Ch. Steinbrüchel

Etch mechanism in the reactive ion etching of silicon nitride

Reactive ion etching of silicon nitride with CHF3/O2 plasmas has been studied in a hexode reactor and compared to silicon dioxide etching. Measurements of etch rates as a function of gas composition and pressure were combined with Langmuir probe data for the ion flux to the substrate to give etch yields (number of substrate atoms removed per bombarding ion). At low oxygen content, the etch yields of both materials are about 2 atoms/ion and are essentially independent of pressure between 10 and 60 mTorr. At higher oxygen content, the etch yield is considerably larger for silicon nitride, and for both materials the etch yields increase with increasing pressure. From these results it can be concluded that in a CHF3/O2 plasma at low oxygen content, the etch mechanism is mostly direct reactive ion etching for both silicon nitride and silicon oxide. On the other hand, at higher oxygen content the etching is ion enhanced for both materials, but to a much greater extent in the case of silicon nitride.

Heat dissipation in catalytic reactions on supported crystallites

Abstract Local temperatures produced by catalytic reactions on small metal particles on insulating supports are examined. The mode of energy release, the size of the particle, and the thermal boundary resistance at the crystallite support interface are shown to be important in determining temperatures. Explicit solutions are obtained for particles large compared to the mean free path of the heat carriers, using the continuum equation for heat conduction, and also for particles so small that boundary scattering dominates. Numerical results predict significant local temperature increases for conservatively estimated parameters. Heating is found to depend sensitively on particle size and on the properties of the metal-support interface.

Condensation of gases on metals: The one-dimensional square well model*

Abstract Trapping probabilities of gas atoms at surfaces are calculated assuming a classical onedimensional square well potential as a function of gas and surface temperatures. It is shown that initial sticking coefficients of chemisorbed gases on transition metal surfaces can in most cases be fit fairly well by this model using reasonable values of the interaction energy, although the model does not predict observed behavior for surface temperatures>600°K. For some systems the initial sticking coefficients are higher than predicted by this model, indicating that other mechanisms of energy transfer are probably operative. The angular dependent sticking coefficient which would be measured in a molecular beam experiment is also computed.

Reactive ion etching of copper with BCl3 and SiCl4: Plasma diagnostics and patterning

The reactive etching of copper in BCl3 and SiCl4 based plasmas has been investigated and empirically modeled using statistically designed experimentation. Gas mixtures with N2 exhibit considerably higher etch rates than mixtures with Ar. In the temperature range of 225 to 275 °C, the copper etch rate in BC13/N2 can be as much as an order of magnitude larger than in BCl3/Ar. Plasma diagnostics [optical emission spectroscopy (OES) and Langmuir probe measurements] were used to study the effect of the diluent gas on the discharge characteristics and the resulting etch process. OES indicates that N2 does not act merely as a diluent but rather increases the concentration of Cl atoms. Langmuir probe data show very little difference in the ion densities between the various gas mixtures. These results suggest, in combination with the empirical model, that the etch rate is limited by the amount of available reactant rather than by product desorption in the temperature range investigated. High‐resolution patterning ...

Analysis of Langmuir probe data: Analytical parametrization, and the importance of the end effect

The ion current to a cylindrical or spherical Langmuir probe can be represented as Ii=I0a(−X)b, where I0 is the ion flux to the edge of the probe sheath, X is the dimensionless probe potential, and a and b are parameters depending on the ratio of the probe radius rp to the Debye length λD and the probe geometry. We present analytical expressions for a and b which are valid over a wide range of rp/λD. We also discuss under which conditions the end of a cylindrical probe contributes appreciably to the probe current. Neglecting the end effect may lead to a significant overestimate of the ion density.

Parametrization of Laframboise’s results for spherical and cylindrical Langmuir probes

Three new aspects regarding the analysis of Langmuir probe data are presented. First, we demonstrate that the numerical results of Laframboise for spherical probes can be parametrized easily for arbitrary ratios of the probe radius rp to the Debye length λD. The ion current can be expressed in the form a(−X)b, where a and b are parameters depending on rp/λD, and X is the dimensionless probe voltage. This functional form is the same as the one for cylindrical probes reported previously, but the values of a and b are different. Second, we use numerical simulations to show that unless the plasma potential Vs is known, it is in general difficult to determine accurately the form of the ion current characteristic Ii(Vp), and thus the ion density Ni, from typical probe data. This is because Ii(Vp), Ni, and rp/λD are interdependent. Third, the simulations indicate that the apparent electron energy distribution is very sensitive to the exact form of Ii(Vp) and to the method by which Ii(Vp) is subtracted from the t...

Kinetics of particle formation in the sputtering and reactive ion etching of silicon

Particle formation from a Si substrate in an Ar sputtering plasma and a reactive ion etching plasma of 10% CCl2F2 in Ar has been investigated by laser light scattering. The kinetics of particle growth are studied as a function of rf power and chamber pressure. Threshold behavior for particle generation has been observed in that minimum values of both rf power and pressure are necessary for particles to appear. In both reactive ion etching and sputtering, the smallest particles detected on the wafer are ∼0.2 μm in size. In sputtering most particles are spherical and must have been deposited onto the substrate from the plasma. On the other hand, in reactive ion etching most particles are hemispherical or conical and seem to have grown on the substrate itself. Particle formation is correlated with significant redeposition in sputtering and with deposition of a surface film in reactive ion etching. Si atoms removed from the substrate are probably responsible for the nucleation of particles in both types of pl...

On the sputtering yield of molecular ions

It is shown that the sputtering yield of a molecular ion can be expressed in terms of the sputtering yields of the constituent atoms, irrespective of the details of the process by which the molecular ion dissociates upon impact on the target surface. It then follows that, in general, sputtering yields on a clean surface should be about the same for molecular and atomic ions of the same mass and energy, exceptions being very light targets or very heavy projectiles, in which cases sputtering yields for molecular ions should be somewhat higher. For ions which are chemically reactive with respect to a target and thus change its surface composition, these conclusions need to be modified. Sputtering yields at very low ion energy, i.e., near the sputtering threshold, are also discussed and compared for atomic and molecular ions.

Growth of plasma‐generated particles and behavior of particle clouds during sputtering of silicon and silicon dioxide

We have observed particles nucleated and grown in an Ar sputtering plasma from Si and SiO2 substrates. With both materials, particle clouds forming near the plasma sheath boundary were studied by laser light scattering. Roughly spherical particles ranging down to 70 nm were collected on transmission electron microscopy grids installed downstream from the wafer even before the onset of laser light scattering. Electron microscopy revealed the structure as well as the size distribution of the particles. The size distribution is essentially monodisperse for very small particles but becomes much wider and spatially varying as the average particle size increases with time. Pressure and flow rate determined the spatial position and the extent of development of the particle cloud above the wafer. Spatial dependence of clouds on pressure and flow rate can be used to control particle formation or particle deposition on the wafer.

Relative sticking coefficients of H2 and D2 on tungsten

The relative sticking coefficients of H2 and D2 on the (100) and (110) planes of W have been measured. It is found that on (100) W, SD2SH2 ≅ 1.4 in agreement with a previous determination by Tamm and Schmidt, but that on the close packed (110) plane SD2SH2≅1.0. This result suggests that qualitatively different mechanisms probably control condensation rates on the two planes. It is also shown that the coverage dependence of S for the tightly bound β2 state on (100) W is constant rather than decreasing as 1−θ as had been suggested previously by Madey.