Chemistry Letters Vol.32, No.6 (2003)
555
Figure 4. SEM images of the two resists with TPSTf (resist E) and
ꢂ2
CpDPSTf (resist F) with a each thickness of 0.4 mm at a dose of 10 mJ cm
.
to the bottom of the resist, and the exposed region be dissolved
clearly. As a result, the photoresist using CpDPSTf showed better
performance than the resist using TPSTf in the finer pattern.
In conclusion, four PAGs containing a cyclopropyl group
have been developed. These novel PAGs show less absorbance
at 193 nm than the conventional PAG, TPSTf. Moreover, the cy-
clopropyl-containing PAGs show the photobleaching effect, while
TPSTf shows the antiphotobleaching effect. From the GC–mass
experiment of CpDPSTf, cleavage of S–C(cyclopropyl) is much
more preferential than that of S–C(phenyl), and this cleavage re-
sults in the photobleaching effect. The resist containing CpDPSTf
gave better resolution than that containing TPSTf which has been
widely used.
Scheme 1. Plausible decomposition mechanism of CpDPSTf upon exposure.
sulfide (1), 34.6% of diphenyl sulfide (2), 37.8% of allylphe-
nylthiobenzene (3), 15.1% of cyclopropylthiophenylbenzene (4),
and 3.0% of the shoulder peak (5), which is another isomer of cy-
clopropylthiophenylbenzene. There are also very minute amounts
of other products. The first two peaks 1 and 2 came from direct
photolysis and the latter three peaks 3, 4 and 5 came from the re-
combination reactions.
The plausible decomposition mechanism of CpDPSTf is de-
picted in Scheme 1. When the PAG is irradiated, the homolytic
cleavage of the sulfur–carbon bond leads to the formation of a cat-
ion-radical on the sulfur–containing fragment and a neutral radi-
cal. And the homolysis must be strongly affected by the stability
of both radicals. The diphenylsulfonium cation radical can be
much more stabilized than the cyclopropylphenylsulfonium cat-
ion-radical because the former has more resonance forms. The cy-
clopropyl radical is known to be converted very fast to more
The authors would like to acknowledge the financial support
of Ministry of Commerce, Industry & Energy, the Center for Ad-
vanced Functional Polymers at Korea Advanced Institute of
Science and Technology, and Brain Korea 21 (BK21) project.
9
stable allyl radical and the allyl radical is certain to be more
References and Notes
stable than phenyl radical. From these two reasons, the path 1 is
expected to be predominant over the path 2. The mechanism is
consistent with the results of the GC–mass experiment.
1
2
3
I. Hiroyuki, U. Shinji, D. Katsuji, K. Toru, C. Hitoshi, and S. Tsutomu, SPIE,
3999, 1120 (2000).
K. Nakano, K. Maeda, S. Iwasa, T. Ohfuji, and E. Hasegawa, SPIE, 2348, 433
(1995).
1
0
Two resist solutions were formulated by dissolving the
polymer (Figure 3, 15 wt% polymer in propylene glycol mono-
methyl ether acetate) and TPSTf or CpDPSTf (1 wt% relative to
the resin), and then the solutions were filtered with a 0.2 mm teflon
membrane filter. Figure 4 shows the comparison between resist E,
the resist with the conventional PAG, and resist F, the resist with
the photobleachable PAG. There are two distinguishing features.
First, the side-wall shape of resist F is relatively rectangular while
that of resist E is slightly sloped. Generally, the resists containing
TPSTf show a somewhat tapered pattern because of the strong ab-
sorbance. Second, at the pattern of 120 nm, comparing with the
shape of resist E, there is a notable improvement in the shape
of resist F. In the case of resist E, even though we increased the
dose of exposure and developed for a longer time, the scum did
not disappear. The main reason of these differences between the
two resists is related to the bleaching effect of CpDPSTf. The
strong absorbance of TPSTf prevents the light from reaching the
bottom of the resist, leading to scum at the resolution limit. But
the photobleaching effect of CpDPSTf makes the light go through
Z. Wenhui, M. K. Stephen, C. Dave, W. P. Joshep, and R. M. Seth, J. Am.
Chem. Soc., 124, 1897 (2002).
T. Shono and Y. Matsumuro, J. Org. Chem., 94, 7892 (1972).
4
5
Alkylcyclopropylphenylsulfonium triflate was obtained in a yield of 67–75%.
ChCpPSTf: 1H NMR (CDCl3, ppm): d 0.8–1.0 (4H, m), 1.0–1.3 (6H, m),
1.7–2.0 (5H, m), 3.9 (2H, m), 7.7–8.0 (5H, m); F NMR: singlet. CpPPSTf:
19
1
H
NMR (CDCl3, ppm):
.75(3H, m), 3.75 (1H, m), 3.85 (2H, m), 7.7–8.0 (5H, m); F NMR: singlet.
d 1.0–1.2(4H, m), 1.25(3H, m), 1.6(1H, m),
19
1
1
CpNPSTf: H NMR (CDCl3, ppm): d 0.9(1H, m), 1.1(1H, m), 1.4(2H, m), 3.75
19
(1H, m), 5.2(1H, d), 5.4(1H, d), 7.4–8.0 (10H, m); F NMR: singlet.
F. G. Kathawala, U. S. Patent 4233292 (1980). CpDPSTf was obtained in a
6
7
yield of 82.5% as a white crystal. 1H NMR (CDCl , ppm): d 1.4 (2H, m),
3
1
.65 (2H, m), 3.95 (2H, m), 7.7–8.0 (10H, m);19F NMR: singlet.
UV spectra were recorded as spin-coated films on quartz plates with a Hewlett-
Packard Model 8453 spectrophotometer. Four resist solutions were made by dis-
solving 0.2 g of poly(tert-butyl methacrylate) (PTBMA) and 5 wt% of a PAG in
cyclohexaneone. These solutions were filtered through a 0.2 mm filter and spin-
coated onto quartz plates. For each sample, the UV absorbance was observed
before and after exposure using a mercury-xenone lamp. Absorbance data were
normalized to one micron and obtained by subtracting the absorbance of the
PTBMA film from the observed data.
8
Steady-state irradiation was performed in a Rayonet Photochemical Reactor
equipped with RPR 254-nm lamps from the Southern New England Ultraviolet
Co. Surpasil quartz cells (10 mm ꢃ 10 mm) were used in these experiments.
Each 10 mM solution of TPSTf and CpDSPTf in 3 mL of degassed acetonitrile
was irradiated with 254-nm light for 3 min. The irradiated solutions were diluted
with acetonitrile and analyzed by GC–mass. Gas chromatography was taken
with HP6890 plus, and mass analysis was taken with JEOL SX102A.
S. Olivella, A. Sole, and J. M. Bofill, J. Am. Chem. Soc., 112, 2160 (1990).
The resist solutions were coated to yield 0.4 mm-thick films on silicon wafers
9
1
0
ꢁ
and soft-baked at 120 C for 90 s. Exposure was carried out with an ArF stepper
(ISI, NA = 0.6) using a conventional illumination method. Exposed resists were
baked at 110 C for 90 s on a hot-plate and developed in a 2.38 wt%
tetramethylammonium hydroxide aqueous solution for 60 s.
ꢁ
Figure 3. SAS-4.
Published on the web (Advance View) May 27, 2003; DOI 10.1246/cl.2003.554