Screening of oxidation resistant capping layers for EUV multilayerseuvlsymposium.lbl.gov/pdf/2004/presentations/day1/Co07_Sasa_Bajt.pdf · S. Bajt et al., Lawrence Livermore National

S. Bajt et al., Lawrence Livermore National Laboratory 2004 EUVL Symposium

Screening of oxidation resistantScreening of oxidation resistantcapping layers for EUV capping layers for EUV multilayersmultilayers

Saša BajtErik J. Nelson, Zurong Dai, Giles A. Graham, Jennifer Alameda,

Sherry Baker, Nhan Nguyen, Cheryl Evans, Art J. Nelson,John S. Taylor,

Lawrence Livermore National Laboratory

Miles Clift, Dean BuchenauerSandia National Laboratories

Andy Aquila, Eric M. GulliksonLawrence Berkeley National Laboratory

and

N. V. Ginger Edwards, Stefan Wurm and Obert WoodSEMATECH

This work was performed under the auspices of the U.S. Department of Energy by University of California Lawrence Livermore NationalLaboratory under contract No. W-7405-Eng-48. This project is supported by SEMATECH under Project LITH 160.


Experimental: ML testing methodology & benchmarking

- In Progress- Planned- Not Funded

- Complete

ExperimentalScaling DataScaling

Behavior

ML1 Fabricated &

Pre-screened

EuropeLifetime Tests

JapanLifetime Tests

SandiaE-beam

Exposures

RutgersModeling of Ru Oxidation

Sample Characterization

Postmortem

ComparisonSimulation vs. Experiment

Simulated Oxidation Data

Material Properties

FundamentalUnderstanding

ParametricDependencies

TestingMethodology

MLBenchmarking

Data

ML2 Fabricated &

Pre-screened

In-situ Optical Metrology

Modeling: Fundamental understanding and parameter scaling

NIST&SandiaLifetime Testing

SEMATECH LITH160 – Projection Optics LifetimeProject Plan

NISTEUV Exposures

Ru capped MLs

Other cap layers

Degradation Mechanisms


Screening tests overview

Material 2 Material 3 … Material NMaterial 1

Screening tests

12.5 13.0 13.5 14.00

1 0

2 0

3 0

4 0

5 0

6 0

7 0

12.90 12.95 13.0067.5

68.0

68.5

69.0

Ref

lect

ance

(%)

Wavelength (nm)

PTB CXRO

10-8

10-6

10-4

10-2

100

0 2 4 6 8 10

As depositedAnnealed at 450C

Rel

ativ

e In

tens

ity

2Θ (deg)

400

600

800

1000

1200

1400

1600

525 530 535 540

background57 hrs96 hrs192 hrs

Inte

nsity

Binding Energy (eV)

OH

oxide

O 1s

0 10 20 30 40 500

20

40

60

80

100

Depth in Å

Si

O

Si oxide

MoC

Reflectometry

AFM

mea

sure

men

tsTh

erm

al a

nnea

ling

E-

be

am

e

xp

os

ure

s

Depth profiling

Surface chemistry

Report to SEMATECH


Multilayer Selection Criteria

Functional requirement: Capping layer requirements:

EUV reflectivity, optical constants

Noble metals and alloys(Ru, Pd, PdAu)

Ceramics(SiC, MoSi2, YSZ)

• Long term stability EUV multilayers• Impervious to diffusion of oxygen• Limited thickness• Complete coverage• Chemically inert to the material underneath• Thermally stable

With the exception of Ru all these materials were only screened, not optimized for EUVL application.


Multilayer 1 Sample Set Details

Preparation 1 (power change)Preparation 4, 5 and 6 (gas mixture variation)Preparation 7 (material variation)

58

60

62

64

66

68

Prep 1 Prep 4 Prep 5 Prep 6 Prep 7

Sample Name

Ref

lect

ivity

(%)

Pre-Exposure

Post-Exposure

• Candidate for accelerated life-testingprotocol development needed

• ML1 was a large set of candidatesamples

• Deposition parameters stronglyinfluenced EUV reflectivity & lifetimeresponse

• Sample with best combination ofreflectivity, thermal stability, and e-beamlifetime chosen

ML1 EUV Reflectivity

*

*Exposure = electron beam exposure; 1 KeV;5µA/ mm2; 5 x 10-7 Torr water; Time = 40 hours


5 nm5 nmPrep 5

5 nm5 nmPrep 7

5 nm5 nm Prep 4 5 nm5 nm Prep 1

-~0.23 nm / Ru/(1010)~0.23 nm / Ru/(1010)-

AA BB

CC DD

Rucap

Mo

Si

Ru cap

Mo

Si

Ru cap

Mo

Si

Ru cap

Mo

Si

silicide

silicide

transitionlayer

transitionlayer

transitionlayer

silicide

silicide

transitionlayer

High lifetime of ML1 (Prep 1) is associated with adense, crystalline capping layer

More results on the lifetime performance inPO Contamination Workshop


EUV reflectivity is one of the selection criteria forcapping layer candidates

RuRu benchmark benchmark


Annealed Pd-capped MLs show high reflectivityloss with notable period change

Pd- and PdAu-capped MLs show significantlylarger period change than other multilayersalthough all of them consists of the sameMo/Si stack with 50 bilayers.

Multilayers exposed to 200ºC for 30 minutes


Thermal annealing considerably increasedsurface roughness of Pd- and PdAu-capped MLs

0.26

0.20

0.17

0.57

0.70

After 200ºC for 30 minHSFR (nm)

0.21MoSi2

0.22YSZ

0.17SiC

0.39PdAu

0.22Pd

As-depositedHSFR (nm)Material

No change

Large change


What is the cause of the reflectance drop inthese materials?

Large variation in reflectance drop fordifferent capping layer materialssuggests different degradationmechanisms


Surface roughness in e-beam exposed areas of Pd-and PdAu-capped MLs increased dramatically

These bright spots are ~14 nm talland their chemical compositioncould not be determined

Pd

PdAu


Exposed areas in Pd and PdAu-capped MLs showincrease in oxygen peak

4500

4000

3500

3000

2500

2000

1500

Cou

nts/

s

545 540 535 530

Binding Energy (eV)

unexposed exposed

Oxygen 1s / Palladium 3p3/2

180

160

140

120

100

80

60C

ount

s/s

35 30 25 20

Binding Energy (eV)

unexposed exposed

Oxygen 2s

7000

6000

5000

4000

3000

2000

1000

Cou

nts/

s

350 345 340 335

Binding Energy (eV)

unexposed exposed

Palladium 3d

350

300

250

Cou

nts/

s

240 235 230 225

Binding Energy (eV)

unexposed exposed

Molybdenum 3d

1000

800

600

400

200

Cou

nts/

s

115 110 105 100

Binding Energy (eV)

unexposed (survey) exposed (high res.)

Silicon 2p

1000

800

600

400

Cou

nts/

s

295 290 285 280

Binding Energy (eV)

unexposed exposed

Carbon 1s

Pd Capping Layer # 1

3500

3000

2500

2000

1500

Cou

nts/

s

545 540 535 530

Binding Energy (eV)

unexposed exposed

Oxygen 1s /Palladium 3p3/2

600

550

500

450

400

Cou

nts/

s

245 240 235 230 225

Binding Energy (eV)

unexposed exposed

Molybdenum 3d

5000

4000

3000

2000

1000

Cou

nts/

s

345 340 335 330

Binding Energy (eV)

unexposed exposed

Palladium 3d

7000

6000

5000

4000

3000

2000

1000

Cou

nts/

s

100 95 90 85

Binding Energy (eV)

unexposed exposed

Gold 4f

800

700

600

500

400

Cou

nts/

s

115 110 105 100

Binding Energy (eV)

unexposed exposed

Silicon 2p

1100

1000

900

800

700

600

500

Cou

nts/

s

300 295 290 285 280

Binding Energy (eV)

unexposed exposed

Carbon 1s

Pd0.5Au0.5 Capping Layer # 4 - LLNL exposure

Pd-capped multilayer PdAu-capped multilayer

O 1s/Pd 3 p3/2

O 2s

Pd 3d Mo 3d

Si 2p C 1sSi 2p C 1s

Au 4fPd 3d

O 1s/Pd 3 p3/2

Mo 3d


Depth Auger profiles reveal diffusion barrierbreakdown in the e-beam exposed areas

Boron carbideBoron carbidediffusion barrierdiffusion barrier

Where isWhere iscarbon?carbon?

Pd peak increase belowPd peak increase belowdiffusion barrierdiffusion barrier

Pd capping layer - as-deposited

Pd capping layer - exposedNo boron -No boron -

Boron possiblyBoron possiblyreacted withreacted withhydrogen tohydrogen to

form gaseousform gaseousspeciesspecies

Pd oxidized and diffusedPd oxidized and diffusedfurther into the MLfurther into the ML

PdAu capping layer – as-deposited

PdAu capping layer – exposed

Presence of AuPresence of Audeeper in the MLdeeper in the ML

Partly oxidizedPartly oxidizedPd even in as-Pd even in as-deposited statedeposited state

Both Pd and AuBoth Pd and Audiffused into the MLdiffused into the ML

Highly oxidizedHighly oxidizedsurface surface –– Some Somesilicon formedsilicon formedan unidentifiedan unidentifiedalloyalloy


25 nm25 nm

25 nm25 nm

25 nm25 nm

M3-031211BA1/08-10-2004-CM300

Cross section TEM image clearly shows coverageproblems on PdAu-capped multilayer


H2O + e- H2O + e-

Oxidation mechanism in Pd-capped multilayers• Pd island growth

• Pd oxidation

• Expansion of SiO2 into spaces between islands

• Oxidation of Si and Mo layers underneath

SiPdOH2O

Similar mechanism expected in PdAu-capped multilayers


H2O + e- CO

Oxidation mechanism in SiC-capped multilayers

• SiC converts to SiO 2 + C + CO(g)• O diffuses into SiC, CO gas escapes through SiO 2, C left at interface• M. Di Ventra and S. T. Pantelides, Phys. Rev. Lett. 83 , 1624 (1999).

O diffusion into SiCO diffusion into SiC

O accumulationO accumulation

C C –– O O3 3 complexcomplex

CO ejection out of SiOCO ejection out of SiO22

SiCMoOCOH2O


Oxidation mechanism in MoSi2-capped multilayers

• Oxidation accelerated by non-stoichiometry, defects• Protective SiO 2 formation on smooth stoichiometric surface, no MoO3

• MoO 3 and SiO 2 formation starts at defects (pores, cracks)• Volume increase at defects -> pesting

H2O + e-

SiMoOH2O


H2O + e-

Oxidation mechanism in YSZ-capped multilayers

• Yttria-stabilized Zirconia (YSZ) unchanged

• Y stabilizes fluorite structure and introduces vacancies

– 3% Y doping makes 0.75% of O sites vacant

• Mobile vacancies -> Enhanced oxygen diffusion in YSZ

• Oxidation of Si and Mo layers underneath

ZrYSiO VacancyOH2O


Si layer partially oxidized to SiO2,Mo layer partially oxidized to MoO3

YSZ unchangedYSZ

Si layer fully oxidized to SiO2,Mo layer partially oxidized to MoO3

Partial Pd oxidation,Au & Pd diffusion into bulk

Au0.5Pd0.5

Si layer fully oxidized to SiO2,Mo layer partially oxidized

Si oxidized to SiO2,Mo removal or diffusion into bulk

MoSi2

Si layer fully oxidized to SiO2,Mo layer unchanged

SiC converts to SiO2 + C + COSiC

Si layer fully oxidized to SiO2,Mo layer partially oxidized

Partial Pd oxidation,Pd diffusion into bulk

Pd

Underlying multilayer XPS resultsCapping layer XPS resultsCappinglayer

E-beam exposed samples

XPS and depth Auger results summaryNon-destructive XPS technique was used to obtain local chemical environmentanalyzing up to 5 nm into the multilayer


SummarySummary

• Fabricated and pre-screened ML1 and ML2 samples. No cappinglayer development efforts were funded.

• Oxidation/EUV reflectivity degradation mechanisms determined forselection of novel capping layer materials for EUV multilayer mirrors.

• Ruthenium capping layer still a leading candidate for oxidationprotection. Further improvements are required, however, needfundamental understanding of Ru surface science.

• The differences in the mechanisms demonstrate that test protocolswill have materials dependence that cannot be ignored.

Documents

Screening of oxidation resistant capping layers for EUV multilayerseuvlsymposium.lbl.gov/pdf/2004/presentations/day1/Co07_Sasa_Bajt.pdf · S. Bajt et al., Lawrence Livermore National