Hatano Y., Katsumura Y., Mozumder A. (Eds.) Charged Particle and Photon Interactions with Matter - Recent Advances, Applications, and Interfaces

Подождите немного. Документ загружается.

720 Charged Particle and Photon Interactions with Matter

to be scavenged at acid yields of approximately 5. The observation of abnormality of some diphe-

nyliodonium salts agrees with that in previous studies using γ-ray (Nagahara et al., 2000; Nakano

etal., 2005), EUV (Brainard et al., 2008; Hirose et al., 2008a), and deep UV (Ablaza et al., 2000)

as exposure tools. Side reactions induced by reactive decomposition products have been suggested

the reason for the high efciency of DPI-tf.

The

parameter has been estimated by pulse radiolysis for acid generators. Figure 26.5 shows

representative kinetic traces of the decay of tetrahydrofuran-(THF) solvated electrons with and

without triphenylsulfonium tris(triuoromethansulfonyl)methane (TPS-tsm) (Natsuda et al., 2009).

The decrease in the initial yield and the increase in the decay rate were observed with increasing

TPS-tsm concentration. The increase in the decay rate indicates the reaction of acid generators with

solvated electrons. The rate constants for the reactions of acid generators with solvated electrons

have been used to estimate the effective reaction radii of acid generators (Kozawa et al., 2004).

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0 0.2 0.4 0.6 0.8 1 1.2

Acid generator concentration (M)

Yield of C6 proton adduct (M)

TPS-tf TPS-nf

TPS-PF TPS-Sb

TPS-cpdi TPS-tsm

DPI-tf DPI-nf

BBI-tf BBI-cpdi

BBI-tsm NAI-tf

NDI-tf NDP

Figure 26.4 Relationship between the concentration of acid generators and the yield of C6 proton adducts

(acids).

(From Natsuda, K. et al., Jpn. J. Appl. Phys., 48, 06FC05, 2009. With permission.)

0.2

0.4

0.6

0.8

0 200 400 600 800 1000

Time (ps)

Optical density

0 mM

10 mM

30 mM

50 mM

100 mM

Figure 26.5 Kinetic traces at 1300 nm in a TPS-tsm solution in THF. The concentrations of TPS-tsm are

0, 10, 30, 50, and 100mM. (From Natsuda, K. et al., Jpn. J. Appl. Phys., 48, 06FC05, 2009. With permission.)

Radiation Chemistry of Resist Materials and Processes in Lithography 721

The radii estimated from the reaction with solvated electrons did not have sufcient accuracy to

explain the difference between acid generators as described above. This is because the acid gen-

erators react not with solvated electrons but with (unsolvated) thermalized electrons in chemically

amplied

resists upon exposure to ionizing radiation.

The

decrease in the initial yield is expressed by the C

parameter. Hunt and coworkers (Wolff

et al., 1970, 1975; Aldrich et al., 1971; Lam and Hunt, 1975; Hunt and Chase, 1977) focused on

the decrease in the initial yield of hydrated electrons in water upon addition of various solutes and

demonstrated that the fraction of decrease, f, is expressed by an exponential as a function of solute

concentration:

−













exp

[ ]

(26.8)

where

[S]

is the concentration of the scavenger

is a constant that depends on the solvent and the solute

The C

parameter can be understood as the solute concentration at which the initial yield of

solvated electrons is decreased to 1/e (37%). C

is useful when the concentration is rather high.

When an electron pulse enters the THF solution, THF molecules are ionized and secondary

electrons are emitted. The emitted electrons lose their kinetic energy through their interaction

with surrounding molecules and are nally thermalized. The thermalized electrons are solvated

with THF molecules and become detectable, as shown in Figure 26.5. Considering the time reso-

lution of the pulse radiolysis system used, the initial decrease can be assumed to occur before

solvation. The C

parameter is believed to reect the reactions with electrons before solvation

(Hamill, 1969; Lam and Hunt, 1975; Hunt and Chase, 1977; Jonah et al., 1977, 1989; Lewis and

Jonah, 1986; Chernovitz and Jonah, 1988; Saeki et al., 2007). In THF, presolvated states, such

as an excited state of solvated electrons, have not been reported. Therefore, C

in THF is likely

related to the reaction with thermalized and possibly epithermal electrons. Figure 26.6 shows

the dependence of the initial yield of THF-solvated electrons on the TPS-tsm concentration

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0 0.05 0.1 0.15 0.2

TPS-tsm concentration (M)

Optical density

Figure 26.6 Relationship between the initial yield of THF-solvated electrons and TPS-tsm concentration.

(From

Natsuda, K. et al., Jpn. J. Appl. Phys., 48, 06FC05, 2009. With permission.)

722 Charged Particle and Photon Interactions with Matter

(Natsuda et al., 2009). The C

parameter can be calculated from the slopes in Figure 26.6. It

has been reported that the C

parameter does not correlate to the rate constant for the reaction

with solvated electrons (Natsuda et al., 2008). Figure 26.7 shows the relationships between the

parameter and the acid yield. The acid yield was evaluated at an acid generator concentra-

tion of 0.4 M. A good correlation was observed between the C

parameter and the acid yield,

although chemical reactions in the solid phase are generally different from those in the liquid

phase. Because the thermalization of secondary electrons is completed within 0.1–1ps (Rassolov

and Mozumder, 2001), this thermalization and the immediate reactions of the secondary elec-

trons before solvation are unlikely to be signicantly affected by the translation and rotation of

molecules even in the liquid phase. The C

parameter is a good indicator for estimating the reac-

tivity of acid generators with thermalized electrons generated in resist lms. Also, the observed

correlation demonstrated the validity of previously reported reaction mechanisms of chemically

amplied resists (Kozawa et al., 1992).

In EB lithography, a 2–100keV EB is generally used. A single-spur model (Kozawa et al., 2004),

which assumes that each pair does not interact, has been reported to closely reproduce the acid

yields generated in chemically amplied resists upon exposure to a 75keV EB (Kozawa et al.,

2006b, 2007a). The acid generation efciency per ionization is calculated to be 0.74 in a PHS lm

with 10wt% TPS-tf (Kozawa et al., 2005). The single-spur model for chemically amplied EB

resists was constructed on the basis of the fact that it closely reproduces the dynamics of interme-

diate species generated in the model solutions of resist materials upon exposure to a 28MeV EB

(Saeki

et al., 2002; Okamoto et al., 2003, 2007).

The

thermalized electrons migrate in the resist matrix under electric elds produced by charged

species within the Onsager length. Acid generators react with these electrons and release anions.

The mobile electrons are replaced with immobile anions, which act as counteranions of the acid.

40 50 60 70 80 90 100

(mM)

Acid yield (mM)

O S

t-Bu

-C

SbF

t-Bu

-N

t-Bu

OSO

Figure 26.7 Relationship between the C

parameter and the acid yield generated in PHS lms with

5wt% C6 and 0.4 M acid generators. (From Natsuda, K. et al., Jpn. J. Appl. Phys., 48, 06FC05, 2009. With

permission.)

Radiation Chemistry of Resist Materials and Processes in Lithography 723

The calculated distribution of counteranions is shown in Figure 26.8 (Kozawa et al., 2009). The

distribution calculated using the single-spur model is also shown in Figure 26.8. The integral of the

graph indicates the acid generation efciency, which decreased when multispurs were taken into

account. The relationship between the average number of overlapped ion pairs and acid generation

efciency (per ionization) was calculated as shown in Figure 26.9. The relationship with the resolu-

tion blur is also shown in Figure 26.9. The acid generation efciency decreases with the increase

in the average number of overlapped ion pairs because of the strong electric elds generated by

multiple cations and the cross recombination. The strong electric elds promote the recombination

between cations and electrons thermalized near cations and decrease the probability of the reaction

of acid generators with electrons. Similarly, the resolution blur increases with the average number

of ion pairs. The acid generation efciency was calculated to be 0.70 per ionization in EB resists at

10wt% TPS-tf loading (0.74 for the single-spur model). The resolution blur was 5.8nm (5.6nm for

the single-spur model).

In nanoscale patterning using a highly sensitive resist, such as a chemically amplied resist,

the statistical effect is not negligible. The fact that ion pairs overlap indicates an increase in the

inhomogeneity of the distribution of intermediate species. However, the long electron migration

in the subexcitational energy region moderates this inhomogeneity, although it induces resolution

blur (Kozawa and Tagawa, 2006). The subsequent proton migration also contributes to reducing the

inhomogeneity

of the acid distribution (Saeki et al., 2006).

26.5 deprotonation oF radiCal Cations

26.5.1 dynaMicS of phS radical cation (phS

•+

)

When onium salts react with low-energy electrons generated by ionization, they decompose into a

neutral radical and an anion. This decomposition path of onium salts is similar to that induced by

the electron transfer from the excited state of polymers observed in photoresists in KrF excimer

laser lithography (Hacker et al., 1992; Cameron et al., 2001). The decomposed products of neutral

triphenylsulfonium radicals were suggested by Dektar and Hacker (1990a). Considering that the

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0 10 20 30

Distance (nm)

Anion generation probability per ionization (nm

–1

)

Single-spur model

Multispur model (EB)

Multispur model (EUV)

Figure 26.8 Distribution of probability of anion generation per spherical-shell thickness (per ionization).

The horizontal axis represents the distance from the rst ionization point (the origin). (From Kozawa, T. et al.,

Jpn. J. Appl. Phys.,

48, 056508, 2009. With permission.)

724 Charged Particle and Photon Interactions with Matter

decomposed products are diphenylsuldes and benzenes (phenyl radicals), it is clear that protons are

not generated from onium salts in this reaction path. Instead, protons are generated from the radi-

cal cations of polymers (Kozawa et al., 1997). Figure 26.10 shows the transient absorption spectra

obtained in the pulse radiolysis of 100mM of PHS-benzonitrile solution, monitored at the pulse end,

and 200 and 400 ns after the electron pulse (Nakano et al., 2006b). Note that the spectrum at the

pulse end largely overlapped with the transient absorption spectra of solvent intermediates (Tagawa

et al., 1972). An absorption peak was observed at 390nm and 200 and 400ns after the pulse. The

absorption at 390nm does not decay within the observed time range. The absorption from 400 to

550nm shows a slight decay. Comparing the absorptions at 200 and 400ns, this decaying compo-

nent is considered to have an absorption maximum at 430–450nm. By adding triethylamine (TEA)

to the sample, changes in transient spectra were observed, as shown in Figure 26.11 (Nakano et al.,

2006b). Although the absorption intensity decreased, the kinetics at 390nm did not change from

200 to 400nm after the pulse. The decaying component observed between 400 and 550nm almost

disappeared

with TEA addition.

Although

the transient absorption spectra of PHS intermediates have not yet been reported, the

radiation chemistry of phenol derivatives has been well studied. Phenoxy radicals, which are stable

decomposition intermediates of phenols, were rst observed in the 1950s (Porter and Wright, 1955).

Land

et al. reported that phenoxy radicals are very weak bases (Land et al., 1961) and show a char-

acteristic

absorption band at 400 nm (Land and Ebert, 1967). The absorption maxima of phenoxy

0 1 2 3 4 5

0.0

0.2

0.4

0.6

0.8

1.0

0 1 2 3 4 5

Average number of ion pairs(a)

(b)

Acid generation eciency

per ionization

Average number of ion pairs

Resolution blur (nm)

Figure 26.9 Relationship of average number of ion pairs in an isolated space with (a) acid generation ef-

ciency (per ionization) and (b) resolution blur. (From Kozawa, T. et al., Jpn. J. Appl. Phys., 48, 056508, 2009.

With

permission.)

Radiation Chemistry of Resist Materials and Processes in Lithography 725

radicals of phenol and p-cresol in water are 400 and 405 nm, respectively (Land and Porter, 1963;

Land and Ebert, 1967). Radical cations of phenol derivatives, which are precursors of phenoxy

radicals in many cases, have been more elusive than phenoxy radicals (Land et al., 1961; Dixon and

Murphy, 1976, 1978; Holton and Murphy, 1979; Maruyama et al., 1986; Tripathi and Schuler, 1987;

Shoute and Neta, 1990; Kesper et al., 1991; Brede et al., 1996). The absorption maxima of radical

cations of phenol and p-cresol are 440 and 430nm, respectively (Ganapathi et al., 2000). These

reports are consistent with the results for PHS. The absorption at 390nm and the decaying compon-

ent

were assigned to the phenoxy radical and the radical cation of PHS, respectively.

The

deprotonation of radical cations of phenol derivatives has been investigated. Dixon and

Murphy reported that the pK

values of radical cations of phenol and p-cresol were −2.0 and −1.6,

Δ O.D.

350 400 450 500 550 600

Pulse end

200 ns

400 ns

0.1

0.2

0.3

0.4

0.5

0.6

Wavelength (nm)

Figure 26.11 Transient absorption spectra obtained in pulse radiolysis of 100mM of PHS solution in ben-

zonitrile with 2.5 mM TEA. The sample was deaerated by SF

bubbling before irradiation. (From Nakano, A.

al., Jpn. J. Appl. Phys., 45, 6866, 2006b. With permission.)

Δ O.D.

350 400 450 500 550 600

Pulse end

200 ns

400 ns

0.1

0.2

0.3

0.4

0.5

Wavelength (nm)

Figure 26.10 Transient absorption spectra obtained in pulse radiolysis of 100mM of PHS solution in

benzonitrile. The sample was deaerated by SF

bubbling before irradiation. (From Nakano, A. et al., Jpn.

J.Appl. Phys., 45, 6866, 2006b. With permission.)

726 Charged Particle and Photon Interactions with Matter

respectively (Dixon and Murphy, 1978). Also, Bordwell and Cheng reported that those of phenol and

p-cresol were −8.1 and −7.1, respectively (Bordwell and Cheng, 1991). The radical cation of 2,4-dime-

thoxyphenol was reported to be deprotonated by 2,6-lutidine at a rate constant of 6.5 × 10

−1

in an acetonitrile solution (Gadosy et al., 1999). The radical cation of phenol is deprotonated by

ethanol at a rate constant of 9.16 × 10

−1

(Ganapathi et al., 2001). This is in contrast to the

reaction of the poly(4-methoxystyrene) (PMS) radical cation with ethanol, as discussed in the fol-

lowing text. For p-cresol, the absorption coefcients of the phenoxy radical and the radical cation

are 2.30 × 10

and 2.88 × 10

M cm

−1

at each peak wavelength (Ganapathi et al., 2000). For the

other phenol derivatives, the absorption coefcients of radical cations observed at 400–600 nm

are generally larger than those of phenoxy radicals observed at 400–500 nm. Therefore, the radi-

cal cation of PHS is considered to be mostly deprotonated 200 ns after the pulse. Considering that

ethanol and p-cresol have similar proton afnities (PAs), the efcient deprotonation of PHS is

probably due to the ion-molecular reaction between base units.

26.5.2 dynaMicS of pMS radical cation (pMS

•+

)

PMS is a model compound of a typical resist polymer (partially protected PHS). The pulse radi-

olysis of a PMS solution in dichloromethane has been reported. Figure 26.12 shows the transient

absorption spectra obtained in the pulse radiolysis of 100 mM of PMS solution in dichlorometh-

ane deaerated by Ar bubbling (Nakano et al., 2006b). The transient absorption spectra shown in

Figure 26.12 have a peak at approximately 460 nm. The peak did not decay within the observed

time range. The transient absorption spectra of PMS have not been reported. However, the dynam-

ics of radical cations of methoxylated benzenes, such as methoxybenzene (anisole), methoxytolu-

ene, and dimethoxybenzene, have been intensively investigated using pulse radiolysis (O’Neill

et al., 1975; Holcman and Sehested, 1976; Takamuku et al., 1989; Jonsson et al., 1993), pho-

tolysis (Ito et al., 1989; Baciocchi et al., 1993), chemical oxidations (Forbes and Sullivan, 1966;

Sullivan and Brette, 1975; Schlesener and Kochi, 1984), and electrochemical methods (Ronlan

et al., 1974; Buck and Wagoner, 1980). Radical cations of p-methoxytoluene are observed at 440

(Baciocchi et al., 1993) and 450nm (Ito et al., 1989) in acetonitrile and at 460 nm in 1,2-dichlo-

roethane (Takamuku et al., 1989), and those of methoxybenzene are observed at 430 nm in water

(O’Neill et al., 1975; Holcman and Sehested, 1976; Jonsson et al., 1993). Also, in an aromatic

Δ O.D.

400 450 500 550 600

Pulse end

200 ns

400 ns

0.1

0.2

0.3

Wavelength (nm)

Figure 26.12 Transient absorption spectra obtained in pulse radiolysis of 100mM of PMS solution in

dichloromethane.

(From Nakano, A. et al., Jpn. J. Appl. Phys., 45, 6866, 2006b. With permission.)

Radiation Chemistry of Resist Materials and Processes in Lithography 727

compound solution in a halocarbon, a π-molecular complex forms through a reaction with a radical

cation of aromatic compounds with a chlorine anion. The absorption maxima of the π-molecular

complex band of methoxybenzene are 509 nm in dichloromethane (Zweig, 1963) and 524nm in

carbon tetrachloride (Raner et al., 1989). As for methoxytoluene, Takamuku et al. reported that

the π-molecular complex of a p-isomer is not observed in the visible wavelength region, although

those of o- and m-isomers are observed at 550 and 490nm, respectively (Takamuku et al., 1989).

To clarify the assignment of the absorption at 460 nm (Figure 26.12), TEA was added to the solu-

tion. TEA is a well-known cation scavenger. Figure 26.13 shows the transient absorption spectra

of the solution with 10 mM of TEA and the kinetic curves of absorption at 460nm at several

TEA concentrations. The decay of the absorption curve became faster with an increase in the TEA

concentration. The observed rate constants are plotted in the inset of Figure 26.13b. These rate

constants show a linear dependence on the TEA concentration. The bimolecular rate constant cal-

culated from the slope is 1.3 × 10

−1

. By analogy with previous research results mentioned

0.1

0.2

0.3

100 200 300 400 500

400 450 500 550 600

Pulse end

200 ns

400 ns

Wavelength (nm)

Kinetic curves

Transient spectra

(a)

(b)

∆ O.D. ∆ O.D.

0.1

0.2

0.3

Time (ns)

1050

Concentration (mM)

Rate constant

(×10

–1

)

0 mM

2.5 mM

5.0 mM

10.0 mM

Figure 26.13 (a) Transient absorption spectra obtained in pulse radiolysis of 100 mM of PMS solution

in dichloromethane with 10mM TEA and (b) changes in kinetic curves at 460 nm induced by TEA addition

(0, 2.5, 5.0, and 10.0 mM). The inset shows the relationship between the observed rate constant and the

concentration of TEA. (From Nakano, A. et al., Jpn. J. Appl. Phys., 45, 6866, 2006b. With permission.)

728 Charged Particle and Photon Interactions with Matter

above and from our experimental results, we conclude that the absorption band at 460 nm is

ascribed not to the π-molecular complex but to the radical cation of PMS.

The deprotonation of radical cations of methoxylated benzenes has been investigated (Schlesener

and Kochi, 1984; Baciocchi et al., 1993). The pK

of the dimethoxytoluene radical cation is 0.45 in

acetonitrile (Baciocchi et al., 1993), which is signicantly higher than those of phenol and p-cresol.

Dimethoxytoluene radical cations are deprotonated at a rate constant of 3.0 × 10

−1

by 2,6-lutidine

(Baciocchi et al., 1993), whose PA is 955kJ mol

−1

(Lias et al., 1984). This is two orders lower than that

for the reaction of 2,4-dimethoxyphenol radical cations with 2,6-lutidine. It is important to determine

whether the PMS radical cation is deprotonated in the absence of strong proton acceptors, such as

amine. Using ethanol as a proton acceptor, the deprotonation of the PMS radical cation has been inves-

tigated (Nakano et al., 2006b). Ethanol has a PA (788kJ mol

−1

) similar to that of the oxygen atom of

p-cresol (PA = 756kJ mol

−1

) (van Beelen et al., 2004), which is a model compound of PHS. By adding

ethanol to 100mM of PMS-dichloromethane solution, it was examined whether PMS

•+

is deproton-

ated. However, no evident decay was observed within the observed time range. This means that PMS

•+

does not release a proton, at least spontaneously. Also, because the PA of p-methoxytoluene is

801kJ mol

−1

(van Beelen et al., 2004), the deprotonation through ion-molecular reaction with a neigh-

boring unit is considered to be ineffective. This is consistent with the experimental result that PMS

•+

hardly decays within the observed time range, as shown in Figure 26.12. It has been conrmed that the

acid yield in the PHS lm with TPS-tf was about two times higher than that in the PMS lm (Nakano

et al., 2006b). The decrease of the acid yield agrees with the difference in deprotonation dynamics

between PMS and PHS radical cations observed in the pulse radiolysis experiments.

On the basis of the experiments in solutions, the deprotonation mechanism of PHS is suggested

follows:

MOH MOH e

→ +

−

(26.9)

MOH MOH MO MOH



+ → +

(26.10)

Here, MOH, MOH

•+

, MO

•

, and MOH

represent a PHS molecule, its radical cation, its phenoxy

radical, and its proton adduct, respectively. This deprotonation mechanism is similar to that of

novolak (Kozawa et al., 1997). The PHS radical cation generated by ionization is deprotonated by

inter- and intramolecular ion-molecular reactions. The reaction mechanism in the solution may be

different from that in the solid lm because of the different conditions of polymers. However, most

PHS hydroxyl groups form inter- and intramolecular hydrogen bonds with a neighboring hydroxyl

group in solid resist lms (Li and Brisson, 1998; Singh et al., 2005). Therefore, reaction (26.10) is

considered

to efciently take place even in solid lms.

26.5.3 pMMa

Upon exposure to ionizing radiation, PMMA molecules are either ionized or excited. For keV

electrons, the cross section of ionization is larger than that for excitation. Through ionization,

a PMMA radical cation and an electron with excess energy are generated. The ejected electron

reacts with PMMA molecules or recombines with its parent radical cation after losing a suf-

cient amount of energy. A PMMA anion radical is formed through the reaction of PMMA with

an electron (Ogasawara et al., 1987; Nakano et al., 2004a). An excited state of PMMA is also

formed through the recombination of an electron with its parent radical cation. The radical cat-

ions (Ichikawa and Yoshida, 1990), radical anions (Tabata et al., 1983; Sakai et al., 1995), and

excited PMMA molecules (Fox et al., 1963; Gupta et al., 1980; Torikai et al., 1990) are reported

to decompose. In either case, the side chain is rst detached and a macroradical is formed.

This macroradical results in β-scission. During the decomposition, protons are considered to be

Radiation Chemistry of Resist Materials and Processes in Lithography 729

generated. It is known that PMMA becomes acidic upon irradiation. Proton generation has been

observed in PMMA lms without an acid generator. The G-value of acid generation is 1.4 with-

out any acid generators (Nakano et al., 2005). Although the counteranion has not been identied,

it is possibly a formic (Tatro et al., 2003) or acetic (Gupta et al., 1980) anion. The main-chain

scission of PMMA is also induced by UV light (Fox et al., 1963; Gupta et al., 1980; Torikai

et al., 1990). PMMA decomposes through the n-π* singlet state. Upon exposure to ionizing

radiation, the contribution of the excited state of PMMA is, however, low because the recombi-

nation reaction (the main generation path for the excited state) is suppressed due to the reaction

of electrons with PMMA. In the presence of acid generators, the generation of an excited state

through a recombination reaction is suppressed even more because acid generators are strong

electron scavengers. The decomposition of PMMA radical anions also leads to main-chain scis-

sion (Sakai etal., 1995). However, the efciency has not been reported. If the decomposition

of PMMA anion radicals is effective, this path causes signicant resolution blur, because the

ejected electrons migrate in the PMMA matrix before they are trapped by PMMA. However,

PMMA is known as the highest-resolution resist. The ultimate resolution of the PMMA resist has

been intensively investigated and is reported to be 3–5 nm (Chen and Ahmed, 1993; Cumming

et al., 1996; Yasin et al., 2001). Therefore, PMMA radical anions are unlikely to decompose

effectively. This is also consistent with the fact that PMMA radical anions are observed to be

relatively stable (Nakano et al., 2004a).

PMMA radical cations, which are obviously a primary product brought about by ionization,

have not been observed even at 4.2K (Tanaka et al., 1990). The identication of PMMA radical

cations was previously reported (Tabata et al., 1983). They later turned out to be PMMA radical

anions (Ogasawara et al., 1987). This elusiveness of PMMA radical cations is widely believed to

be due to their short lifetime. Although the decomposition path has not been completely clari-

ed, some intermediates and products have been identied. Besides PMMA radical anions, neutral

radical species such as main-chain radical –C

•

H–, side-chain radical –COOC

•

–, and β-scission

products –CH

–C

•

(CH

)COOCH

have been identied (Geuskens and David, 1973). It has been

widely accepted that the main-chain radical leads to main-chain scission. The G-values of main-

chain scission reported by many researchers were scattered. The average of the reported values

is 1.9 ± 0.3 (Chapiro, 1962). The main-chain radical is generated through the detachment of the

side chain. Among these reported neutral radicals, the side-chain radical is a direct product gener-

ated through the deprotonation of PMMA radical cations. The G-value of –COOC

•

– has been

reported to be approximately 2 (Tanaka et al., 1990). The G-value of triuoromethanesulfonic acids

in PMMA with 10wt% TPS-tf is shown in Figure 26.14 (Kozawa et al., 2007a). The G-value of

G-value (molecules/100 eV)

Acid generator concentration (M)

0.0

0.5

1.0

1.5

2.0

2.5

0.0 0.1 0.2 0.3

Figure 26.14 Dependence of the G-value of acid generation in PMMA on acid generator concentration.

(From

Kozawa, T. et al., J. Photopolym. Sci. Technol., 20, 577, 2007a. With permission.)