Study of a two-stage photobase generator for photolithography in microelectronics

Article abstract of DOI:10.1021/jo302149u

The investigation of the photochemistry of a two-stage photobase generator (PBG) is described. Absorption of a photon by a latent PBG (1) (first step) produces a PBG (2). Irradiation of 2 in the presence of water produces a base (second step). This two-photon sequence (1 + hν → 2 + hν → base) is an important component in the design of photoresists for pitch division technology, a method that doubles the resolution of projection photolithography for the production of microelectronic chips. In the present system, the excitation of 1 results in a Norrish type II intramolecular hydrogen abstraction to generate a 1,4-biradiacal that undergoes cleavage to form 2 and acetophenone (Φ ～ 0.04). In the second step, excitation of 2 causes cleavage of the oxime ester (Φ = 0.56) followed by base generation after reaction with water.

Full text of DOI:10.1021/jo302149u

Article
pubs.acs.org/joc
Study of a Two-Stage Photobase Generator for Photolithography
in Microelectronics
Nicholas J. Turro,*,† Yongjun Li,† Steffen Jockusch,† Yuji Hagiwara,‡ Masahiro Okazaki,‡ Ryan A. Mesch,‡
David I. Schuster,§ and C. Grant Willson*,‡
†Department of Chemistry, Columbia University, New York, New York 10027, United States
‡Chemistry Department, University of Texas, Austin, Texas 78741, United States
§Department of Chemistry, New York University, New York, New York 10003, United States
S
* Supporting Information
ABSTRACT: The investigation of the photochemistry of a
two-stage photobase generator (PBG) is described. Absorption
of a photon by a latent PBG (1) (first step) produces a PBG
(2). Irradiation of 2 in the presence of water produces a base
(second step). This two-photon sequence (1 + hν → 2 + hν →
base) is an important component in the design of photoresists
for pitch division technology, a method that doubles the resolu-
tion of projection photolithography for the production of microelectronic chips. In the present system, the excitation of 1 results
in a Norrish type II intramolecular hydrogen abstraction to generate a 1,4-biradiacal that undergoes cleavage to form 2 and
acetophenone (Φ ∼ 0.04). In the second step, excitation of 2 causes cleavage of the oxime ester (Φ = 0.56) followed by base
generation after reaction with water.
INTRODUCTION
■
The main method for increasing computational speed in modern
computers is the reduction of the size of the electronic features
on the processor chip. Several techniques have been proposed
an intermediate species that is an active PBG (B), which in a
second photochemical step produces a base (C). Here we report
the study of a proof-of-principle system that demonstrates the
feasibility of this two-stage PBG proposal.
to achieve higher resolution in optical lithography for micro-
electronic chip manufacturing.1 A new technique, called pitch
division technology, has demonstrated doubling of the resolu-
tion in the projection printing process using conventional photo-
lithography tools without additional processing steps.2 In con-
ventional photolithography, a photoacid generator (PAG) is
used, which upon light exposure generates acid that catalyzes a
deprotection reaction which in turn modifies the solubility of
the resist polymer. In the new pitch division technology, a com-
bination of a PAG and a photobase generator (PBG) is used.
By carefully matching the kinetics and stoichiometry of the
photoreactions, the generated base can neutralize the acid at
areas of highest UV light exposure, which leaves highly exposed
areas and the unexposed areas undesolved during resist develop-
ment. In areas of weaker light exposure, not enough base is
produced to neutralize the generated acid, and therefore, the
resist polymer is dissolved in these weakly exposed areas in
the development step. This leads to a division of the pitch and
doubling of the resolution.2,3 However, line edge roughness of
the developed resist features prevents commercial implementa-
tion of pitch division.3 It has been proposed that the line edge
roughness can be improved by creating a delay in the onset of
base generation. One method to achieve this delayed base gen-
eration is through the use of a two-stage PBG,3,4 one that requires
two sequential photolysis steps to produce a molecule of base
(eq 1). In the first photochemical step, a latent PBG (A) produces
Our design of such a system is shown in Scheme 1.4 The first
photochemical step is an example of a Norrish type II reaction,5
which involves intramolecular hydrogen atom abstraction by a
ketone to form a 1,4-biradical (1,4-BR) (reaction a, Scheme 1).
The 1,4-biradical has several possible reaction pathways that
lead to stable products: cleavage of the 2,3 bond to produce a 2,4-
cyclohexadienone and an enol (reaction b) that tautomerize to
PBG (2) (reaction c) and acetophenone (3) (reaction d). For-
mation of a cyclobutanol (not shown) by intramolecular radical
recombination of the 1,4-BR could occur as a side reaction. The
second photochemical step involves cleavage of the N−O single
bond of 2 to produce a radical pair (primary RP) (reaction e,
Scheme 1), which then undergoes decarboxylation to form a
secondary radical pair (secondary RP) (reaction f) followed
by radical pair recombination to produce the imine 4 (reaction g).6,7
The imine, already a weak base, can undergo hydrolysis during
the resist development to generate the stronger base 3-aminophenol
(5) (reaction h).
Special Issue: Howard Zimmerman Memorial Issue
Received: September 30, 2012
Published: October 29, 2012
© 2012 American Chemical Society
1735
dx.doi.org/10.1021/jo302149u | J. Org. Chem. 2013, 78, 1735−1741

The Journal of Organic Chemistry
Article
Scheme 1. Overall Pathway for Sequential Photolysis of 1 (Latent PBG) and 2 (PBG) to Form Base 5
These possibilities are considered quantitatively in the next sec-
tion using data available in the literature.
RESULTS AND DISCUSSION
■
The mechanism of photochemical formation of 2 from 1 de-
serves some detailed attention since 1 possesses two potentially
active chromophores, an acetophenone (AP) moiety and a cyclo-
hexenone (CH) moiety, each possessing its own unique photo-
chemistry. Figure 1 shows absorption spectra of 1, acetophenone
3, and a model compound 6 which does not contain the aceto-
phenone chromophore. For our photochemical experiments, we
selected two different excitation wavelength, 254 and 300 nm.
The acetophenone chromophore absorbs strongly at 254 nm
(ε3 = 2780 M−1 cm−1), whereas the cyclohexenone-containing
model compound (6) shows only weak absorption (ε6 = 41 M−1
cm−1). This ensures predominant light absorption at 254 nm
by the acetophenone chromophore in 1. We also looked at
300 nm excitation because cleaner photoreactions with less side
reactions are often observed at longer wavelengths. However, the
cyclohexenone (CH) chromophore shows significant absorp-
tion at 300 nm (ε6 = 22 M−1 cm−1) vis-a-vis the acetophenone
chromophore (ε3 = 54 M−1 cm−1), which would lead to less
efficient photoconversion.
Intramolecular triplet−triplet energy transfer has been inves-
tigated for the bichromophoric donor−acceptor systems D−(CH2)n-
O−A (D = benzoyl, A = 2-naphthyl) where n =3−14.10 The values of
ET for the benzoyl and naphthyl groups are ∼73 and 62 kcal mol−1,
respectively. Therefore, intramolecular triplet−triplet energy transfer
is strongly exothermic. The highest energy transfer rate constant has
been observed for the shortest linker (n = 3; kET = 1.6 × 109 s−1).10
This rate constant serves us as an estimate of the triplet energy transfer
rate constant from the AP moiety to the CH moiety in 1. Since energy
transfer occurs by the exchange mechanism for both S1 and T1,
we can assume kET ∼ 2 × 109 s−1 is the rate constant for energy
transfer from either S1 or T1 of AP to CH.
The state energy diagram for these chromophores is shown in
Figure 2. The singlet and triplet energies (APES1 = 80 kcal mol−1
and APET1 = 74 kcal mol−1)8 of the acetophenone (AP) moiety
are higher than those of the cyclohexenone (CH) moiety
To simplify our experimental studies on the phototransforma-
tion from 1 to 2, we synthesized a model compound (1M, Scheme2)
which does not contain the oxime ester moiety. Scheme 2 shows
the rate constants for the competing processes leading from APS1
to the triplet 1,4-BR produced by type II hydrogen abstraction,
based on data from the literature. Starting from APS1, the rate
constant of intersystem crossing to APT1 (kisc = 4 × 1010 s−1)11 is
estimated to be about 10 times faster than the rate constant of
energy transfer from APS1 to CHS1. Therefore, it is expected that
the acetophenone APT1 is the starting point of the photochemistry
after photoexcitation of 1M. In this analysis, we assume that only
triplet type II reaction and triplet−triplet energy transfer provide
the major deactivation processes on the pathway to reaction
products. The rate constant for type II (kH) is assumed to be faster
than that for the valerophenone system (kH = 1.3 × 108 s−1)5 since
9
CH
(
E = 75 kcal mol−1 and CHET1 = 63 kcal mol−1). Therefore,
S1
singlet and triplet energy transfer from excited states of the AP to
CH moiety is energetically favorable. From the state energy
diagram, photoexcitation of the acetophenone moiety generates
APS1, which could undergo intersystem crossing to APT1 or energy
transfer to CH. If the CHS1 state is populated, it can undergo
intersystem crossing to CHT1 or radiationless deactivation to
CHS0. In order for the type II process to occur with substantial
quantum efficiency, the yield of APT1 must be high and the rate
constant of the type II reaction from APT1 must be competitive
with the rate constant for energy transfer from APT1 to CHT1.
1736
dx.doi.org/10.1021/jo302149u | J. Org. Chem. 2013, 78, 1735−1741

The Journal of Organic Chemistry
Article
Figure 1. Absorbance of 1, 2, 3, and model compound 6 lacking the acetophenone chromophore in acetonitrile solutions.
Figure 2. Energy diagram of acetophenone and cyclohexenone, the two chromophores of 1.
this process for 1M involves the abstraction of an activated allylic H.
From the literature, the rate constant kH of ∼5 × 108 s−1 is rep-
resentative for abstraction of an allylic H.5 From the D−(CH2)nO−A
systems for n = 3−14, a value of kET ∼ 2 × 109 s−1 is selected as
the maximum value of kET for intramolecular triplet−triplet energy
transfer for a flexible chain. Under these assumptions, the ratio of
the competitive rate constants kH and kET is ∼0.5 × 109 s−1/∼2 ×
109 s−1 = 0.25, and the maximum quantum yield of forma-
tion of 3 (Φ3) could be as high as 0.25, a significant value.
The experimental value of Φ3 will be lower if kH has been over-
estimated or if kET is underestimated or if the conversion of the
1,4-biradical to the final products is inefficient.
Laser flash photolysis experiments were performed to verify
that triplet states of the acetophenone moiety in 1M are effi-
ciently quenched by energy transfer to the cyclohexenone moiety
and/or Norrish type II reaction. Pulsed laser excitation at 266 nm
of acetophenone (3) in deoxygenated acetonitrile showed the
typical triplet−triplet absorption with maximum at 320 nm and a
1737
dx.doi.org/10.1021/jo302149u | J. Org. Chem. 2013, 78, 1735−1741

The Journal of Organic Chemistry
Article
Scheme 2. Estimated Rate Constants for the Photoconversion of 1M to 3 and 7
lifetime of ∼18 μs under our experimental conditions (for details,
see Supporting Information, Figure S1, blue line).12 Solutions
of 1M under the same experimental conditions (the same ab-
sorbance at 266 nm and the same laser power) did not show any
significant triplet−triplet absorption with the time resolution
limited by our experimental setup (Supporting Information,
Figure S1, red line), which suggests that the triplet lifetime of 1M
is shorter than 10 ns at room temperature.
Low-temperature phosphorescence experiments were per-
formed in an ethanol matrix at 77 K to investigate if the triplet
states of the AP moiety of 1M are detectable by its phos-
phorescence. Excitation of AP (3) in an ethanol matrix at 77 K with
light of 315 nm showed the typical phosphorescence spectrum
(Supporting Information, Figure S2, blue line). However, only a
small broad luminescence signal (<5% compared to 3) was
observed from 1M (Figure S2, red line) under the same experi-
mental conditions (same absorbance at 315 nm). The absence of
phosphorescence from the AP moiety in 1M is an indication of
efficient triplet quenching, even in a rigid matrix at 77 K. Because
an activation energy is associated with Norrish type II H-abstractions,
the process is typically not observed at 77 K;13 thus triplet quenching
by energy transfer to CH is probably occurring.
side reactions. Photolysis experiments were also performed at
254 nm. However, more side products were observed according
to HPLC analysis (Supporting Information, Figure S5).
The quantum yield of the photoconversion of 1M to 3 and 7
was determined using valerophenone as actinometer. Valerophenone
shows similar absorption properties to 1M and undergoes efficient
Norrish type II reaction upon photolysis with a quantum yield
of Φ = 1.0 (loss of valerophenone) in acetonitrile.14 Using
valerophenone as actinometer (for full details, see the Supporting
Information), a quantum yield of Φ3 = 0.04
0.01 for the
production of 3 from 1M was determined at 254 nm. A smaller
quantum yield, Φ3 = 0.02 0.01, was observed for photolysis at
300 nm because of competing light absorption by the cyclo-
hexenone chromophore (see Figure 1). The experimental quantum
yields are lower than those estimated from the literature values of
kET and kH (Scheme 2). Thus, we conclude that there are either
competing or unaccounted deactivation pathways or inaccurate
estimations of the values of kET and kH.
Steady-state photolysis experiments in deoxygenated acetoni-
trile solutions were conducted to see if the predicted products 3
and 7 are formed upon photolysis of 1M at 300 nm. Figure 3
shows the HPLC analysis of these irradiated solutions. Two new
peaks with retention times of 3.5 and 4.8 min appeared and were
assigned to the predicted photoproducts 7 and 3, respectively, based
on co-injection with authentic samples (Supporting Information,
Figure S3). Considering the extinction coefficients of the three com-
pounds (1M, 3, and 7) at 288 nm, the analyzing wavelength used in
these HPLC experiments, molar equivalents of both products
(3 and 7) are formed for every lost molecule of starting material
(1M). This is an indication of a clean photoprocess with negligible
The rate constant for type II reactions in valerophenone is
known to be affected by substitution at the phenyl ring.15
A
p-OCH3 substituent changes the nature of the lowest triplet state
from a nπ* (unsubstituted valerophenone) to a ππ* state. Type
II reactions are typically not observed from ππ* states.15
A
p-OCH3-substituted derivative of 1M (p-OCH3-1M) should
show negligible type II reaction (kH) because of its ππ* triplet
state. However, energy transfer from the p-OCH3 acetophenone
moiety to the cyclohexenone moiety (kET) should not be signif-
icantly affected by the p-OCH3 substituent. Laser flash photolysis
1738
dx.doi.org/10.1021/jo302149u | J. Org. Chem. 2013, 78, 1735−1741

The Journal of Organic Chemistry
Article
Figure 3. HPLC analysis of irradiation at 300 nm of 1M in deoxygenated acetonitrile solutions. Analyte detection by UV spectroscopy at 288 nm.
Figure 4. HPLC analysis of photolysis at 254 nm of 2 in deoxygenated acetonitrile solutions. Analyte detection by UV spectroscopy at 288 nm. See
Supporting Information, Figure S8, for full HPLC traces.
at room temperature (Figure S6, Supporting Information)
and low-temperature phosphorescence experiments (Figure S7)
show efficient triplet quenching of p-OCH3-1M by the CH moiety.
Steady-state photolysis and HPLC analysis showed only negligible
amounts of photoproducts. A low quantum yield of p-OCH3-1M
loss of Φ= 0.002 0.001 was determined by HPLC analysis. There-
fore, photolysis experiments with p-OCH3-1M verify that our
proposed reaction mechanism generates the observed photo-
products (2 and 3) through type II reactions from triplet excited
states of the AP moiety (Scheme 2) and not through reactions of
triplet excited states of the CH moiety.
To study the efficiency of the second photolysis step, the
generation of the base precursor 4 from 2 (Scheme 1, reactions
e−g), deoxygenated acetonitrile solutions of 2 were irradiated at
254 nm and analyzed by HPLC. Figure 4 shows HPLC traces
after different photolysis times. The peak at 12.8 min (assigned
to 2) decreased with increasing photolysis time, while the major
product peak at 1.9 min increased. This new peak was assigned to
1739
dx.doi.org/10.1021/jo302149u | J. Org. Chem. 2013, 78, 1735−1741

The Journal of Organic Chemistry
Article
Figure 5. Kinetics of photoconversion of 1 (25 mM) to 2, 3, and 4 determined by HPLC analysis of photolyzed acetonitrile solutions a 254 nm. See
Supporting Information, Figure S11, for HPLC traces.
the expected photoproduct 4 based on HPLC-MS analysis
(Supporting Information, Figures S9 and S10). A quantum yield
for loss of 2 of Φ = 0.56 0.02 was determined using valerophen-
one as actinometer, which demonstrates the efficiency of this
photoreaction.
ester 2 (PBG) undergoes N−O cleavage with a high quantum
yield (Φ = 0.56 for direct photolysis of 2, λex = 254 nm) leading to
the imine 4, which generates the desired base upon reaction with
water. Because the production of 4 from 1 involves two sequen-
tial photolysis steps, the desired delayed generation of 4 was
observed. However, for the material to be used in practical resist
formulations for photolithography with 193 nm light, the relative
rates of both sequential steps should be of similar magnitude.
The rates are determined by the extinction coefficients of 1 and 2
at the excitation wavelength and the quantum yields for each
photoreaction step. While the extinction coefficients of 1 and 2
are similar, the quantum yield for the second step (2 + hν →
base) is 1 order of magnitude larger compared to the first step
(1 + hν → 2). Therefore, the overall rate is dominated by the
kinetics of the slowest step (1 + hν → 2), and the desired delay in
the onset of base generation is reduced. Further chemical mod-
ification of the system is needed before it meets the requirements
for industrial use in resist formulations for pitch division photo-
lithography.
Using the model compounds 1M and 2, both photochemical
steps have been investigated separately as described above. To
study the entire sequential two-step base generation as outlined
in Scheme 1, acetonitrile solutions of 1 were irradiated at 254 nm
and analyzed by HPLC (Supporting Information, Figure S11).
Using authentic samples and quantitative HPLC calibration, the
concentrations of the major product (3) and intermediate (2)
were determined. Figure 5 shows the concentrations of 1, 2,
and 3 with increasing photolysis time. With consumption of 1,
the concentrations of 2 and 3 increase. As soon as enough 2 is
generated to compete for the excitation light at 254 nm (ε2 =
2360 M−1 cm−1 and ε1 = 7070 M−1 cm−1; Figure 1), photo-
product 4 is formed and the concentration of 2 decreases.
We were unable to determine the concentrations for the genera-
tion of 4 due to the lack of an authentic sample of 4 for quan-
titative HPLC calibration. Therefore, only relative concentrations of
4 are plotted in Figure 5 (right, green line). The formation of sig-
nificant amounts of 4 is observed after an inhibition time of ∼20 min
(Figure 5, green line), which shows that only at higher light doses
the second photolysis step gains importance. This delayed base
generation is a desired characteristic for potential application of 1 in
photoresist systems using pitch division technology.3
EXPERIMENTAL SECTION
■
Materials. The synthesis and characterization of compounds 1, 2, 6, 1M,
and p-OCH3-1M have been described previously.4
Laser flash photolysis experiments employed the pulses from a Nd:YAG
laser (266 nm, 5 ns) and a computer-controlled system, which has
been described previously.16 Acetonitrile solutions containing 1M, 3,
p-OCH3-1M, or p-OCH3-acetophenone were prepared at concentrations
to have an absorbance of 0.3 at 266 nm (1 cm optical path length). The
sample solutions were deoxygenated by argon purging.
High-performance liquid chromatography (HPLC) was performed on a
system equipped with a photodiode array detector and an ESI mass
detector. The injection volume was typical 10 μL. A reverse-phase C18
column (C18 5 μm 4.6 × 100 mm column) was used for all HPLC runs.
HPLC-grade water and acetonitrile were used as mobile phases with
either isocratic or gradient modes. The flow rate for all HPLC runs was
1 mL/min. The system was equilibrated with the mobile phases for at
least 20 min before each run. The samples being injected were dissolved
in acetonitrile if not specified.
CONCLUSIONS
■
We have demonstrated that 1, a latent PBG, functions as a two-
stage PBG. Upon photolysis of 1, the generated triplet excited
states of the acetophenone moiety undergo Norrish type II reac-
tion with the CH moiety, leading to generation of 2 (PBG). The
quantum yield of this type II reaction of the AP triplet is relatively
low (Φ ∼ 0.04 at λex = 254 nm) due to competition with the more
efficient triplet energy transfer from the acetophenone to the
cyclohexenone moiety. However, this low quantum yield for type
II reaction is sufficient to generate significant amounts of 2 on
steady-state irradiation. In the second photolysis step, the oxime
Quantum yield measurement. Valerophenone with a quantum yield of
1 (loss of valerophenone in acetonitrile solutions)14 was used as actinometer
for quantum yield measurements. The loss of valerophenone was quantified
1740
dx.doi.org/10.1021/jo302149u | J. Org. Chem. 2013, 78, 1735−1741

The Journal of Organic Chemistry
Article
by gas chromatography (GC) equipped with a flame ionization detector
and a capillary column (Cp-Sil 5 CB low bleed/MS, 0.05 mm × 25 m,
0.25 μm particle size). A calibration curve for each compound was
generated with at least five points, and the linear range of the detector
was determined. In a typical quantum yield measurement, the ab-
sorbances of sample and the standard in acetonitrile solutions were
matched at the irradiation wavelength. Two milliliters of the sample or
standard solution was deoxygenated (argon bubbling) and irradiated in
sealed 1 × 1 cm Suprasil quartz cells under stirring with a magnetic
stirring bar using a low-pressure Hg lamp emitting at 254 or 300 nm as
light source. Aliquots of the solution (15 μL) at different irradiation
times were withdrawn from the photolysis cell for GC and HPLC anal-
ysis and quantified by calibration curves for each compound being anal-
yzed. The conversions in these photoreactions were kept below 30% to
minimize interference from secondary products.
ASSOCIATED CONTENT
* Supporting Information
Additional HPLC traces, flash photolysis, and low-temperature
phosphorescence results. This material is available free of charge
via the Internet at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
*E-mail: njt3@columbia.edu, willson@che.utexas.edu.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors at Columbia thank the National Science Foundation
for financial support through grant NSF-CHE-11-11398 and
Intel Corporation for their support.
REFERENCES
■
(1) French, R. H.; Tran, H. V. Annu. Rev. Mater. Res. 2009, 39, 93−126.
(2) Gu, X.; Bates, C.; Cho, Y.; Coster, E.; Marzuka, F.; Nagai, T.; Ogata,
T.; Shi, C.; Sundaresan, A. K.; Turro, N. J.; Bristol, R.; Zimmerman, P.;
Willson, C. G. J. Photopolym. Sci. Technol. 2009, 22, 773−781.
(3) Gu, X.; Cho, Y.; Kawakami, T.; Hagiwara, Y.; Rawlings, B.; Mesch,
R.; Ogata, T.; Kim, T.; Seshimo, T.; Wang, W.; Sundaresan, A. K.; Turro,
N. J.; Gronheid, R.; Blackwell, J.; Bristol, R.; Willson, C. G. Proc. SPIE
2011, 7972, 79720F.
(4) Hagiwara, Y.; Mesch, R. A.; Kawakami, T.; Okazaki, M.; Jockusch,
S.; Li, Y.; Turro, N. J.; Willson, C. G. J. Org. Chem. 2012, DOI: 10.1021/
jo3021488.
(5) Wagner, P. J. Acc. Chem. Res. 1971, 4, 168−177.
(6) Lalevee, J.; Allonas, X.; Fouassier, J. P.; Tachi, H.; Izumitani, A.;
Shirai, M.; Tsunooka, M. J. Photochem. Photobiol. A 2002, 151, 27−37.
(7) Mallavia, R.; Sastre, R.; Amat-Guerri, F. J. Photochem. Photobiol. A
2001, 138, 193−201.
(8) Montalti, M.; Credi, A.; Prodi, L.; Gandolfi, M. T. Handbook of
Photochemistry, 3rd ed.; CRC Press LLC: Boca Raton, FL, 2006.
(9) Schuster, D. I.; Dunn, D. A.; Heibel, G. E.; Brown, P. B.; Rao, J. M.;
Woning, J.; Bonneau, R. J. Am. Chem. Soc. 1991, 113, 6245−6255.
(10) Wagner, P. J.; Klan, P. J. Am. Chem. Soc. 1999, 121, 9626−9636.
(11) McGarry, P. F.; Doubleday, C. E., Jr.; Wu, C.-H.; Staab, H. A.;
Turro, N. J. J. Photochem. Photobiol. A 1994, 77, 109−117.
(12) Carmichael, I.; Helman, W. P.; Hug, G. L. J. Phys. Chem. Ref. Data
1987, 16, 239−260.
(13) Wagner, P. J.; May, M. J.; Haug, A.; Graber, D. R. J. Am. Chem. Soc.
1970, 92, 5269−5270.
(14) Wagner, P. J. J. Am. Chem. Soc. 1967, 89, 5898−5901.
(15) Wagner, P. J.; Kemppainen, A. E.; Schott, H. N. J. Am. Chem. Soc.
1973, 95, 5604−5614.
(16) Yagci, Y.; Jockusch, S.; Turro, N. J. Macromolecules 2007, 40,
4481−4485.
1741
dx.doi.org/10.1021/jo302149u | J. Org. Chem. 2013, 78, 1735−1741