|
Early technology learning for O.25m lithography should confirm predictions of tools and processes needed for printhg these small image sizes. Improvements are felt to be needed in reticle critical dimension (CD) control, stepper overlay, lens design, and resists for imaging O.25.im. Etch technology is also tested for substrate as well as organic ARLs at these line sizes. This paper will present results from efforts to produce O.25jm poly, O.301um contacts/vias, and O.3OMm metal images on a device of realistic topography applications. Poly patterning has been improved by using off axis illumination but the process window is still small. There is a trade off between focus latitude and isolated to dense bias when using annular illumination. While organic ARLs have improved imaging CD control, there is a corresponding loss in CD control during ARL etch. Effects of ARL thickness variation over topography will also cause CD variations when etching to clear the thicker areas. Poly etch itself also contributes to the CD uniformity variations. At the contact masking level the wafer has a fairly planar oxide surface but even with a chemical mechanical polish the oxide thickness will vary due to the underlying topography. Contact patterning studies will include sensitivities of contact imaged CDs to variations in film stack thicknesses and reflectivities. Process windows will be shown for several film stacks with and without ARC for different resist thicknesses. The relationship between oxide thickness and contact CDs will be shown for contacts at O.3Om. For metal etch, results will be shown for msitu and dedicated ARL etch tools using various over-etch conditions. The relationship between the resist thickness required for the current ARL and metal etch selectivity and litho patterning capabilities will be discussed. Results will be shown for metal patterns post imaging, ARL etch, and metal etch at O.301um.
Keywords: annular illumination, etch bias, process window, deep ultraviolet, process bias
|
