Fundamental mechanisms in advanced resist systems in optical lithography

This thesis deals with the understanding of and models for chemically amplified resists (CAR) and resist silylation processes. It presents a mechanism-based resist model which quantifies the effects of acid reaction and diffusion in chemically amplified resist image formation for both Fickian and nonlinear acid transport. Parameter extraction techniques for CAR characterization and modeling which are compatible with a production environment are also presented. Moreover, this work presents the first formulation of a mechanism-based, computationally efficient silylation model; and the implementation of both the CAR and top surface imaging (TSI) silylation models into STORM (Simulation of TSI Organic Resist Mechanisms), co-developed by Ebo Croffie.; Accurate resist engineering models have been an evolving concern in optical lithography. This work approaches the problem through mechanism based models for both CAR and silylation systems. In the case of CARs, final resist imaging is quantified by studying the time evolution of extent of reaction contours (EERC) resulting from a coupled reaction/diffusion process. Such an approach provides an accurate means of quantifying acid transport, which cannot be directly observed, as well as provides a basis for parameter extraction techniques which relate post exposure bake (PEB) parameters to the dissolution behavior of the film.; Such modeling and parameter extraction techniques are applied to three different resists exhibiting different mechanisms and processing conditions. The resulting model can accurately capture the profile time evolution as a function of initial exposure and PEB conditions. In the case of APEX-E, a positive tone resist, the critical dimension (CD) dependence on PEB time is described by an exponential diffusion model. Fickian acid transport is observed in UVIIHS and ENR, positive and negative resists respectively. In all cases, the final resist image is a complex function of the initial exposure, reaction and subsequent diffusion of the acid within the film. Estimates of profile formation based on estimated of acid transport alone do not accurately capture behavior observed experimentally.; Although many silylation systems have been proposed and investigated, the general mechanisms describing the uptake of a silicon-containing agent to provide subsequent oxygen plasma etch resistance are not well understood. Profile predictions of silylated resists based on bulk silylation studies do not adequately explain thickness and profile shapes observed. This thesis proposes a general reaction diffusion model which exponentially relates silylating agent diffusivity to the local degree of uptake and linearly relates the reaction activation energy to the stresses present in the film due to silylating agent induced swelling consistent with polymer diffusion and deformation studies. Application of this model in simulation shows that both the silylation depth and sidewall angle are complex functions of the initial exposure of the film, the degree of diffusion nonlinearity present, and the amount of swelling-induced reaction rate retardation. These mechanisms work in conjunction to achieve a sharper sidewall angle and shallower silylation depth than expected from bulk silylation uptake and aerial image studies. Predicted behavior is in agreement with experimental studies in which the final silylated profile is both a function of feature size and feature type imaged. These results suggest that the resist material properties can be the limiting factors determining the ultimate resolution achievable with TSI techniques.