Technical Physics, Vol. 49, No. 1, 2004, pp. 104–107. Translated from Zhurnal Tekhnicheskoœ Fiziki, Vol. 74, No. 1, 2004, pp. 104–107. Original Russian Text Copyright © 2004 by Burdovitsin, Burachevsky, Oks, Fedorov.

ELECTRON AND ION BEAMS, ACCELERATORS

Specific Features of the Formation of a Uniform Ribbon Electron Beam by a Plasma Source in the Forevacuum Pressure Range V. A. Burdovitsin, Yu. A. Burachevsky, E. M. Oks, and M. V. Fedorov Tomsk State University of Control Systems and Radio Electronics, Tomsk, 634050 Russia e-mail: [email protected] Received February 21, 2003

Abstract—Results are presented from experimental studies of the formation of a ribbon electron beam during the extraction of electrons from the plasma of a steady-state hollow-cathode discharge in the forevacuum pressure range. It is shown that the main reason for the nonuniformity of the current density is the increase in the local nonuniformity of the emission plasma density caused by the return flow of ions from the accelerating gap. Taking this feature into account when developing a system of beam extraction provides for the generation of a ribbon beam with a nonuniformity of the current density along the beam of less than 10%. © 2004 MAIK “Nauka/Interperiodica”.

INTRODUCTION One of the promising applications of ribbon electron beams is the formation of large-area (up to 1 m2) “plasma sheets” that can be used in various technological processes (such as plasmochemical and ion etching and the deposition of various coatings in the course of decomposition and fusion reactions in the generated plasma) and as moving microwave mirrors [1]. A fairly high gas pressure (10–100 Pa) is required to generate such a plasma. As a result, it is very difficult to use sources with a thermionic cathode for this purpose, so that there is, in fact, no alternative to the use of plasma electron sources based on the extraction of electrons from the plasma of low-pressure discharges with nonincandescent electrodes [2–4]. In forming a ribbon electron beam, one of the most important problems is that of attaining a highly uniform current density. This problem was considered by Bugaev et al. [5] who analyzed the main reasons for the nonuniformity of the emission current in large-cross-section beams and suggested ways of eliminating these reasons. At the same time, Bugaev et al. [5] largely treated “standard” plasma sources of electrons, whose working pressure range was, as a rule, below 0.1 Pa. At such pressures, the degree of uniformity of the current density of an electron beam extracted from the plasma is largely determined by the uniformity of the emission plasma. The transition to the forevacuum pressure range results in an increase in the effect of the return flow of ions formed in the accelerating gap and the electron-beam transport region on the emissive properties of the plasma [6]. Therefore, for plasma sources of electrons operating in the forevacuum pressure range, it seems insufficient to achieve the initial uniformity of the emission plasma. It was shown in [6, 7] that an increase in

the working pressure makes it necessary to take into account the ionization processes in the accelerating gap and attendant phenomena. In this paper, we describe the results of investigations of the formation of a ribbon electron beam with a highly uniform current density in a forevacuum plasma source of electrons based on a discharge with an extended hollow cathode [8]. EXPERIMENTAL SETUP The experimental forevacuum plasma source of electrons for the generation of a ribbon beam comprised the same basic elements as the source of a cylindrical electron beam described by us in [6], namely, a hollow cathode, a flat anode with an emission opening, an accelerating electrode, and a collector. Rectangular hollow cathode 1 (Fig. 1) 300 × 80 × 40 mm in size maintained a steady-state discharge with a current of up to 1.5 A. The beam size was determined by a 250 × 10 mm emission slot in anode 2. The slot was overlapped by fine-mesh metal grid 3. Cathode 1, anode 2, and accelerating electrode 4 were electrically separated from one another by caprolan insulators 5 and 6. The parameters of the emitting plasma were measured with cylindrical probes 7 introduced into the plasma via channels in insulator 5. The probes were arranged so that fast ions from the accelerating gap could not fall on their collecting surfaces. The working gas was air. Figure 1 also shows how the sources of the discharge and accelerating voltages (Ud aqd Ua, respectively) were connected to the electrodes. The electron current distribution along the electron beam was measured using movable molybdenum collector 8 located behind a grounded grid having a slot with a width of 1 mm and

1063-7842/04/4901-0104$26.00 © 2004 MAIK “Nauka/Interperiodica”

SPECIFIC FEATURES OF THE FORMATION

length exceeding the beam size. The distance from the emission grid to the collector was 15 cm.

1

EXPERIMENTAL RESULTS In order to determine the distribution of the plasma density along the hollow cathode, the currents to probes 7 (Fig. 1) were measured in the ion segment of the current–voltage characteristic. In the absence of electron emission, the nonuniformity of the plasma density along the hollow does not exceed 5–10%, except for the density maxima at the edges. At the same time, the electron beam extracted when applying the accelerating voltage is significantly nonuniform and, as visual observations demonstrate, turns out to consist of at least ten fine beams (jets). The distribution of the current density i(x) at the movable collector under different experimental conditions is given in Fig. 2. In this case, a much smaller number of experimentally recorded maxima is due to the effect of individual jets in the region where the distribution of the electron beam current is measured. A decrease in the pressure resulted in the disappearance of the beam nonuniformities. A significant decrease in nonuniformities was also observed when the grid cell size was reduced. We performed a special experiment with a composite grid in order to more clearly establish a correlation between the beam current density and the density of the emitting plasma. The middle 6-cm-long part of the composite grid was a grid with a 0.8 × 0.8 mm mesh, and the remainder was a grid with a 0.4 × 0.4 mm mesh. The effect of the electron emission on the plasma density distribution in the hollow for this situation is illustrated by Fig. 3. Figure 4 gives the corresponding distributions of the current density i(x) along the beam. It can be seen that, in the absence of emission, the nonuniformity of the plasma density along the hollow does not exceed 20%. At the same time, the electron emission results in a several-fold increase in the plasma density in the middle part of the hollow. The results presented in this figure clearly demonstrate the agreement between the positions of the maximum of the beam current density and the maximum of the plasma density in the hollow. We also note that, as the pressure decreases, the nonuniformities of the beam current density and the plasma density are smoothed out.

5

ANALYSIS OF THE RESULTS The basic experimental results can be formulated as follows. In the absence of electron emission, the nonuniformity of the plasma density along the hollow does not exceed 10%. The extraction of electrons from the plasma at elevated pressures leads to the emergence of more significant nonuniformities in both the plasma and the electron beam. In this case, the spatial positions of the maxima of the emission current density and the density of the emitting plasma coincide. The nonuniformity of the current density of the electron beam exceeds TECHNICAL PHYSICS

Vol. 49

No. 1

2004

105 –Ud

7

3

2

+Ud –Ua

6

4

x

8

Fig. 1. Schematic of an electron source.

1 2 3

i, A/m 4 3 2 1 0 0

5

10

15

20

25

30 x, cm

Fig. 2. Distribution of the electron current along the beam for pressures of p = (1, 3) 4 and (2) 1 Pa and grid cell sizes of (1, 2) 0.8 × 0.8 and (3) 0.4 × 0.4 mm.

J, A/m2 8 1 2

6 4 2 0 0

5

10

15

20

25

30 x, cm

Fig. 3. Distribution of the probe current density in the anode plasma for accelerating voltages of Ua = (1) 0 and (2) 3 kV at a pressure of p = 4 Pa.

BURDOVITSIN et al.

106 i, A/m 8

i, arb. units

6

1 2 3

2.5

1 2

2.0 4 1.5 2 1.0 –0.06 –0.04 –0.02

0 0

5

10

15

20

25

30 x, cm

Fig. 4. Distribution of the current density along the beam for pressures of p = (1) 4 and (2) 2.4 Pa at Ua = 3 kV.

the corresponding nonuniformity of the plasma density. The effect of emission on the nonuniformity of the current and plasma densities significantly reduces as the gas pressure and the size of the emission hole (grid cell) decrease. Based on the experimental results, we can assume the following mechanism for increasing the nonuniformity of the electron emission current. In the initial stage of the extraction of electrons from the plasma, the nonuniformity of the current density is mainly due to the nonuniformity of the plasma density in the hollow. The current density distribution may also be affected by the nonuniformity of the emission grid, namely, by the differences in the local curvature and the scatter in the sizes of elementary cells. The ionization of the residual gas in the accelerating gap and in the region of electron beam transport, which is significant in the forevacuum pressure range, leads to the emergence of a significant return flow of ions. Since the ionization rate is proportional to the electron current density, the density profile of the return ion flow must correspond to the initial distribution of the current density of the electrons emitted by the plasma. Fast ions, which get into the plasma and exchange charges with gas molecules, bring with them a positive space charge, which is neutralized by the plasma electrons. This results in a local increase in the nonuniformity of the plasma density and in the corresponding increase in the nonuniformity of the emission current. The increase in the emission current density with increasing plasma density is also due to the increase in the area of the open plasma surface within each cell of the anode grid because of the narrowing of the space charge layer separating the plasma from the grid. Therefore, a minor local variation in the plasma density results in a disproportionate increase in the local density of the electron emission current. The return ion flow associated with the electron current causes a further local increase in the plasma density and the corresponding further disproportionate increase in the density of the electron emission current

0

0.02

0.04

0.06 x, cm

Fig. 5. Calculated dependence of the electron current density on the x coordinate in the perturbation region for pressures of p = (1, 2) 6.6 and (3) 4 Pa and grid cell sizes of h = (1, 3) 0.8 × 0.8 and (2) 0.4 × 0.4 mm.

at this site. This positive feedback reaches saturation and the plasma density ceases to increase when the formation rate of slow ions is balanced by their diffusion from the perturbation region. In order to qualitatively estimate the possibility of the existence of a local maximum in accordance with the above mechanism, we will write balance equations for slow ions formed in the cathode hollow due to charge exchange of fast ions arriving at the plasma from the accelerating gap. The production of ions is balanced by their departure from the perturbed plasma region due to diffusion. In the one-dimensional case, the balance equation has the following form: l

1 8kT e 2 --- -----------n n Q e Q r d exp ( – Q r n n y ) dy 4 πm

∫ 0

(1)

Xb

×

n( x ) ∫ n ( x ) dx = – D d------------dx

l,

i

– Xb

Xb

where Di is the ionic diffusion coefficient; n(x) is the plasma density in the perturbation region; ±Xb are the coordinates of the boundaries of the perturbation region; dn(x)/d x X b is the gradient of the plasma density at the edge of the perturbation region; l is the depth of the hollow, which is less than the mean free path of charge-exchange ions, whereby a one-dimensional model may be employed; nn is the density of neutral particles; Qe is the effective cross section for the ionization of gas particles by electrons; d is the accelerating gap length; and Qr is the effective cross section for the charge exchange of ions. In order to take into account the variation in the area of the emitting surface, the coefficient K(x) is introTECHNICAL PHYSICS

Vol. 49

No. 1

2004

SPECIFIC FEATURES OF THE FORMATION

duced in the left-hand side of Eq. (1), ( h – 2l s ( x ) ) -, K ( x ) = ----------------------------2 h 2

(2)

where h is the inner size of the grid cell, ls(x) = –1

2 ε 0 U ( n ( x ) ekT e ) is the thickness of the ion layer separating the plasma from the grid [2], and U is the plasma potential relative to the anode. It was assumed in the calculations that the dependence of the plasma density on the x coordinate within the perturbation region has the form of a Gaussian function. It follows from experiment that the value of Xb remains constant. The value of dn(x)/d x X b is determined from the condition of equality of the perturbed (n) and unperturbed (n0) densities at x = Xb. The numerical results presented in Fig. 5 demonstrate a local increase in the emission current density i(x) with increasing both gas pressure and cell size of the emission grid. Therefore, the results of our calculations qualitatively confirm the possibility of the existence of local nonuniformities of the emission current due to the above physical mechanism. 3/2

CONCLUSIONS A specific feature of the formation of a ribbon electron beam by a plasma source in the forevacuum pressure range is the high probability of the emergence of a nonuniformity in the distribution of the emission current along the beam. The results of our investigations indicate that this nonuniformity is caused by the return ion flow from the accelerating gap into the emitting plasma. The positive feedback that arises between the

TECHNICAL PHYSICS

Vol. 49

No. 1

2004

107

electron emission current and the return ion flow causes a disproportionate increase in the primary nonuniformity of the beam due to local nonuniformities of the plasma density, as well as to the local bending of the cells and difference in their sizes. The effect of the return ion flow may be reduced, e.g., by its defocusing accomplished by varying the grid shape in the emission electrode. This provides the generation of a ribbon electron beam with a nonuniformity of 10% or less. The investigation results provide one with more assurance in developing plasma electron sources generating electron beams in the forevacuum pressure range. REFERENCES 1. W. M. Manheimer, R. F. Fersner, M. Lampe, et al., Plasma Sources Sci. Technol. 9, 370 (2000). 2. Yu. E. Kreœndel’, Plasma Electron Sources (Atomizdat, Moscow, 1977). 3. M. A. Zav’yalov, Yu. E. Kreœndel’, A. A. Novikov, and L. P. Shanturin, Plasma Processes in Electron Guns (Énergoatomizdat, Moscow, 1989). 4. E. M. Oks and P. M. Schanin, Phys. Plasmas 6, 1649 (1999). 5. S. P. Bugaev, Yu. E. Kreœndel’, and P. M. Shchanin, Wide Electron Beams (Énergoatomizdat, Moscow, 1984). 6. Yu. A. Burachevskiœ, V. A. Burdovitsin, E. M. Oks, et al., Zh. Tekh. Fiz. 71 (2), 48 (2001) [Tech. Phys. 46, 179 (2001)]. 7. V. A. Burdovitsin, M. N. Kuzemchenko, and E. M. Oks, Zh. Tekh. Fiz. 72 (7), 134 (2002) [Tech. Phys. 47, 926 (2002)]. 8. V. A. Burdovitsin, Yu. A. Burachevskiœ, E. M. Oks, et al., Prib. Tekh. Éksp., No. 2, 1 (2003).

Translated by A. Bronshteœn