Diagnosing phenomena that influence missing contacts and pillars

Abstract

Algebraic models using Poisson statistics are used to investigate phenomena contributing to missing contacts and pillars, assess progress in their mitigation, and provide guidance for technology pathfinding for sub-10nm. Fitting the missing rate versus dose clearly shows the dominant EUV photon absorption stochastics, enables scaling from 10 -4 to 10 -13 error rates, and gives an equivalent number of dissolution aiding events for diagnosing physical phenomena. This number for 20 nm contacts in CAR is about 300 and suggest the possible presence of reaction enhanced diffusion or reaction enhanced solvent penetration and gel extrusion prior to disentanglement dissolution. Metal oxide resists are more complex and data on missing pillars is just becoming available. The use of similar characterization techniques can pursue phenomena that are more Gaussian in nature and more beneficial for reducing error rates.