Compact Helicon Plasma Source with Permanent Magnets in Various Configurations

Yu. V. Virko, K. P. Shamrai, V. F. Virko and A. I. Yakimenko
Institute for Nuclear Research NAS of Ukraine, 47 Prospect Nauki, 03680 Kiev, Ukraine
Abstract. We report the results of measurements of plasma and emergent ion beam characteristics in a compact helicon source equipped with a changeable multi-component system of permanent magnets combined of the annular ferrite, electromagnet, cylindrical ferrite array, and iron shield. Optimization of magnetic configuration is shown to give rise to substantial growth of plasma density and to generation of emergent beam of ions accelerated up to energies exceeding 100 eV. Elimination of the magnetic cusps, the rf antenna placement in the region of lower magnetic field, and working gas pressure reduction are found to be critical factors for enhancement of the source efficiency.
Keywords: Helicon discharge, permanent magnets, emergent ion beam.
PACS: 52.35.Hr, 52.40.Fd, 52.50.Dg.
Helicon sources can serve as advanced tools for various applications, such as materials processing, decontamination of hazardous substances, plasma propulsion, etc. They normally operate with electromagnets producing relatively uniform magnetic fields. Devices with permanent magnets (PMs), which are preferred for some applications, have been also developed [1-5] but are much less examined. Magnetic fields of the PMs are strongly nonuniform and include the null points (cusps). While moderate field nonuniformity can enhance considerably the plasma generation at the same power input [6,7], strong nonuniformity is expected to hinder free wave propagation and plasma outflow and thus to reduce the source efficiency. We report the results from a compact helicon source that was equipped with a flexible, variable magnetic system in order to find optimal magnetic configurations enabling to maximize plasma production and to control efficiently the parameters of emergent ion beam.
The discharge chamber is a 4.5-cm-inner-diameter, 30-cm-long quartz tube attached to a 14.5-cm-diameter quartz drift chamber (Fig. 1). Magnetic fields of various shape were produced by the left section of an electromagnet (EM) and by a set of PMs. The basic magnetic field was created by a 13.5-cm-outer-diameter and 1.8-cm-wide annular ferrite (PM in Fig. 1) magnetized axially. An iron shield (IS) of the same size could be attached to the PM. We also used a 12-cm-long, 12-cm-inner-diameter cylindrical ferrite array (CFA) magnetized radially. To examine the effect of discharge chamber length, we used an axially movable quartz plate (QP). The discharge in Ar was excited by a triple-turn (m = 0), axially movable antenna supplied from an rf generator of frequency 13.56 MHz and power up to 1 kW. Plasma parameters and rf field characteristics were measured by the Langmuir probe, plane probe with a guard ring, and magnetic probe, all being movable axially. Emergent ion beam was examined with the five-grid retarding field energy analyzers (RFEA). It was positioned 7 cm from the discharge chamber outlet and could rotate over the angle of 90°, to face to the source outlet or to the sidewall of the drift chamber.
We used four different magnetic configurations created by (1) a single PM; (2) a combination of the PM with the EM; (3) a combination of the PM with the CFA; and (4) a combination of the PM with the adjoined.

Axial profile of the magnetic field of the single PM has two inversion points (cusps), as seen from Fig. 2(a), which both lower the source efficiency. The left cusp (between the PM and the drift chamber) impedes plasma ejection from the discharge chamber. The right cusp (between the PM and the antenna) hinders from penetration of the waves into the region of strong magnetic field. Plasma density has maximum in the discharge chamber, under the antenna [Fig. 3(a)], of the order of 2×1012 cm−3 at high Ar pressure 3−5 mTorr, and twice lower at low pressure of 0.7 mTorr. Mean electron temperature is 6−8 eV at higher pressures, and raises up to 14−18 eV at lower pressures. The RFEA characteristics are shown in Fig. 4(a); solid curves relate to the case with the analyzer facing to the source outlet, while the broken ones to the case of perpendicular facing. Plasma potential in the drift chamber is about 30 and 60 V, for Ar pressures of 5 and 0.7 mTorr, respectively. The RFEA characteristics are close for both cases, independently of pressure. Plasma flow from the source is weak and emergent beam of accelerated ions is absent.
We eliminated the cusps by adding the EM field [see Fig. 2(b)]. Plasma density was found to grow, in both discharge and drift chambers, if the antenna is put as close as possible to the PM. Further density increase was obtained by varying the EM current and by moving the EM, to create the field growing monotonically from the antenna to the PM. In final configuration, the plasma density was enhanced substantially, with maximum under the PM, near the source outlet, of the order of (4−5)×1012 and 2×1012 cm−3 at high (3−5 mTorr) and low (0.7 mTorr) Ar pressures, respectively [Fig. 3(a)]. Mean electron temperature was practically the same as with the single PM. The RFEA characteristics [Fig. 4(b)] show substantial enhancement of plasma ejection from the source. Below 2.5 mTorr, there arises the emergent beam of accelerated ions with energy growing with pressure decrease, as a result of potential growth in the discharge chamber. At 0.55 mTorr [Fig. 4(b)], the potential is ~125 V in the discharge chamber while ~70 V in the drift chamber. Net acceleration of emergent beam, about 55 eV, is produced by a steep potential drop that arises at the source outlet due to strong gradients of the magnetic field and density. However, we failed to detect in this region a structure like the double layer that was observed in other experiments [8,9].

We tried to approximate the favorable magnetic configuration with using the CFA instead of the EM. In resulting magnetic configuration shown in Fig. 2(b), the cusp at the source outlet is reduces, although not eliminated, while the cusp between the antenna and the PM is eliminated completely. The magnetic field in the discharge chamber grows from the antenna to the outlet, which is favorable for enhanced plasma production [7]. As a result, the plasma density rises, as compared with previous configurations, both near the source outlet and in the drift chamber [Fig. 3(a)]. This is apparently due to more efficient rf power deposition into the region of strong magnetic field, as seen from enhancement of the rf Bz-field under the PM and in the drift chamber [Fig. 3(b)]. Electron temperature is about 9 and 15 eV at 5 and 0.7 mTorr, respectively. The RFEA characteristics [Fig. 4(c)] demonstrate the intensity of the ion flow to enhance by 4−5 times as compared with previous configurations. However, a distinct emergent ion beam is absent, though some tail with energies in the range 40−60 eV arises at Ar pressures below 1 mTorr [Fig. 4(c)]. Suppression of beam generation is thought to result from incomplete elimination of the cusp at the source outlet.
The magnetic cusp at the source outlet can be eliminated by increasing the CFA magnetization, or by attaching the IS to the PM. As the magnetic field gradient is increased at the IS location, the potential drop at the outlet is expected to grow thus giving rise to more efficient ion acceleration. Axial profiles of the ion saturation current and the rf magnetic field for the PM+CFA+IS configuration are shown in Figs. 3(c) and 3(d), and the RFEA characteristics in Fig. 4(d). As seen, a tail of accelerated ions, with energies up to 80 eV, does become more pronounced, but its intensity falls. The reasons for insufficient efficiency of this configuration may be the following: insufficient diameter of the IS, which reduces shielding effect, and considerable power dissipation in the IS, which is detected as strong heating of the outlet flange. The way for increasing the efficiency of the IS utilization is thought in increasing its diameter, slitting it radially, and/or removing it off the antenna.
We have ascertained that optimization of the magnetic configuration enables to enhance considerably the plasma density, both in the discharge and drift chambers, and to get the emergent flux of accelerated ions. Elimination of the magnetic field cusp between the antenna and the PM, which was incarnated by combining the PM with the EM and/or CFA, leads to discharge localization in a region of strong magnetic field adjoining to the source outlet and rises maximum plasma density up to 5×1012 cm−3, at Ar pressures 3−5 mTorr and rf input power of 600 W.

Elimination of the cusp at the source outlet gives rise to generation, at reduced Ar pressures, of the emergent beam of accelerated ions with current density up to 1 mA⋅cm−2 (measured 7 cm from the outlet) and energy above 100 eV (relative to the grounded electrode). Next, we are planning to try the single CFA and to modify the field by the iron circuit.
Experiments carried out at various lengths of the discharge chamber, with use of the movable quartz plate (QP), have shown that shortening twice the discharge chamber does not alter considerably the discharge modes, including the values of the ion saturation currents and the ion flux characteristics. This is a result of directional discharge burning towards the source outlet in configurations with the magnetic field growing downstream from the antenna.
This work was supported by the Science and Technology Center in Ukraine under contract No. 3068. The authors are grateful to Prof. G. S. Kirichenko for continuous encouragement and discussions.
1. K. P. Shamrai, A. F. Aleksandrov, G. E. Bougrov, V. F. Virko, V. P. Katyukha, S. K. Koh, E. A. Kralkina, G. S. Kirichenko and A. A. Rukhadze, J. Physique Coll. IV 7, C4-365 (1997).
2. I. S. Hong, Y. S. Hwang, G. H. Lee, D. Y. Kim, H. Y. Won, G. S. Eom and W. Choe, Rev. Sci. Instrum. 71, 1385 (2000).
3. K. Sasaki, H. Kokubu, D. Hayashi and K. Kadota, Thin Solid Films 386, 243 (2001).
4. V. I. Miroshnichenko, S. N. Mordik, V. V. Olshansky K. N. Stepanov, V. E. Storizhko, B. Sulkio-Cleff and V. Voznyy, Nucl. Instr. and Meth. B 201, 630 (2003).
5. F. F. Chen, Private communication (2005).
6. G. Chevalier and F. F. Chen, J. Vac. Sci. Technol. A 11, 1165 (1993).
7. V. F. Virko, K. P. Shamrai, G. S. Kirichenko and Yu. V. Virko, Phys. Plasmas 11, 3888 (2004).
8. S. A. Cohen, N. S. Siefert, S. Stange, R. F. Boivin, E. E. Scime and F. M. Levinton, Phys. Plasmas 10, 2593 (2003).
9. C. Charles and R. W. Boswell, Phys. Plasmas 11, 17063 (2004).

Опубликовано в рубрике Documents