The Institute of Scientific and Industrial Research, Osaka University

Status and Challenge of Chemically Amplified Resists

for Extreme Ultraviolet Lithography

Takahiro Kozawa

1. Status and challenge

2. Anion-bound resist

3. Wavelength diffusion

Lithography roadmap

1

10

100

1000

g line i line KrF ArF EUV EB

Ener

gy

(eV

)

Exposure tool

Sensitivity

Resolution

LWR(LER)

Trade-off Ionization energy (~10eV)

Radiation chemistry

Two keywords in the development of resist materials and processes

1.2

ITRS2010 Update

2001 04 07 10 13 16 19 22

130 90 65 45

Year

Line width (nm)

LWR (nm)

Lithography

Solution

KrF excimer

(248 nm)

ArF excimer

(193 nm)

ArF excimer

Immersion (+DP)

32

2.2 1.6

EUV

(13.5 nm)

22

EB for mask production

1.1

16 11

0.8

Transition of basic science for material design

Photochemistry

Trade-off relationships between resolution, sensitivity, and LER

Chemically amplified resist In

tensi

ty

Typical components: Polymer, Acid generator, Quencher

0.0

0.2

0.4

0.6

0.8

1.0

0 1 2 3 4 5 SEM image of resist Position

Role: Conversion of energy modulation to binary image

Development Formation of latent image

Formation of acid image

Energy deposition

Acid diffusion, deprotection

Decomposition of acid generator

Exposure source

Photon/electron interaction with matter

Dissolution of molecules

Resist image

Thermal energy

Energy modulation of quantum beam

Si

Solubility change through chemical reaction

LER ∝ width of intermediate region ≒ Chemical gradient

1

EUV

Relationship between LER and chemical gradient

Aerial image

Acid Image Latent image

Pattern

Insoluble

Soluble

(Spatially) partially soluble

OH

CH CH2

Mixture of soluble and insoluble molecules

LER formation

OR

CH CH2

Cross section

Top-down view

After development

OH

CH CH2

OR

CH CH2

+ Absolutely stochastic

Can not be controlled

Can be controlled

dxdm

fLER LER

/

Protected unit

Resist screening

IMEC

Selete (later EIDEC)

Sematech

A large number of resist materials have been tested in these sites.

IMEC, Sematech, and Selete (EIDEC) have supported the development of

EUV lithography including resist material and processes.

Sensitivity

Resolution

LWR(LER)

Trade-off Evaluation of resist performance is tricky

because of the trade-off relationship.

Performance (efficiency) of resist

①How many photons can be absorbed?

②How many acids can be generated by a single photon?

Quantum efficiency: 2-3

Absorption coefficient: ~4 /mm

③How many dissolution inhibitor (protecting group) can be removed by a

single acid during the diffusion of unit length?

Effective reaction radius

The number of incident photons is limited because of the sensitivity requirement.

④How smoothly are the polymers dissolved in developer?

Relationship between LER and chemical gradient, fLER

dxdm

fLER LER

/

The chemical gradient is determined by ①－③

Generally, resists are evaluated using resolution, LER, and sensitivity.

We in collaboration with Selete evaluated the effective reaction radius and fLER

to understand the current status of chemically amplified resists.

Small Field Exposure Tool : SFET

Source (Xe DPP)

Exposure

chamber

Mask loader

Wafer loader

Wafer

track

Items Target

Specifications

NA 0.3

Illumination

mode

Annular(0.3/0.7),

x-slit

Field size 0.2 x 0.6 mm

Magnification 1/5

Wavefront error ＜0.9 nm rms

Flare ＜7% （MSFR）

Source power 0.5W @IF

Wafer size 300 mm

Selete Standard Resists (SSR1 to SSR7)

Resist screening at Selete (EIDEC)

Selete has reported highest performance resists among

those tested using SFET as

Representative patterning results SSR4

SSR5

24 nmHP, 10 mJ cm-2

(Top down)

26nmHP, 10 mJ cm-2

32nmHP, 10 mJ cm-2

32nmHP, 11 mJ cm-2

32nmHP, 8.5 mJ cm-2 Line width, LER

(0 degree)

Thickness

Changing HP

Chan

gin

g d

ose

24nmHP, 10 mJ cm-2 26nmHP, 10 mJ cm-2 32nmHP, 10 mJ cm-2

32nmHP, 8 mJ cm-2 32nmHP, 12 mJ cm-2

(b) SSR5

22 23 24 25 26 28 30 32 35 40 45 50 60

Exposure dose (mJ cm-2)

-30

30

Dev

iati

on

fro

m h

alf-

pit

ch

(no

min

al l

ine

wid

th)

(nm

)

-10

-20

0

20

10

0

15

LE

R (

nm

)

10

5

LER

Line width

7.9 mJ cm-2

22 23 24 25 26 28 30 32 35 40 45 50 60

Exposure dose (mJ cm-2)

(a) SSR4

22 23 24 25 26 28 30 32 35 40 45 50 60

Exposure dose (mJ cm-2)

9.4 mJ cm-2

22 23 24 25 26 28 30 32 35 40 45 50 60

Exposure dose (mJ cm-2)

Hal

f-p

itch

(n

m)

Hal

f-p

itch

(n

m)

Hal

f-p

itch

(n

m)

Hal

f-p

itch

(nm

)

(b) SSR5

22 23 24 25 26 28 30 32 35 40 45 50 60

Exposure dose (mJ cm-2)

-30

30

Dev

iati

on

fro

m h

alf-

pit

ch

(no

min

al l

ine

wid

th)

(nm

)

-10

-20

0

20

10

0

15

LE

R (

nm

)

10

5

LER

Line width

22 23 24 25 26 28 30 32 35 40 45 50 60

Exposure dose (mJ cm-2)

(a) SSR4

22 23 24 25 26 28 30 32 35 40 45 50 60

Exposure dose (mJ cm-2)

22 23 24 25 26 28 30 32 35 40 45 50 60

Exposure dose (mJ cm-2)

Hal

f-p

itch

(n

m)

Hal

f-p

itch

(n

m)

Hal

f-p

itch

(n

m)

Hal

f-p

itch

(nm

)

Footing

Top loss

Simulation result

0.0

0.2

0.4

0.6

0.8

1.0

0 10 20

Reconstruction of latent images from does-pitch matrices of line width and LER

24 nm

13.5mJ

11.5mJ

27.8, *

26.3, 8.1 26.5, 6.9 24.2, 6.4

22.5, *

Dissolution point

No

rmal

ized

pro

tect

ed u

nit

co

nce

ntr

atio

n

Distance (nm)

32.6, 7.6 28.0, 6.6 23.8, 6.4

19.7, 6.9 Bridge (Line width, LER)

dxdmLER

/

31.0

Chemical gradient

2.4 nm s-1

Dissolution rate

11.5

12.0 12.5 13.0

13.5

(SSR3)

Evaluated parameters

SSR3

SSR4

SSR5

SSR7

0.1

0.08-0.1

0.09

0.06

0.24-0.31

0.17-0.24

0.20

0.14

Matrix

Polymer

Polymer

Polymer

Fullerene

Resist Effective reaction radius (nm) fLER

0

2

4

6

8

10

-16 -8 0 8 16

Distance, x (nm)

Nu

mb

er o

f pro

tect

ed u

nit

, C

Average

-0.4s

+0.4s

s

Dissolution line

dx dx

dC/dx: 0.48 dx: 5.6

dC/dx

Q conc.: 0.02

LER Chemical reaction

0.3 nm

Anion

Proton

Reaction site

0

1

2

3

4

5

0 20 40 60 80 100

Challenges Half-pitch: 16 nm

Optical contrast: 0.8

Eff. reaction radius: 0.1 nm

fLER: 0.2

AG 5wt%

AG 10wt% AG 20wt%

AG 30wt%

Required LER

Target

LE

R (

nm

)

Exposure dose (mJ/cm2)

15 nm features has been reported to be resolved with 30 mJ/cm2 sensitivity.

Absorption coefficient: Sum of absorption cross section of atoms

Quantum efficiency: Acid generator concentration

Effective reaction radius: Activation energies for deprotection and diffusion

Required sensitivity

Summary (1)

We have investigated resist patterns fabricated using an EUV exposure

tool on the basis of reaction mechanisms. To simultaneously meet the

requirements of resolution, LER, and sensitivity, it is essential to

enhance the absorption coefficient of the resist, the quantum efficiency

of acids, and the effective reaction radius of catalytic chain reaction.

Especially, the absorption enhancement is the most important factor.

Actually, it is not difficult to increase the absorption coefficient of the

polymer, because the absorption coefficient against EUV is determined,

not by chemical bonds but the photoabsorption cross sections of atomic

elements. The problem is that the introduction of new atomic elements

to the polymer significantly changes the chemistry induced in the resist

films. It is not an easy task to increase the effective reaction radius of

chemical reactions in the new resist platform to the same level as that

in well-studied organic resist polymers.

Anion bound resist

It has been considered that the acid diffusion is the most serious problem for the improvement of resist performance in terms of trade-off relationship.

Recently, a chemically amplified resist with anion-bound acid generator attracted much attention.

Promising material for 16 nm node and beyond

Reaction mechanism is unknown.

An acid diffusion model in a chemically amplified resist with anion-bound acid generator

New simulation code was developed on the basis of radiation chemistry.

For the development of resist materials, particularly, used in the 16 nm node and beyond, it is important to understand the reaction mechanism of catalytic reactions induced in the anion-bound resists.

Basic idea

Similar sensitivity to that of blend type resists

Protons can diffuse. (Experimentally confirmed)

High resolution

Protons cannot diffuse freely.

24 nm 18.16 mJ cm-2 22 nm 20 nm

SFET

Annular

The efficiency of acid catalytic reaction is not bad.

Initial distribution

Proton Anion

PEB PEB

－ ＋ －

＋

－ ＋ － ＋ －

＋

－

＋

－

＋ － ＋

－

＋

－

＋

×

Proton diffusion model

○

Protons diffuse under the electric

field produced by anions.

Anion-bound resist

H. Yamamoto et al., Jpn. J. Appl. Phys. 43 (2004) L848.

Representative patterning results

26 nmHP, 18.16 mJ cm-2

Line width, LER

Changing HP

Chan

gin

g e

xposu

re d

ose

22 nmHP 18.16 mJ cm-2

Chemically amplified resist with anion-bound AG

32 nmHP 18.16 mJ cm-2

45 nmHP 18.16 mJ cm-2

28 nmHP 16.50 mJ cm-2

28 nmHP 17.33 mJ cm-2

28 nmHP 18.16 mJ cm-2

Conventional dose-pitch analysis Relationship between LER and chemical gradient

-20

20

Dev

iati

on

fro

m h

alf-

pit

ch

(no

min

al l

ine

wid

th)

(nm

)

0

0

10

LE

R (

nm

)

5

(b) LER (a) Line width

22 23 24 25 26 28 30 32 35 40 45 50 60

Exposure dose (mJ cm-2)

Hal

f-p

itch

(n

m)

22 23 24 25 26 28 30 32 35 40 45 50 60

Exposure dose (mJ cm-2)

Hal

f-p

itch

(n

m)

Analysis of dose-pitch matrices of anion-bound resist

-20

20

Dev

iati

on

fro

m h

alf-

pit

ch

(no

min

al l

ine

wid

th)

(nm

)

0

0

10

LE

R (

nm

)

5

22 23 24 25 26 28 30 32 35 40 45 50 60

Exposure dose (mJ cm-2)

Hal

f-p

itch

(nm

)

22 23 24 25 26 28 30 32 35 40 45 50 60

Exposure dose (mJ cm-2)

Hal

f-p

itch

(nm

)

Exp

erim

en

tal

Sim

ula

tion

dxdmLER

/

20.0

m: normalized protected unit concentration

EUV photon

Ionization

Ionization Next ionization or excitation

Resist

Thermalization

photon

electron

E < Eth

E > Eth

hn

hn-Ie

-Ie

Sensitization mechanism of EUV resists

+

-

+

+

- - Acid generator

(1) Absorption

(2) Deceleration

Eth < E < hn-Ie

(3) Deceleration

25 meV < E < Eth

Eth: Threshold energy for electronic excitation

Ie: Ionization energy

Simulation processes

(4) Electron diffusion and reaction

E=25 meV

Proton generation

Anion generation Post exposure baking (PEB)

Simulation of proton diffusion and deprotection

+

nE

r tDtkT

eDddd 6

Electric field

Random vector Displacement of proton

Monte Carlo simulation

+

+

+

+

+

Fixed

Mobile

Coulomb interaction

Immediately after exposure, protons are separated from anions because of their different origins.

Protected unit (polymer)

If a proton reaches a protected unit within the reaction radius, Rp, the deprotection is induced. Rp

Rq

If a proton reaches a quencher within the reaction radius, Rq, the proton is neutralized.

Quencher

Deprotection

Neutralization

0.0

0.5

1.0

1.5

2.0

-22 -11 0 11 22

Comparison with SFET exposure (22 nm half-pitch)

0.0

0.5

1.0

1.5

2.0

-22 -11 0 11 22

0.025 nm-3

0.01 nm-3

Distance (nm)

Pro

tect

ed u

nit

co

nc.

(n

m-3

) Dose: 18.16 mJ cm-2

PEB 50 s

PEB 100 s

PEB 150 s

Distance (nm)

Pro

tect

ed u

nit

co

nc.

(n

m-3

)

Dissolution point

Chemical gradient（Exp.）

Pattern boundary (Exp.)

PEB time dependence

0.025 nm-3

0.020 nm-3

0.015 nm-3

Quencher concentration dependence

0.010 nm-3

Quencher concentration

PEB time

Space Line Line

Space Line Line

0.0

0.5

1.0

1.5

2.0

-26 -13 0 13 26

0.0

0.5

1.0

1.5

2.0

-26 -13 0 13 26

Dose dependence of latent image (26 nm half pitch)

Distance (nm)

Pro

tect

ed u

nit

con

c. (

nm

-3)

Distance (nm)

Pro

tect

ed u

nit

con

c. (

nm

-3)

Dose: 18.16 mJ cm-2

Dose: 17.33 mJ cm-2 Dose: 18.99 mJ cm-2

0.0

0.5

1.0

1.5

2.0

-26 -13 0 13 26

Dissolution line

Chemical gradient（Exp.）

Pattern boundary (Exp.)

LER (Exp.)

Distance (nm)

Pro

tect

ed u

nit

co

nc.

(n

m-3

)

Space Line Line

Space Line Line Space Line Line

0.0

0.5

1.0

1.5

2.0

-22 -11 0 11 22

0.0

0.5

1.0

1.5

2.0

-25 -12.5 0 12.5 25

0.0

0.5

1.0

1.5

2.0

-24 -12 0 12 24

0.0

0.5

1.0

1.5

2.0

-23 -11.5 0 11.5 23

Half-pitch dependence of latent image

Distance (nm)

Pro

tect

ed u

nit

con

c. (

nm

-3)

Distance (nm)

Pro

tect

ed u

nit

con

c. (

nm

-3)

24 nm HP 25 nm HP

Distance (nm)

Pro

tect

ed u

nit

co

nc.

(n

m-3

)

Distance (nm)

Pro

tect

ed u

nit

co

nc.

(n

m-3

)

22 nm HP 23 nm HP

Dissolution line

Chemical gradient

（Exp.）

Pattern boundary (Exp.)

Dose: 18.16 mJ cm-2

LER (Exp.)

Space Line Line Space Line Line

Space Line Line Space Line Line

The catalytic chain reaction induced in a chemically

amplified resist with anion-bound AG was modeled.

The calculated latent images were compared with the

resist patterns fabricated using SFET of EIDEC. In the

preliminary examination, the calculated images well

agreed with the experimental results. The detailed

analysis of resist patterns using the developed

simulation code is ongoing to obtain the material

design strategy for 16 nm node and beyond.

Summary (2)

This work was partially supported by the New Energy and

Industrial Technology Development Organization (NEDO).

Acknowledgement

EUV photon

Ionization

Ionization Next ionization or excitation

Resist

Thermalization

photon

electron

E < Eth

E > Eth

hn

hn-Ie

-Ie

Reduction of wavelength

+

-

+

+

- - + Acid generator

The electron with thermal energy can sensitize acid generators.

Activation energy for dissociative electron attachment : ~0

(1) Absorption

(2) Deceleration

Eth < E < hn-Ie

(3) Deceleration

25 meV < E < Eth

Eth: Threshold energy for electronic excitation

Ie: Ionization energy

Simulation processes

Wavelength dependent

(4) Electron diffusion and reaction

E=25 meV

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

0 10 20 30

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

0 10 20 30

Distance, (nm) 222 zyx Distance, x (nm)

Pro

bab

ilit

y d

ensi

ty (

nm

-1)

×

qu

antu

m e

ffic

ien

cy

Pro

bab

ilit

y d

ensi

ty (

nm

-1)

×

qu

antu

m e

ffic

ien

cy

2 nm

13.5 nm

12 nm

2 nm

13.5 nm

e- e-

e-

e-

+ + + +

EUV photon

Multiple scattering at <Eth

line & space

Wavelength dependence of resolution blur caused by secondary electrons

Spherical coordinate x coordinate

Spherically symmetric

Electron migration after thermalization

per unit spherical

shell thickness

0

1

2

3

4

5

6

7

8

0 5 10 15

0

1

2

3

4

5

0 5 10 15

Wavelength (nm)

Res

olu

tio

n b

lur

(nm

)

Optical

Secondary electron

Total

Wavelength (nm)

Res

olu

tio

n b

lur

(nm

) Optical

Secondary electron

Total

Acid image resolution (acid diffusion length does not depend on wavelength)

Performance of conventional chemically amplified resists –Resolution-

Optical blur, boptical

Secondary electron blur, belectron

NAkCD

1

22

electronopticalt bbb

2

CDboptical

Total blur,

11 NA

k5.01

NA

k

Average distance

0

1

2

3

4

5

6

0 5 10 15

Wavelength (nm)

Lin

ear

abso

rpti

on

co

effi

cien

t (m

m-1

)

G v

alu

e [(

10

0 e

V)-1

]

0

1

2

3

4

5

6

Inner shell excitation

Performance of conventional chemically amplified resists –Sensitivity-

Absorption coefficient

Acid generation efficiency per unit absorbed dose (G value)

Auger electron (light element)

Fluorescence X-rays (heavy element)

Photoelectron emission

Compton scattering

Carbon K-edge

Acid concentration (acid diffusion length does not depend on wavelength)

Positively affect resolution

Negatively affect resolution

The W-value increases by 10-20% at the absorption edge of the inner shell.

0

1

2

3

4

5

0 5 10 15

0

2

4

6

8

10

0 10 20 30

Distance, x (nm) Wavelength (nm)

Pro

bab

ilit

y d

ensi

ty

×

qu

antu

m e

ffic

ien

cy (

nm

-1)

Res

olu

tio

n b

lur

(nm

) 0

5

10

15

20

25

0 5 10 15

Wavelength (nm)

Rat

io (

%)

Optical

Secondary electron

Total

e-

e-

e-

e-

+ + +

+

EUV photon

E > Eth

E > 25 meV

5.01 NA

k

High-resolution conventional resist such as PMMA

utilize this part for the chemical reaction for pattern

formation (main chain scission).

Non-chemically amplified resists

Chemically amplified resists utilize this part for

the decomposition of acid generators.

Distribution of radical cation Resolution blur of L&S pattern

Summary (3)

The wavelength dependence of lithography resolution was

investigated in the wavelength region of extreme

ultraviolet. The resolution is expected to be highest at a

wavelength of 3-5 nm, depending on NA of exposure tools.

In the case of low-NA tools, the merit of wavelength

reduction from 13.5 nm is significant. However, the merit

of wavelength reduction is lost in the case of high-NA

tools, particularly when the increase in transparency of the

resist with the reduction in wavelength is taken into

account. One of the keys to the realization of 6.67nm

lithography is the development of high-absorption resists.

This work was partially supported by the New Energy and

Industrial Technology Development Organization (NEDO).

Acknowledgement