A continuing improvement in resist process is a necessity for high-end photomask fabrication. In advanced chemically amplified resist systems the lithographic performance is strongly influenced by diffusion of acid and acid quencher (i.e. bases). Beside the resist properties, e.g. size and volatility of the photoacid, the process conditions play important roles for the diffusion control. Understanding and managing these properties influences lithographic characteristics on the photomask such as CD uniformity, CD and pitch linearity, resolution, substrate contamination, clear-dark bias and iso-dense bias. In this paper we have investigated effects on the lithographic characteristics with respect to post exposure bake conditions, when using the chemically amplified resist FEP-171. We used commercially available mask blanks from the Hoya Mask Blank Division with NTAR7 chrome and an optimized resist thickness for the 248 nm laser tool at 3200Å. The photomasks were exposed on the optical DUV (248nm) Sigma7300 pattern generator.Additionally, we investigated the image stability between exposure and post exposure bake. Unlike in wafer fabrication, photomask writing requires several hours, making the resist susceptible to image blur and acid latent image degradation.

Tighter requirements on mask resolution, CD and image positioning accuracy at and beyond the 45 nm technology node push the development of improved photomask blanks. One such blank for attenuated phase-shift masks (att-PSM) provides a thinner chrome film, named TF11, with higher chrome etch rate compared to the previous generation Att-PSM blank (NTAR5 chrome film) from the same supplier. Reduced stress in the chrome film also results in less image placement error induced by the material. FEP-171 is the positive chemically amplified resist (PCAR) that is most commonly used in advanced mask manufacturing with both 50 keV variable shaped e-beam (VSB) and DUV laser pattern generators. TF11 allows an FEP-171 resist film down to about 2000 Å thickness with sufficient etch resistance, while the standard resist thickness for NTAR5 is around 3000 Å.This work has experimentally evaluated the use of TF11 chrome and FEP-171 resist together with a 248 nm DUV laser pattern generator, the Sigma7500. First, patterning performance in resist with thicknesses from 2000 Å to 2600 Å, in steps of 100 Å, was tested with respect to swing curve and basic lithographic parameters including resolution, CD linearity, CD iso-dense bias and dose sensitivity. Patterning results on mask showed a swing minimum at around 2200 Å and a swing maximum at around 2500 Å, which correspond to reflectivity measurements for 248 nm wavelength performed by the blank supplier. It was concluded that the overall patterning performance was best close to the swing maximum. Thereafter the patterning performance using TF11 at two resist thicknesses, 2000 Å and 2550 Å, was studied in more detail and compared to performance using NTAR5 with 3200 Å resist. The evaluation showed that the Sigma7500-II offers good compatibility with TF11, especially using the optimized FEP-171 resist thickness of 2550 Å. It also showed that the patterning capability of the Sigma7500-II using TF11 and 2550 Å resist is improved compared to using NTAR5 and 3200 Å resist.

No abstract

No abstract