Design of very transparent fluoropolymer resists for semiconductor manufacture at 157 nm

Article abstract of DOI:10.1016/S0022-1139(03)00075-7

Photolithography at 157 nm requires development of new photoresists that are highly transparent at this wavelength. Transparent fluoropolymer platforms have been identified which also possess other materials properties required for chemically amplified imaging and aqueous development. Polymers of tetrafluoroethylene (TFE), a fluoroalcohol-substituted norbornene and an acid-labile acrylate ester show the best combination of properties. A solution, semibatch, free-radical polymerization process was developed allowing synthesis of the terpolymers on a multikilogram scale. Further property enhancements may arise from replacing the norbornene with functionalized tricyclononenes. Formulated resists have been imaged in a 157 nm microstepper.

Full text of DOI:10.1016/S0022-1139(03)00075-7

Journal of Fluorine Chemistry 122 (2003) 11–16
Design of very transparent fluoropolymer resists for
semiconductor manufacture at 157 nm
A.E. Feiring*, M.K. Crawford, W.B. Farnham, J. Feldman, R.H. French,
K.W. Leffew, V.A. Petrov, F.L. Schadt III, R.C. Wheland, F.C. Zumsteg
DuPont Central Research and Development, Experimental Station E328/231, P.O. Box 80328, Wilmington, DE 19880-0328, USA
Abstract
Photolithography at 157 nm requires development of new photoresists that are highly transparent at this wavelength. Transparent
fluoropolymer platforms have been identified which also possess other materials properties required for chemically amplified imaging and
aqueous development. Polymers of tetrafluoroethylene (TFE), a fluoroalcohol-substituted norbornene and an acid-labile acrylate ester show
the best combination of properties. A solution, semibatch, free-radical polymerization process was developed allowing synthesis of the
terpolymers on a multikilogram scale. Further property enhancements may arise from replacing the norbornene with functionalized
tricyclononenes. Formulated resists have been imaged in a 157 nm microstepper.
# 2003 Elsevier Science B.V. All rights reserved.
Keywords: Tetrafluoroethylene copolymers; Photolithography; Transparent fluoropolymers; Fluorinated alcohols
1. Introduction
been faithfully reproduced on the chip surface. This process
is repeated multiple times in the construction of the hundreds
of million transistors on modern computer chips.
1.1. Photolithography
Because minimum achievable feature size is directly
proportional to the imaging wavelength, the decrease in
feature size has been enabled by the development of imaging
technology using shorter wavelengths of light. Thus, leading
edge chip manufacture has gone from using 365 nm
(Hg i-line) to 248 nm (KrF laser) to the emerging 193 nm
(ArF laser) light for imaging. Modern optics allow genera-
tion of feature sizes roughly one-half the size of the imaging
wavelength. It is anticipated that the next step in this devel-
opment will be to use of light at 157 nm, provided by the F2
laser, which should emerge as a commercial process in the
middle of this decade.
The astonishing advances in computer technology over
several decades derive largely from the continuing devel-
opment of smaller and faster computer chips. This devel-
opment has been enabled by use of ever smaller features on
the chips, going from micron-sized to about 120 nm at the
current state-of-the-art for critical features.
The complex patterns on a computer chip are generated by
photolithography [1] in which light of a given wavelength is
passed through a mask containing the desired pattern,
focused and allowed to impinge on a photoresist applied
as a thin film on the wafer substrate surface. The light causes
a chemical change in the photoresist which (in a positive
working resist) renders the material more soluble so that the
irradiated areas can be removed by washing with a given
solvent. The wafer with the patterned photoresist on its
surface is then subjected to a ion-etching process. In the
unprotected areas, the underlying substrate is etched
whereas the photoresist prevents etching in the unimaged
areas. After removal of the remaining photoresist, the overall
result is that the pattern originally present in the mask has
1.2. Photoresist technology
A modern photoresist contains multiple components but
the most important is a polymeric binder. The binder poly-
mer must have many critical properties [2] including:
ꢀ functionality for image formation and subsequent disso-
lution;
ꢀ transparency at the imaging wavelength so that sufficient
light can penetrate through its full thickness;
ꢀ etch resistance;
* Corresponding author. Tel.: þ1-302-695-1841; fax: þ1-302-695-9799.
E-mail address: andrew.e.feiring@usa.dupont.com (A.E. Feiring).
0022-1139/03/$ – see front matter # 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0022-1139(03)00075-7

12
A.E. Feiring et al. / Journal of Fluorine Chemistry 122 (2003) 11–16
Table 1
Absorption coefficients at 157 nm for selected polymersa
ꢀ high contrast;
ꢀ a glass transition temperature (Tg) above typical proces-
sing temperatures so that flow does not occur;
ꢀ solubility in selected organic solvent for application by
spin coating;
Polymer
A (mmÀ1
)
Poly(hydrosilsequioxane)
Amorphous perfluoropolymer
Poly(dimethylsiloxane)
Poly(vinyl alcohol) (99.7%)
Poly(methylmethacrylate)
Poly(vinylphenol)
0.06
0.70
1.61
4.16
5.69
6.25
6.73
6.80
11.00
ꢀ molecular weight sufficiently high for acceptable
mechanical properties but low enough for rapid dissolu-
tion in developer;
ꢀ ability to stick to the silicon or other substrates; and
ꢀ extremely high purity, usually measured in parts per
billion levels.
Poly(norbornylmethacrylate)
Poly(norbornene)
Poly(acrylic acid)
a Date reprinted with permission from [7]. Copyright 1999, American
Vacuum Society.
The photoresist binder polymer must, of course, also be
manufacturable on appropriate scale at an acceptable cost.
Lithography at 248 nm is based on a chemically amplified
photoacid-catalyzed process (Fig. 1). The binder polymer is
generally poly(para-hydroxystryene) in which some hydro-
xyls are protected as acid-labile tert-butyl carbonates [3].
The protected polymer is insoluble in aqueous base. Irradia-
tion causes decomposition of a photoacid generator (PAG),
typically a triarylsulfonium perfluoroalkylsulfonate, to gen-
erate strong acid in the irradiated areas. During a post
exposure bake step, the acid catalyzes removal of sufficient
protecting groups to render the exposed areas soluble in
aqueous base for washout. A key feature is that the depro-
tection is catalytic in protons so one photon can cause many
deprotection steps. This chemical amplification of the light
is especially important at shorter wavelengths due to the cost
of generating high energy photons.
At 157 nm the situation becomes even more severe.
Although the maximum tolerable absorbance of a resist
binder polymer is still under discussion, most agree that
it must be under about 2 mmÀ1 and preferably under 1 mmÀ1
.
Studies showed that known resists are far too absorbing for
use at 157 nm at a reasonable film thickness (Table 1) [6,7].
Styrenic and (meth)acrylate polymers are above 6 mmÀ1
,
orders of magnitude higher than the goal value. Interestingly,
even the homopolymer of norbornene which is devoid of
unsaturated groups was found to have an absorbance of
6.8 mmÀ1
.
1.3. Role of fluoropolymers
The hydroxystyrene polymers used at 248 nm have
another important feature: high etch resistance due to the
presence of the aromatic rings. Carbon-rich organics tend to
have better etch resistance than materials with larger
amounts of hydrogen or oxygen [4].
However, aromatic polymers cannot be used at shorter
wavelenghts due to their high absorbances. For 193 nm
photolithography, several resist polymers have been devel-
oped, based on norbornene/maleic anhydride copolymers
or poly(meth)acrylates having polycyclic (e.g. adamantyl)
groups in their side chains [5]. The polycyclic groups are
suitably transparent at 193 nm and provide etch resistance.
Typically, tert-butyl (meth)acrylate or related groups are
incorporated to provide the acid-labile solubility switch.
Two classes of polymers with low absorbance at 157 nm
were identified: certain fluoropolymers and polysiloxanes
[7]. In particular, Teflon1AF, an amorphous copolymer of
tetrafluoroethylene (TFE) and perfluoro(2,2-dimethyldiox-
ole), was found to have an absorbance of about 0.7 mmÀ1 at
157 nm. We found that the alternating copolymer of TFE
and norbornene has an absorbance of 1.3 mmÀ1 at 157 nm,
some five orders of magnitude more transparent than the
norbornene homopolymer [8]. This polymer is amorphous,
has a desirable Tg of about 150 8C and was shown to
have good etch resistance, comparable to existing photo-
resists. Our work has since focused on turning the TFE/
norbornene copolymer into a fully-functional 157 nm photo-
resist [9–12]. This paper will be focused on synthesis of the
Fig. 1.

A.E. Feiring et al. / Journal of Fluorine Chemistry 122 (2003) 11–16
13
(hexafluorisopropanol-substituted polymers have been
used for photolithography at longer wavelengths [13]). In
contrast to the phenols, however, the hexafluoroisopropanol
group is highly transparent at 157 nm. This is illustrated by
the absorbance data in Fig. 4 showing that incorporation of
the fluoroalcohol side chain actually improves the transpar-
ency of the already highly transparent TFE/norbornene
copolymer.
Fig. 2.
The TFE/NB–F–OH dipolymer is soluble in aqueous base
but its rate of dissolution is too slow for a practical photo-
resist so some t-BuAc is used as the solubility switch.
However, the required level of acrylate is significantly
reduced by the presence of the fluoroalcohol groups so that
the terpolymers have absorbance values in the range of 1.0–
2.5, depending on acrylate content.
resist polymers; properties relating to their lithography
performance have been described elsewhere [11].
2. Results and discussion
2.1. Resist polymer structure
These TFE/NB–F–OH/t-BuAc terpolymers have glass
transition temperatures in the range of 140–150 8C and
weight average molecular weight of around 10,000, a sui-
table value for balancing mechanical properties and dissolu-
tion rate. The polymers are readily soluble in moderately
polar organic solvents such as acetone, 2-heptanone, THF
and chloroform and insoluble in hexane and methanol.
Consideration of the properties needed in a polymer for a
157 nm photoresist led to design of a terpolymer (Fig. 2) of
tetrafluoroethylene, tert-butyl acrylate (t-BuAc) and a nor-
bornene with a pendant fluoroalcohol group. Each compo-
nent provides one or more of the required properties, but
may impart certain disadvantages as well. The carbon-rich
norbornane ring provides etch resistance but, as noted above,
must be spaced by TFE units for good transparency at
157 nm. The t-BuAc units provide acid-labile solubility
switches because they are rapidly cleaved to isobutylene
and polymer bound carboxylic acids when heated in the
presence of photogenerated acid. However, the carbonyl
groups are also highly absorbing at 157 nm. We found that
incorporation of sufficient t-BuAc groups to make the
hydrophobic TFE/norbornene polymer soluble in aqueous
base after imaging increased the polymer absorbancy to
above 3.0 mmÀ1. Hence the need for the fourth component,
the pendant hexafluoroisopropanol groups. These are intro-
duced as side chains on the norbornene monomer, NB–F–
OH (Fig. 3), and help to provide solubility in aqueous base.
The hexafluoroisopropanol group is known to be as acidic
as phenol so it readily ionizes for dissolution in aqueous base
2.2. Polymerization technology
There are several key issues with respect to the polymer-
ization technology for preparation of the fluorinated photo-
resist polymers including reactivity ratios of the monomers
and reactivity of the substituted norbornene monomer. The
terpolymers are prepared by free-radical solution polymer-
izations using a peroxydicarbonate initiator and a hydro-
fluorocarbon solvent. TFE and norbornene have a strong
tendency to alternate as expected for the combination of a
highly fluorinated olefin with a hydrocarbon olefin [14] so
control of polymer composition is simple with the dipoly-
mer. The situation is made more complex by the acrylate
termonomer as illustrated by a series of batch polymeriza-
tions of TFE, norbornene and t-BuAc (Table 2). t-BuAc feed
concentrations of 0, 2, 5, 10 and 20 mol% were used and
each recipe was run to low and high conversion by varying
the amount of initiator. The results at low conversion show
that the acrylate content in the polymer is 2.8–4.0 times
greater than that present in the monomer mixture. At high
conversion, the overall polymer composition is nominally
closer to the monomer feed composition but this suggests
the formation of a very inhomogeneous blend during the
Fig. 3.
Fig. 4.

14
A.E. Feiring et al. / Journal of Fluorine Chemistry 122 (2003) 11–16
Table 2
Batch polymerizations of TFE, norbornene and tert-butyl acrylate (t-BuAc)a
Entry
Monomer feed (mol%)
Initiatorb
(mol%)
Yield (%)
Polymer composition (mol%)c
Mn
TFE
NBE
t-BuAc
TFE
NBE
t-BuAc
1
2
70
70
68
68
65
65
60
60
50
50
30
30
30
30
30
30
30
30
30
30
0.25
1.0
31
89
26
91
24
89
27
87
29
63
52
70
49
64
37
56
28
53
16
42
48
30
43
32
44
40
35
35
28
40
6700
4900
5200
4500
6200
4300
6300
5600
9900
6400
3
2
2
0.25
1.0
8
4
4
5
5
0.25
1.0
19
5
6
5
7
10
10
20
20
0.25
1.0
37
13
56
18
8
9
0.25
1.0
10
a Polymerizations in a 200 ml shaker tube at 50 8C using 1,1,2-trichlorotrifluoroethane solvent.
b Initiator is di-(tert-butylcyclohexyl)peroxydicarbonate (Perkadox1 16N).
c Compositions determined by 13C NMR.
course of the polymerization as the more reactive acrylate is
depleted. Thus, a semibatch process, described in Section 4,
was developed. Polymerization is initiated with an acrylate–
lean mixture in the reactor. Initiator and a more acrylate-rich
monomer mixture are fed continuously to the reactor during
the course of the polymerization. This process has now been
scaled-up to the production of multikilogram batches.
A second issue regarding polymerization is the reactivity
of substituted norbornenes. Although norbornene itself poly-
merizes with TFE as described [15], we found that the
available 5-substituted norbornenes were substantially less
reactive, a situation previously observed with other norbor-
nene copolymerizations [16]. We then discovered that the
polymerizability of 5-substituted norbornenes depends
critically on the stereochemisty of the substituent at the
5-position. This is illustrated by the data in Table 3, which
compares copolymerization of TFE with a commercially
available mixture of endo- and exo-norborn-2-en-5-yl acet-
ates and a copolymerization with the pure exo-acetate [17].
The latter is seen to give substantially improved conversion
and higher molecular weights.
The available substituted norbornenes are generally a
mixture of endo- and exo-isomers, mostly endo, due to their
syntheses by cycloadditions to cyclopentadiene. The need
for a norbornene with a purely exo-fluoroalcohol substituent
at the 5-position led to the synthesis shown in Fig. 5. The
known [17] exo-acetate, prepared by addition of acetic
acid to norbornadiene, is hydrolyzed to the exo-alcohol.
The sodium salt of the alcohol then reacts readily with
hexafluoroisobutylene oxide [18] to afford NB–F–OH which
polymerizes with TFE to give polymer with acceptable
conversion and molecular weight.
2.3. Tricyclononene copolymers
Table 3
Copolymerization of TFE and norborn-2-en-5-yl acetatesa
Could incorporation of additional fluorine result in further
enhancement in polymer properties? As noted above, the
cyclic hydrocarbon monomer and TFE have a strong ten-
dency to alternate so simply increasing the TFE content in
the backbone would be difficult. In addition, runs of TFE are
likely to decrease transparency [19] and, if too long, could
Norborneneyl acetate
Conversion (%)
Mn/Mw
Endo/exo mixture
Exo
8
2100/3100
3200/8300
33
a Batch copolymerization of a 60/40 mixture of TFE/norbornenyl
acetate in Solkane1 365 solvent using 0.8 mol% Perkadox1 16N initiator.
Fig. 5.

A.E. Feiring et al. / Journal of Fluorine Chemistry 122 (2003) 11–16
15
2950 TGA, respectively. Tg is reported as the midpoint in the
deflection on a second heat from À25 to 275 8C at a heating
rate of 20 8C minÀ1. Thermal decomposition temperature is
reported as the temperature of 10 wt.% loss under nitrogen
by TGA with a heating rate of 20 8C minÀ1
.
Fig. 6.
Exo-5-norbornen-2-ol was prepared by addition of acetic
acid to norbornadiene, followed by alkaline hydrolysis.
TCN–(F)2(F)2 was prepared by reaction of tetrafluoroethy-
lene (1 equivalent) and norbornadiene (1.25 equivalents) at
180 8C for 8 h in a sealed metal pressure vessel. The product
was purified by spinning band distillation to provide pure
cuts of the tricyclic comonomer, having a boiling point of
67 8C at 45 mm.
lead to undesirable crystallinity in the polymer. Adding
fluorine to the norbornene ring in proximity to the double
bond would have a negative effect on polymerizability.
One approach which we have begun to explore is through
the development of a family of fluorinated tricyclononene
monomers. The prototypical example is the monomer TCN–
(F)2(F)2 (Fig. 6) readily prepared by cycloaddition of TFE
and norbornadiene.
4.1. Synthesis of NB–F–OH
In practice, we found that this tricyclononene monomer
copolymerizes as well or better with TFE than does nor-
bornene, presumably due to the fact that the four membered
ring is on the exo face of the bicycloheptene system and the
fluorines are sufficiently remote from the double bond.
Interestingly, the TFE/TCN–(F)2(F)2 copolymer has a sig-
nificantly higher glass transition temperature (228 8C versus
150 8C) and is more transparent (A ¼ 0:69 mmÀ1 versus
1.3 mmÀ1) than the TFE/norbornene copolymer. We are
currently exploring copolymers of TFE with various func-
tionalized TCN monomers for this application.
A dry round bottom flask equipped with mechanical
stirrer, addition funnel and nitrogen inlet was swept with
nitrogen and charged with 19.7 g (0.78 mol) of 95% sodium
hydride and 500 ml of anhydrous DMF. The stirred mixture
was cooled to 5 8C and 80.1 g (0.728 mol) of exo-5-nor-
bornen-2-ol was added dropwise so that the temperature
remained below 15 8C. The resulting mixture was stirred for
1/2 h. Hexafluoroisobutylene oxide [20] (131 g, 0.728 mol)
was added dropwise at room temperature. The resulting
mixture was stirred overnight at room temperature. Metha-
nol (40 ml) was added and most of the DMF was removed on
a rotary evaporator under reduced pressure. The residue
was treated with 200 ml water and glacial acetic acid was
added until the pH was about 8.0. The aqueous mixture was
extracted with 3 ꢁ 150 ml ether. The combined ether
extracts were washed with 3 ꢁ 150 ml water and 150 ml
brine, dried over anhydrous sodium sulfate and concentrated
on a rotary evaporator to an oil. Kugelrohr distillation at
0.20–0.27 ꢁ 10À4 MPa and a pot temperature of 30–60 8C
3. Conclusions
Fluoropolymers already have significant applications in
the semiconductor industry, mainly as materials for handling
the corrosive and/or high purity chemicals used for making
computer chips. These applications are mainly passive and
protective and derive from the exceptional chemical and
thermal stability of highly fluorinated polymers. It now
seems likely that fluoropolymers will play a more active
role in chip manufacture by functioning as the photoresist
binder polymer in the actual process of chip manufacture.
The new role for fluorine chemistry comes from the impact
of fluorination on the absorption spectrum of materials at
very short wavelengths and from the effect of fluorination of
acidity of a key functional group. However, careful selection
of comonomers and polymerization conditions are required
for achieving the desired balance of polymer properties.
1

gave 190.1 g (90%) of product. H NMR (d, CD2Cl2) 1.10–
1.30 (m, 1H), 1.50 (d, 1H), 1.55–1.65 (m, 1H), 1.70 (s, 1H),
1.75 (d, 1H), 2.70 (s, 1H), 2.85 (s, 1H), 3.90 (d, 1H), 5.95 (s,
1H), 6.25 (s, 1H). Anal. Calcd. for C11H12F6O2: C, 45.53; H,
4.17; F, 39.28. Found: C, 44.98; H, 4.22; F, 38.25.
4.2. Synthesis of a TFE/NB–F–OH/t-BuAc terpolymer
A metal pressure vessel of approximately 1 l capacity was
charged with 206.6 g NB–F–OH, 4.80 g t-BuAc and 75 ml
of Solkane 365 mfc (1,1,1,3,3-pentafluorobutane). The ves-
sel was closed, cooled to À15 8C and pressured to 400 psi
with nitrogen and vented several times. The reactor contents
were heated to 50 8C and TFE was added from a reservoir
to a pressure of 2.21 MPa. A regulator was set to maintain
pressure at 2.21 MPa throughout the polymerization. A
solution of 18.45 g Perkadox1 16N which had been dis-
solved in 100 ml of methyl acetate and then diluted to 200 ml
with Solkane 365 mfc was added at 6.0 ml minÀ1 for 6 min
and then at 0.24 ml minÀ1 for 8 h. Simultaneously, with the
start of initiator feed, a solution of 203.9 g NB–F–OH and
4. Experimental details
GPC data were obtained on Waters Alliance 2690 instru-
ment with a model 410 refractometer using THF as the
mobile phase at 1 ml minÀ1. Column set consisted of three
PL gel 5 mm particle size columns, two mixed C and one
500A. Column temperature was 35 8C and data is reported
versus polystyrene standards. DSC and TGA data were
obtained on TA Instruments model 2920 MDSC and model

16
A.E. Feiring et al. / Journal of Fluorine Chemistry 122 (2003) 11–16
30.0 g t-BuAc diluted to 250 ml with Solkane 365 mfc was
added at 0.28 ml minÀ1 for 12 h. After 16 h reaction time, the
vessel was cooled to room temperature and vented to 1 atm.
The reaction solution was recovered using additional Solkane
365 mfc to rinse. The combined reactor solution and rinse
was concentrated on a rotary evaporator to a thick oil. The oil
was added in a thin stream to 5.1 l of vigorously stirred
heptane. The resulting slurry was stirred for 30 min, then
filtered. The solid was washed with heptane and allowed to
dry on the filter for 90 min (unreacted NB–F–OH is readily
recovered by evaporation of the filtrate and distillation). The
solid polymer (219.6 g) was dissolved a mixture of 94 ml
THF and 350 ml Solkane 365 mfc and this solution was
added slowly to 11 l heptane. The resulting slurry was stirred
for 35 min, then filtered. The solid was washed with 300 ml
heptane and dried in a vacuum oven at 80 8C to give 162.4 g
of polymer. The polymer composition was determined to be
32% TFE, 46% NB–F–OH and 22% t-BuAc by 13C NMR (in
CD2Cl2) using integrals of peaks at 64.6 (OCH2 of NB–F–
OH), 105.9–132.7 (CF2 of TFE and CF3 of NB–F–OH), and
73.6–87.8 (one quaternary carbon of t-BuAc and two carbons
was calculated to be 53% TFE and 47% NB–TFE.
The absorbance at 157 nm was found to be 0.69 mmÀ1 from
measurements on spin cast films of 88.4 and 102.9 nm thick-
ness. Anal. Found: C, 46.26; H, 2.90; F, 49.80.
References
[1] G.M. Wallraff, W.D. Hinsberg, Chem. Rev. 99 (1999) 1801.
[2] K. Patterson, M. Somervell, C.G. Willson, Solid State Technol. 43
(3) (2000) 41.
[3] H. Ito, C.G. Willson, J.M.J. Frechet, US Patent 4,491,628 (1985).
[4] H. Gokan, S. Esho, Y. Ohnishi, J. Electrochem. Soc. 130 (1983) 143.
[5] E. Reichmanis, O. Nalamasu, F.N. Houlihan, Acc. Chem. Res. 31
(1999) 659.
[6] T.M. Bloomstein, M.W. Horn, M. Rothschild, R.R. Kunz, S.T.
Palmacci, R.B. Goodman, J. Vac. Sci. Technol. B 15 (6) (1997) 2112.
[7] R.R. Kunz, T.M. Bloomstein, D.E. Hardy, R.B. Goodman, D.K.
Downs, J.E. Curtin, J. Vac. Sci. Technol. 17 (1999) 3267.
[8] M.K. Crawford, A.E. Feiring, J. Feldman, R.H. French, M.
Periyasamy, F.L. Schadt III, R.J. Smalley, F.C. Zumsteg, R.R.
Kunz, V. Rao, L. Liao, S.M. Hall, in: Proceedings of SPIE—The
International Society for Optical Engineering 3999, 2000, pp. 357–
364.
[9] A.E. Feiring, J. Feldman, PCT Int. Appl. WO 2000 17712 (2000).
[10] A.E. Feiring, J. Feldman, PCT Int. Appl. WO 2000 67072 (2000).
[11] M.K. Crawford, W.B. Farnham, A.E. Feiring, J. Feldman, R.H.
French, K.W. Leffew, V.A. Petrov, F.L. Schadt III, F.C. Zumsteg, J.
Photopolym. Sci. Tech. 15 (2002) 677–687.
from NB–F–OH). 19F NMR (d, CDCl3) À76 to À77 (CF3),

À95 to À125 (CF2). GPC: Mn ¼ 6900, Mw ¼ 9900,
Mw=Mn ¼ 1:44. DSC: Tg ¼ 150 8C, TGA: Td ¼ 292 8C.
Anal. Found: C, 46.51; H, 4.51; F, 34.35.
[12] M.K. Crawford, A.E. Feiring, J. Feldman, R.H. French, V.A. Petrov,
F.L. Schadt III, R.J. Smalley, F.C. Zumsteg, in: Proceedings of
SPIE—The International Society for Optical Engineering 4345. Part
1. Advances in Resist Technology and Processing XVIII, 2001,
p. 428.
4.3. Synthesis of TFE/TCN–(F)2(F)2 copolymer
A 200 ml stainless steel pressure vessel was charged with
46.1 g of TCN–(F)2(F)2, 75 ml of 1,1,2-trichlorotrifluor-
oethane and 1.0 g of Perkadox1 16N. The vessel was closed,
cooled in dry ice, evacuated and charged with 36 g of tetra-
fluoroethylene. The vessel was then agitated with its contents
at 50 8C for 18 h. The vessel was cooled to room temperature
and vented to one atmosphere. The translucent, gel-like
solution was removed from the vessel using additional
1,1,2-trichlorotrifluoroethane to rinse. The mass was allowed
to air-dry. The polymer was dissolved in tetrahydrofuran
and precipitated into excess methanol. The solid was dried
in a vacuum oven at 85 8C giving 18.0 g of white polymer;
GPC: Mn ¼ 9400, Mw ¼ 13,100; Mw=Mn ¼ 1:40. DSC:
Tg ¼ 228 8C. 19F NMR À95 to À122 (multiplets, 4F from
TFEand2FfromNB–TFE),À124.4(dd,2FfromNB–F–OH).
From integration of the spectrum, the polymer composition
[13] K.J. Przybilla, H. Roschert, G. Pawlowski, Adv. Mater. 4 (1992) 239;
H. Ito, G.M. Wallraff, N. Fender, P.J. Brock, W.D. Hinsberg, A.
Mahorowala, C.E. Larson, H.D. Truong, G. Breyta, R.D. Allen, J.
Vac. Sci. Technol. B 19 (2001) 2678.
[14] A.E. Feiring, in: R.E. Banks, B.E. Smart, J.C. Tatlow (Eds.),
Organofluorine Chemistry, Principles and Commercial Applications,
Plenum Press, New York, 1994, pp. 339–372.
[15] M. Kobo, H. Inukai, T. Kitahara, K. Sugioka, US Patent 5,177,166
(1993).
[16] H. Ito, D.C. Miller, M. Sherwood, J. Photopolym. Sci. Technol. 13
(2000) 559.
[17] S.J. Cristol, T.C. Morrill, R.A. Sanchez, J. Org. Chem. 31 (1966)
2733.
[18] V.A. Petrov, Synthesis (2002) 2225.
[19] R.L. Waterland, K.D. Dobbs, A.M. Rinehart, A.E. Feiring, R.C.
Wheland, B.E. Smart, J. Fluorine Chem. 122 (2003) 37–46.
[20] V.A. Petrov, A.E. Feiring, J. Feldman, PCT Int. Appl. WO 2000
66575 (2000).