PROCEEDINGS ARTICLE | May 14, 2004
Proc. SPIE. 5376, Advances in Resist Technology and Processing XXI
KEYWORDS: Lithography, Etching, Capillaries, Ions, Silicon, Scanning electron microscopy, Critical dimension metrology, Photoresist processing, Semiconducting wafers, Standards development
Exposure wavelength is being reduced significantly, along with design rule reductions. The sub-100-nm node process is currently underway with 193-nm lithography. The problems that need to be solved for the shift in wavelength from 248-nm to 193-nm lithography are those attributed to resist materials, such as plasma resistance, SEM (scanning electron microscope) shrink, and problems attributed to processes, such as pattern collapse and deposition defects (Fig. 1). Although thin films are preferable to improve resist resolution limits, pattern collapse is more likely to occur in 193-nm and 157-nm processing due to DIW (deionized water) rinse surface tension during the drying step after development. This is because of the increased A/R (aspect ratio) of the resist used to improve etching durability and lower the rigidity of 193-nm resist compared to the 248-nm resist. We had focused on controlling the capillary effect between the resist pattern and the rinse solution to avoid swelling. We evaluated the method with the use of DIW with additives rinse, and named its process “FIRM (Fishing-up by improved rinse materials)”. In this paper, we report the effectiveness of the FIRM treatment for each resist by using a dispenser of track system. We had confirmed the pattern collapse within the wafer, the process margin, CD (critical dimension) variation, CDU (CD uniformity), Defect test and, the effectiveness of the FIRM treatment in the etching process. Results indicated that the FIRM process could be used in mass production. Additionally, we had investigated application of this method to the sub-65-nm node process. We created a 55-nm line (Pitch 200-nm), with A/R = 4.47 by overdosing and performed the FIRM treatment. We were able to confirm that the FIRM treatment improved the results while all patterns had collapsed after a standard development. We believe that the FIRM treatment will be applicable to the 65-nm node.
