David Drain

Dual-layer dye-filled developer-soluble BARCs for 193-nm lithography

A family of dye-filled developer-soluble bottom anti-reflective coatings (BARCs) has been developed for use in 193-nm microlithography. This new dye-filled chemical platform easily provides products covering a wide range of optical properties. The light-sensitive and positive-working BARCs use a transparent polymeric binder and a polymeric dye in a thermally crosslinking formulation, with the cured products then being photochemically decrosslinked prior to development. The cured BARC films are imaged and removed with developer in the same steps as the covering photoresist. Two dye-filled BARCs with differing optical properties were developed via a series of DOEs and then used as a dual-layer BARC stack. Lithography with this BARC stack, using a 193-nm resist, gave 150-nm L/S (1:1). A 193-nm dual-layer BARC stack (gradient optical properties) from the well-established dye-attached family of light-sensitive BARCs also gave 150-nm L/S (1:1) with the same resist. However, the latter provided much improved line shape with no scumming. The targeted application for light-sensitive dual-layer BARCs is high-numerical aperture (NA) immersion lithography where a single-layer BARC will not afford the requisite reflection control.

Customer Expectation and Warranty Cost — Nominal-the-Best Case

A fixed target, be it at zero or infinity, is assumed in Taguchis method for formulating the quality loss function (QLF). The QLF only accounts for immediate issues within manufacturing facilities whereas warranty cost occurs during customer use. Variable customer expectation has not been considered in the Taguchi methodology. This article presents a methodology to predict warranty probability, the probability of customer complaint, on the basis of two independent variables; product performance and consumer expectation. It is expected that the formulation presented will serve as a basic model for predicting warranty loss using warranty probability due to a single characteristic under certain assumptions. The nominal-the-best case is considered in this article and warranty cost is estimated for an automotive example to demonstrate the methodology.