A new look at the antenna effect

It is a widely held belief that plasma process-induced damage (PID) is caused, among other things, by plasma nonuniformities. These nonuniformities are caused by either nonoptimum reactor design and operation or nonflat wafer topography, the latter giving rise to electron shading that selectively inhibits one charge specie over another causing local surface charging. This charging builds over many cycles. While these phenomena are of serious concern, it is the object of this work to point out that there is an additional way to conceptualize the mechanism for PID in ac glow discharges. In this model, it is proposed that the net charge flux in each power supply half cycle is enough to induce a surface potential of sufficient magnitude to damage thin dielectrics. Semiquantitative arguments are made that outline how this mechanism is capable, in principle, of explaining a wide range of published empirical data. Notable among these predictions is that plasma PID can depend upon the size and shape of the antenna, behaves as though the stress were from a current source, and is only proportional to antenna ratio (AR) under special circumstances.