ule glass with improved thermal properties for euvl masks ...euvlsymposium.lbl.gov/pdf/2012/pres/k....

ULE® Glass with Improved Thermal Properties for

EUVL Masks and Projection Optics Substrates

Ken Hrdina and Carlos Duran

Corning Incorporated

2012 International Symposium on Extreme Ultraviolet Lithography

Brussels, Belgium, Sept. 30 – Oct. 4, 2012

2 Hrdina/Duran / ULE® Glass for EUV Lithography © Corning Incorporated - 2012

Introduction – Ultra Low Expansion ULE® Glass

• Corning has made ULE® Glass for 4 decades: – A high-purity, single-phase SiO2 glass doped with TiO2

– Used as a substrate for high performance mirrors:

• EUV Lithography driving material improvements: – “Polishability”

– Absolute CTE better than ±5 ppb/K

– Detailed metrology and uniformity requirements

– Push for lower thermal expansion

Hubble Telescope

2.4 m Primary Mirror

(1978)

Subaru Telescope

8.2 m Primary Mirror

(1994)

3 Hrdina/Duran / ULE® Glass for EUV Lithography © Corning Incorporated - 2012

Introduction – ULE® Glass (continued)

• Most outstanding feature of ULE® Glass is its extremely low coefficient of

thermal expansion:

– Mean CTE5-35C = 0±30 ppb/K (more than 100x better than PYREX®)

• Single-Phase Glass enables other important properties:

– Long Term Dimensional stability

– No hysteresis when cycled up to ~300 °C • Room-Temperature shape and CTE are unchanged

• No specific thermal cycle or cooling rate is needed

• No temporal drift in properties is triggered

• Material is not affected by standard optical manufacturing and coating

techniques

– Mechanical strength • No delayed elastic effects have been observed (J.W. Pepi & D. Golini,

Adv. Optics 20, 3087 (1991)

4 Hrdina/Duran / ULE® Glass for EUV Lithography © Corning Incorporated - 2012

ULE® Glass for EUVL – Why?

• High throughput in EUVL requires a bright light source (100’s of Watts).

• Mirrors are expected to heat up:

– Single mirror reflectivity is only ~70% at 13.5 nm.

– Vacuum makes heat removal difficult.

– In a 6 mirror + mask system, >90% of the light is absorbed before it

gets to the wafer.

– Ultra low expansion material in mask and mirrors is a must in order to

preserve sub-nm wavefront distortion.

• ULE® Glass can be polished to sub nm roughness.

• Temperature of CTE=0 (crossover temperature Tzc) can be tuned.

• Metrology capable of non-destructively certifying material compliance

within a narrow range.

5 Hrdina/Duran / ULE® Glass for EUV Lithography © Corning Incorporated - 2012

Thermal Expansion in ULE® Glass

• CTE in ULE® Glass is controlled by composition: – Temperature dependence of expansivity curve is known.

– Curve shifts in a predictable manner with changes in TiO2 concentration.

– Manufacturing adjustments in a narrow range around room temperature are

understood, and compatible with production process.

– Non-destructive metrology allows us to define the actual CTE of all glass we

make (ultrasound velocity technique).

-80

-60

-40

-20

0

20

40

60

80

-10 0 10 20 30 40 50

Temperature in [°C]

Ex

pa

ns

ivit

y i

n [

pp

b/K

]

Less TiO2

Nominal ULE

More TiO2δ(ΔL / L)

δT

Th

erm

al e

xp

an

sio

n

ΔL / L

[p

pm

]

Temperature [°C]

Tzc = 20 C

6 Hrdina/Duran / ULE® Glass for EUV Lithography © Corning Incorporated - 2012

Minimizing Distortion in EUVL Optics and Masks

CASE I

Standard Grade

ULE®Glass

7 Hrdina/Duran / ULE® Glass for EUV Lithography © Corning Incorporated - 2012

Minimizing Distortion in EUVL Optics and Masks

Case II

EUVL Grade

ULE®Glass

With TZC set at ~20⁰C.

• Tune Tzc to expected temperature range in the optic:

• Done by adjusting TiO2 content during glass forming.

• It is difficult to control when Tzc is tightly specified.

8 Hrdina/Duran / ULE® Glass for EUV Lithography © Corning Incorporated - 2012

Minimizing Distortion in EUVL Optics and Masks

Case III

EUVL Grade

ULE®Glass.

With TZC set at

~midpoint.

9 Hrdina/Duran / ULE® Glass for EUV Lithography © Corning Incorporated - 2012

Minimizing Distortion in EUVL Optics and Masks

Case IV

EUVL Grade

ULE®Glass.

With TZC set at

~midpoint.

And

Reduced

expansivity

slope. •Reduce slope of expansivity curve:

• In addition to tuning Tzc

• Gain becomes more important as the optic temperature range grows.

10 Hrdina/Duran / ULE® Glass for EUV Lithography © Corning Incorporated - 2012

Expansivity Improvement Definition

TEMPERATURE (°C)

EX

PA

NS

IVIT

Y (

ppb/K

) Current

behavior

Ideal behavior

(“100%” Improved)

Goal of this project was to optimize the expansivity through thermal

treatment only, without introducing changes to a well established ULE®

Glass production process.

11 Hrdina/Duran / ULE® Glass for EUV Lithography © Corning Incorporated - 2012

Expansivity Improvement –

Required Metrology

• Characterization of glass structural state after annealing:

• We use Fictive Temperature (Tf) as parameter to describe the

structural state of the glass.

• Measure Tf through FTIR measurements (method by A. Agarwal,

K.M. Davis and M. Tomozawa, J. Non-Crystalline Solids 185 (1995),

p191-198). (J.E. Shelby, Phys. Chem. Glasses., 2005, 46 (5), 494-

499).

• Specific implementation and calibration for ULE® Glass submitted to

J. Non-Crystalline Solids – Tingley et al. (2012).

• Measurement of effect on expansivity:

• Sandwich Seal measurements

• Other production-compatible techniques

• Modeling of Tf

• Calibration of Ultrasound Velocity technique for improved material.

12 Hrdina/Duran / ULE® Glass for EUV Lithography © Corning Incorporated - 2012

Glass Tf Characterization

FTIR Measurements and Empirical Modeling

• Quenched samples after isothermal

holds at various temperatures to

calibrate Tf vs. FTIR Peak Position.

• Empirical model to calculate Tf after

anneal and obtain correlation

between model predictions and

measurements.

Model Predictions vs. Measurements

y = 1.000x

R2 = 0.976

860

880

900

920

940

960

980

860 880 900 920 940 960 980

SIMULATED TfM

EA

SU

RE

D T

f

13 Hrdina/Duran / ULE® Glass for EUV Lithography © Corning Incorporated - 2012

Measuring Changes in Expansivity -Sandwich Seal Differential Expansion Measurement

Incoming Light

Center – reference glass

Top & Bottom – unknown glass

Outgoing Light (Elliptically Polarized)

Seal with stress (due to difference in CTE)

(Linearly Polarized)

• Use polarized light to measure stress, and evaluate thermal expansion.

• Technique by H. Hagy and collaborators, J. Opt. Soc. Am. A3, P83 (1986), and

references therein.

• New setup incorporates computer control and data analysis: • Can detect small differences in thermal expansion.

• Requires smaller samples than other techniques.

• Temperature range from ~200 to 425 K (-70 to 150 °C).

14 Hrdina/Duran / ULE® Glass for EUV Lithography © Corning Incorporated - 2012

Sandwich Seal Data Analysis (Conceptual) - Seals made with standard production ULE® Glass -

-600

-400

-200

0

200

400

600

0 10 20 30 40

TEPERATURE (C)

EX

P D

IFF

(p

pb

)

-60

-40

-20

0

20

40

0 10 20 30 40

TEMPERATURE (C)

EX

PA

NS

IVIT

Y (

pp

b/C

)

Seal Signals Material Expansivity

• Expansivity curves in production ULE® Glass have a constant shape vs. T

• Only effect of changes in TiO2 concentration is to shift the curves

vertically.

• Stress measured in sandwich seals made with this glass will be linear: • Sign of stress indicates whether “bread” expands more or less than “meat”.

• Slope is a direct measurement of difference in CTE between bread and meat.

15 Hrdina/Duran / ULE® Glass for EUV Lithography © Corning Incorporated - 2012

Sandwich Seal – Correlation to Ultrasound

(Actual Measurements)

• Set of six sandwich seal samples were made using ULE® Glass with different amounts of TiO2 doping.

• System has excellent sensitivity to small expansion differences.

• Expansion Difference vs. Temperature plots are straight lines:

– Explicitly shows that Expansivity is independent of CTE offset.

• Uncertainty in CTE difference from Sandwich Seal is ~±2-3ppb/K – Excellent tool for evaluation of new materials.

-2500

-2000

-1500

-1000

-500

0

500

1000

1500

2000

2500

200 250 300 350 400

TEMPERATURE (K)

Exp

an

sio

n D

iff

(pp

b)

Seal#6

Seal#5

Seal#4

Seal#3

Seal#2

Seal#1

R2 = 0.98

-20

-10

0

10

20

30

-20 -10 0 10 20 30

Ultrasound CTE difference (ppb/K)

Seal C

TE

diff

ere

nce

(ppb/K

)

16 Hrdina/Duran / ULE® Glass for EUV Lithography © Corning Incorporated - 2012

Sandwich Seal Data Analysis (Conceptual)

- Seals made with non-Standard ULE® Glass -

-50

-25

0

25

50

75

0 10 20 30 40

TEMPERATURE (C)

EX

PA

NS

IVIT

Y (

pp

b/C

)

-600

-400

-200

0

200

400

600

0 10 20 30 40

TEMPERATURE (C)

EX

P D

IFF

(p

pb

)

Seal Signals Material Expansivity

• Seals are prepared using Standard ULE® Glass for the “meat”, and non-

standard for the “bread”

• If expansivity curves are not parallel, measured stress is no longer linear

with temperature.

• Fit measured seal stress adding a quadratic term:

• Sign and magnitude of quadratic term indicates whether the bread

expands slower than or faster than standard ULE® Glass.

17 Hrdina/Duran / ULE® Glass for EUV Lithography © Corning Incorporated - 2012

Sandwich Seal Measurement Results

(Actual Measurements)

-2500

-2000

-1500

-1000

-500

0

500

1000

1500

2000

2500

200 250 300 350 400

TEMPERATURE (K)

Exp

an

sio

n D

iff

(pp

b)

Seal#6

Seal#5

Seal#4

Seal#3

Seal#2

Seal#1

S1S1

S1S2

S1S5

Seals made with “bread” slices of non-Std Anneal ULE®

• Curved lines indicate expansivity slope change.

• Measurement of curvature allows quantification of

improvements in expansivity (next slide).

18 Hrdina/Duran / ULE® Glass for EUV Lithography © Corning Incorporated - 2012

Change in Expansivity Slope

(Quadratic Term in Sandwich Seal Fits)

• Samples prepared from different Boules of ULE® Glass.

• Consistent behavior seen for all seals measured.

• US Patent Applications 2011/0207592 and 2011/0207593

19 Hrdina/Duran / ULE® Glass for EUV Lithography © Corning Incorporated - 2012

Thermal Expansion in Improved ULE® Glass

Improvement in

Thermal Expansion

• Improvement shown in this figure is ~ 24%

• US Patent Applications 2011/0207592 and 2011/0207593

20 Hrdina/Duran / ULE® Glass for EUV Lithography © Corning Incorporated - 2012

Summary

• Until now, Tzc of ULE® Glass has been determined during the glass

forming process

• Using controlled annealing, we can simultaneously improve the

expansivity and fine tune the Tzc of ULE® Glass (US Patent Application

2011/0048075) – Process can be tuned to meet specific customer requirements.

– We have demonstrated improvements in expansivity greater than 30%.

– Tzc tuning will allow us to supply glass within narrower specification ranges

• Corning ULE® Glass is an enabler for EUV Lithography today – We are committed to continue to improve material properties and

manufacturing processes towards future needs

21 Hrdina/Duran / ULE® Glass for EUV Lithography © Corning Incorporated - 2012

ACKNOWLEDGMENTS

• D. Sears

• J. Tingley

• T. Carapella

• B. Tuttle

• J. Maxon

• B. Whispell

• D. Gauthier