EUV-induced carbon growth at contaminant pressures between 10-10 mbar and 10-6 mbar: Experiment and model.

Carbon contamination induced by ultraviolet (UV) radiation affects precision optics in applications as diverse as semiconductor lithography and satellite observations of the Sun. Our previous experiments have shown that low-intensity UV-induced surface contamination depends quasi-logarithmically on the partial pressure of the organic contaminant due to the poly-dispersive nature of the surface-adsorbate system. This complex dependence presents difficulties because, without a physically motivated model, it cannot be extrapolated to low pressures. We present measurements and a model of carbon growth induced by UV exposure in the presence of tetradecane vapor. The model, which includes a coverage-dependent adsorption energy, describes the measurements over four orders of magnitude in pressure, and we expect that it can be extrapolated to the lower pressures of interest to the extreme ultraviolet (EUV) lithography and solar astronomy communities. Our experience with other contaminants leads us to expect that other organic contaminants will behave similar to tetradecane. The results also provide insights into the kinetics governing coverage isotherms at extremely low partial pressures.

Note: Thermally stable thin-film filters for high-power extreme-ultraviolet applications.

We investigated several types of thin-film filters for high intensity work in the extreme-ultraviolet (EUV) spectral range. In our application, with a peak EUV intensity of 2.7 W cm(-2), Ni-mesh-backed Zr filters have a typical lifetime of 20 h, at which point they suffer from pinholes and a 50% loss of transmission. Initial trials with Si filters on Ni meshes resulted in rupture of the filters in less than an hour. A simple thermal calculation showed that the temperature rise in those filters to be about 634 K. A similar calculation indicated that using a finer mesh with thicker wires and made of Cu reduces the temperature increase to about 60 K. We have exposed a Si filter backed by such a mesh for more than 60 h with little loss of transmission and no leaks.

Frequency-dependent viscosity of xenon near the critical point.

We used a novel, overdamped oscillator aboard the Space Shuttle to measure the viscosity eta of xenon near its critical density rho(c) and temperature Tc. In microgravity, useful data were obtained within 0.1 mK of Tc, corresponding to a reduced temperature t=(T-Tc)/Tc=3 x 10(-7). Because they avoid the detrimental effects of gravity at temperatures two decades closer to T(c) than the best ground measurements, the data directly reveal the expected power-law behavior eta proportional, variant t(-nuz(eta)). Here nu is the correlation length exponent, and our result for the viscosity exponent is z(eta)=0.0690+/-0.0006. (All uncertainties are one standard uncertainty.) Our value for z(eta) depends only weakly on the form of the viscosity crossover function, and it agrees with the value 0.067+/-0.002 obtained from a recent two-loop perturbation expansion [H. Hao, R.A. Ferrell, and J.K. Bhattacharjee, (unpublished)]. The measurements spanned the frequency range 2 Hz< or = f < or =12 Hz and revealed viscoelasticity when t < or = 10(-5), further from Tc than predicted. The viscoelasticity's frequency dependence scales as Aftau, where tau is the fluctuation-decay time. The fitted value of the viscoelastic time-scale parameter A is 2.0+/-0.3 times the result of a one-loop perturbation calculation. Near Tc, the xenon's calculated time constant for thermal diffusion exceeded days. Nevertheless, the viscosity results were independent of the xenon's temperature history, indicating that the density was kept near rho(c) by judicious choices of the temperature versus time program. Deliberately bad choices led to large density inhomogeneities. At t>10(-5), the xenon approached equilibrium much faster than expected, suggesting that convection driven by microgravity and by electric fields slowly stirred the sample.