Paul Kirsch, S.C.Song, & Prashant Majhi, SEMATECH & Raj Jammy, IBM assignee to SEMATECH
ABSTRACT
Historically, both logic and memory have been front-end process technology drivers. For example, logic development resulted in shallow trench isolation eliminating “bird’s beak” isolation oxide. Ion implantation increased doping densities needed for high mobility nchannel transistors. More recently, embedded SiGe, dual stress nitride, Hf-based highk and metal gates have been logic performance drivers. On the other hand, memory development has spawned plasma doping, nitrided oxides, Al-based high-k and borderless contacts, as well as pushed lithography critical dimension. Motivation for memory development work is multi-fold. First, memory revenues grow faster than logic revenues. Second, the growth potential of memory materials and processes in embedded applications holds tremendous promise to improve system level performance. Finally, many new materials are emerging to address memory-scaling challenges but screening these many options is difficult. Building on extensive knowledge of gate dielectric and metal gate materials gained through several years of experience, SEMATECH is pursuing the challenge of materials development for DRAMs, Flash, PRAM, and ReRAM as described below. SEMATECH has successfully provided solutions to the semiconductor industry [1,2,3 and 4] and this deep knowledge accelerates memory progress.