Vacuum-UV influenced design of polymers and dissolution inhibitors for next generation photolithography

Article abstract of DOI:10.1016/S0022-1139(03)00076-9

An overview of our 157 nm photoresist development activities is presented. Examination of the vacuum ultraviolet (VUV) absorbances of fluorinated monomers and polymers has provided knowledge that influenced copolymer design so that resist transparency in the vacuum-UV can be maximized. Partially fluorinated norbornenes and tricyclononenes (TCNs) have been incorporated into copolymers using metal-catalyzed addition and radical initiators. These materials have orders of magnitude higher transparency at 157 nm compared to their hydrocarbon analogues as measured by variable angle spectroscopic ellipsometry (VASE). We have also synthesized fluorinated dissolution inhibitors for use in a three-component resist system. The results of preliminary lithographic evaluations of resists formulated from these polymers are presented.

Full text of DOI:10.1016/S0022-1139(03)00076-9

Journal of Fluorine Chemistry 122 (2003) 17–26
Vacuum-UV influenced design of polymers and dissolution
inhibitors for next generation photolithography
Brian C. Trinquea, Charles R. Chambersa, Brian P. Osborna, Ryan P. Callahana,
Geun Su Leea, Shiro Kusumotoa, Daniel P. Sandersb, Robert H. Grubbsb,
Willard E. Conleyc, C. Grant Willsona,*
aDepartment of Chemistry and Biochemistry, University of Texas, Austin, TX 78712, USA
bArnold and Mabel Beckman Laboratories of Chemical Synthesis, Division of Chemistry and Chemical Engineering,
California Institute of Technology, Pasadena, CA 91125, USA
cInternational SEMATECH, 2706 Montopolis Drive, Austin, TX 78741, USA
Abstract
An overview of our 157 nm photoresist development activities is presented. Examination of the vacuum ultraviolet (VUV) absorbances of
fluorinated monomers and polymers has provided knowledge that influenced copolymer design so that resist transparency in the vacuum-UV
can be maximized. Partially fluorinated norbornenes and tricyclononenes (TCNs) have been incorporated into copolymers using metal-
catalyzed addition and radical initiators. These materials have orders of magnitude higher transparency at 157 nm compared to their
hydrocarbon analogues as measured by variable angle spectroscopic ellipsometry (VASE). We have also synthesized fluorinated dissolution
inhibitors for use in a three-component resist system. The results of preliminary lithographic evaluations of resists formulated from these
polymers are presented.
# 2003 Elsevier Science B.V. All rights reserved.
Keywords: 157 nm lithography; 157 nm photoresist; Fluoronorbornenes; Metal-catalyzed addition polymerization; Tricyclononene; 2-Trifluoromethyl-
acrylate; Dissolution inhibitor
1. Introduction
If the photoresist is too absorbing at the exposure wavelength,
only the top portion of the exposed area will experience the
crucial solubility change, resulting in a poor resist profile.
Therefore, the demand for smaller features requires the
development of a highly transparent resist for each new
wavelength reduction. The most common wavelengths used
in current semiconductor lithography are I-line (365 nm) and
DUV (248 nm). Resist technology for 193 nm has been
extensively explored [1–3], and it is likely that this exposure
wavelength will be utilized in the near future in order to keep
pace with Moore’s Law.
As device geometries shrink below 100 nm, an enormous
amount of effort has been placed on the development of
chemically amplified resists for 157 nm lithography. Creat-
ing such a material is a significant task since water, oxygen,
and simple hydrocarbons absorb very strongly in this region.
Major progress has been made in decreasing the absorbance
of carbon-based copolymers by partial fluorination, but this
process may limit at the transparency of perfluorocarbon
polymers [4]. Our group has focused our polymer synthesis
efforts on learning how to maximize transparency at 157 nm.
Vacuum ultraviolet (VUV) gas phase studies of model
As semiconductor device dimensions continue to decrease,
new constraints are continually placed on photolithography,
which is the core process for the volume production of
microchips. Underlying this process is the Rayleigh equation,
stating:
l
F /
NA
which relates the feature size (F) to the wavelength of the
exposure light (l) and numerical aperture (NA): the sine of
half of the angle of the image-forming cone of light at the
image. The majority of the advances in photolithography
technology can be attributed to decreases in the exposure
wavelength. Each new generation of exposure tools intro-
duces wavelength reductions that limit the materials available
for design. This limitation results from the increasing
opacity of conventional resist materials at lower wavelengths.
* Corresponding author. Tel.: þ1-512-471-4342; fax: þ1-512-471-7222.
E-mail address: willson@che.utexas.edu (C.G. Willson).
0022-1139/03/$ – see front matter # 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0022-1139(03)00076-9

18
B.C. Trinque et al. / Journal of Fluorine Chemistry 122 (2003) 17–26
compounds has proved to be extremely valuable. Norbor-
nene monomers were hydrogenated and the relative absor-
bance of these monomers in the gas phase has, in every
instance, correlated with the absorbance of the correspond-
ing polymers as measured by variable angle spectroscopic
ellipsometry (VASE). We have extensively studied the ima-
ging capabilities of metal-catalyzed addition polymers made
from a variety of fluorinated monomers, and have studied
how altering protective groups can improve the transparency
of these materials. Radical polymerizations have also been
studied. This approach has produced a variety of transparent
copolymers that have successfully produced positive tone
structures with 157 nm exposure. We discovered that the
dissolution rate of the new, acidic polymers can be drasti-
cally altered by addition of small amounts of appropriate
monomeric species when functioning as dissolution inhibi-
tors. Therefore, we are studying a three-component resist
system that utilizes photolabile dissolution inhibitors to
induce the solubility switching of the resist. A variety of
dissolution inhibitors have been synthesized, and their
effects on polymer dissolution rate and 157 nm imaging
have been examined.
absorbance at 157 nm. Using a variety of synthetic path-
ways, we have prepared a number of mono- and disubsti-
tuted fluoronorbornenes (1–4). This study has demonstrated
that absorbance in this region relies not only on the number
of fluorines substituted on the norbornene ring but also, and
perhaps more importantly, on the position of the substitution.
As seen in Fig. 1, substituting two hydrogens for two
fluorines leads to a higher transparency at 157 nm than
substitution with only one fluorine. Early in this study, we
discovered the surprising lability of the monofluoro deriva-
tives to ‘‘glass catalyzed’’ elimination [5]. Vicinal substitu-
tion was therefore less attractive from a practical perspective
than the more stable, geminally disubstituted isomers. The
VUV spectra show that introduction of geminal fluorosub-
stituents on the two-carbon bridge (4), rather than the seven-
carbon bridge (3) is the most effective substitution pattern.
This led to the conclusion that transparency at 157 nm can be
maximized by synthesizing monomers with geminal elec-
tron-withdrawing groups such as fluorine or trifluoromethyl
at the two-carbon bridge.
These gas phase data led to the synthesis of the highly
fluorinated norbornene 5, which boasts both geminally
substituted fluorines and geminally substituted trifluoro-
methyl and hydroxyl groups. Because of this advantageous
substitution pattern, we anticipated that this molecule would
be even more transparent than 4. The gas phase data in Fig. 2
shows that this is indeed the case. Compound 5 has obvious
advantages over 1–4 from a resist development standpoint
as the hydroxyl group is an important synthetic handle for
further synthesis (i.e. addition of a protecting group for
chemical amplification). We are currently investigating the
incorporation of 5 into a copolymer.
2. Results and discussion
2.1. Absorbances of fluorinated norbornenes
To gain a better understanding of the effect that fluorina-
tion exerts on the absorbance of the norbornene skeleton,
a systematic replacement of hydrogen atoms with fluorine
was proposed in order to measure the effect upon their
Fig. 1. Vacuum-UV absorbances of fluorinated norbornenes.

B.C. Trinque et al. / Journal of Fluorine Chemistry 122 (2003) 17–26
19
Fig. 2. Vacuum-UV absorbance of compound 5.
2.2. Synthesis and imaging of addition copolymers
This copolymer was made from poly(3-(bicyclo[2.2.1]hept-
5-en-2-yl)-1,1,1-trifluoro-2-(trifluoromethyl)propan-2-ol
(PNBHFA, 9), which was partially t-BOC protected. This
platform has a higher fluorine incorporation, thereby low-
ering its absorbance to 1.9 mmÀ1 at 157 nm. Preliminary
157 nm imaging of 8 is shown in Fig. 4.
Preliminary 157 nm imaging studies on formulations
derived from copolymer 6 using the trifluoromethyl-contain-
ing carbon monoxide copolymer 7 as a dissolution inhibitor
have been previously reported [6]. This resist system has
geminal trifluoromethyl groups, which provides an absor-
bance of 2.3 mmÀ1 at 157 nm. This is a better transparency
than polynorbornene by at least three orders of magnitude.
We have continued to improve the quality of the images
obtained from formulations prepared from these materials.
Representative images are shown in Fig. 3.
A series of imaging studies were also carried out
on formulations based on the vinyl addition copolymer 8.
Fig. 3. SEMs of vinyl addition copolymer 6 with CO copolymer 7 (DI).

20
B.C. Trinque et al. / Journal of Fluorine Chemistry 122 (2003) 17–26
Fig. 4. SEMs of vinyl addition copolymer 8.
The aforementioned vinyl addition copolymers each have
measured. We have previously demonstrated the litho-
graphic performance of acetal-protected analogs of 12 for
157 nm applications [7]. While this new material had only a
slight decrease in absorbance at 157 nm when compared to
its t-BOC analog (1.78 mmÀ1 versus 1.9 mmÀ1, respec-
tively), we believe that with improved purification of this
material the absorbance can be significantly decreased.
Preliminary imaging studies of this new material are shown
in Fig. 7. The contrast of 12 is improved when blended with
copolymer 7.
their advantages: copolymer 8 has a higher transparency at
157 nm. However, copolymer 6 exhibits superior contrast
due to deprotection of a carboxylic acid moiety, which is
considerably more soluble in aqueous base developer com-
pared to the hexafluoroalcohol moiety. In order to take
advantage of the strengths of each of these copolymers,
they were combined into one resist and imaged at 157 nm.
The resist based on the blended material had an absorbance
of 2.08 mmÀ1 at 157 nm, and imaging efforts have produced
structures of the sort shown in Fig. 5. While we are still
optimizing formulations and imaging conditions, we believe
a blend of these materials will have a far superior litho-
graphic performance compared to the performance of each
of the individual materials.
Though the lithographic performance of these copolymers
shows promise, there are still improvements to be made in
polymer transparency. The increase in absorbance observed
in copolymers 6 and 8 when compared to PNBHFA was
attributed to the carbonyl functionality in protecting groups.
In order to investigate this theory, model compounds 10 and
11 were synthesized and their gas phase spectra acquired. As
expected, the acetal-protected alcohol 10 had a significantly
higher transparency at 157 nm than the carbonate-protected
11 (Fig. 6). Due to this result, it was envisaged that incor-
poration of a protecting group that contained no carbonyl
group would offer a norbornene-based addition copolymer
with an even lower transparency at 157 nm. The acetal-
protected copolymer 12 was synthesized and its absorbance
2.3. Tricyclononene (TCN) copolymers
The incorporation of geminally CF3-substituted norbor-
nenes into a polymer using the metal addition catalysts that
made polymers 6, 8, and 12 is difficult. The reason for this
may stem from the reduced electron density of the double
bond resulting from the close proximity of the electron-
withdrawing groups, from steric inhibition, or both. In order
to incorporate these monomers into a polymer system, we
have moved the trifluoromethyl group further away from the
double bond by synthesizing the corresponding tricyclono-
nene (TCN) analogues. TCN compounds behave much like
norbornenes in polymerizations. The fused cyclobutane ring
allows the electron-withdrawing trifluoromethyl and/or
fluorine substituents to be positioned further away from
the double bond, reducing the steric and electronic effects
these groups may have. In a previous report, our group
demonstrated that a model TCN polymer that contained a
Fig. 5. SEMs of blend of vinyl addition copolymers 6 and 8.

B.C. Trinque et al. / Journal of Fluorine Chemistry 122 (2003) 17–26
21
Fig. 6. Vacuum-UV absorbances of compounds 10 and 11.
Fig. 7. SEMs of vinyl addition copolymer 12.
Fig. 8. SEMs of vinyl addition copolymer 14.

22
B.C. Trinque et al. / Journal of Fluorine Chemistry 122 (2003) 17–26
geminal trifluoromethyl group could be synthesized with a
Pd(II) addition catalyst and was relatively transparent at
157 nm. Monomer 13 was synthesized, and copolymerized
with bicyclo[2.2.1]hept-5-ene-2-(1,1,1-trifluoro-2-trifluoro-
methylpropane-2-ol (NBHFA) to produce copolymer 14,
which has a lower absorbance than 6 at 157 nm (2 mmÀ1
versus 2.3 mmÀ1, respectively). This result is evidence of the
significant impact that a geminally-substituted trifluoro-
methyl group can have on polymer transparency, even when
two extra carbons are required for the substitution. The
results of preliminary imaging work utilizing resist formula-
tions that contain 14 and the carbon monoxide copolymer 7
as a dissolution inhibitor is shown in Fig. 8. We have
previously reported [6], a TCN molecule with additional
difluoro-substitution on the cyclobutane ring, showing a
further decrease in absorbance of a TCN monomer. That
result has inspired the synthesis of monomer 15, and efforts
are now underway to incorporate this more transparent and
lithographically useful molecule into a copolymer.
2.4. 2-Trifluoromethylacrylate platform
Our group has also explored 2-trifluoromethylacrylate/
norbornene copolymers as a way to incorporate highly fluori-
nated norbornenes in resist materials. Most recently, we have
successfully produced 16, which is a copolymer of the t-BOC
protected hexafluoroalcohol moiety (17) and a highly fluori-
nated2-trifluoromethylacrylate(18). Monomer18containsan
ester that compliments published efforts by Vohra et al. [8].
Interestingly, this radical copolymer deviated from the 2:1 (2-
trifluoromethylacrylate/norbornene) copolymerization incor-
porationratiofirstreportedbyItoetal.[9,10].Thismaterialhad
an extremely high transparency at 157 nm (1.3 mmÀ1), how-
ever, attempted imaging at 157 nm revealed that deprotection
of the acid labile moiety did not render the copolymer soluble
in 0.26N tetramethylammonium hydroxide (TMAH) devel-
oper. Knowing that the hexafluoroalcohol moiety (pKa ꢀ 11)
hasprovidedsufficientsolubilityinaqueousbasedeveloperfor
resists 6, 8, and 12, we wondered if the aliphatic ester on 18
resulted in a high degree of hydrophobicity in copolymer 16.
In order to test this, we anionically polymerized 18 to yield
the homopolymer 19. We examined this material’s dissolution
rate and found that when a 600 nm thick film of this material
was coated on a silicon wafer and placed in 0.26N TMAH
developer there was no change in thickness over time (Fig. 9).
Fig. 9. Dissolution rate monitor (DRM) measurements of 9 versus 19 in 0.26N TMAH developer.

B.C. Trinque et al. / Journal of Fluorine Chemistry 122 (2003) 17–26
23
Fig. 10. SEMs of radical copolymer 21.
Although 18 was found to be insoluble in standard
developer, we recognized its importance from a transpar-
ency standpoint. We radically copolymerized 18 with 20, a
norbornene molecule with a geminally-substituted trifluor-
omethyl group and an acetal-protected carboxylic acid. We
hoped that upon deprotection, the highly acidic carboxylic
acid would provide sufficient dissolution so that when a
formulated 21 was exposed it would be base soluble.
Copolymer 21 had the expected 2:1 (2-trifluoromethylacry-
late/norbornene) monomer incorporation [9,10], and had an
absorbance of 2.2 mmÀ1 at 157 nm. This material did pro-
vide positive-tone structures with 157 nm exposure (Fig. 10).
when loaded with varying percentages of 22, and 1 wt.%
triphenylsulfonium nonaflate (TPS-Nf) the difference in
dissolution rate between exposed and unexposed PNBHFA
is significant. Much to our surprise, 22 had a greater
inhibition effect on PNBHFA than it did on novolac, a
traditional I-line resist polymer [11] (Tables 1 and 2). This
Table 1
Exposed and unexposed dissolution rates of PNBHFA with varying
loadings of 22 (1 wt.% TPS-Nf)
wt.% 22 in
PNBHFA
Unexposed dissolution
rate (nm/s)
Exposed dissolution
rate (nm/s)
0
1
3
5
71
32
8
161
195
190
196
2.5. Dissolution inhibitors for 157 nm applications
As an alternative to traditional chemically amplified
resists, our group has also been exploring a resist platform
based on the addition of various dissolution inhibitors to
PNBHFA. This approach poses the advantage of using only
an intrinsically soluble homopolymer as the base resin and
thereby avoids the synthetic inconvenience of having to tune
the extent of co-monomer incorporation into a copolymer.
The viability of such a three-component system (homopo-
lymer, dissolution inhibitor and photoacid generator) would
create a greater ease of resist formulation and thereby speed
resist development at 157. We first examined the effect that
bis-BOC–bis-phenol A (22) had on PNBHFA. We found that
4
Table 2
Exposed and unexposed dissolution rates of novolac with varying loadings
of 22 (1 wt.% TPS-Nf)
wt.% 22 in
novolac
Unexposed dissolution
rate (nm/s)
Exposed dissolution
rate (nm/s)
0
1
3
5
23
17
12
9
26
24
20
17

24
B.C. Trinque et al. / Journal of Fluorine Chemistry 122 (2003) 17–26
Fig. 11. SEMs of polymer 9 with 10 wt.% 22.
Table 3
Exposed and unexposed dissolution rates of PNBHFA with varying
inspired preliminary 248 nm imaging experiments of this
three-component resist system, which produced high con-
trast positive tone structures (Fig. 11).
loadings of 24 (1 wt.% TPS-Nf)
wt.% 24 in
PNBHFA
Unexposed dissolution
rate (nm/s)
Exposed dissolution
rate (nm/s)
0
71
6
161
151
150
135
0.8
2
3
3
2
we are also exploring the use of fluorinated dissolution
inhibitors with increased etch resistance. For example, the
tert-butyl ester protected molecule 24 contains norbornene
rings that have demonstrated high etch resistance. The effect
of this material on the dissolution rate of exposed and
unexposed films of PNBHFA was studied, and the results
(Table 3) indicate that it too could serve as a dissolution
inhibitor for 157 nm lithography. Imaging experiments with
this material are currently underway.
In keeping with transparency concerns with 157 nm resist
technology, we are also investigating fluorinated materials as
dissolution inhibitors in a three-component resist system. A
number of cyclic alkanes that contain a protected hexafluor-
oalcohol moiety such as 23 have been synthesized. Like 22,
23 also had a significant impact on the dissolution rate of
PNBHFA, decreasing its dissolution rate from 165 to <5 nm/s
with 10 wt.% addition of the dissolution inhibitor. When 23
was formulated with PNBHFA and photoacid generator and
imaged at 157 nm, it produced the high-contrast positive-
tone structures shown in Fig. 12.
While 23 is highly fluorinated and an effective dissolution
inhibitor for 157 nm lithography, its relatively low carbon to
hydrogen ratio could conceivably diminish the etch resis-
tance of the base resin in the formulation. In light of this,
3. Conclusions and future outlook
While reported fluorinated resist materials do not cur-
rently match the industry standard absorbances at 157 nm,
great strides have been made toward improving transpar-
ency. The fluoropolymers designed in our labs and others are
orders of magnitude more transparent than their hydrocar-
bon analogues. Considering that the 157 nm lithographic
process will utilize thinner resist films to allow for more
absorbing materials, this technology is close to meeting the
absorbance requirements for 250 and 300 nm thick films
(Table 4). Recently, Kodama et al. [12] have designed an
acidic polymer with a still higher fluorine content that has a
Fig. 12. SEMs of PNBHFA with 23.

B.C. Trinque et al. / Journal of Fluorine Chemistry 122 (2003) 17–26
25
Table 4
gas cell attachment. The details of the cell design and
implementation are the subject of a separate paper [13].
VUV spectra of polymer films were calculated from mea-
surements made with a J.A. Woollam VU301 variable
angle spectroscopic ellipsometer. The films were cast on
bare silica from solutions in propylene glycol monomethyl
ether acetate (PGMEA) and baked at 110 8C for at least
90 s prior to analysis. All absorbance data reported are in
base 10.
Absorbance requirements for varying resist thicknesses for 157 nm
photoreists
Exposure
wavelength (nm)
Absorbance
(mmÀ1
Thickness
(nm)
Absorbance at this
thickness (mmÀ1
)
)
248
193
157
157
0.37
0.47
0.62
0.74
500
400
300
250
0.185
0.185
0.185
0.185
4.3. Synthesis
0.6 mmÀ1 absorbance at 157 nm, which has set a new
standard in this technology. Our group has learned a great
deal about the effect that fluorine-substitution has on absor-
bance and demonstrated the value of acquiring the VUV data
of a variety of model compounds and polymers. Model
studies showed that a transparency price is being paid for
the incorporation of carbonyl functionality in the resist.
These data influenced the design of our addition copolymers
that incorporate geminal trifluoromethyl groups on each
monomer unit. The absorbance of these materials may
ultimately be further improved by introducing an elec-
tron-withdrawing group on the two-carbon bridge, we have
not been able to incorporate such monomers into a copo-
lymer using any addition catalyst. This has inspired us to
pursue TCN monomers and 2-trifluoromethyacrylates in
order to incorporate these intriguing materials into copoly-
mers. Both of these monomers have been incorporated into
resists that have provided images with 157 nm exposure. A
three-component resist system has also been explored, and
should prove to be a viable option for 157 nm resist technol-
ogy. Avariety of dissolution inhibitors have been synthesized
for this purpose, and their design has also been influenced by
VUV studies in order to maximize transparency.
The synthesis of 4, 6–9, 13, 14, 17, and 23 have been
reported previously [14]. The synthesis of molecules 1–3
will be reported elsewhere. Polymer 12 was synthesized
from 9 by reacting the homopolymer with an appropriate
amount 2.5 M n-butyl lithium and chloromethyl ethyl ether
in THF. Monomer 5 was prepared by the cycloaddition of
1,1-difluoro-2-acetoxy-3,3,3-trifluoropropene and cyclo-
pentadiene, hydrolyzed with base, and hydrogenated with
Pd/C under hydrogen atmosphere. Model compound 11 was
synthesized from hexfluoroisopropanol by reacting the alco-
hol with 1.1 equivalents of di-tert-butyl dicarbonate and a
catalytic amount of 4-dimethylamino pyridine in THF. The
same starting material was reacted with 2.5 M n-butyl
lithium and chloromethyl ethyl ether in THF to produce
model compound 10. The synthesis of polymers 16 and 21
were carried out at 80 8C in bulk under N2 after thorough
degassing using AIBN or V601 as the initiator. Monomer 15
was prepared by the reaction of quadracyclane and methyl
3,3-difluoro-2-(trifluoromethyl)acrylate. The methyl ester
was hydrolyzed with NaOH and methanol to the carboxylic
acid, with was then reacted with isobutene and acid to
produce the target molecule. Monomer 18 was prepared
by reacting 1,4-bis(2-hydroxyhexafluoroisopropyl)cyclo-
hexane with 2-trifluoromethylacryloyl chloride (from 2-tri-
fluoromethylacrylic acid and phthaloyl dichloride) and
triethylamine in DCM. Homopolymer 19 was prepared by
reacting 1,4-bis(2-hydroxyhexafluoroisopropyl)cyclohexane
with 1 equivalent of di-tert-butyl dicarbonate and a catalytic
amount of 4-dimethylamino pyridine in THF, isolating the
mono-blocked alcohol, reacting that with 2-trifluoromethy-
lacryloyl chloride, polymerizing the monomer with a cata-
lytic amount of potassium acetate and 18-crown-6 [15], and
finally deprotecting the resulting polymer with trifluoroace-
tic acid. Monomer 20 was prepared by the cycloaddition
of 2-trifluoromethylacrylic acid and cyclopentadiene. The
resulting adduct was deprotonated with 2.5 M n-butyl
lithium and reacted with chloromethyl ethyl ether in THF.
Dissolution inhibitor 22 was synthesized by reacting the
appropriate alcohol with 2 eq. of di-tert-butyl dicarbonate
and a catalytic amount of 4-dimethylamino pyridine in THF.
Dissolution inhibitor 24 was prepared by hydroboration of
the cycloaddition adduct of tert-butyl-2-trifluoromethylacry-
late and cyclopentadiene, and then reaction of that alcohol
with 0.5 eq. of phosgene.
4. Experimental
4.1. Materials
All starting materials were purchased from Aldrich and
were used as received. NBHFA was received as a gift from
JSR Corp. and Central Glass, PNBHFA and 1,1-difluoro-2-
acetoxy-3,3,3-trifluoropropene was received as a gift from
AZ Clariant and 2-trifluoromethylacrylic acid was pur-
chased from Honeywell. All solvents were dried by frac-
tional distillation either over Na/K metal or CaH2. All liquid
reagents used for VUV measurements were distilled from
appropriate drying agents, thoroughly degassed by freeze–
pump–thaw cycles, and sealed in glass ampoules under
vacuum.
4.2. Vacuum-UV spectroscopy
Gas-phase VUV measurements were made on an Acton
CAMS-507 spectrophotometer fitted with a custom-made

26
B.C. Trinque et al. / Journal of Fluorine Chemistry 122 (2003) 17–26
4.4. Imaging
References
All imaging work was performed on an Exitech 157 nm
small field (1:5 mm  1:5 mm) mini-stepper (NA ¼ 0:6)
using either a binary mask (s ¼ 0:7) or phase-shift mask
(s ¼ 0:3) at International SEMATECH in Austin, TX. Scan-
ning electron micrographs were collected on a JEOL JWS-
7550, and cross-sectional data were collected on a Hitachi
4500 microscope. Coating, baking, and development of
resist films were performed using an FSI Polaris 2000 track.
Thickness measurements were made on a Prometrix inter-
ferometer. A typical resist formulation was prepared by
mixing the polymer with 6 wt.% (relative to polymer)
photoacid generator (TPS-Nf) and 0.3 wt.% tetrabutylam-
monium hydroxide (TBAH) as the base to control acid
diffusion and reduce T-topping. Dissolution inhibitors were
mixed with the polymer to the desired ratio. The entire
mixture was diluted in PGMEA to provide a viscosity that
provides resist thicknesses of approximately 100–200 nm
after spinning the resist at 2500 rpm onto a silicon wafer that
had been previously coated with ꢀ80 nm bottom anti-reflec-
tive coating (BARC, Shipley AR19). The post-apply bake
was 140 8C for 60 s and the post-exposure bake was 130 8C
for 90 s, unless stated otherwise. The exposed resists were
developed in the industry-standard 0.26N TMAH developer.
[1] R.D. Allen, G.M. Wallraff, D.C. Hofer, R.R. Kunz, Polym. Mater.
Sci. Eng. 72 (1995) 100–101.
[2] U. Okoroanyanwu, J. Byers, T. Shimokawa, C.G. Willson, Chem.
Mater. 10 (1998) 3328–3333.
[3] K. Patterson, M. Yamachika, S. Cho, T. Rager, S. Yamada, J. Byers,
C.G. Willson, Polym. Mater. Sci. Eng. 81 (1999) 43–44.
[4] R.R. Kunz, T.M. Bloomstein, D.E. Hardy, R.B. Goodman, D.K.
Downs, J.E. Curtin, Proc. SPIE-Int. Soc. Opt. Eng. 3678 (1999)
13–20.
[5] T. Chiba, R.J. Hung, S. Yamada, B. Trinque, M. Yamachika, C.
Brodsky, K. Patterson, A.V. Heyden, A. Jamison, S.-H. Lin, M.
Somervell, J. Byers, W. Conley, C.G. Willson, J. Photopolym. Sci.
Technol. 13 (2000) 657–664.
[6] R.J.-P. Hung, H.V. Tran, B.C. Trinque, T. Chiba, S. Yamada, D.
Sanders, E.F. Connor, R.H. Grubbs, J.M. Klopp, J.M.J. Frechet, B.H.
Thomas, G.J. Shafer, D.D. DesMarteau, W. Conley, W.C. Grant,
Proc. SPIE-Int. Soc. Opt. Eng. 4345 (2001) 385–395.
[7] K. Patterson, M. Yamachika, R.J. Hung, C.J. Brodsky, S. Yamada,
M.H. Somervell, B. Osborn, D. Hall, G. Dukovic, J. Byers, W.
Conley, C.G. Willson, Proc. SPIE-Int. Soc. Opt. Eng. 3999 (2000)
365–374.
[8] V.R. Vohra, K. Douki, Y.-J. Kwark, X.Q. Liu, C.K. Ober, Y.C. Bae,
W. Conley, D. Miller, P. Zimmerman, Proc. SPIE-Int. Soc. Opt. Eng.
4690 (2002) 84–93.
[9] H. Ito, H.D. Truong, M. Okazaki, D.C. Miller, N. Fender, P.J. Brock,
G.M. Wallraff, C.E. Larson, R.D. Allen, J. Photopolym. Sci. Technol.
15 (2002) 591–602.
[10] H. Ito, G.M. Wallraff, P.J. Brock, N. Fender, H.D. Truong, G. Breyta,
D.C. Miller, M.H. Sherwood, R.D. Allen, Proc. SPIE-Int. Soc. Opt.
Eng. 4345 (2001) 273–284.
Note added in proof
[11] C.G. Willson, R.D. Miller, D.R. McKean, L.A. Pederson, M. Regitz,
Proc. SPIE-Int. Soc. Opt. Eng. 771 (1987) 2–10.
Since this article was written we have found that com-
pound 24 is surprisingly volatile and is sublimed out of films
rapidly during the post-apply bake process.
[12] S.-i. Kodama, I. Kaneko, Y. Takebe, S. Okada, Y. Kawaguchi, N.
Shida, S. Ishikawa, M. Toriumi, T. Itani, Proc. SPIE-Int. Soc. Opt.
Eng. 4690 (2002) 76–83.
[13] C. Brodsky, J. Byers, W. Conley, R. Hung, S. Yamada, K. Patterson,
M. Somervell, B. Trinque, H.V. Tran, S. Cho, T. Chiba, S.-H. Lin, A.
Jamieson, H. Johnson, T. Vander Heyden, C.G. Willson, J. Vac. Sci.
Technol. B 18 (2000) 3396–3401.
Acknowledgements
[14] H.V. Tran, R.J. Hung, T. Chiba, S. Yamada, T. Mrozek, Y.-T. Hsieh,
C.R. Chambers, B.P. Osborn, B.C. Trinque, M.J. Pinnow, S.A.
MacDonald, C.G. Willson, D.P. Sanders, E.F. Connor, R.H. Grubbs,
W. Conley, Macromolecules 35 (2002) 6539–6549.
The authors gratefully acknowledge International SEMA-
TECH for financial support of this work. SEMATECH
employees Danny Miller, Shashi Patel, Vicky Graffenberg,
and Georgia Rich are all acknowledged for their help with
the imaging work.
[15] T.E. Hogen-Esch, J. Smid (Eds.), Recent Advances in Anionic
Polymerization, New York, 1987.