As critical dimensions continue to shrink in line with the SIA roadmap, the ratio of printed feature size and accepted wavelengths for optical lithography is driving inexorably towards the theoretical limitation of 0.25 for the Raleigh equation constant, k1. With the drive to lower k1 values fundamental limitations start to impact optical lithography. One example is the inability to simultaneously print features at different duty cycles with acceptable process windows. In the k1 regime down to 0.5, dense and isolated features could be printed in one with acceptable process windows. Today advanced lithography is operating at k1 values of 0.42-0.37 using KrF excimer laser light sources at a wavelength (lambda) of 248nm High lens Numerical Aperture (NA) is required to obtain sufficient aerial image contrast for dense lines, but results in reduced depth of focus which scales proportional to lambda /NA(2). Using off-axis illumination techniques such as annular illumination can compensate the reduction in depth of focus for dense lines. For isolated lines high NA has only limited impact on the aerial image contrast due to the difference in the diffraction pattern and only serves to reduce the limited depth of focus which, unlike dense lines, does nor benefit from the application of off-axis illumination. Use of increasingly strong imaging enhancement techniques will be required at lower k1 values resulting in further trade-offs to be addressed in pattern dependency. For example, quadrupole and di-pole off-axis illumination provides stronger enhancement to the available process window than annular illumination but only for features with specific orientations. Tn this paper an overview of the different imaging enhancement techniques will be given and examples of the trade-offs between enhancement and detoriation for different feature types will be given. In order to manufacture devices at low k1 using strong enhancement techniques, constraints on orientations and duty cycles will have to be applied in the device design. Alternatively, individual device layers will have to be separated by feature type, duty cycle and orientation to allow optimum enhancement techniques to be applied for each feature using multiple exposures. These approaches will be required if optical lithography at k1 values around 0.3 is to be realized. In the paper we will compare the use of two very strong enhancement techniques, dipole illumination and alternating (Levinson type) Phase Shift Mask with respect to process latitude, complexity and aberration sensitivity. To complete the review of low k1 the economic viability of optical lithography utilizing these strong enhancement techniques will be analyzed in terms of Cost of Ownership.
