Ellipsometry

Data Analysis: a Tutorial

G. E. Jellison, Jr.

Solid State Division

Oak Ridge National Laboratory

WISE 2000

University of Michigan

May 8-9, 2000

 Motivation

The Opportunity:

Spectroscopic Ellipsometry (SE) is sensitive to

many parameters of interest to thin-film science,
such as

Film thickness

Interfaces

Optical functions (n and k).

But

SE data is not meaningful by itself.

Therefore

One must model the near-surface region to get

useful information.
 Outline

Data representations

Calculation of reflection from thin films

Modeling of optical functions (n and k)

Fitting ellipsometry data

Examples
 Introduction
What we measure:

Stokes Vector of a light beam

I I
o
o

Q I0 I
S = =
90

U I I
45 45

V I I
rc lc

Io (Q2 + U2 + V2 )1/2

To transform one Stokes vector to another:

Sout = M Sin

M = Mueller matrix (4X4 real)

 Introduction
What we measure:

light PSG PSD

source p p
Detector
s s
Sample

Intensity of beam at the detector = SPSDT M SPSG

The matrix M includes:

Sample reflection characteristics

Intervening optics (windows and lenses)

 Introduction
Isotropic Reflector:
Epin Epout

Esin Esout
Isotropic reflector

Mueller Matrix:

1 N 0 0

N 1 0 0
M reflector =
0 0 C S

0 0 S C

)
N = cos(2
) sin(
S = sin(2 )
) cos(
C = sin(2 )
N2 + S2 + C2 = 1
 Introduction
What we calculate:

light PSG PSD

source p p
Detector
s s
Sample

rpp rps op / ip so / ip
J = = o i
rsp rss p / s o
s / s i

ps tan( )e i i ps
tan( ps )e
= rss =
sp 1 tan( sp )e i sp
1 .

There is no difference between Mueller and Jones IF

there is no depolarization!
 Data Representations
Mueller-Jones matrices:

M = A ( J J* ) A-1,

1 0 0 1

1 0 0 1
A=
0 1 1 0
.
0 i 0
i

Assumes no depolarization!

E out C + iS
in
rp E
= = p p
= tan( ) e i
=
rs E out
s E in
s 1+ N
 Data Representations
Anisotropic samples:

1 N ps C sp + 1 S sp + 2

N sp 1 sp ps C sp + 1 S sp + 2
M=
C + 1 C ps + 1 C + 1 S + 2
ps
S + S ps + 2 S + 2 C 1
ps 2

Cps + iSps ips

ps = = tan( ps )e
1+ N
Csp + iSsp i sp
sp = = tan( sp )e
1+ N

N2 + S2 + C2 + Sps2 + Cps2 + Ssp2 + Csp2 = 1

ps tan( )ei tan( ps )e ps

i

J = rss =

sp 1 tan(sp )e
isp
1
 Calculation of Reflection Coefficients
Single Interface: Fresnel Equations (1832):
0
0

1 cos( 0 ) 0 cos( 1 )
r pp =
1 cos( 0 ) + 0 cos( 1 )

0 cos( 0 ) 1 cos( 1 )
rss =
0 cos( 0 ) + 1 cos( 1 )

Snells Law (1621):

= 0 sin 0 = 1 sin 1.
 Calculation of Reflection Coefficients
Two Interfaces: Airy Formula (1833):

r1, pp ,ss + r2, pp ,ss exp(2ib)

rpp ,ss =
1 + r1, pp ,ss r2, pp ,ss exp( 2ib)

2d f
b= f cos( f )

 Calculation of Reflection Coefficients
Three or more interfaces: Abeles Matrices (1950):

Represent each layer by 2 Abeles matrices:

cos( j )
cos(b j ) i sin(b j )
j
P j , pp =
j
i sin(b j ) cos(b j )
cos( )
j

sin( b j )
cos( b j ) i
P j , ss = j cos( j )
i cos( ) sin( b ) cos( b j )
j j j

Matrix multiply:

N
M pp = 0, pp ( Pj , pp ) sub, pp
j =1

N
M ss = 0,ss ( P j ,ss ) sub ,ss
j =1
 Calculation of Reflection Coefficients
Three or more interfaces: Abeles Matrices:

Substrate and ambient characteristic matrices:

cos( ) 1
1 1
1 0 1 0 cos( )
0 , pp = 0 , ss =
2 1 cos( ) 2 1 1
0 0 cos( )

cos( sub ) 1
0 0
sub, pp = sub sub,ss = sub cos( sub )

1 0 1 0

Final Reflection coefficients:

M 21 , pp M 21 , ss
rpp = rss =
M 11 , pp M 11 , ss
 Calculation of Reflection Coefficients
Anisotropic materials:

2 2X2 Abeles matrices become one 4X4 Berreman

matrix (1972)

Reason: s- and p- polarization states are no longer

eigenstates of the reflection.

Inhomogeneous layers:

If a layer has a depth-dependent refractive index, there

are two options:

1) Build up many very thin layers

2) Use interpolation approximations

 Models for dielectric functions
Tabulated Data Sets:

1) Usually good for substrates

2) Not good for thin films

3) Even for substrates: problems

a) surface roughness

b) surface reconstruction

c) surface oxides

4) Most tabulated data sets do not include error

limits

Measured optical functions of silicon depend on the face!

[near 4.25 eV = 292 nm, 2(100)<2(111)]

 Models for dielectric functions
Lorentz Oscillator Model (1895):

A j 2
( ) = ( ) 2 = 1 +
j 2 o2 , j + i j

Bj
( E ) = ( E ) 2 = 1 +
j E o2, j E 2 + i j E

Sellmeier Equation:

A j 2
= n2 = 1+
j 2 2o , j For Insulators

Cauchy (1830):

Bj
n = B0 + j 2 j

Drude (1890):

Bj 1
(E) = 1
j E E ij For Metals
 Models for dielectric functions
Amorphous Materials (Tauc-Lorentz):

A( E E g ) 2 ( E E g )
2 ( E ) = 2n ( E ) k ( E ) =
( E 2 Eo2 ) 2 + 2 E

2 ( )
1 ( E ) = 1 ( ) + P 2 2 2 d
Rg E

Five parameters:

Eg Band gap

A Proportional to the matrix element

E0 Central transition energy

Broadening parameter

1(
) Normally = 1
 Forouhi and Bloomer

(Forouhi and Bloomer Phys Rev. B 34, 7018 (1986).)

Extinction coefficient:
A( E E g ) 2
k FB ( E ) =
E 2 BE + C

Refractive index: (a Hilbert transform)

1 k ( ) k ( )
n FB ( E ) = n ( ) + P d
E

Problems:

kFB(E)>0 for E<Eg. Unphysical.

kFB(E) constant as E
. Experiment states that
.
k(E) 0 as 1/E3 as E
Time-reversal symmetry required [k(-E) = -k(E)].
Hilbert Transform Kramers Kronig [k(
) = 0].
 Models for dielectric functions
Models for Crystalline materials:

Critical points, excitons, etc. in the optical spectra

make this a very difficult problem!

Collections of Lorentz oscillators:

i j
Aj e
(E) = o +
j E E j + i j

Can end up with MANY terms

 Models for dielectric functions
Effective Medium theories:

< > h j h
=fj
< > +h j j + h

Choice of host material:

1) Lorentz-Lorentz: h = 1

2) Maxwell-Garnett: h = 1

3) Bruggeman h = <>

Depolarization factor ~2.

 Fitting Models to Data
Figure of Merit:

Experimental quantities: exp(

)

Calculated quantities: , z)
calc(,

z: Vector of parameters to be fit (1 to m)

film thicknesses,

constituent fractions,

parameters of optical function models, etc.

Minimize

N [ ( ) ( , z )]2
1
2 =
N m 1 j =1
exp j calc

( j ) 2
j

(
) = random and systematic error
 Fitting Models to Data
Calculation Procedure:

1) Assume a model.

a. Number of layers

b. Layer type (isotropic, anisotropic, graded)

2) Determine or parameterize the optical functions of

each layer

3) Select reasonable starting parameters.

4) Fit the data, using a suitable algorithm and Figure

of Merit

5) Determine correlated errors in z and cross

correlation coefficients

If the Figure of Merit indicates a bad fit (e.g. 2>>1), go

back to 1).
 Fitting Models to Data

fitting of parameters is not the end-all of parameter

estimation. To be genuinely useful, a fitting procedure
should provide (i) parameters, (ii) error estimates on
the parameters, and (iii) a statistical measure of
goodness-of-fit. When the third item suggests that the
model is an unlikely match to the data, then items (i)
and (ii) are probably worthless. Unfortunately, many
practitioners of parameter estimation never proceed
beyond item (i). They deem a fit acceptable if a graph
of data and model looks good. This approach is
known as chi-by-eye. Luckily, its practitioners get what
they deserve.

Press, Teukolsky, Vettering, and Flannery, Numerical

Recipes (2nd ed., Cambridge, 1992), Ch. 15, pg. 650.
 Fitting Models to Data
Levenberg-Marquardt algorithm:

Curvature matrix:

1 2 2 N
1 calc ( j , z ) calc ( j , z )
kl =
2 z k z l
=
j =1 ( j )
2
z k z l

Inverse of curvature matrix: = -1

Error in parameter zj = jj.

Cross correlation coefficients: proportional to the

elements of .
 Fitting Models to Data
Meaning of fitted parameters and errors:

Assume air/SiO2/Si structure

Parameterize SiO2 using Sellmeier dispersion

o=93 nm):
(

A j 2
= n2 = 1+
j 2 2o , j

Two fit parameters: df, Af

Af





d

df
 Fitting Models to Data
An Example: a-SixNy:H on silicon:

3
SiN / c-Si
2

0
Re(pp)

-1
2
-2 T-L =0.96
2
-3 Lorentz =3.84
2
-4 F&B =47.8

1
Im(pp)

-1

-2
1 2 3 4 5
Energy (eV)
 Fitting Models to Data
An Example: a-SixNy:H on silicon:

Model: 1) air

2) surface roughness

Bruggeman EMA (50% air, 50% a-SiN)

3) a-SiN (3 models)

Lorentz

Forouhi and Bloomer

Tauc-Lorentz

4) interface

Bruggeman EMA (50% Si, 50% a-SiN)

5) silicon

Fitting parameters: d2, d3, d4, A, Eo, , (

) and

Eg (F&B and T-L)

 Fitting Models to Data
An Example: a-SixNy:H on silicon:

Lorentz F&B T-L

Roughness thick (nm) 2.10.3 4.90.7 1.90.3
Film thickness (nm) 197.80.7 195.61.1 198.20.4
Interface thick (nm) 0.60.3 -0.60.6 -0.10.2
A 201.94.6 4.561.9 78.412.7
o (eV) 9.260.05 74.430.5 8.930.47
(eV) 0.010.02 0.7423.5 1.820.81
Eg (eV) ---- 2.850.48 4.350.09
() 1.000.02 0.930.51 1.380.26
2 3.64 47.5 0.92
Roughness thick (nm) 2.40.3 4.70.6 1.80.2
Film thickness (nm) 198.60.4 195.30.9 198.10.3
A 202.51.4 5.030.35 97.73.2
o (eV) 9.260.02 70.72.2 9.610.0.03
(eV) 0.010.01 40.110.8 3.070.33
Eg (eV) ---- 2.970.27 4.440.04
2 3.75 48.0 0.96
 Fitting Models to Data
An Example: a-SixNy:H on silicon:

0.2
exp-calc

0.1
Re(pp)

0.0

T-L
-0.1
Lorentz
F&B
-0.2

0.2

0.1
Im(pp)

0.0

-0.1

-0.2
1 2 3 4 5
Energy (eV)
 Optical Functions from Ellipsometry
Optical Functions from Parameterization:

2.1 T-L
refractive index (n)

Lorentz
F&B
2.0

1.9

1.8

0.015
extinction coefficient (k)

0.010

0.005

0.000

1 2 3 4 5
Energy (eV)

Error limits:

Use the submatrix s from the associated fitted

parameters.
 Optical Functions from Ellipsometry
Newton-Raphson algorithm:

Solve:

Re( , , nf, kf, )) - Re(

calc( exp(
)) = 0

calc(
Im( , , nf, kf, )) - Im(
exp(
)) = 0

Jacobian:

re re

J = n k
im im
n k

nnew = nold + n; where n = -J-1

Propagate errors!
 Optical Functions from Ellipsometry
Optical functions of semiconductors:

Dielectric function from air/substrate system:

1 2
= 1 + i 2 = sin 2 ( ){1 + [ ] tan 2 ( )}
1+

Only valid if there is no overlayer (almost never true)

If there is a thin film, Drude showed:

4 d n f 1
2

= o (ns , k s ) + K
n 2f
 Optical Functions from Ellipsometry
Optical functions of semiconductors:

Pseudo-dielectric functions of silicon with 0, 0.8, and 2.0

nm SiO2 overlayers.

<n> 0.0
6
0.8
4 2.0

0
6
<k>
4

0
6
10
< > (1/cm )
5
10

4
10

3
10
1 2 3 4 5
Energy (eV)
 Optical Functions from Ellipsometry
Optical functions of semiconductors:

Error limits of the dielectric function of silicon:

0.05

n
k
0.04
Error in n and k

0.03

0.02

0.01

0.00
1 2 3 4 5
Energy (eV)
 Optical Functions from Ellipsometry
Optical functions of thin films:

Sm all grain poly silicon

1
Re()

-1

0.5
Im()

0.0

-0.5

1 2 3 4 5
Energy (eV)
 Optical Functions from Ellipsometry
Optical functions of thin films:

Method of analysis:

A. Restrict analysis region to interference oscillations.

Parameterize the optical functions of the film.

1 air

2 surface roughness (BEMA)

3 T-L model for film

4 Lorentz model for a-SiO2

5 c-Si

B. Fit data to determine thicknesses and Lorentz

model parameters of a-SiO2.

C. Calculate optical functions of thin film using

Newton-Raphson.
 Optical Functions from Ellipsometry
Optical functions of thin films:

n
6

k
4

6
10
(1/cm)
5 c-Si
10
lg p-Si
4 sg p-Si
10
a-Si
3
10
1 2 3 4 5
Energy (eV)
 Parting Thoughts

SE is a powerful technique, but modeling is critical.

Modeling should use an error-based figure of merit

Does the model fit the data?

Calculate correlated errors and cross-correlation

coefficients.

When used properly, SE gives very accurate thicknesses

and values of the complex dielectric function.