ITRS2003FrontEndProcessingChallenges.pptVIP

ITRS 2003Front End Processing Challenges David J. Mountain *Gate Stack Leff Control *Memory Cells Dopant Control Contacts *Starting Material FEP Grand Challenge – Material Limited Scaling Why are materials a grand challenge? MOSFET material development 1965 (Moore’s Law) – MOS is really Aluminum/SiO2/Silicon Materials: Aluminum, SiO2, Silicon, Boron, Phosphorous 1985 (CMOS) – MOS is really Polysilicon/SiO2/Silicon New materials: Arsenic, Refractory Metals (Ti, Ta, W), Nitride 2005 (Today) – MOS is really Polycide/Nitrided SiO2/Modified Silicon New materials: Copper, Germanium 2015 (End of Moore’s Law?) – MOS will be ??/??/Modified silicon New materials: metal gate, gate dielectric, nickel silicide, others?? MOSFET device structure 1965 Leff ~20 microns: Al gate, SiO2 dielectric and isolation 1985 Leff ~1.25 microns Self-aligned and silicided Source/Drain Polysilicon gate with silicide LOCOS isolation 2005 Leff ~0.032 microns Oxynitride gate dielectric Strained Silicon (Ge implants) Lightly-doped source/drain extensions with nitride spacer 2015 Leff ~0.01 microns FinFET structure Metal gates, unknown material – multi-gate designs Unknown high-k gate dielectric Elevated Source/Drain Strained SOI starting material Looking ahead “... solution to a large number of process integration issues arising from a fully 3-D transistor design must be dealt with before the successful introduction of multi-gate, fully depleted CMOS production. These challenges, added to the anticipated shifts to high-K gate dielectrics and dual-workfunction, metal-gate materials, constitute a revolutionary change in transistor technology in the coming 5-7 years.” Why is this revolutionary change? We are adding new materials faster than ever ... to the most critical part of the transistor for the first time ... while the basic device structure is being substantially modified Gate Stack Need high-K material to replace SiO2 because direct tunneling caused gate leakage current to explo