Distributions of Plasma Parameters, Electron Energy Characteristics and LF Wave Activity in a Magnetized ICP

V. M. Slobodyan, V. F. Virko, K. P. Shamrai and G. S. Kirichenko
Institute for Nuclear Research NAS of Ukraine, 47 Prospect Nauki, 03680 Kiev, Ukraine
Abstract. We report the results of probe measurements of stationary plasma parameters, electron energy characteristics, and wave activity in a magnetized inductively coupled plasma (ICP). The discharge regimes are shown to change with abrupt jumps at increasing magnetic field and to depend strongly on magnetic field shape and antenna design. Probe characteristics, space and floating potentials, and energy distributions reveal a population of fast electrons. Spectra and spatial distributions of low-frequency waves depend on profiles of static plasma parameters and magnetic configuration.
Keywords: Inductively coupled plasma, abrupt density jumps, electron energy distribution, low-frequency turbulence.
PACS: 52.35.Hr, 52.40.Fd, 52.50.Dg.
1. INTRODUCTION
Inductively coupled plasmas (ICPs) have being developed as advanced tools for various applications. These devices, even without an external magnetic field, demonstrate intricate physics originating from numerous cooperative phenomena (e.g., [1]). Applying the magnetic field considerably enhances the efficiency of plasma generation and, on the other hand, diversifies greatly physical processes due to occurrence of waves [2-6]. We performed correlative measurements of plasma parameters, electron energy distributions, and wave characteristics in order to determine the effect of magnetic configuration on efficiency of plasma generation and performance of the ICP discharge and to reveal the driving mechanisms of LF wave activity that is inherent to plasma. We also examined the effect of antenna design (outer diameter and number of turns) on plasma characteristics.
2. EXPERIMENTAL DEVICE AND DIAGNOSTICS
The source consists of a 20-cm-diameter, 30-cm-long cylindrical stainless-steel discharge chamber limited by a substrate table from below and by a 22-mm-thick quartz window from above and equipped by two probe ports (Fig. 1). To excite the discharge in Ar, we normally used a 17.5-cm-diameter single-loop (m = 0) flat antenna, which was put 6 mm above the window and supplied from an rf generator of frequency 13.56 MHz and power up to 2.5 kW. Some experiments were performed with smaller antennas, an 11.5-cm-diameter double-turn antenna and a 6.5-cm-diameter four-turn one. Three magnetic coils, the upper, the medium, and the lower (not shown in Fig. 1) ones with separately controlled currents (Iup, Imid та Ilow, respectively) can produce the magnetic field of varying strengths and shape. Just the magnetic configuration, which depends mostly on Iup, was found to control significantly plasma parameters and wave processes. Axial profiles of the magnetic field are shown in Fig. 2 for fixed values of Imid and Ilow and various Iup. Plasma parameters and the rf and LF oscillations were measured by electric, thermo-emissive and magnetic probes, and by the probe biased with modulated potential. Correlation of the LF oscillations was measured by the double probe that was located at a radius of 6 cm and could rotate around its axis.
3. CHARACTERIZATION OF PLASMA PARAMETERS
The discharge demonstrates the variety of regimes depending on Ar pressure (pAr), rf power (Pforw), and, substantially, on the magnetic configuration. To ascertain the effect of the latter factor, we performed experiments at Pforw = 750 W and pAr = 3.8 mTorr and with fixed Imid = 0.3 A and Ilow = 2 A, but varying Iup = (0−400) mA. Axial profiles of the magnetic field are shown in Fig. 2, for various values of Iup. Considering nonuniformity of the total magnetic field, further results are shown as dependent on Iup.
Strong dependence of the discharge regimes on magnetic configuration is displayed, particularly, in abrupt jumps of the ion saturation current at continuous variation of Iup (Fig. 3). Characteristic density jumps are measured over the whole plasma cross-section thus evidencing global changes of the discharge performance. Simultaneously with the jumps, radial plasma profiles alter considerably, as seen from Fig. 4. At small Iup (≤ 30 mA), the discharge intensity is relatively low and the density profiles are nearly flat, with considerable density gradient only near the sidewall. With increasing Iup, the discharge intensity grows, the density profiles become more and more peaked, and the range of considerable density gradient spreads over the whole cross-section. In this regime, the discharge demonstrates bright blue core near the axis; this regime is similar to the so-called “blue mode” of the normal helicon discharge, but does not need very high power or magnetic field. On approaching to some critical values of Iup, the discharge regime jumps with abrupt density falls. After the strongest jump, at Iup ≈ 200 mA, the density profile becomes slightly hollow. With further increase of Iup, the profile is still hollow but is steeper at the periphery. Finally, the discharge disruption occurs on reaching the critical current Iup ≈ 360 mA. Overall discharge intensity is found to grow with use of the smaller antennas, the 11.5-cm-diameter double-turn antenna and the 6.5-cm-diameter four-turn one. In particular, at the same input power the ion saturation current increases along with critical values of Iup at which density jumps and the discharge disruption occur.
The probe floating potential Vf is negative practically in all the discharge regimes (Fig. 5). Its radial profile has on-axis minimum in low-intense discharge regimes, and off-axis one in intense regimes. In some conditions, the floating potential on the axis can be as low as −40 V. Typical semi-logarithmic probe characteristic shown in Fig. 6 demonstrates two groups of electrons with different temperatures; it differs from the probe characteristic for normal two-temperature electron distribution [7] in that the lower (higher) temperature corresponds to more (less) energetic electron distribution function has “bump-on-tail”. Then, the lower temperature relates to the beam while the higher one is the “effective” temperature in the region between the beam and the “main body” of electrons. The beam temperature is 2.5− 4.5 eV and depends only slightly on radius while the temperature of the main body is uncertain.
To determine the electron energy distribution function, we applied the low-frequency modulation to bias potential (so called the second-derivative method). Measurements do confirm that the electron distribution involves a population of fast particles with approximately isotropic distribution on velocities. Mean energy of this population grows with increasing magnetic field and input power and can be as large as 40 eV. The density of fast population can amount to 1% of the mean plasma density.
To determine the plasma potential Vp, we used the thermo-emissive probe whose characteristics in cold and heated regimes are shown in Fig. 7. Radial profiles of the plasma potential are shown in Fig. 8. Considerable variation of Vp evidences that quite strong radial electric fields (up to 5 V⋅cm−1) exist in plasma. One can see from comparison of Figs. 5 and 8 that difference between the floating and plasma potentials can be as large as 70 V on the axis, which is the further confirmation of non-equilibrium electrons. Indeed, the relation Vp ≈ Vf + 5Te, which is valid for equilibrium (maxwellian) plasma [8], is not satisfied under any reasonable assumption regarding Te.
4. CHARACTERIZATION OF LF TURBULENCE
LF wave activity was found to arise in plasma in all the discharge regimes. Its characteristics, as well as plasma parameters, depend strongly on magnetic configuration. LF spectra are shown in Fig. 9 for various values of Iup. At small Iup = 23 mA, the oscillations has narrow spectrum with maximum around 650 kHz and are localized at the periphery of the discharge column, where the density gradient is quite strong (cf. Figs. 4 and 10). Around the location of LF noise the dc radial electric field, as well as the rf field [6], is small. Thus, these oscillations are, most probably, driven by the azimuthal diamagnetic electron drift current. Measurements with the double probe show that LF oscillations are well correlated and propagate azimuthally with the ion-acoustic velocity.
With increase of Iup and related increase of the discharge intensity, LF noise arises over the whole cross-section, grows considerably in amplitude, and shows continuous and broad spectrum of width 1−1.5 MHz (see Figs. 9 and 10). Correlation of oscillations falls sharply, and wave propagation direction becomes uncertain. Finally, at Iup > 200.

5. DISCUSSION AND CONCLUSIONS
Abrupt changes in the discharge regimes at continuous variation of magnetic configuration arise, most probably, from transitions between helicon eigenmodes excited in plasma (see [6,7]). With varying Iup, the magnetic field changes considerably near the quartz window while only slightly in the remote region where the measurements were conducted. Thus, the discharge performance is essentially controlled by the rf power input under the antenna.
Substantial fall of the floating potential at the discharge center is thought to result from population of fast electrons that was detected in measurements of temperature and electron distribution function. This population can be generated due to stochastic acceleration in wave fields.
LF activity is intrinsic for plasma in any regime. At small Iup, LF waves have spiky spectrum, are localized around maximum of the radial density gradient, and are apparently driven by azimuthal diamagnetic electron drift. At larger Iup, the LF noise is broadband and much more intense and is apparently driven by electric electron drift and/or parametric instability. LF wave activity was not revealed to contribute considerably in plasma generation, as long as its intensity is stronger at larger Iup, i.e., when the plasma density is reduced.
ACKNOWLEDGMENTS
This work was supported by the Science and Technology Center in Ukraine under contract No. 3068.
REFERENCES
1. V. A. Godyak, Plasma Phys. Control. Fusion 45, A399 (2003).
2. J. E. Stevens, M. J. Sowa and J. L. Cecchi, J. Vac. Sci. Techhnol. A 13, 2476 (1995).
3. S. Shinohara, S. Takechi, Y. Kawai, Jpn. J. Appl. Phys. 35, 4503 (1996).
4. S. S. Kim, C. S. Chang, N. S. Yoon and K. W. Whang, Phys. Plasmas 6, 2926 (1999).
5. V. M. Slobodyan, V. F. Virko, G. S. Kirichenko and K. P. Shamrai, Problems of Atomic Sci. Technol. #4, 235 (2003).
6. K. P. Shamrai, S. Shinohara, V. F. Virko, V. M. Slobodyan, Yu. V. Virko and G. S. Kirichenko, Plasma Phys. Control.
Fusion 47, A307 (2005).
7. F. F. Chen, Introduction to Plasma Physics and Controlled Fusion, 2nd ed., Vol. 1 (Plenum, NY, 1984).
8. F. F. Chen, Lecture notes on Langmuir probe diagnostics. IEEE ICOPS Mini-Course on Plasma Diagnostics (Jeju, Korea, 2003).

Опубликовано в рубрике Documents