dbit dehradun
TRANSCRIPT
-
8/7/2019 DBIT Dehradun
1/24
A Presentation on
Growth kinetics for thermal oxidation of silicon, thespecial case for VLSI/MEMS technology
ByPrachi Sharama
Dept. of ECE, Banasthali University, Banasthali.
4/11/2011 1
Compiled by:-
1. Kulwant Singh2. Soney Varghese3. Uvais V N4. Prachi Sharama5. Ashish Kumar
-
8/7/2019 DBIT Dehradun
2/24
Overview
SiO 2 and the Si/SiO 2 interface are the principal reasons for silicons dominancein the IC industry.
10 nm
0.1 m
1 m
1 nm
Masking Oxides
Gate OxidesTunneling Oxides
Field Oxides
Pad Oxides
Chemical Oxides from CleaningNative Oxides
Thermally Grown OxidesOxide
Thickness
Deposited Oxides
Backend InsulatorsBetween Metal Layers
Masking Oxides
4/11/2011 2
T hickness & uses of SiO 2 in Si Technology
-
8/7/2019 DBIT Dehradun
3/24
4/11/2011 3
-
8/7/2019 DBIT Dehradun
4/24
A dvantages/ applications SiO 2:-
Easily selectively etched usinglithography. Masks most common impurities(B, P, A s, Sb).
Excellent insulator ( ).
High breakdown field( ) Excellent junction passivation. Stable bulk electrical properties. Stable and reproducible interface
with Si.
V "1016 ; cm, E g " 9 eV
10 7 Vcm -1
No other known semiconductor/insulator combination has properties thatapproach the Si/SiO 2 interface.
4/11/2011 4
-
8/7/2019 DBIT Dehradun
5/24
Future Projection Si TechnologyYear of Production 1998 2000 2002 2004 2007 2010 2013 2016 2018
Technology Node (half pitch) 250 nm 180 nm 130 nm 90 nm 65 nm 45 nm 32 nm 22 nm 18 nm
MPU Printed Gate Length 100 nm 70 nm 53 nm 35 nm 25 nm 18 nm 13 nm 10 nm
DRAM Bits/Chip (Sampling) 256M 512M 1G 4G 16G 32G 64G 128G 128G
MPU Transistors/Chip (x10 6) 550 1100 2200 4400 8800 14,000
Gate Oxide T ox Equivalent (nm)
MPU
1.2 0.9 0.7 0.6 0.5 0.5
Gate Oxide T ox Equivalent (nm)
Low Operating Power
1.5 1.2 0.9 0.8 0.7 0.7
Gate Dielectric Leakage
(nA/m @ 100 C) MPU
170 230 330 1000 1670 1670
Thickness Control (% 3 W) < 4 < 4 < 4 < 4 < 4 < 4
Min Supply Voltage (volts) 1.8-2.5 1.5-1.8 1.2-1.5 0.9-1.2 0.8-1.1 0.7-1-0 06-0.9 0.5-0.8 0.5-0.7
Adapted from NTRS SIA, 1997
4/11/2011 5
-
8/7/2019 DBIT Dehradun
6/24
Manufacturing Methods & Equipment
Qu artz
ub eWafers
Qu artz
arrier
esista ce Heati g
H2
2Conceptual Silicon Oxidation System
Oxidation systems are conceptuallyvery simple.
In practice today, vertical furnaces,RT O systems and fast ramp furnacesall find use.
T hermal oxidation potentially being used in many places in chipfabrication. In practice, deposited SiO 2 layers are increasingly being used
in case of back-end process.
4/11/2011 6
Si O 2 p SiO 2
Si 2H 2O p SiO 2 2H 2
-
8/7/2019 DBIT Dehradun
7/24
Charge Associated with the SiO 2/Si
System
++++ +xxxxxx
+++ -
--
K +
TransitionRegion
Na +SiO 2
Q m
Q ot
Q f
Q itSilicon
Four charges are associated withinsulators and
insulator/semiconductor interfaces.
Q f - fixed oxide charge Q it - interface trapped charge Q m - mobile oxide charge Q ot - oxide trapped charge
4/11/2011 7
-
8/7/2019 DBIT Dehradun
8/24
Measurement Methods
Physical measurement Scanning tunneling microscope(ST M)
Atomic Force microscope (A FM) SEM etc..
Optical Measurement Color chart by Pliskin and Conrad
Electrical measurement- The C MOS Capacitor
4/11/2011 8
-
8/7/2019 DBIT Dehradun
9/24
A . Planar growth kinetics
Ox i e
I
O
C S
C I
x O
Gas
0.01 - 1 m - 500 m
Silico
1 2 3
T he basic model for oxidationwas developed in 1960 by Dealand Grove.
Si O 2 SiO 2
Si 2H 2O S iO 2 2H 2
(1)
(2)
F 1 ! G C G C S
F2 !
xxx
! D C O C Ix O
F 3 ! k S C I
(3)
(4)
(5)4/11/2011 9
-
8/7/2019 DBIT Dehradun
10/24
Under steady state conditions, F 1 = F 2 = F 3 so
C I !C *
1k Sh
k S x OD
$ C*
1k S xO
D
(6)
C O
C * 1k S x O
D
1k S
h
k S x O
D
C * (7)
Note that the simplifications are made by neglecting F 1 which is a very goodapproximation.
Combining (6) and (7), we have
d xd t !
F
N1 !
k S C*
N1 1
k Sh
k S xOD
Integrating this equation, results in the linear parabolic model.
(8)
4/11/2011 10
-
8/7/2019 DBIT Dehradun
11/24
x O2 x i
2 xO x i/
! t (9)
where (parabolic rate constant) (10)
(11)
B
A! C
*
N1
1k
S
1h
$ C*k S
N1
(linear rate constant)and
(9) can also be written with oxide thickness as a function of time.
x O !A
21
t XA 2 / 4 B
1
(12)
where X x i
2 Ax iB
(13)
4/11/2011 11
B ! 2DC*
N 1
-
8/7/2019 DBIT Dehradun
12/24
T he rate constants B and B/ A have physical meaning (oxidant diffusion andinterface reaction rate respectively).
B ! C 1 exp E1 / k
B
A! C 2 exp E2 / k
(14)
(15)
A m b ie t B B /A
Dry O 2 C 1 = 7.72 x 102 2 h r -1
E 1 = 1.23 eC 2 = 6.23 x 10
6 h r -1
E 2 = 2.0 e
Wet O 2 C 1 = 2.14 x 102 2 h r -1
E 1 = 0.71 eC 2 = 8.95 x 10
7 h r -1
E 2 = 2.05 e
H 2O C 1 = 3.86 x 102 2 h r -1
E 1 = 0.78 eC 2 = 1.63 x 10
8 h r -1
E 2 = 2.05 e
Plots of B, B/ A using thevalues in the above Table.
Numbers are for (111) silicon,for (100) divide C 2 by 1.68.
.
0.0001
0.001
0.01
0.1
1
10
100
0.65 0.7 0.75 0.8 0.85 0.9 0.95 1
B
m2
hr- 1
B/A m hr
- 1
1000/ (Kelvi )
800900100011001200
( C )
B /A H2O
B /A Dry O 2B Dry O2
B H 2O
4/11/2011 12
-
8/7/2019 DBIT Dehradun
13/24
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 2 4 6 8 10
Oxide hickess - micros
ime - hou rs
1200 C
1100 C
1000 C
900 C800 C
0
0.5
1
1.5
2
0 1 2 3 4 5 6 7 8 9 10
Oxide hickess - micros
ime - hou rs
1100 C
700 C
1000 C
900 C
800 C
Calculated (100) silicon dry O 2oxidation rates using Deal Grove.
Calculated (100) silicon H 2Ooxidation rates using Deal Grove.
4/11/2011 13
-
8/7/2019 DBIT Dehradun
14/24
B. T hin Oxide Growth KineticsA major problem with the Deal Grove model was recognized when it was firstproposed - it does not correctly model thin O 2 growth kinetics. Experimentally O
2oxides grow much faster for 20 nm than Deal Grove predicts
M A NY models have been suggested in the literature.1. Reisman et. al. Model
x O ! a t t i b or xO ! a t x ia
1b
b
(16)
Power law fits the data for all oxide thicknesses. a and b are experimentally extracted parameters. Physically - interface reaction controlled, volume expansion and viscous flow
of SiO 2 control growth.2. Han and Helms Model
dx Odt
B12x O A 1
B22x O A 2
(17)
Second parallel reaction added - fits the data for all oxide thicknesses.T hree parameters (one of the A values is 0).
4/11/2011 14
-
8/7/2019 DBIT Dehradun
15/24
3. Massoud et. al. Model
d x Od t
!B
2x O AC exp
x OL
(18)
Second term added to Deal Grove model - higher dx/dt during initial growth. L 7 nm, second term disappears for thicker oxides. Easy to implement along with the DG model, @ used in process simulators. Data agrees with the Reisman, Han and Massoud models. (800C dry O 2 model
comparison below.)
.
0
0.005
0.01
0.015
0.02
0.025
0.03
0 2 4 6 8 10
Oxide Thickness - microns
an &
elms Model
Reisman et. al. Model
Time - hours
Deal Grove Model ( X 0)
Deal Grove Model ( X 8 hrs)
4/11/2011 15
-
8/7/2019 DBIT Dehradun
16/24
1
1
1.3
11
1
1.2
1
1
1
1.3
1.3
Si substrate Si substrate
Si 2
e site si i
Si Substrate
ri i a Si Sur a eu e
a si
a t i S i 3 a s
Oxidation involves a volumeexpansion
( 2.2X). Especially in 2D and 3Dstructures, stresseffects play a dominant role.
4/11/2011 16
C. 2D SiO 2 Growth Kinetics
-
8/7/2019 DBIT Dehradun
17/24
Etched Si Ring
Si Substrate
Side Views Top Views
Polysilicon
SiO 2
Si
a)
b)
c)
d)
T hese effects were investigated indetail experimentally by Kao et. al.about 15 years ago.
Typical experimental results below.
(Kao et.al)4/11/2011 17
-
8/7/2019 DBIT Dehradun
18/24
.
0 1 2 3 4 5 6 7 8 1/r m -1
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1.1
Normalized Oxide Thickness
Convex Radii
Concave Radii
1200 C
1100 C
1000 C
900 C
800 C
1100 C
1000 C900 C
1 m 0.2 m 0.125 m
Several physical mechanisms seem to beimportant: Crystal orientation 2D oxidant diffusion Stress due to volume expansion
To model the stress effects, Kao et. al.suggested modifying the Deal Groveparameters.
k S (stress ) ! k S exp Wn VR k
exp Wt V T
k T
D (stress ) ! D expP VD
k T
C *(stress ) ! C * exp P VS k T
(19)
(21)
(20)
where and are the normal andtangential stresses at the interface.VR, VT and V S are reaction volumes andare fitting parameters.
(Kao et.al)
Wn
Wt
4/11/2011 18
-
8/7/2019 DBIT Dehradun
19/24
In addition, the flow properties of the SiO 2 need to be described by a stressdependent viscosity
L(stress ) ! L(T) WS VC / 2k Tsinh W
SV
C/ 2k T
(22)
Where is the shear stress in the oxide and V C is again a fitting parameter.
WS
arameter alue0.0125 m 3 0.0065 m 3
S, 00.3 m 3 @ 850
0.72 m 3 @ 1050 ( ) - Si 2 3.13 10 10 e (2.19 e /k ) oise( ) - Si 3N4 4.77 10 10 e (1.12 e /k ) oise
T hese models have been implemented in modern process simulators and allowthem to predict shapes and stress levels for VLSI structures (above right).
AT HEN A simulation: Left - no stress dependent parameters, Right - includingstress dependence.
.
Silico n
SiO 2Si 3 N 4
0
0.2
0.4
0.6
0.8
-0.2
-0.4
icro ns
icrons
0 0.4 0.8-0.4-0.8 icro ns
0 0.4 0.8-0.4-0.8
4/11/2011 19
-
8/7/2019 DBIT Dehradun
20/24
D. Point Defect Based ModelsT he oxidation models we have considered to this point are macroscopic models(diffusion coefficients, chemical reactions etc.).
*
*
O 2
H 2O
D iff u sion
Ox id e Silico n
II
V
1 2K Si Si 2O I 2FV m SiO 2 2KI stress
T here is also an atomistic picture of oxidation that has emerged in recentyears.
Most of these ideas are driven by thevolume expansion occurring duringoxidation and the need for free volume.
In Chapter 3 we described internal oxidation in the following way:(23)
Surface oxidation can be thought of in the same way.
4/11/2011 20
-
8/7/2019 DBIT Dehradun
21/24
T he connection between oxidation and other processes can then be modeled asshown below.
B u rie d Dopan t arker Layer
G R
B u lk R ecom b inatio n
Su rfaceR ecom b inatio n
*
OEDInertDiff u sion
O2
I
.
Inert Diff u sion
OED
Microns
0
0.5
1.0
1.5
M icro ns0 1 2-1-2
Si 3 N 4SiO 2
Example - AT HEN A simulation of OED.
Oxidation injects interstitials to create free volume for the oxidation process. Oxidation can also consume vacancies for the same reason.T hese processes increase I concentrations and decrease V concentrations innearby silicon regions.
A ny process (diffusion etc) which occurs via I and V will be affected.
4/11/2011 21
-
8/7/2019 DBIT Dehradun
22/24
E. Complete Process Simulation of Oxidation
Many of these models (and others in Chapter 6), have been implemented inprograms like SUPREM.
.
0
0.4
0.8
Microns
0
0.4
0.8
Microns
Microns1-1 0
Simulation of an advancedisolation structure (the SW A MIprocess originally developed byHewlett-Packard), usingSSUPREM IV. T he structureprior to oxidation is on the topleft. A 450 min H 2O oxidation at1000 C is then performed whichresults in the structure on thetop right. A n experimental
structure fabricated with asimilar process flow is shownon the bottom right. T he stresslevels in the growing SiO 2 areshown at the end of theoxidation on the bottom left.
4/11/2011 22
-
8/7/2019 DBIT Dehradun
23/24
Summary of Key Ideas
T hermal oxidation has been a key element of silicon technology since itsinception.
T hermally, chemically, mechanically and electrically stable SiO 2 layers on silicondistinguish silicon from other possible semiconductors.
T he basic growth kinetics of SiO 2 on silicon are controlled by oxidant diffusionand Si/SiO 2 interface chemical reaction.
T his simple Deal-Grove model has been extended to include 2D effects, highdopant concentrations, mixed ambients and thin oxides.
Process simulators today include all the physical effects (and more) and arequite powerful in predicting oxidation geometry and properties.
4/11/2011 23
-
8/7/2019 DBIT Dehradun
24/24
4/11/2011 24