ISSN 10637842, Technical Physics, 2015, Vol. 60, No. 2, pp. 213–216. © Pleiades Publishing, Ltd., 2015. Original Russian Text © A.V. Kazakov, A.V. Medovnik, V.A. Burdovitsin, E.M. Oks, 2015, published in Zhurnal Tekhnicheskoi Fiziki, 2015, Vol. 85, No. 2, pp. 55–58.

PLASMA

Behavior of an Arc Discharge in a Forevacuum Plasma Source of Electrons A. V. Kazakov*a, A. V. Medovnika, V. A. Burdovitsina, and E. M. Oksa,b a

Tomsk State University of Control Systems and Radio Electronics, pr. Lenina 40, Tomsk, 634050 Russia b Institute of HighCurrent Electronics, Siberian Branch, Russian Academy of Sciences, Akademicheskii pr. 2/3, Tomsk, 634055 Russia *email: [email protected] Received December 13, 2013; in final form, May 22, 2014

Abstract—The parameters and characteristics of a pulsed arc discharge with a cathode spot used in a fore vacuum plasma source of electrons are investigated. It is shown that the accelerating voltage influences arc initiation to a lesser extent compared with an electron source based on a hollowcathode glow discharge. At forevacuum pressures, two stages of the arc discharge may arise within the current pulse. At the beginning of the pulse (first stage), the arcing voltage is high and the fraction of residual gas ions in the plasma is signif icant. At the second stage, the arcing voltage drops and ions of the cathode material dominate in the plasma. The duration of the first stage grows with rising gas pressure and decreasing arc current. DOI: 10.1134/S1063784215020103

INTRODUCTION Forevacuum plasma sources of electrons effectively generate continuous and pulsed electron beams in the pressure range 1–100 Pa [1–4]. When these devices operate in the pulsed mode, a hollowcathode glow discharge [3] or an arc discharge with cathode spot [4] are used. Using the arc, designers tend to raise the electron beam current. It is known [5] that a plasma forming medium in an arc discharge results from the evaporation and ionization of the material in the cath ode spot. In addition, a plasma is generated in a small nearcathode region of the discharge. In this region, the partial pressure of cathode metal vapor far exceeds the residual pressure of the gas. Nevertheless, a rise in the pressure in the discharge gap of the arc from the purely “vacuum” case (10–4 Pa) to the “gas” case (10–1 Pa) influences the parameters of the arc plasma (primarily its mass and charge compositions), causing a sharp reduction of the fraction of multiply charged cathode material ions [6] and gas ionization [7]. Obvi ously, when an arc is initiated at higher pressures in the forevacuum range, the gas may have a more consider able effect on the arc parameters, and hence, on the emissivity of the arc plasma. In this work, the behavior of an arc discharge in the forevacuum range of pressures, more specifically, in a plasma source of electrons, is studied.

and hollow cylindrical anode 2 100 mm in length and 80 mm in diameter. Both electrodes are made of cop per. The cathode is put in ceramic tube 3, which limits its working area by the end face. An emission aperture 90 mm in diameter is provided on the base of the cylindrical anode opposite to the cathode. This aper ture is covered by a fine grid made of stainless steel. An arc was initiated between cathode 1 and anode 2 using an auxiliary discharge on the ceramic surface, which separates cathode 1 and ignitor electrode 4. The igni tor electrode is connected to the anode through a resistor, so that the same pulsed power supply 5 can be used for arc initiation and arc current maintenance

EXPERIMENTAL The electrode system of a pulsed forevacuum plasma electron source based on an arc discharge (Fig. 1) includes rodlike cathode 1 with a diameter of 5 mm 213

1

5

3

Id to oscilloscope 100 Ω

4 2

Ud 8 Ua

7 11 6

9

10

Fig. 1. Electron source and parameter measurement cir cuit: (1) cathode, (2) anode, (3) ceramic insulator, (4) ignitor electrode, (5) pulsed power supply of discharge, (6) accel erating electrode (extractor), (7) Caprolon insulator, (8) highvoltage power supply, (9) optical probe, (10) opti cal fiber, and (11) spectrometer.

214

KAZAKOV et al. (a) 2

20 15

1

−40

10

−60

5

−80

0

(a)

600 τ, μs

−20

750

Id, A

Ud, V

0

25

2

450 300

1 200 300 400 500 600 t, μs

15

(b) 2

0 Ud, V

−15 Ud2 −30 −45

1

Ud1

−60 −75 −90

τ1 0

100

τ2

70 60 50 40 30 20 10 0

0 10

20

30

40

50

60

70

P, Pa 500 (b) 400 τ, μs

100

Id, A

0

150

300

2 1

200

200 300 400 500 600 t, μs 100

Fig. 2. Waveforms of the (1) arcing voltage pulse and (2) arc current pulse for Id = (a) 20 and (b) 60 A. P = 30 Pa (air).

(the socalled triggerless mode [8]). Before arcing, when the arc current is zero, anode–cathode voltage Ud might reach 1000 V. After arc initiation, it dropped to the arc voltage (several tens of volts). In experiments, the amplitude and duration of the arc current pulse were varied within 20–60 A and 300–700 μs, respectively. In all experiments, except for special experiments in which the parameters and characteristics of the discharge were studied as func tions of pulse repetition rate ν, ν was kept equal to 1 pps. The extraction of plasma electrons to form an electron beam was accomplished by applying constant voltage Ud = 1–12 kV between anode 2 and accelerat ing electrode (extractor) 6. An electron source was mounted on the flange of the vacuum chamber, which was evacuated by a mechanical pump with an evacua tion rate of 11 L/s. The pressure in the range 4–70 Pa was controlled by direct supply of a working gas (argon or air) to the chamber. The design of the electron source, as well as its parameters and characteristics, is detailed elsewhere [4]. Discharge current Id and the arc voltage were mea sured using a current transformer (Rogowski loop) with a sensitivity of 50 A/V and an HVP15HF resis tive voltage divider with a division ratio of 1 : 1000.

20

30

40 Id, A

50

60

Fig. 3. Times (1) τ1 and (2) τ2 vs. (a) pressure P (Id = 30 A) and (b) discharge current Id (P = 30 Pa). Air is used as a working gas.

The optical spectrum of the discharge glow was taken with optical probe 9. A signal was passed to Ocean Optics 2000USB spectrometer 11 with the wavelength interval 200–1100 nm through optical fiber 10. When taking the spectra, the accelerating electrode was removed and the optical probe was placed at a distance of 15 mm from the axis of the cathode. The configuration when the optical probe was placed strictly opposite to the cathode proved to be inefficient because of illumination from the cathode spot the intensity of which far exceeded the intensity of the plasma glow. The emission lines observed in the experiment were identified according to the technique described in [9, 10]. EXPERIMENTAL DATA It was shown [1] that the key point in the operation of forevacuum plasma sources of electrons is that a backward ion flux from the region of electron beam acceleration and transport influences the discharge initiation and maintenance conditions, plasma emis sivity, and ultimate parameters of the beam. Backward TECHNICAL PHYSICS

Vol. 60

No. 2

2015

BEHAVIOR OF AN ARC DISCHARGE Cu

I, arb. units

500

N O Ar

215

(a)

400 Ar

300 200 100 0

200

300

400

500

600

700

800

I, arb. units

(a) 400 350 300 250 200 150 100

Cu

(b)

O Ar

(b)

Fig. 5. Cathode spot tracks at Id = 60 A, τ = 300 μs, and P = (a) 50 (stage 1) and (b) 4 Pa (stage 2). Air is used as a working gas.

N

200

300

400

500 600 λ, nm

700

800

Fig. 4. Optical spectra of the discharge radiation. (a) P = 50 Pa (stage 1), Id = 60 A and (b) P = 4 Pa (stage 2), Id = 60 A. Argon and residual air are used as working gases.

ions are generated in the beam plasma. Another con tributor to the ion flux, along with the beam plasma, is the nearanode plasma of a “parasitic” highvoltage glow discharge (HVGD). This discharge stably persists in the accelerating gap at a high forevacuum pressure. Even in the absence of the electron beam, the back ward ion flux of the HVGD is significant [11]. Our pre vious observation that the initiation voltage of a hol lowcathode glow discharge in a forevacuum plasma source decreases after the application of the accelerat ing voltage [12] is explained exactly by the influence of this ion flux. Accelerating voltage Ua and its respective HVGD have an influence on arcing as well. However, com pared with the hollowcathode glow discharge, the accelerating voltage governing the HVGD current influences the arcing voltage to a lesser extent. In the case of the hollowcathode discharge, almost all ions of the HVGD penetrating the discharge system fall into the cathode region and, knocking out extra addi tional electrons from its surface, provide a twofold decrease in the discharge voltage [12]. In the case of the arc, only a small fraction of HVGD ions (those trapped by the end face of the cathode) favor arcing. That is exactly why the application of the accelerating voltage decreases the arcing voltage in the discharge cell of the electron source by no more than 10–15%. A change in the residual gas pressure also slightly influences the arc initiation in the forevacuum plasma source of electrons. Specifically, when the pressure rose from 0.01 to 4 Pa, the arcing voltage dropped by no more than 50–100 V. The rise time of the arc cur rent is influenced by the gas pressure to a much greater TECHNICAL PHYSICS

Vol. 60

No. 2

2015

extent. Experiments showed that in the same pressure range, the leading edge of the current pulse shrinks about twofold from 9–6 to 5–3 μs in the arc current range 10–30 A. The same influence of the pressure on the setting time of the pulsed arc current is described in [13]. In that work, the arc current rise time decreased from 1.0 to 0.4 μs when the pressure grew from 10–3 to 10–2 Pa. The significant difference between the current rise times may be associated with the difference in the geometries of the discharge sys tems and operating features of pulsed power supplies maintaining the arc. If the pulse of arc current Id is almost rectangular, two discharge modes (stages) can be distinguished in the waveform of arcing voltage Ud in which the values of Ud differ (Fig. 2). When the first (initial) stage with voltage Ud1 changes to the second stage with voltage Ud2, the arcing voltage markedly drops. Duration τ1 of the fist stage and time instant τ2 of the onset of the sec ond stage depend on the gas pressure and discharge current. As the pressure grows and the discharge cur rent diminishes, times τ1 and τ2 grow (Fig. 3), whereas the time interval between the first and second stages (τ2–τ1) either slightly extends (Fig. 3a) or remains the same (Fig. 3b). At a pressure above 15 Pa and an arc current below 20 A, the transition from first to second stage was not observed in the experiment. However, at a pressure below 10 Pa, the transition to the second stage takes place at the leading edge of the pulse at any arc current. In both arcing modes, the arcing voltage changes little with arc current and gas pressure. It was shown [14] that when an arc discharge was initiated at a gas pressure of about 0.1 Pa, the rise in the pulse repetition rate had a considerable effect on the arc plasma parameters. In particular, the fraction of gas ions in the plasma drops because of a higher rate of gas desorption from the cathode surface in this case. The same influence of the pulse repetition rate could also be expected for the arc discharge at higher fore vacuum pressures. Experiments showed, however, that the variation of the pulse rate in the interval 0.2– 25 pps has a negligible effect both on the duration of the first stage and on the time of the firsttosecond

216

KAZAKOV et al.

stage transition. It seems that more detailed investiga tions are necessary to explain such differences in the influence of the pulse rate. When taking the spectra of optical radiation from the arc plasma, we studied two extreme cases. (i) High pressure (50 Pa), when the arc is at the first stage with a higher arcing voltage throughout the pulse and (ii) low pressure (4 Pa), when the arc almost imme diately passes to the second stage with a lower arcing voltage. The spectra of optical radiation from the arc plasma taken at different pressures, gases, and currents are depicted in Fig. 4. In the first extreme case or at stage 1 (Fig. 4a), the intensity of gas spectral lines exceeds that of cathode material (copper) lines. In the second limiting case or at stage 2, the situation reverses. In the plasma glow spectrum, copper lines dominate and their intensity is several times higher than the intensity of residual gas lines (Fig. 4b). The tracks of the cathode spot on the cathode surface for these two extreme cases of arcing are shown in Fig. 5. At a pressure of 50 Pa corresponding to the first stage of arcing (Fig. 5a), the track occupies a larger surface area than at the second stage (Fig. 5b). At stage 2, the tracks of the cathode spot are deeper but shorter. CONCLUSIONS Arcing modes with different values of Ud observed experimentally can be attributed to the processes of gas desorption from the cathode surface during which both types of cathode spots are observed within one pulse [15]. At the first (initial) stage of arcing, the cathode spots function on the “contaminated” surface of the cathode. It seems that cathode spots of the first type and some spots of the second type form on the cathode surface at the first stage. At the second stage of arcing (with a lower arcing voltage), only spots of the second type exist.

ACKNOWLEDGMENTS This work was supported by the Russian Founda tion for Basic Research, grant nos. 1308000175_a and 140831075mol_a. REFERENCES 1. V. A. Burdovitsin and E. M. Oks, Laser Part. Beams 26, 619 (2008). 2. A. A. Zenin, A. S. Klimov, V. A. Burdovitsin, and E. M. Oks, Tech. Phys. Lett. 39, 454 (2013). 3. Yu. G. Yushkov, V. A. Burdovitsin, A. V. Medovnik, and E. M. Oks, Prib. Tekh. Eksp., No. 2, 85 (2011). 4. A. V. Kazakov, V. A. Burdovitsin, A. V. Medovnik, and E. M. Oks, Prib. Tekh. Eksp., No. 6, 50 (2013). 5. G. A. Mesyats, Ectons (Nauka, Yekaterinburg, 1993). 6. P. Spadtke, H. Emig, B. H. Wolf, and E. Oks, Rev. Sci. Instrum. 65, 3113 (1994). 7. A. G. Nikolaev, E. M. Oks, and G. Yu. Yushkov, Tech. Phys. 43, 1031 (1998). 8. A. Anders, I. G. Brown, R. A. Mac Gill, and M. R. Dick inson, J. Phys.: Appl. Phys. 31, 584 (1998). 9. A. R. Striganov and G. A. Odintsova, Tables of Spectral Lines of Atoms and Ions: A Handbook (Energoizdat, Moscow, 1982). 10. A. R. Striganov and N. S. Sventitskii, Tables of Spectral Lines of Neutral and Ionized Atoms (Plenum, New York, 1968). 11. A. F. Medovnik, V. A. Burdovitsin, and E. M. Oks, Izv. Vyssh. Uchebn. Zaved., Fiz. 53 (2), 27 (2010). 12. I. S. Zhirkov, V. A. Burdovitsin, E. M. Oks, and I. V. Osi pov, Tech. Phys. 51, 1379 (2006). 13. N. N. Koval, Yu. E. Kreindel, V. S. Tolkachev, and P. M. Schanin, IEEE Trans. Electr. Insul. 20, 735 (1985). 14. G. U. Yushkov and A. Anders, IEEE Trans. Plasma Sci. 26, 220 (1998). 15. A. Anders, Cathodic Arcs: From Fractal Spots to Ener getic Condensation (Springer, New York, 2008).

Translated by V. Isaakyan

TECHNICAL PHYSICS

Vol. 60

No. 2

2015