Mask distortion issues for next generation lithography

Roxann L. Engelstad ( engelsta-AT-engr-DOT-wisc-DOT-edu.gif ) and Edward G. Lovell ( lovell-AT-engr-DOT-wisc-DOT-edu.gif )
UW Computational Mechanics Center (UW-CMC), 1513 University Ave., University of Wisconsin, Madison, WI 53706, USA

Advanced lithographic technologies face a difficult challenge in order to meet the stringent requirements for the 70 nm node and below. Extending 157 nm lithography to these lower nodes will indeed require innovative technical approaches. Consequently, it may be necessary to accelerate research and development on NGL alternatives in order to meet the expectations of the new ITRS roadmap. A key issue for both optical lithography and the NGL technologies is the design and implementation of a low-distortion mask. Research at the University of Wisconsin Computational Mechanics Center (UW-CMC) is addressing the development of advanced lithographic masks for the sub-70 nm regime. Modeling and simulation is being used to investigate materials, fabrication processing and general system parameters in order to achieve the necessary pattern placement accuracy. This presentation will detail our latest predictions on the overlay performance of the EUV and EPL mask systems. For the EUV mask, finite element (FE) models have been used to simulate the complete fabrication process flow, illustrating the importance of employing the same type of chucking within each tool. Thermomechanical models were subsequently used to identify the thermal response of the mask during scanning exposure. The flatness requirements for EUV reticle have been investigated in conjunction with the SEMI EUV Reticle Standard. One source of image placement error (during electrostatic chucking) is the distortions induced when a particle becomes lodged between the mask and the chuck. Predictive modeling has been used to assess the influence of particulates of various sizes, shapes and materials on the micro- and macro-scale response of the EUV reticle. The UW-CMC has also been participating in a study of continuous versus stencil membrane formats for the fabrication of EPL masks. The effects of pattern density gradients have been identified both numerically and experimentally; correction methods to alleviate gradient effects have been evaluated in detail and will be presented. Additional models to simulate the exposure heating (and corresponding distortion) have been used to investigate stitching issues associated with EPL-type masks. The results emphasize the benefits of the modeling and simulation programs for design and optimization before implementation.