An overview is presented of transport phenomena in non-transferred plasma arc systems. It has been shown that while the modelling of these systems involves the standard application of computational fluid mechanics techniques, numerous pitfalls may be encountered in practice. A detailed discussion of these is presented, with emphasis on the torch inlet conditions, the effect of swirl, the temperature dependence of the system properties and the laminar-turbulent transition in the system. A critical comparison is presented with experimental measurements, and the necessary future work is also outlined.

A mathematical model for the analysis and design of inductively coupled plasma torches Is presented. The model is based upon a solution of the electromagnetic vector potential equation and is capable of predicting the two-dimensional velocity, temperature, and electromagnetic fields as well as the reaction kinetics inside the torch for any axi-symmetric coil configuration. The model is used to study the effect of coil geometry on the thermal decomposition of silicon tetrachloride to silicon. The coil geometry Is found to affect both the temperature field and the flow field and to have a significant effect on the reaction kinetics in the torch. It is demonstrated that through fundamental changes in the coil design It is possible to control the location of the reaction zone and to prevent silicon deposition on the wall of the reactor.

The variation of etch rate with power, pressure and flow is studied using a coupled electron and chemical kinetics model. The electron kinetics model involves a solution of the electron continuity and current continuity equations in conjunction with the Boltzmann equation. The temporal variation of the electric field, electron energy distribution function (EEDF) and electron number density in the bulk of an RF discharge, is calculated using measured current waveforms, and calculated species concentrations. Electron generation through electron-impact ionization, is balanced by attachment and diffusion losses. A time-dependent solution of the Boltzmann equation is employed to investigate the problem of non-equilibrium, between the EEDF and the instantaneous field. Rates for electron impact processes are calculated using the EEDF.

Rate equations for the different species are solved to obtain steady state species concentrations. Radicals and ions produced by electron-impact processes are lost through neutral recombination, ion-ion neutralization, diffusion to reactor surfaces and flow losses. The calculated ion number densities far exceed the electron number density. A transport model that considers the diffusion of etchant species to the wafer and subsequent reaction, is used to compute the etch rate.

The dependences of the electrode self-bias voltage and the ratio of ion energies on electrode area ratio are calculated for a model of capacitively coupled rf discharges. It is assumed that concentric spherical elecrodes with fluid-like radial ion flow adequately models the ion motion, that sheath impedances are dominant, and that ionization processes in the glow are due to ohmically heated electrons. Results show that the ratio of ion energies impacting the smaller electrode to those on the larger depends on the ratio of electrode areas in a more complex manner than a power law.The reason for this is that sheath impedances are more resistive or capacitive at different times in the rf cycle. The self-bias ratio is found to depend relatively little on the ionization model or the pressure but differs substantially from the “power law” result. The agreement of measurements with the model is fairly good.

The prediction of temperature and fluid flow associated with a d.c. plasma jet exiting from the nozzle has been an important issue for-some years. Modeling efforts have mainly relied on incompressible flow formulations and various turbulence models to predict the d.c. plasma at atmospheric conditions. The primary assumption of such models is that the plasma is in local thermodynamic equilibrium, steady state, and that no other species from the ambient are entrained. In the subsonic parabolic approach, the plasma is treated as a free jet with no downstream influence on the upstream calculations. The elliptic approach needs conditions to bespecified along all boundaries, and calculations are dependent on the extent of the solutiondomain and the exit boundary conditions.

Very few attempts have been made to model plasma jets which are supersonic upon exiting the nozzle. The problem is important due to the advantages of low pressure plasma deposition but is complex and difficult to analyze. Specifically one has to acount for compressibility and viscous dissipation effects. Off-design operating conditions (over or underexpanded conditions) greatly influence and complicate the plasma temperature and velocity profiles.

Numerical models describing the plasma physics in the column and plume of a nontransferred, dc arc torch were used in conjunction with two experiments to quantify the extent of turbulence in the free plume. Numerical predictions were made by solving the governing conservation equations with suitable boundary conditions. The solutions indicate a high temperature, high velocity region starting at the anode which decays rapidly as the plume interacts with the atmosphere. The turbulent viscosity is predicted to be up to 140 times higher than the molecular viscosity at specific locations. Modeling results for the plume compare favorably with spectroscopic temperature measurements, and with the cone angles measured with holographic interferometry, which quantifies the amount of turbulence in the plume.

The dependence of ion current density on the sheath voltage and thickness, and ion energy distributions as functions of the pressure and sheath voltage amplitude have been calculated using a self-consistent computer code, ISHEATH. The current density is found to obey an approximate power law dependence on the amplitude of the sheath voltage, Vs, and the sheath thickness, d, and, when charge exchange collisions are considered, on the collision parameter α ΞNgas · σcx ·d. The ion energy distribution (IED) is found to mainly depend on α: for α ≃l there is only a slight (≃10%) change from the collisionless case; for a ≤ 10 the IED is peaked at an energy which is much less than the time-average sheath potential.

A mathematical formulation and computed results are presented to describe the velocity fields, temperature fields and concentration of the activated species in an ion nitriding process, operated at 1–5 torr pres-sure. The theoretical predictions, which are based on the two-dimensional trasnport equations and on a model for computing the electron number density, gave results in broad agreement with experimental findings reported by others for similar systems.

In modelization of plasma jet profiles and calculating individual particle trajectories and the local temperature profile of the injected particles it has generally been assumed that the particles injected into the plasma plume constitute a dilute system. The latter allows one to decouple the jet profile calculations from the particle trajectory and temperature calculations. However, in practice, low pressure plasma deposition is carried out under high loading conditions to minimize thermal energy losses. In addition, the plasma experiences local cooling and deceleration due to the high loading of the injected particles. The effect of powder injection rates (10 kg/hr -30 kg/hr) on the plasma and the particles was studied and numerically modeled. An axisymmertrical parabolic supersonic flow model with compressibility corrected k-… turbulence formulation was developed and used in this study to describe the jet flow. The exchange of energy and momentum between the injected particles and the plasma flow was treated by considering the source term of the governing equations. The effect of particle loading on the resulting jet profiles, particle trajectories and temperature profiles are presented and discussed.

The modeling of powders melting in a plasma jet has received increasing attention as the plasma spraying is being developed to produce high density, high performance materials. These models are being used to predict process capability for melting powders and to develop plasma processing parameters for complete particle melting. Most models assume splherical powder particles which heat and melt uniformly during their travel in the plasma jet. The results of this research indicate this assumption may be invalid. Several experiments conducted to evaluate the melting behavior of nickel and nickel alloy powders (Ni,Ni2OCu, Ni2OCr, Ni5O Cr, Rene'80) are presented.

Spectroscopic temperature measurements of spraying plasma jets (produced by D.C. arcs) are studied (equilibrium conditions, influence of the surrounding environment of plasma jets at 1 atm). For solid particles in flight in these jets are presented the measurement methods of surface temperature, particle vapors densities and excitation temperatures. The “limitation of these measurements are discussed and examples are given.

Optical Emission Spectroscopy (OES) with argon actinometry has been used to study the influence of machine parameters on the composition of a BCl3 RF plasma discharge in the absence and presence of aluminium. Two steady state models are proposed to account for the appearance of the various species seen, and to explain their relative abundances in response to changes in power and pressure. The validity of the actinometric technique for measuring relative changes in ground state concentrations is discussed also.

Two coincidence techniques for simultaneously obtaining particle size and temperature in a plasma torch are described. The absolute magnitude of scattered laser light is used to determine particle size and particle temperature is determined from a two-color pyrometer technique. Typical results on alumina particles injected into a 7 kW plasma torch are presented.

Laser induced fluorescence of CF2 has been observed in plasmas of CF4 and its mixtures with O2 and H2. Surface removal rates of the radical in pure CF4 were measured by observing the decay of the radical when the plasma is switched off. The reduction in CF2 concentration, and the increase in F atom concentrations (the latter measured by optical emission spectroscopy) on the addition of O2 is reproduced by a model of the plasma in which gas phase chemical reactions play a dominant role. The increase in CF2 concentration on the addition of H2 to a CF4 plasma is shown to be due to a reduction in the surface removal rate.

Infrared video thermography was used to evaluate the plasma plume and particle streams created by dc plasma jets operating in air and spraying several types of materials. The temperatures of the plasma plume were measured with the aid of a graphite ‘witness.’ By using the thermography system in an 8khz line scan mode, we were able to obtain data relative to the average velocity of particle groups injected into the plasma. The effects of process parameters could be observed in real time using these methods, showing the potential for real time process control.

This paper reports an application of the planar laser-induced fluorescence (PLIF) technique to visualize species in a plasma jet. For the experiment, 10 μm manganese particles were injected through a center feed in the cathode of a 8 kW plasma torch. A tunable laser beam (−20 mJ/pulse) was focused by a cylinderical lens into a sheet, approximately 2.4 cm high and 550 μm thick, which passed through the center of the plasma plume, exciting manganese lying within the plane defined by the laser. The fluorescence generated was imaged onto a gated, intensified CID camera, converted to digital format, and processed by an image analyzer to form a 2-D map of the relative manganese vapor concentration within the plasma. The results of this proof-of-principle experiment demonstrate that the PLIF measurement holds considerable promise for obtaining simultaneous spatially resolved measurements of species concentration.

Thin carbon films of ˜ 600 Å have been deposited on Si <111> wafers by striking an RF discharge in gas mixtures of hydrogen and methane. The deposition rate increased with increasing methane fraction. The peak rate was ˜ 1 Å/sec at an applied power density of 0.4 W cm−2. The films, with an average density of 2.54 gm cm−3, are amorphous in nature but exhibit broad diffraction maxima corresponding to interatcidc spacings of 2.05Å and 1. 15Å. Measurements of hydrogen concentration in the films showed that the hydrogen at. % [H/(H+C)] increased from 30 to 40% as the hydrogen fraction in the feed gas increased. By using a D2-CH4, we were also able to deduce that hydrogen molecules can be a large source of hydrogen trapped in the films.

The application of plasma techniques is emerging as one of the more efficient means of processing materials. Technology transfer of plasma systems, both thermal and nonequilibrium, however, is severely limited by a lack of understanding of the basic mechanisms involved which govern the plasma processes used in many applications including advanced coatings. The aspects of plasma gas temperature, and species concentration and distribution (molecular, atomic, excited) in the gas phase and associated synergistic effects are fragmentary and unclear. Laser/optical diagnostic techniques represent a critical starting point for providing this needed information; they are in-situ, non-intrusive,species and quantum level specific and give excellent spatial and temporal resolution. A survey of Laser/optical diagnostic techniques applicable to plasma materials processing needs is presented. Some examples of the use of different techniques are also given.

An experimental method using laser-induced fluorescence (LIF) to determine the distribution of neutral manganese in a gas tungsten welding arc is described. The theoretical consideration and experimental difficulties encountered when extending LIF techniques from flames to thermal plasmas are discussed, and preliminary results are presented.

A broad overview is given of diagnostic techniques applicable to thermal plasmas used for materials processing, chemical synthesis and plasma spraying.