Invited Paper VT+EN+TF-TuA7
Optics Contamination in EUV Lithography: Measurement, Modeling and Mitigation

Tuesday, October 29, 2013, 4:00 pm, Room 202 C

Over the past decade the semiconductor industry has relied upon increasingly complex optical and processing techniques to reduce the feature size of microelectronic devices from 100 nm to 22 nm without reducing the wavelength of the lithographic light (193 nm). Since the cost to continue this trend is rapidly becoming prohibitive, the industry is poised to undergo a dramatic shift to Extreme-Ultraviolet Lithography (EUVL) using 13.5 nm (92 eV) radiation. One of the many technical challenges to this transition has been the degradation of throughput and optical resolution caused by the buildup of carbonaceous deposits on optic surfaces resulting from EUV-induced decomposition of organic species which are continuously outgassed during EUV irradiation of the chemically complex photoresists. The National Institute of Standards and Technology (NIST) is uniquely positioned to address this problem using our Synchrotron Ultraviolet Radiation Facility (SURF III) which has a peak output near 13.5 nm. We have directly measured EUV-induced contamination rates of various species over a wide range of partial pressures as a function of EUV intensity, dose and wavelength. Among other findings, these investigations revealed two regimes of contamination. For intensities high enough that the EUV-induced reaction rate far exceeds the thermal desorption rate, the contamination rate is independent of intensity and scales linearly with partial pressure since every adsorbed molecule photoreacts. In the opposite limit, the surface coverage of adsorbed precursor molecules in thermal equilibrium with the gas phase is only slightly perturbed by EUV irradiation. The contamination rate in this case scales linearly with intensity, as expected, but also displays a quasi-logarithmic dependence on partial pressure over many decades. Although initially surprising, this sub-linear pressure scaling was found to be consistent with temperature-programmed desorption measurements (in collaboration with Rutgers University) and with models of desorption kinetics on non-ideal surfaces. I will summarize these results and discuss how fundamental investigations such as these enabled the development of practical methods to systematically measure the contamination produced by outgassing of EUV resists. Together with additional measurements and models, this allowed the makers of EUVL tools to define the maximum allowable level of resist outgas contamination that could be managed by in situ mitigation and cleaning techniques. I will outline the resulting outgas testing protocols first fully implemented at NIST in 2011 to which all resists must be subjected before use in EUVL tools.