Experimental evaluation of the impact of EUV pellicles on ......transmission at EUV, low reflectivity, low scattering, low thermal expansion coefficient, and high mechanical stability

Experimental evaluation of the impact of EUV pellicles on reticle imaging Iacopo Mochi

a, Marina Timmermans

b, Emily Gallagher

b, Marina Mariano

b, Ivan Pollentier

b,

Rajendran Rajeeva, Patrick Helfenstein

a, Sara Fernandez

a, Dimitrios Kazazis

a and Yasin Ekinci

a

aPaul Scherrer Institut, 5232 Villigen, Switzerland

bIMEC, Kapeldreef 75, 3001 Leuven, Belgium

ABSTRACT

The purpose of EUV pellicles is to protect the surface of EUV lithography masks from particle contamination. Currently

several pellicle prototypes are being developed. It is important to ensure that the optical characteristics of the pellicle

membrane do not critically affect the reticle image quality. We present here a study of the impact of a few selected

EUV pellicle prototypes on the quality and the contrast of the reticle image obtained with an actinic lensless

microscope.

Keywords: EUV, pellicle, actinic pattern inspection, defect inspection, lensless microscopy, coherent diffraction

imaging, carbon nanotubes.

1. INTRODUCTION

Extreme ultraviolet (EUV) masks will eventually be equipped with protective pellicles. The pellicle, installed at a

distance of about 2.5 mm from the mask surface, will hold fall-on particles out of the imaging plane, thereby

minimizing their impact on wafer yield. Although non-pellicle lithography is being considered, this is a risky and non-

ideal solution. The challenge is to find a membrane with suitable optical and mechanical characteristics: high

transmission at EUV, low reflectivity, low scattering, low thermal expansion coefficient, and high mechanical stability.

The pellicle characteristics have an impact on the imaging performance of a scanner or of an inspection system, and for

this reason, an accurate evaluation of their optical properties must be performed.

We tested different pellicle samples on RESCAN (Reflective-mode EUV mask SCANning microscope) to verify their

impact on the aerial image formation. RESCAN is an actinic pattern inspection platform currently under development at

Paul Scherrer Institute (PSI). [1-2] It consists of a lensless scanning microscope based on coherent diffraction imaging

(CDI). In RESCAN, the sample is illuminated with a coherent EUV beam and the reflected diffraction beam is recorded

by a pixel detector. Multiple diffraction patterns are combined together to reconstruct the sample image and phase,

using ptychography. [3] Since RESCAN does not have any imaging optics close to the reticle, there is no physical

obstacle to mount a pellicle on top of the sample and this makes it a flexible and effective tool for prototyping and

investigating the performance of new membranes and materials.

In this work, we present the results of the image reconstruction of an EUV reticle test sample with different carbon

nanotubes (CNT) pellicles with and without coating and we compare the images to a reference obtained with the bare

sample. We also present an estimate of the scattering introduced by different types of EUV-transparent membranes

and its effect on the image formation. Finally, we demonstrate how RESCAN can be used not only as an actinic mask

inspection platform, but also as an efficient tool to inspect the pellicle surface for defects.

2. MEASUREMENT SETUP

RESCAN

The measurements were carried out using RESCAN, a lensless microscope dedicated to EUV mask defect inspection

operating at the XIL-II beamline of the Swiss Light Source. RESCAN’s working principle consists in collecting images of

the sample and comparing them to the image of a reference die or to an aerial image calculated from the reticle’s

design to identify the presence of defects. [1] The images of the sample are obtained through CDI which allows for

retrieval of both the phase and the magnitude of the reticle’s surface. The CDI approach used in RESCAN is based on

Photomask Technology 2018, edited by Emily E. Gallagher, Jed H. Rankin, Proc. of SPIE Vol. 10810, 108100Y · © 2018 SPIE · CCC code: 0277-786X/18/$18 · doi: 10.1117/12.2502480

Proc. of SPIE Vol. 10810 108100Y-1Downloaded From: https://www.spiedigitallibrary.org/conference-proceedings-of-spie on 18 Apr 2019Terms of Use: https://www.spiedigitallibrary.org/terms-of-use

Incoming beam

Support Pellicle

EUV reticle sample

utgoing beam

cattered light

ptychography and consists of scanning the sample with an EUV illumination probe, collecting the relative diffraction

patterns and propagate them back to the sample plane with an iterative procedure to retrieve the phase and

magnitude of the reticle. The sample scan is performed moving the sample along x and y axes with a piezo-motor with

about 5 nm accuracy. RESCAN is a research prototype and it is currently located in a small vacuum chamber that limits

the size of the samples to 20×20 mm2 and is equipped with a stage with a travel range of 200 μm.

The optical layout of RESCAN, shown in Figure 1A, consists of an aperture, a focusing toroidal mirror and a folding

mirror. The XIL-II beamline is equipped with a monochromator that allows RESCAN to operate with a narrow

bandwidth Δ⁄ = 1500, which is an essential requirement for coherent diffraction imaging systems [4]. The beam is

focused by the toroidal mirror and folded onto the reticle sample with a variable numerical aperture (NA) ranging from

0.002 to 0.015. The beam footprint on the sample can be tuned by changing the position of the folding mirror or the

focusing mirror. The beam, reflected and diffracted by the sample, is collected by a 2048×2048-pixel detector at a

distance d of 62 mm. The maximum resolution of RESCAN depends on the NA of the detector defined as: NA = sin(arctan / ), Where s is the detector half size. The detector used for the measurements presented here has a pixel size of 13.5 μm

and its half size s is 13.8 mm. The image resolution is therefore: ≈ 2NA = 35nm. RESCAN is equipped with the adjustable pellicle mount sketched in Figure 1B. It consists of a 35×35 mm

2 support that

can be installed as close as 0.5 mm from the sample surface and can hold a pellicle sample mounted on a 30×30 mm2

wafer.

Figure 1. A. RESCAN optical layout. The EUV beam from the synchrotron is focused and folded onto the sample plane by a

toroidal mirror (M1) and by a plane mirror (M2). Both mirrors are coated with a multilayer to ensure maximum EUV

reflectivity at their respective incidence angle. The reflected and diffracted beam is collected by a CCD camera at a

distance of 62 mm from the sample. B. RESCAN pellicle mount detail.

M1M2

Detector

EUV beam

EUV reticle sample EUV reticle sample

Support Pellicle

A B


10 nm

--

--- _..J1

P-

::kr" -P

7

J V,

t11

I.I

1

REM

Figure 2. S

Pellicle samp

The developm

evaluated an

[5-7] Among

reflectivity ~

sample). [8]

radiation insi

study, we ex

different pro

mesh. The su

explored to s

pellicles, but

capabilities. A

currently und

Through ana

presence of d

the impact of

is a 20×20 mm

6° incidence

silsesquioxan

Pellicle defec

The size of pe

micrometers

illuminating t

shadow on th

shows how a

ratio betwee

considering t

SEM measure

Scanning electr

ples

ment of suitab

d tested both

them, carbon

0.001% and

The lifetime

ide a scanner

xamined diffe

otective meta

upport mesh

strengthen th

the large st

All the pellicle

der developm

3.

lysis of RESCA

defects on the

f the pellicle o

m2 square silic

angle. The

ne (HSQ). We n

ct inspection

ellicle defects

[9]. The raw

the sample th

he diffraction

CNT support

en the distan

he detector p

ed value of 7 μ

ron micrograph

ble EUV pellic

h experimenta

n nanotube (

good mecha

of CNT mem

r. To mitigate

erent CNT-ba

l coatings an

in these sam

he membrane

rands of the

e samples sho

ent.

THROUGH

AN raw data,

e surface are

on the imagin

con wafer wit

sample is pa

note that this

that can caus

w data collec

hrough the pe

pattern as it

t mesh is visib

nce of the de

pixel size, we e

μm.

hs of a random C

cle samples is

ally and with

CNT) pellicles

anical stability

mbranes can b

this effect, it

ased pellicles

nd aligned mu

mples are a sp

e. MW-CNT m

mesh provid

own here are

H-PELLICLE IM

the direct stu

evaluated. M

ng contrast an

th a multilaye

atterned with

sample has lo

se a significan

ted in RESCA

ellicle and sca

t propagates t

ble in the reco

etector and

estimated the

CNT membrane

currently ong

simulations t

s proved to b

y (<0.08 mm

be limited by

t is possible to

including ra

ulti-walled (M

parse net of m

membranes w

e a useful ex

e based on la

MAGING OF

udy of the pe

Moreover, from

nd on the patt

r coating desi

h a random,

ower contrast

nt CD error in

AN consist of

anning the EU

through the m

orded diffracti

the pellicle f

e size of the C

e (left) and an a

going. Various

to assess their

be a promising

deflection un

y the exposur

o coat the me

andom single

MW) CNTs (fig

macro-CNT bu

with support m

xample to de

st year’s tech

EUV REFLEC

llicle is enabl

m the reconst

tern fidelity. T

igned for max

logic-like de

t than the stat

the aerial ima

f a set of far

UV reticle. A

membrane fro

ion pattern. C

from the foc

CNT bundle to

aligned CNT me

s types of pell

r mechanical

g option with

nder a 2 Pa p

re to hydroge

embrane with

-walled (SW)

gure 2-right)

undles with a

mesh are no

emonstrate RE

hnology. More

CTIVE SAMPL

ed and its sca

tructed sampl

The EUV mask

ximum reflecti

esign on a 14

te-of-the-art E

age is of the o

r-field diffrac

defect on the

om the sampl

Calculating the

cal plane (loc

o be 6.5 μm, in

embrane (right)

icle membran

and optical c

h EUV transm

pressure for

en ions gene

h a protective

CNTs (figure

with and wit

a diameter of

longer candid

ESCAN’s pellic

e advanced p

LES

attering prope

le images, we

k sample used

ivity (>60%) a

40 nm layer

EUV reticles.

order of magn

tion patterns

e pellicle can

e to the dete

e pixel magnif

cated on the

n good agreem

) [5].

nes have been

haracteristics

mission ~>95%

a 10×10 mm

rated by EUV

e layer. In this

e 2-left) with

thout support

f about 7 μm

dates for EUV

cle inspection

prototypes are

erties and the

e can estimate

d for this work

t 13.5 nm and

of hydrogen

itude of a few

s obtained by

cast a visible

ector. Figure 3

fication as the

sample) and

ment with the

n

s.

%, 2

V

s

h

t

,

V

n

e

e

e

k

d

n

w

y

e

3

e

d

e


Figure 3. A. Diffraction pattern of a reticle sample covered by a MW-CNT pellicle with support mesh. The picture is cropped

to 1800×1800 pixels and displayed on a logarithmic intensity scale. The diffraction pattern corresponds to a Manhattan-

geometry reticle with a 45°orientation with respect to the CCD axes. The black square delimits a region of the diffraction

pattern where two crossing strands of the support mesh cast a visible shadow. B. Detail of the diffraction pattern

delimited by the black square on A showing the shadow of the two crossing strands. The image is displayed on a

compressed logarithmic scale.

Pellicle scattering

In the absence of the pellicle, the diffraction patterns recorded by RESCAN provide an immediate measure of the

scattering irradiance coming from the sample surface and from the condenser optics. The pellicle generates two

additional diffraction contributions from the interaction with the incoming and outgoing beam. In addition, large

defects on the pellicle surface can cast a shadow on the diffraction pattern generated by the sample. In RESCAN it is

relatively straightforward to evaluate these contributions by comparing the diffraction patterns recorded through

pellicle to a reference obtained from the bare sample. We inspected a region free of absorber patterns on the EUV

mask sample to avoid the strong diffraction signature from the reticle pattern and simplify the data interpretation. The

images in the top row of Figure 4 show the recorded diffraction patterns in log scale. The irradiance intensity is

normalized to the maximum irradiance of the no-pellicle case (A). The images in the bottom row of Figure 4 show the

irradiance intensity of the reflected beam that corresponds to the condenser pupil projection on the detector plane.

From a visual observation of the diffraction patterns, there seems to be little or no difference between the random SW-

CNT pellicle (B) and the reference (A) indicating that the main effect of this membrane is the beam attenuation, visible

from the comparison between the average intensities of the reflected beam. The aligned MW-CNT pellicle (C, D top

images in Figure 4) shows instead a visible diffraction contribution that corresponds to the dominant scattering

direction of the aligned CNTs. Although the individual CNTs are not visible in the diffraction patterns, in the pupil

projection image it is possible to distinguish their alignment direction from the presence of the stripes due to the non-

uniformity of the CNT’s distribution.

6.5 μm

A B


5

10

1.8

1.6

1.4

1.2

B

iC

-10

1

0.8

0.6

0.4

0.2

L Sca

tterin

g in

ten

sity

(log

)R

efle

ctio

n in

ten

sity

(lin)

OW O 7 N

OIL

)r

r-O

O O

O

Figure 4. T

zero-orde

Reference

in the pup

surfaces.

C rotated

To make a m

intensity fro

convenient to

scattering int

in the middle

Where I is th

respectively.

Where λ is th

The top row sh

er diffraction or

e case without

pil projection in

B. Random SW

by 90°.

more quantitat

m the diffrac

o use an estim

tensity as a fu

e of the pupil p

e measured d

Expressing r a

he wavelength

ows the log sca

r simply the pro

pellicle. The im

ntensity and the

CNT pellicle w

tive evaluatio

ction pattern

mator that do

unction of the

projection as

diffraction irra

as a function o

h and z is the d

ale diffraction p

ojection of the e

mage was collect

e large blemish

ith metal coatin

n of the pelli

ns along a sp

oes not depen

e diffraction. T

shown in figu( ) =adiance and r

of the spatial

distance betw

patterns observ

exit pupil of the

ted on a flat reg

in the lower pa

ng. C. Aligned M

cle scattering

pecific axis. H

nd on the sca

This is done a

re 5. The scat1(2 ∆r + ∆r and Δ are t

frequency of ( ) = .ween the samp

ved in RESCAN.

e condenser op

gion of the sam

art of the image

MW-CNT pellicl

g contribution

However, wh

attering direct

veraging the

ttering intensi

) d . the inner radi

the scattering

ple and the de

The bottom row

tics displayed o

mple, covered w

e come from th

e (uncoated). D

, it is sufficien

hen comparin

tion. In this ca

intensity ove

ty S(r) is then

us and thickn

g features we

etector.

w shows the de

on a linear scale

with HSQ. The ro

he condenser m

D. Same pellicle

nt to extract t

ng different s

ase, we repor

r a circular cr

:

ness of the cir

obtain:

etail of the

e. A.

oughness

mirrors

e shown in

the irradiance

samples, it is

rt the average

own centered

rcular crown C

e

s

e

d

C


Figure 5. Evaluation of the scattering intensity as a function of the scattering angle is carried out averaging the recorded

irradiance I over a circular crown C centered in the middle of the entrance pupil projection. The radius r of the circular

crown corresponds to the scattering angle. The right panel shows a plot of the scattering intensity as a function of the

spatial frequency of the scattering features for different pellicles.

In Figure 5 we show the average scattering intensity as a function of the spatial frequency of the diffracting features for

the same pellicles shown in Figure 4. While the scattering from the random SW-CNT pellicle has the same behavior of

the no-pellicle reference case, the aligned MW-CNT pellicle shows a higher scattering intensity in the spatial frequency

range between 0.6 and 2 μm-1

. This scattering will superimpose to the diffraction of features with the same spatial

frequency range and reduce their contrast in the aerial image.

Image quality evaluation

Mask defect detection in RESCAN requires the reconstruction of the reticle image. Although this image does not

necessarily correspond to the aerial image generated by a scanner where the illumination is engineered to optimize the

process window, we can still use it to estimate the impact of EUV pellicles on the contrast and on the pattern fidelity.

We performed through-pellicle inspections of the same sample using a constant parameter set, including illumination

NA, scanning area, scanning step size and image reconstruction algorithm. As for the scattering measurements, we

collected a dataset without pellicle as a reference. Figure 6 shows the reconstructed image magnitude for the

reference case (A), for a MW-CNT pellicle with support mesh (B), and for two random SW-CNT pellicles with different

metal coatings (C, D). From a qualitative inspection, it is evident that the two random SW-CNT pellicles have less impact

on the image than the aligned MW-CNT. While images C and D do not show any difference from A, a clear contrast loss

and some artifacts can be observed in image B. To evaluate quantitatively the pellicle’s impact on the reconstructed

image quality, we measured the image contrast and a fidelity metric based on the comparison of the images with a

reference calculated from the reticle design.

The random scattering from the pellicle will contribute to the system flare and reduce the image contrast. To evaluate

the contrast of the reconstructed pattern, we separated the bright regions of the image corresponding to the

multilayer areas from the dark ones corresponding to the absorber. The selection was carried out using an arbitrary

threshold value. We calculated the average and the standard deviation of the intensity in the two regions and we used

the traditional contrast definition: = −+ ± 2 ∆ − ∆( + ) , where and are the average intensities of the bright and the dark regions respectively and ∆ and ∆ are the

corresponding standard deviations. Since there is a focal gradient in the reconstructed image, this calculation was

carried out in a 4×4 μm2 region centered on the best focus spot.

No pellicle

SW CNT (Coating 1)

Aligned MW CNT (support mesh)

Aligned MW CNT (support mesh) 90°

Scatt

ering in

ten

sity [

a. u

.]

10−6

10−5

Spatial frequency [μm-1]

0.5 1 2 5 10


EJ

COUNo)wE

0.60

0.55

0.50

0.45Mw

Firet,zn

ry.,),

J CNT 1 MW CNT 2

B

0.5

ro

D

10.5

J0

No P'ellicle SV

J

V CNT Coating1

2 SW CNT C:oating 1

Figure 6. T

is 200 nm

an aligned

protective

coating.

Figure 7. A

the b

EUV

scatt

The values sh

also have a lo

Through-pellicl

. The sample h

d MW-CNT mem

e metal coating

Average contra

bright and dark

multilayer desi

tering show the

hown in Figur

ower contrast

e image recons

as a tilt that ma

mbrane with su

g. D. Image reco

ast of through p

regions in the

igned for maxim

e highest image

re 7 indicate

with respect

structions. A. Re

anifests with a

upport mesh. C

onstruction wit

pellicle reconstr

images. The ma

mum reflectivity

e contrast loss.

that if aligne

to random co

eference image

defocus in the

. Image reconst

h a random SW

ructed images.

ask sample use

y at 6° incidenc

d MW-CNT p

oated SW-CNT

e reconstructed

lower part of th

truction with a

W-CNT membran

The contrast is

ed is generated

ce angle. The m

pellicles have

T pellicles. The

d without pellic

he image. B. Im

random SW-CN

ne with a differ

calculated from

patterning a 14

membranes exhi

a stronger sc

e overall cont

le. The CD of th

mage reconstruc

NT membrane w

rent protective

m the average v

40 nm HSQ film

ibiting the stron

cattering cont

rast level is re

he pattern

ction with

with a

metal

values of

m on an

ngest

tribution, they

emarkably low

y

w


for all these cases. This is partly due to the fact that this is an average contrast and it is not based on the absolute

maximum and minimum values in the images, but also to the fact that the absorber of the mask sample under

investigation consists of a 140 nm layer of HSQ. The absorption coefficient of HSQ ranges from 4.7 e-3

to 7.4 e-3

nm-1

.

Considering a total thickness of 2⋅140/cos(6°)nm, the EUV intensity transmitted through the absorber ranges from

12% to 27% of the intensity reflected by the bare multilayer leading to a contrast cap that goes from 73% to 88%.

Figure 8. Reconstructed image fidelity is estimated as the standard deviation of the difference between the reconstructed

image and a simulated aerial image calculated from the sample design. The smaller the value of the standard deviation,

the higher the reconstructed image fidelity. The reconstruction obtained through the random SW-CNT membrane with

metal coating 1 showed little or no difference in image fidelity with respect to the no pellicle case.

To evaluate the reconstructed image quality, we defined a fidelity metric based on an ideal reference image. We

calculated the reference aerial image by propagating the binary GDS design of the sample through a perfect lens with

an aperture equal to the one defined by RESCAN's detector. The reference aerial image does not include mask 3D

effects or imperfections in the mask pattern. The reconstructed image and the reference image are aligned using a

least square registration algorithm, normalized for best contrast matching and cropped to a 4×4 μm2 area centered in

the best focus spot as done for the contrast evaluation case. The image fidelity is defined as the standard deviation of

the difference between the two images. The lower the standard deviation, the closer the reconstructed image is

assumed to be to the ideal case. The results of the image fidelity evaluation are shown in Figure 8, where we observed

that one of the random SW-CNT pellicles with metal coating (corresponding to image C in figure 6) exhibits little or no

difference from the reference image reconstructed without the pellicle.

4. CONCLUSIONS AND OUTLOOK

We used the RESCAN defect inspection platform to investigate the effects of different CNT-based pellicles on the

actinic image of EUV mask samples. In particular, we demonstrated a method to measure the pellicle-induced

scattering in double-pass mode, we evaluated the contrast loss caused induced by the pellicle membrane and we

defined an image fidelity metric to compare the effects of the pellicles on the reconstructed image quality.

We observed that the metal-coated random SW-CNT membranes show a scattering behavior similar to the bare sample

case, in line with the results described by previous studies [7]. Aligned MW-CNT membranes show instead a higher

scattering contribution in the spatial frequency region between 0.6 and 2 μm-1

, suggesting that reticle structures in this

spatial frequency range may suffer a contrast loss in the aerial image.

The image contrast and the fidelity metric we defined showed that metal-coated random SW-CNT pellicles perform

better than the aligned MW-CNT ones. In particular, one of the two metal coated pellicles, shows little or no

performance difference from the no-pellicle reference case.

We also demonstrated that RESCAN can detect pellicle defects with size below 6.5 μm which is approximately the

smallest dimension that can cause a critical CD error in the reticle aerial image. This is an inherent advantage of CDI-

based inspection tools that directly record the diffraction patterns of the sample under investigation. Thanks to this


capability, RESCAN can be used not only as an actinic mask inspection platform, but also as an efficient tool to inspect

the pellicle surface for defects, to characterize its scattering properties and to evaluate the impact of pellicles on the

reticle aerial image contrast and quality.

ACKNOWLEDGMENTS

The authors wish to thank the pellicle suppliers, Canatu Oy and the Nano-Science & Technology Center of Lintec of

America Inc., who agreed to share the samples used in this work, and the PSI ALM technical team, Markus Kropf,

Michaela Vockenuber and José Gabadinho for their contribution to the experiments at the XIL-II beamline.

REFERENCES

[1] Iacopo Mochi, Patrick Helfenstein, Rajendran Rajeev, Sara Fernandez, Dimitrios Kazazis, Shusuke Yoshitake, Yasin

Ekinci, “Actinic inspection of EUV reticles with arbitrary pattern design”, Proc. SPIE 10450, International

Conference on Extreme Ultraviolet Lithography 2017, 1045007 (16 October 2017).

[2] Iacopo Mochi, Patrick Helfenstein, Istvan Mohacsi, Rajeev Rajendran, Dimitrios Kazazis, Shusuke Yoshitake, Yasin

Ekinci, “RESCAN: an actinic lensless microscope for defect inspection of EUV reticles,” Journal of

Micro/Nanolithography, MEMS, and MOEMS 16(4), 041003 (27 July 2017).

[3] Patrick Helfenstein, Rajendran Rajeev, Iacopo Mochi, Armin Kleibert, C. A. F. Vaz, and Yasin Ekinci, “Beam drift and

partial probe coherence effects in EUV reflective-mode coherent diffractive imaging,” Opt. Express 26, 12242-

12256 (2018).

[4] Kewish CM, Thibault P, Dierolf M, Bunk O, Menzel A, Vila-Comamala J, Jefimovs K, Pfeiffer F., “Ptychographic

characterization of the wavefield in the focus of reflective hard X-ray optics”. Ultramicroscopy. 2010

Mar;110(4):325-9. doi:10.1016/j.ultramic.2010.01.004.

[5] Marina Y. Timmermans, Ivan Pollentier, Jae Uk Lee, Johan Meersschaut, Olivier Richard, Christoph Adelmann,

Cedric Huyghebaert, Emily E. Gallagher, "CNT EUV pellicle: moving towards a full-size solution", Proc. SPIE 10451,

Photomask Technology, 104510P (16 October 2017)

[6] Ivan Pollentier, Johannes Vanpaemel, Jae Uk Lee, Christoph Adelmann, Houman Zahedmanesh, Cedric

Huyghebaert, Emily E. Gallagher, “EUV lithography imaging using novel pellicle membranes”, Proc. SPIE 9776,

Extreme Ultraviolet (EUV) Lithography VII, 977620 (18 March 2016)

[7] Frank Scholze, Christian Laubis, Michael Krumrey, Marina Y. Timmermans, Ivan Pollentier, Emily E. Gallagher, “EUV

optical characterization of alternative membrane materials for EUV pellicles “, Proc. SPIE 10451, Photomask

Technology, 104510R (16 October 2017)

[8] Emily Gallagher, Marina Y. Timmermans, Ivan Pollentier, Jae Uk Lee, Marina Mariano, Christoph Adelmann, Cedric

Huyghebaert, Frank Scholze, Christian Laubis, “CNTs in the context of EUV pellicle history,” Proc. SPIE 10583,

Extreme Ultraviolet (EUV) Lithography IX, 105831E (23 April 2018)

[9] Guk-Jin Kim, In-Seon Kim, Sung-Gyu Lee, Michael Yeung, Min-Su Kim, Jin-Goo Park, Hye-Keun Oh, "Influence of a

wrinkle in terms of critical dimension variation caused by transmission nonuniformity and a particle defect on

extreme ultraviolet pellicle," J. Micro/Nanolith. MEMS MOEMS 16(4) 041008 (12 October 2017)


Documents

Experimental evaluation of the impact of EUV pellicles on ......transmission at EUV, low reflectivity, low scattering, low thermal expansion coefficient, and high mechanical stability