Gas Cluster Ionized Beam Etching of Fluropolymers

摘要

There are two broad lines of research in the pursuit of low dielectric constant polymeric interlayer dielectrics. Each has well known limitations. The first is the introduction of free space into the polymer. But this may result in a degree of loss of hardness and a risk of increased permeability to diffusion of contaminants through this free space. The second involves the introduction of Fluorine into the polymer. The danger with this approach includes possible reduced adhesion, and the chance of chemically active species containing the Fluorine being released under various process or environmental situations. Combining these two approaches appears to be the most logical method of lowering the dielectric constant while mitigating the negative features of each approach. However, once the decision to use a fluoro-polymer is made, this demands that the Fluorine be tightly bonded to the polymer so that it will not be released into the bulk under any circumstances. Hence the fluoro-polymer must be chemically inenrt and highly thermally stable to coexist with modern processing steps. The one place where this cannot be assured is in the process of etching or planarization. Here the process actively seeks to dissociate the dielectric for patterning or surface texture modification, and we have a contradictory demand that the polymer break down. Similarly for adhesion purposes it is essential to have the polymer bond to substrates during these destructive process steps. Traditionally, plasma etching has been used to pattern polymers or to texture their surfaces for increased adhesion. Here the damage done to the polymer is primarily at the erosion surface where the dissociating reactions take place. There are, some secondary effects that can promote damage on a much larger volumetric scale deep within the body of the polymer, relatively far from the erosion surface. For plasma etching, radiation damage deep under the surface may be responsible for promoting dissociating reaction. UV radiation damage is suspected as a key dissociation mechanism. In this paper we examine a novel, new process for eroding a surface using a Gas Cluster Ionized Beam or GCIB. The mechanisms of GCIB erosion are significantly different from Plasma etching to merit close examination. Vastly reduced levels of UV and other radiation are possible with this scheme. In addition, the way in which the surface texture is affected by GCIB exposure is quite different.