1

Investigation of lithographic metrics

for the characterization of intrinsic

resolution limits in EUV resists

Patrick Naulleau

Lawrence Berkeley National Laboratory

University at Albany

Chris Anderson

University of California, Berkeley

Ryoung-han Kim, Bruno La Fontaine, Tom Wallow

AMD

2

Outline

• Motivation

• Intrinsic bias

• MTF

• LER Spectral Characteristics

• Corner Rounding

• Contacts

• Summary

3

Motivation

• Resist now one of biggest challenges facing EUV

• Problem lies in simultaneous achievement of

resolution, LER, and sensitivity goals

• While, LER and sensitivity are easily quantified

entities, resolution often remains subjective

• Which metrics are best for comparing and ultimately

quantifying the “intrinsic resolution” of a resist?

• Can we extract a resist blur (point-spread) function?

4

Which of these resists has the best

intrinsic resolution?35

35-- n

mnm

30

30-- n

mnm

KRSKRS MET 1KMET 1K

Severe

collapse

Supplier CSupplier C

5

Intrinsic bias as a measure of intrinsic

resolution

Iso-focal

position

Iso-focal

position

“Reaction front”

6

To find intrinsic bias we must first

find expected iso-focal CD

5050--nmnm 4545--nmnm 4040--nmnm

Bossungs based on

10% dose increments

7

KRS resist shows intrinsic bias of ~19 nm

5050--nmnm 4545--nmnm 4040--nmnm

Intrinsic bias = 19 nm

Bossungs based on 5% dose increments

Nominal dose = 19 mJ/cm^2

32.532.5--nmnm

8

Supplier A shows low intrinsic bias but poor

process latitude below 45 nm half pitch

Intrinsic bias = 4 nm

5050--nmnm 4545--nmnm 4040--nmnm

Bossungs based on 5% dose increments

Nominal dose = 11 mJ/cm^2

32.532.5--nmnm

9

Intrinsic bias not found to be well

correlated to resolution limit

>16>163535MET 1KMET 1K

191932.532.5KRSKRS

443535Supplier CSupplier C

443535Supplier ASupplier A

Intrinsic Intrinsic

Bias (nm)Bias (nm)Res.Res.

(nm)(nm)ResistResist

3535--nm lines and spacesnm lines and spaces

KRSKRS MET 1KMET 1K Supplier ASupplier A Supplier CSupplier C

10

Resist based MTF measurements provide

insight into resist and system properties

minmax

minmax

minmax

minmax

DD

DD

II

IIContrast

+

−=

+

−=

• MTF = pitch-dependent contrast

• Contrast determined from:

- Dmax, the dose at which resist lines

first begin to clear

- Dmin, the dose at which resist lines

disappear

11

Resist performance has strong impact

on measured contrast

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 20 40 60 80 100 120

Feature Size (nm)

Contr

ast

ModelingEUV 2DKRSMET 1KSupplier DSupplier ASupplier C

12

Modeling Resist Using Simple

Point-Spread-Function (PSF) Method

“Deprotection blur” function PSF

* C. Ahn, H. Kim, K. Baik, “A novel approximate model for resist

process,” Proc. SPIE 3334, (1998).

** Gregg Gallatin, “Resist Blur and Line Edge Roughness,” Proc. SPIE

5754, (2005)

• PSF resist modeling* is

fast and convenient

• Model easily generated

• Provides intuitive link to

resist resolution limit

• Few parameters makes

model less susceptible

to extrapolation errors

• Resist process well

approximated by

deprotection function**

13

Extracting the deprotection blur from MTF data

0

0.2

0.4

0.6

0.8

1

20 40 60 80 100Feature Size (nm)

Contr

ast

Aerial ImageMET 1KBlurred

45455050EUV 2DEUV 2D

24243535Supplier ASupplier A

22223535Supplier CSupplier C

21213535MET 1KMET 1K

141432.532.5KRSKRS

Blur Blur

(nm)(nm)Res.Res.

(nm(nm))ResistResist

35-nm dense

14

LER roll-off as a resolution metric

26263535Supplier ESupplier E

25253030Supplier FSupplier F

28283535Supplier CSupplier C

24245050EUV 2DEUV 2D

20203535Supplier ASupplier A

22223535MET 1KMET 1K

181832.532.5KRSKRS

LLcc

(nm)(nm)Res.Res.

(nm(nm))ResistResist

LER roll-off (correlation length) is

NOT a good indicator of resolution

15

Roughness exponent as a resolution metric

18180.590.5932.532.5KRSKRS

26260.650.653535Supplier ESupplier E

25250.630.633030Supplier FSupplier F

0.420.42

0.630.63

Roughness exp.Roughness exp.

24245050EUV 2DEUV 2D

22223535MET 1KMET 1K

LLcc (nm)(nm)Res.(nmRes.(nm))ResistResist

Roughness exponent is NOT a

good indicator of resolution

16

Comparing MTF and Correlation Length

Metrics for Process Studies

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

20 25 30 35 40 45 50

PEB = 90 oCPEB = 100 oCPEB 110 oCAerial Image MTF

CD (nm)

Co

ntr

ast

3232

2929

2222

Blur (nm)Blur (nm)

22.622.6

23.823.8

20.720.7

LLcc (nm)(nm)

9090°° CC

100100°° CC

110110°° CC

PEBPEB

17

Corner Rounding as a Resolution Metric

• Use fine-corner detail in large feature to determine resolution limit

18

Performance of the Corner Rounding Metric

4545

6060

5656

8787

Corner Radius Corner Radius

(nm)(nm)Blur* Blur*

(nm)(nm)

3535Supplier ESupplier E

3030Supplier FSupplier F

5050EUV 2DEUV 2D

3535MET 1KMET 1K

Res.(nmRes.(nm))ResistResist

Corner-rounding appears to be

good predictor of resolution

19

Through-Dose Contact Printing as

Resolution Metric

• 50-nm contacts with different levels of resist blur

0 nm 15 nm 30 nm

20

0

20

40

60

80

100

120

140

0.5 1 1.5 2 2.5 3

Supplier E

MET 1K

Preliminary Results with Contact Metric

Normalized Dose

CD

(nm

)

• Results consistent with corner rounding

• Need more data to compare performance relative to other metrics

• Can be viewed as inverse Fourier equivalent of MTF method

Supplie

r E

21

Corner-Rounding Analysis Showed Supplier

F to be Best Sample: Printing Results

Coded 45Coded 45--nm:90nm:90--nmnm

Actual 38Actual 38--nm:90nm:90--nmnm

LER 3.0 nm: L = 403 nmLER 3.0 nm: L = 403 nm

Coded 40Coded 40--nm:80nm:80--nmnm

Actual 33Actual 33--nm:80nm:80--nmnm

LER 3.1 nm: L = 403 nmLER 3.1 nm: L = 403 nm

Coded 35Coded 35--nm:70nm:70--nmnm

Actual 27Actual 27--nm:70nm:70--nmnm

LER 3.0 nm: L = 366 nmLER 3.0 nm: L = 366 nm

Dose to size

(50-nm 1:1)

= 19 mJ/cm2

Y-monopole

22

More Supplier F Printing Results

Coded 30Coded 30--nm:60nm:60--nmnm

Actual 24Actual 24--nm:60nm:60--nmnm

LER 4.0 nm: L = 350 nmLER 4.0 nm: L = 350 nm

Coded 30Coded 30--nm:60nm:60--nmnm

Actual 21Actual 21--nm:60nm:60--nmnm

LER 4.0 nm: L = 350 nmLER 4.0 nm: L = 350 nm

Coded 22.5Coded 22.5--nm:67.5nm:67.5--nmnm

Actual 22.7Actual 22.7--nm:67.5nm:67.5--nmnm

LER 4.0 nm: L = 512 nmLER 4.0 nm: L = 512 nm

Coded 32.5Coded 32.5--nm:65nm:65--nmnm

Actual 28Actual 28--nm:70nm:70--nmnm

LER 4.4 nm: L = 407 nmLER 4.4 nm: L = 407 nm

Y-monopole

23

Summary

•• Various potential resolution metrics have been studiedVarious potential resolution metrics have been studied

•• Intrinsic bias and LER spectral characteristics do not Intrinsic bias and LER spectral characteristics do not

show good correlation with resolution limitshow good correlation with resolution limit

•• MTF and cornerMTF and corner--rounding analysis appear good to be rounding analysis appear good to be

good metrics of intrinsic resolutiongood metrics of intrinsic resolution

•• More modeling required to extract actual PSF from More modeling required to extract actual PSF from

corner analysiscorner analysis

•• Contact analysis also looks promising, more complete Contact analysis also looks promising, more complete

comparison to other metrics still requiredcomparison to other metrics still required

•• A new resist outperforming KRS has been identifiedA new resist outperforming KRS has been identified

24

Acknowledgments

Supported by:

Paul Denham Paul Denham

Brian HoefBrian Hoef

Gideon JonesGideon Jones

JerrinJerrin Chiu Chiu

LBNLLBNL

Jim ThackerayJim Thackeray

Katherine SpearKatherine Spear

Rohm and HaasRohm and Haas

SEMATECH and the SEMATECH logo are registered servicemarks of SEMATECH, Inc.

AMD, the AMD Arrow logo and combinations thereof are trademarks of Advanced Micro Devices, Inc.