ISSN 1063-7842, Technical Physics, 2008, Vol. 53, No. 4, pp. 432–435. © Pleiades Publishing, Ltd., 2008. Original Russian Text © A.S. Klimov, Yu.A. Burachevsky, V.A. Burdovitsyn, E.M. Oks, 2008, published in Zhurnal Tekhnicheskoœ Fiziki, 2008, Vol. 78, No. 4, pp. 43–46.

GAS DISCHARGES, PLASMA

Inhomogeneous Extended Hollow Cathode Discharge for Raising the Current Density in a Forevacuum Plasma Source of a Ribbon Electron Beam A. S. Klimov, Yu. A. Burachevsky, V. A. Burdovitsyn, and E. M. Oks Tomsk State University of Control Systems and Radio Electronics, pr. Lenina 40, Tomsk, 634050 Russia e-mail: [email protected] Received May 21, 2007

Abstract—The plasma parameters and the emissivity of a ribbon electron beam source based on a discharge with an inhomogeneous extended hollow cathode are measured. A constriction in the cathode cavity increases the plasma density near the emitting area boundary, which adds to the electron current density in the beam. The reason for the above effect is the formation of the plasma density distribution nonuniform across the cavity with a maximum in the middle. This maximum is caused primarily by a plasma electron flow from the constriction, which is generated by the electric field and is directed toward a slit emission-extracting aperture. PACS numbers: 52.25.Tx DOI: 10.1134/S1063784208040063

INTRODUCTION Equipment designed for electron-beam processing of large areas needs beams with a large cross section [1, 2], such as ribbon beams [3, 4]. Of special interest among appropriate sources are those generating a beam under a reduced pressure provided by mechanical fore pumps [5, 6]. The plasma source of electrons developed in our laboratory on the basis of an extended hollow cathode discharge can generate steady ribbon-shaped (25 × 1-cm) electron beams with an energy of up to 8 keV and a current density to 15–20 mA/cm2 at a pressure of 1–10 Pa [7]. Despite the parameters of the ribbon electron beam generated under forevacuum conditions being unique, the beam current densities achieved are nevertheless insufficient for a number of critical applications. Efforts to raise the beam current density by merely narrowing the cathode cavity or its exit aperture have failed because of the heavily nonuniform current distribution over the beam’s cross section [8, 9]. Therefore, the problem of increasing the current density in the electron beam generated under forevacuum conditions seems to be of both fundamental and applied importance. The aim of this work is to study the feasibility of raising the current density in the ribbon electron beam under forevacuum conditions by refining the cathode cavity geometry.

low cathode 1 made of stainless steel (width d of its upper half can be varied with copper inserts 2), plane anode 3 with a 280 × 10-mm emission-extracting window, and insulators 4 and 5 to fix the electrodes. The inserts in the cathode cavity were electrically insulated from the walls, so that currents passing through the narrow and remaining parts of the cavity can be measured separately. Discharge, Ud, and accelerating, Ua, voltages were applied to the respective electrodes of the 1 2 6

Ud

Z 5 d 3

Ua

4 7 Y 8

V

X

Up

V

EXPERIMENTAL Experiments were conducted on a model plasma source of a continuous ribbon electron beam that was designed for operation under forevacuum conditions (Fig. 1). The source consists of 280 × 75 × 40-mm hol432

Fig. 1. Design of the electron source: 1, hollow cathode; 2, inserts; 3, anode; 4 and 5, insulators; 6, probe; 7, accelerating electrode; and 8, movable collector.

INHOMOGENEOUS EXTENDED HOLLOW CATHODE DISCHARGE

source as shown in Fig. 1 (for details concerning the design and parameters of the source, see [7]). Plasma density n in the cathode cavity was measured by moving Langmuir probe 6, potential Up of which was set by a stand-alone power source. The design of the probe made it possible to measure the plasma parameters along the width of the cathode at different distances z from the slit aperture. The probe measured plasma potential ϕ in the floating mode [10]. During the measurements, the grid was removed from the extracting window and voltage Ua was not applied to accelerating electrode 7. The current density distribution in the beam was measured by movable collector 8. The vacuum chamber was evacuated using an AVZ-20 mechanical fore pump. The pressure in the chamber was varied in the interval 3–10 Pa by directly delivering air.

433

Ud, V 700

1 500

4 3

300

2 0

400

800 Id, mA

Fig. 2. I–V characteristic of the discharge for constriction width d = (1) 40, (2) 12, (3) 14, and (4) 16 mm. P = 6 Pa.

EXPERIMENTAL RESULTS The experiments showed that the properties of the inhomogeneous-hollow-cathode discharge differ from those of the discharge initiated with a homogeneous hollow cathode (i.e., a cathode having a rectangular cross section). Distinctions show up in both the I–V characteristics of the discharges and the parameters of the discharge plasma. The I–V curve of the homogeneous-cathode discharge is monotonic (Fig. 2, curve 1), while that of the inhomogeneous-cathode discharge has a region where the discharge voltage drops stepwise and the discharge current rises (Fig. 2; curves 2, 3). The plasma density in the symmetry plane totally duplicates the behavior of the current. Simultaneously with the step increase in the current and plasma density, an increased luminosity of the plasma is visually observed within the narrow part. The threshold current initiating the step growth depends on the gas pressure and width of the constriction. The lower the pressure and the narrower the constriction, the higher the threshold current. The ratio of currents In and Iw (Fig. 3) passing through the narrow and wide parts suggests that the step of the current is due to its redistribution, a larger fraction being accounted for by the narrow part. The spatial distribution of the plasma density, n(x), was measured at different values of coordinate z (Fig. 4). For the homogeneous cathode, curve n(x) is smooth (curve 1). For the inhomogeneous one, the shape of this curve taken in the wide part depends on the discharge current. For currents much lower than the threshold value, curve n(x) has the same shape as for the homogeneous cavity. Above-threshold discharge currents lead to z-dependent distributions. At large z, i.e., in the immediate vicinity of the boundary between the two parts of the cavity, the curve n(x) exhibits a clear-cut peak roughly narrow part wide (curve 2). As z decreases, the peak lowers and expands (curve 3). The plasma peak density exceeds the density in the homogeneous cavity by 1.5– 2.0 times at the same values of the discharge current. The increase in the plasma density leads to an associated increase in the current density in the electron TECHNICAL PHYSICS

Vol. 53

No. 4

2008

In, Iw 0.8 2 0.6

0.4 1 0.2 200

400

600

800 Id, mA

Fig. 3. Ratio of the currents passing through the wide (Iw) (curve 1) and narrow (In) (curve 2) parts of the cathode cavity to discharge current Id. P = 5 Pa and d = 16 mm.

n, 1016 m–3 12 2 3

8

4

1

0 0

1

2

3

4 x, cm

Fig. 4. Transverse distribution of plasma density n in the (1) homogeneous cavity and inhomogeneous cavity at depths z = (2) 3 and (3) 2 cm. Id = 800 mA, d = 16 mm, and P = 6 Pa.

434

KLIMOV et al. j, mA/cm2

ϕ, V 0

n, 1016 m–3 4

30

10

–4 1

4

20

–8 6

3 10 1 0

2

2

0

0.8

1.6 x, cm

0

Fig. 5. Distribution of current density j in the beam for discharge current Id = (1, 2) 100 and (3, 4) 800 mA in the (1, 3) homogeneous and (2, 4) inhomogeneous cathode cavities with d = 16 mm. P = 6 Pa and Ua = 2 kV.

DISCUSSION The results obtained can be explained under the assumption that the discharge in the inhomogeneous cathode cavity burns in two modes. At low currents and, accordingly, low plasma densities, the discharge is initiated only in the wide part of the cavity, since cathodic layers cover the constriction and the plasma cannot penetrate into the narrow part. As the discharge current is raised via external control, cathodic sheaths ultimately break and the plasma penetrates into the constriction. The penetration conditions can be written as [12] 1/2

3/4

1/4

4 z, cm

6

8

Fig. 6. Plasma (1, 3) concentration n and (2, 4) potential ϕ vs. coordinate z in the (1, 2) inhomogeneous (d = 16 mm) and (3, 4) homogeneous cavities. Id = 800 mA and P = 6 Pa.

beam extracted from the discharge plasma (Fig. 5). As follows from Fig. 5, the current density in the beam rises in the same proportion as the plasma density and reaches 35 mA/cm2. To gain insight into physical reasons for the increase in the plasma density, we made probe measurements along the z axis in the symmetry plane of the cathode cavity (Fig. 6) and found that plasma density n rises noticeably with penetration depth into the narrow part of the cavity. The plasma potential in the constriction is much lower than in the rest of the cavity. A local stepwise increase in the potential, an attribute of the double electrical layer at the boundary of the constriction [11], was not observed. The decrease in the potential in the narrow part with increasing coordinate z was fairly smooth and roughly equaled 10 V. The potential drop calculated for the discharge system of an electron plasma source based on a hollow-cathode reflection-mode discharge was the same [12].

d/2 > l c = ( ε 0 /n ) ( U c ) / ( ekT e ) ,

2

–12

2

3

(1)

where d is the width of the narrow part, lc is the cathode layer extent, n and Te are the density and electron temperature of the plasma, and Uc is the cathode potential

drop. From the measured parameters of the plasma and discharge as applied to the threshold current, we find that lc ≈ 0.5 cm, which meets well condition (1). The stepwise growth of the current, along with the decrease in the discharge voltage (Fig. 2), as well the redistribution of the current among the parts of the cavity (Fig. 3) in such a way that a major part of the cathode current passes through the narrow part, is unambiguous evidence that ionization in a discharge system with an inhomogeneous hollow cathode is more efficient. This result seems to be somewhat unexpected if one keeps in mind that, when the width of the entire cavity equals that of the narrow part, the discharge cannot be sustained because of the nonuniform plasma distribution along the cavity in the working pressure range [8]. As a major fraction of the current switches over to the narrow part, the plasma density within this part rises. The discharge current redistribution among the two parts of the cavity can be explained using the following line of reasoning. Let us compare free path length λ of γ electrons with the width of the cavity. For the pressures used (3–10 Pa), λ = 1–3 cm. This means that electron oscillations are weak in a cavity 4 cm wide and, correspondingly, the hollow cathode effect does not show up in full measure. As the cavity shrinks to ≈1.5 cm, conditions for oscillation apparently become more favorable and electron–ion pairs may arise in the cathode layer [14]. Since the effects discussed are observed in the inhomogeneous cavity alone, which is in essence a combination of narrow and wide parts, one can suppose that these two discharge regions interact with each other, sustaining the discharge and effective ionization. The appreciable drop of the potential in the narrow part of the cathode provides diffusion and/or drift of plasma electrons toward the wide part. This additional electron flow stabilizes the discharge in the wide part, as in the case of electron injection from the outside of the cathode region of the low-pressure glow discharge TECHNICAL PHYSICS

Vol. 53

No. 4

2008

INHOMOGENEOUS EXTENDED HOLLOW CATHODE DISCHARGE

[15]. Also, the electron flow from the narrow part leads to a plasma redistribution in the wide part, giving rise to a peak in the curve n(x) at large z and causing this peak to expand with decreasing z. Next, comparing the diffusion, jd, and drift, jf, currents given by dn j d = – D ------ , dz

dϕ j f = – µn -----dz

(2)

(D and µ are the diffusion coefficient and mobility of ions, respectively), one sees that the drift flux of ions toward the narrow part is roughly two orders of magnitude higher than the diffusion counterflux. Hence, the ions moving from the wide to the narrow part diminish the ionization rate in the wide part but favor ionization in the narrow part by means of γ electron emission from the surface of the cathode. This may be one more reason why the current switches over to the narrow part. CONCLUSIONS The discharge system with an inhomogeneous extended hollow cathode provides conditions for more effective ionization of the gas and, thereby, an increase in the plasma density in the middle of the cathode cavity. As a result, the current density in the ribbon-shaped electron beam generated by a plasma source of electrons under forevacuum conditions grows by 1.5–2.0 times. It seems that the beam current density achieved is not ultimate. It can be raised further by optimizing the geometry of the inhomogeneous extended cathode cavity and discharge parameters. ACKNOWLEDGMENTS This work was supported by the Russian Foundation by Basic Research, grant nos. 05-02-98000 and 05-0801319.

TECHNICAL PHYSICS

Vol. 53

No. 4

2008

435

REFERENCES 1. S. P. Bugaev, Yu. E. Kreindel’, and P. M. Shchanin, Electron Beams of Large Cross Section (Energoatomizdat, Moscow, 1984), p. 10 [in Russian]. 2. N. N. Koval, E. M. Oks, Yu. E. Kreindel, et al., Nucl. Instrum. Methods Phys. Res. A 321, 417 (1992). 3. S. W. A. Giclens, P. J. M. Peters, W. J. Witterman, et al., Rev. Sci. Instrum. 67, 2449 (1996). 4. V. V. Osipov, N. V. Gavrilov, O. A. Bureyev, et al., Laser Phys. 16, 68 (2006). 5. D. Leonhardt, S. G. Walton, C. Muratore, et al., J. Vac. Sci. Technol. A 22, 2276 (2004). 6. V. A. Burdovitsin, E. M. Oks, and M. V. Fedorov, Izv. Vyssh. Uchebn. Zaved., Fiz., No. 3, 74 (2004). 7. V. A. Burdovitsin, Yu. A. Burachevskii, E. M. Oks, et al., Prib. Tekh. Eksp., No. 2, 127 (2003). 8. Yu. A. Burachevsky, V. A. Burdovitsin, A. S. Klimov, et al., Zh. Tekh. Fiz. 76 (10), 62 (2006) [Tech. Phys. 51, 1316 (2006)]. 9. V. A. Burdovitsin, Yu. A. Burachevskii, E. M. Oks, et al., Zh. Tekh. Fiz. 74 (1), 104 (2004) [Tech. Phys. 49, 104 (2004)]. 10. R. H. Huddlestone and S. L. Leonard, Plasma Diagnostic Techniques (Academic, New York, 1965; Mir, Moscow, 1967). 11. A. S. Metel’, Zh. Tekh. Fiz. 54 (2), 241 (1984) [Sov. Phys. Tech. Phys. 29, 141 (1984)]. 12. V. A. Burdovitsin, V. L. Galanskii, V. A. Gruzdev, et al., Zh. Tekh. Fiz. 63 (1), 184 (1993) [Tech. Phys. 38, 46 (1993)]. 13. Yu. E. Kreindel’, Plasma Electron Sources (Atomizdat, Moscow, 1977), p. 44 [in Russian]. 14. A. S. Metel’ and A. I. Nastyukha, Izv. Vyssh. Uchebn. Zaved., Radiofiz. 19, 1884 (1976). 15. A. V. Vizir’, E. M. Oks, M. V. Shandrikov, and G. Yu. Yushkov, Prikl. Fiz., No. 6, 115 (2004).

Translated by V. Isaakyan