Reactive ion etching of silicon substrates in a plasma containing chlorinated species does not result in undercut of a permanent mask. When the silicon is very highly doped it behaves as a different material and undercut has been observed. This phenomenon will be discussed. For use in chlorinated plasmas, there is a choice of nonerodible masks that sputter slowly but will not be redeposited on the substrate surface. Both CCl4/Ar and Cl2/Ar plasmas will be described. The variation in etch rate of silicon with rf power, frequency, reactant concentration, reactant flow rate, gas presure, crystal orientation, and batch size will be presented. The possibilities of polymer formation and surface roughening will be discussed.

An etching method capable of producing vertical-walled high-resolution patterns in aluminum and aluminum alloy films is described. The process consists of rf sputter etching in a plasma containing ion species, which react with the metal to form volatile or easily sputtered compounds. The presence of reactive species greatly enhances the etch rate, while the electric field maintains the directionality inherent in the sputtering process. Halogen ion species, as obtained in rf plasmas containing a partial pressure of Cl2, Br2, HCl, HBr, or CCl4, were used to produce reactive ion etching of Al. The etch rate in a CCl4 plasma at a power input of 0.6 W/cm2 is as high as 5000 Å/min. Variations in reaction rate with rf power, reactant concentration, reactant flow rate, temperature, gas pressure, batch size, and residual gas contamination are discussed. Etch rate data for various materials found suitable for masking are also presented.

The application of a suitable negative bias to a film that is being deposited by dc sputtering produces a film of substantially lower resistivity (than without bias). The action of the bias is shown to result from positive ion bombardment of the film during deposition, leading to the selective removal of adsorbed impurities. An analytical expression describing the process is derived and found to give good agreement for tantalum films sputtered in argon containing oxygen as a contaminant. The effect of bias on contaminants originating both at the cathode and in the gas is compared. Comparison of bias sputtering and asymmetric ac sputtering in terms of an ion bombardment model shows that bias sputtering leads to purer films because ion clean-up is in operation during the entire deposition time only for the case of bias sputtering. Good adhesion of bias-sputtered Ta films to glass substrates is obtained by initially applying a small positive bias to the substrate.

We have demonstrated that it is feasible to reactively ion etch (RIE) copper films.The etch rate is about an order of magnitude higher than the sputter etch rate.There is no undercut of an inert mask and the patterns formed have vertical sidewalls.

Reactive ion etching of aluminum and its alloys in CCl4/argon plasmas produces vertical etch profiles and residue-free etched regions. However, under certain conditions, preferential lateral chemical etching occurs, which produces pits in the surface of the metal along the edges of the lands. The lateral etching is due to a thermally activated process and is related to the structure of the films. The extent of attack can be reduced by annealing the films at 450°C before etching. Al–Cu films deposited onto heated substrates are more susceptible to attack than those evaporated at lower temperatures. This apparent contradiction has not been resolved. The effects of etching temperature, plasma parameters, and film structure will be discussed. Scavenging of the responsible species by additions of H2 or O2 to the plasma is not successful.

Today, reuse in software engineering is usually supported by component libraries, such as Java packages. Components are self-contained, ready-to-use building blocks, which are selected and composed. They are usually associated with the implementation phase, a result of practical experience rather than of existing limitations. In this paper, we shed some new light on the concept of components from the protocol engineering point of view. In particular, we describe a conceptual framework for the protocol design phase, and introduce a specific type of protocol design components called micro protocols. We then instantiate and apply this framework to a subset of SNMP, the Simple Network Management Protocol, using SDL as design language.

No abstract

The results of a study of the influence of arsenic doping of silicon and Cl2 and CCl4 reactant concentrations on etch rate and etch profile are described. From approximately 1016 to 1020 As/cm3, the etch is anisotropic and the etch rate does not increase in either Cl2 or CCl4 plasmas. At concentrations above 1020 As/cm3, mask undercutting occurs and the etch rate rapidly increases (approximately 300% for 1.5×1021 normalAs/cm3 ), although the functional form of the relationship between etch rate and As concentration depends on the reactant, reactant supply, and silicon load. Temperature effects are also described.