Thin Damaged Layer Research Articles

Directional emission of nonthermal halogen atoms by electron bombardment of alkali halides.

We present the first experimental results on angle-resolved kinetic-energy distributions of halogen atoms desorbed from single crystals of alkali halides because of electron bombardment. We found that the ejection of nonthermal Br atoms from the (100) surface of KBr is strongly forward peaked along the normal. We suggest that this effect is caused by a thin damaged layer on the surface due to a strong nonstoichiometry of the erosion process itself.

Inert gas reactive ion etching damage to GaAs using inverted heterojunctions

Selectively doped inverted heterojunctions containing a two-dimensional electron gas were used as a sensitive vehicle for monitoring dry processing damage. We found that the electron sheet concentration, strongly dependent on the total number of carriers in the GaAs cap layer, and the mobilities were significantly depressed even for very short exposures to low-voltage helium plasmas. Argon, which caused less degradation than helium, was found to increase the sheet carrier concentration and hence the mobility after prolonged exposure. The damage mechanism responsible for the carrier loss in both cases is most likely the production of traps. The subsequent carrier increase seen for the argon case is probably attributable to the creation of a very thin donorlike damage layer on the surface of the GaAs cap layer.

The surface properties of mechanically polished (100) manganese zinc ferrites

Mechanically polished mono-crystalline (100) manganese zinc ferrite samples have been investigated using magneto-optical spectroscopy and hysteresis loops. The as-polished (100) surface shows a thin damaged layer on top of a relatively thick highly compressed layer. The stress is close to the maximum strength and gives rise to an in-plane anisotropy, with an anisotropy field of up to 460 kA/m. The in-plane coercive field is about 1 kA/m. A Syton treatment can remove the stressed layer and reduce the coercivity.

Phonon emission from a metallic film excited by nanosecond electrical pulses

We report the first measurements of heat pulses generated by electrical excitation of a thin metallic film on a nanosecond time scale. Using heater power densities of between 8 and 800 Wmm -2, we found a significant variation of the phonon pulse length reaching the bolometer and also of the ratio of transverse to longitudinal polarisation amplitudes. The phenomena were attributed to diffusive propagation due to highly frequency dependent phonon scattering within a thin damaged layer at the surface of the sapphire substrate.

Sputter Etching Effects on GaAs Schottky Junctions

The electrical properties of Schottky junctions with rf sputter etching, immediately before sputter deposition of metals, and their thermal recovery characteristics are described. Sputter etching not only results in removal of surface layers, but also in formation of a thin (∼100Å) damage layer, which decreases the Schottky junction effective barrier height. Thermal recovery characteristics strongly depend on sputter‐etching conditions, such as rf‐power and etching time. The activation energy for thermal recovery is continuously distributed, even in one sample. The results suggest that the damage is in lattice disorder and the degree of disorder depends on energy and/or dose of incident ions. The damage introduced during low power (50W) sputter‐etching is completely annihilated by thermal annealing at 350°C. The barrier height uniformity after annealing is improved by sputter etching at low power. It has been confirmed that sputter etching is a useful technique for surface treatment in the fabrication of Schottky junctions with uniform barrier height.

Tapered windows in phosphorus-doped SiO2by ion implantation

Phosphorus-doped SiO <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> is frequently used as a dielectric coating in silicon integrated circuits. It is important that windows in this dielectric have sufficiently tapered walls so that the subsequent metallization has good step coverage. It is shown here that tapered windows can be made in both Nitrox-deposited ∼ 1-percent phosphorus-doped SiO <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> and Silox-deposited ∼ 7-percent phosphorus-doped SiO <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> as well as undoped SiO <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> by an ion implantation which produces a thin damaged layer at the top of the oxide. The damaged layer etches at a faster rate than the undamaged oxide. This fast-etching layer undercuts the photoresist which serves as the etching mask and results in window walls having slopes in the range of 30-40° with respect to the wafer surface. Tapering windows by ion implantation is a dependable process that gives reproducible results without having to rely on the art of photoresist liftoff methods.

Characterization of real surfaces by ellipsometry

The results of ellipsometric measurements on single-crystal silicon surfaces subjected to various surface treatments such as mechanical polishing, chemical etching, sputtering by argon ions, cleavage, and annealing, are presented. Because of the consistency and reproducibility of the results on chemically etched silicon samples by other workers as well as our own, these results are believed to be truly representative of the bulk sample of silicon without any damaged surface layer even though there is always a thin oxide film present on the surface of the sample. Ellipsometric measurements on both mechanically polished and argon ion bombarded surfaces have shown considerable increase in the values of the ellipsometric parameter ψ and the pseudo-normal reflectance R as compared with those of the etched samples. This is obviously caused by the presence of the damaged surface layers on these samples. It is observed that the variation in ψ between different cleaved samples is much larger than the possible experimental errors. This discrepancy has been attributed to the presence of dislocations on or near the cleaved surfaces created in the process of cleavage. These fresh dislocations can be taken into account by an effective average value of the refractive index of silicon or, more meaningfully, by an effective thin damaged surface layer with different optical parameters, on the bulk silicon. The annealing of cleaved silicon samples at high temperatures and ellipsometric measurements on such annealed samples show that the fresh dislocations introduced in the sample during cleavage can be removed by annealing. These studies reveal that ellipsometry provides another means of detecting surface defects and yields information concerning the optical properties and average film thickness of the damaged surface layer as well as those of the film and the substrate.

Bolometer for Detecting Heat Pulses in a Magnetic Field

The bolometer depends on the rapid temperature variation of the current in an avalanching semiconductor. Typically, at 1.7°K, it has a sensitivity of 0.05 V W−1, a response time of 2×10−7 sec, and a noise equivalent power of 10−7 W Hz−1/2. The rapid response is possible because only a thin damaged layer at the surface of the semiconductor is involved in the detection process. We have used the device to detect heat pulses in magnetic fields up to 30 kOe.