Timothy John Sommerer

The 2012 Plasma Roadmap

Low-temperature plasma physics and technology are diverse and interdisciplinary fields. The plasma parameters can span many orders of magnitude and applications are found in quite different areas of daily life and industrial production. As a consequence, the trends in research, science and technology are difficult to follow and it is not easy to identify the major challenges of the field and their many sub-fields. Even for experts the road to the future is sometimes lost in the mist. Journal of Physics D: Applied Physics is addressing this need for clarity and thus providing guidance to the field by this special Review article, The 2012 Plasma Roadmap.

Numerical investigation of the kinetics and chemistry of rf glow discharge plasmas sustained in He, N2, O2, He/N2/O2, He/CF4/O2, and SiH4/NH3 using a Monte Carlo-fluid hybrid model

Capacitively coupled radio‐frequency (rf) glow discharges are standard sources in plasma assisted materials processing. Theoretical analyses of rf discharges have been hampered by the computational difficulty of simultaneously resolving nonequilibrium electron transport and plasma chemistry. We have developed a hybrid Monte Carlo‐fluid simulation that can simulate nonequilibrium electron transport while executing with the speed of a fluid simulation. An electron Monte Carlo simulation (EMCS) is used to calculate the electron energy distribution (EED) as a function of position and phase in the rf cycle. Collision rates and transport coefficients are calculated from the EED and used in a self‐consistent fluid model (SCFM) of charged particle behavior and a neutral chemistry/transport model. Electric fields from the SCFM are cycled back to the EMCS, and the process is iterated until convergence. All pertinent heavy particle (charged and neutral) reactions can be included as well as collisions of electrons wi...

Two‐dimensional hybrid model of inductively coupled plasma sources for etching

Inductively coupled plasmas (ICPs) are currently being investigated as high density (≳1011–1012 cm−3), low pressure (<1–20 mTorr) sources for semiconductor etching and deposition. We have developed a two‐dimensional (r,z) hybrid model for ICP sources and have used the model to investigate Ar/CF4/O2 mixtures for etching applications. The simulation consists of electromagnetic, electron Monte Carlo, and hydrodynamic modules with an ‘‘off‐line’’ plasma chemistry Monte Carlo simulation. The model produces the temporally and spatially dependent magnetic and electric fields (both inductively and capacitively coupled), plasma densities, and the energy resolved flux of ions and radicals to the substrate. We discuss results for densities, power deposition, and ion energies to the substrate as a function of position.

Monte Carlo‐fluid hybrid model of the accumulation of dust particles at sheath edges in radio‐frequency discharges

Particulate contamination (dust) has been observed to accumulate near the sheath‐plasma boundary in both radio‐frequency (rf) and direct‐current (dc) discharges. We have developed and applied a hybrid Monte Carlo‐fluid simulation of electron, ion, and charged dust transport in rf discharges to investigate the dynamics of particulate contamination. The processes governing the transport of charged dust in the model are drift of partially shielded particles in the electric field, collisions with the fill gases, and viscous ion drag arising from Coulomb interactions of particles with ions drifting and diffusing in the plasma. We find that negatively charged dust particles accumulate near the sheath‐plasma boundary, and that transport of the particles is dominated by ion drag.

Monte Carlo‐fluid model of chlorine atom production in Cl2, HCl, and CCl4 radio‐frequency discharges for plasma etching

Chlorine chemistries are often used for the radio‐frequency (rf) discharge plasma etching of compound semiconductors, metals, and silicon. A variety of gas mixtures are used as Cl atom donors, many of which have different electron transport coefficients. In this article we computationally investigate 13.56 MHz rf discharges sustained in He/Cl2, He/HCl, and He/CCl4 gas mixtures in the context of plasma etching. The study is performed using a Monte Carlo‐fluid hybrid model of rf discharges. We find that the Cl atom production efficiency is surprisingly similar in these mixtures, while the details of the electron transport (sources of ionization, locations of attachment, electronegativity) dramatically differ. We also find that even at the low pressures of interest (0.25–1 Torr) attachment in He/HCl mixtures is dominated by vibrationally excited HCl, in analogy to high‐pressure discharge devices.

Translationally hot neutrals in etching discharges

The presence and influence of translationally energetic ions in low‐pressure etching discharges is well known. Neutral atoms and molecules, though known to be chemically reactive, are not generally considered to be otherwise activated in these plasmas. Neutral species may, however, become translationally hot through either charge exchange collisions or by dissociative excitation caused by electron impact. These species are important in etching discharges because they may bring an isotropic source of activation energy to the substrate which may compromise anisotropic etching mechanisms. In this paper we present a theoretical study of the sources and effects of translationally hot neutral atoms and molecules in CF4 etching plasmas. We find that ballistically hot F atoms comprise a significant fraction of the radical flux striking the substrate at pressures of <100 mTorr. In CF4 plasmas, the maximum flux of translationally hot F incident on the substrate of a parallel‐plate rf etching discharge occurs betwee...

Model of a weakly ionized, low-pressure xenon dc positive column discharge plasma

A model of a weakly ionized, low-pressure positive column discharge is described. Of particular interest is the generation of atomic resonance radiation. The model computes species densities as a function of radial position using fluid continuity equations. Electron impact collision rates are found from the zero-dimensional Boltzmann equation. The appropriate electron impact cross sections, heavy particle collision rate coefficients, and radiative decay rates for xenon have been assembled from the literature, and are summarized. Model predictions compare favourably with measured axial electric field values in a xenon positive column discharge, while the model is found to overpredict the electron temperature at low xenon pressure.

Why shallow defect levels alone do not cause high resistivity in CdTe

It is usually assumed that deep defect levels are responsible for the high resistivity in detector-grade CdTe, however, it has been recently reported that shallow defects alone can explain high resistivity. In order to resolve this contradiction we analyze different high-temperature compensation regimes and we particularly show that donor–acceptor self-compensation is not a sufficient condition for high resistivity. We also analyze the dependence of the Fermi level on shallow donor concentration at both high temperature and room temperature using analytical solution of the charge neutrality equation and the graphical method based on the formation energy diagrams. We derive the limits of the high-temperature Fermi level values that lead to the high resistivity at room temperature in a system with shallow defect levels only and find that it is theoretically possible to obtain high resistivity using only shallow defects, but only under an extremely narrow range of physical parameters that is unlikely to occur in practice. Finally, we show that the transition levels of cadmium vacancy acceptor are not deep enough to provide high resistivity.

Radiometric characterization of xenon positive column discharges

We use both experimental diagnostics and computational modelling to characterize the electrical properties and vacuum ultraviolet emission of positive column discharges in both pure xenon and mixtures of xenon with other rare gases. Discharge conditions include xenon pressures of 10-100 mTorr at room temperature, discharge currents of 100-500 mA and positive column diameters of 1.1-5 cm. Our primary interest is the efficient generation of large quantities of ultraviolet radiation for use in fluorescent light sources. We therefore focus on the total output of atomic xenon resonance radiation at 147 nm, reported in units of watts of 147 nm radiation per unit length of positive column, and also the efficiency by which electrical input power is converted into 147 nm radiation. Over the range of discharge conditions considered, we observe outputs in excess of of discharge length and efficiencies in excess of 80%, but we also find that there is a strong tradeoff between efficiency and output, and that these two figures of merit cannot be simultaneously maximized. A change in the behaviour of the discharge is observed to occur near 25 mTorr in 2.2 cm diameter tubes and is attributed to the transition from a discharge sustained by multistep electron-impact ionization to one sustained by single-step electron-impact ionization.

Validation and benchmarking of two particle-in-cell codes for a glow discharge

The two particle-in-cell codes EDIPIC and LSP are bench-marked and validated for a parallel-plate glow discharge in helium, in which the axial electric field had been carefully measured, primarily to investigate and improve the fidelity of their collision models. The scattering anisotropy of electron-impact ionization, as well as the value of the secondary-electron emission yield, are not well known in this case. The experimental uncertainty for the emission yield corresponds to a factor of two variation in the cathode current. If the emission yield is tuned to make the cathode current computed by each code match the experiment, the computed electric fields are in excellent agreement with each other, and within about 10% of the experimental value. The non-monotonic variation of the width of the cathode fall with the applied voltage seen in the experiment is reproduced by both codes. The electron temperature in the negative glow is within experimental error bars for both codes, but the density of slow trapped electrons is underestimated. A more detailed code comparison done for several synthetic cases of electron-beam injection into helium gas shows that the codes are in excellent agreement for ionization rate, as well as for elastic and excitation collisions with isotropic scattering pattern. The remaining significant discrepancies between the two codes are due to differences in their electron binary-collision models, and for anisotropic scattering due to elastic and excitation collisions.