IEUVI Optics Contamination TWG, 10/02/08 1

Adv

ance

d M

ask

Tech

nolo

gy/C

R

Ted LiangComponents Research

Intel Corporation10/02/2008

EUV Mask Contamination During Use(A presentation at the IEUVI Optics Contamination TWG)

Introduction

Analysis of a MET-contaminated mask─ Line width change by CD-SEM

─ Thickness by AFM

─ Composition and depth by AES

Summary and future direction

IEUVI Optics Contamination TWG, 10/02/08 2

Adv

ance

d M

ask

Tech

nolo

gy/C

RMask Contamination

Issue of C contamination in vacuum under EUV illumination─ Hydrocarbons everywhere; deposit readily on most surfaces under EUV and e-beam illum.─ Form of PID encountered today in 193nm lithography

Multiple effects on EUV mask: Refl loss, contrast change, CD change, phase (?)

C contamination on EUV mask relatively straightforward to understand─ EUV absorption is atomic, independent of chemical bonding structures

─ Prevention a challenge

EUV R loss due to C on M L R=7 0%

0%1%2%3%4%5%6%7%8%9%

10%

0 2 4 6 8 10

C thickness (nm)

R lo

ss

Line CD increase

IEUVI Optics Contamination TWG, 10/02/08 3

Adv

ance

d M

ask

Tech

nolo

gy/C

RMask Contamination (cont’d)

Complex characteristics of possible contamination make specification difficult─ Uniform C deposition, causing global refl loss ( Global CD change)─ Structure-dependent C deposition, causing local refl loss (local CD change)─ Material-dependent C deposition, causing pattern and side-wall change─ Other manifestations in aerial imaging

Current status of understanding – Some, but not enough data─ What we know today

Observed heavy contamination, estimated 1nm C growth per 50 J/cm2 (44 wafers!)Patterned CD was affectedContamination is non-uniform over space vs. patternLikely to be cleaned off

─ What we do not knowHow close does this observations from MET represent reality in a full-field exposure tool?What are the requirements?How to monitor in-situ? Is ex-situ sufficient?What is the best recovery method – wet or dry?

Sematech’s role in answering these questions─ Baseline, testing, infrastructure, …

IEUVI Optics Contamination TWG, 10/02/08 4

Adv

ance

d M

ask

Tech

nolo

gy/C

RDescription of METDescription of MET--contaminated Maskscontaminated Masks

Mask taken off MET@Albany ~Q1’08– Most used mask at Albany

– Estimated total dose ~1000J/cm2

– POR mask materials, jointly made by LBL and Intel for the Albany MET in 2005» LBL patterned resist and Intel etched, Delivered ~ Q1’05

» Hoya blank: 70nm LR-TaBN on 2.5nm Ru-ML blank with quartz substrate

– Contamination analysis done at Intel

– Contamination: 15-20nm C build up and line CD widening

Mask taken off MET@Berkeley ~Q2’07– Used about 2 years on Berkeley MET, a relatively clean system (SR source)

– Estimated total dose ~600J/cm2

– Saw about 20-30% refl loss

– Old mask materials, jointly made for MET by LBL and Intel» LBL patterned resist and Intel etched

» Hoya blank: TAR70 Cr absorber with 20nm SiO2 buffer and 11nm Si-cap ML/Quartz

– Contamination analysis done by AMD (Uzo Okoroanyanwu,)

– Contamination: ~6-10nm C build up and line CD widening

Adv

ance

d M

ask

Tech

nolo

gy/C

R

Top right die (reference die)

Middle right die (contaminated die)

Die size 3mm x 3mm

Distance 3mm12mm

Mask off Mask off MET@AlbanyMET@Albany

Layout– No repeated dies on this mask Contaminated die visible

Visible dark area by naked eye

IEUVI Optics Contamination TWG, 10/02/08 6

Adv

ance

d M

ask

Tech

nolo

gy/C

RAdditional Information

MET field 10 is primary exposure field

This field has seen 99% of EUV that has been exposed in RTC (resist test center)

Carbon build up is evident on MET field 10

CD’s in this field compared to CD’s in other fields are larger due to the carbon build up in the area

IEUVI Optics Contamination TWG, 10/02/08 7

Adv

ance

d M

ask

Tech

nolo

gy/C

RSEM Images Taken at Sematech

Nominal 500nm (5X) 1:1 dense lines MET Field 3Average of 5 areas in this field measured 545.1nm.

Nominal 500nm (5X) 1:1 dense lines from Field 10500nm design CD, measured 608.9nm64nm greater than 1:1 lines in other fields

IEUVI Optics Contamination TWG, 10/02/08 8

Adv

ance

d M

ask

Tech

nolo

gy/C

RSEM Images Taken at Sematech

Nominal 175nm 1:1 dense lines from Field 3

Average of 5 areas in this field measured 188.5nm

Nominal 175nm 1:1 dense lines from Field 10

175nm design CD, measured 233.1nm

45nm greater than 1:1 lines in other fields

Adv

ance

d M

ask

Tech

nolo

gy/C

RSEM Images Taken at Intel

Since there are no repeated dies on this mask, images were taken on lines with same coding and labeling (i.e., all labeled 40nm)From mask as received in April 2008‘Dark’ patterns clearly indicate carbon-containing contamination

40nm elbow on reference die 40nm elbow on contaminated die

Adv

ance

d M

ask

Tech

nolo

gy/C

RSEM Images Taken after SC1 Clean at Intel

■ Extremely high level of particulates on the as-received mask■ In order to use clean room tools for measurements, mask was cleaned with SC1

– Intended to only remove fall-on particles and maybe some ‘loose’ carbons– Did not cause significant CD change (304nm after clean vs. 296nm before clean)

Some contaminations were cleaned off; Some obviously remained (dark areas on lines)

Elbows patterns on contaminated die

Labeled as 40nm Labeled as 45nm

Adv

ance

d M

ask

Tech

nolo

gy/C

RAFM on Reference Die

Pattern height ~68nm (same as incoming nominal 70nm LR-TaBN thickness)Pattern top is flat

40nm elbow on reference die

AFM height of reference pattern

-10

0

10

20

30

40

50

60

70

80

0 0.5 1 1.5 2

Distance (um)

Hei

ght (

nm)

68nm

Absorber

Ru-ML

Adv

ance

d M

ask

Tech

nolo

gy/C

RAFM on Contaminated Die – 40nm Elbows

At least 16nm thick C build up– Pattern height ~52nm (Δt = -16nm)– Contamination on Ru-surface remained after SC1– Some contamination remained on AR-TaBN after SC1

40nm elbows on contaminated die

SEM AFM

AFM line scan

-10

0

10

20

30

40

50

60

70

0 0.5 1 1.5 2 2.5

Distance (um)

Hei

ght (

nm)

68nm

‘Rabbit ear’ on pattern top: ~10nm (from 16nm)─ Contamination on pattern edge remained after SC1─ Different characteristics? – TaBN vs. AR-TaBN, high SE

Contamination on pattern top– Most was cleaned off with SC1– Remaining patches ~16nm thick

Pattern height

-10

0

10

20

30

40

50

60

70

80

0 0.5 1 1.5 2

Distance (um)

Hei

ght (

nm)

C ~16nm

68nm68nm

16nm C on Ru-ML

Space Line

Adv

ance

d M

ask

Tech

nolo

gy/C

RAFM on Contaminated Die – 45nm Elbows

45nm elbows on contaminated die

45nm pattern height

-10

0

10

20

30

40

50

60

70

0 1 2 3 4 5

Distance (um)

Hei

ght (

nm)

Same observations─ Pattern height ~52nm (Δt = 16nm)

─ Same “Rabbit ear’ on pattern top

Adv

ance

d M

ask

Tech

nolo

gy/C

RAES Depth Profiling – to Verify C Layer

■ On Reference die■ Area 1 is Ru-ML, Area 2 is TaBN

Adv

ance

d M

ask

Tech

nolo

gy/C

RAES Depth Profiling – Reference Die (cont’d)

On Reference (non-contaminated) die, Area 1 (Ru-ML)Typical of a normal mask with Ru-ML blank

0 2 0 4 0 6 0 8 0 1 0 0 1 2 0 1 4 0 1 6 0 1 8 0 2 0 00

1 0

2 0

3 0

4 0

5 0

6 0

7 0

8 0

9 0

1 0 0C 0 8 U 4 0 0 8 1 1 _ 1 . p ro

S p u t t e r D e p th (Å )

Ato

mic

Con

cent

ratio

n (%

)

IM O 1 2 9 2 0 2 . 6 /0 3 /0 8T o p R ig h t, A re a 1

M o

S i

R u

N

O

S i

S i

O

M o

M o

O 1 .ls 1S i1 . ls 2N1 .ls 3

M o 1 .ls 4Ru 2. ls 6

Note: Ru sputtering rate not calibrated

Adv

ance

d M

ask

Tech

nolo

gy/C

R

On Reference die, Area 2 (LR_TaBN absorber)

Typical of a normal mask with ARC-TaN based absorber

AES Depth Profiling – Reference Die (cont’d)

0 20 40 60 80 100 12 0 140 160 180 2000

10

20

30

40

50

60

70

80

90

100C08U400811_2.pro

Sputter Depth (Å )

Ato

mic

Con

cent

ratio

n (%

)

IMO 129202. 6/03/08Top R ight, Area 2

O

Ta

CSi

N

Ta

C C

O

O

Ta

Ta1O1.ls1Si1.ls2N1.ls4C1.ls5

Adv

ance

d M

ask

Tech

nolo

gy/C

R

On contaminated die

AES Depth Profiling – Contaminated Die

Ru-ML

TaBNL

Adv

ance

d M

ask

Tech

nolo

gy/C

R

■ On Contaminated die, Area 1 (Ru-ML)■ Surface was completely covered by C contamination■ No significant O signal increase

AES Depth Profiling – Contaminated Die

0 50 100 150 20 0 250 300 3500

10

20

30

40

50

60

70

80

90

100C08U400831_1.pro

S put ter Depth (Å )

Ato

mic

Con

cent

ratio

n (%

)

IM O129202. 6/ 03/08Middle Right , Ar ea 1

C

NO

R u

S i

Mo

O

Mo

Mo

S i

S i S i

O1.ls 1S i1 .ls 2N1.ls 3

Ru2. ls 6Mo1.ls 7

C1.ls 1

Adv

ance

d M

ask

Tech

nolo

gy/C

R

■ On Contaminated die, Area 2 (LR_TaBN absorber)

■ Very little C at this particular site (Most likely cleaned off by SC1)

AES Depth Profiling – Contaminated Die

0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 00

1 0

2 0

3 0

4 0

5 0

6 0

7 0

8 0

9 0

1 0 0C 0 8 U 4 0 0 8 3 1 _ 2 .p ro

S p u t te r D e p th (Å )

Ato

mic

Con

cent

ratio

n (%

)

IM O 129202 . 6 / 03 /08

Midd le - R igh t D ie , Area 2

T a

N

C , O

O

C

S i NC

T a

O 1N1

T a 1C1

S i1 .ls 1

IEUVI Optics Contamination TWG, 10/02/08 20

Adv

ance

d M

ask

Tech

nolo

gy/C

RSummary and Future Directions Summary and Future Directions

EUV exposure results in (unavoidable) carbonaceous contamination– Up to at least 16nm C build up from the mask off MET@Albany– Contamination was not uniform: Ru vs. TaN, edge vs. top, cleanability by SC1– No significant oxidation was observed (a good thing - to avoid irreversible damage)

C contamination on Ru-ML is different than that on TaN absorber– Different responses to SC1 cleaning– Even though this may not be entirely unexpected, further assessment needed

Further determination of mask lifetime due to contamination is needed in a exposure tool with well-emulated and controlled conditions

– Understand implications to manufacturability and productivity

Sematech project in 2009 shall be designed to address/confirm the following, in collaboration with exposure tool suppliers and MCs

– Specification on contamination based on imaging requirements (Dose, CD, flare…)– Growth rate– Pattern and surface dependency– Monitoring – in-situ or ex-situ– Reduction and recovery Ref: Roman Caudillo: Poster and TWG presentations

IEUVI Optics Contamination TWG, 10/02/08 21

Adv

ance

d M

ask

Tech

nolo

gy/C

RAcknowledgements Acknowledgements

Sematach:

– Andrea Wüest and Co.

CXRO – LBNL

– Patrick Naulleau

Intel:

– Armando Cobarrubia

– Joseph Rodriguez

– Guojing Zhang

– Roman Caudillo

– Gilroy Vandentop

IEUVI Optics Contamination TWG, 10/02/08 22

Adv

ance

d M

ask

Tech

nolo

gy/C

R

Backup

IEUVI Optics Contamination TWG, 10/02/08 23

Adv

ance

d M

ask

Tech

nolo

gy/C

R

Similar dark regions of those exposed fields

Mask off Mask off MET@BerkeleyMET@Berkeley

IEUVI Optics Contamination TWG, 10/02/08 24

Adv

ance

d M

ask

Tech

nolo

gy/C

RSEM Shows Same Line Edge Widening

Reference Contaminated