In order to succeed with such low-k1 lithography at EUV wavelength, we need to be able to print a grating at high contrast similar to ArF immersion tools, where a contrast exceeding 0.95 is achieved routinely. All 2d printing is composed of interference of x and y-directed diffraction orders and high contrast in 2d thus depends on such 1d grating contrast. Any low-k1 imaging will use either dipole or some other sort of extreme off-axis illumination such as cross-quad (cQuad). The two relevant magnitudes for any high contrast are the intrinsic contrast due to a monopole, and the spatial shift of the two images that are generated by the two monopoles making up the dipole exposure. In EUV with current absorbers, high contrast can currently only be achieved using monopole illumination, a technique that does not lend itself to process integration due to removal of wafer side telecentricity and resulting overlay problems at all but preferred pitch. For dipole illumination at low-k1 pitches, we collect only 0th order light and only one 1st diffracted order for each pole. This means that for a dipole at the resolution limit, the final image for horizontal l/s patterns consists of only four incident waves, one TE and one TM wave for each of the poles. In this paper, we screen absorber by n and k values. In the process, we introduce phasor notation in order to gain insight into the behavior of the absorber and try to understand the metrics. We investigate intrinsic contrast and image blur due to monopole image shift.
|
