EE-612: Lecture 22: CMOS Process Steps Mark Lundstrom Electrical and Computer Engineering Purdue University West Lafayette, IN USA Fall 2006 NCN www.nanohub.org Lundstrom EE-612 F06 1
outline 1) Unit Process Operations 2) Process Variations Lundstrom EE-612 F06 2
unit process operations 1) Oxidation 2) Diffusion 3) Ion Implantation 4) RTA/RTP 5) Chemical Vapor Deposition 6) Lithography 7) Etching 8) Metalization 9) Well Structures 10) Isolation 11) Source / Drain structures Lundstrom EE-612 F06 3
useful references 1) J.D. Plummer, M.D. Deal, P.B. Griffin, Silicon VLSI Technology, Fundamentals, Practice, and Modeling, Prentice Hall, Upper Saddle River, NJ, 2000. 2) S.A. Campbell, The Science and Engineering of Microelectronic Fabrication, 2nd Ed., Oxford Univ. Press, New York, 2001. Lundstrom EE-612 F06 4
oxidation Si + O 2 SiO 2 Si+ H 2 O SiO 2 (dry) (wet) wafers heater O 2 or H 2 O +carrier gas T = 800-1100 o C fused quartz furnace tube Lundstrom EE-612 F06 5
oxidation and doping C phophorous m > 1 boron m < 1 x m = C Si C SiO2 Lundstrom EE-612 F06 6
local oxidation Si 3 N 4 SiO 2 Si Lundstrom EE-612 F06 7
local oxidation (LOCOS) Si 3 N 4 SiO 2 field oxide Si 'bird's beak' Lundstrom EE-612 F06 8
constant source diffusion dopant-containing gas (e.g. POCl 3 ) C C S C(x) = C S erfc( x /2 Dt) Q = C x,t dx = 2C S Dt π 0 time ( ) x Lundstrom EE-612 F06 9
Limited source diffusion C C( x,t)= ( Q / π Dt )exp x /2 Dt ( ) 2 C S t = 0 predep time x Lundstrom EE-612 F06 10
diffusion Si Si Si Si Si Si Si Si Si Si Si Si Si Si Si Si Si Si Si Si Si Si Si Si Si substitutional interstitialcy interstitial D(T ) = D 0 e E A /k B T oxidation enhanced diffusion Lundstrom EE-612 F06 11
ion implantation energetic ions bombard silicon wafer B magnet F = Q r υ r B acceleration deflection wafer I ion source Lundstrom EE-612 F06 12
ion implantation Si R P (E) N x ΔR P (E) ()= N p exp x R p ( ) 2 2ΔR p 2 implant damage (anneal) Q = 2π N p ΔR p Lundstrom EE-612 F06 13
ion implantation (ii) 1.0 B Projected range (μm) 0.1 As 0.01 10 100 1000 Acceleration energy (kev) Lundstrom EE-612 F06 14
channeling C() x ΔR P tilted 3 deg R P x Lundstrom EE-612 F06 15
rapid thermal annealing reflector lamps quartz window thermal budget Dt wafer gas inlet Lundstrom EE-612 F06 16
chemical vapor deposition reaction chamber Dt gas inlet wafer susceptor gas exhaust 2SiH 4 + 4NH 3 Si 3 N 4 + 12H 2 SiH 4 Si+2H 2 SiH 4 +O 2 SiO 2 +2H 2 silicon nitride poly silicon silicon dioxide Lundstrom EE-612 F06 17
plasma CVD / etching RF power in gas inlet heater wafer lower temperature reduces thermal budget Dt gas exhaust Lundstrom EE-612 F06 18
lithography optical source wavelength, λ lens shutter mask resist contact or proximity wafer expose, develop, etch Lundstrom EE-612 F06 19
projection printing α UV source lens 1 mask lens 2 wafer Lundstrom EE-612 F06 20
registration errors EE E E E EE EE misalignment run out Lundstrom EE-612 F06 21
phase shift lithography conventional mask phase shift mask electric field at mask electric field at mask intensity at wafer intensity at wafer Lundstrom EE-612 F06 22
pattern transfer negative resist (less soluble after exposure) wafer wafer positive resist (more soluble after exposure) resist: optically sensitive polymer which, when exposed to UV changes its solubility in specific chemicals wafer Lundstrom EE-612 F06 23
etching wet chemical etching (isotropic) dry etching (plasma or reactive ion etching - RIE) (anisotropic) wafer wafer undercut chemicals react with underlying material, but not resist ionized gases react with underlying material, but not resist Lundstrom EE-612 F06 24
pattern transfer (ii) mask chrome L Drawn lithography bias resist etch bias L Gate (physical) Lundstrom EE-612 F06 25
metalization Tungsten (W) plugs for first layer metal dep CMP www.itrs.net 2005 Edition Lundstrom EE-612 F06 26
outline 1) Unit Process Operations 2) Process Variations Lundstrom EE-612 F06 27
discrete doping effects V = W L x j N + N + example: L = 50 nm W = 100 nm x j = 25 nm N A = 10 18 cm -3 N TOT = 125 P (N A cm-3) Number of dopants in the critical volume is a statistical quantity Lundstrom EE-612 F06 28
discrete doping effects (ii) source drain Effects: 1) σ VT (10 s of mv) 2) lower avg. V T (10 s of mv) 3) asymmetry in I D 3D transport leads to inhomogeneous conduction (see Wong and Taur, IEDM, 1993, p. 705) Lundstrom EE-612 F06 29
discrete doping effects (iii) 35 nm MOSFET AFM measurements, Fujitsu (simulations from A. Asenov group, Univ. of Glasgow) Lundstrom EE-612 F06 30
statistical variability Line edge roughness discrete dopants From A. Asenov, Univ. of Glasgow Lundstrom EE-612 F06 31
variability is becoming a major issue G. Declerck, Keynote talk, VLSI Technol. Symp. 2005 Lundstrom EE-612 F06 32
outline 1) Unit Process Operations 2) Process Variations For a basic, CMOS process flow for an STI (shallow trench isolation process), see: http://www.rit.edu/~lffeee/advcmos2003.pdf Lundstrom EE-612 F06 33
CMOS process flow For a basic, CMOS process flow for an STI (shallow trench isolation process), see: http://www.rit.edu/~lffeee/advcmos2003.pdf The author is indebted to Dr. Lynn Fuller of Rochester Institute of Technology for making these materials available. What follows is a condensed version of a more complete presentation by Dr. Fuller. I regret any errors that I may have introduced by shortening these materials. -Mark Lundstrom 10/19/06 Lundstrom EE-612 F06 34