w.elsevier.com/locate/tsf

Thin Solid Films 504 (

Characterization of chemically amplified resist for X-ray lithography by

Fourier transform infrared spectroscopy

T.L. Tan *, D. Wong, P. Lee, R.S. Rawat, S. Springham, A. Patran

Natural Sciences and Science Education, National Institute of Education, Nanyang Technological University, 1 Nanyang Walk, Singapore 637616, Singapore

Available online 20 October 2005

Abstract

SU-8 resist was characterized for X-ray lithography from a plasma focus source by studying its cross-linking process using Fourier transform

infrared (FT-IR) spectroscopy. The cross-linking process of the resist during post-exposure bake (PEB) was accurately monitored using the infrared

absorption peaks at 862, 914, and 1128 cm� 1. Results showed that the cross-linking of SU-8 was effectively completed at the exposure dose of 2500

mJ/cm2 for resist thickness of 25 Am. Reliable processing conditions consisted of an intermediate PEB at 65 -C for 5 min, with the PEB temperature

ramped up to 95 -C over 1.5 min and then followed by a final PEB at 95 -C for 5 min. Test structures with aspect ratio 20:1 were obtained.

D 2005 Elsevier B.V. All rights reserved.

Keywords: Chemically amplified resist SU-8

1. Introduction

Various lithographic studies on SU-8 have been performed

using a variety of radiation sources—ultraviolet [1–3], X-rays

[4–7], proton beam [8] and electron beam [9–11]. In

particular, the short wavelength and high penetration power

of X-rays allow microstructures with submicron resolution and

high aspect ratio to be produced. A previous study [12] on the

characterization of SU-8 for X-ray Lithography (XRL) using

the dense plasma focus (DPF) device was made on resist

thickness of 3–15 Am. In this investigation, the processing

parameters of the 25-Am-thick SU-8 resist for X-ray litho-

graphic applications are derived.

2. Experimental details

For our study, we used the SU-8 2010 formulation with a

viscosity of 1050 cst (at 25 -C), obtained commercially from

MicroChem Corporation. The SU-8 resist consists of a multi-

functional, highly branched polymeric epoxy resin dissolved in

an organic solvent, gamma-butyrolactone (GBL). Along with

the formulation is a triaryl sulfonium salt which acts as a

photoacid generator. The resist has on average 8 epoxy ring

0040-6090/$ - see front matter D 2005 Elsevier B.V. All rights reserved.

doi:10.1016/j.tsf.2005.09.151

* Corresponding author. Tel.: +65 67903837; fax: +65 68969414.

E-mail address: tlatan@nie.edu.sg (T.L. Tan).

groups in each molecule; hence the name FSU-8_. Its chemical

structure is shown in Fig. 1.

2.1. Thickness measurements of resist layer

The SU-8 resist was spin-coated at various speeds of 500–

2500 rpm. Care was taken to remove any excess resist from the

edge and bottom of the substrate. The resist film was then sub-

jected to a 15-min soft-bake at 95 -C on a level hot plate in order

to remove the solvent from the resist. It was found that the spin

speed of 500 rpm gave about 25-Am-thick samples. The thick-

ness of the resist was measured using the oscillating patterns

along the baseline of the FT-IR spectrum [12], as shown from

Fig. 2.

2.2. Resist exposure

The 25 Am resist samples were irradiated by soft X-rays

(SXRs) emitted from a DPF device. Such a device operates with

a pulsed capacitive discharge, where a dense, magnetically

compressed plasma is produced at the end of the two coaxial

electrodes. This is followed by the decay of the plasma column,

giving rise to X-rays, amongst a host of other radiation. The

details of the set-up of the DPF device are found elsewhere [13].

A 10-Am Be filter was used to shield off visible and ultra-

violet light from irradiating the samples. Based on the X-ray

2006) 113 – 116

ww

CH3

CH3

O

O

O

O

CH3

O

O

O

O

CH3

H3C H3C

O

O

O

O

CH3

CH3

O

O

O

O

Fig. 1. Chemical structure of SU-8.

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.8 1.0 1.2 1.4 1.6Energy (keV)

Tra

nsm

issi

on

10µm15µm25µm

Fig. 3. Estimated X-ray transmission curves of SU-8 (density=1.2 g/cm3) for

resist thickness of 10, 15 and 25 Am.

T.L. Tan et al. / Thin Solid Films 504 (2006) 113–116114

spectral distribution given off by the DPF device, the average

weighted energy is found to be about 1.0 keV. However,

because of the coupled effect of the Be filter and Ne gas

transmission characteristics, the average weighted energy

falling on the resist is actually slightly higher.

Previous experiments [12] showed that for a resist of

thickness 10 Am, a dosage of about 250 mJ/cm2 was sufficient

for structures to be formed, although with slightly incomplete

cross-linking at the bottom surface. The estimated X-ray

transmission curves of SU-8 (density=1.2 g/cm3) for resist

thickness of 10, 15 and 25 Am are shown in Fig. 3. For X-ray

energy of 1.1 keV, the transmission after passing through 10

Am of SU-8 resist is about 0.1. Hence we estimate that to

maintain the same minimum exposure dose at the bottom

surface for a 25-Am resist, we need about ten-fold an amount of

X-rays to be irradiated at the top surface (2500 mJ/cm2).

3. Results and discussion

3.1. Degree of cross-linking

The FT-IR spectrum for pre-exposed SU-8 is given in Fig. 4.

The infrared absorption peaks of the epoxide ring modes at 914

and 862 cm� 1 were observed to gradually reduce in intensity

4000 3500 3000

Wavenumber (cm-1)

Abs

orba

nce

22322717

20002500

Fig. 2. FT-IR spectra of SU-8 resist at spin speed of 500 rpm.

with increased exposure dose after PEB. Conversely, intensity

of the 1128 cm� 1 absorption peak associated with the ether

bond gradually increased. The epoxide peaks at 914 and 862

cm� 1 were selected for quantitative determination of the

degree of cross-linking.

In an effort to eliminate errors due to uneven thickness, an

infrared peak, which does not participate in the cross-linking

process is normally used as an internal standard, and all other

peaks are normalized against this peak. The aromatic ring C–C

stretch mode at 1608 cm� 1 was selected for this purpose. The

intensity of the peak at 1608 cm� 1 is comparable to that at 914

cm� 1 and hence the value of the normalized intensity given as

the ratio (I914/I1608) gives the most reliable measure for the

degree of cross-linking.

3.2. Post-exposure bake

The samples were subjected to varying PEB temperatures of

65–120 -C in order to initiate the cross-linking process. After

PEB, FT-IR spectra of the resist samples were recorded. To

prevent resist cracking on large exposed areas, we applied an

intermediate bake at 65 -C for 5 min for each sample. This was

followed by ramping at the rate of about 20 -C/min and with

the samples subjected to the desired temperature for another 5

9141128

8621608

1100 900 700

Wavenumber (cm-1)

Abs

orba

nce

130015001700 500

Fig. 4. FT-IR spectra of non-cross-linked SU-8.

0.30

0.35

0.40

0.45

0.50

0.55

60 70 80 90 100 110 120 130

PEB Temperature (°C)

I 914/

I 1608

Fig. 5. The dependence of the degree of cross-linking on PEB temperature for

25 Am SU-8 resist.

Fig. 7. SEM micrograph showing a close-up view of test structures with an

aspect ratio of up to 20:1 on 25-Am-thick SU-8 resist.

T.L. Tan et al. / Thin Solid Films 504 (2006) 113–116 115

min. The samples were then left on the hotplate to allow

gradual cooling (at a rate of 2 -C/min) to room temperature.

In the experiments, the SXR exposure dose was fixed at

2500 mJ/cm2 while varying the final PEB temperature from 65

to 120 -C. The dependence of the degree of cross-linking on

the PEB temperature for 25 Am SU-8 resist is shown in Fig. 5.

As expected, the degree of cross-linking increases with

increasing PEB temperature. This happens because a higher

temperature increases the diffusion rate of the photogenerated

acid responsible for catalyzing the cross-linking reaction.

3.3. X-ray lithographic experiments

The resist samples were irradiated with X-rays, with a gold

mesh as a test mask, via contact printing, at an exposure dose

of 2500 mJ/cm2. The gold mesh consists of wires of 5-Amdiameter, forming square grids with 11 Am width and the SEM

micrograph of it is shown in Fig. 6a.

The set of reliable experimental conditions comprised an

exposure dose of 2500 mJ/cm2, with the final PEB temperature

set at 95 -C. An intermediate bake at 65 -C was applied for 5

min, followed by another 1.5 min for ramping the temperature to

95 -C. The sample was left at 95 -C for 5 min and then left to

cool on the hotplate gradually to room temperature. The sample

was developed by dipping it into cyclohexanon for 5 min

a b

Fig. 6. SEM micrographs of (a) gold mesh, (b) SU-8

without agitation and rinsed with fresh developer. The SEM

micrograph is shown in Fig. 6b. Square structures of 11 Amwidth were produced, giving a direct imprint of the gold mesh

patterns, without affecting the resolution. Our results show that

for successful imprint of X-ray lithographic structures on 25 AmSU-8 resist, a dose of 2500 mJ/cm2 from our DPF device is

sufficient with the processing conditions described above.

In the final phase of the X-ray lithographic work, the SU-

8 resist was processed using the above experimental condi-

tions, using a specially designed mask. The resist was then

developed using cyclohexanon at room temperature. The resist

layer was found to be relatively free from air-borne particles.

The SEM micrograph of the 25 Am patterned SU-8 resist in

Fig. 7 shows the mask pattern accurately transferred to the SU-

8 resist layer. The test structure shown in Fig. 7 possesses clear

and sharp edges implying complete cross-linking under the

derived lithographic conditions, with excellent aspect ratio of

20:1 in a 25-Am thick resist.

4. Conclusion

The SU-8 resist, commercially used in DUV lithography, is

now characterized for X-ray lithographic applications. A spin

resist test structures at optimized PEB conditions.

T.L. Tan et al. / Thin Solid Films 504 (2006) 113–116116

coater speed of 500 rpm gave about 25-Am-thick resist

samples, measured by FT-IR spectroscopy. The samples were

soft-baked for 15 min. Our results showed that an X-ray

exposure dose of 2500 mJ/cm2 from a neon-filled DPF device

is sufficient for complete cross-linking of the 25 Am resist.

Reliable processing conditions consisted of an intermediate

PEB at 65 -C for 5 min, with the PEB temperature ramped up

to 95 -C over 1.5 min and then followed by a final PEB at 95

-C for 5 min. These experimental conditions for the X-ray

lithographic processing of SU-8 successfully produced resist

test structures with an aspect ratio of up to 20:1. FT-IR

spectroscopy was used to accurately monitor the extent of

cross-linking in SU-8 by measuring the normalized peak

intensity of the absorption peaks at 914 and 1126 cm� 1.

Further studies can be made using thicker SU-8 resist layer

which has vast applications in MEMS devices. The present 25

Am resist thickness of SU-8 can be further increased using SU-

8 of higher viscosities, higher X-ray irradiation power, and

shorter X-ray wavelength.

Acknowledgments

We gratefully acknowledge Nanyang Technological Uni-

versity, Singapore for the research grant in financial support of

our project.

References

[1] J. Zhang, K.L. Tan, G.D. Hong, L.J. Yang, H.Q. Gong, J. Micromechanics

Microengineering 11 (2001) 20.

[2] P.M. Dentinger, K.L. Krafcik, K.L. Simison, R.P. Janek, J. Hachman,

Microelectron. Eng. 61–62 (2002) 1001.

[3] J.D. Williams, W. Wang, Soc. Photo-Opt. Instrum. Eng. 3 (2004) 563.

[4] A.L. Bogdanov, S.S. Peredkov, Microelectron. Eng. 53 (2000) 493.

[5] C. Cremers, F. Bouamrane, L. Singleton, R. Schenk, Microsyst. Technol.

7 (2001) 11.

[6] C. Khan Malek, Microelectron. J. 33 (2002) 101.

[7] B.-Y. Shew, J.-T. Hung, T.-Y. Huang, K.-P. Liu, C.-P. Chou, J.

Micromechanics Microengineering 13 (2003) 708.

[8] A.A. Bettiol, I. Rajta, E.J. Teo, J.A. van Kan, F. Watt, Nucl. Instrum.

Methods Phys. Res., B Beam Interact. Mater. Atoms 158 (1999) 179.

[9] A.K. Nallani, S.W. Park, J.B. Lee, Micro Nano Devices and Systems

Laboratory, University of Texas, Dallas, USA, 2002.

[10] V. Kudryashov, X.-C. Yuan, W.-C. Cheonga, K. Radhakrishnana,

Microelectron. Eng. 67–68 (2003) 306.

[11] A. Pepin, V. Studer, D. Decanini, Y. Chen, Microelectron. Eng. 73–74

(2004) 233.

[12] T.L. Tan, D. Wong, P. Lee, A. Patran, Appl. Spectrosc. 58 (2004) 1288.

[13] D. Wong, A. Patran, T.L. Tan, R.S. Rawat, P. Lee, IEEE Trans. Plasma

Sci. 32 (2004) 2227.