Photochemistry and photophysics of (p-benzoylphenyl)diphenylmethyl and (p-benzoylphenyl)bis(4-tert-butylphenyl)methyl radicals in different solvents

Article abstract of DOI:10.1021/jp000227g

The photochemical reactions of (p-benzoylphenyl)diphenylmethyl (1) and (p-benzoylphenyl)bis(4-tert-butylphenyl)methyl (2) in various solvents were investigated. The photophysical parameters of the first excited doublet state of the radicals were measured using spectroscopic and kinetic methods and led to a 'molecular rotor' model to characterize the excited-state behavior. The charge-transfer excited state for both radicals was observed. Photoproducts separated from the photolysis of 1 and 2 in benzene suggest photodecomposition proceeds via H-abstraction (55%), fragmentation (20%), cyclization (10%), and addition (10%).

Full text of DOI:10.1021/jp000227g

J. Phys. Chem. A 2000, 104, 5131-5140
5131
Photochemistry and Photophysics of (p-Benzoylphenyl)diphenylmethyl and
(p-Benzoylphenyl)bis(4-tert-butylphenyl)methyl Radicals in Different Solvents
Viktor V. Jarikov, Alexandre V. Nikolaitchik, and Douglas C. Neckers*
Center for Photochemical Sciences’, Bowling Green State UniVersity, Bowling Green, Ohio 43403
ReceiVed: January 18, 2000; In Final Form: March 23, 2000
The photochemical reactions of (p-benzoylphenyl)diphenylmethyl (1) and (p-benzoylphenyl)bis(4-tert-
butylphenyl)methyl (2) in various solvents were investigated. The photophysical parameters of the first excited
doublet state of the radicals were measured using spectroscopic and kinetic methods and led to a “molecular
rotor” model to characterize the excited-state behavior. The charge-transfer excited state for both radicals
was observed. Photoproducts separated from the photolysis of 1 and 2 in benzene suggest photodecomposition
proceeds via H-abstraction (55%), fragmentation (20%), cyclization (10%), and addition (10%).
Introduction
In the present work we focus on the photophysical properties
and photochemical reactions of (p-benzoylphenyl)diphenyl-
Resonance stabilization and steric hindrance are major factors
in determining the stability of a free radical. Since 1900, when
Gomberg first prepared a stable compound with trivalent
carbon,2 numerous triarylmethyl (TAM), aryloxyl, and nitroxyl
stable free radicals have been discovered and characterized.3
The photochemistry and photophysics of arylmethyl radicals
has been reviewed.4-10
methyl (1)22 and its di-tert-butylated derivative, (p-benzoylphe-
nyl)bis(4-tert-butylphenyl)methyl (2).15,23,24 The resonance struc-
tures of 1 and 2 can be divided into three major groups. In the
first the unpaired electron is located on the carbon atoms of the
radical; in the second it is localized on the oxygen atom; and
the third contains charge-transfer (Scheme 1) type structures.
The “g-values” of 1 (2.0029),18 2, and other substituted TPM
radicals (2.0026-2.0035)17,18 indicate that they are carbon-
centered radicals; hence the first group of resonance structures
may dominate. The photochemistry of 1 in solution is compli-
cated due to significant formation of the dimer. However, 2
should be fully dissociated in solution because all the reactive
para-positions are hindered by bulky substituents.23
The intramolecular chemistry of excited radicals has been
shown to involve cleavage, photoionization, and cyclization
reactions, which are normally not observed for ground-state
radicals.4 The intramolecular photochemistry of triphenylmethyl
(TPM) was first investigated by Letsinger et al.11 and recently
revisited in detail by Siskos et al.12 9-Phenylfluorene-derived
photoproducts were identified from the photolysis of this and
similar radicals.13 On the other hand, excitation of tri- and
diarylmethyl radicals in polar solvents results in formation of
the corresponding cations and esolv, as reported by several
workers.12,14 Although suggested on the basis of the photoprod-
uct studies,15 carbene formation via excited-radical chemistry
in solution at room temperature remains to be proved, for
example, by the way of trapping.4
Results and Discussion
I. (p-Benzoylphenyl)diphenylmethyl (1). (a) Characteriza-
tion. Radical 1 was synthesized from chloride 4 using Ag
powder in a number of solvents such as hexane, benzene,
acetone, acetonitrile, and adiponitrile. Solvent polarity has a
minor effect on the absorption spectrum of 1 (λmax ) 582-584
nm in acetonitrile or benzene, Figures 1A and 2A), indicating
that insignificant change in the radical dipole moment is
associated with the absorption process. On the other hand, a
red shift and change of the spectral shape of the fluorescence
of 1 occurs in polar solvents (Figure 1B). This points to a
significant dipole moment of the excited D1 state as well as
possible existence of a state with different energy.
The intermolecular photochemistry of radicals has been shown
to include enhanced electron-donor and -acceptor properties,
H-abstraction, and physical quenching by oxygen.4,5 Photoin-
duced electron-transfer reactions of the first excited doublet
states of TAM radicals with amines (k ∼ 108-1010 M-1 s-1),13
halogenated hydrocarbons and arenes (φcation ∼ 100%),4 and
methyl viologen16 have been reported.
Recently, a number of reports on the chemistry of TAM
radicals have been published.17-20 Renewed research efforts in
the area are directed at the investigation of the resonance
stabilization and “captodative” stabilization of stable radicals.17,18
ESR and ENDOR are the major methods used for the investiga-
tions. The hyperfine coupling constants of TAM radicals depend
on the overall conformation of the molecule. A twist angle of
about 32° for the benzene rings in the TPM radical was
determined by Adrian.21 Since no significant change in hyperfine
coupling constant as well as in width of the ESR signal occurs
upon para-substitution, the conformation of the para-substituted
TPM radicals is the same as that of the TPM radical.
In fluid solution 1 undergoes reversible exothermic dimer-
ization. The structure of dimer 5 was determined by NMR
spectroscopy (Scheme 2).18 Compound 6, obtained via an acid-
catalyzed H[1,5] shift in 5, serves as additional evidence for
the structure of 5.25 The thermodynamics of dimerization was
studied by Neumann et al. using an ESR technique.17,18 By
dissolving a known amount of the dimer in toluene, measuring
the intensity of the ESR signal, and comparing it to a standard
signal from diphenylpicrylhydrazyl, a degree of dissociation of
the dimer (R) at a given temperature could be calculated. A
value of 33% was obtained for a 0.01 M solution of 5 in benzene
10.1021/jp000227g CCC: $19.00 © 2000 American Chemical Society
Published on Web 05/02/2000

5132 J. Phys. Chem. A, Vol. 104, No. 21, 2000
Jarikov et al.
SCHEME 1: Structure Corresponding to the CT State
of 2
Figure 2. Absorption (A) and fluorescence (B, λexc ) 400 nm) spectra
of 1 in acetonitrile, [1]0: (1) 1.125 × 10-3 M, (2) 2.745 × 10-3 M, (3)
4.191 × 10-3 M, (4) 6.764 × 10-3 M.
Assuming that Beer’s law is valid in the 10-4-10-3 M range,
the dimerization equilibrium constant in acetonitrile, K ) 20 300
M-1 and the extinction coefficient at 584 nm, ꢀ ) 6300 could
be calculated from the absorption spectra in Figure 2. This
indicates that the equilibrium shifts toward the dimer in more
polar solvents.
Both the ground-state D0 and the first excited-state D1 of 1
have doublet multiplicity. To determine whether the quartet state
of 1 is populated by intersystem crossing from D1, emission
from excited 1 in rigid organic glass was studied. Dimerization
of 1 is an exothermic process, thus the lower the temperature
the greater [5] and the lower [1]. Therefore, a cooling step was
carried out as quickly as possible. Two methods were used to
study 1 in the methyltetrahydrofuran (MeTHF) glass at 77 K.
In method #1 1 was produced by addition of Ag to the solution
of chloride 4 in degassed MeTHF at 25 °C. This was followed
by rapid cooling in liquid nitrogen, and a fraction of 1 was
trapped in the solid. In method #2, chloride precursor 4 was
dissolved in degassed MeTHF at 25 °C, the solution was cooled
in liquid nitrogen and irradiated with 366 nm light. The
irradiation of the chloride led to homolysis and formation of 1.
Both samples were examined for emission in the 600-800 nm
range with different delay times (λexc ) 580 nm). Absence of
the observed emission signal suggests that the quartet state of
1 is not populated.
The two samples differed depending on the method of
preparation. Radical 1 synthesized at 25 °C in solution was red
(#1), while that prepared in MeTHF glass at 77 K via irradiation
was blue (#2). The fluorescence excitation and emission spectra
of the two samples are shown in Figure 4. The red 15 nm shift
in the fluorescence excitation spectrum of 1 in sample #2
explains the difference in color between the samples. This
difference could come from the presence of the chlorine atom
in the matrix of sample #2. However, when the same experi-
Figure 1. (A) Absorption spectrum of 1 in benzene ([4] ) 1.12 ×
10-3 M). (B) Normalized fluorescence spectra of 1 (λexc ) 540 nm):
(1) hexane, (2) mineral oil, (3) benzene, (4) acetonitrile.
at 25 °C. A value of 10.3 ( 0.1 kcal/mol was obtained for the
∆H of dissociation of 5 based on the study of the temperature
effect.
The equilibrium concentrations of 1 and 5 can be calculated
on the basis of the dimerization equilibrium constant,18,26 K )
308 M-1, and [1]0 ([1]0 ) [4] assuming quantitative conversion
of the starting chloride 4; negative Beilstein27 test). Thus, one
can plot [1]eq/[5]eq as a function [1]0 in benzene (Figure 3).23
This plot indicates that the dimer is the dominating species in
benzene when [1]0 > 3 × 10-3 M.
The presence of dimer 5 can be observed indirectly by
measuring the fluorescence of 1 in solution. The absorption and
fluorescence spectra of 1 in acetonitrile are shown in Figure 2.
The fluorescence spectra of 1 were measured with 400 nm
excitation where all the light was absorbed. As [1]0 increased
from sample #1 to sample #4, the fluorescence intensity
decreased. This is explained by the presence of quinoid dimer
5, which absorbs at 400 nm as well. As [1]0 increases, so does
[5]eq. At higher [1]0, the ratio [1]eq/[5]eq is smaller than at lower
[1]0 (Figure 3); hence [21*] and subsequently fluorescence
intensity decrease. Lower fluorescence intensity at higher [1]0
can also be explained by self-quenching of 1, which may
produce even more of 5 via the Pschorr reaction.28,29

Photochemistry of Triarylmethyl Radicals
J. Phys. Chem. A, Vol. 104, No. 21, 2000 5133
SCHEME 2: Dimerization of 1
When deuterated methanol was added to the solution of 1 in
acetonitrile, a spontaneous dark reaction led to formation of (p-
benzoylphenyl)diphenylmethyl methyl ether (12) (major) and
(p-benzoylphenyl)diphenyldeuteriomethane (11) (minor) (Scheme
3). However, a solution of 1 in benzene in the presence of 1 M
MeOD was stable in the dark. When the latter was irradiated,
11 and 12 were found in a 1:1 ratio. Deuteriomethane 11 and
(p-benzoylphenyl)diphenylmethanol (13) in a 1:1 ratio were the
products of the reaction in acetonitrile in the presence of 1 M
D2O. This reaction was slow in the dark but fast upon irradiation
with visible light. The GC/MS analysis of methane 11 (deu-
terated 7) formed in the presence of MeOD and D2O revealed
a significant degree of deuteration. We conclude that addition
to excited 1 took place indicating significant polarity of 1*.
Figure 3. Ratio [1]eq/[5]eq as a function of [1]0 in benzene at 25 °C.
The investigation of 1 is complicated by the 1-5 equilibrium,
which is affected by solvent. Processes such as fluorescence
and photochemical transformations depend on [5], thus numer-
ous corrections are required to obtain concentration-independent
values for φfl and the rate constants of the transient decay.
Therefore, we determined it preferably to eliminate the quinoid
dimer. Dissociation of the dimer is an endothermic process with
∆H ) 10.3 kcal/mol at 25 °C in benzene. Therefore, at 70 °C
and 10-5 M concentration of 1, [5] is reduced to a negligible
level. The disadvantage of this approach is the thermal instability
of the radical in polar solvents such as acetonitrile.
Magnetic susceptibility and absorption measurements indicate
that tris(4-tert-butylphenyl)methyl is 100% dissociated in solu-
tion at room temperature.30 Therefore, (p-benzoylphenyl)bis(4-
tert-butylphenyl)methyl (2) is likely to be 100% dissociated in
solution because all three para-positions are blocked with bulky
substituents.
Figure 4. Fluorescence excitation (1a, 1b) and emission (2a, 2b) spectra
of 1 in MeTHF glass at 77 K: (1a, 2a) synthesized in solution (#1);
(1b, 2b) synthesized in glass (#2).
II. (p-Benzoylphenyl)bis(4-tert-butylphenyl)methyl (2). (a)
Characterization. A relatively small blue shift (4 nm) is
observed in the absorption spectra of 2 as the solvent is changed
from benzene to acetonitrile (Figure 5A). On the other hand, a
significant change in the shape of the fluorescence spectrum of
2 (Figure 5B), as well as in the efficiency of fluorescence, was
observed with the increase in solvent polarity. In benzene, the
fluorescence spectrum of 2 is a mirror image of the absorption
spectrum. Therefore, the first excited doublet state D1 of 2 may
have a rigid structure similar in geometry to that of the ground-
state D0. In benzonitrile and acetonitrile the shape of the spectra
differs significantly from that in benzene. A broad emission is
observed along with a 50-70 nm red shift in the maximum.
Such a large red shift of fluorescence corresponds to an excited
state with a significant degree of charge transfer. Quantum yields
of fluorescence of 2 vary from 0.44 in benzene to 0.024 in
benzonitrile and 1 × 10-3 in acetonitrile. The fluorescence
excitation spectrum of 2 in acetonitrile (observation at 712 nm)
was identical with the absorption spectrum of 2, which indicates
that the same excited D1 state is formed initially in acetonitrile.
ments were performed with the TPM radical, no difference was
detected between the samples. We propose that the spectral shift
is brought about by the difference in the conformation of 1. In
the case of sample #2, the chloride precursor 4 was trapped in
a matrix and its conformation was transferred into 1 formed
after irradiation, while 1 was trapped in its normal conformation
in solution in sample #1.
(b) Photochemistry of 1 in Solution. The photochemistry
of 1 in benzene was initially studied by Neckers et al.15 The
products of photolysis with both visible and UV light were found
to be essentially the same. They were identified as (p-
benzoylphenyl)diphenylmethane (7) (major), benzophenone (8),
tetraphenylethylene (9), and 1,4-dibenzoylbenzene (10) (Scheme
3). When photolysis was performed in benzene-d6 or CD3CN,
no deuterium was incorporated into 7 and 8. This points to the
bimolecular H-abstraction reaction between 1 and 5 or the
photolysis products. Quinoid dimer 5 seems one of the most
likely H-donors. It appears that products 8-10 are formed as a
result of a photodissociation of 1. Ethylene 9 can be formed
via coupling reaction of diphenylcarbene, while 10 can come
from a recombination reaction of benzophenone radical and
benzoyl radical. The efficiency of the reaction is low upon
irradiation with both UV and visible light. The quantum yield
of the reaction was not measured because of the presence of
the dimer, which most probably interfered with the reaction.
On the basis of these data the existence of a charge-transfer
(CT) state for 2 can be proposed (Scheme 1). In polar solvents
this CT state is lower in energy than the D1 state, thus a rapid
transition between the two states takes place. Observation of
low φfl and broad structureless emission in polar solvents is

5134 J. Phys. Chem. A, Vol. 104, No. 21, 2000
Jarikov et al.
SCHEME 3: Photolysis of 1 (A) in Benzene with Visible Light, (B) in the Presence of MeOD, and (C) in Acetonitrile in
the Presence of D2O
SCHEME 4: Examples of “Molecular Rotor”
Fluorescence Probes
glass at -78 °C. A quartz test tube containing a solution of 2
in either mineral oil or methylcyclohexane was cooled to -78
°C in a spectroscopic Dewar equipped with a thermocouple.
Mineral oil forms a stable glass in the -78 to -90 °C
temperature range, while methylcyclohexane is a viscous liquid
at this temperature. Both are nonpolar solvents. The sample
compartment of the fluorometer was purged with nitrogen to
eliminate moisture, and the fluorescence spectra were recorded
at -78 °C. The solutions were then allowed to warm to room
temperature in the fluorometer sample compartment. Then the
room-temperature fluorescence spectra were recorded at +23
°C (Figure 6A). The fluorescence maximum of 2 is at 615 nm
in methylcyclohexane and 613 nm in mineral oil, and the shape
of the emission spectrum is a mirror image of the absorption
spectrum. An increase in the emission intensity at lower
temperature is common. An increase in φfl of 1.3 times was
observed in methylcyclohexane, which is a mobile liquid at both
+23 and -78 °C. However, the 5.6-fold increase in φfl of 2 in
mineral oil at lower temperature is partially due to a more rigid
environment, which implies that the geometry of the D1 state
of 2 is somewhat flexible.
Figure 5. (A) Absorption spectra of 2 in (1) benzene (λ/ꢀ - 400/
3400, 550/850 , 594/1120) and (2) acetonitrile (λ/ꢀ 400/2100, 550/
530, 594/690). (B) Normalized fluorescence spectra of 2 (λexc ) 400
nm) in (1) methylcyclohexane, (2) benzene, (3) benzonitrile, and (4)
acetonitrile.
The fluorescence spectra of 2 in polar 2-cyanoethyl acetate
at different temperatures are shown in Figure 6B. Although at
-78 °C 2-cyanoethyl acetate is a viscous oil, a 12-fold increase
in φfl was observed at lower temperature. This indicates a greater
sensitivity of φfl in polar solvents toward the viscosity of the
surrounding media. Therefore, the “molecular rotor” model can
be applied toward the emitting state of 2 in polar solvents.
An organic glass of 2 in methylcyclohexane at 77 K was
examined for phosphorescence emission. Since no phosphores-
cence was detected, there is no evidence that the quartet state
of 2 is populated.
explained by the fact that emission comes mostly from the CT
state. Molecules with flexible geometry in the excited state
exhibit low values of φfl because the energy is dissipated into
rotational and vibrational activity.31 Thus, excitation of 2 in polar
solvents produces a CT state with flexible geometry, as opposed
to the rigid geometry of the ground state. A number of
molecules, used as fluorescence probes, exhibit a dramatic
increase in φfl with an increase of the viscosity-rigidity of the
surrounding medium. Loutfy32 reported a class of compounds
that display a 100-fold difference between the values of φfl in
solution and in a rigid polymer such as poly(methyl methacryl-
ate). He named them “molecular rotors” (Scheme 4).
The electrochemistry of 2 was examined in acetonitrile in
the presence of 0.1 M tetrabutylammonium perchlorate as the
supporting electrolyte. Oxidation and reduction curves were
To test whether 2 exhibits similar behavior, fluorescence
spectra were compared in solution at -78 °C and in organic

Photochemistry of Triarylmethyl Radicals
J. Phys. Chem. A, Vol. 104, No. 21, 2000 5135
Figure 7. (A) Transient decay of D1 state of 2 in benzene: absorption
(467 nm), lifetime 26 ( 2 ns; fluorescence (680 nm), lifetime 25 ( 2
ns ([2] ) 1.3 × 10-3 M). (B) Transient absorption spectra of the D1
state of 2 in acetonitrile (6.4 × 10-3 M).
Figure 6. (A) Fluorescence of 2 in (1) methylcyclohexane, -78 °C;
(2) methylcyclohexane, +23 °C; (3) mineral oil glass, -78 °C; (4)
mineral oil liquid, +23 °C. (B) Fluorescence of 2 in 2-cyanoethyl
acetate: (1) at -78 °C; (2) at +23 °C (λexc ) 590 nm).
trile and acetonitrile. For this reaction to be possible at [2] )
0.01 M, the lifetime of D1 should be at least a few hundred
nanoseconds. However, transient absorption and fluorescence
studies described next yielded the much shorter values of the
D1 lifetime.
(b) Laser Flash Photolysis. A solution of 2 in benzene (1.3
× 10-3 M) was investigated with nanosecond LFP. Decay of
the fluorescence signal at 680 nm had a lifetime of 26 ( 2 ns,
while the lifetime of the transient absorption centered at 467
nm was 25 ( 2 ns (Figure 7A). Therefore, the observed transient
absorption in the 400-500 nm range is that of the D1 state of
2 and corresponds to D1 f D2 transition.
Fluorescence lifetime measurements were performed for 2
in acetonitrile (2.3 × 10-3 M) and benzonitrile (3 × 10-3 M).
Because of the low φfl of 2 in acetonitrile and the 8 ns resolution
of the instrument, the lifetime of fluorescence at 712 nm was
estimated to be on the order of 1-2 ns in acetonitrile. The fact
that the observed fluorescence was that of 2 was confirmed when
the fluorescence spectrum in acetonitrile was recorded 10 ns
after the laser pulse. It was identical to that measured under the
steady-state conditions. Experiments in benzonitrile yielded a
value of 11 ( 2 ns for the observed fluorescence lifetime of 2.
No transient absorbance in the 400-500 nm region was detected
in acetonitrile and benzonitrile on the nanosecond time scale.
This suggests that an emitting state of 2 other than D1 is
populated in polar solvents.
measured vs Ag/0.01 M AgNO3 reference electrode. A value
of +0.21 V vs SCE for the oxidation potential of 2 was
calculated from the oxidation curve. This is comparable with
the oxidation potential for 4,4′,4′′-trimethyltriphenylmethyl
(+0.29 V vs SCE in 1,2-dimethoxyethane).33 The reduction
potential of 2 in acetonitrile (-0.87 V vs SCE) is comparable
with that of 4,4′,4′′-trichlorotriphenylmethyl (-0.87 V vs Ag/
AgCl in DMSO).34 The oxidation and reduction potentials for
2 in acetonitrile vs SCE were calculated using the oxidation
potential of ferrocene Eox ) 0.063 V in acetonitrile vs Ag,
AgNO3 and E(Ag/0.01 M AgNO3//TEAP) ) 0.291 V vs SCE
as a reference.35
The Rehm-Weller equation gives a free energy value for an
electron-transfer reaction between the first excited doublet state
of radical R(D1) and the ground doublet state R(D0), which
produces solvated ions.
R(D1) + R(D0) f R- + R+
∆G ) Eox - Ered - E00 + 2.6/ꢀ - 0.13
fl
The excited-state energy E00 of 2 in benzene is 1.96 eV (λmax
fl
) 633 nm), in benzonitrile E00 ) 1.77 eV (λmax ) 700 nm),
and in acetonitrile E00 ) 1.74 eV (λmaxfl ) 712 nm). Therefore,
the free energy of the electron-transfer reaction in benzene
∆GPhH is +0.14 V, in benzonitrile ∆GPhCN ) -0.73 V, and in
acetonitrile ∆GMeCN ) -0.72 V. Thus, the electron-transfer
reaction is endothermic in benzene and exothermic in benzoni-
Absorption of the D1 state was detected in acetonitrile, when
532 nm laser pulses with 30 ps width were employed for the
excitation ([2] ) 6.4 × 10-3 M, Figure 7B). A satisfactory fit
of the transient absorption decay at 467 nm was obtained using
first-order kinetics. The lifetime of the D1 state of 2 in

5136 J. Phys. Chem. A, Vol. 104, No. 21, 2000
Jarikov et al.
TABLE 1: Photophysical Properties of 2 in Various
Solventsa
solvent ꢀb
ηb λmax λfl
φfl
τ(D1)
τfl
104φr
PhH 2.27 0.649 596 633 0.44
PhCN 25.7 1.24 593 700 0.024 1.60 ( 0.07 11 ( 2 3.9
MeCN 37.5 0.345 592 712 0.001 0.126 ( 0.005 1-2
25 ( 2
26 ( 2 9.0
1.0
a λmax is the absorption maximum, nm; λfl is the fluorescence emission
maximum, nm; φfl is the quantum yield of fluorescence; τ(D1) is the
lifetime of the of the D1 state, ns; τfl is the lifetime of the observed
fluorescence, ns; φr is the quantum yield of the photoreaction. b From
Murov, S. L.; Carmichael, I.; Hug, G. L. Handbook of Photochemistry;
1993. ꢀ is the dielectric constant, and η is viscosity, cP.
acetonitrile was 126 ( 5 ps. The 640 nm peak at in Figure 7B
belongs to D1 as well, since the same decay parameters were
obtained at this wavelength. In benzonitrile, the lifetime of D1
was determined as 1.60 ( 0.07 ns.
Since the lifetimes of the D1 state of the radical are in the
0.1-25 ns range, bimolecular reactions of D1 with any other
species but solvent molecules are of little significance.
(c) Summary. Photophysical parameters measured for 2 in
various solvents are compiled in Table 1. Different values for
the lifetime of the observed fluorescence and the lifetime of
the first excited doublet state of 2 in polar solvents indicate the
existence of two excited states: D1 and CT. In nonpolar solvents
such as benzene, excitation of the radical produces D1, which
then decays via fluorescence and internal conversion. The
absorption of D1 (D1 f D2) is observed, the lifetime of which
equals the observed lifetime of fluorescence. The CT state is
not populated in benzene because it is of higher energy than
the D1 state. In polar solvents such as acetonitrile, transition
between the D1 state of 2 and the CT state occurs, since the
latter is lower in energy. The CT state of 2 has a flexible
geometry. A tendency in φfl and φr (next section) shows that
fluorescence and the photoreaction are more efficient for the
D1 state, while efficient internal conversion is the primary decay
pathway for the CT state.
Figure 8. (A) Photolysis of 2 in benzene (0 s f 2500 s at 360 s
increments). (B) Photolysis of 2 (2.3 × 10-3 M) in benzonitrile.
In the second run, a 300 mL benzene solution was irradiated
with a 500 W tungsten halogen lamp (350-800 nm) and water
was used as a filtering-cooling solution. Upon exposure of a
0.012 M solution of 2 (1.7 g) in degassed benzene to visible
light for 31 h, the red color of the radical slowly disappeared.
The yellow solution of the photolyzate was exposed to air and
analyzed. The best separation of the roughly 20 spots of which
four were major (TLC) was achieved with ethyl acetate/hexanes
) 1/15. Eluting ethyl acetate/hexanes ) 1/33 over a silica
column provided the best separation by HPLC and showed a
total of 22 peaks of which 9 were major. After evaporation of
solvent, the photolyzate was subjected to preparative TLC
followed by Soxhlet extraction. (p-Benzoylphenyl)bis(4-tert-
butylphenyl)methanol (18) was recovered, and the conversion
of 2 was presumed to be 91%. The photoproducts are listed in
Table 2. Products 16, 18, 20, and 21 were isolated, purified,
and identified. Compounds 18 and 21 were found identical to
the known materials. The rest of the photoproducts were
separated in mixtures of two to three compounds and not
purified but characterized primarily by GS/MS and TLC. An
inseparable mixture of the unidentified low-yield products
constituted about 20% of the photolyzate.
(d) Photoreaction of 2 in Various Solvents. Upon exposure
of a solution of 2 to visible or UV light, a slow bleaching process
is observed. The rate of bleaching varies in different solvents.
Quantum yields of reaction (φr) were determined in benzene,
benzonitrile, and acetonitrile. When determining φr, it is
necessary to account for the change in the absorption of the
radical. Equation 5 for φr is derived in the Experimental Section.
To determine the quantum yield of radical disappearance,
solutions of 2 were irradiated with a 514 nm line of an Ar-ion
laser and the absorption of 2 was measured at certain times
(Figure 8A). A plot of -log((1 - T)/T) vs time is shown in
Figure 8B for 2 in benzonitrile. Substitution of values of ꢀ of 2
at 514 nm, I0 (light intensity, 3.5 × 10-7 einstein/s), and path
length into eq 5 gave the values of φr: 9 × 10-4, 3.9 × 10-4
,
and 1.0 × 10-4 in benzene, benzonitrile, and acetonitrile,
respectively. Therefore, the efficiency of the photoreaction is
generally low and is lower in more polar solvents.
Benzene was chosen as a solvent for two separate preparative
photolyses.23,24 In the first, a 100 mL solution (5 × 10-3 M)
was irradiated for 12 h with light from a 300 W tungsten lamp,
which was filtered through a NaNO2 solution (400-800 nm).
The photolysis was stopped when the red color of 2 disappeared.
The residue obtained after evaporation of the solvent was treated
with ethyl acetate, and an insoluble white precipitate was
filtered. It was identified as tetrakis(4-tert-butylphenyl)ethylene
(21) (yield 20%). GC/MS analysis of the photolyzate showed
no benzophenone. The photolyzate was not analyzed further.
A trapping experiment was attempted. Degassed toluene was
added to a benzene solution of 2. This resulted in a ground-
state reaction since in 2 h the color of radical 2 disappeared in
the dark.
Isolation of a dominant H-abstraction product, triarylmethane
20, came as a surprise because we assumed absence of H-donors
such as the dimer of 2 in the system. This assumption might
have been incorrect. The color of 2 could disappear during the
preparative photolysis partially due to dimerization (ground state
and/or photoinduced). Upon exposure to air, the dimer of 2 could

Photochemistry of Triarylmethyl Radicals
J. Phys. Chem. A, Vol. 104, No. 21, 2000 5137
TABLE 2: Products from Irradiation of 2 in Degassed
Benzene, Molar %
Interestingly, phenanthrene 27, which is the photocyclization
product of ethylene 21, was detected.
The apparent extinction coefficients of 2 in benzene and
acetonitrile (Figure 5A) were calculated assuming that no
dimerization of 2 occurred. Their values are significantly lower
than ꢀ of 1 in acetonitrile that we have derived above. This
indicates that the dimerization of 2 might have persisted. One
possible reason is that the p-benzoyl group is not bulky enough.
Though it was unexpected, the photoproducts of 2 in benzene
differ slightly from those of 1. Products of H-abstraction account
for 50-60% of the converted radical, while products of
fragmentation account for ∼20%, and cyclization products
account for only ∼10%. Although our mechanistic study is far
from complete, we propose formation of diarylcarbene and
p-benzoylphenyl takes place. The products of this fragmentation
in benzene should be formed via initial C-C bond cleavage in
the excited radical. On the basis of the average value for a C-C
bond dissociation energy (85 kcal/mol) and the energy of the
excited radical (45 kcal/mol in benzene), as well as the low
value of the quantum yield of fragmentation (2 × 10-4), we
conclude that this reaction may require absorption of at least
two visible-light photons. To confirm this, the dependence of
the reaction rate on the intensity of the absorbed light must be
studied.
Conclusions
Investigation of the photophysics of 1 and 2 resulted in
development of a molecular model characterizing their behav-
ior: both radicals follow the “molecular rotor” model in their
excited-state chemistry.
The photochemistry of these stable radicals is complicated
by the propensity to dimerize. The difference in the photoprod-
ucts of 1 and 2 is minor. We postulate that the photochemistry
of 1 is dominated by H-abstraction due to reaction of 1* with
dimer 5. But in 2, where dimer formation was thought to have
been eliminated, similar products were obtained. The H-
abstraction reaction (50-60%) dominates in the case of 2 as
well. Other reactions occur: fragmentation (10-20%), cycliza-
tion (5-15%), and possibly addition (photoinduced dimerization
of two radicals, 10-20%). To occur to an appreciable extent,
the rate constant of the addition reaction should be at least 107-
108 M-1 s-1
.
Experimental Section
Materials and Equipment. Reagents and solvents were
purchased from Aldrich. Benzene (99%+, thiophene-free) and
THF (anhydrous, 99.8%) were distilled under argon from Na-
benzophenone prior to use. Acetonitrile (99.9%+, HPLC grade)
was distilled from P2O5 under argon, and benzonitrile (99%+,
anhydrous) was purged with argon at 75 °C. Solvents used
without purification were spectrophotometric grade. NMR
spectra were taken with either a Varian Gemini 200 NMR or a
Varian Unity Plus 400 NMR spectrometer. Chemical shifts are
in ppm with either TMS or residual nondeuterated solvent as
the internal standard. GC/MS and DIP/MS were taken on an
HP 5988 mass spectrometer coupled to an HP 5880A GC with
a 30 m × 0.25 mm i.d. × 0.25 mm film thickness DB-5 ms
column, interfaced to an HP 2623A data processor. UV-visible
spectra were obtained using an HP 8452 diode array spectro-
photometer. Infrared spectrometry was performed using a
Mattson Instruments 6020 Galaxy Series FT-IR spectrometer.
Elemental analysis was carried out by Atlantic Microlab Inc.
HRMS was performed at the University of Illinois Urbana/
Champaign, School of Chemical Sciences. Melting points were
produce carbinol 18 (yield 9%) as well as other products of
oxidation. Alternative H-donors could be intermediates derived
from cyclization reactions of 2 that lead to the formation of
fluorene derivatives 23, 26, 28, and 29. Formation of ethylene
21 suggests the intermediacy of bis(4-tert-butylphenyl)carbene
and p-benzoylphenyl, although neither the dimer of p-ben-
zoylphenyl nor benzophenone was detected. Instead, 4-hydroxy-
benzophenone (22) was isolated in 2:1 molar ratio to 21.

5138 J. Phys. Chem. A, Vol. 104, No. 21, 2000
Jarikov et al.
determined using a capillary melting point apparatus (Uni-melt,
Arthur H. Thomas Co., Philadelphia) and were uncorrected.
HPLC analysis was performed on a HP 1050 series instrument
using a multiple wavelength diode array detector. Separations
at a flow rate of 1 mL/min were performed on either a Nucleosil
AB (Altech) 15 × 4.6 mm column, using H2O/MeCN ) 1/5 as
eluent or Hypersil (HP) 20 × 4.6 mm column with hexane/
EtOAc ) 15/1.
Fluorescence Measurements. Fluorescence spectra were
recorded with a SPEX Fluorolog 2 spectrofluorometer equipped
with both excitation and emission double beam monochroma-
tors. Spectra were measured in perpendicular geometry using a
1 cm quartz cuvette. All spectra were corrected (band-pass 2.8
nm). Fluorescence quantum yields of radicals in solution were
measured using cresyl violet acetate in ethanol (φfl ) 0.50 (
0.02)36 as a reference solution. The values of φfl were calculated
as: φfl ) φrefIfl(1 - Tref)/(Iref(1 - T)), where φref is a quantum
yield of fluorescence of the reference, Ifl and Iref are fluorescence
intensities of the unknown and the reference, and T and Tref are
the transmittances of the unknown and the reference at the
wavelength of excitation.
generated by the second harmonic (532 nm) of Nd:YAG laser
(Quantel YG571C). The 50 mL sample solution was placed in
an optical glass cuvette connected with a 100 mL Schlenk tube.
The large volume of the sample solution allowed minimization
of the effect of radical bleaching during the measurements. The
solution in the cuvette was stirred continuously during the
experiment.
Steady-State Photolysis. Photolyses were performed under
a dry argon atmosphere. For UV-light irradiation, a benzene
solution of the radical in a 250 mL Pyrex test tube was placed
in a Rayonet RPR-100 photoreactor (a merry-go-round ap-
paratus) equipped with 14 350 nm filter-coated GE F8T5 (24
W) low-pressure mercury UV bulbs (λexc ∼ 330-370 nm), a
stirring plate, and a cooling fan. For visible-light photolysis a
tungsten halogen lamp was used (λexc ) 350-800 nm).
Preparative Photolysis of 2 (#1). Irradiation was performed
in an Ace Glass photochemical reactor (100 mL) using a 300
W tungsten lamp. The reactor was cooled internally by
circulating a solution of NaNO2 (75 g/L) in H2O to cut off UV
light (<400 nm) generated by the lamp. Irradiation was
continued until complete disappearance of the radical was
assured. The residue obtained after evaporation of benzene was
dissolved in 20 mL of ethyl acetate, and the insoluble white
solid was filtered.
Quantum Yield of Radical Disappearance. A solution of
the radical in a cuvette was irradiated with 514 nm line of an
air-cooled Ar-ion laser (OmNichrome, model 543). This line
was isolated with an interference filter (Oriel, 6 nm band-pass).
Laser power was measured with a power meter (Scientech 365).
Absorbance of the radical in solution at 514 nm was measured
at certain times during irradiation.
Preparative Photolysis of 2 (#2). Radical 2 (1.7 g, 3.7 mmol)
was prepared via addition of Cu powder (17:1) to chloride 19
(1.92 g, 3.9 mmol). The synthesis, handling, and photolysis were
done under argon. Preparative photolysis of 0.012 M solution
in benzene (300 mL) was performed in an Ace Glass photo-
chemical reactor (500 mL) using a 500 W tungsten halogen
lamp. The reactor was cooled internally by circulating water.
Irradiation was continued until the red color of the radical
disappeared. After irradiation the solution was exposed to air,
and oxidation occurred. The residue obtained after evaporation
of benzene was dissolved in 10 mL of dichloromethane. The
unreacted starting material and products were separated on 20
× 20 cm silica gel plates (2000 µm, 5-17 µm, 60 Å, Altech)
with hexane/ethyl acetate ) 15/1, purified, weighed, and
identified. Compounds were characterized by 1H and 13C NMR,
mass, IR, and UV-vis spectroscopy, elemental analysis, HRMS,
HPLC, and TLC. For low-yield and inseparable compounds,
TLC and GC-DIP/MS results are reported.
The following factors were considered in the calculation of
φr. Let us consider a photochemical reaction of a compound R
excited by monochromatic light: R + hν f R* f products.
The rate of disappearance of R is proportional to the number
of photons absorbed by R in a unit of time: -dc/dt ) φrIa (1),
where c is concentration of R in M, t is time in seconds, φr is
quantum yield for the photochemical reaction of R, and Ia is
photon flux absorbed by R in einsteins/s. From Beer’s law: c
) A/lꢀ (2), where A is the absorbance of R at a given
wavelength, ꢀ is an extinction coefficient of R at this wavelength
in M-1 cm-1, and l is a path length in cm. Thus, -dA/dt ) lꢀφrIa
(3). Ia can be expressed as Ia ) I0(1 - T), where I0 is a total
photon flux and T is transmittance of solution of R at the
wavelength of excitation. T ) 10-A and A ) -log T.
Substituting these expressions into eq 3 results in: d(log T)/dt
) lꢀφrI0(1 - T) (4). Integrating yields: log(T/(1 - T)) )
2.3lꢀφrI0t (5).
Photoproducts. (p-Benzoylphenyl)bis(4-tert-butylphenyl)-
1
methane (20): H NMR (200 MHz, CDCl3) δ 7.81-7.66 (m
(d, dd), 4H), 7.59-7.47 (m, 1H), 7.47-7.36 (d, 2H), 7.35-
7.27 (d, J ) 8.4 Hz, 4H), 7.27-7.18 (d, J ) 8.4 Hz, 2H), 7.10-
Electrochemical Characterization. A solution of radical 2
(1 × 10-3 M) and tetrabutylammonium perchlorate (0.1 M) in
acetonitrile was placed in a 15 mL electrochemical cell equipped
with a working platinum electrode, an auxiliary electrode
(platinum wire), and a reference electrode (Ag, 0.01 M AgNO3//
0.1 M TBAP in MeCN). The solution was kept under argon
during the measurements, which were performed on a BAS
100A Electroanalytical analyzer with a scan rate of 200 mV/s.
At the end of the experiment a small amount of ferrocene
(99.8%) was added to the cell, and the oxidation potential was
determined. This value was used to correct the values of redox
potentials measured for 2.
6.98 (d, J ) 8.4 Hz, 4H), 5.54 (s, 1H(CH)), 1.29 (s, 18H); 13
C
NMR (APT, 50 MHz, CDCl3) δ 196.3, 149.4, 149.2, 140.1,
137.7, 135.4, 132.1, 130.1, 129.9, 129.3, 128.9, 128.1, 125.2,
56.0, 34.3, 31.3; MS (EI) 460, 445, 403, 355, 278, 215, 176,
162, 117, 105 (b), 77, 57. Anal. Calcd for C34H36O: C, 88.65;
H, 7.87. Found: C, 88.71; H, 7.82.
Tetrakis(4-tert-butylphenyl)ethylene (21):39 1H NMR (200
MHz, CDCl3) δ 7.33-7.25 (d, J ) 8.4 Hz, 8H), 7.13-7.03 (d,
J ) 8.4 Hz, 8H), 1.24 (s, 36H); C6D6 δ 7.10-7.02 (d, J ) 8.4
Hz, 4H), 6.96-6.87 (d, J ) 8.4 Hz, 4H), 1.14 (s, 36H); 13C
NMR (APT, 50 MHz, CDCl3) δ 148.8, 141.1, 140.2, 130.9,
124.1, 34.3, 31.3; C6D6 δ 149.1, 142.0, 140.7, 131.7, 124.9,
34.4, 31.3; MS (EI) 556, 541, 263, 57 (b), 41; mp 342 °C, dec;
UV-vis 312 nm.
Laser Flash Photolysis. Nanosecond pulsed laser photolyses
were carried out on a setup described by Ford and Rodgers using
the second harmonic of a Q-switched Nd:YAG laser (Con-
tinuum, YG660) as a pump light (fwhm 8 ns).37 The instrument
was used in both absorption and emission modes. Picosecond
LFP experiments were performed on a setup described by
Logunov and Rodgers.38 Laser pulses of 30 ps width were
Other: MS (EI): (22) 198, 121 (b), 105, 93, 77, 51, 39;
(23) 474, 417, 402, 361, 222, 161, 105 (b), 77, 57; (25) 342,
327 (b), 271, 209, 161, 152, 105, 77; (26) 292, 277 (b), 249,

Photochemistry of Triarylmethyl Radicals
J. Phys. Chem. A, Vol. 104, No. 21, 2000 5139
178, 131, 117, 103, 57; (27) 554, 539, 262, 57; (28) 458, 402,
353, 214, 105 (b), 77, 57.
1314, 1283, 730, 600. MS (EI): 476, 459-61, 403, 343, 315,
295 (b), 209, 161, 105 (b), 77, 57, 41. HRMS Calcd for
C34H36O2: 476.2710. Found: 476.2715. Purity: 99%, HPLC.
(p-Benzoylphenyl)bis(4-tert-butylphenyl)methyl Chloride
(19). Carbinol 18 (1 g, 2.1 mmol) in anhydrous benzene (20
mL) was heated to reflux. Thionyl chloride (99.9%+, 3 mL) in
5 mL of anhydrous benzene was added through the reflux
condenser. The mixture was refluxed for 4 h. The solvent was
evaporated, and the crude product recrystallized from anhydrous
petroleum ether (yield 95%). Mp: 212-213 °C (dec). IR (KBr),
cm-1: 2965, 2904, 2868, 1655, 1599, 1506, 1477, 1462, 1449,
1401, 1317, 1310, 1276, 703, 610. MS (EI): 459, 402, 326.
Anal. Calcd for C34H35OCl: C, 82.52; H, 7.08. Found: C, 82.68;
H, 7.13.
Synthesis of 1. (p-Benzoylphenyl)diphenylmethanol (3) was
synthesized from 4-(trifluoromethyl)benzophenone via Friedel-
Crafts reaction in benzene using a procedure described by
Neckers et al.40 and purified by recrystallization from methanol
(91% yield, mp 131 °C).22 (p-Benzoylphenyl)diphenylmethyl
chloride (4) was prepared from carbinol 3 via reaction with
thionyl chloride in refluxing benzene for 4 h. Chloride 4 was
purified by repeated recrystallization from glacial AcOH/AcCl
(5/1 by volume) according to Wittig et al. (yield 64%, mp 122
°C).22 Radical 1 was prepared from 4 by Gomberg’s method2
with Ag powder (2.5 µm).
Synthesis of 2. The Friedel-Crafts reaction of 4-(trifluo-
romethyl)benzophenone with 2 equiv of tert-butylbenzene in
the presence of 2 equiv of AlCl3 led to many byproducts due
to migration of tert-butyl group.41 However, reaction of the
Grignard reagent formed from acetal 14 with benzophenone 16
was successful.
4-Bromobenzopheneone Acetal (14).42 Solution of 4-bro-
mobenzophenone (11 g, 42.1 mmol), ethylene glycol (20 mL,
354 mmol), and a catalytic amount of p-toluenesulfonic acid
(500 mg) in 150 mL of toluene was refluxed for 10 h. Water
was removed from the reaction mixture by azeotropic distilla-
tion. The reaction mixture was washed with 10% Na2CO3 and
aqueous NaCl and dried over MgSO4. The solvent was
evaporated, and the product was recrystallized from ethanol
(yield 86%, mp 56 °C).
Radical 2 was produced from 19 by reaction with either Cu
powder (99%, submicron) in benzene at 80 °C for 2 h or Ag
powder (99.99%, 2.5 µm) at room temperature in acetonitrile.2
Heating of the radical in acetonitrile resulted in its decomposi-
tion. The reaction was stopped after Beilstein test produced
negative result. Sensitivity of the test is in the microgram
1
region,27 thus quantitative conversion of 19 was assumed. H
NMR (200 MHz, C6D6): δ 7.75-7.50 (m (d, d), 4H), 7.50-
6.95 (m, 13H), 1.15 (s, 18H).
Synthesis of 2-Cyanoethyl Acetate. 3-Hydroxypropionitrile
(20 g, 280 mmol) was reacted with acetic anhydride (300 g) in
anhydrous pyridine (500 mL) for 24 h. Excess acetic anhydride,
acetic acid, and pyridine were evaporated under reduced
pressure. The remaining yellow oily liquid was distilled under
1
1 mmHg to give 28 g (89% yield) of 2-cyanoethyl acetate. H
4,4′-Di-tert-butylbenzophenone (16).43 4-tert-Butylbenzoyl
chloride (10 g, 51 mmol), tert-butylbenzene (8 g, 60 mmol),
and AlCl3 (7.3 g, 55 mmol) were dissolved in 30 mL of dry
nitrobenszene. The solution was stirred at 25 °C for 20 h and
then at 60 °C for 2 h. Nitrobenzene was removed by steam
distillation, and the resulting solid was recrystallized twice from
NMR (CDCl3): δ 2.11 (3H, s), 2.708 (2H, t, J ) 6.34 Hz),
4.272 (2H, t, J ) 6.34 Hz).
Acknowledgment. We thank Drs. George S. Hammond and
Thomas H. Kinstle for insightful discussions. We are indebted
to Dr. Michael A. J. Rodgers for use of his LFP facilities. Both
V.J. and A.N. are grateful to the McMaster Endowment for a
Fellowship.
1
ethanol (yield 79%, mp 131-132 °C). H NMR (200 MHz,
CDCl3) δ 7.81-7.71 (d, J ) 8.8 Hz, 4H), 7.53-7.44 (d, J )
8.8 Hz, 4H), 1.37 (s, 18H); 13C NMR (APT, 50 MHz, CDCl3)
δ 155.9, 135.1, 130.0, 125.1, 35.1, 31.1; MS (EI) 294, 279(b),
237, 161, 146, 132, 118, 104, 91, 77, 57, 41.
References and Notes
(1) Contribution No. 402 from the Center for Photochemical Sciences.
(2) Gomberg, M. J. Am. Chem. Soc. 1900, 22, 757.
(3) Forrester, A. R.; Hay, J. M.; Thompson, R. H. Organic Chemistry
of Stable Free Radicals; Academic Press: New York, 1968.
(4) Johnston, L. J. Chem. ReV. 1993, 93, 251.
(5) Wilson, R. M.; Schnapp, K. A. Chem. ReV. 1993, 93, 223.
(6) Johnston, L. J.; Scaiano, J. C. Chem. ReV. 1989, 89, 521.
(7) Scaiano, J. C.; Johnston, L. J. Org. Photochem. 1989, 10, 309.
(8) Caldwell, R. In Kinetics and Spectroscopy of Carbenes and
Biradicals; Platz, M. S., Ed.; Plenum Publishing Corp.: New York, 1990;
p 77.
(9) Smirnov, V. A.; Plotnikov, V. G. Russ. Chem. ReV. 1986, 55, 929.
(10) Duxbury, D. F. Chem. ReV. 1993, 93, 381.
(11) Letsinger, R. L.; Collat, R.; Magnusson, M. J. Am. Chem. Soc.
1954, 76, 4185.
(12) Siskos, M. G.; Zarkadis, A. K.; Steenken, S.; Karakostas, N.; Garas,
S. K. J. Org. Chem. 1998, 63, 3251 and references therein.
(13) Fox, M. A.; Gaillard, E.; Chen, C. J. Am. Chem. Soc. 1987, 109,
7088.
(p-Benzoylphenyl)bis(4-tert-butylphenyl)methanol (18). So-
lution of acetal 14 (3.05 g, 10 mmol) in THF (20 mL) was
added dropwise over 30 min to a stirred suspension of activated
magnesium in THF (20 mL), which was prepared by the
procedure of Rieke and Bales44 from anhydrous MgCl2 (1.92
g, 20 mmol), KI (1.68 g, 10 mmol) and K (1.39 g, 35.8 mmol).
The mixture was further stirred for 2 h, and a solution of 16
(2.94 g, 10 mmol) in 20 mL of THF was added. The reaction
mixture was stirred at 50 °C for 24 h. After treatment with
saturated NH4Cl at 0 °C, the mixture was extracted with ether.
After removal of the solvent, a yellow oil was obtained, to which
50 mL of water, 100 mL of THF, and 0.8 g of TsOH were
added. The mixture was refluxed for 8 h. After evaporation of
THF, 200 mL of ether was added, and the solution was extracted
with 10% Na2CO3, water, and aqueous NaCl. The solvent was
evaporated, and the crude product was chromatographed on
silica gel with hexane/ethyl acetate (20/1). The separated
carbinol was recrystallized from petroleum ether (yield 49%).
(14) Faria, J. L.; Steenken, S. J. Phys. Chem. 1993, 97, 1924.
(15) Neckers, D. C.; Rajadurai, S.; Valdes-Aguilera, O.; Zakrzewski,
A.; Linden, S. M. Tetrahedron Lett. 1988, 29, 5109.
(16) Scaiano, J. C.; Tanner, M.; Weir, D. J. Am. Chem. Soc. 1985, 107,
4396.
1
Mp: 151-152 °C. H NMR (200 MHz, CDCl3): δ 7.8-7.66
(17) Neumann, W. P.; Uzick, W.; Zarkadis, A. K. J. Am. Chem. Soc.
(m (d, dd), 4H), 7.59-7.48 (m, 1H), 7.48-7.35 (m (d, d), 4H),
7.35-7.24 (d, J ) 8.6 Hz, 4H), 7.23-7.12 (d, J ) 8.6 Hz,
4H), 3.12 (s, 1H(OH)), 1.3 (s, 18H). 13C NMR (APT, 50 MHz,
CDCl3): δ 196.4, 151.6, 150.2, 143.4, 137.5, 135.9, 132.3,
130.0, 129.6, 128.2, 127.7, 127.5, 124.9, 81.5, 34.4, 31.3. IR
(KBr), cm-1: 3490, 2962, 2902, 2866, 1656, 1605, 1446, 1319,
1986, 108, 3762.
(18) Neumann, W. P.; Penenory, A.; Stewen, U.; Lehnig, M. J. Am.
Chem. Soc. 1989, 111, 5845.
(19) Bordwell, F. G.; Zhang, X.; Alnajjar, M. S. J. Am. Chem. Soc.
1992, 114, 7623.
(20) Lehnig, M.; Stewen, U. Tetrahedron Lett. 1989, 30, 63.
(21) Adrian, F. J. J. Chem. Phys. 1958, 28, 608.

5140 J. Phys. Chem. A, Vol. 104, No. 21, 2000
Jarikov et al.
(22) Wittig, G.; Kairies, W.; Hopf, W. Chem. Ber. 1932, 65, 767.
(23) Nikolaitchik, A. V. Ph.D. Dissertation, Bowling Green State
University, Bowling Green, OH, 1996.
(24) Jarikov, V. V. Ph.D. Dissertation, Bowling Green State University,
Bowling Green, OH, 1999.
(25) Takeuchi, H.; Nagai, T.; Tokura, N. Bull. Chem. Soc. Jpn. 1971,
44, 753.
(26) Zarkadis, A. K.; Neumann, W. P.; Uzick, W. Chem. Ber. 1985,
118, 1183.
(34) Breslow, R.; Chu, W. J. Am. Chem. Soc. 1973, 95, 411.
(35) Mann, C. K.; Barnes, K. K. Electrochemical Reactions in Non-
aqueous Systems; Dekker: New York, 1970.
(36) Lopez-Arbeloa, I.; Rohatgi-Mukherjee, K. K. Chem. Phys. Lett.
1986, 129, 607.
(37) Ford, W. E.; Rodgers, M. A. J. J. Phys. Chem. 1994, 98, 3822.
(38) Logunov, S. L.; Rodgers, M. A. J. J. Phys. Chem. 1992, 96, 2915.
(39) Smith, W. B. Org. Magn. Reson. 1981, 15, 317.
(27) Dittrich, H.; Vorberg, B. Anal. Chim. Acta 1982, 140, 237.
(28) Nonhebel, D. C.; Walton, J. C. Free-radical Chemistry; University
Press: Cambridge, 1974.
(40) Neckers, D. C.; Linden, S.-M.; Williams, B. L.; Zakrzewski, A. J.
Org. Chem. 1989, 54, 131.
(29) Kupchan, S. M.; Wormser, H. C. J. Org. Chem. 1965, 30, 3792.
(30) Selwood, P. W.; Dobres, R. M. J. Am. Chem. Soc. 1950, 72, 3860.
(31) Turro, N. J. Modern Molecular Photochemistry; University Science
Books: Mill Valley, CA, 1991.
(32) Loufty, R. O. Pure Appl. Chem. 1986, 58, 1239.
(33) Bank, S.; Ehrlich, C. L.; Zubieta, J. A. J. Org. Chem. 1979, 44,
1454.
(41) Kulka, M. J. Am. Chem. Soc. 1954, 76, 5469.
(42) Masuhara, H.; Maeda, Y.; Nakajo, H.; Mataga, K. N.; Tomita, K.;
Tatemitsu, H.; Sakata, Y.; Misumi, S. J. Am. Chem. Soc. 1981, 103, 634.
(43) Wagner, P. J.; Truman, R. J.; Puchalski, A. E.; Wake, R. J. Am.
Chem. Soc. 1986, 108, 7727.
(44) Rieke, R. D.; Bales, S. E. J. Am. Chem. Soc. 1974, 96, 1775.