast, the carbon fiber electrode could absorb electrons to develop the electron avalanches generated from the insulated electrode smoothly. Consequently, the instantaneous current amplitudes of negative half cycles are larger than those of the positive half cycles. Moreover, the formation of an electric field caused by charged particles in the dielectric layer is in the opposite direction to the applied electric field, which breaks the discharge. Further, there are preferential discharge points in the non-uniform spatial electric field formed by this electrode structure, thus representing a non-synchronous discharge in time and space, resulting in a combination of multiple current pulses with many peaks (each pulse width is several tens of nanoseconds). According to the Lissajous figure in Fig. 6(b), it can be calculated that the discharge power of a 20-cm-long electrode is 1.378 W. When the thickness of the luminescent region is estimated to be 0.6 mm, the power density is 3.046 W/cm$^{3}$, which is slightly larger than that of the APGD in rare gases owing to the strong insulating property of air and the higher frequency in this study. Both the current and discharge power have typical characteristics of dielectric barrier glow discharge. In summary, we have proposed a method of using microdischarge electrodes to form a macro glow discharge phenomenon and designed a carbon fiber spiral-contact electrode structure to achieve an APGD in air. The theoretical characteristics of the electrode are as follows: (1) the application of asymmetric electrode structure, formed by the carbon fiber, creates a special electric field distribution. By bending the electric field lines and generating a non-uniform electric field with a gradient change, the development of the electron avalanches in the $\alpha$ process is effectively suppressed, creating the conditions for the formation of glow discharge in air. (2) The formation of a three-dimensional and open-type discharge space: using the carbon fiber as an electrode, the distribution of the electric field lines is in a wrapped shape, and the plasma can be generated outside the electrode. The gradient change of the electric field strength provides favorable conditions for the charged particles in the small gaps to diffuse from the inner to the outer regions, enhancing the uniformity of discharge. (3) The field emission of microdischarge facilitates easier discharge. The carbon fiber has a strong capability of field emission. Acting as the non-insulated electrode in the single DBD, it can provide more initial electrons to the microdischarge space, playing a role in reducing the voltage and average electric field strength of discharge. The discharge demonstrates the advantages of low discharge voltage, low energy consumption, good diffusion, and less ozone generation, which is very suitable for air purification. In the future, the structural parameters of the carbon fiber spiral-contact electrode will be further optimized. Moreover, a large APGD plasma can be generated by arranging and combining a plurality of electrodes to satisfy the requirements of air purification and other applications. References An indoor air purification technology using a non-thermal plasma reactor with multiple-wire-to-wire type electrodes and a fiber air filterSurface Modification by Low Temperature Plasma: Sterilization of Biodegradable MaterialsHigh Speed Deposition of SiO ₂ Films with Plasma Jet Based on Capillary Dielectric Barrier Discharge at Atmospheric PressureStudy on Glow Discharge Plasma Used in Polyester Surface ModificationObservation of the transition from a Townsend discharge to a glow discharge in helium at atmospheric pressureSimulation of transition from Townsend mode to glow discharge mode in a helium dielectric barrier discharge at atmospheric pressureA Large Gap of Atmospheric Pressure RF-DBD Glow Discharges in Ar and Mixed GasesThe Key Factor for Uniform and Patterned Glow Dielectric Barrier DischargeA uniform glow discharge plasma source at atmospheric pressureA study of the glow discharge characteristics of contact electrodes at atmospheric pressure in airA single electrode room-temperature plasma jet device for biomedical applicationsAn Atmospheric-Pressure Glow-Discharge Plasma Jet and Its ApplicationCapillary plasma jet: A low volume plasma source for life science applicationsOzone generation using atmospheric pressure glow discharge in airBehaviour of atomic oxygen in a pulsed dielectric barrier discharge measured by laser-induced fluorescenceThe role of the field emission effect in direct-current argon discharges for the gaps ranging from 1 to 100 ? mThe Breakdown Phenomena in Micrometer Scale Direct-Current Gas DischargesA mathematical model of the modified Paschen?s curve for breakdown in microscale gapsComparison of carbon fiber and cesium iodide-coated carbon fiber cathodes

[1]	Jung J S and Kim J G 2017 J. Electrostat. 86 12
[2]	Stepczyńska M 2016 Plasma Processes Polym. 13 1080
[3]	Ito Y et al 2008 Appl. Phys. Express 1 067009
[4]	Liu W Z, Lei X and Zhao Q 2016 Plasma Sci. Technol. 18 35
[5]	Luo H Y, Liang Z, Lv B, Wang X X, Guan Z C and Wang L M 2007 Appl. Phys. Lett. 91 221504
[6]	Li X C, Niu D Y, Xu L F, Jia P Y and Chang Y Y 2012 Chin. Phys. B 21 075204
[7]	Li B, Chen Q, Liu Z W and Wang Z D 2011 Chin. Phys. Lett. 28 015201
[8]	Ouyang J T, Duan X X, Xu S W and He F 2012 Chin. Phys. Lett. 29 025201
[9]	Moon S Y, Choe W and Kang B K 2004 Appl. Phys. Lett. 84 188
[10]	Packan D 2004 PhD Dissertation (Stanford: Stanford University)
[11]	Liu W Z, Sun G L, Li C H and Zhang R R 2014 Phys. Plasmas 21 043514
[12]	Lu X P, Jiang Z H, Xiong Q, Tang Z Y and Pan Y 2008 Appl. Phys. Lett. 92 151504
[13]	Li X, Tao X M and Yin Y X 2009 IEEE Trans. Plasma Sci. 37 759
[14]	Topala I and Nagatsu M 2015 Appl. Phys. Lett. 106 054105
[15]	Buntat Z, Smith I R and Razali N A M 2009 J. Phys. D 42 235202
[16]	Ono R, Yamashita Y, Takezawa K and Oda T 2005 J. Phys. D 38 2812
[17]	Wang X X, Lu M Z and Pu Y K 2002 Acta Phys. Sin. 51 2781 (in Chinese)
[18]	Radmilović Radjenović M, Matejčik Š Klas M and Radjenović B 2013 J. Phys. D 46 015302
[19]	Radmilović Radjenović M, Radjenović B, Matejčik Š and Klas M 2014 Plasma Chem. Plasma Process 34 55
[20]	Go D B and Pohlman D A 2010 J. Appl. Phys. 107 103303
[21]	Shiffler D A, Lacour M J, Sena M D and Mitchell M D 2000 IEEE Trans. Plasma Sci. 28 517
[22]	Li Q and Liang E J 2005 Acta Phys. Sin. 54 5933 (in Chinese)