SPIE Advanced Lithography 2009
Conference paper

Quantitative measurement of resist outgassing during exposure

View publication


Determination of both the identity and quantity of species desorbing from photoresists during exposure at any wavelength - 248nm, 193nm and EUV - has proved to be very challenging, adding considerable uncertainty to the evaluation of risks posed by specific photoresists to exposure tool optics. Measurements using a variety of techniques for gas detection and solid film analysis have been reported but analytical results have not in general been easy to compare or even in apparent agreement, in part due to difficulties in establishing absolute calibrations. In this work we describe two measurement methods that can be used for any exposure wavelength, and show that they provide self-consistent quantitative outgassing data for 2 all-organic and 2 Si-containing 193 nm resists. The first method, based upon gas collection, uses two primary chromatographic techniques. Organic products containing C, S and Si are determined by collection of vapors emitted during exposure in a cold trap and analysis by Gas Chromatography-Flame Ionization Detector-Pulsed Flame Photometric Detector-Mass Spectrometry (GC-FID-PFPD-MS). Inorganic products such as SO2 are identified by adsorbent bed with analysis by Gas Particle-Ion Chromatography (GP-IC). The calibration procedure used provides reasonable accuracy without exhaustive effort. The second method analyzes the elemental concentrations in resist films before and after exposure by secondary ion mass spectrometry technique (SIMS), which requires only knowledge of the resist compositions to be quantitative. The extent of outgassing of C and S determined by the two methods is in good agreement for all 4 resists, especially when taking their fundamentally different characters into account. Overall, the gas collection techniques yielded systematically lower outgassing numbers than did SIMS, and the origins of the spread in values, which likely bracket the true values, as well as detection limits will be discussed. The data for Si were found to differ significantly, however, and we show that the discrepancy is due to photo-induced reactions at the polymer surface with the gas atmosphere present above the resist during exposure. For example, photolytic oxidation of the C-Si bonds in air causes volatile Si-containing products to be formed from an otherwise stable polymer, showing it is important to take the gas environment during exposure into account when designing resist polymers for low Si outgassing. ©2009 SPIE.