Transcript Process Simulation

School of Microelectronic Engineering
Semiconductor Physics 2007/2008
• 2D cross-section of wafer
– X-coordinate: parallel to the wafer surface
– Y-coordinate: depth into the wafer
• Grid structure:
– The continous physical process are modeled
numerically by using finite difference (for
diffusion) and finite element (for oxide flow)
solution techniques.
– Each region is divided into a mesh of nonoverlapping triangular elements
– Solution values are calculated at the mesh nodes
(at the corners of the triangular elements), value
between the nodes are interpolated
School of Microelectronic Engineering
Semiconductor Physics 2007/2008
STRUCTURE SPECIFICATION
•
INITIALIZE – sets up the initial structure (mesh, background doping,
orientation, resistivity, etc.) for a simulation
INITIALIZE
{ ( IN.FILE=<c> { ( [SCALE=<n>] [FLIP.Y] ) | TIF {REGION=<c>
NEW.MATERIAL=<c> [KEEP] } } ) ] ( [WIDTH=<n> [DX=<n>] ] [ { <111> |
<110> | <100> | ORIENTAT=<n> } ] [ { ROT.SUB=<n> | X.ORIENT=<n> } ]
[RATIO=<n>] [LINE.DAT] ) } [IMPURITY=<c> {I.CONC=<n> | I.CONC=<c> |
I.RESIST=<n>} ] [MATERIAL=<c>] [ANTIMONY=<c>] [ARSENIC=<n>]
[BORON=<n>] [PHOSPHOR=<n>] [ { CONCENTR | RESISTIVITY } }
Note:
[ ], { }, | and ( ) are not part of the syntax. [ ] indicates optional parameters, { |
} indicates alternate choices, and ( ) indicate group or single item in higher
level
School of Microelectronic Engineering
Semiconductor Physics 2007/2008
STRUCTURE SPECIFICATION
• A structure must be initialized before any processing
steps (after setting up the grid automatically, manually, or
from a file)
• Grid can be generated automatically using INITIALIZE
• Manual grid is generated using LINE statement and also
along with REGION, BOUNDARY and ELIMINATE
statements
• Grid can be read from a saved structure using
INITIALIZED or LOADFILE statement
• MESH specifies grid spacing factor and default values
for controlling automatic grid generation
School of Microelectronic Engineering
Semiconductor Physics 2007/2008
STRUCTURE SPECIFICATION
•
•
MESH [GRID.FAC=<n>] [DX.MAX=<n>] [DX.MIN=<n>] [DX.RATIO=<n>]
[LY.SURF=<n>] [DY.SURF=<n>] [LY.ACTIV=<n>] [DY.ACTIV=<n>]
[LY.BOT=<n>] [DY.BOT=<n>] [DY.RATIO=<n>] [FAST]
LINE {X | Y } LOCATION=<n> [SPACING=<n>] [TAG=<c>]
WIDTH
DX.MIN
DX.MAX
x
DY.SURF
y
LY.SURF
DY.ACTIV
DY.RATIO
LY.ACTIV
LOC
LINE Y
SPAC
LINE X
DY.BOT
SPAC
LY.BOT
DX
LOC
DX.RATIO
School of Microelectronic Engineering
Semiconductor Physics 2007/2008
STRUCTURE SPECIFICATION
Example:
LINE X LOC=0.0 SPAC=0.15
LINE X LOC=1.25 SPAC=0.05
LINE X LOC=1.5 SPAC=0.1
LINE Y LOC= 0 SPAC=0.03
LINE Y LOC=0.5 SPAC=0.1
INITIALIZE
School of Microelectronic Engineering
Semiconductor Physics 2007/2008
PROCESS STEPS
• DIFFUSION
– To perform:
• Diffusion – non-oxidizing ambient
• Oxidation – oxidizing ambient
• Salicidation – user-specified material, impurity and reaction
– Ambients:
• DRY (100% O2), WETO2 (92% H2O+8% N2), STEAM (100% H2O),
INERT (100% N2 or other inert gasses), AMB.1, … AMB.5 (userdefined)
– Examples:
• 30-minute boron pre-deposition
DIFFUSION TIME=30 TEMP=1000 BORON=1E20
• 60-minute dry oxidation at 900oC with ambient containing 2% HCl
DIFFUSION TIME=60 TEMP=900 HCL
School of Microelectronic Engineering
Semiconductor Physics 2007/2008
PROCESS STEPS
• IMPLANTATION
– To simulate the implantation of ionized impurities (dose
and range) into the simulation structure for accurate
subsequent thermal cycling
– Impurity distribution can be obtained from:
• Analytical models: Gaussian or Pearson
• Table of range statistics
• Numerical model: Monte Carlo
– Implant damage model – models the transition from
crystalline to amorphous material
– Example:
IMPLANT BF2 DOSE=1E13 ENERGY=50 TILT=7
School of Microelectronic Engineering
Semiconductor Physics 2007/2008
PROCESS STEPS
• DEPOSITION
– To perform conformal deposition of materials at
temperature low enough that impurity diffusion can be
ignored
– Materials: SILICON, OXIDE, OXYNITRI, NITRIDE,
POLYSILI, ALUMINUM, PHOTORES or by specifying
material name with MATERIAL statement
– Deposited layer can be doped with one or more impurities
– Example:
DEPOSIT OXIDE THICK=0.02
School of Microelectronic Engineering
Semiconductor Physics 2007/2008
PROCESS STEPS
• MASKING, EXPOSURE & DEVELOPMENT OF PHOTORESIST
– Masking information is read from a mask file created by Taurus
LAYOUT.
– Idealized model – photoresist lines always have vertical sidewalls,
positioned directly beneath mask edges.
– The EXPOSE statement uses the x coordinates to determine which
portions of the photoresist should be marked as exposed
– The DEVELOP statement removes all the positive photoresist that has
not been marked as exposed
– Example:
MASK IN.FILE=CMOS3.TL1
…………………..
DEPOSIT POLY THICKNESS=0.2
DEPOSIT POSITIVE PHOTORES THICKNESS=1
EXPOSE MASK=POLY SHRINK=0.05
DEVELOP
School of Microelectronic Engineering
Semiconductor Physics 2007/2008
PROCESS STEPS
• ETCHING
– To remove structure but not intended to simulate a physical
etching process
– Types:
• Trapezoidal etch model – dry etching with masked undercutting and angled
sidewalls
ETCH MATERIAL TRAPEZOI THICKNES (depth) ANGLE (sidewall slope)
UNDERCUT (horizontal penetration masking layer edges)
• Removal of a region to the left or right of a line
ETCH MATERIAL {LEFT|RIGHT} P1.X P1.Y P2.X P2.Y
• Removal of arbitrary region
ETCH MATERIAL {START|CONTINUE|DONE} X Y
• Removal of the entire structure
ETCH ALL
• Removal using Taurus-Topography (physical etching simulator)
School of Microelectronic Engineering
Semiconductor Physics 2007/2008
PROCESS STEPS
• ETCHING
– EXAMPLES:
Etch nitride to the left of 0.5m to a depth of 1m
ETCH NITRIDE LEFT P1.X=0.5 P2.Y=-1
Etch oxide region defined by (0,0), (1,0), (1,1), (0,1)
ETCH OXIDE START X=0 Y=0
ETCH CONTINUE X=1 Y=0
ETCH CONTINUE X=1 Y=1
ETCH DONE X=0 Y=1
School of Microelectronic Engineering
Semiconductor Physics 2007/2008
PROCESS STEPS
• EPITAXY
– To grow silicon on exposed silicon surface
– Example:
$ Epitaxy done by alternately depositing 1m Si and diffusing antimony
with 1E19 cm-3 in 10 steps
EPITAXY THICK=1 TIME=180 TEMP=1100 ANTIMONY=1E19
SPACES=10
• STRESS
– To calculate the stresses caused by thermal mismatch between
materials or due to intrinsic stress in deposited films
– STRESS [TEMP1 TEMP] [NEL]
– Example:
$ Define intrinsic stress of nitride and calculate stresses in substrate
and nitride.
Material nitride intrin.s=1.4e10
STRESS
School of Microelectronic Engineering
Semiconductor Physics 2007/2008
MODELS & COEFFICIENTS
• Model – a mathematical abstraction of a physical phenomenon (e.g.,
diffusion equation, Deal an Grove model for oxidation, etc.)
• Coefficients – parameters used in a model
• Choosing/executing models
– METHOD, DIFFUSION, DEPOSITION, IMPLANTATION, …
• Setting coefficients
– AMBIENT, MOMENT, MATERIAL, IMPURITY, REACTION,
MOBILITY, INTERSTITIAL, VACANCY, ANTIMONY, ARSENIC,
BORON, PHOSPHORUS, …
School of Microelectronic Engineering
Semiconductor Physics 2007/2008
MODELS & COEFFICIENTS
• ERFC – simplest oxidation model, accurate for 1D, limitations: no
PolySi oxidation and does not take masking layer into account
• ERFG – oxidation model under nitride mask. ERF1 for Thicknitride <
Thickpadox, ERF2 if otherwise, same limitations as ERFC.
• Vertical – simplest and fastest numerical ox. Model; useful for
uniform planar oxidation; default.
• Compress – simulates viscous flow of oxide; 2D movement of
interface; crystal orientation effect; allow polySi oxidation; suitable
for non-planar, small amount of oxide (no stress effect) and noncritical oxide shape.
• Viscous – as compress with surface tension added.
• Viscoela – adds elastic component to viscous and compress;
recommended for 2D where shapes are important; use with
STRESS.D; faster than viscous when stress dependence is
considered
School of Microelectronic Engineering
Semiconductor Physics 2007/2008
MODELS & COEFFICIENT
• PD.FERMI –simplest and fastest model where
point defects concentrations are assumed at their
thermal equilibrium; does not model oxidation –
enhanced diffusion, high concentration or implant
damage effects; set as default
• PD.TRANS – performs full, transient simulation of
two dimensional point defect distributions
• PD.FULL – complete diffusion model for maximum
accuracy, when using implant damage model or
when high concentration effect is important
School of Microelectronic Engineering
Semiconductor Physics 2007/2008