Microsoft PowerPoint - 20090525_ppt.ppt []Yia-Chung Chang
Research Center for Applied Sciences, Academia Sinica Taipei, Taiwan, R.O.C.p , ,
May 22 2009 at NCKUMay 22, 2009 at NCKU
Collaborators: Hanyou Chu, George Li, J. Opsal (Thermawave Inc., USA) Shih-Hsin Hsu, Pei-Kuen Wei, Chih-Ming Wang (RCAS, Taiwan) Ding-Ping Tsai (NTU)
1
IntroductionIntroduction Nanometrology allows optical inspection of the
geometry of nanostructures down to 10nm scale.
It uses a best fit to the measured ellipsometric spectraIt uses a best fit to the measured ellipsometric spectra via theoretical simulation (with efficient software) to determine the critical dimensiondetermine the critical dimension.
If done correctly, one can reconstruct images of nm resolution by using an optical instrument (with wavelength longer than 100nm).
It is noninvasive and capable of probing buried structures
2
structures..
3
Buried Self-assembled Quantum Dots [Shih-Yen Lin (RCAS)]
Tpyro=485 oC Tpyro=500 oCTpyro=485 oC Tpyro=500 oC
Tpyro=515 oCTpyro=515 oC
pyropyro
0.6
0.8
1.0
1.2
1.4
en si
ty (a
0.2
0.4
O
SEM studies
analyzer
E p-plane
plane of incidence
In Jones Matrix representation, ellipsometry measures Ψ and Δ in:
( ) ( )spi
s
p
s
n = εn =
)()()()( λελελλ iinn ++
Band structure of GaAs & Dielectric functionBand structure of GaAs & Dielectric function
Eo’
Si ε1 + i ε2
refracted at each interface. Amplitude, phase & polarization
are changed by
Film Stack
CD, Profile)CD, Profile) Contributes to Unique Ellipsometry Spectral
Amplitude of Four Fourier Coefficients,
DC Si (2 ) Si (4 )DC, Sin(2ωt), Sin(4ωt), Cos(4ωt))
vs Wavelength
Rigorous Coupled-Wave Analysisg p y
• Most used method for rigorous analysis of optical g y p diffraction by periodic gratings.
• For multilayer system with definite periodicityFor multilayer system with definite periodicity • EM fields in each layer are expanded in terms of a
finite number (N) of plane wavesfinite number (N) of plane waves E(x,z)=Σn Cn(z) exp{iknx}; kn=2πn/p
• The coefficients Cn(z) are dertermined by matching boundary conditions at all interfaces.
• The method is accurate (as long as enough plane waves are used) by not necessarily efficient (scale
12
waves a e used) by ot ecessa y e c e t (sca e like N3)
A typical fitting to the data of the scatterometry components of sin2ω, sin4ω, and cos4ω
1 01 0
200 300 400 500 600 700 800 -0.6 -0.4
200 300 400 500 600 700 800200 300 400 500 600 700 800 -0.6 -0.4
13
200 300 400 500 600 700 800200 300 400 500 600 700 800 Wavelength (nm)
200 300 400 500 600 700 800
The RT/CDTM technology combines the standard Opti-Probe® rotating compensator spectroscopic ellipsometer (RCSE) 4 and a high performance (10 GHz) server. RCSE spectra from samples with periodic line/space structures taken on the Opti-Probe®spectra from samples with periodic line/space structures taken on the Opti Probe® are sent to the server that performs the real-time regression analysis and returns calculated CD profiles as well as underlying film thicknesses.
CD jobfile
H ei
H ei
gh t (
0Resist ARC
0
0
(Data provided by Themawave Inc )
15 * This data is wafer dependent. Customer specific data must be gathered with customer wafers.
(Data provided by Themawave Inc.)
Application 2Application 2 Shallow Trench Isolation
1CD Sidewall angle C
16
(Data provided by Themawave Inc.)
Evolution of the real time regression fitting from beginning (top figures) to final solution (bottom figures) illustrating how adding feature detail improves the quality of the fit.(Symbols: : DC, : Sin(2wt), : Sin(4wt),: Cos(4wt)) to measurement results (Lines)
400
100
200
300
400
500
Wavelength (nm)
300 400 500 600 700 -0.8
Single {CD1 4 tPR tArc } recipe for monitoring a DI wafer under
nm )
nm )
nm )
0
0
50
Single {CD1-4, tPR, tArc,} recipe for monitoring a DI wafer under focus exposure matrix (FEM)
-150-100 -50 0 50 100 150
H ei
gh te
H ei
gh te
H ei
gh te
-250
-200
-150
-100
-50
H ei
gh te
X Data
X Data
X Data
H ei
gh te
X Data
0
50
100
H ei
X Data
-150-100 -50 0 50 100 150 -150-100 -50 0 50 100 150
X Data
X Data
X Data
X Data
100
150
200
250
300
X Data
-150-100 -50 0 50 100 150-150-100 -50 0 50 100 150 -50
0
50
X Data
-150-100 -50 0 50 100 150-150-100 -50 0 50 100 150
-150-100 -50 0 50 100 150
CD (nm)
-150-100 -50 0 50 100 150-150-100 -50 0 50 100 150-150-100 -50 0 50 100 150
18
Comparison of various methods method complexity scaling best for
RCWA (rigorous coupled wave analysis): 1 10MN3 straight grating( g p y ) g g g
Finite-difference (transfer-matrix method): 1 MN3 any 2D grating
G0 (multi-layer Green’s function): 3 3KMN2 low contrast 2D/3D
G1 (ideal grating Green’s function): 5 10N3+5kMN2 high contrast 2D/3D
BIM (boundary-integral method): 8 (2L+fN)(2L)2 isolated structure
BEM (bo ndar element method) 10 K(2L+fN)(2L) isolated str ct reBEM (boundary-element method): 10 K(2L+fN)(2L) isolated structure
N=number of coupled waves
M=number of slices
K=number of iterations
f=prefactor for evaluating GF ~ 20
19
CD metrology
20
[Y. C. Chang et al., J. Opt. Soc. Am. A 23, 638 (2006)]
Numerical procedures
22
23
24
25
Ellipsometric measurements of Au nanoparticles on a substratenanoparticles on a substrate
Motivation Au nanoparticles of definite size can be obtained commerciallyAu nanoparticles of definite size can be obtained commercially. They can be placed on insulating substrate and serve as a mask for making nanostructures.g It is desirable to have a characterizing tool that can detect the density and uniformalty of the distribution of Au nanoparticles on a substrate. Ellipsometry is an optical characterizing tool that is noninvasivce, can penetrate deep into the material, and can be used in-situ during growth.
Here we try to explore the capability of elliposmetry for characterizing nanoscale structures on surface or in burried
27
Nanorods Sample PreparationNanorods Sample Preparation (by Gong-Ru Lin, National TaiwanUniversity)
Developed pattern of Ni particles on o ide on Siparticles on oxide on Si
Transfer oxide pattern
28
0c m
20 0
Sample Preparationp p APTES (3-aminopropyl triethoxysilane)*
purity: >=98 0%purity: >=98.0% linear formula: H2N(CH2)3Si(OC2H5)3
formula weight: 221.37g refractive index: 1.422 @ λ0 = 589 nm
Sample Preparation Steps:Sample Preparation Steps: silanization: immersed in 1% APTES + 1mM acetic acid for 10 min. rinsed with D.I. water and bake at 120ºC for 30 min. immersed in aqueous solution with Au nanoparticles for 2 hours. rinsed with D.I. water and dried with N2 gas.
30* http://www.sigmaaldrich.com/catalog/search/ProductDetail/FLUKA/09324
d = 20nm d = 60nm
31d = 40nm d = 80nm
Rotating analyzer ellipsometer configuration (RAE) with the addition of AutoRetarderaddition of AutoRetarder
Dry nitrogen purge to eliminateDry nitrogen purge to eliminate absorption from ambient water vapor and oxygen
Spectral range: 140nm ~ 1700nm
32
8
0 80 nm
Ph t ( V) 1 2 3 4 5 6 7 8
4
Model fits
• Here we assume Au nannoparticles are sphere like p p (with diameter d) and placed uniformly on a square lattice with pitch p. The spheres are modeled by 5 to p p p y 15 slices of discs.
• We use fixed optical dielectric constants of bulk AuWe use fixed optical dielectric constants of bulk Au and substrate.
• d=20 40 60 80 nm• d=20, 40, 60, 80 nm • picth varies between 40nm-1000nm • Square lattice + hexagonal lattice • Calculations performed by GF & RCWA methods
34
-20
0
20
14
16
18
ε1
-60
-40
10
12
-140
120
0
2
4
35
data from E. D. Palik, Handbook of Optical Constants of Solids
Models Fittingg Au nanoparticles (20 nm), angle of incidence: 60º, 65º
Pitch =80nm
6 60o
36
1 2 3 4 5 6 7 8 Photon energy (eV)
1 2 3 4 5 6 7 8 Photon energy (eV)
Model Fittingg Au nanoparticles (40 nm), angle of incidence: 60º, 65º
Pitch =160nm 20
1 2 3 4 5 6 7 8 9 4
6
0
10
37
Photon energy (eV) Photon energy (eV)
Model Fittingg Au nanoparticles (60 nm), angle of incidence: 60º, 65º
Pitch =200nm
80
exp.
8
10 65o
Photon energy (eV) Photon energy (eV)
Model Fittingg Au nanoparticles (80 nm), angle of incidence: 60º, 65º
Pitch =300nm 28
exp. d l
8
10 65o
Ellipsometry Measurement vs. simulation
Au nanoparticles: 20, 40, 60, 80 nm angle of incidence: 60º
24
26
24 20 nm 40 nm 60 nm 80 nm 20
22
14
16
18
6
8
1 2 3 4 5 6 7 8 9 4
6
8
40
1 2 3 4 5 6 7 8 Photon energy (eV) Photon energy (eV)
Ellipsometric measurement vs. simulation Au nanoparticles: 20, 40, 60, 80 nm angle of incidence: 60º
80 Au NP, 60o
Δ (
30
40
Δ (
0
1 2 3 4 5 6 7 8 9 10
20 80 nm
41
1 2 3 4 5 6 7 8 Photon energy (eV) Photon energy (eV)
Ellipsometric measurement vs. simulation Au nanoparticles: 20, 40, 60, 80 nm angle of incidence: 55º
22
22 20 nm 40 nm 60 nm 80 nm 18
20
12
14
16
4
6
1 2 3 4 5 6 7 8 9 2
4
6
42
1 2 3 4 5 6 7 8 Photon energy (eV)
1 2 3 4 5 6 7 8 9 Photon energy (eV)
Ellipsometric measurement vs. simulation Au nanoparticles: 20, 40, 60, 80 nm angle of incidence: 55º
180 20 nm
140
100
120
40
60
40
60
43
Photon energy (eV) 1 2 3 4 5 6 7 8
Photon energy (eV)
Ellipsometric Measurement vs. simulation
Au nanoparticles: 20, 40, 60, 80 nm angle of incidence: 65º
28
26 20 nm 40 nm 60 nm 80 nm 24
26 20 nm
18
20
22
12
14
12
14
44
1 2 3 4 5 6 7 8 Photon energy (eV)
1 2 3 4 5 6 7 8 9 Photon energy (eV)
Ellipsometric Measurement vs. simulation
Au nanoparticles: 20, 40, 60, 80 nm angle of incidence: 65º
60 20 nm
40
20
30
0
Photon energy (eV) 1 2 3 4 5 6 7 8 9
0
AFM tipAFM tip
imaging lens
Summary Samples with different sizes of Gold nanoparticles immobilized on
a glass substrate are investigated by variable-angle spectroscopic
ellipsometry (VASE) in the UV to near IR region.
Both the Green’s function method and rigorous coupled-wave
analysis (RCWA) were used to model the ellipsometric spectra
GF method is 10 – 100 times more efficient than RCWA in most
cases.
ellipsometric measurements.
This demonstrates that the spectroscopic ellipsometry could be a
f l l id i f i b h i d d i fuseful tool to provide information about the size and density of
nanoparticles deposited on insulating substrate.
Th t h i b t d d t i t b i d t t
47
Plasmonic Infrared emitter
ε
2.5
x1
y2 In te
0
2
4H
0.5
1.0
x3
x5
48
-1500 -1000 -500 0 500 1000 1500x (nm) 0 25 50 75 100 125 150 175 200 225 250 0.0
x3
z (nm)
Angle-dependent reflectance spectrag p p L=3000nm, fL=1900nm, tg=100nm, tw=25nm
E i t RCWA i l ti
49
Relation between the reflection and emissionRelation between the reflection and emission / 2
( ) ( ) 1 ( , ) cos( )Emission TR R d π
λ λ λ θ θ θ
= − ∫
Reflectance Emission 0
∫
0.4
0.6
Em
0.0
0.2
50
Each resonant mode in the reflectance spectrum corresponds to a emission peak.
Summary
experimentally and theoreticallyexperimentally and theoretically.
The grating-coupled and localized SPP modes were identified.
Multiple LSPP peaks of the metal/dielectric/metal cavity areMultiple LSPP peaks of the metal/dielectric/metal cavity are
suppressed in the T-gate structure.
The angular dependent reflectance spectrum of T-gate arrayThe angular dependent reflectance spectrum of T gate array
shows a clear resonant dip, whose wavelength can be tuned by the
permitivity of the filler material.p y
. Using Kirchhoff’s law, the thermal radiation of the proposed
structure shows a sharp peak at 3.5mm with low sideband emission.p p
This narrow bandwidth emission is very useful for studies of
reactions of biological systems, environmental surveillance, and
51