Download - Polysilanes for semiconductor fabrication
Polysilanes for semiconductor fabrication
Shuzi Hayase*
Corporate Research and Development Center, Toshiba Corporation, Komukai-toshiba-cho, Saiwai-ku, Kawasaki 212-8585, Japan
Received 5 November 2001; revised 26 May 2002; accepted 10 June 2002
Abstract
Polysilanes have various interesting properties and many applications have been proposed. However, low durability at elevated
temperature and that toward light exposure on polysilanes have limited their application fields. The most promising application is
lithography for LIS fabrication. The unusual photo-reactivity and high etching durability render polysilanes the candidate of most
promising material for the LSI lithography in future, where new materials have been required in order to fabricate micro-patterns
less than 0.2 mm. In this report, application of polysilanes to semiconductor fabrication, particularly, bilayer resists and anti-
reflection layers with high etching properties, is reviewed. This review includes synthesis of new polysilanes, photo-reactivity,
amplification of the photo-reactivity, bilayer resists developable with aqueous solutions, pattern fabrication with all dry process,
UV absorptions, anti-reflection properties, durability toward plasma etchings and the pattern transfer.
q 2002 Published by Elsevier Science Ltd.
Keywords: Polysilane; Lithography; Bilayer; Resist; Anti-reflection; Photo-reactivity; Amplification; Electron transfer; Benzophenone; Pattern
transfer
Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 360
2. Problems that current LSI lithography technologies are facing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 360
3. Bilayer resist systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 360
3.1. Polysilane resists developed with dry processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361
3.2. Wet development with base aqueous solutions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362
4. Anti-reflection layers and problems to be improved . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
4.1. Syntheses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367
4.2. UV absorption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367
4.3. Relationship between n, k values and polysilane structures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369
4.4. Polysilane etching behaviors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 370
4.4.1. Relationship between etching selectivity toward resists and polysilane structures. . . . . . . 370
4.4.2. Oxygen atoms in polysilane backbones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371
4.4.3. Effects of pressures in an etching chamber . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371
4.4.4. Reaction mechanism. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372
4.5. Etching durability of polysilanes during SiO2 etching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373
5. Pattern fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374
5.1. Resist patterns on anti-reflection layers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374
5.2. Patterning of polysilane layers by dry etching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375
6. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378
0079-6700/03/$ - see front matter q 2002 Published by Elsevier Science Ltd.
PII: S0 07 9 -6 70 0 (0 2) 00 0 34 -5
Prog. Polym. Sci. 28 (2003) 359–381
www.elsevier.com/locate/ppolysci
* Present address: Graduate School of Life Science and Systems Engineering, Kyushu Institute of Technology, 1-1 Hibikino Wakamatsu-ku,
Kitakyushu 808-0196, Japan. Tel.: þ81-93-695-3479; fax: þ81-93-695-3470.
E-mail address: [email protected] (S. Hayase).
1. Introduction
Polysilanes are high molecular weight polymers contain-
ing consecutive silicon main chains synthesized first in
1980s [1]. The consecutive silicon chains are expected to be
completely different from the corresponding consecutive
carbon chains. As results of intensive researches, various
properties such as liquid crystal properties, thermochro-
mism, piezochromism, photo-reactivity, non-linear optical
properties, emission properties and semiconductive proper-
ties have become clear, some of which are very important
for future electronics. For example, non-linear optical
properties, emission properties and semiconductive proper-
ties are suggested to be useful for swift optical switches and
light emission displays [2–4].
Low durability at elevated temperatures and toward light
exposures are serious problems for polysilanes to be
successfully applied to these electronic devices. Promising
applications would be to use polysilanes as materials for
fabrications of patterns, because polysilanes are easily
changed to materials having completely opposite properties
when these are exposed to lights [5] and do not remain in
these devices. In LSI (large scale integrated circuit)
fabrication processes, patterns are fabricated by differences
in etching rates of various materials. For example, SiO2 and
Si layers are dry-etched faster than organic polymers by
using fluorine and chlorine gases. Therefore, SiO2 and Si
patterns can be fabricated by using these organic polymers
as masks. SiO2 layers are dry-etched faster than Si layers by
using fluorine gases, which makes it possible to fabricate
SiO2 layers by using Si layers as masks. Similarly, dry
etching rates for polysilane layers are different from the
corresponding polysiloxane layers which are formed by
irradiating polysilanes. The dry etching rates can be
controlled by the exposure doses. This makes polysilanes
attractive as lithography materials. At the end of the
lithography process, these polysilanes are removed and do
not remain in these devices. Durability for long time is not
required for the purpose. Another advantage for polysilanes
to be useful for LSI manufacturing processes is that silicon
atoms contained in polysilanes are familiar with LSI
manufacturing processes and are not recognized to be
impurities. Other metal atoms are apt to cause serious
contamination problems.
In this paper, first of all, problems that LSI lithography
technologies are facing are discussed. Then, potential
applications to bilayer resist and anti-reflection layers are
reviewed. Bilayer photoresists developed with organic
solvents and laser ablation have been extensively investi-
gated since 1980s and there are many reviews [5–14].
Therefore, they are omitted. Bilayer resists, which can be
developed with dry etchings, are added in this review
because these results are closely related to the fabrication of
anti-reflection layers.
2. Problems that current LSI lithography technologies
are facing
KrF and ArF excimer laser lithography has been
investigated for geometry down 0.2 mm, but the lithography
process windows become narrower as the design rule
reaches below the wavelength. Precise critical dimension
(CD) controls become very important in order to achieve
high performance of devices and good yields. Reflection
between resist/air interface or resist/substrate interface
causes two serious problems. One is line-width variation,
and the other is standing wave as shown in Fig. 1.
The interference between the upper and lower resist
interfaces strongly affects the energy absorbed in the resist
film, and the line-width of the resist patterns varies with the
changes in the thickness of substrates [15,16]. Standing
waves appear strong in high resolution resists, because the
high resolution resists have high dissolution contrast and
small acid diffusion length. These lithographic problems can
be solved by using bilayer resist systems or anti-reflective
layers (ARCs).
3. Bilayer resist systems
Bilayer resist processes using polysilanes provide a
solution to the topography and the reflectivity. The top layer
acts as the photo-imaging layer and a hard mask for the
etching of the underlying substrate. The bottom layer acts as
an ARC as shown in Fig. 2 [5–14,17–38]. The following
properties are required. (1) High sensitivity, (2) high
resolution, (3) development without using organic solvents,
(4) high Tg, (5) high oxygen-reactive ion etching (O2-RIE)
durability toward underlying organic layers, and (6) able to
be removed easily from the substrate. The high sensitivity
relates to the through-put, and the exposure dose under
100 mJ/cm2 is needed. Developments using much solvent
should be avoided from the environmental point of view.
The high Tg is needed in order to avoid pattern deformation
Fig. 1. Problems caused by multi-reflections. ARC: anti-reflection
layers; standing waves: see Fig. 24.
S. Hayase / Prog. Polym. Sci. 28 (2003) 359–381360
during O2-RIE. Commonly, the thickness of the imaging
layer and the bottom organic layer are around 0.1–0.2 and
1.0 mm, respectively. Therefore, while the bottom layer is
etched, the etching rate for the bottom layer should be more
than 20 times higher than that for the imaging layer. The
etching rate has been reported to increase with an increase in
the amount of silicon atoms contained in the imaging layer
[39].
Their unique photo-reactivity and high silicon content,
facilitating excellent pattern transfer, make polysilanes
ideally suited for this process. Polysilane layer patterns are
fabricated by its exposure to a deep UV light, followed by
wet development. Then, the organic polymer layer under the
polysilane layer is etched with oxygen-reactive ions. In this
process, SiO2 layers formed at the surface of the polysilane
during the O2-RIE, act as the hard mask.
Recent reports have focused on the development process
which does not use organic solvents. Conventional alkyl or
aryl substituted polysilane resists had to be developed with
organic solvents, such as alcohol/methylethylketone mix-
tures [41–43]. The development is not compatible with
current production lines, where dilute basic aqueous
solutions have been employed. Fig. 3 summarizes poly-
silane resists in terms of development methods.
Two methods have been presented to solve the problem.
One is a dry development and the other is a development
with aqueous solutions.
3.1. Polysilane resists developed with dry processes
The dry development is carried out using Cl2 or HBr gas.
The unexposed polysilane area is etched faster than the
photo-oxidized polysilane area, providing the negative
image [18,20,21,24,26,27,44].
Surface-imaged polysilane layers for 193 nm excimer
laser lithography has been reported [21]. Penetration depth
of lights is apt to become crucial problems for lithography
using shorter wavelength, because conventional organic
resists have inevitable UV absorptions in the wavelength
region below 200 nm. The lithography using surface-
imaging methods is an ideal process. The oxidized polymer
exposed to 193 nm light inhibits bromine-initiated etching,
however, the parent polysilane film was etched to yield a
negative-tone image. Linear and branched polysilanes are
evaluated for the surface-imaging lithography. Patterns of
0.2 mm were fabricated with 50 mJ/cm2 exposure. The
addition of triazine as the sensitizer increased the sensitivity
up to 20 mJ/cm2. The etching rate of the unexposed area was
three times higher than that of the exposed area when
cyclohexylsilyne–butylsilyne copolymer (2:8) was used.
The etching selectivity increased up to 30 when bromine and
hydrogen gases were added to HBr gas. A 100 nm thick
polysilyne film was patterned and the pattern was trans-
ferred to the underlying organic polymer layer by O2-RIEs.
The extent of photo-oxidation is linear with dose, and the
latent image depth is controlled by altering the degree of the
photo-bleaching. During the etching by use of HBr-based
plasma, non-linearity was observed with respect to the
oxidation degree, yielding the development contrast of more
than five. Polybutylsilyne had highest sensitivity toward
HBr etching. Large substituents such as phenyl and
cyclohexyl groups decreased the sensitivity [21].
Polymethylsilyne is expected to have highest sensitivity
toward the HBr dry etching. However, the polymethylsilyne
has problems with solubility and oxidation. Therefore, it
was difficult to deal with and make a high quality film.
There are a few reports on the fabrication of polysilynes
by chemical vapor deposition (CVD). For example, the film
has been fabricated from hexamethylsilazane, hexamethyl-
disilane, tetramethylsilane [20] and methyltrihydrosilane
[24,44]. The polymethylsilyne film (thickness: 120–
170 nm) was plasma-polymerized by using methyltrihydro-
silane gas on a 700 nm layer of a hard-baked photoresist
[44]. After the film was irradiated with 248 nm light, the
film was exposed to chlorine plasma to remove unexposed
regions selectively. The etching selectivity as low as three
between exposed area (100 mJ/cm2) and unexposed area
was obtained. The etching selectivity has been reported to
be sufficient to achieve good lithographic process latitude.
Due to their high silicon content and efficient photo-
oxidation chemistry, the polymethylsilyne film was
readily patterned to give thin-oxide-like-hard masks.
Patterns of 200 nm have been successfully fabricated
without phase shift masks. The results show the high
potential of the polymethylsilyne film for 248 nm
lithography.
A new negative-tone dry development process has been
proposed [45]. First of all, a polysilane layer is exposed to
soft X-rays in order to oxidize the polysilane layer.
Unoxidized areas are removed by KrF laser ablation.
Unexposed areas of the polysilane film had a high ablation
rate while exposed areas cannot be ablated. A bilayer resist
using a polysilane film as a surface-imaging layer has been
patterned using this dry development process. A 0.12 mm
thickness of poly(methylhexyl)silane was coated on a
Fig. 2. Process flow for bilayer resist process.
S. Hayase / Prog. Polym. Sci. 28 (2003) 359–381 361
hard-baked novolac resin layer with 1 mm thickness. The bi-
layer resist was exposed to X-rays with a proximity gap of
0.15 mm. The X-ray exposure dose was 59 J/cm2. After the
exposure, the resist was irradiated in air by KrF laser with a
pulse intensity of 39 mJ/cm2. Total irradiated dose was
120 mJ/cm2. Patterns of 0.4 mm have been reported to be
replicated (Fig. 4).
3.2. Wet development with base aqueous solutions
Polysilane films developable with basic aqueous sol-
utions have been trying to synthesize, because developments
with organic solvents are apt to be avoided in production
lines. Polysilane structures developable with aqueous
solutions are summarized in Fig. 5.
Fig. 3. Polysilane resists and development methods.
S. Hayase / Prog. Polym. Sci. 28 (2003) 359–381362
Polysilanes could be developed with basic aqueous
solutions when hydrophilic substituents such as hydroxy and
carboxylic acid groups are introduced into polysilanes
[33–35]. Polysilanes with phenol groups synthesized for the
first time by us were evaluated as resist materials for KrF
lithography. The polysilane with a phenol group had little
solubility in an aqueous solution of tetramethylammonium-
hydroxide (TMAH), a common developer in recent LSI
lithography. Carboxylic acid groups were successfully
introduced into the polysilanes bearing phenol groups. The
polysilane was proved to be developable with the TMAH
aqueous solution [33,34].
A new copolysilane with carboxylic acid was syn-
thesized by hydrosilylation reaction of poly(phenylhydro)-
silane with vinyl acetic acid [35,36]. This polysilane could
be developed with an aqueous solution of sodium bicar-
bonate. It has been reported that the introduction of t-butyl
groups on phenyl groups improve the sensitivity. The
sensitivity was 540 mJ/cm2 to fabricate patterns with
1.5 mm thickness. The increase in the sensitivity is due to
the t-butyl group blocking the aryl ring from the attack by
silyl radicals making it more difficult to form cross-linking
[35,36].
These resists have hydrophilic substituents in the side
chains to promote the dissolution to basic aqueous solutions.
Poly(phenylhydrosilane) not bearing the corresponding
hydrophilic substituents was found to be developed with
the TMAH aqueous solution [37,38]. This has an advantage
that the introduction of these hydrophilic substituents are
not needed. The poly(phenylhydro)silane has been reported
to be synthesized directly from phenylhydrodichlorosilane
by using zirconocene complexes [46].
The reaction mechanism was explained as follows:
Poly(phenylhydrosilane) photodecomposes to form silyl
compounds having Si–OH groups, which are acidic and
become soluble in an aqueous base solutions as shown in
Fig. 6. The parent poly(phenylhydrosilane) is stable in the
aqueous solution. Therefore, the polysilane film provides
positive-tone patterns. The results proved that poly(phenyl-
hydrosilane) has potentials to give the positive tone patterns
with aqueous base solutions. However, the sensitivity and
the pattern shape had to be improved. The dose to make
patterns reached about 1 J/cm2.
It was found that the addition of 3,30,4,40-tetra (t-
butylperoxycarbonyl)benzophenone (EA1) increases
specifically the resist sensitivity to 100–200 mJ/cm2. The
increase in the resist sensitivity was discussed in terms of
photolysis rates and dissolution rates into aqueous base
solutions. The increase in the photolysis occurs through
photo-induced electron transfers from polysilanes to
electron acceptors.
The electron transfer was analyzed by employing Stern-
Volmer kinetics, which is often followed by purely
diffusional quenching of fluorescent materials by electron
acceptors [47]
F0=F ¼ 1 þ Ksv½Q� Ksv ¼ Kqðp0Þ
Kq is the bimolecular quenching constant; F0, the fluorescenceFig. 5. Polysilane structures developable with base aqueous
solutions. PR1: Refs. [33,34]; PR2: Ref. [35]; PR3: Refs. [37,38].
Fig. 4. Process flow for polysilane pattern fabrication by flat KrF
exposure. Reprinted with permission from Jpn J Appl Phys 1995;34:
6800. q1995 Japan Society of Applied Physics [45].
Fig. 6. Dissolution mechanism of poly(phenylhydrosilane) film into
base aqueous solution Poly(phenylhydrosilane) itself is hydro-
phobic and is insoluble in base aqueous solution. But after exposure
to KrF laser, compounds bearing SiOH groups form. These low
molecular weight oligomers or monomers bearing SiOH are soluble
in base aqueous solution. because SiOH moieties are acidic.
S. Hayase / Prog. Polym. Sci. 28 (2003) 359–381 363
in the absence of electron acceptors; F, the fluorescence in
the presence of electron acceptors; Ksv, the Stern-Volmer
constant; p0 is the fluorescence life time of poly(methyl-
phenylsilane), 86ps [48,49].
The fluorescence intensity of poly(methylphenylsilane)
should decrease when electrons transferred from photo-
excited poly(methylphenylsilane) to electron acceptors.
The decrease in the polysilane fluorescence implies the
ease of electron transfer reactions.
Fig. 7 shows the relationship between F0=F and electron
acceptor (quencher) concentration. Poly(methylphenylsi-
lane) was used instead of poly(phenylhydrosilane), because
the fluorescence of poly(methylphenylsilane) was easy to
measure. EA1, EA2, EA4 and EA7 were used as electron
acceptors, where EA4 and EA7 have been reported to be
efficient photosensitizers for polysilane photo-decompo-
sitions [48,49]. The Stern-Volmer plot shows slightly
positive deviation from linearity at higher acceptor
concentration. Therefore, we fitted it at low acceptor
concentration by using the method of least squares in
order to obtain Ksv and Kq. (Kq)s for various electron
acceptors are summarized in Table 1. Fig. 7 clearly shows
that the poly(methylphenylsilane) is quenched more
efficiently by the addition of EA1 and EA2, compared
to other electron acceptors. EA1 and EA2 show
anomalous behavior for fluorescence quenching of
poly(methylphenylsilane). (Kq)s of the polysilane in
the presence of EA1 and EA2 were one order higher
than expected [48,49]. We assume that EA1 and EA2
may locate close to polysilanes with configurations
suitable for electron transfers by the weak interaction,
Fig. 7. Relationship between F0=F and quencher concentration.
Reprinted with permission from Macromolecules 1998;31:8794.
q1998 American Chemical Society [38].
Table 1
Kq values for polysilane fluorescence quenching in the presence of
various electron acceptors
Electron acceptor Kq ( £ 10210) (M21 s21)
EA2 212
EA4 17
EA1 221
EA5 12
EA7 20
EA6 6
EA3 12
Fig. 8. Relationship between decrease in UV absorption and
irradiation time for poly(methylphenylsilane) film in the presence of
electron acceptors. UV exposure, 365 nm; observation, 340 nm.
Reprinted with permission from Macromolecules 1998;31:8794.
q1998 American Chemical Society [38].
S. Hayase / Prog. Polym. Sci. 28 (2003) 359–381364
such as hydrophobic interactions between a benzophe-
none structure and polysilane structures or hydrophilic
interactions between a benzophenon carbonyl and Si–
OHs of the polysilane chain ends.
The fluorescence quenching rate of poly(phenylhydrosi-
lane) in the presence of EA1 was almost the same as that in
the presence of benzophenone (EA2). EA1 did increase both
photolysis rates and resist sensitivities, but EA2 did not.
Fig. 8 shows the decrease in the UV absorption for a
poly(methylphenylsilane) film when 340 nm light are
irradiated. The decrease in the UV absorption corresponds
to polysilane photo-decomposition. The difference between
EA1 and EA2 was discussed in terms of back-electron
transfer reactions as shown in Fig. 9. When polysilanes are
photo-excited, electrons shift from polysilanes to acceptors
to form polysilane radical cations and acceptor radical
anions. The radical anion should exist as a counteranion of
the polysilane radical cation (R2 in Fig. 9). Radical anions
of EA1 have peroxide moieties and should decompose to
carboxylate anions and tert-butyloxy radicals (R5 in Fig. 9).
Therefore, the electron does not get back to the polysilane
radical cation (R6 in Fig. 9), leading to Si–Si bond
cleavages, where nucleophilic attacks may be needed from
the viewpoint of the energy diagram [38]. In case of EA2,
because the EA2 radical anion should be very stable, the
polysilane radical anion would have a chance to obtain an
electron from the EA2 radical anion to form the parent
polysilane by back-electron transfers. Miller and co-workers
have reported that the photolysis of poly(methylphenylsi-
lane) was showed to be stabilized in the presence of
perchloropentacyclododecane in spite of the fact that
effective fluorescence quenching of poly(methylphenylsi-
lane) was observed in the presence of it. They implied that
this may be due to rapid back-electron transfer [48,49].
This may provide a useful method to stabilize polysilane
films against UV irradiations. We found that poly(methyl-
phenylsilane) was strongly quenched by C60. The Kq was
300 times larger than that of EA1. It has been reported that
photo-induced electron transfers actually occur from photo-
excited polysilanes to C60 [50–52]. When poly(methyl-
phenylsilane) was exposed to a 254 light in the presence of
C60 (0.5 wt% to polysilanes), we found that the photolysis
rate actually decreased, compared with that in the absence of
C60. This would also be explained by back-electron
transfers from very stable C60 radical anions to polysilane
radical cations. The C60 radical anion has been reported to
survive for more than 1.5 s [53].
In case of EA1, the cleavage of peroxide bonds on the
benzophenone moiety would effectively inhibit the back-
electron transfer to the polysilane radical cation. In addition,
EA1 photodecomposes to form compounds bearing car-
boxylic acids which also facilitate the dissolution of the
resist layer into aqueous base solutions. Fig. 10 shows the
contrast curves for polysilane resist (poly(phenylhydrosi-
lane)) in the presence of EA1 and EA2. The development
was carried out by using a 2.38% tetramethylammonium
hydroxide aqueous solution. The figure shows clearly that
the addition of EA1 increased the sensitivity.
4. Anti-reflection layers and problems to be improved
It has been reported that bottom ARCs can eliminate
lithographic problems discussed before. ARCs are deposited
by spin-coating [54], by CVD [55] and sputtering. CVD
ARCs have advantages of highly conformal coatings and
minimize the amount of over-etch. Recently, plug vias and
damascene metalization process are used and pattern
formations on chemical mechanical polished (CMP)
dielectric films are required. As semiconductor topographies
become more planar by using the CMP process, the high
conformality obtained from CVD ARCs becomes less of an
advantage. Spin-coating type ARCs are superior to CVD
ARCs in terms of their simplicity.
Wide lithographic margin could be obtained by thin
Fig. 10. Contrast curves of polysilane resist (poly(phenylhydrosi-
lane)) in the presence of EA1 and EA2. Exposure, 254 nm;
development, 2.38% tetramethylammonium hydroxide aqueous
solution for 1 min. Reprinted with permission from Macromol-
ecules 1998;31:8794. q1998 American Chemical Society [38].Fig. 9. Electron transfer from polysilanes to electon acceptor EA1.
S. Hayase / Prog. Polym. Sci. 28 (2003) 359–381 365
resist systems, because they prevent these resist pattern from
collapsing [55]. Thick ARCs with moderate absorbance
have a good anti-reflection ability. However, conventional
organic RACs cannot be used as a thick film, because resist
thickness loss occurs during the RAC etching. Spin-coatable
ARCs with high etching selectivity toward resists are
strongly desired.
We would like to propose a potential use of polysilanes
for ULSI fabrication, where the polysilane acts as an ARC
as well as a hard mask for fabricating underlying substrates.
The multi-layer is composed of a resist layer, a polysilane
layer and a substrate. These structures and the process flow
are summarized in Fig. 11.
An organic resist is spin-coated on a polysilane layer and
is exposed to UV lights through photo-masks. The
polysilane layer absorbs the UV light passing through the
resist layer and prevents reflections from the underlying
substrate. After the resist pattern is fabricated, the multi-
layer is exposed to Cl2 plasma. Because the polysilane layer,
consisting of consecutive Si–Si bonds, is etched faster than
the resist layer, the resist profile is transferred exactly to the
polysilane layer. This process has two advantages. Firstly,
high resolution organic resists can be used freely, without
the synthetic limitations experienced when using silicon-
containing resists. Secondly, the thickness of the resist for
this process can be thinner than that for a single layer resist
process, which improves the process latitude during
exposure. Anti-reflective properties are attainable by
employing conventional organic polymers not bearing
silicon atoms. However, it is difficult to etch the organic
anti-reflective layer (ORGARC) selectively using the resist
mask, because both the anti-reflective polymer and the resist
polymer have aliphatic hydrocarbon structures.
In order to complete the ARC process successfully, a
number of features are important, for example, the solubility
properties of the polysilane as it is necessary, to prevent the
polysilane layer mixing with the resist layer. We require the
ability to make a planar film, high glass transition
temperatures (Tg), anti-reflective properties, a high etching
selectivity toward resists, and a high etching durability
toward underlying substrates. These requirements are also
summarized in Fig. 11.
In this paper, we focus on clarifying the following
relationships.
1. The relationship between polysilane structures and UV
absorptions and n, k values at 248 nm.
2. The relationship between polysilane structures and
etching selectivity toward resist polymers.
Fig. 11. Process flow for polysilane anti-reflection layer and
requirements for each step.
Fig. 12. Abbreviations for linear polysilanes, cross-linked poly-
silanes and branched polysilanes.
S. Hayase / Prog. Polym. Sci. 28 (2003) 359–381366
3. The relationship between polysilane structures and
etching selectivity toward SiO2 layers.
Finally, SiO2 patterns are fabricated by using polysilane
anti-reflection layers to discuss the potential.
In order to fulfill these requirements, a few new
polysilanes were synthesized. Before discussing properties
of anti-reflection layers, we want to describe the synthesis.
4.1. Syntheses
Polysilanes have been reported to be synthesized by
Wurt coupling reaction [68], reducing metals less active
than sodium [56–58], electrochemical reactions [59,60],
ring-opening reactions [61–65], masked disilane methods
[66,67], and so on. Among three kinds of polysilanes,
namely, linear polysilanes, polysilanes cross-linked with
ethylene units, and branched polysilanes, as shown in Fig.
12, linear polysilanes were prepared by conventional Wurtz
coupling reaction [68]. Polysilanes cross-linked with
ethylene units were also synthesized in the same ways.
Instead of dichlorosilane monomers, 1,2-bis (dichloro-
methyl)ethane were employed for the syntheses of PX1,
PX2, PX3 and PX4. Poly(phenylhydrosilane) PB1 was
synthesized according to the method reported by Li et al.
[69].
Some of the branched polysilanes have Q units (Si units
bonding to four silicon atoms). Polysilanes containing Q
units have been synthesized by reacting mixtures of organic
halides, trichlorosilane and tetrachlorosilane with Na, or by
reacting tetrachlorosilane with Mg under ultrasonic
irradiation, followed by adding organic halides [57,58,
70–72]. Polysilanes containing Q units are apt to make gels
during the syntheses. We synthesized these polymers by a
modified method as follows.
In a four necked round bottomed flask equipped with a
dropping funnel and a condenser, 100 ml of dimethox-
yethane (DME), 13.3 g of 15-crown-5-ether, 0.1 g of CuCl,
48.63 g of Mg and a small amount of iodine were placed.
Then 84.9 g of tetrachlorosilane in 30 ml DME was added
dropwise. After exothermic reaction started, 370 ml of DME
was added in the flask at once. The reaction mixture was
stirred for 20 h, followed by stirring the mixture for 2 h for
75 8C. Butylbromide of about 137 g in 20 ml DME was
added dropwise to the mixture for 60 min and the mixture
was stirred for 1 h at 60 8C. Then toluene was added to the
reaction mixture in order to precipitate salts. The mixture
was filtered under vacuum and the filtrate was washed with
water. The organic layer was separated and dried over
anhydrous MgSO4. The supernatant solution was separated
and solvents were removed under vacuum. The residue was
resolved in 50 ml of toluene and the mixture was poured into
500 ml of ethanol to obtain polymer PB3. Yield, 33%.
Polysilanes cross-linked with ethylene units were also
synthesized for the first time. PX2 was synthesized as
follows. In a four necked round bottomed flask equipped
with a dropping funnel and a condenser, 130 ml of xylene,
20 ml of anisol, 0.35 g of CuCl and 20.6 g of Na were
placed. The sodium was dispersed by stirring the mixture
vigorously at 120 8C. Then, diphenyldichlorosilane
(63.38 g, 0.25 mol) and 1,2-bis(dichloromethylsilyl)ethane
(15.37 g, 0.06 mol) were added for 20 min and stirred for
2 h at 120 8C. After the reaction mixture was cooled down to
room temperature, 500 ml of toluene was added. After the
precipitate was removed under an Ar atmosphere, the
solvent in the filtrate was removed under a reduced pressure.
The residue was again dissolved in 100 ml of toluene. Then
the polysilane was precipitated from 500 ml of ethyllactate.
The polymer was purified three times in the same manner,
and dried at 90 8C for 1 day. PX2: Yield, 33%.
4.2. UV absorption
The reasons why these three structures are selected are as
follows. Linear polysilanes have unique UV absorptions in
the region from 300 to 400 nm assigned to s–sp transitions.
These UV absorptions vary, depending on substituents on
silicone atoms, the main chain conformation, the branching
of the main chain, and molecular weights [73–86].
The introduction of phenyl substituents on silicon atoms
has been reported to shift the UV absorption peak up to
around 400 nm, and to increase the UV absorbance at
around 250–300 nm assigned to the p–pp transition of the
phenyl group [37–41,46]. Linear polysilanes were syn-
thesized in order to know the effectiveness of these aromatic
groups on the UV absorbance at 248 nm. The distortion of
the silicon main chains disturbs the overlap of the silicon s
orbital, and shifts the UV absorption peak to shorter
wavelengths [44,45]. Distorted polysilanes were given by
the cross-linking of polysilane chains by ethylene units. The
branching of the silicon main chain and the decrease in the
number of the consecutive silicon atoms also shift the UV
absorption peak to shorter wavelengths [39–41,47–52]. Me
(SiMe2)5Me has a UV absorption peak at 250 nm. However,
it was difficult to make the high quality flat film with the
polydimethylsilane oligomers. Therefore, the introduction
of aromatic substituents, branched structures and distorted
structures was taken into account to increase the UV
absorbance at 248 nm.
Fig. 13. Representative UV absorptions for linear polysilanes, cross-
linked polysilanes and branched polysilanes.
S. Hayase / Prog. Polym. Sci. 28 (2003) 359–381 367
Anti-reflective properties of polysilanes were compared
each other, by measuring the UV absorbance of their films
with 0.1 mm thickness. The UV absorbance was monitored
at 248 nm, the wavelength employed for KrF excimer laser
lithography. Details for the anti-reflective property of
polysilanes are described later.
UV absorption peaks assigned to the silicon s–sp
transition were clearly observed for linear polysilanes
from PS1 to PS6 at around 300–400 nm, but those for
other polysilanes were observed as shoulders or were
overlapped with other UV absorptions at around 250–
300 nm. This leaded to the increase in the UV
absorbance at 248 nm. Representative UV absorption
spectra for the linear polysilane (PS1), the cross-linked
with ethylene units (PX2) and the polysilanes with
branched structures are shown in Fig. 13. These spectra
imply that the introduction of the branching and the
chain distortion is effective for the increase in the UV
absorption at 248 nm.
In order to examine the effect of the aromatic group, the
UV absorbance for polysilane films with 0.1 mm thickness
at 248 nm was plotted against the percentage of the aromatic
group (Fig. 14). In case of methylphenylsilane–diphenylsi-
lane copolymers (PS2–PS6, Area A in Fig. 14), the UV
absorbance at 248 nm increased with increasing the
percentage of the phenyl group.
The UV absorbance of polysilanes cross-linked at the
ethylene group (PX1, PX2 and PX3) was found to be higher
than that of linear polysilanes (Area B in Fig. 14),
suggesting that partial cross-linking increases UV absor-
bance at 248 nm. The UV absorbance increased in the order
of PX1, PX3 and PX2. PX3 had almost the same UV
absorbance with PX2, even though the percentage of the
phenyl group of PX3 (35%) was much less than that of PX2
(65%). The introduction of the silicon branched structure in
PX3 may well explain the increase in the UV absorbance.
Fig. 14 shows that branched polysilanes are in the region
(Area C) above polysilanes cross-linked with ethylene units
(Area B). For example, the UV absorbance of PB2 was higher
than that of the linear polysilane, PS3, even though the
percentage of the phenyl group of PS3 is the same as that of
PB2. In each area, there was the tendency that the UV
absorbance increases with increasing the amount of the
aromatic group with a few exceptions. PB1 was in the region
under Area C. This is explained as follows: Hydrogen atoms
of PB1 have been reported to be partially removed to make
lower branched silicon backbones during the synthesis [56,
69]. The degree of the branching for PB1 should be lower
Fig. 14. Relationship between UV absorbance for polysilane films and percentage of aromatic groups. 0.1 mm thickness at 248 nm.
Fig. 15. Relationship between UV absorbance for polysilane films and Si content. 0.1 mm thickness at 248 nm.
S. Hayase / Prog. Polym. Sci. 28 (2003) 359–381368
than that for PB2. Therefore, PB1 has linear polysilane
structures and lower branched structures. This explain the
results that PB1 is in the region between A and C areas. PB4
has large UV absorbance even though it has no phenyl
groups. PB4 has 42% of Q structures, the structure that a
silicon atom is bonded to four Si atoms. This may increase the
UV absorbance, instead of phenyl groups. Fig. 14 shows that
UV absorbance increased with increasing the amount of
aromatic groups for all three kinds of polysilanes, and the UV
absorbance increased in the order of linear polysilanes,
polysilanes cross-linked with ethylene units and branched
polysilanes.
Since the etching selectivity of polysilanes toward resists
is expected to increase with the increase in the Si content, it
is important to know how the UV absorbance changes,
depending on the Si content. Fig. 15 shows the relationship
between them.
The UV absorbance decreased with increasing the Si
content for all three kinds of polysilanes. The tendency is
shown as circles D, E and F in Fig. 15. When compared at
the same Si content, the UV absorbance increased in the
order of linear polysilanes, polysilanes cross-linked with
ethylene units and branched polysilanes. PX3 has a higher
UV absorbance than other polysilanes cross-linked with
ethylene units. This may be explained by the specific
structure containing both cross-linking and branched
structures. These relations imply that branched polysilanes
should be the best structure in order to obtain both high Si
content and UV absorbance.
In conclusion, the introduction of aromatic groups, the
chain distortion and the branching of the silicon main chain
turned out to be effective for the increase in the UV
absorption of the polysilane film at 248 nm.
4.3. Relationship between n, k values and polysilane structures
Refractive index is defined as follows:
N ¼ n 2 ik
N is the complex index of refraction; n, the refractive index;
k is the extinction coefficient.
The k means the decrease in the wave amplitude when
the wave moves in a media by one wavelength. Both n and k
correlate with a (absorbance) as follows
a ¼ 4pnk=l
The intensity reflectance, R, is defined by the following
equation [87]
R ¼ ððnr 2 npÞ2 þ ðkr 2 kpÞ
2Þ=ððnr þ npÞ2 þ ðkr þ kpÞ
2Þ
nr is the refractive index of resist; np, the extinction
coefficient of resist; kr, the refractive index of polysilane; kp
is the extinction coefficient of polysilane.
This equation shows that R decreases when nr is close to
np and kr is close to kp. R also decreases when nr, np, kr and kp
have large values.
Fig. 16 shows the relationship between the weight
percentage of phenyl groups and n and k values for linear
polymers, PS2–PS6. Simple copolymers were employed in
order to eliminate other factors associated with n and k
values. The n and k values at 248 nm increased with
increasing the weight percentage of the phenyl group.
Fig. 17 shows the relationship between Si content and n
and k values for polysilanes having Q units and cross-linked
structures. The values of n and k increased with increasing
Si content. The values of n for PB4 was lower than expected.
This may be explained by the fact that the n value for
Fig. 16. Relationship between weight percentage of phenyl groups
and n, k values for model polymers, PS2–PS6.
Fig. 17. Relationship between Si content and n, k values for cross-linked and branched polysilanes.
S. Hayase / Prog. Polym. Sci. 28 (2003) 359–381 369
amorphous silicon is 1.74. PB4 has Q unit structures similar
to the amorphous silicon. The n and k values for
conventional resists are 1.78 and 0.02. The reflectance
should vary, depending on the n and k values for both resists
and polysilanes. The reflectivity should decrease when the n
value is close to 1.78 and the k value increases. The actual
reflectivity will be discussed later.
Table 2 summarizes n and k values for each layer of
multi-layers. One of the advantages of polysilane layers is
that the n value is close to that of resist. This is one of the
requirements in order to decrease the reflection as
mentioned earlier.
4.4. Polysilane etching behaviors
4.4.1. Relationship between etching selectivity toward
resists and polysilane structures
During etching of polysilane layers by chlorine plasma,
the polysilane layer must be etched anisotropically. The
anisotropic etching conditions of polysilane films with a
large aspect ratio have not been previously reported in
detail. Fluorine-containing gases, Cl2 and HBr gases have
been used for inorganic silicon etching [88]. The silicon
layer is etched by chemical reactions and physical
sputtering. The latter may contain chemical reactions
assisted with ion bombardments [88]. The chemical reaction
tends to etch the silicon layer isotropically while the
physical sputtering tends to etch it anisotropically [89].
It has been reported that fluorine-containing gases etch
silicon layers isotropically, because atomic fluorine in the
plasma reacts with the silicon layer spontaneously. On the
other hand, the process using Cl2 and HBr gases provides
anisotropic etching because spontaneous reactions of these
gases with the silicon layer are very few and the etching occurs
in the presence of ion bombardments. The difference between
fluorine and chlorine gases has been explained by the boiling
point of silicon halides and the ease with which the atomic
halides can penetrate the silicon surface [89]. The boiling point
of SiF4 is 178 K and is particularly volatile, while SiCl4 and
SiBr4 have boiling points of 330 and 427 K, respectively.
Therefore, SiCl4 and SiBr4 are less desorbed from the silicon
surfaces: the desorption is accelerated by bombardment with
ions which attack the silicon surfaces anisotropically with the
aid of a self-bias potential. The isotropic etching observed for
fluorine gas has also been explained by the small size of the
fluorine atom, 0.7 A, making it possible to penetrate 10–20 A
in depth from the silicon surface. The covalent radius of Br is
1.1 A and is comparable in size to a silicon atom (1.2 A).
Therefore, in this study, Cl2 gas was employed for the
anisotropic etching of polysilane layers.
Fig. 18 shows the relationship between the silicon
content in the polymer and the etching selectivity of the
polysilane toward the resist (R1) containing polyvinylphe-
nol polymers. The etching selectivity increased with
increasing silicon content for linear polysilanes having
methyl and phenyl groups, such as PS1, PS2, PS3, PS4, PS5
and PS6 (Area G in Fig. 18). This result is expected from the
results of inorganic silicon etchings.
The etching selectivity of polysilanes partially cross-
linked at the ethylene unit, such as PX1, PX2 and PX3, was
not as large as that for linear polysilanes with comparative
silicon content (Area H in Fig. 18). Branched polysilanes
were also in Area H. Etching selectivity for both cross-
linked polysilanes and branched polysilanes tended to
increase with increasing Si content. The etching selectivity
reached 5.2 for PB3, highly branched polysilanes bearing
methyl groups.
A significantly large deviation was observed for PX1.
They contain a propyl group. The etching selectivity for
PX4, having a cyclohexyl and a naphthyl group, was also
much less than that for PX1, not shown in Fig. 18 because
the etching condition was slightly different from the others
Table 2
n and k values for each layer of multi-layers
Resist Organic ARC Polysilane SiO2 Si
n 1.8 1.5 1.8–2 1.52 1.5
k 0.01–0.03 0.42 0.2–0.4 0.02 3.42
Fig. 18. Relationship between silicon content and etching selectivity of polysilane toward the resist (R1). Etching selectivity: etching rate of
polysilane films/etching rate of resist. Etching gas: C12.
S. Hayase / Prog. Polym. Sci. 28 (2003) 359–381370
shown in Fig. 18. This also supports the fact that alkyl
substituents longer than a methyl group decrease the etching
selectivity.
Highly branched silicon backbones or cross-linking
through ethylene groups increased (Tg)s of polysilanes.
This increase in (Tg)s is needed in order to prevent
polysilane patterns from being deformed thermally during
reactive ion etchings. For example, the Tg of 128 8C for PS3
increased to 155 8C for PX2 by the introduction of partial
cross-linking by ethylene groups. Although the introduction
of the cross-linking decreased the etching selectivity, the
decrease was much less than that brought about by the
introduction of large alkyl groups. Large or bulky alkyl
groups should therefore be avoided. As far as etching
selectivity is concerned methyl groups would be the best
choice, however, phenyl groups are needed to make the
polymers soluble in organic solvents and to fabricate
uniform films.
Fig. 19 shows the relationship between etching selectiv-
ity toward the resist and UV absorbance at 248 nm. If
polysilanes have linear structures and are substituted with
methyl and phenyl groups, the etching selectivity increases
and the UV absorbance decreases with an increase in the
silicon content. Therefore, it is difficult to find a linear
polysilane structure with both a high etching selectivity and
high UV absorbance. UV absorption groups, except for
phenyl groups, have to be incorporated into the polysilane
structure in order to satisfy the high etching selectivity and
high UV absorption criteria. Highly branched polysilanes,
such as PB2, PB3 and PB4, are such candidates.
4.4.2. Oxygen atoms in polysilane backbones
The presence of oxygen atoms in the polysilane structure
was found to be crucial. Etching conditions of polysilane
layers referred to silicon wafer etching conditions. There-
fore, the etching gasses are selected so as to enlarge the
etching selectivity of consecutive silicon atoms against
organic resists. The introduction of other atoms into these
consecutive silicon chains have to be avoided. The most
probable case is the insertion of oxygen atoms into these
chains. In the case of PB1 containing hydrogen groups,
oxidation occurs while the film is baked unless this process
is carried out under argon. Oxidation may occur during
synthesis for all polysilanes. The etching selectivity for PB1
decreased to 80% when 60% of Si–Si bonds were oxidized
to Si–O–Si bonds, and decreased to 40% when all of the
Si–Si bonds were oxidized. PB1 has Si–H groups and is
easily oxidized. The oxidation of polysilanes should be
avoided during their synthesis and the film preparation.
4.4.3. Effects of pressures in an etching chamber
The dependence of etching pressure on etching rates
provides the information on etching mechanisms. As ion
energies are lowered by surrounding molecules, it is
expected that at lower pressures, etching will be dominated
by physical processes, while the etching at higher pressures
emphasizes chemical processes [76]. The etching rate for
the resist film R1 and the PX2 film decreased with
increasing pressure. However, the etching selectivity of
the PX2 film against the resist film R1 increased with an
increase in the pressure as shown in Fig. 20. This implies
that the etching behavior of the PX2 film is similar to that of
the organic resist R1 but the etching of the PX2 film is a
little more pronounced as a result of the chemical reaction,
probably because Si–Si bonds are more chemically reactive
than C–C bonds.
Fig. 19. Relationship between etching selectivity toward resist and UV absorbance at 248 nm. See Fig. 18 for etching conditions.
Fig. 20. Dependence of etching pressure on etching selectivity of
PX2 film against resist film R1. See Fig. 18 for etching conditions.
Etching selectivity: etching rate for PX2/etching rate for R1.
S. Hayase / Prog. Polym. Sci. 28 (2003) 359–381 371
4.4.4. Reaction mechanism
Etching rates of silicon substrates by Cl2 gas have been
discussed in terms of their reactivity with chlorine radicals,
the volatility of reaction products, their reactivity towards
radicals assisted by ion bombardments, and physical
sputtering [89]. The physical sputtering is affected mainly
by the film density. The film density of polysilane films was
almost the same as that of the organic resist film R1, namely,
1.2–1.4 g/cm3. Therefore, we assumed that the etching of
polysilane films by physical sputtering should be almost the
same as that for the resist R1.
Polysilanes have Si–Si, Si–C and C–C bonds, among
which, Si–C and Si–Si bonds are characteristic to
polysilanes. In order to elucidate the etching mechanism,
the polysilane surface was analyzed by X-ray photoelectron
spectroscopy (XPS). The PS6 film has a peak at 101 eV
assigned to the Si 2p for polysilane main chain silicon
atoms. When the PS6 film was exposed to Cl2 gas, the peak
did not change. The peak shifted to 103 eV and became
broader after the S6 film was exposed to Cl2 plasma. The
shift in the binding energy due to one Si–Cl has been
estimated to be about 1.1 eV [89,93]. Therefore, the peak
shift of about 2 eV and peaks overlapped with broad peaks
in the lower energy region could be assigned to silicon
atoms bearing a few Cl groups [89,93]. We are not able to
exclude the possibility that the peak may contain com-
ponents attributed to Si–O linkages which should be formed
as a result of the reaction of Si–Cl with H2O. A similar peak
shift has been observed previously in the case of fluorine and
HBr-etched silicon surfaces when silicon–fluorine species
and silicon–bromine species were formed [90–92]. Actu-
ally, SiCl4 gas was detected by mass spectroscopy in the
exhaust gases from the etching chamber. These results
support the reaction mechanism that during polysilane
etchings, cleavages occur at Si–Si linkages predominantly
to produce Si–Cl groups.
Reaction products have to be removed from film surfaces
swiftly in order to promote dry etchings. Therefore, the
volatility of reaction products has been reported to be one of
the factors determining the dry etching rates [89]. When PS3
is exposed to chlorine plasmas, a number of small molecules
containing one silicon atom, e.g. tetrachlorosilane (Si1),
methyltrichlorosilane (Si2), phenyltrichlorosilane (Si3), and
methylphenyldichlorosilane (Si4) can result. Their boiling
points are 57, 66, 80 8C/10 mm Hg, and 55 8C/1 mm Hg,
respectively. Since etchings were carried out at 80 8C under
a reduced pressure, Si1 and Si2 would be removed
effectively. Si3 and Si4 may need further bond cleavages
for volatilization. The boiling point of propyltrichlorosilane
(Si5) which is characteristic of PX1 decomposition, is
123 8C. This is higher than that of Si2 containing a methyl
group. This may partially explain the phenomena that the
etching rate for polysilanes bearing a propyl or cyclohexyl
group is extremely low. In addition, this phenomena may be
explained by the fact that long alkyl groups bonded to
polysilane backbones are easily cross-linked by ion attacks,
resulting in non-volatile. For poly(di-n-hexylsilane), cross-
linking had been reported to occur by ion beam irradiation,
containing H þ , He þ , and He2 þ ions [94].
Let me explain the reason why polysilane films are
etched faster than organic resist layers by chlorine plasmas.
The anisotropic etching behaviors suggest that sputtering
mechanisms are predominant. When Si radicals formed by
Fig. 21. Supposed etching mechanism.
Fig. 22. Relationship between etching selectivity of polysilane films toward SiO2 layers and Si content. Etching selectivity: etching rate for SiO2
layer/etching rate for polysilane layer; etching gas: C4H8/CO/O2/Ar.
S. Hayase / Prog. Polym. Sci. 28 (2003) 359–381372
sputtering of charged atoms, the Si radicals react with other
Si radicals to form polysilane structures again, or react with
Cl radicals to form Si–Cl as shown in Fig. 21. Because Si
radicals are more stable than C radicals, Si radicals may
have larger chances to react with Cl radicals than C radicals.
In case of C radicals, recombination reactions between two
C radicals to form C–C linkages are more likely than those
between two Si radicals. The fact that the boiling point of
SiCl4 (57 8C) was lower than that of CCl4 (77 8C) may have
something to do with the difference in selective etching
rates.
4.5. Etching durability of polysilanes during SiO2 etching
After polysilane patterns are fabricated, the underlying
SiO2 layer has to be etched by using the patterned polysilane
layer as a hard mask. Mixtures of C4F8, CO, O2 and Ar gases
were used for the etching. The gas mixture was employed
commonly in order to etch SiO2 layers. The C4F8 gas etches
the SiO2 layer. The CO and O2 gases eliminate residues
deposited on the etched surface of the SiO2 layer during the
etching, which retards the swift etching. The C4F8 gas
deposits on polymer layers or Si layers selectively to make
fluoride carbon polymer layers which protect these layers
from the etching. The deposition occurs selectively on
layers whose oxygen content is lower. This is because
fluorocarbon polymers deposit selectively on polymer films
and Si layers [95]. Since polysilane films do not contain
oxygen atoms ideally, high etching durability was expected.
Fig. 22 shows the relationship between Si content and
etching selectivity, where the etching selectivity means the
value of etching rates of SiO2 layers/the etching rate of
polysilanes layers. The etching selectivity defined here is
opposite to that defined in case of polysilane etchings, where
the etching selectivity is defined as etching rates of
polysilane layers/the etching rate of resist layers. The
etching selectivity of conventional polymer anti-reflection
layers is about 5 and that of resists is around 6. The etching
selectivity of polysilane films is higher that that of organic
resist films. Polysilane films are easily etched with chlorine
gas plasmas, however, they are hardly etched with fluorine
gas plasmas. This may be surprising, considering that both
are halogen gases. This difference can be explained by the
selective deposition during the etching mentioned earlier. In
case of polysilane etchings, chlorine gases are used and
deposition does not occur. But, in case of SiO2 etching,
decomposed fluoride gases deposit on polysilane layers to
protect them. As shown in Fig. 21, the etching selectivity
depends on the polysilane structures. The etching selectivity
did not change among PB4, PB5 and PB6 when silicon
content varied. PH2 had higher etching selectivity than the
others. We were not able to elucidate the relationship
between Si content and the etching selectivity from their
limited number of the results. Probably, the etching
mechanism associated with the deposition may make the
explanation difficult. Further works on the relationship
between oxygen content in polysilane films and the etching
selectivity would be needed. Fig. 23 shows the relationship
between the etching selectivity during polysilane etchings
and that during SiO2 etchings. The upper right area in Fig.
23 is the best position. PX2 and PB4 are candidates to fulfill
the requirement.
Fig. 23. Relationship between polysilane etching selectivity toward resist film during polysilane etching and SiO2 etching selectivity toward
polysilane films during SiO2 etching. Etching selectivity during polysilane etching: etching rate for PX2/etching rate for R1; gas: Cl2; etching
selectivity during SiO2 etching: etching rate for SiO2 layer/etching rate for polysilane layer; etching gas: C4H8/CO/O2/Ar.
Table 3
Etching selectivity for anti-reflection layer
Constitution Etching gas Etching selectivity
Conventional ORGARC/resist Fluorocarbon/O2 1.2
This time Polysilane/resist Cl2 2.4
ARC, anti-reflection layer; etching selectivity, etching rate for ARC/etching rate for resist layer.
S. Hayase / Prog. Polym. Sci. 28 (2003) 359–381 373
Tables 3 and 4 summarize the etching selectivity during
anti-reflection layers and that during SiO2 layers. During the
etching of ARCs, polysilane layers are etched more than two
times faster than that of conventional ORGARCs. On the
other hand, polysilane films have durability toward SiO2
layer etchings. Etching rates of polysilane films are 1.2–1.5
times slower than that of conventional organic layers. These
data show the advantages of polysilanes against conven-
tional ORGARCs.
5. Pattern fabrication
5.1. Resist patterns on anti-reflection layers
The effectiveness of the polysilane ARC was verified by
using actual lithography processes [96–98].
PX2 in anisole was spin-coated on the substrate and
baked on a hot plate at 190 8C for 60 s. The resist (R1) for
KrF excimer laser lithography consisting polyvinylphenol
as the main polymer was spin-coated on the polysilane layer
and dried at 100 8C for 90 s. The film was insoluble in
common resist solvents such as ethyl lactate, butyrolactone
and so on. Therefore, any mixing layers between the resist
and the polysilane layer were not observed. The resist film
was exposed by a KrF scanning stepper (numerical aperture
(NA) ¼ 0.6, s ¼ 0:75), followed by developed with a
2.38% tetramethylammoniumhydroxide aqueous solution
(TMAH) using a 60 s puddle process. Fig. 24 shows the
resist patterns on the polysilane PX2 layer. The multi-layer
consists of a resist layer (470 nm)/a polysilane layer
(n ¼ 2.0, k ¼ 0.23) (240 nm)/a SiO2 layer (240 nm)/a Si
substrate. For comparison, the resist pattern on a commer-
cially available organic anti-reflective material (n ¼ 1.55,
k ¼ 0.42) (80 nm) (ORGARC) on the SiO2 layer (150 nm)
is shown in Fig. 25. Good profiles are obtained without any
footing or residue on the polysilane ARCs. Standing waves
are observed on the pattern wall on the ORGARC layer,
however, the resist wall on the PX2 layer was flat.
The anti-reflective performance of the PX2 layer was
simulated and compared with that of the ORGARC layer.
The reflectivity was calculated by an in-house software,
considering the multi-player interference effect. Optical
parameters of stacked films for the calculation are listed
below.
Thickness: resist (470 nm)/ARC/SiO2 (150–300 nm)/Si;
Anti-reflective layer: PX2, 240 nm; ORGARC, 80 nm;
Refractive index at 248 nm. Resist: n ¼ 1:78; k ¼ 0:02;
SiO2: n ¼ 1:52; k ¼ 0:02;
Table 4
Etching selectivity for SiO2 etching
Constitution Etching gas Etching selectivity
Conventional SiO2/ORGARC C4F8/CO/O2/Ar ,5
Conventional SiO2/resist C4F8/CO/O2/Ar ,6
This time SiO2/Polysilane C4F8/CO/O2/Ar ,8
Etching selectivity: etching rate for SiO2 layer/etching rate for ARC layer.
Fig. 24. 0.18 mm L=S resist patterns on polysilane layer. Consti-
tution: resist (470 nm)/PX2 (240 nm) (n ¼ 2.00, k ¼ 0.23)/
SiO2(150 nm). Reprinted with permission from J Vac Sci Technol
1999;17(6):3398. q1999 American Vacuum Society [96].
Fig. 25. 0.18 mm L=S resist patterns on conventional ORGARC.
Constitution: resist (470 nm)/ORGARC(80 nm) (n ¼ 1.55,
k ¼ 0.42)/SiO2(150 nm). Reprinted with permission from J Vac Sci
Technol 1999;17(6):3398. q1999 American Vacuum Society [96].
S. Hayase / Prog. Polym. Sci. 28 (2003) 359–381374
Si: n ¼ 1:50; k ¼ 3:42; PX2: n ¼ 2:00; k ¼ 0:23;
ORGARC: n ¼ 1:55; k ¼ 0:42:
Fig. 26 shows the reflectivity at the resist/ARC interface
when the SiO2 thickness was varied from 150 to 300 nm.
When the thickness of the ARC is thin, the reflectivity
varied, depending on the thickness of anti-reflection and
SiO2 layers. For thicker films, the intrinsic refractive index
governs the total reflectivity because the ARC absorbs all
lights. The intrinsic reflectivity of the ORGARC layer was
1.7% and that of the PX2 film was 0.7%. The reflectivity less
than 1% is required. In the case of the PX2 film, much
thicker film can be employed than that of the ORGARC film
because of the larger etching selectivity. The etching
selectivity of the ORGARC layer was only 0.70. If thicker
ORGARC layers were used as the ARC, organic resist
layers should be lost before the ORGARC layer was etched.
An advantage of polysilane ARCs is that thick films can be
employed because of the larger etching selectivity. The
reflectivity variation for the thick polysilane film is small
even though the thickness of anti-reflection layers and the
underlying SiO2 layers varied during actual manufacturing
processes.
Fig. 27 shows the relationship between the CD variation
and the thickness of SiO2 layers. The CD variation means
the change of pattern widths. The CD variation of the resist
pattern on the polysilane film was 4% which was smaller
than that on the ORGARC film (8%) even though the SiO2
thickness changed.
Polysilane anti-reflection layers exhibit good litho-
graphic performances because the high etching selectivity
toward resist makes it possible to use thick anti-reflection
layers which decrease the deflection itself and increase
latitudes for variations of each layer thickness.
5.2. Patterning of polysilane layers by dry etching
In order to fabricate deep trenches, a fairly thick resist is
needed because of aggressive etching conditions. The thick
resist reduces lithographic process windows. Polysilane ARC
is one candidate to solve this problem. The PX2 was coated on
the substrate and the following constitution was fabricated
R1 resistð430 nmÞ=PX2 filmð240 nmÞ=SiO2 layerð700 nmÞ=
£ SiN layerð220 nmÞ=Si substrate
The deep trench has a twin hole pattern separated by
100 nm lines. The hole width is 300 nm. The SiO2 and SiN
layers act as a hard mask when the Si substrate is etched.
Fig. 28 shows the deep-trench pattern (0.2 mm design rule)
of the resist R1 before etching. Any standing waves on the
resist R1 wall were not observed. The PX2 film was dry-etched
through the resist pattern mask by using Cl2 gas. During
the etching of PX2 film (t: 240 nm), the resist thickness
decreased by 96 nm. The etching selectivity of the PX2 film
Fig. 26. Relationship between reflectivity at resist/ARC interface and thickness of anti-reflection layers. (A) ORGARC, constitution: resist
(470 nm)/ORGARC (80 nm) (n ¼ 1.55, k ¼ 0.42)/SiO2 thickness was varied from 150 to 300 nm. (B) Polysilane anti-reflection layer.
Constitution: resist (470 nm)/PX2 (240 nm) (n ¼ 2.00, k ¼ 0.23)/SiO2. Reprinted with permission from J Vac Sci Technol 1999;17(6):3398.
q1999 American Vacuum Society [96].
Fig. 27. Relationship between CD and SiO2 thickness; Polysilane
anti-reflection layer (PX2). Constitution: resist (470 nm)/PX2
(240 nm) (n ¼ 2.00, k ¼ 0.23)/SiO2 (l50 nm); ORGARC. Consti-
tution: resist (470 nm)/ORGARC (80 nm) (n ¼ 1.55, k ¼ 0.42)/SiO2
(150 nm). Reprinted with permission from J Vac Sci Technol
1999;17(6):3398. q1999 American Vacuum Society [96].
S. Hayase / Prog. Polym. Sci. 28 (2003) 359–381 375
toward the resist R1 film was 3.75 in the deep trench, which
was much higher than that in the flat surface (the etching
selectivity: 2.1). The resist pattern turns out to be transferred to
the polysilane layer with exactly the same dimension as shown
in Fig. 29.
The transfer of the deep trench pattern to the Si substrate
was conducted by a dry etching process with C4H8/CO/O2/Ar
gas mixtures. For comparison, the pattern was transferred by
using conventional ORGARCs. The etching condition was
as follows
R1 resistð430 nmÞ=ORGARC filmð80 nmÞ=
SiO2 layerð700 nmÞ=SiN layerð220 nmÞ=Si substrate
Figs. 30 and 31 show top-down profiles of SiO2/SiN
patterns when PX2 and ORGARC films were employed
for ARCs. The residual resist and ARCs were removed by
etching and striping, respectively. The polysilane layer
was removed by O2 asher followed by washing with a
dilute fluoric acid solution. When the ORGARC layer was
used as the ARC, striations around the twin holes were
observed. The PX2 anti-reflective film improved this
striation. It has been reported that the significant rough-
ness in the resist surface during etchings is responsible for
the striations, because the roughness is transferred to
SiO2/SiN patterns. Polysilane may prevent the resist
roughness from transferred to SiO2/SiN layers [99,100].
An advantage of polysilane ARCs is that thinner resist
layers can be employed. The thinner resists make the
exposure latitude wider.
Fig. 32 shows surface profiles measured by AFM for films
of PB4, PS2 and resist R1 after exposure to C4H8/CO/O2/Ar
gas plasma. While PS2 and resist R1 were denatured, the P1
film exhibited a smoother surface. The peak to valley heights
of PS2, resist 1 and PB4 films were 105, 25.4 and 6.5 nm,
respectively. The phenomena can be explained as follows.
The denatured surface is formed by ion bombardment during
etching, because relatively high rf power is applied for the
etching. The hardness of PB4 was 740 N/mm2, which is much
harder than that of PS2, 260 N/mm2. The hardness brought
about by the tightly cross-linked Q structures should be
responsible for the roughness improvement.
The roles for each part of polysilane molecules are
summarized in Fig. 33. Solubility is controlled by organic
substituents so as to fabricate flat films by spin-coating and
Fig. 28. Resist pattern of a deep trench with a 0.20 mm design rule on PX2 anti-reflection layer. Constitution: resist R1
(430 nm)/PX2(240 nm)/SiO2(700 nm)/SiN(220 nm)/Si. Further explanation: see text. Reprinted with permission from J Vac Sci Technol
1999;17(6):3398. q1999 American Vacuum Society [96].
Fig. 29. Resist and PX2 patterns of a deep trench with a 0.20 mm design rule after the PX2 layer was etched. Constitution: resist
R1(430 nm)/PX2(240 nm)/SiO2(700 nm)/SiN (220 nm)/Si. Etching gas: Cl2, for further explanation: see text. Reprinted with permission from J
Vac Sci Technol 1999;17(6):3398. q1999 American Vacuum Society [96].
S. Hayase / Prog. Polym. Sci. 28 (2003) 359–381376
Fig. 30. Top-down profile of SiO2/SiN patterns when PX2 film was employed for ARCs. Constitution: resist R1(430 nm)/PX2(240 nm)/SiO2
(700 nm)/SiN (220 nm)/Si. Etching gas for PX2: Cl2; etching gas for SiO2/SiN layer: C4F8/CO/O2/Ar. The remaining resist and PX2 layers were
removed with an asher using O2/CF4 gas mixtures. Reprinted with permission from J Vac Sci Technol 1999;17(6):3398. q1999 American
Vacuum Society [96].
Fig. 31. Top-down profile of SiO2/SiN patterns when ORGARC film was employed for ARCs. Constitution: resist R1(430 nm)/ORGARC
(60 nm)/SiO2(700 nm)/SiN(220 nm)/Si. Etching gas for ORGARC: CF4/O2/Ar gas mixtures; etching gas for SiO2/SiN layer: C4F8/CO/O2/Ar.
The remaining resist and ORGARC layers were removed with an asher using O2 gases. Reprinted with permission from J Vac Sci Technol
1999;17(6):3398. q1999 American Vacuum Society [96].
Fig. 32. AMF profiles of various films after C4H8/CO/CO2/Ar gas RIE. (A) PB4, (B) PS2, (C) resist 1. Area measured: 1 mm £ 1 mm. The peak
to valley height of the roughness for PB4, PS2 and resist 1 were 6.5, 105, and 25.4 nm, respectively. Reprinted with permission from Jpn J Appl
Phys 2000;39:6781. q2000 Japan Society of Applied Physics [97].
S. Hayase / Prog. Polym. Sci. 28 (2003) 359–381 377
not to make mixing layers. UV absorptions and etching
selectivity relate to Q units and phenyl silyl groups. High Tg
is brought about by cross-linking and branched structures.
Optimizations of polysilane molecular structures are needed
to cope with all requirements.
6. Conclusion
We discussed the potential application of polysilanes for
LSI fabrication process. Since LSI manufacturing processes
dislike metal impurity, materials employed are restricted.
One of the silicon polymers, polysilanes, should be accepted
for the process without any difficulty. The most interesting
property for polysilanes is that etching rates can be changed
with oxidations of UV-exposures. This makes it possible to
fabricate fine patterns by anisotronic-dry etching without
other resist materials. Bi-layer resists using polysilanes
would become important for deep UV exposures, the
wavelength shorter than 200 nm, because this would
eliminate the problem on penetration depth of the light.
Anti-reflection layer for deep UV lithography is another
interesting application. Polysilanes were found to have a
unique property on dry etchings. Polysilane films are etched
faster than organic resists by using chlorine gas plasmas.
But, the polysilane were etched slower than organic resists
and SiO2 layers by using fluorine gas plasmas. The anti-
reflective property was better than that of conventional
ORGARCs. Advantages for polysilane ARC are that the
resist thickness can be reduced and the ARC thickness can
be increased. Both increased the lithography process
latitude. The etching selectivity and optical properties
make polysilanes the most promising anti-reflection film
for future lithography.
Acknowledgements
The author would like to thank Y. Sato, E. Shiobara,
S. Miyoshi, A. Asano, H. Matsuyama, Y. Onishi, and
T. Ohiwa at Microelectronics Engineering Laboratory,
Toshiba Corporation, and Y. Nakano, S. Yoshikawa,
H. Ohota at Corporate Research and Development Center,
Toshiba Corporation, for their help during this work.
References
[1] West R. The polysilane high polymers. J Organomet Chem
1986;300:327–46.
[2] Fujii A, Yoshimoto K, Yoshida M, Ohmori Y, Yoshino K.
Ultraviolet electro luminescent diode utilizing poly(methyl-
phenylsilane). Jpn J Appl Phys 1955;34:L1365–7.
[3] Bleyl I, Ebata K, Hoshino S, Furukawa K, Suzuki H.
Conformational phase transition in a high-efficiency near-
ultraviolet electro-luminescent diarylpolysilane. Synth Met
1999;105(1):17–22.
[4] Lacave-Goffin B, Hevesi L, Demoustier-Champagne S,
Devaux J. Synthesis and properties of polysilane: versatile
new organic materials. ACH-Models Chem 1999;136(3):
215–36.
[5] Miller RD, Michl J. Polysilane high polymers. Chem Rev
1989;89:1359–410.
[6] Michl J, Downing JW, Karatsu T, McKinley AJ, Poggi G,
Wallraff GM, Sooriyakumaran R, Miller RD. Solution
photochemistry of poly(dialkylsilanes): a new class of
photoresists. Pure Appl Chem 1988;60(7):959–72.
[7] Miller RD, Fickes GN, Hofer D, Sooriyakumaran R, Willson
CG, Guillet JE, Moore J. Polym Mater Sci Engng 1986;55:
599–603.
[8] Rosilio C, Rosilio A, Serre B. Contribution to the study of
polysilanes for high-resolution photolithography. Microelec-
tron Engng 1988;8:55–78.
[9] Miller RD, Hofer D, McKean RD, Willson CG, West R,
Fig. 33. Optimization of polysilane structures.
S. Hayase / Prog. Polym. Sci. 28 (2003) 359–381378
Trefonas IIIPT. Soluble polysilane derivatives in materials
for microlithography. In: Thompson LG, Willson CG,
Frechet JMJ, editors. ACS Symposium Series 266, Washing-
ton, DC: American Chemical Society; 1984. Chapter 3.
[10] Zeigler JM, Harrah LA, Johnson AW. Self-developing
polysilane deep-UV resists. Proc SPIE 1985;539:166–74.
[11] Miller RD, Hofer D, Fickes GN, Willson CG, Marinero E.
Soluble polysilanes: an interesting new class of radiation-
sensitive materials. Polym Engng Sci 1985;539:166–74.
[12] Miller RD, Rabolt JF, Sooriyakumaran R, Fleming W, Fickes
GN, Farmer BL, Kuzmany H. Soluble polysilane derivatives:
chemistry and spectroscopy in inorganic and organometallic
polymers. In: Zeldin M, Wynne KL, Allcock HR, editors.
ACS Symposium Series 360, Washington, DC: American
Chemical Society; 1988. p. 43–60.
[13] Miller RD, Wallraff GM, Clecak N, Sooriyakumaran R,
Michl J, Karatsu T, Makinley AJ, Klingensmith KA,
Downing L. Polysilanes: solution photochemistry and deep-
UV lithography in polymers. In: Reichmanis E, MacDonald
SA, Iwayanagi T, editors. Microlithography: materials and
processes. ACS Symposium Series 412, Washington, DC:
American Chemical Society; 1989. p. 115–32.
[14] Wallraff GM, Miller RD, Clecak N, Baer M. Polysilanes for
microlithography. SPIE 1991;1466:211–7.
[15] Dijkstra HJ, Jufferamans CAH. Optimization of anti-reflec-
tion layers for deep UV lithography. Proc SPIE 1993;1927:
275–86.
[16] Brunner TA. Optimization of optical properties of resist
processes. Proc SPIE 1991;1466:297–308.
[17] Kunz RR, Rothschild M, Ehrlich DJ, Sawan SP, Tsai YG.
Controlled-ambient photolithography of polysilane resist at
193 nm. J Vac Sci Technol B 1989;7:1629–33.
[18] Kunz RR, Horn MW, Goodman RB, Bianconi PA, Smith
DA, Freed CA. Polysilane thin films as resists for deep
ultraviolet lithography. J Vac Sci Technol B 1990;8:1820–5.
[19] Kunz RR, Bianconi PA, Horn MW, Paladugu RR, Shaver
DC, Smith DA, Freed CA. Polysilyne resists for 193 nm
excimer laser lithography. Proc SPIE 1991;1466:218–26.
[20] Horn MW, Pang SW, Rothschild M. Plasma-deposited
organosilicon thin films as dry resists for deep ultraviolet
lithography. J Vac Sci Technol B 1991;8:1493–6.
[21] Kunz RR, Horn MW, Wallraff GM, Bianconi PA, Miller RD,
Goodman W, Smith DA, Eshelman JR, Ginsberg EJ. Surface-
imaged silicon polymers for 193 nm excimer laser lithogra-
phy. Jpn J Appl Phys, Part 1 1992;31:4327–31.
[22] Kunz RR, Horn MW, Wallraff GM, Bianconi PA, Miller RD,
Goodman W, Smith DA, Eshelman JR. Wet-developed
bilayer resists for 193 nm excimer laser lithography. J Vac
Sci Technol B 1992;10:2554–9.
[23] Wallraff MG, Miller RD, Baier M, Ginsburg EJ, Kunz RR.
Polysilane photoresists for 193 and 248 nm lithography.
J Photopolym Sci Technol 1992;5:111–21.
[24] Weidman TW, Joshi AM. New photodefinable glass etch
masks for entirely dry photolithography. Appl Phys Lett
1993;62:372–4.
[25] Joubert O, Joshi AM, Weidmann TW, Lee JT, Taylor GN.
Plasma polymerized organosilane network polymers. Proc
SPIE 1994;2195:358–71.
[26] Horn MW, Maxwell BE, Kunz RR, Hibbs MS, Eriksen LM,
Palmateer SC, Forte AR. All-dry resist processes for 193-nm
lithography. Proc SPIE 1995;2438:760–1.
[27] Weidman TW, Joubert O, Joshi AM, Kostelak RL.
Application of plasma polymerized methylsilane as a resist
and silicon dioxide precursor for 193 nm and 248 nm
lithography. Proc SPIE 1995;2438:496–503.
[28] Kubiak GD, Outka DA. Soft X-ray resist characterization:
studies with a laser plasma X-ray source. Proc SPIE 1990;
1263:272–81.
[29] Kubiak GD. XUV resist characterization: studies with a laser
plasma source. Proc SPIE 1991;1343:283–91.
[30] Yamaguchi A, Ogawa T, Soga T, Tachibana H, Matsumoto
M, Oizumi H, Takeda E. Ablation of Si-containing polymers:
application to X-ray lithography. J Photopolym Sci Technol
1994;7:607–14.
[31] Ogawa T, Yamaguchi A, Soga T, Tachibana H, Matsumoto
M, Oizumi H, Takeda E. New dry surface-imaging process
for X-ray lithography. Jpn J Appl Phys, Part 1 1994;33:
1577–82.
[32] Yamaguchi A, Ogawa T, Tachibana H, Oizumi H, Soga T,
Matsumoto M, Matsuzaka T, Takeda E. Self-developing
characteristics of Si containing polymers and their appli-
cation to X-ray lithography. J Electrochem Soc 1996;143:
657–65.
[33] Horiguchi R, Onishi Y, Hayase S. High molecular weight
polysilanes with phenol moieties. Macromolecules 1988;21:
304–9.
[34] Hayase S, Horiguchi R, Onishi Y, Ushigogouchi T. Synthesis
of polysilanes with functional groups. Macromolecules 1988;
22:2933–8.
[35] Beach JV, Loy DA, Hsiao YL, Waymouth RM. Envir-
onmentally friendly polysilane photoresists. In: Reichmanis
E, Ober CK, MacDonald SA, Iwayanagi T, Nishikubo T,
editors. Microelectronics technology. ACS Symposium
Series 614, Washington, DC: American Chemical Society;
1995. p. 355–66.
[36] Beach JV, Loy DA. Environmentally friendly polysilanes.
Polym Mater Sci Engng 1995;72:205–6.
[37] Kani R, Nakano Y, Yoshida H, Mikoshiba S, Hayase S.
Poly(phenylhydrosilane):base-developable polysilane resist.
J Polym Sci, Polym Chem Ed 1997;35:2355–64.
[38] Kani R, Nakano Y, Yoshida H, Hayase S. Photosensitization
of base-developable poly(phenylhydrosilane) with 3,30,4,40-
tetra (tert-butylperoxycarbonyl)benzene. Macromolecules
1998;31:8794–801.
[39] Tayler GN, Wolf TM. Oxygen plasma removal of thin
polymer films. Polym Engng Sci 1980;20:1087–92.
[40] Hofer DC, Miller RD, Willson CG. Polysilane bilayer UV
lithography. Proc SPIE 1984;469:16–23.
[41] Miller RD, Hofer D, Fickes GN, Willson CG, Trefonas III P,
West R. Soluble polysilanes: an interesting new class of
radiation sensitive materials. Polym Engng Sci 1986;26:
1129–34.
[42] Taylor GN, Hellman MY, Wolf TM. Lithographic, photo-
chemical and O2 RIE properties of three polysilane
copolymers. Proc SPIE 1988;920:274–90.
[43] Miller RD, Wallraff G, Clecak N, Sooriyakumaran R, Mickl
J, Karatsu T, Mckinley AJ, Klingensmith KA, Downing J.
Polysilanes: photochemistry and deep UV lithography.
Polym Engng Sci 1989;29(13):882–6.
[44] Joubert O, Weidman T, Joshi A, Cirelli R, Stein S, Lee JTC,
Vaidya S. Plasma polymerized all-dry resist process for
S. Hayase / Prog. Polym. Sci. 28 (2003) 359–381 379
0.25 mm lithography. J Vac Sci Technol B 1994;122(6):
3909–13.
[45] Yamaguchi A, Tachibana H, Oizumi H, Soga T, Tanaka T,
Ogawa T, Matsumoto M, Takeda E. Negative tone dry
development of Si containing resists by laser ablation. Jpn J
Appl Phys 1995;34:6800–4.
[46] Banovetz JP, Stein KM, Waymouth RM. Stereo-selectivity in
the catalytic oligomerization of phenylsilane. Organometal-
lics 1991;10(10):3430–1.
[47] Turro NJ. Modern molecular photolysis. Menlo Park, CA:
Benjamin/Cummings; 1978. Chapter 8.
[48] Wallraff GM, Baier M, Diaz AF, Miller RD. Fluorescence
quenching of poly(methylphenylsilane) in solution. J Inorg
Organomet Polym 1992;2:87–102.
[49] Miller RD, Baier M, Diaz AF, Ginsburg EJ, Wallraff GM.
The electronic structure of high-molecular-weight substi-
tuted silane polymers in solution. Pure Appl Chem 1992;64:
1291–8.
[50] Wang Y, West R, Yuan CH. Fullerene-doped polysilane
conductor. J Am Chem Soc 1993;115:3844–5.
[51] Kepler RG, Cahill PA. Photoinduced charge transfer and
charge carrier generation in polysilane films containing C60
fullerene molecules. Appl Phys Lett 1993;63:1552–4.
[52] Yoshino K, Yoshimoto K, Hamaguchi M, Kawai T,
Azkhivod AAA, Ueno H, Kakimoto M, Konuma H.
Photoluminescence quenching in polysilanes by fullerene
doping and effective photoinduced charge transfer depending
on aromatic side group. Jpn J Appl Phys 1995;34:L141–4.
[53] Watanabe A, Ito O. Photoinduced electron transfer between
C60 and polysilane studied by laser flash photolysis in the
near-IR region. J Phys Chem 1994;98:7736–40.
[54] Sethi S, Distasio R, Siger D, Lamb J, Flaim T. Use of
antireflective coatings in deep UV lithography. Proc SPIE
1991;1463:30–40.
[55] Azuma T, Matsunaga K, Kawamura D, Mimotogi S, Onishi
Y. Viability of conventional KrF imaging for 150 nm
lithography. J Vac Sci Technol B 1997;15(6):2434–8.
[56] Choi N, Onozawa S, Sakakura T, Tanaka M. Dehydrocou-
pling of polysilane catalyzed by (dimethylamino)alkyl- or
branched-alkyl-substituted Zirconocene complexes. Organo-
metallics 1997;16:2765–7.
[57] Watanabe A, Fujitsuka M, Ito O, Miwa T. Soluble three-
dimensional polysilane with organosilicon nanocluster struc-
ture. Jpn J Appl Phys 1997;36:L1265–7.
[58] Watanabe A, Fujitsuka M, Ito O, Miwa T. Control of silicon
dimensionality of polysilanes and their optical properties.
Mol Cryst Liq Cryst 1998;316:363–6.
[59] Kanai A, Kawakami T, Toyoda E, Ishikawa M. Electro-
chemistry of organosilane compounds. Organometallics
1991;10:2001–3.
[60] Shono T, Kashimura S, Murase H. Electroreductive synthesis
of polygermane and germane–silane copolymer. J Chem
Soc, Chem Commun 1992;896–7.
[61] Chruscel J, Matyjaszewski K. Ring opening of 1,2,3,4-
tetraethyl-1,2,3,4 tetraphenylcyclotetrasilane in the presence
of transition metal catalysts. J Polym Sci, Part A: Polym
Chem 1996;34:2243–52.
[62] Cypryl M, Chrusciel J, Fossum E, Matyjaszewski K. Ring
opening polymerization of strained cyclotetrasilanes as a new
route towards well defined polysilanes. Makromol Chem,
Macromol Symp 1993;73:167–76.
[63] Fossum E, Matyjaszewski K. Ring-opening polymerization
of cyclotetrasilanes. Macromolecules 1995;28:1618–25.
[64] Cypryl M, Gupta Y, Matyjaszewski K. J Am Chem Soc 1991;
113:1046–7.
[65] Suzuki M, Kotani J, Gyobu S, Kaneko T, Saegusa T.
Synthesis of sequence-ordered polysilane by anionic ring-
opening polymerization of phenylnonamethylcyclopentasi-
lane. Macromolecules 1994;27:2360–3.
[66] Sakamoto K, Yoshida M, Sakurai H. Alternating polysilylene
copolymer prepared by anionic polymerization of masked
disilene. Macromolecules 1990;23:4494–6.
[67] Sakurai H, Yoshida M, Sakamoto K. Polymerization of in
situ generated disilenes. J Organomet Chem 1996;521:
287–93.
[68] Trefonas III P, Djurovich PI, Zhang X-H, West R, Miller RD,
Hofer D. Organosilane high polymers: synthesis of formable
homopolymers. J Polym Sci, Polym Lett Ed 1983;21:
819–22.
[69] Li H, Gauvin F, Harrod JF. Observations concerning the
inactivity of dimethylhafnocence as a catalyst for the
dehydrocoupling of phenylsilane. Organometallics 1993;12:
575–7.
[70] Maxka J, Chrusciel J, Sasaki M, Matyjaszewski K.
Polysilanes with various architectures. Macromol Symp
1994;77:79–92.
[71] Matyjaszewski K, Chrusciel J, Maxka J, Sasaki M. Branched
polysilanes from tetrafunctional monomers. J Inorg Organo-
met Polym 1995;5:261–79.
[72] Trefonas III P, West R, Miller RD, Hofer D. Organosilane
high polymers. Electronic spectra and photodegradation.
J Polym Sci, Polym Lett Ed 1983;21:823–9.
[73] Pitt CG, Bursey MM, Rogerson DF. Catenates of the Group
IV elements. Correlation of s-electron energies. J Am Chem
Soc 1970;92:519–22.
[74] Kumada M, Tamao K. Aliphatic organopolysilanes. In: Stone
FGA, West R, editors. Advances in organometallic chem-
istry. New York: Academic Press; 1968. p. 80.
[75] Pitt CG, Carey RN, Toren Jr. EC. Nature of the electronic
interactions in aryl-substituted polysilanes. J Am Chem Soc
1972;95:3806–11.
[76] Harrah LA, Zeigler JM. Rod-to-coil transition in solutions of
poly(di-n-hexylsilane). J Polym Sci, Polym Lett Ed 1985;23:
209–11.
[77] Harrah LA, Zeigler JM. Electronic spectra of polysilanes.
Macromolecules 1987;2:601–8.
[78] Miller RD, Hofer D, Rabolt J, Fickes GN. Anomalously long-
wavelength transition in conformationally locked organosi-
lane high polymers. Am Chem Soc 1985;107:2172–4.
[79] Trefonas III P, Damewood JR, West R, Miller RD.
Organosilane high polymers. Thermochromic behavior in
solution. Organometallics 1985;4:1318–9.
[80] Zhang X-H, West R. Organopolysilane polymers: formable
copolymers containing dimethylene units. J Polym Sci,
Polym Chem Ed 1984;22:159–70.
[81] Bianconi RA, Schilling FC, Weidman TW. Ultrasound-
mediated reductive condensation. Synthesis of silicon–
silicon bonded network polymers. Macromolecules 1989;
22:1697–704.
[82] Furukawa K, Fujino M, Matsumoto N. Optical properties of
silicon network polymers. Macromolecules 1990;23:
3423–6.
S. Hayase / Prog. Polym. Sci. 28 (2003) 359–381380
[83] Sekuguchi A, Nanjo M, Kabuto C, Sakurai H. Polysilane
dendrimer. J Am Chem Soc 1995;117:4195–6.
[84] Suzuki H, Kimata Y, Satoh S, Kuriyama A. Polysilane
dendrimer. Chem Lett 1995;293–4.
[85] Miller RD, Sooriyakumaran R. The synthesis and character-
ization of the first soluble, high molecular weight, substituted
polydiphenylsilane homopolymers. J Polym Sci, Part C:
Polym Lett 1987;25:321–6.
[86] Harrah LA, Zeiglar JM. Rod-to-coil transition of solutions of
poly(di-n-hexyl silane). Macromolecules 1987;20:2037–9.
[87] Suda Y, Motoyama T, Harada H, Kanazawa M. A new
antireflective layer for deep UV lithography. Proc SPIE 1992;
1674:350–61.
[88] Flamm DL, Donnelly VM. The design of plasma etchants.
Plasma Chem Plasma Proc 1981;1:317–63.
[89] Bestwick TTD, Oehrlein GS. Reactive ion etching of silicon
using bromine containing plasma. J Vac Sci Technol A 1990;
8(3):1696–701.
[90] Grunz KJ, Ley L, Johnson RL. Photoelectron spectra of
fluoride-doped amorphous silicon. Phys Rev B 1981;24:
2069–80.
[91] Bakke AA, Chen H-W, Jolly WL. A table of absolute core-
electron binding-energies for gases atoms and molecules.
J Electron Spectrosc Relat Phenom 1980;20:333–66.
[92] Bestwick TD, Oehrlein GS. Reactive ion etching of silicon
using bromine containing plasmas. J Vac Sci Technol A
1990;8:1696–701.
[93] Briggs D. In: Clarke DR, Suresh S, Ward IM, editors. Surface
analysis of polymers by XPS and static SIMS. Cambridge,
UK: Cambridge University Press; 1998.
[94] Seki S, Kanzaki K, Kunimi Y, Tagawa S, Yoshida Y, Kudoh
H, Sugimoto M, Sasuga T, Seguchi T, Shibata H. LET effects
of ion beam irradiation on poly(di-n-hexylsilane). Radiat
Phys Chem 1997;50:423–7.
[95] Coburn JW, Winters HF, Chuang TJ. Ion-surface interactions
in plasma etching. J Appl Phys 1977;48:3532–40.
[96] Sato Y, Shiobara E, Miyoshi S, Asano M, Matsuyama H,
Onishi Y, Nakano Y, Hayase S. Novel antireflective layer
using polysilane for deep ultraviolet lithography. J Vac Sci
Technol 1999;17(6):3398–401.
[97] Sato Y, Shiobara E, Abe J, Onishi Y, Yoshikawa S, Nakano
Y, Hayase S. Application of organic silicon clusters to
pattern transfer process for deep UV lithography. Jpn J Appl
Phys 2000;39:6781–5.
[98] Onishi Y, Sato Y, Shiobara E, Miyoshi S, Matsuyama H, Abe
J, Ichinose H, Ohiwa T, Nakano Y, Yoshikawa S, Hayase S.
Application of polysilanes for deep UV antireflective coat-
ing. Proc SPIE 1999;3678:205–12.
[99] Ramanathan V, Chen S, Lai K, Brongo M, Samarakone N.
Etch integration issues in development of deep sub-micron
contacts utilizing DUV resists and organic BARC. Proc SPIE
1998;333:909–15.
[100] Vinet F, Pazaniz PJ. Micro-patterning for advanced inter-
connections. J Photopolym Sci Technol 1996;9:541.
S. Hayase / Prog. Polym. Sci. 28 (2003) 359–381 381