Eur. Phys. J. D 60, 455–463 (2010) DOI: 10.1140/epjd/e2010-00266-5 THE EUROPEAN PHYSICAL JOURNAL D Regular Article DC atmospheric pressure glow microdischarges in the current range from microamps up to amperes V.I. Arkhipenko, A.A. Kirillov, Y.A. Safronau, and L.V. Simonchika Gas Discharge Physics Laboratory, Stepanov Institute of Physics of the NAS of Belarus, ave. Nezavisimostsi 68, 220072 Minsk, Belarus Received 30 April 2010 / Received in final form 16 August 2010 c EDP Sciences, Società Italiana di Fisica, Springer-Verlag 2010 Published online 20 October 2010 – Abstract. Electrical and spectroscopic results of comprehensive investigations of microdischarges in atmospheric pressure helium, argon, nitrogen and air at gaps less than 1 mm and over a large current range from microamperes up to amperes are presented. At distance shorter than 1 mm, electrodes influence the normal glow discharge structure and their electrical and optical parameters. Two oscillation discharge regimes were revealed in discharge current range comparable or less than the low-current limit of normal glow discharge. 1 Introduction Among the atmospheric pressure glow discharges (APGD), microdischarges (MD) have demonstrated the possibility of generating stable diffuse, non-thermal plasmas with characteristic dimensions in the 0.01–1 mm range [1–9]. The unique properties and several selected applications of microplasmas are discussed in review article [1]. Among the most attractive features of microplasmas are applications such as plasma cathodes, which can support larger-volume DC atmospheric-pressure glow discharges in various gases including air [2,3,9–12], the realization of plasma reactors for chemical/bacterial decontamination [13], volume plasma chemistry [14,15] and the enormous potential for the development of novel photonic devices [16,17]. In MD devices, the processes in the cathode fall are more dominant than in the case of macroscopic glow discharges. As a consequence, plasmas at low gas temperatures can be generated with electron energy distributions which contain large concentrations of high energy electrons due to non-local effects. This can lead to a change of the plasma chemical processes promoting the three-body reactions. In this case, spontaneous emission rates and the breakdown of quasi-neutrality will undoubtedly become of greater importance. This is a new interesting field of plasma science at pd values that are comparable to those for large-volume, low-pressure plasmas, promising new fundamental physics results. A number of configurations and microplasma devices including coaxial cathode and anode microcavities, spirals or slits in the cathode and/or anode, etc. have been developed in the past decade. We will focus on DC MDs at atmospheric pressure. As an example of the cathode a e-mail: simon@imaph.bas-net.by boundary layer discharge, Stark and Schoenbach [2,3] and Sankaran and Giapis [4,5] used a micro hollow cathode with dimensions of several hundred microns. The current values in such discharges were up to ∼20 mA. Another example of microdischarge is the DC glow microdischarge investigated in [6,7] between a planar cathode and a wire end, separated by a distance on the order of 0.4 mm in free space in different atmospheric pressure gases like helium, argon, hydrogen, nitrogen and air. The discharge current did not exceed 5 milliamps. The third kind of DC MD, dissimilar to these two in its geometry was presented in [8]. It is the MD between two parallel planar electrodes separated from each other by a distance of about 0.2 mm. As can be seen, the discharge current did not exceed several milliamps in the DC MD investigations and was larger than 10 mA in isolated experiments. This is quite clear, because the cathode fall is characterized by high volumetric power densities in the range of tens to hundreds of kW cm−3 . The energy losses to the surface surrounding the plasma contribute to enhanced plasma stability. Such microplasmas in air, for example, are non-thermal plasmas with temperatures ranging from several hundred to over 1000 K [7]. Due to the elevated temperature, the regimes of stable operation for microplasma are often limited to low currents and simple gas mixtures. The elevated gas temperature can thus pose problems for the application of such discharges to processes. The small dimensions of the MDs often limit their applications as well. In [18–22] we presented the results of DC self-sustained normal APGD investigations in helium, argon and nitrogen over a wide current range. These investigations were mainly concerned with the discharge cathode fall and positive column for interelectrode gaps of 1 and 10 mm, when the APGD was in normal glow discharge mode. As shown, the cathode region parameters stay constant in the current 456 The European Physical Journal D range from hundreds of microamps up to ten amps and do not practically depend on discharge gaps in the range of 1–10 mm. In this paper, we focus our attention mostly on the APGD at small interelectrode gaps, comparable with the dimensions of near cathode layers (cathode fall, negative glow, Faraday dark space), in the current range corresponding to the transition from normal glow discharge to subnormal (or Townsend) one and ten milliamps. We can characterize this discharge as microdischarge. At what current and gap is there an influence of the electrodes on the electrical parameters and discharge structure? What is the lower limit of current that corresponds to the normal microdischarge mode? When do non-local effects begin to influence the discharge structure formation? In this paper, we will try to answer these and some another questions. Although the dc normal microdischarge is only one type of microdischarge we hope that the most basic results obtained here may be applicable to other forms of microdischarges. The paper is organized as follows. Section 2 is a short presentation of the experimental setup and diagnostic equipment. Experimental results including the illustrative photographic descriptions, electrical and spectroscopic measurements and some discharge parameter calculations are described in Section 3. Obtained experimental results are discussed in Sections 4. Section 5 presents the conclusions. 2 Experimental setup A schematic of the experimental setup used for the generation of MDs is shown in Figure 1. Glow discharge is ignited between two electrodes in air-locked chamber 1 with quartz glass windows used for spectroscopy and visualization of the discharge. The external dimensions of the discharge chamber are 100 × 100 × 70 mm3 . The weakly rounded tungsten anode 2 is 6 mm in diameter and flat cathode 3, made of copper, is 36 mm in diameter. An image of the electrode in the discharge chamber is shown in Figure 1 as well. The interelectrode gap can be varied from 0 to 10 mm. A weak flow of helium, argon, nitrogen or air through the discharge chamber (40 cm3 /s) was provided. The impurity concentrations in the gas flows did not exceed 0.02%. Depending on the experimental situation, two power supplies (PS in Fig. 1) were used. These were stepup transformers with typical bridge rectifiers and capacitive filtering. Voltage variation was performed by an autotransformer. One power supply with output voltages of about 0–5 kV and a ballast resistor of several hundreds of kΩ was used at low discharge currents (less than 100 mA); the other power supply with output voltages of 0–1.5 kV and a ballast resistor of about 100 Ω was used at high discharge currents (more than 100 mA). The interelectrode voltage was measured with a digital voltmeter CH300 (instrumental error 0.1%) and the discharge current was controlled using ammeter M1105 (instrumental error 0.2%). Resistors R1 and R2 (Fig. 1) formed a voltage divider with the ratio of 1000 : 1 connected to an oscilloscope Fig. 1. (Color online) The APGD feed circuit and the image of electrodes. 1 – Discharge chamber, 2 – anode, 3 – cathode, PS – power supply, V – voltmeter, A – amperemeter, R – ballast resistance, R1, R2 – divider resistors, R3 – instrument shunt, O – oscillograph. which was used to measure the discharge voltage and oscillations. Resistor R1 = 50 Ω was used to measure the discharge current and its oscillations. Power supply voltage and current reading were also built into the high voltage power supply. A scanning 0.5-meter high resolution monochromator (two gratings with 1800 grooves/mm) produced by SOLAR-TII was used for the spectral decomposition of the APGD emission. The inverse linear dispersion was ∼0.5 nm mm−1 , and the halfwidth of the Gaussian instrumental profile was equal to Δλa ≈ 0.011 nm. A photomultiplier FEU-171 was used as a converter of light intensity to electric signal. An enlarged discharge image was focused onto the entrance slit of the monochromator using a lens. The electrodes were oriented so that the optical axis and the entrance slit were parallel to the cathode surface. To obtain spatial distributions of the light emission intensity the discharge chamber was placed on a specially designed platform which could be moved in two directions along and across to the optical axis by micrometer screws. Experimental data were collected and processed using a personal computer. 3 Experimental results 3.1 Discharge images Images of the APGD in helium, argon, nitrogen and air at different discharge currents and the same interelectrode gap of 1 mm are shown in Figure 2. Electrodes are located in the manner as shown in Figure 1: the cathode is at the bottom of the image and the anode is at the top of the image. All these discharges in general have similar structures all having visible characteristic regions of normal glow discharge. As can be seen, a thin layer of negative glow appears right above the cathode surface; a positive column extends to the anode. There is a Faraday dark space between these two luminous regions. Anode glowing layers for most discharges are constricted or divided V.I. Arkhipenko et al.: DC atmospheric pressure glow microdischarges in the current range... 457 Fig. 2. (Color online) Images of the DC helium, argon, nitrogen and air APGDs at a discharge gap of 1 mm against different currents. Numbers in brackets indicate exposure time. into separate spots on the anode surface. There are obvious differences in the apparent color of the discharges and variations in the relative dimensions and brightness of the regions for different gases and discharge currents. The transverse dimensions of the characteristic discharge regions depend on the discharge current values. Negative glow diameter changes from a few hundred microns at a discharge current of about 1 mA up to a centimeter at one ampere. Negative glows in the helium and nitrogen APGDs are homogeneous in the radial direction. The argon discharge is inhomogeneous and has an asymmetrical shape. Moreover, argon negative glow is not stable and undergoes chaotic motion. The positive columns of all the discharges are constricted. At the same time, as it is shown in [18], the helium positive column at a gap of 10 mm at the same currents is diffuse. The images of discharges at a discharge current of 5 mA at different gaps are shown in Figure 3. The gap was varied from about 0.06 mm up to 1.4 mm. As can be seen, the visible changes in discharge structure are observed at gaps less than 0.5 mm. At gaps less than about 0.1 mm, the negative glows area increases a little. The anode layers become more homogeneous and their dimensions become similar to the cathode ones. 3.2 Voltage-current characteristics Voltage-current (V -I) characteristics reflect the dependence of interelectrode voltage on discharge current. The V -I characteristics for the investigated APGDs in helium, argon, nitrogen and air at the same interelectrode gap of 1 mm are shown in Figure 4. Before the discussion of these characteristics, it is necessary to speak about the stability of both the voltage and discharge current over the entire measurement range. The stabilities of both the current and the voltage were controlled by recording of the corresponding signals using a digital oscillograph C8-40 (Belvar) and subsequent Fig. 3. (Color online) Images of the DC helium, argon, nitrogen and air APGDs at discharge current of 5 mA against different gaps. Fourier-conversion of the registered signals. It should be noted that oscillations of the discharge parameters at the mains frequency (50 Hz) and its harmonics are practically absent (<1%). Different kinds of oscillations in helium, argon and nitrogen APGDs are observed at currents less than about 1 mA. Their existence is due to the falling part of the V -I characteristic at the transition from Townsend discharge to a normal glow one. At low pd (less than 1–2 cm torr), the mechanisms of these fluctuations were described in [23,24] for low pressure glow discharge. At atmospheric pressure, current and voltage relaxation pulsations caused by the constrictedglow-discharge to transient-spark transition in the constricted discharge regime in turbulent flow of nitrogen, are presented in recent work [25]. The authors of paper [7] note the existence of similar oscillations in such micro APGDs as well, although they have chosen discharge regimes without oscillations for their investigations. Of course, the parameters of these oscillations strongly depend on the ballast resistance value, power lead capacitance and interelectrode gap. In our case, the fluctuation frequency changes from a few kilohertz to hundreds of kilohertz. Taking into account the high frequency of fluctuations and the stationary state of the discharge, the APGD characteristics will be used at their average values in Figure 4 and hereinafter. As it follows from Figure 4, the V -I characteristics of all the discharges for small gaps (1 mm) and discharge currents larger than 1–2 mA are slowly falling or close to flat. Such V -I characteristic behavior is typical for normal glow discharge [26,27]. According to [21], voltage drops in the 458 The European Physical Journal D Fig. 4. (Color online) The V -I characteristics of the APGDs in helium (triangles), argon (circles), nitrogen (snowflakes) and air (stars) at an interelectrode gap of 1 mm. Fig. 6. (Color online) Images of helium discharge at planar copper electrodes at different gaps. Fig. 5. (Color online) Images of the dc APGD in helium, argon, nitrogen, and air at a discharge gap of 1 mm against currents. The upper row of images corresponds to oscillating regimes and lower one corresponds to normal glow discharge regimes. Numbers in brackets indicate exposure time (second). cathode fall layer are 145, 190 and 202 V for helium, argon and nitrogen APGDs, respectively. Differences between the interelectrode voltages and these cathode drops are divided between other characteristic discharge regions (negative glow, Faraday dark space, anode region). At currents larger than about 1 mA for all studied discharges, the APGDs preserve the structure of normal glow discharge as shown in Figure 5 (bottom row of images). If the current values are 2–3 times lower than 1 mA, the discharge structures are somewhat different. For helium, for example, negative glow becomes dimmer (left image in upper row in Fig. 5) and its area increases. At the same time, the interelectrode voltage slightly increases (Fig. 4). The interelectrode voltage of argon, nitrogen and air discharges significantly increases as the current decreases below 1 mA (Fig. 4) and discharges transfer into oscillating modes (Fig. 5, upper row). As can be seen, a significant number of short-living thin discharges like filaments is observed during the exposure time indicated in brackets. 3.3 Peculiarities of microdischarges Let us consider the peculiarities of discharges at atmospheric pressure at small gaps (less than 1 mm) and low currents (less than 10 mA). At such discharge parameters, we define these discharges as microdischarges. The discharge peculiarities described here are inherent to the four discharges presented here. Therefore, we will study the most common among them, microdischarge in helium, as an example. This discharge operates at smaller voltage at low currents and is the most stable. The discharge structure as a function of the interelectrode gap can be traced using Figure 3 (upper row), where images of helium discharge at different gaps are shown. In the experiment, the discharge current was 5 mA. Unfortunately, the changes in discharge structure with decreasing gap are not clearly defined in this picture. On the other hand, in our experiments, we used a weakly rounded (radius of about 3 mm) anode. For this reason, one can claim that all the differences at small gaps are due to the nonplanar anode configuration. In order to overcome this we performed the experiment with two planar copper electrodes of 36 mm diameter and at larger current. The discharge current was 0.5 A and the gap was varied from 0.4 up to 4.5 mm. Discharge images are shown in Figure 6. Here, we can see the changes in discharge structure with gap variation. These changes are even more distinct. At large gaps, the positive column is constricted and only its length changes. It is observed up to gaps of about 1 mm. At a gap of about 1 mm the positive column is practically absent and the luminous glow in the middle of the gap becomes more constricted. At all these large gaps, no changes are observed in negative glow. Simultaneously, with decreasing gap, the anode spots become clearly defined. V.I. Arkhipenko et al.: DC atmospheric pressure glow microdischarges in the current range... 459 (a) (b) Fig. 7. (Color online) Interelectrode voltage vs. gap at different currents. Significant changes in discharge structure are observed at gaps of less than 1 mm. Discharge becomes diffuse, the color becomes more bluish. Negative and anode glows now look like thin planar layers close to the electrode surfaces. Negative glow slightly increases in diameter and is homogeneous. Anode glow is inhomogeneous and consists of a large number of small fuzzy spots. Between these two layers there is a faint glow. The area of the anode layer practically corresponds to the negative glow one. Thus, at discharge gaps of less than 1 mm, which are typical for operation of atmospheric pressure microdischarges, electrodes strongly influence the normal glow discharge structure. 3.3.1 Interelectrode voltage vs. gap Such discharge changes should manifest themselves in the discharge parameters’ behaviour. In Figure 7, the dependences of interelectrode voltage on gap at different discharge currents are shown. As one can see, the experimental points for currents of 10 and 3 mA at gaps larger than 1 mm are located on straight lines (dashed lines in Fig. 7) which cross the vertical axis at a voltage slightly exceeding the cathode fall value (about 145 V for the helium-copper pair [21]). However, the experimental points at gaps less than 1 mm are located considerably below these straight lines. The difference can be about 20 V. As it is seen, on the other hand, interelectrode voltage does not practically change at these small gaps when the discharge current is varied from 0.5 up to 10 mA. 3.3.2 V-I characteristics at small gap As indicated above, there are some peculiarities in the voltage dependences for small gaps (less than 1 mm). In addition, at low currents, discharge can be unstable (see Fig. 4). Let us to study these facts in more detail. For these experiments, we used a discharge gap of about 0.55 mm. Our standard electrode combination (planar cathode and a weakly-rounded anode) was used as well: weakly rounded tungsten anode and a planar copper cathode. The ballast resistor was 85 kΩ. Fig. 8. (Color online) V -I characteristic (a) and discharge images (b) with a ballast resistor of 85 kΩ. The interelectrode gap is 0.55 mm. Arrows indicate the directions of current change. In Figure 8, two V -I characteristics are shown. They were obtained for two cases of discharge current variation: increasing current and decreasing current. They are practically the same at discharge currents larger than 1 mA. Some differences in voltage (about 10 V) take place at currents below 1 mA and correlate with changes in the external appearance of the discharge. They can be seen in Figure 8b (images at a current of 0.6 mA). Arrows in this figure point out the directions of current variation: a current increase for the upper row of images and a current decrease for the lower one. As one can see, with increasing current, the discharge is constricted in the anode region and in the middle part. Here, its color is in general granite and does not change up to small current (about 0.3 mA). In the case of decreasing current both the discharge structure and its color change beginning from a current of 1 mA. It becomes bluish. The discharge blows out at currents below 0.3 mA. Let us now increase the ballast resistor up to 650 kΩ. This results in an increase of power supply voltage. The corresponding V -I characteristic is shown in Figure 9a. We emphasize once more that in these characteristics the values of both the voltage and current are averaged over time, because in this case the discharge is in oscillation mode at small current (less than 3 mA). Depending on the direction of current variation we obtain two curves. As the discharge current increases from small values (about 0.1 mA) up to 3 mA, the voltage steadily decreases to about 200 V. Upon further current increase, an abrupt voltage decrease of about 50 V is observed. After that, the voltage negligibly changes with increasing current. In the opposite direction of current variation this voltage level is practically the same. So, at a current of about 0.6–0.8 mA, the voltage is only about 160 V. At the same time, it is about 225 V when the current decreases further. Thus, we observe a hysteresis of voltage in the current range of 0.6–3 mA. The rows of images in Figure 9b illustrate corresponding changes in discharge structure. The upper row of 460 The European Physical Journal D (a) (a) (b) (b) Fig. 9. (Color online) V -I characteristic (a) and discharge images (b) at ballast resistor of 650 kΩ. Interelectrode gap is 0.55 mm. Dashed arrows indicate the directions of current changes. images is specific to the case of current increase. As one can see, at current variation from 0.1 mA up to 3 mA the discharge structure is constant, but the glow intensity increases. Discharge images have constricted column and negative glow dimensions of which do not practically change. At a discharge current of about 3 mA, when an abrupt transition in voltage occurs, the discharge structure changes. We observe constricted anode glow, a weak glow in the middle part of the gap and negative glow with a smaller transverse dimension. This discharge structure persists with further current increase. The bottom row of images illustrates how discharge structure changes with current decrease. As can be seen, the structure corresponding to higher currents persists up to a current of about 0.5 mA with decreasing current. At this point, the abrupt changes in discharge structure occur. These changes take place in all discharge regions. This especially concerns the changes in the negative glow. Its area increases by a factor of 5. At first glance it seems that we have discharges at low currents with the current density on the cathode much lower than the normal one. However, this is not true. These discharges have normal current density on the cathode. This fact is confirmed by the measurement of discharge parameter dynamics. 3.3.3 Discharge parameter dynamics To trace the changes in the discharge parameters over time, the interelectrode voltage, the discharge current and total discharge luminosity in the visible spectrum region were simultaneously recorded by two oscillographs C8-40. The current signal was taken from resistor R3 (50 Ω, Fig. 1) and the voltage one was taken from resistor R2 (40 kΩ, Fig. 1). The photomultiplier FEU-171 was used to convert light intensity to electric signal. (c) Fig. 10. (Color online) Waveforms of the interelectrode voltage (a), discharge current (b) and visible light intensity (c) at different currents: 1–0.3 mA, 2–1 mA and 3–3 mA. Gap is 0.7 mm, ballast resistor is 650 kΩ. In Figure 10, the waveforms of voltage, current and light intensity are shown for the three current values 0.3, 1 and 3 mA. The interelectrode voltage waveform (Fig. 10a) is saw-toothed with a slow voltage rise and a sharp decay. At a current of 0.3 mA, the frequency is about 25 kHz. The peak voltage is about 340 V and the minimum is at roughly 120 V. As the current increases up to 1 mA, the peak voltage decreases by about 40 V and the sawtooth frequency increases up to about 125 kHz. At the same time, the minimum voltage remains practically constant. This voltage value, in our opinion, is defined by the cathode fall. Upon further current increase (more than 3 mA), the discharge has a typical structure with normal current density in the cathode region and no oscillations are observed. In this case, the voltage waveform is a horizontal line which indicates that interelectrode voltage is constant. Current waveforms corresponding to these voltage ones are presented in Figure 10b. As can be seen, there are current surges corresponding to the voltage decays in oscillating discharge regimes. The peak current values of these surges are about 5–7 mA. Between two consecutive current surges a background current of about 1 mA flows. At discharge currents of more than 3 mA, the currents are constant. The average discharge glow intensity observed in images (Fig. 9) in oscillating regimes, is probably determined by the discharge light intensity at the surge moments, because light intensities between two consecutive light surges are very weak (Fig. 10c). Naturally, the intensity is constant at high current, when oscillations are V.I. Arkhipenko et al.: DC atmospheric pressure glow microdischarges in the current range... 461 (a) (b) Fig. 12. (Color online) Axial gas temperature profiles in the cathode region at different discharge currents and a discharge gap of 0.7 mm. (c) Fig. 11. (Color online) Waveforms of interelectrode voltage and visible light intensity of helium APGD: gap is 0.7 mm, ballast resistor is 10 MΩ, discharge current is 0.15 mA. absent. Since surge frequency increases with average current and the surge peak current changes insignificantly, we observe an increase in the discharge glow without changes in their characteristic regions (Fig. 9b). Therefore, using the surge current value and transverse dimension of the negative glow in corresponding images in Figure 9b it is possible to determine the current density on the cathode. Obtained values are close to normal current density for helium-copper which equals about 2.3 A/cm2 . This means that helium discharge at small currents in the fluctuating regime has normal current density on the cathode. Due to the narrow glow column in this case, maybe, its luminosity is determined by electrical breakdown like spark or streamer channels. Let us estimate the breakdown voltage, for example, at an average discharge current of 0.3 mA. The sawtooth peak voltage is about 340 V and the gap is 0.7 mm. Thus, the electric field strength is about 5 kV/cm. It is, of course, less than the corresponding breakdown for helium. However, we should take into account the fact that a weak plasma exists between breakdowns and an increase of its density, probably, leads to the decrease of the breakdown voltage (Fig. 10). Upon further increase of the power supply output voltage, the discharges transfer into spark regimes. Electrical parameters in this experiment were as follows: average current was about 0.15 mA, the sawtooth peak voltage is about 360 V, power supply voltage was about 2600 V and the ballast resistor was equal to 10 MΩ. As can be seen in Figure 11, the frequency of saw-tooth voltage oscillations is about 12 kHz. The voltage drops practically to zero. Simultaneously, at the breakdown moment, the high surges of both the light intensity and the current are observed. At the same time, no light intensity and no current are registered in the period between two consecutive surges. Oscillations similar to those registered in helium discharge (Figs. 10 and 11) are observed in glow discharges in other gases at small currents as well. However, their waveforms have some differences from corresponding helium discharge. 3.3.4 Gas temperature The gas temperature Tg was determined using the resolved 2 + 2 + rotational bands of nitrogen ions (0, 1) N+ 2 (B Σu –X Σg ) 2 + 2 and hydroxyl radical (0, 0) OH (A Σ –X Πi ). The methods of gas temperature determination using both rotational bands are described in detail in [21]. They are based on the measurements of the relative rotational line intensities of corresponding bands of nitrogen ions and hydroxyl under the assumptions that the rotational level populations of excited electronic states follow a Boltzmann distribution and that the rotational temperature Trot corresponds to the gas temperature Tg . The axial temperature profiles in the cathode region at different discharge currents are shown in Figure 12. The gap was 0.7 mm in this experiment. It is observed that the gas temperature is maximal in the cathode region only at a discharge current of 1 and 3 mA. At a discharge current of 10 mA gas heating becomes appreciable in the region adjoining the anode. In this case we have practically constant temperature on the discharge axis. The gas temperature at the end of the cathode fall doesn’t practically change. Temperature increase close to the anode is observed at a discharge current of more than 10 mA. This leads to the temperature increase in the cathode fall at discharge currents above 40 mA. Figure 13 shows the axial gas temperature profiles at discharge currents of 10 mA and 5 mA at gaps of 0.7 and 5 mm. As can be seen, the maximal gas temperature (∼700 K) at a discharge current of 10 mA and a gap of 5 mm takes place in the positive column. Gas temperature close to the cathode is about 540 K. The same occurs at a discharge current of 5 mA. At the same time, at discharge current of 5 mA the positive column temperature 462 The European Physical Journal D Fig. 13. (Color online) Axial gas temperature profiles at discharge currents of 10 mA at gaps of 0.7 mm (circles), 5 mm (trianges) and 5 mA at a gap of 5 mm (snowflakes). decreases to about 580 K. Since the positive column temperatures for both cases are higher than in the cathode region, this means that a heat flux from the positive column to the cathode region exists. According to the two slanting dotted lines in Figure 13, the temperature gradient is about 2000 K/cm for these cases. Thus, the transition from discharge with large interelectrode gap to microdischarge leads to the reduction of the heat flux from the positive column to the cathode fall region. 4 Discussion The classification of gas discharges at low pressure is performed according to their V -I characteristics in wellknown books and reviews [26–30]. In these references one can find a statement that normal glow discharges exist only at low and intermediate pressures and in a discharge current range from hundreds of microamps up to half an ampere. In Figure 4, the APGDs characteristics in helium, argon, nitrogen and air for a gap of 1 mm are shown. As can be seen, the normal glow discharges at atmospheric pressure exist from a low current of several milliamps up to about 10 A. Unlike normal low pressure glow discharges (LPGDs), the V -I characteristics of the APGDs are dropping. This can be explained by the fact that the near-electrode regions occupy only a fraction of the 1 mm gap. The pd product for typical normal LPGD varies from unity up to tens of torr cm. If we recalculate these values to atmospheric pressure, we obtain an interelectrode gap of less than 1 mm. According to [30], the transition from normal glow discharge to subnormal one is observed when the transverse dimension of the cathode fall becomes comparable to its thickness. The parts of the V -I characteristics at low currents shown in Figure 4 are characterized by a sharper voltage increase with discharge current decrease below about 1 mA. Let’s define the low current limits, at which normal APGDs takes place. This can be done using the following expression: In-sn = jc πd2c , (1) where jc – is the normal current density and dc – is the cathode fall thickness. By substituting the corresponding values of jc and dc for helium APGD in (1), which according to [19] are equal to 2.3 A/cm2 and 0.06 mm, we will obtain a low-current limit equal to about 0.25 mA. If we suppose that the cathode fall thicknesses for argon, nitrogen and air APGDs are the same as in the case of helium discharge we obtain the following current limits: 0.6, 1 and 1.2 mA, respectively. If we look at Figure 4, the current values corresponding to abrupt rises of interelectrode voltage with current decrease occur at discharge currents of about 0.3. 0.6, 1.5 and 2 mA. There is good agreement between calculated and experimental data. These low-current limits do not depend on gas density within the scaling laws, because, according to these laws, the normal current density is proportional to the gas density squared and the cathode fall thickness is inversely proportional to the gas density. The low-current limits of normal APGDs existence are shifted to higher current values in comparison with the LPGD ones. According to [19], this can be related to scaling laws nonfulfilment at the transition from low pressure up to atmospheric pressure due to changes in elementary processes. When discharge current decreases below the estimated low-current limits of the normal glow discharge the following behaviour of the APGD is observed depending on the ballast resistor value. At small ballast resistance (less than 100 kΩ) the APGD blows out. In this case, the power supply output voltage is insufficient to produce the breakdown of an actual interelectrode gap. If we apply larger ballast resistance (about 800 kΩ, for example) an increase in the power supply voltage is needed in order to reach the same current obtained at lower ballast resistance. In this case, the discharge is oscillating. There are transitions between two normal glow discharges with currents differing by a factor of 5–10. The oscillation frequency depends on the capacitance of the power supply leads and the interelectrode gap, discharge current, ballast resistor and etc. The oscillation dynamics is defined by a hysteresis of the V -I characteristic like the one shown in Figure 9. The voltage waveforms are sawtooth. A breakdown occurs at a voltage significantly lower than the corresponding one for gas breakdown in an actual gap. The reason for breakdown, probably, is connected with ionization instability. The breakdown voltage decreases while the oscillation frequency increases. This happens, in our opinion, due to discharge relaxation between breakdowns developing with discharge current decrease while voltage is being increased. In the case when the ballast resistance is equal to a few megaohms the power supply output voltage increases higher than the breakdown one for an actual interelectrode gap. Due to large resistance the power supply can not provide a current higher than the low-current limit for normal APGD existence. The voltage decreases to zero. The breakdown has a spark character, because the spark hissing is heard. Between two breakdowns, no discharge current and no glow are registered. V.I. Arkhipenko et al.: DC atmospheric pressure glow microdischarges in the current range... If a positive column exists in discharge, the heat flux from the positive column into the cathode region takes place even at a discharge current of a few milliamps. Additional heat flux in the cathode region leads to changes in both the gas temperature in the cathode region and the interelectrode voltage. 5 Conclusions Comprehensive electrical and spectroscopic investigations of the DC glow microdischarges (interelectrode gaps less than 1 mm) in atmospheric pressure helium, argon, nitrogen and air over a large current range from microamps up to ten amps were performed. Our attention in these investigations was focused mainly on the characterization of the APGD peculiarities at gaps less than 1 mm. In this case, the electrodes are located a short distance apart and influence the normal glow discharge structure and its electrical and optical parameters. It was shown that the positive columns in all gases used in this investigation are constricted. Two oscillation (or pulsing) discharge regimes were revealed in the discharge current range comparable or less than the lower current limit of normal glow discharge. One regime is defined by oscillation between two normal glow discharges with currents differing by a large factor. The reason for these oscillations, probably, is connected with ionization instability. The oscillation frequency depends on the capacitance of the power supply leads, interelectrode gap, discharge current, ballast resistor, etc. In the second regime the fluctuations at extremely low current are caused by periodical breakdown of the interelectrode gap when the power supply voltage exceeds the breakdown voltage of an actual gap. The heat flux from the positive column into the cathode region takes place even at a discharge current of a few milliamps. Additional heat flux in the cathode region leads to changes in both the gas temperature in the cathode region and the interelectrode voltage. One should take into account an additional heat flux in the cathode region in simulation of microdischarges at atmospheric pressure at a discharge current of a few milliamps when the discharge gap exceeds the cathode fall layer thickness a few times over. This work was partially supported by the Belarusian Republican Foundation for Fundamental Research under grants F09F-006 and F09-076. References 1. K.H. Becker, K.H. Schoenbach, J.G. Eden, J. Phys. D 39, R55 (2006) 2. R.H. Stark, K.H. Schoenbach, Appl. Phys. Lett. 74, 3770 (1999) 3. R.H. Stark, K.H. Schoenbach, J. Appl. Phys. 85, 2075 (1999) 4. R.M. Sankaran, K.P Giapis, Appl. Phys. Lett. 79, 593 (2001) 463 5. R.M. Sankaran, K.P. Giapis, J. Phys. D 36, 2914 (2003) 6. D. Staack, B. Farouk, A.F. Gutsol, A.A. Fridman, Plasma Source. Sci. Technol. 14, 700 (2005) 7. D. Staack, B. Farouk, A.F. Gutsol, A.A. Fridman, Plasma Source. Sci. Technol. 17, 025013 (2008) 8. Q. Wang, I. Koleva, V.M. Donnelly, D.J. Economou, J. Phys. D 38, 1690 (2005) 9. E.E. Kunhardt, IEEE Trans. Plasma Sci. 28, 189 (2000) 10. G. Bauville, B. Lacour, L. Magne, V. Puech, J.P. Boeuf, E. Munoz-Serrano, L.C. Pitchford, Appl. Phys. Lett. 90, 031501 (2007) 11. V.I. Arkhipenko, A.A. Kirillov, Th. Callegari, Ya.A. Safronau, L.V. Simonchik, IEEE Trans. Plasma Sci. 37, 740 (2009) 12. V.I. Arkhipenko, Th. Callegari, Ya.A. Safronau, L.V. Simonchik, IEEE Trans. Plasma Sci. 37, 1297 (2009) 13. G. Fridman, G. Friedman, A. Gutsol, A.B. Shekhter, V.N. Vasilets, A. Fridman, Plasma Process. Polym. 5, 503 (2008) 14. H. Qui, K. Martus, W.Y. Lee, K. Becker, Intern. J. Mass Spectrom. 233, 19 (2004) 15. V.I. Arkhipenko, A.A. Kirillov, L.V. Simonchik, S.M. Zgirouski, in Proceedings of 19th International Symposium on Plasma Chemistry (ISPC-19) (Bochum, 2009) 16. O. Sakai, T. Sakaguchi, Y. Ito, K. Tachibana, Plasma Phys. Control. Fusion 47, B617 (2005) 17. J. Lo, Th. Callegari, J. Sokolov, J.P. Boeuf, in Contributed papers of VI International Conference Plasma Physics and Plasma Technology, Minsk, edited by V.M. Astashynski, V.S. Burakov, L.E. Kratko (Polifank, Minsk, 2009), Vol. 1, p. 268 18. V.I. Arkhipenko, A.A. Kirillov, Ya.A. Safronau, L.V. Simonchik, S.M. Zgirouski, Plasma Source. Sci. Technol. 18, 045013 (2009) 19. V.I. Arkhipenko, S.M. Zgirovskii, A.A. Kirillov, L.V. Simonchik, Plasma Phys. Rep. 28, 858 (2002) 20. V.I. Arkhipenko, A.A. Kirillov, L.V. Simonchik, S.M. Zgirouski, Plasma Source. Sci. Technol. 14, 757 (2005) 21. V.I. Arkhipenko, S.M. Zgirouski, A.A. Kirillov, L.V. Simonchik, in Spectroscopy of Plasma and Natural Objects, edited by V.I. Arkhipenko, V.S. Burakov, A.F. Cherniavskiy (Belaruskaja Navuka, Minsk, 2007), Vol. 10 (in Russian) 22. V.I. Arkhipenko, A.A. Kirillov, Ya.A. Safronau, L.V. Simonchik, S.M. Zgirouski, Plasma Source. Sci. Technol. 17, 045017 (2008) 23. Z.Lj. Petrovic, A.V. Phelps, Phys. Rev. E 47, 2806 (1993) 24. Z.Lj. Petrovic, A.V. Phelps, Phys. Rev. E 56, 5920 (1997) 25. Yu. Akishev, M. Grushin, V. Karalnik, A. Petryakov, N. Trushkin, J. Phys. D 43, 075202 (2010) 26. G. Francis, in Handbuch der Physik, edited by S. Flugge (Berlin, 1956), Vol. 22, p. 53 27. Y.P. Raizer, in Gas Discharge Physics (Springer, New York, 1991) 28. A.V. Phelps, Plasma Source. Sci. Technol. 10, 329 (2001) 29. B.N. Klyaiferd, L.G. Guseva, A.S. Pokrovskaya-Soboleva, Zh. Tekh. Fiz. 36, 704 (1966) (in Russian) 30. A. von Engel, M. Steenbeck, in Elektrische Gasentladungen (Springer, Berlin, 1934), Vol. 2