Investigations on plasma physics for an extreme ultraviolet lithography based on laser-produced high-Z plasmas

Summary form only given. We will present experimental and numerical investigations on the plasma physics dominating generation and transport of in-band (2% bandwidth) 13.5 nm EUV emission in laser- produced Sn-based plasmas. Atomic data for Sn at the temperature (30-60 eV) of interest to 13.5 nm EUV emission generation, have rarely been benchmarked by experiment. Plasma physics of the EUV Sn plasmas, including radiation transport and hydrodynamic expansion etc, have never been understood completely and are still critical barriers to design a practical EUVL source. Our researches have shown that for Sn plasma the reabsorption of EUV emission induced by the EUV plasma itself is a key process in the generation and transport of inband 13.5 nm EUV emission. It was pointed out that there is an optimum plasma density profile to obtain high inband CE, in which there is maximum ion density to emit EUV light while the re-absorption of the EUV light still keeps to be negligible. It was shown that 2-D Sn plasma expansion results in a specific angular distribution of the EUV light. The effect of driving laser wavelength on the opacity was investigated. It was found that re-absorption of the EUV light can be avoid by using long wavelength laser, like CO₂ laser with a wavelength of 10.6 mum. High in-band CE, i.e., 3%, was observed with a CO₂ laser pulse with various pulse durations from 15 to 110 ns. Higher in-band CE, 5 %, was achieved by confining CO₂ laser-produced Sn plasma. It was found that the hydrodynamic expansion of CO₂ laser-produced Sn plasma shows a deviation from the well-agreed isothermal model. Another issue still remaining in the road to scale the plasma EUV source to high averaged power is the energetic ions and neutral particles from the EUV plasma. Ions with kinetic energy as high as several 10 keV have been observed from the EUV plasma. These energetic ions could damage the expansive EUV optics permanently. Our rese- rches showed that ion kinetic energy strongly depends on the initial plasma density profiles. Ion energy could be reduced 30 to 100 times by introducing a low energy pre-pulse which introduces a gentle initial density profile. This work was supported by Cymer Inc. and by the University of California (UC) under the UC Industry-University Cooperative Research Program (ele06-10278), and by Gigaphoton in Japan.