Transcript n - Physics and Astronomy - Brigham Young University

Optical Constants of Sputtered
Thoria Thin Films Useful in EUV
Optics from IR to EUV
David D. Allred
Brigham Young University
16 Aug 2006
Our Goal – EUV Applications
• Extreme Ultraviolet Optics has
many applications.
• These Include:
– EUV Lithography
– EUV Astronomy= image mission
– Soft X-ray Microscopes
• A Better Understanding of
EUV Optics & Materials for
EUV applications is needed.
21 Feb. 2007
EUV Lithography
EUV Astronomy
The Earth’s magnetosphere in the EUV
Soft X-ray Microscopes
2
Participants
• William R. Evans: senior
(honors) thesis:
Spectroscopic ellipsometry
1- 6.5 eV
• Niki F. Brimhall: senior
(honors) thesis: EUV optical
constants of Thoria
– (also Guillermo Acosta &
Jed E. Johnson. )
• Sarah C. Barton 10.2 eV
reflectance (Monarch)
• R.S. Turley: most
everything spectroscopic
>10 eV.
• Michael Clemens: AFM
21 Feb. 2007
• XPS Amy B. Grigg
• and The BYU EUV Thin
Film Optics Group, past
and present who went to
ALS : Jacque Jackson,
Elise Martin, Lis Strein,
Joseph Muhlestein
• Dr. Thomas Tiwald: JA
Woollam Co:
interpretation & extending
range of Spec. Ellips. To
IR and 9.5 eV
• Matt Linford’s Group
(Chem.)
3
Financial + Other Assistance
• BYU Department of Physics and Astronomy
shop & electronics
• BYU Office of Research and Creative Activities
• Rocky Mountain NASA Space Grant Consortium
• V. Dean and Alice J. Allred, Marathon Oil
Company
• ALS time (DOE) and help @ beamline
6.3.2: Eric Gullikson, Andy Aquila
21 Feb. 2007
4
• XUV Optics
Outline
– Applications -Production
• Review of Optics for EUV/ x-rays (E>15 eV)
• Why Actinides in EUV? Why Oxides?
– besides ML there are low-angle front surface mirrors
• Optical constants from R and T
• Measuring XUV OC with Reflectance &
Transmission on “Absolute” X-ray diodes
• Real Surfaces: Characterizing & Improving them.
• Spectroscopic Ellipsometry 1- 6.5 eV:
– Thoria has some leftover problems in solid state.
– Index and band gap.
21 Feb. 2007
5
Extreme Ultraviolet Optics—What is
our end goal?
Multilayer Mirrors
Astronomy
Lithography
21 Feb. 2007
Microscopy
6
Optics like n-IR, visible, & nUV? First you need a light.
21 Feb. 2007
7
Optics like n-IR, visible, & n-UV?
• How to manipulate light?
• Lens? Prisms? Mirrors? Diff Gratings? ML
interference coatings?
• We need to have optical constants;
• How to get in EUV?
– Kramers-Kronig equations n ()  k ()
– Variable angle of reflection measurements,
– Real samples aren’t good enough.
Roughness
21 Feb. 2007
8
Absorption and Refraction
• Optical properties characterized by index
of refraction n
• Visible
– n real (often >>1)
– n >0 (total internal reflection)
• XUV and X-Rays
– n complex; n=1-δ+iβ
– Re(n) < 1 (but not by much)
21 Feb. 2007
9
Reflectance (normal)
n1  n2
r
n1  n2
R | r |
21 Feb. 2007
2
10
Complex Index of Refraction
• Real n
 2nx 
E  sin 

  
• Complex n=1-δ+iβ
• β =k
 2 (1   ) x 
  2kx 
E  sin 
 exp 




  
21 Feb. 2007
11
Multilayer Mirrors
• Problems
– Need constructive interference
– Absorption in layers
21 Feb. 2007
12
Image Mirror
21 Feb. 2007
13
U/Si ML coating for EUV instrument
• Picture (41 eV) is from EUV imager on the
IMAGE Spacecraft. He (II) in magnetosphere
• This was student powered project 1997-98
• Designed: needed 7 degree width off normal, 7.5
layer U/Si ML with U Oxide cap- peak R 25%
• Coated &
• Tested
• Launched 2000 March 25
21 Feb. 2007
14
EUV Multilayer Optics 101
High reflectivity multilayer coatings require:
• Refractive index (n = 1-δ+iβ) contrast at the interfaces:
for most materials, these optical constants are not well
known in this region.
• Minimal absorption in the low-Z material
• Interfaces which are chemically stable with time
• Minimal interdiffusion at the interfaces
• Thermal stability during illumination
• Chemically stable vacuum interface
Even with the very best designs, multilayer mirrors have
21 Feb. 2007
only
achieved a reflectivity of around 70% in the EUV. 15
The solution? Research of new
materials with these properties
Uranium:
 Highly reflective in the region from 124-248 eV [1]
 Not chemically stable with time
Uranium Oxide:
 Highly reflective in the region from 124-248 eV [1]
 Not chemically stable with time
Thorium:
 Highly reflective in the region from 138-177 eV [2]
 Not chemically stable with time, tho better than U.
[1] RL Sandberg, DD Allred, JE Johnson, RS Turley, " A Comparison of Uranium Oxide and Nickel as
Single-layer Reflectors", Proceedings of the SPIE, Volume 5193, pp. 191-203 (2004).
[2] J. Johnson, D. Allred, R.S. Turley, W. Evans, R. Sandburg, “Thorium-based thin films as highly
reflective mirrors in the EUV”, Materials Research Society Symposium Proceedings 893, 207-213, 2006.
16
• n=1-δ+iβ
Solutions
• Find materials with big δ and small β
• Good candidates: High Density, High -Z
materials like U. But Oxidation occurs.
– Th as ThO2 has entrée.
21 Feb. 2007
17
How to Get OC from Data
• Measure reflectance and/or transmission
– Multiple wavelengths
– Multiple angles
• Fit data to a theoretical Model
– film thicknesses
– optical parameters
• But reflectance is sensitive to surfaceinhomogeneities roughness; oxidation
21 Feb. 2007
18
Transmissionk?
•
•
•
•
T = (Corrections) exp (-αd);
Corrections are due to R and can be small
At normal incidence R goes as [2 + β2]/4
If film is close to detector scattering due to
roughness etc. is less important.
• But how to get an even, thin film?
– A very thin membrane?
21 Feb. 2007
19
Measurements of reflectance and
transmittance
~20 nm reactively sputtered ThO2 on a polyimide membrane
(~100 nm, Moxtek) and a naturally oxidized silicon substrate.
21 Feb. 2007
20
Better procedures for fitting
 Take several measurements—use each measurement
to constrain those parameters to which it is most
sensitive
21 Feb. 2007
21
A major problem with our first try
 Measurements of thorium dioxide deposited on
polyimide films gave unreliable data.
Reflectances measured with different filter sets differed by as much
as 32% of total reflectance. Absolute transmission measurements
21 Feb. 2007
22
were
uncertain by as much as 19%.
Optical Constants
Even though our absolute transmission was uncertain to this
degree, the energy of the incident light was known to 0.012%, and
so even if the exact values of delta and beta are off, the edges
won’t
21 Feb. be.
2007
23
A second method that worked
 Thorium dioxide deposited on AXUV-100
silicon photodiodes (IRD).
21 Feb. 2007
24
Verification and a surprise
 In delta: a peak shift to lower energies by 3 eV
from 92.8 eV
21 Feb. 2007
25
Verification and a surprise
 In beta: absorption edge shifts to lower energies
from those of thorium by 4 eV from 105.6 eV
and 2 eV from 91.5 eV
21 Feb. 2007
26
Summary and Conclusions
 We report the optical constants of ThO2 from 50108 eV
 We have used constraining techniques to fit optical constants
including fitting film thickness using interference fringes in
highly transmissive areas of the spectrum and fitting
reflectance and transmittance data simultaneously
 In delta we observed a peak shift to lower energies from that
of thorium by 3 eV from 92.8 eV
 In beta we observed absorption edge shifts to lower energies
from those of thorium by 4 eV from 105.6 eV and 2 eV from
91.5 eV
21 Feb. 2007
27
Transmission thru a film on PI
21 Feb. 2007
28
But reflectance is a problem
21 Feb. 2007
29
The problem is waviness of
substrate. Sample on Si does fine.
21 Feb. 2007
30
The Solution: Deposit the film
on the detector
• Uspenskii, Sealy and Korde showed that
you could deposit a film sample directly
onto an AXUV100 silicon photodiode (IRD)
and determine the films transmission ( by
) from the ratio of the signals from
various coated diodes with identical
capping layers.
• JOSA 21(2) 298-305 (2004).
21 Feb. 2007
31
Our group’s 1st approach
1. Measure the reflectance of the coated
diode at the same time I am measuring
the transmission. And
2. Measure both as a function of angle. And
3. Get the film thickness from the (R and T
data) to check ellipsometry of witness.
21 Feb. 2007
32
Fitting T() to get dead layer thickness
(6-7nm) on bare AXUV diode
@=13.5nm
21 Feb. 2007
33
Focusing on the high reflectance &
transmission had a problem
21 Feb. 2007
34
Comments
1. Either T or R have n and k data, but
2. Transmission has very little n data when δ is
small (the EUV).
3. Reflection  n, k and when interference
fringes are seen, and
4. It has thickness (z) data.
What follows shows how we confirmed
thickness for air-oxidized Sc sputtercoated AXUV diodes.
21 Feb. 2007
35
Our recent group’s approach
1. Measure the reflectance of the coated
diode at the same time I am measuring
the transmission. And
2. Measure both as a function of angle. And
3. Get the film thickness from the (R)
interference fringes (@ high angles).
21 Feb. 2007
36
0
Interference in R (50<φ<70 )
 zfit=19.8 nm @ =4.7 nm
21 Feb. 2007
37
The complete set of R data
(6<θ<200) zfit =28.1 nm @ =4.7 nm
21 Feb. 2007
38
We might gone with z= 24 nm, but
21 Feb. 2007
39
We looked at another = 7.7nm;
needs z=29 nm
21 Feb. 2007
40
And the =4.7nm data is OK
21 Feb. 2007
41
Reflectance and transmittance of a ThO2-coated
diode at 15 nm fitted simultaneously to obtain n&k
• Green (blue) shows
reflectance
(transmission) as a
function of grazing
angle ()*
• Noted the interference
fringes at higher angles
in R.
* is always from grazing
incidence
21 Feb. 2007
42
R &T of a ThO2-coated diode at 12.6 nm fitted
simultaneously to obtain optical constants.
• The fits were not very
good at wavelengths
where the
transmission was
lower than 4%.
• All of these fits were
trying to make the fit
of transmission
narrower than the
data was.
21 Feb. 2007
43
“Intermediate Conclusions”
• Thin films of scandium oxide, 15-30 nm thick, were
deposited on silicon
• photodiodes by
– Sputtering Sc from a target & letting it air oxidize OR
– reactively sputtering scandium in an oxygen
environment.
• Similar thing was done with Thorium to make Thoria
• R and T Measured using synchrotron radiation at the als
(Beamline 6.3.2), at LBNL
– over wavelengths from 2.5-40 nm at variable
– angles, were taken simultaneously.
21 Feb. 2007
44
ThO2
• A number of studies by our group have
shown that thorium and thorium oxide
(ThO2) have great potential as highly
reflective coatings in the EUV.
• In certain regions, ThO2 may be the
best monolayer
reflector that has
yet been studied.
21 Feb. 2007
45
XUV Optics Production
• Sputtering or Evaporation
21 Feb. 2007
46
Biased Sputtering
• Our films were deposited by
biased RF Magnetron
Sputtering.
• ThO2 was reactively sputtered
off of a depleted thorium target
with oxygen introduced in the
chamber.
• Chamber sputtering pressures
were about 10-4 torr.
• Bias voltages were between 0
and -70 V DC.
21 Feb. 2007
47
Film Characterization
• Film composition was measured using x-ray photoelectron
spectroscopy. Th % stayed between 60% and 70% with
oxygen making up the balance of the composition. Only
traces of other elements were detected.
• X-ray diffraction was used 1) as a first measurement of film
thickness and 2) to measure crystal structure. Orientations
(111), (200), (220),
and (311) were clearly
visible, with other
orientations being
largely absent.
21 Feb. 2007
48
Spectroscopic Ellipsometry
• Optical characteristics were measured using
spectroscopic ellipsometry in the visible and
near UV.
• Ellipsometric data were taken from samples
deposited on silicon between 1.2 and 6.5 eV at
angles of every degree between 67° and 83°.
• Normal incidence
transmission data were
taken over the same range
of energies, from samples
deposited on quartz slides.
21 Feb. 2007
49
Data Fitting
• The data were modeled using the J. A. Woollam
ellipsometry software.
– n is modeled parametrically using a Sellmeier model
which fits ε1 using poles in the complex plane.
– The Sellmeier model by itself doesn’t
account for absorption.
(i.e. All poles are real.)
– k can be added in
separately, either
by fitting point by point,
or by modeling ε2 with
parameterized
oscillators.
21 Feb. 2007
50
n
Results: n
ThO2 050429 -- 28 nm -- 0 V -- (Bottom Band)
ThO2 050503 -- 28 nm -- 50 V -- (Middle Band)
ThO2 050526 -- 57 nm -- 0 V -- (Top Band)
ThO2 050527 -- 47 nm -- 0 V -- (Middle Band)
ThO2 050604 -- 24 nm -- 64 V -- (Middle Band)
ThO2 050604-2 -- 357 nm -- 0 V -- (Middle Band)
ThO2 050818 -- 578 nm -- 65 V -- (Top Band)
From Liddell (1974) - Homog. Film - d=92.4nm
From Mahmoud 2002 (fig.7)
2.7
2.6
2.5
2.4
2.3
2.2
2.1
2
1.9
1.8
1.7
1
2
3
4
5
n vs E (eV)
6
E (eV)
21 Feb. 2007
51
21 Feb. 2007
Average -- 1.28 eV
Biased
Unbiased
Thick
Thin
Average -- 2.50 eV
Biased
Unbiased
Thick
Thin
Average -- 3.00 eV
Biased
Unbiased
Thick
Thin
Average -- 4.00 eV
Biased
Unbiased
Thick
Thin
2.5
2.4
2.3
2.2
2.1
2
1.9
1.8
1.7
1.6
1.5
Average -- 5.49 eV
Biased
Unbiased
Thick
Thin
Average -- 6.00 eV
Biased
Unbiased
Thick
Thin
n
n not related to Bias Voltage or
Thickness
Average n and Standard Deviations at
Different Energies
52
Results: Absorption
ThO2 050429 -- 0 V -- 28 nm
ThO2 050503 -- 50 V -- 28 nm
ThO2 050520 -- 68 V -- 69 nm
ThO2 050527 -- 0 V -- 47 nm
ThO2 050604 -- 64 V -- 24 nm
ThO2 050604-2 -- 0 V -- 357 nm
ThO2 050818 -- 65 V -- 578 nm
3
2.5
alpha*d
• There is a
narrow
absorption
feature at
about 6.2 eV,
with full width
half max of
about 0.4 eV.
2
alpha*d vs E
1.5
1
0.5
0
3.5
21 Feb. 2007
4
4.5
5
5.5
E (eV)
6
6.5
7
53
Comparing to the Literature
• In reviewing the literature,
there seems to be a
couple of different band
gaps that people detect:
21 Feb. 2007
54
Graphic From: Rivas-Silva, et. al.
Comparing to the Literature
• Sviridova & Suikovskaya measured sample
thickness and absorption for several different
wavelengths near where thorium goes
transparent, for thorium chloride and thorium
nitrate.
10
• From this we
determine a band 8
gap of ~5.92 eV.
6
• We obtained a
value in the same 4
range.
Alpha*d
y = 16.383x - 96.609
R2 = 0.9883
y = 8.1226x - 48.174
R2 = 0.94
y = 7.31x - 43.423
R2 = 0.9155
2
y = 5.2168x - 30.944
R2 = 0.8306
NO3 300 C
400 C
500 C
700
Linear (400 C)
Linear (NO3
300 C)
Linear (500 C)
Linear (700)
0
21 Feb. 2007
5.6
5.8
6
6.2
Energy eV
6.4
6.6
6.8
55
Comparing to the Literature
• Mahmoud
reports a very
clear band gap of
about 3.84 eV.
• However, his
samples were
deposited on
glass of
unspecified
composition.
21 Feb. 2007
56
What we think might be going on...
• If the middle band were centered at
about -9.8 eV in stead of -11.8 eV, the
~6 eV band gap reported in the
majority of the thin film sources would
be explained as a jump from the
valence band to the middle band.
• Also, if the conduction band started at
about -6 eV in stead of about -7 eV,
the ~4 eV band gap reported by
Mahmoud and others could be
explained by a transition from the
middle band, which had some
electrons in it due to mild doping,
transitioning into the conduction band.
21 Feb. 2007
57
Measurements at 10.2 eV
• We used a
McPherson
Vacuum
monochromator
at BYU to
measure optical
constants of our
ThO2 thin films at
the 10.19 eV Kα
line of Hydrogen.
21 Feb. 2007
58
Measurements at 10.2 eV
ThO2 050429 -- 0 V -- 28 nm
ThO2 050503 -- 50 V -- 28 nm
ThO2 050520 -- 68 V -- 69 nm
ThO2 050527 -- 0 V -- 47 nm
ThO2 050604 -- 64 V -- 24 nm
ThO2 050604-2 -- 0 V -- 357 nm
ThO2 050818 -- 65 V -- 578 nm
3
alpha*d
2.5
2
alpha*d vs E
1.5
1
0.5
0
3.5
21 Feb. 2007
4
4.5
5
5.5
E (eV)
6
6.5
7
59
Measurements at 10.2 eV
ThO2
ThO2
ThO2
ThO2
ThO2
ThO2
ThO2
40
35
alpha*d
30
050429 -- 0 V -- 28 nm
050503 -- 50 V -- 28 nm
050520 -- 68 V -- 69 nm
050527 -- 0 V -- 47 nm
050604 -- 64 V -- 24 nm
050604-2 -- 0 V -- 357 nm
050818 -- 65 V -- 578 nm
alpha*d vs E
25
20
15
10
5
0
3.5
21 Feb. 2007
4.5
5.5
6.5
7.5
E (eV)
8.5
9.5
10.5
60
Conclusions: Vis & UV
• First of all, we have shown that DC Biased
sputtering cannot be expected to significantly
affect the optical constants of ThO2 thin films.
– This is not surprising considering the extremely
high melting point of ThO2.
• Secondly, exactly what is going on with the
band gap of ThO2 is still not really
understood.
– It appears that there are two fundamental band
gaps in ThO2, but more research is needed.
– We are in the process of making additional
measurements on ThO2 between 6 and 9 eV.
21 Feb. 2007
61
• Questions?
21 Feb. 2007
62
EUV measurements
• Our project was to see if we could get n as well as k
from samples set up to measure transmission in the
EUV.
• The films were deposited directly on Absolute EUV
silicon photodiodes. IRD
21 Feb. 2007
63
Total Reflection
• Snell’s Law
n1 sin 1  n2 sin 2
• Total Internal
Reflection
n1
n2
• Total External
Reflection
21 Feb. 2007
n1
n2
64