1 MURI Plans S E Thompson March 27

1 MURI Plans S. E. Thompson March 27, 2005

OUTLINE • SRC and AMD, AMAT, IBM, Intel, TSMC, TI, UMC funded device modeling/characterization work • Plans: Single event transient – Start with SRC strained Si modeling/calibration – SET on State-of-the-art uniaxial strained Si (90 -45 nm) – SET on Future strained devices • Strained Ge transistor • Strained Si and Ge on (110) wafers TI Fellows Forum

Many Ways to Do Strain Intel 2004 EDL TI 2004 VLSI 45 nm AMAT 2004 IEDM Si 1 -x. Gex IBM 2005 VLSI Si 1 -x. Gex TSMC/Freescalse 2005 Samsung 2005 VLSI Nitride stress a-Si Gate stress Hoyt Removable film pre-anneal 1 -2. 5 GPa stress film Post salicide TI Fellows Forum Even more on high stress layers

Strain Being Adopted by All Strained Si Fujitsu 45 nm Si 1 -x. Gex p-type MOSFET Tensile Nitride stress Source: Ti Compressive Nitride Gate STI NMOS TI Fellows Forum PMOS Source Chipworks: 90 nm Intel, IBM, AMD, TI, Fujitsu stress

Why Strain: Very Impressive Performance 2004 IEDM Intel TI Fellows Forum

Significantly Alters Band Structure/Transport Biaxial Tension Heavy Hole Uniaxial Longitudinal Compression E K <110> Light Hole TI Fellows Forum Uniaxial Longitudinal Tension Valance Band warping, changes m*, m

Stress Contours 140 nm 45 nm 120 nm 30 nm -536 (μm) Si 0. 83 Ge 0. 17 -83 95 STI 40 3 31 MPa Source FLOOPs TI Fellows Forum (μm) STI

Device Level Calibration: SRC/Intel Funded • Industrial samples – – – 30 nm to 1 um Si trasistors from 3 companies Unstressed, uniaxial and biaxial stressed wafer Bulk and SOI Fully depleted SOI /Metal Gate High k/metal gate and sub-micron Ge channel devices TI Fellows Forum

Four-Point Bending Set-Up TI Fellows Forum

Mobility Enhancement (%) Strain Enhanced Mobility: Model / Measured 50 Uniaxial 40 Longitudinal 30 20 Biaxial 10 0 -300 -500 -10 -20 Uniaxial -30 Transverse -40 -50 Stress / MPa TI Fellows Forum Si. Ge S/D [4] Data Model 300 500 Biaxial Rim

6 Band K P Including Confinement Schrodinger’s Equation and Poisson’s Equation solved self-consistently using the Finite-Difference Method. TI Fellows Forum

Si and Ge Band Structure Si HH Ge LH HH LH No Stress Top Biaxial Stress Longitudinal compression TI Fellows Forum Bottom Top Bottom

In and Out-of-Plane Masses (Ge) Uniaxial Stress Biaxial Stress m| m|| kz kx TI Fellows Forum m| ky m|| kz ky kx

Si and Ge Band Structure on (100) and (110) Longitudinal compression Si Top (100) (110) hybrid TI Fellows Forum Ge Bottom Top Bottom

Full Transport Model: Calculation of Density of States Si is confined in kz direction. 2 -dimensional density of state is given by: And total charge density over all possible bands: TI Fellows Forum

Density of States Mass Longitudinal compressive 2. 5 0 Effective Mass m*/m 3 2 1. 5 Production level stress 1 0. 5 0 TI Fellows Forum Biaxial tensile 0. 1 1 Stress / GPa 10

Summary / Next Steps • First-principles quantum mechanical methods for strained Si band structure • Spatially dependent strain-induced band structure • Model charge transport and collection due to single event in FLOOPS/FLOODS – Start with existing MURI developed models – Add strain for Si and Ge transistors TI Fellows Forum