Low-pressure Inductive Discharge Research Articles

Instabilities of Ar/SF6 inductive plasma discharges

Relaxation oscillations are studied in three low-pressure inductive plasma discharges operated with argon and sulfur hexafluoride gas mixtures. Two distinct phenomena, downstream instabilities and source oscillations, occur in certain domains of gas pressure, radio frequency power, and electronegative gas chemistry. The downstream instabilities develop at some location well below the plasma source. They are consistent with ion two-stream instabilities, in inductively coupled plasmas (ICPs) with sufficiently long downstream regions. Source oscillations consist of large amplitude density variations within the ICP plasma. They are consistent with capacitive to inductive mode transitions, in ICPs with sufficiently large capacitive currents.

Dynamics of steady and unsteady operation of inductive discharges with attaching gases

Relaxation oscillations in charged particle densities are seen in low-pressure inductive discharges, if attaching gases such as SF6 and Ar/SF6 mixtures are used, such that the plasma contains negative ions. These oscillations occur in the neighborhood of the transition between lower power capacitive operation and higher power inductive operation. An instability, which occurs around this transition, can be described by a global model, that consists of time-dependent equations for the electrons, ions, and electron temperature. The model qualitatively agrees with experimental observations, but leaves significant quantitative differences to be explained. To understand the reasons for the model’s inability to make quantitative predictions, the local conditions for instability are determined for a model somewhat simplified from the one that approximates the experiment, but retains most of the important physics. The simplified model allows explicit determination for the local transition to instability and identifies two parameters that govern the transition. The local criteria are related to the global criterion that all fixed points are unstable. The importance for the instability of the physical parameters of the ion wall loss and of the ratio of capacitive to inductive power transfer has led to a reexamination of the size of these factors, and gives better quantitative agreement of the model to experiments.

Downstream instabilities of electronegative plasma discharges

Relaxation oscillations are observed in low-pressure inductive plasma discharges operated with argon and SF6 gas mixtures. The data suggest that instabilities may develop periodically at some downstream location. Linear fluid and kinetic calculations predict ion two-stream instabilities when the positive and negative ions develop sufficiently large drift velocities of opposite sign in the downstream region. The thresholds and the growth rates of the instabilities are calculated and are found in fair agreement with the observations. The instability models can explain several key experimental features.

Radiation transport coupled particle-in-cell simulation of low-pressure inductive discharges

Low pressure (1–5 Torr) argon discharges driven by an inductive radio frequency wave are simulated with a one-dimensional radiation transport coupled particle-in-cell model. The discharge is maintained by an induced azimuthal electric field which is self-consistently coupled with plasma dynamics. The radiation efficiency is investigated for the variations of input power, gas pressure, and cylinder radius, and compared with that of positive column discharges. The radiation efficiency is improved up to 8% compared with that of conventional positive column discharges by virtue of reduced radiation trapping resulting from enhancement of excitation collisions near the wall for inductive discharges.

Experimental study of diffusive cooling of electrons in a pulsed inductively coupled plasma.

Langmuir probe measurements of the temporal behavior of the electron distribution function in a low-pressure inductive discharge are presented. The structure of the measured distribution functions suggests that the loss of high energetic electrons to the wall of the discharge chamber is the main energy loss mechanism. Electron-heavy-particle collisions play only a secondary role for the energy loss. The rapid loss of energetic electrons--while low energy electrons remain confined in the space charge potential field--leads to a fast cooling of the electron distribution function. We also present a simple model to describe the evolution of the mean kinetic energy and plasma potential on the basis of a distribution function that is cutoff at energies above the potential electron energy at the wall.

SiO2 film deposition in a low-pressure RF inductive discharge SiH4 + O2 plasma

SiO2 films were deposited in an rf inductive discharge plasma (13.56 MHz) at a substrate temperature of-550 K and an ion energy of 35 eV. Variable parameters were silane and oxygen flow rates, pressure (0.3–1.2 Pa), and ion current density (1–2.5 mA/cm2). In these ranges of the process parameters, good silicon dioxide films with a composition close to that of thermally oxidized SiO2 were obtained. The film growth rate was found to be 65 nm/min. The deposition nonuniformity was <4% over 100-mm silicon wafers.

Instabilities in low-pressure inductive discharges with attaching gases

Plasma instabilities at frequencies 1 Hz–900 kHz have been observed in low-pressure inductive processing discharges with attaching gases. Instability windows in pressure and driving power are found. A volume-averaged (global) model of the instability is developed, considering idealized inductive and capacitive energy deposition. As pressure or power are varied to cross a threshold, the instability is born at a Hopf bifurcation, with relaxation oscillations between inductive and capacitive modes causing modulations of charged particle densities, electron temperature, and plasma potential. The oscillations can be so strong that the potential collapses and negative ions flow to the walls.

Tailoring of electron energy distributions in low-pressure inductive discharges

Assuming a background Maxwellian electron distribution, the effects of adding an energetic tail upon the electron temperature and density in an inductive low-pressure argon discharge are investigated theoretically. The principal effect of the tail is to modify the rate constants for ionization, excitation, and momentum transfer. Using particle and energy conservation, the possibility of suppressing the electron temperature and increasing the electron density is demonstrated. It is found that under typical conditions of operation the electron temperature can be decreased from 2.5 to 1.5 eV, and the electron density increased by at least a factor of 2.4.

Effect of Collisionless Heating on Electron Energy Distribution in an Inductively Coupled Plasma

Significant frequency dependence of the electron energy distribution has been found in a low-pressure inductive discharge experiment. The observed frequency dependence reveals the presence of collisionless electron heating and appears as a result of finite electron-transit time through the skin layer. The energy distributions calculated from a kinetic equation accounting for nonlocality of electron kinetics and electrodynamics are in good agreement with the experiment.

Low-pressure inductive rf discharge in a rare gas-halogen mixture for economical mercury-free luminescence light sources

It is demonstrated that an inductive rf discharge in an Xe+Cl2 mixture may be used as the active medium of an efficient, mercury-free luminescence light source.

Kinetic modeling of the charging of nonconducting walls in a low pressure radio frequency inductively coupled plasma

This article investigates the overall charging of a nonconducting, plane wall (for instance a wafer) in a low pressure inductively coupled plasma. The problem is addressed using a two-dimensional kinetic model for a low pressure inductive discharge. Comparisons to experimental results show good agreement with the charging profiles predicted by the model. It is pointed out that the surface charge profile on a nonconducting wall is determined by the plasma homogeneity and the high energy part of the electron distribution function. An interpretation of the radial profiles of the sheath potential drop and of the surface charge potential in terms of the differential temperature of the electron distribution function in different energy ranges is presented.

The electric field and current density in a low-pressure inductive discharge measured with different B-dot probes

The magnitude and relative phase of the time varying magnetic field in an inductively coupled discharge have been measured with two dB/dt (B-dot) probes: One B-dot probe was enclosed by a dielectric tube (as is commonly used in dB/dt measurements) while the other, a thin wire probe, was immersed directly into the plasma. Each probe was used to measure the radial and axial component of dB/dt. A comparison of rf electric fields and currents obtained by the two probes showed essentially different results. The disagreement is interpreted to be due to a large local disturbance of the plasma density and current caused by the dielectric tube.

Formation of XeI(B) in low pressure inductive radio frequency electric discharges sustained in mixtures of Xe and I2

Low pressure excimer discharges (&amp;lt;1–5 Torr) are of interest for use as ultraviolet sources for lighting applications. XeI(B) is an attractive candidate for excimer lamps due to the low corrosive properties of iodine. The excimer is thought to be formed by either harpoon or ion-ion recombination reactions, the latter of which requires a third body. The formation of the excimer at low pressures is therefore problematic. To address this issue, an investigation was conducted to determine the kinetic processes which produce XeI(B) in a low pressure (0.5–5 Torr) inductive radio frequency discharge sustained in Xe and I2. The diagnostics applied in this study include laser-induced fluorescence, optical absorption spectroscopy, and optical emission spectroscopy. Our results indicate that for the experimental conditions, Xe+I2**→XeI(B)+I is a major reaction producing the excimer. This result differs from studies performed at higher pressures which concluded that the harpoon reaction between Xe* and I2 or ionic recombination between Xe+2 and I− are the major sources of XeI(B).

Magnetic field distribution measurements in a low-pressure inductive discharge

The magnitude and relative phase of the rf magnetic flux density has been measured as a function of radial position using a magnetic (B-dot) probe in an inductively coupled low pressure argon discharge driven at 13.56 MHz. The spatial variation of the rf electric field and the discharge current density were determined from the probe measurements and are compared at gas pressures of 3, 30, and 300 mTorr for a fixed discharge power of 50 W and at discharge powers of 21, 50, and 103 W for a fixed gas pressure of 30 mTorr. In the active zone near the induction coil where the rf field is strong, the electron rf drift velocity derived from the B-dot data and Langmuir probe measurement was found to be considerably less than the electron thermal velocity even at the lowest argon pressure. For an axially symmetric inductive discharge, as is found in many applications, a two-dimensional magnetic probe measurement with space and phase resolution is necessary to correctly infer the rf electric field and the discharge current density distribution.

A biased optical probe method for measurements of electron energy distribution in a plasma

A biased optical probe (BOP) method, i.e. novel technique for measurements of the electron energy distribution function (EEDF) in a range of such high energies as 10-50 eV has been developed. This method is based on the optical emission induced by electrons passing through a negatively biased mesh in a plasma. The BOP method was used to measure the EEDF in a DC glow discharge in He/Xe mixture while the low-energy part (<10 eV) of the EEDF was measured using the conventional Druyvesteyn method. The measured distribution function considerably deviates from an ideal Maxwellian distribution, because of abundant high-energy electrons generated in a cathode fall. On the contrary, the EEDF measured in a low-pressure inductive RF discharge was found to follow a single Maxwellian distribution over a wide energy range.