Photoinduced Inter- and Intra-Ion-Pair Electron Transfer Reactions in N-(p-Benzoylbenzyl)-N,N,N-tri-n-butylammonium Triphenyl-n-butylborate and -Gallate Salts

Article abstract of DOI:10.1021/jp960931q

The protoreduction of N-(p-benzoylbenzyl)-N,N,N-tri-n-butylammonium triphenyl-n-butylborate (1) and gallate (2) was studied using nano- and picosecond laser flash photolysis.An electron transfer reaction from the borate or the galate counteranion to the excited triplet state of the benzophenone moiety (BP^*3) was demonstrated.This reaction leads to the formation of benzophenone radical anion (6) and the boranyl radical, the latter of which dissociates rapidly to form butyl radical.The electron transfer rate was found to depend on the polarity of the solvent.In neat benzene, the short lifetime of the triplet obtained (300 +/- 150 ps) indicates an intramolecular process and that the compounds exist as tight ion pairs.The addition of 1percent MeCN increases the triplet lifetime up to 1.2 ns.The formation of a solvent-separated ion pair was suggested in this solvent mixture.However, in the polar solvent MaCN, 1 and 2 exist partially as free ions as determined by their dissociation constants.A diffusion-controlled intermolecular electron transfer process was shown in this solvent.The p-benzyloxybenzyl radical (BPCH2.), most likely formed by the C-N cleavage bond in the radical anion (6), was identified as an intermediate in the reaction pathway.After the initial electron transfer reaction, this produced tributylamine.The quantum yield obtained for the reaction of 1, 1E-2 M in MeCN, is 0.5 +/- 0.05.Triplet quenching as well as coupling of the radical anion (6) with the butyl radical deriving from decomposition of the borate was suggested to complete with the C-N cleavage reaction.

Full text of DOI:10.1021/jp960931q

12386
J. Phys. Chem. 1996, 100, 12386-12393
Photoinduced Inter- and Intra-Ion-Pair Electron Transfer Reactions in
N-(p-Benzoylbenzyl)-N,N,N-tri-n-butylammonium Triphenyl-n-butylborate and -Gallate
Salts
S. Hassoon,*,† A. Sarker, A. Y. Polykarpov, M. A. J. Rodgers, and D. C. Neckers*
Center for Photochemical Sciences,1 Bowling Green State UniVersity, Bowling Green, Ohio 43403
ReceiVed: March 27, 1996X
The photoreduction of N-(p-benzoylbenzyl)-N,N,N-tri-n-butylammonium triphenyl-n-butylborate (1) and gallate
(2) was studied using nano- and picosecond laser flash photolysis. An electron transfer reaction from the
borate or the gallate counteranion to the excited triplet state of the benzophenone moiety (BP*3) was
demonstrated. This reaction leads to the formation of benzophenone radical anion (6) and the boranyl radical,
the latter of which dissociates rapidly to form butyl radical. The electron transfer rate was found to depend
on the polarity of the solvent. In neat benzene, the short lifetime of the triplet obtained (300 ( 150 ps)
indicates an intramolecular process and that the compounds exist as tight ion pairs. The addition of 1%
MeCN increases the triplet lifetime up to 1.2 ns. The formation of a solvent-separated ion pair was suggested
in this solvent mixture. However, in the polar solvent MeCN, 1 and 2 exist partially as free ions as determined
by their dissociation constants. A diffusion-controlled intermolecular electron transfer process was shown in
•
this solvent. The p-benzoylbenzyl radical (BPCH2 ), most likely formed by the C-N cleavage bond in the
radical anion (6), was identified as an intermedate in the reaction pathway. After the initial electron transfer
reaction, this produced tributylamine. The quantum yield obtained for the reaction of 1, 10-2 M in MeCN,
is 0.5 ( 0.05. Triplet quenching as well as coupling of the radical anion (6) with the butyl radical deriving
from decomposition of the borate was suggested to compete with the C-N cleavage reaction.
Introduction
discrepancy between the theory and experiment.9 For the
intermolecular cases studied, electron transfer rates reached a
plateau value in the highly exothermic region and inverted
behavior is not seen down to a value of ∆Get ∼ 2.5 eV. The
intra-ion-pair electron transfer rates from borates to an exited
cationic acceptor are anticipated to be high and above the
diffusion rate limit. This makes these systems attractive for
investigation of electron transfer theories. The electron transfer
rates obtained for series of cyanine-tetraarylborates in benzene
were insensitive to the significant changes in the hard sphere
radius of the borate anion suggesting the formation of penetrated
ion pairs.10 In an attempt to obtain accurate and thermodynami-
cally meaningful values for the oxidation potentials of tet-
raarylborates, Schuster and co workers developed a kinetic
method11 and found that the application of the Rehm-Weller
relationship leads to an overestimation of the oxidation potentials
when compared with those obtained from the classical Marcus
relationship.
In a recent communication12 we have outlined preliminary
experiments dealing with the photoinduced electron transfer
reaction of N-(p-benzoylbenzyl)-N,N,N-tri-n-butylammonium
triphenyl-n-butylborate (1), in which the borate counteranion
is complexed to the para position of the benzophenone moiety.
Electron transfer from the triphenylbutylborate counteranion to
the excited triplet state of the benzophenone moiety (BP*3) was
demonstrated. The photoreduction of benzophenone by electron
donors such as amines has been studied extensively,13 and early
flash photolysis experiments confirmed the following mecha-
nistic scheme.14,15 Electron transfer from the amine to BP*3
forms the radical anion BP•- and the amino radical cation as a
contact ion pair (CIP). Subsequent reactions of the CIP can be
solvation to form a solvent-separated ion pair (SSIP), proton
transfer to form neutral radicals, or back electron transfer. The
extent of each depends on the rates of proton transfer and back
electron transfer in each solvent.16,17
Photoinduced electron transfer reactions represent an impor-
tant route to free radicals among whose uses is the initiation of
vinyl polymerization.2 It is apparent that borate anions are
especially beneficial electron transfer donors, and these have
many important applications in free-radical polymerizations
initiated by different electron acceptor sensitizers. The one-
electron photooxidation of triarylalkylborates forms a boranyl
radical which lives for a very short time and cleaves rapidly
yielding triarylborane and an alkyl radical, the latter initiating
polymerization. Schuster and his co-workers3,4 have studied
the photochemistry and photophysics of carbocyanine alkylt-
riphenylborate salts and shown that electron transfer occurs from
the borate to the singlet state of the cyanine cation. Since the
lifetime of the cyanine singlet is short (picoseconds), the said
electron transfer is observed only in nonpolar solvents where
the process can occur in the tight ion pair.5 Electron transfer
photooxidation of borates may also occur with uncharged
acceptors.6,7 We have recently reported that photoreduction of
fluorone dyes by triphenylbutylborate salts is an intermolecular
electron transfer reaction from the borate anion to the triplet
state of the dye that leads to the formation of butyl radicals
which initiate polymerization.7
Rate constants for electron transfer are related to the free
energy of the reactions (∆Get) according to classical Marcus
theory.8 An important prediction of Marcus theory is the
existence of the “inverted region”. The electron transfer rate
will increase with decreasing free energy of the reaction up to
a maximum and then decrease as ∆Get becomes more exoergic.
Considerable attention has been made to testing the Marcus
theory approach. Rehm and Weller were first to point out a
† Present address: Department of Chemistry, Technion-Israel Institute
of Technology, Haifa 32000, Israel.
X Abstract published in AdVance ACS Abstracts, July 1, 1996.
S0022-3654(96)00931-8 CCC: $12.00 © 1996 American Chemical Society

Photoinduced Electron Transfer Reactions
J. Phys. Chem., Vol. 100, No. 30, 1996 12387
SCHEME 1: Synthesis of Salts 1-4. (a) Tributylamine,
ether/chloroform, room temperature. (b) Lithium salt of
triphenylbutyl borate, H2O, room temperature.
(c) Lithium salt of triphenylbutylgallate, THF. (d) HBF4,
48 wt %.
O
C
O
C
Br
+
–
CH2Br
CH2 N(C4H9)3
a
3
–
O
C
Ph3B–C4H9
+
–
CH2 N(C4H9)3
b
Figure 1. Ground state absorption spectra of 10-3 M of 1 in MeCN.
1
–
O
–
Ph3Ga–C4H9
+
CH2 N(C2H5)3
O
C
+
–
3
C
CH2 N(C4H9)3
c
Ph3B–C4H9
2
5
O
–
BF4
+
2. Absorption and Emission Spectra. The photophysical
properties of all compounds (1-5) are determined by the
benzophenone (BP) chromophore as can be seen from their
absorption and emission spectra. Their absorption spectra above
300 nm are virtually identical and similar to that of benzophe-
none (Figure 1). The n-π* transitions are slightly red-shifted
by (∼5 nm) relative to benzophenone but the absorption
coefficients are similar (ꢀ ∼ 150 L mol-1 cm-1).
–
CH2 N(C4H9)3
C
d
4
Since electron transfer in (1) is to the long-lived triplet excited
state of the benzophenone moiety, the relative contributions of
the intermolecular or intramolecular processes are found to
depend on the polarity of the solvent. In neat benzene, the
process is very fast as indicated by a measured triplet lifetime
of 300 ( 150 ps which suggests an intramolecular process and
that the compound exists as a tight ion pair. However, in
acetonitrile the compound exists partially as free ions with their
amount relative to the tight pair depending on dissociation
constant and concentration. Accordingly both inter- and in-
tramolecular electron transfer processes were shown in this case.
The present work represents further mechanistic studies of
the benzophenone-borate (1) and the extension of these studies
to the benzophenone-gallate system (2). The unique elimina-
tion of a tri-n-butylammonium radical cation from the radical
anion formed after the initial electron transfer step (see Scheme
2) produces p-benzoylbenzyl radical and free amine. This was
further investigated by comparing the photochemistry to the case
where the borate counteranion was complexed to the meta
position of the benzophenone (5). In the latter case, since free
amine was also obtained, an elimination step forming an amino
radical cation and the anion of the ketone is not a necessary
step in the reaction mechanism, and the elimination of the free
amine occurs directly as a result of intramolecular electron
transfer in the radical anion.
Excitation of 10-3 M solutions of (1-5) at 335 nm in
acetonitrile (MeCN) and at room temperature results in very
weak fluorescence. This is not surprising knowing that
intersystem crossing in benzophenone is very fast (<10 ps) and
triplet quantum yield formation is ∼1. The fluorescence in the
bromide case (3) is ∼3 times less intense when compared to 1
at the above conditions. This can be related to the difference
in the dissociation constants of the two salts as will be discussed
later. The bromide salt under these conditions is less dissociated
and exists mostly as a tight ion pair while the borate salt exists
more as separated ions. Therefore, quenching of the singlet
state either by electron transfer or by enhanced intersystem
crossing will be more efficient in the bromide case.
The phosphorescence spectra obtained at 77 K in methanol/
ethanol (1:1) matrix for the compounds 1-4 were identical but
shifted 4 nm to the red relative to BP. The phosphorescence
quantum yields were about the same.
3. Transient Spectra and Triplet Decay Kinetics. A.
Nanosecond Laser Flash Photolysis. Nanosecond time-resolved
transient absorption spectroscopy with excitation at 355 nm
using the third harmonics of a Q-switched Nd:YAG laser was
used to investigate the interaction mechanism. The triplet decay
rates and transient absorption spectra of 1 and 4 were measured
in acetonitrile (MeCN), benzene, and methanol (MeOH). In 4
when no electron transfer is expected from the -BF4 counter-
anion to the excited triplet state of BP, the triplet transient
absorption spectrum was obtained in the three solvents. The
transient absorption spectrum of 10-3 M solution of 4 obtained
in benzene (1% MeCN by volume),18 0.12 and 4.2 µs after the
pulse is shown in Figure 2a. The two absorptions at 330 and
540 nm decay with the same rate and we assign them to the
triplet absorption of the benzophenone moiety in agreement with
literature.16,17,19 Within this time scale no triplet was detected
for 1 and 2 under the same conditions as can be seen for 1 in
Figure 2b, indicating a very fast quenching of the triplet. The
peak at 320 nm is assigned to the p-benzoylbenzyl radical
Results
1. Synthesis. The synthesis of the compounds 1-4 is
described in Scheme 1. 4-Bromomethylbenzophenone, when
treated with tributylamine in ether/chloroform gave N-(p-
benzoylbenzyl)-N,N,N-tri-n-butylammonium bromide (3). N-(p-
benzoylbenzyl)-N,N,N-tri-n-butylammonium triphenyl-n-butylb-
orate (1) and the gallate analog salt (2) were synthesized from
3 and the lithium salts of triphenylbutylborate or -gallate,
respectively, in water, while 4 was obtained by treating 3 with
fluoroboric acid in water.
N-(m-benzoylbenzyl)-N,N,N-tri-n-ethylammonium triphenyl-
n-butylborate (5) was prepared in similar way to (1) except that
3-bromomethylbenzophenone and triethylamine were used in
step (a). Details of the synthesis are given in the Experimental
Section.
•
(BPCH2 ) as will be discussed later (see Scheme 2). However,
the transient absorption spectra of 1 (10-3 M) in the polar solvent

12388 J. Phys. Chem., Vol. 100, No. 30, 1996
Hassoon et al.
Figure 3. Transient absorption spectra recorded at three delay times
after the laser pulse of 2 × 10-3 M (1:1) mixture of 4-methylben-
zophenone and tetrabutylammoniuom triphenyl-n-butylborate in MeCN.
Figure 2. Transient absorption spectra recorded at two delay times
after the laser pulse of (a) 10-3 M of 4 in benzene (1% MeCN), (b)
10-3 M of 1 in benzene (1% MeCN), and (c) 10-3 M of 1, in MeCN.
MeCN as given in Figure 2c shows that the triplet is forming
immediately after the pulse at 330 and 540 nm, but in contrast
to the case of 4 the two absorptions do not decay at the same
rate. While the peak at 530-540 nm decayed almost completely
after ∼1 µs, the peak at 330 nm does not decay, suggesting
that by quenching the triplet a new transient is forming at this
wavelength with longer lifetime. As mentioned above, this is
Figure 4. Transient absorption spectra recorded at three delay times
of (a) 10-3 M of 1 and (b) 10-3 M of 4 in MeOH.
spectra shown in Figure 4b, where only the transient absorbing
at 550 nm is formed after the triplet decay. In this case no
electron transfer is expected. The only possible quenching
reaction of the excited triplet state of the BP moiety would be
by hydrogen abstraction which leads to the formation of the
related ketyl radical. On the basis of these experiments and
comparing to the BP ketyl radical absorption reported in the
literature, we assign the absorption obtained at 550 nm as the
analog ketyl radical of 1 or 4.
We show the triplet decay rate obtained for 10-3 M solution
of 4 in MeCN as a function of concentrations of tetrabutylam-
monium triphenyl-n-butylborate and -gallate, respectively in
Figures 5 and 6. The triplet decay rate increases as a result of
the intermolecular electron transfer quenching with an increase
in concentration. A similar concentration dependence was
obtained when the triplet decay rate was measured for 1 and 2
at different concentrations. It is clear from these results that
the quenching process in this solvent is bimolecular which
means that the ions exist mostly as a separated ions. The
quenching rate constants obtained by linear fitting of the data
in Figure 5 and 6 are 8.88 × 109 and 1.47 × 109 M-1 s-1 for
1 and 2, respectively. The quenching rate constant is ∼6 times
faster in the borate case relative to the gallate case.
•
the BPCH2 radical formed in the reaction pathway (Scheme
2).
The transient absorption spectra obtained by photolysis of a
mixture (1:1) of 2 × 10-3 M of 4-methylbenzophenone (MeBP)
and tetrabutylammonium triphenyl-n-butylborate in MeCN are
shown in Figure 3. As the triplet at 530 nm decays, a transient
absorbing around 720 nm forms. We assign this transient to
the radical anion MeBP•- which is in agreement with the
literature reported values for BP with amines.16 For comparison,
we were unable to detect any absorption in that spectral range
for the expected analog radical anion from 1 or 2. Alcohols
are known to stabilize BP radical anions through hydrogen
bonding.14,16 The transient absorption spectra obtained for 1,
10-3 M in MeOH, at different time intervals, are shown in
Figure 4a. As the triplet absorption at 540 nm decays, two new
transients absorbing at 550 and 620-630 nm form. It has
previously been shown that BP•- absorption in MeOH is blue
shifted to ∼615 nm compared to its absorption in MeCN (∼720
nm), as a result of hydrogen bonding.14,16 Accordingly, we
assign the absorption at 620-630 nm in Figure 4a to the radical
anion (6) formed after the initial electron transfer step (Scheme
2). Under the same conditions, 4 gives the transient absorption

Photoinduced Electron Transfer Reactions
J. Phys. Chem., Vol. 100, No. 30, 1996 12389
Figure 7. Transient absorption spectra of 1, 5 × 10-3 M in benzene
Figure 5. Concentration dependence of the triplet decay rate of the
benzophenone moiety in 1 in MeCN. Inset depicts the nonlinear
concentration dependence.
(1% MeCN) recorded at 60, 320, and 800 ps after the pulse.
TABLE 1: Quantum Yields for the Photolysis Reaction of
Aerated and Degassed 10-2 M Solutions of (1) in MeCN
degassed
aerated
compound
((10%)
((10%)
1
0.50
0.14
0.12
0.52
0.065
0.057
(BPCH2)2 (9)
BPCH2Bu (10)
this absorption to the BPCH2• radical (8) in agreement with the
reported absorption of related benzyl radicals.20
The quantum yields of disappearance of 1 in MeCN were
measured using an actinometric method based on the known
benzophenone-benzhydrol reaction.21 Though the original
measures were reported for benzene we had to use MeCN due
to the low solubility of the parent compound in benzene. Results
were corrected according to the differences in the absorption
coefficients of the compound in the two solvents and are
summarized in Table 1.
Figure 6. Concentration dependence of the triplet decay rate of the
benzophenone moiety in 2 in MeCN. Inset depicts the nonlinear
concentration dependence.
Two main observations result from the data reported in Table
1. First, the quantum yield for the disappearance of 1 is only
0.5 and is independent in the presence of oxygen. Second, the
quantum yield of formation of products 9 and 10 is oxygen
dependent. On the other hand, photolysis of 4 under the same
conditions, but for longer time (20 min compared to 1 min
photolysis with 1 gives quantum yield of 0.003 for the final
product (10) in the degassed sample. Thus there is almost no
reaction in this case.
B. Picosecond Laser Flash Photolysis. Failure to observe
the triplet transient from 1 in benzene and the radical anion (6)
in MeCN and benzene in the nanosecond experiment may
indicate that their lifetimes are less than the instrument response
time ca 10 ns. The triplet of 2 is quenched very rapidly in
benzene, probably as a result of very fast electron transfer in
this solvent. A picosecond flash experiment was used to
measure the rate of electron transfer in 1 in neat benzene, though
we were able to detect an immediate formation of a transient
having λmax ) 540 nm which we ascribe to the triplet absorption
spectrum of the benzophenone moiety. Since the parent
compound has poor solubility in benzene, the absorbance at 355
nm, the excitation wavelength, was limited. This produced
absorbance signals of poor signal-to-noise ratio which precluded
measurement of a precise lifetime, but our best estimate was
300 ( 150 ps. The addition of 1% (by volume) of MeCN to
benzene caused the solubility to increase such that it was
possible to obtain good kinetic data under these conditions. This
led to a measurement of the lifetime of the triplet in this solvent
mixture of 1.2 ns (Figure 7). However, under these conditions
we still could detect no radical anion.
Discussion
The triplet decay rates in 1 and 2 in MeCN are found to be
bimolecular from their concentration dependence (Figures 5 and
6). In the polar solvent MeCN (ꢀ ) 36), 1 or 2 probably exist
in the form of free ions (depending on their equilibrium
constants) and the quenching process would thus be intermo-
lecular. The observation that almost no triplet quenching or
radical formation occurs in 4 allows us to assume that the triplet
quenching observed in 1 and 2 is due to an electron transfer
process. Accordingly, the rate constants for the intermolecular
electron transfer would be 8.88 × 109 and 1.47 × 109 M-1 s-1
for 1 and 2, respectively, obtained by linear fitting of the data
in Figures 5 and 6. On the other hand, in the nonpolar solvent
benzene (ꢀ ) 2.3) it may be that the compounds exist as tight
ion pairs, in which case the electron transfer process would be
intramolecular and probably the triplet decay is complete on a
subnanosecond time scale as shown in Figure 2. In the
picosecond experiment, the triplet absorption of 1 in benzene
was observed with approximately 300 ps lifetime. This supports
the contention that intramolecular decay is occurring. No
bimolecular reaction would be possible at such short time and
at 1 × 10-4 M, the concentration used in this experiment. The
4. Production Analysis and Quantum Yields. Irradiation
of 10-2 M of 1 in MeCN and MeOH gives mainly p,p′-
(benzoyl)bibenzyl (9) (Scheme 2). Compounds 9 and 10, as
well as tributylamine, were isolated and identified from the
reaction mixture (see Experimental Section). This implicates
the BPCH2• radical (8), an intermediate most likely formed by
the dissociation of the radical anion (6) along with the
tributylammonium radical cation, in the reaction pathway (see
Scheme 2). In support of this we also observed an absorption
peak formed with long lifetime around 320 nm in the transient
absorption spectra as can be seen in Figure 2b, and c. We assign

12390 J. Phys. Chem., Vol. 100, No. 30, 1996
Hassoon et al.
SCHEME 2
O–
3
O
C
*
ket
•
–
Ph CH2N+
(Bu)3
Ph
Ph CH2N+
(Bu)3
B
(Ph)3
+
(Ph)3BBu
Ph
C
•
Bu
kdis
6
initiate
•
(Ph)3B + Bu
kdis
polymerization
O–
Ph
C
CH2 + N–(Bu)3
7
kbet
O
C
•
+
Ph CH2
N–(Bu)3
Ph
8
coupling reactions
O
O
O
+
Ph
C
Ph CH2 CH2 Ph
C
Ph
Ph
C
Ph CH2 Bu
10
9
addition of 1 vol % MeCN to benzene increases the lifetime of
the triplet to 1.2 ns, again consistent with the process being
intramolecular. It seems that the triplet decay rate increases as
the polarity of the solvent decreases. The literature contains a
number of reports dealing with the nature of radical ion pair
systems obtained by photoinduced electron transfer reactions
in various solvents.22 In general, in less polar solvents the
formation of a contact ion pair (CIP) was confirmed while
increasing the polarity of the solvent leads to the formation of
solvent-separated ion pair (SSIP). In a similar way, we could
assume that the addition of 1% MeCN to benzene results in
changing the structure of 1 or 2 from a tight ion pair form to
the solvent-separated ion pair form. This, in turn, may reflect
the increase in the triplet decay rate as given above.
The bimolecular triplet quenching rate constants in 1 and 2
in MeCN, as reported above, were obtained by linear fitting of
the data in Figures 5 and 6. However, a closer look at the data
shows that the concentration dependence of the triplet decay
rates is not linear as shown in the insets in Figures 5 and 6.
This can be explained in terms of an equilibrium between
dissociated ions and the tight ion pair in these systems. The
data in Figures 5 and 6 were obtained by adding increasing
amounts of triphenylbutylborate or -gallate counteranions to a
fixed concentration of 4. As the concentration of the borate or
the gallate counteranion increases, the concentration of the N-(p-
benzoylbenzyl)-N,N,N-tri-n-butylammonium cation will decrease
and the tight ion pair will form depending on the equilibrium
constant (eq 1).
This indicates that the concentration of the benzophenone moiety
is decreasing under these conditions as well. This implies that
the actual concentration of the borate or the gallate anions should
be smaller than the initially added concentration as was plotted
in Figures 5 or 6, and the data points should be shifted to lower
concentrations accordingly. At very low concentrations of the
borate or the gallate counteranions, we could assume that the
concentration of the tight ion pair is very small. Therefore, if
we take the slope of the straight line between the two lowest
concentrations data points in Figures 5 and 6, the triplet
quenching rate constants will be 1.73 × 1010 and 3.29 × 109
M-1 s-1, respectively, and the electron transfer rates would be
diffusion controlled.
The proposed mechanism for the photoreduction of the borate
compound (1) is given in Scheme 2. The same mechanism
pertains for the gallate case (2). Electron transfer from the
borate anion to the excited benzophenone moiety BP*3 gives
the radical anion (6) and the boranyl radical both of which
dissociate rapidly. The estimated lifetime for the triphenylbu-
tylboranyl radical is 250 fs11 and it dissociates to give triph-
enylborane and the butyl radical. Such a short lifetime implies
irreversible dissociation and back electron transfer to the radical
anion to form 1 or 2 would be virtually nonexistent reaction.
On the other hand, we were unable to detect the radical anion
(6) in benzene even in the picosecond experiment. However,
in MeOH where hydrogen-bonding stabilization occurs, it was
possible to detect its transient absorption around 620 nm on
the nanosecond time scale (Figure 4). Benzophenone is known
to be H-bonded to alcohols in the ground state and one should
expect that the radical anion to be H-bonded as its forms. A
possible explanation for the H-bonding effect given by Fes-
senden et al.16 was based on the fact that the lowest excited
triplet state (n-π*) of benzophenone has a lower dipole moment
than the ground state.23 Excitation will shift the electron density
away from the oxygen atom and this could disrupt the
H-bonding. As the triplet state is quenched by electron transfer
to form the radical anion the H-bonding will again be formed,
and this results in the blue shift in its absorption.
O
–
–
C
CH2 N(C4H9)3 + n-Bu B(Ph)3
+
O
C
–
(1)
CH2 N(C4H9)3
+
–
n-Bu B(Ph)3
In the nanosecond experiment, one can follow the quenching
rate of the triplet of the benzophenone moiety in the separated
form only since that process is bimolecular. The quenching
rate is too fast in the tight ion pair and canot be seen. In fact,
we have observed that the initial absorption immediately after
the laser pulse in the transient triplet decay experiments
decreases as the concentration of the counteranion increases.
The observation of the tributylamine along with the coupling
products (9) and (10) in the reaction mixture suggested that the
radical anion (6) reacts rapidly to give the p-benzoylbenzyl
•
radical BPCH2 (8) and the amine. The formation of the radical
(8) was confirmed by its transient absorption spectra at 320 nm

Photoinduced Electron Transfer Reactions
J. Phys. Chem., Vol. 100, No. 30, 1996 12391
SCHEME 3
–
(Ph)3B
O–
C
O
O
(Ph)3B
Ph CH2N+
(Bu)3
•
Ph
C
Ph CH2N+
(Bu)3
Ph
C
+
Ph
Ph CH2N+
(Bu)3
Bu
•
Bu
Bu
OH
(Ph)3B
–
Ph CH2N+
(Bu)3
(Ph)3BOH
Ph
C
H2O
Bu
(Figure 2). In fact, the elimination of the amine from 6 can
occur either by intramolecular electron transfer followed by
immediate cleavage of the carbon-nitrogen bond or through
the formation of the ketone anion (7) and the amino radical
cation as shown in Scheme 2. At this point, we have no
spectroscopic evidence for the formation of either the amino
radical cation or the ketone anion. The free energy estimated
from the reported redox potentials of the amino cation radical
and related benzyl radicals24 predicts that the electron transfer
reaction from ketone anion to the amino cation radical would
be very exothermic (∆G ≈ -2 V) which means that they live
for a very short time and are not detectable under the present
In the actual experiment wherein the quantum yields were
measured, the change in the concentration of borate (1) after
photolysis was estimated from the decrease in its peak area
obtained by HPLC analysis. Unfortunately, under our detection
conditions we were unable to distinguish the different borate
salts. Accordingly, one possible explanation of the low
measured quantum yield is that there is another reaction in which
the radical anion (6) re-forms 1 or another borate salt which
competes with the elimination reaction that releases the amine.
If this is the case the amount of borate (1) which reacts during
the photolysis experiment would be larger than what is obtained
from the HPLC measurement and accordingly the quantum yield
would be higher.
•
experimental conditions. The p-benzoylbenzyl radical (BPCH2 )
In Scheme 3 we describe the coupling reaction between the
butyl radical (formed by the dissociation of the boranyl radical)
with the radical anion as a possible reaction. Similar coupling
reactions have been reported in the literature.25
is stable and couples either with a radical like itself or with the
butyl radical.
In an attempt to further investigate the driving force for this
elimination process, 5, in which complexation of the borate
counteranion is meta to the benzoyl group, was synthesized.
The elimination of the amino radical cation and the formation
of the ketone anion (7) in the para-substituted compound should
be facilitated by aromatic conjugation. This would be impos-
sible in the meta-substituted compound and elimination would
lead to the formation of very unstable intermediate. Therefore,
the dissociation of the radical anion to form the meta analog
anion of 7 and the amino radical cation would be unlikely. The
results obtained for 5, however, were similar to those obtained
for 1, thus indicating the substitution pattern has no effect on
the outcome. Triethylamine was formed along with the expected
coupling products. This result indicates that the dissociation
step of the radical anion to form the ketone anion and the amino
radical cation is not necessary, though it still might be operative
with the para compound (1).
The quantum yield for disappearance of 1 (10-2 M in MeCN)
measured in the steady state experiment is 0.5 ((10%). The
actinometric method used is based on the assumption that the
triplet quantum yield is unity in BP. Since the phosphorescence
spectra obtained for compounds (1-4) were comparable to that
of BP one would assume that the triplet quantum yields in these
compounds are also similar. At the concentration (10-2 M) used
in this experiment, as we have discussed before, the borate salt
would be partially in its tight pair form and the electron transfer
reaction would be partially intramolecular and very fast. The
back electron transfer is expected to be minor as result of the
rapid dissociation of the boranyl radical. Under these conditions
one would expect a higher quantum yield for the disappearance
of 1. For the cyanine triphenylalkylborate, the dissociation
constant obtained from conductivity measurements is 3.2 × 10-2
M in MeCN.4 Assuming a similar dissociation constant for 1,
the degree of dissociation at 10-2 M concentration would be
0.8. Therefore, bimolecular quenching such as triplet-ground
state quenching may be responsible for lowering the quantum
yield if the rate is comparable to the electron transfer rate in
this solvent.
As shown in the scheme the coupling reaction would lead to
the formation of a zwitterionic intermediate. From this inter-
mediate no elimination reaction is possible, but it can react with
Ph3B in the presence or absence of water to give, in both cases,
a borate salt. Under our experimental conditions we could not
isolate and confirm the formation of these borate salts.
The observation that the same quantum yield is obtained for
the disappearance of 1 in the presence or the absence of oxygen
indicates that quenching of the triplet by oxygen could not
compete with the electron transfer reaction. In fact, under our
experimental conditions, we observe no effect of oxygen on
the triplet decay of borate (1) at concentrations higher than 10-3
M. At lower concentrations, where the borate salt exists more
as separated ions, the effect of oxygen becomes more significant.
The presence of oxygen significantly impacts on the formation
of the coupling products 9 and 10 as shown in Table 1. We
observed this effect in the flash experiment where the presence
of oxygen quenches the long lifetime part of the decay curve.
This implies that oxygen reacts with the p-benzoylbenzyl radical
(8) in competition with the coupling reactions. This result is
not unanticipated since the oxidation of radicals by oxygen is
well-known. Thus, we did not investigate the products under
these conditions further.
Conclusions
In this paper we report a new photoinitiator system developed
to initially involve the triplet state of an aromatic ketone
(benzophenone) complexed to borate or gallate anions. Electron
transfer from the borate or the gallate counteranions to the
benzophenone moiety leads to the formation of a radical, an
amine and a Lewis acidsthe original triphenylboron or -gallium
derivative. Since the electron transfer is to the long-lived triplet
excited state, the relative contributions of intermolecular or
intramolecular processes were shown to depend on the polarity
of the solvent. In the nonpolar solvent (benzene) the electron
transfer process is intramolecular and the electron transfer rate

12392 J. Phys. Chem., Vol. 100, No. 30, 1996
Hassoon et al.
constant is very fast far behind the diffusion rate. In the polar
solvent (MeCN) the electron transfer rate constant is diffusion
controlled.
products formed and the disappearance of 1 were determined
using HPLC method calibrated with the synthesized standard
compounds.
Picosecond and nanosecond laser flash photolysis, as well
as steady state experiments, product analysis, and quantum
yields measurements, were used to study the interaction mech-
anism. In these systems the elimination of tertiary amine
through a C-N bond cleavage in the radical anion formed by
the electron transfer reaction was shown. This unique elimina-
tion has already shown a potential use in photoimaging
applications.26
Synthesis. Melting point determinations (uncorrected) were
made using a Thomas Hoover capillary melting point apparatus.
Elemental analysis was performed by Atlantic Microlab, Inc.
Proton NMR spectra were recorded on 200 and 400 NMR
spectrometers. NMR spectrum resonance positions are reported
relative to the tetramethylsilane standard on the δ scale in parts
per million (ppm). 11B NMR were recorded on Varian
Unity+400 (128 MHz) with B(OMe)3 as the external standard.
Unless otherwise mentioned, starting reagents and solvents were
obtained from Aldrich Chemical Co. and used as received.
Experimental Section
N-(p-Benzoylbenzyl)-N,N,N-tri-n-butylammonium Bro-
mide (3). 4-Bromomethylbenzophenone31 (10.0 g; 36.34 mmol)
was dissolved in ether-chloroform (2:1) mixture at room
temperature. Tributylamine (8.76 g; 47.26 mmol) was added
and the solution stirred at room temperature for 24 h during
which a precipitate was formed. The precipitate was filtered
and recrystallized from ethanol-ethyl acetate (1:10) mixture;
11.70 g (70% yield) of white crystals was obtained: mp 160-
161 °C. 1H NMR (acetone, δ): 7.90 (m, 4 H), 7.68 (m, 1 H),
7.58 (m, 4 H), 5.11 (s, 2 H), 3.50 (m, 6 H), 2.06 (m, 6 H), 1.43
(m, 6 H), 1.00 (t, J ) 7 Hz, 9 H). Anal. Calcd for C26H39-
BrNO: C, 67.86; H, 8.26; Br, 17.37. Found: C, 67.79; H, 8.28;
Br, 17.43.
General. Absorption spectra were recorded with a Hewlett
Packard 8452A diode array UV-vis spectrophotometer. Fluo-
rescence and phosphorescence spectra were measured using
SPEX Fluorolog spectrophotometer. High performance liquid
chromatography (HPLC) was used for photoproduct analysis.
A HP 1050 Series instrument equipped with a diode-array
detector (DAD) was used. A good separation between borate
or gallate salts and their photoproducts was obtained using a
reverse phase (HP LiChrosorb RP, 10 µm, 200 × 4.6 mm)
column. Acetonitrile was used as eluent at 0.5 mL/min flow
rate and 40 °C column temperature. All solvents used are HPLC
grade obtained from Aldrich Chemical Co.
Flash Photolysis. Nanosecond laser flash photolysis were
carried out using the apparatus described previously.7,27 The
third harmonics (355 nm) of a Q-switched Nd:YAG laser
(Surelite I), giving approximately 80-90 mJ/pulse and a pulse
width of 8 ns, was used for excitation. Only a small part of
the pulse energy was used when measuring the transient decay
kinetics to eliminate other decay processes such as T-T
annihilation. Sample solutions were contained in 1 cm quartz
cuvettes and were deaerated by continuous purging with argon
during the experiment in the cases where oxygen had to be
removed.
N-(p-Benzoylbenzyl)-N,N,N-tri-n-butylammonium Tetra-
fluoroborate (4). N-(p-Benzoylbenzyl)-N,N,N-tri-n-butylam-
monium bromide (3) (0.21 g; 0.45 mmol) was dissolved in 10
mL of water at 50-55 °C. Undissolved dusty materials were
filtered out (if any present). Fluoroboric acid (0.5 mL 48 wt
%) in water was added at 50 °C dropwise with stirring. A
precipitate was formed immediately. Water (10 mL) was added
and stirred for 30 min. Product was filtered and washed with
water; 0.16 g (80% yield) of white solid was obtained: mp 78-
79 °C. Anal. Calcd for C26H39BF4NO: C, 66.86; H, 8.14; Br,
0.0. Found: C, 66.93; H, 8.19; Br, 0.0.
The third harmonic (355 nm) of a Nd:YAG active/passive
mode-locked oscillator amplified through a double-pass Nd:
YAG amplifier was used for excitation in the picosecond
experiment. Pulse widths were ca. 30 ps with ∼75 mJ/pulse at
1064 nm. The basic features of the pump-continuum probe
coupled with a diode array spectrography system are as
described earlier.28 Multiple wavelength and single wave-
length29 were used to obtain the kinetic waveforms. The sample
solutions were ∼10-4 M in neat benzene or 5 × 10-3 M in
benzene (1% MeCN) and were flowed through a 1 cm quartz
cuvette.
N-(p-Benzoylbenzyl)-N,N,N-tri-n-butylammonium Tri-
phenyl-n-butylborate (1). Lithium triphenyl-n-butylborate32
0.80 g (2.61 mmol) was dissolved in 9 mL of water and
undissolved particles, if any were filtered out. N-(p-Benzoyl-
benzyl)-N,N,N-tri-n-butylammonium bromide (1.22 g; 2.65
mmol) in 25 mL water-methanol (5:1) was added to the borate
solution dropwise with stirring at room temperature. A white
solid was formed immediately and stirring was continued for
another 1 h after the addition of 25 mL of water. The white
solid was filtered and dried overnight. After recrystallization
from ethylacetate/ethanol, 1.20 g (68%) of white crystals were
obtained: mp 133-133.5 °C. 1H NMR (acetone, δ): 7.92 (m,
2 H), 7.80 (m, 4 H), 7.68 (m, 1 H), 7.59 (m, 2 H), 4.73 (s, 2 H,
N-CH2-ph), 3.36 (m, 6 H, N-CH2-CH2), 1.95 (m, 6 H,
N-CH2-CH2), 1.45 (m, 6 H, N-C2H4-CH2), 1.23 (m, 2 H,
B-CH2), 1.02 (m, 13 H, N-CH3 and B-CH2-C2H4), 0.77 (t,
3 H, B-CH3). 11B NMR (DMSO): d -29.41. Anal. Calcd
for C48H62BNO: C, 84.87; H, 9.13; N, 2.06. Found: C, 84.93;
H, 9.21; N, 2.00.
Steady State Photolysis and Quantum Yields. Photolysis
were carried out in a Rayonet model photochemical reactor fitted
with filter-coated mercury lamps (300-400 nm) with output
maximum at 365 nm. The photoproducts were separated using
silica gel column chromatography with hexane-ethyl acetate
(9:1) and characterized by NMR, HPLC, and mass spectroscopy.
The main products 9 and 10 were synthesized independently30
and these samples were used for comparison.
The actinometer solution (2 mL) containing benzophenone
and benzhydrol, 0.1 M each, in benzene was degassed with three
freeze-pump-thaw cycles. The sample was irradiated for 20
min and the amount of the reacting benzophenone was
determined from the change in its absorption at 344 nm. Using
the calibrated φact ) 0.68 from ref 21, the light intensity was
calculated to be 3.45 × 1018 photons/min. Under the same
conditions, 10-2 M solution of borate (1) was irradiated for 1
min only to avoid secondary reactions. The concentration of
N-(p-Benzoylbenzyl)-N,N,N-tri-n-butylammonium Tri-
phenyl-n-butylgallate (2). Triphenylgallium33 (1.88 g; 6.25
mmol) was dissolved in 16 mL of dry benzene under a nitrogen
atmosphere, and 2.5 mL (2.5 M) of butyllithium was added into
it at room temperature. A white precipitate was formed
immediately and stirring was continued for 2 h under nitrogen.
Dry hexane (15 m) was added and stirred for another 1 h. The
precipitate was filtered and dried under vacuum. Lithium salt
of triphenyl-n-butylgallate (1.50 g; 66% yield) was obtained as

Photoinduced Electron Transfer Reactions
J. Phys. Chem., Vol. 100, No. 30, 1996 12393
(11) (a) Murphy, S. T.; Zou, C.; Mires, J. B.; Ballew, R. M.; Dlott, D.
D.; Schuster, G. B. J. Phys. Chem. 1993, 97, 13152. (b) Murpy, D.;
Schuster, G. B. J. Phys. Chem. 1995, 99, 511.
(12) Hassoon, S.; Sarker, A.; Rodgers, M. A. J.; Neckers, D. C. J. Am.
Chem. Soc. 1995, 117, 11369.
(13) (a) Inbar, S.; Linschitz, H.; Cohen, S. G. J. Am. Chem. Soc. 1980,
102, 1419. (b) Inbar, S.; Linschitz, H.; Cohen, S. G. J. Am. Chem. Soc.
1981, 103, 1048.
(14) (a) Simon, J. D.; Peters, K. S. J. Am. Chem. Soc. 1982, 104, 6542.
(b) Schaefer, C. G.; Peters, K. S. J. Am. Chem. Soc. 1980, 102, 1566.
(15) (a) Hoshino, M.; Seki, H.; Shizuka, H. Chem. Phys. 1989, 129,
395. (b) Hoshino, M.; Shizuka, H. J. Phys. Chem. 1987, 91, 714.
(16) (a) Devadoss, C.; Fessenden, R. W. J. Phys. Chem. 1990, 94, 4540.
(b) Devadoss, C.; Fessenden, R. W. J. Phys. CHem. 1991, 95, 7253.
(17) Yamaji, M.; Kiyota, T.; Shizuka, H. Chem. Phys. Lett. 1994, 226,
199.
white powder, mp >250 °C; 0.56 g (1.45 mmol) of the lithium
salt of triphenyl-n-butylgallate and 0.56 g (1.21 mmol) of 3 were
dissolved in anhydrous THF. The solution was stirred overnight
under nitrogen at room temperature. The solution was con-
centrated by evaporating solvent under vacuum, and the product
was precipitated by adding water. After drying under vacuum,
0.42 g (46%) white solid was obtained. Anal. Calcd for C48-
H62GaNO: C, 78.09; H, 8.40; N, 1.90. Found: C, 77.43; H,
8.21; N, 1.89.
N-(m-Benzoylbenzyl)-N,N,N-tri-n-ethylammonium Tri-
phenyl-n-butylborate (5). The compound was synthesized as
(1) in overall yield of 12% after recrystallization. mp 107-
108 °C (from acetone). 1H NMR (DMSO-d6, 200 MHz) d 7.78
(m, 7 H), 7.61 (m, 2 H), 7.19 (br s, 6 H), 6.89 (t, J ) 7 Hz, 6
H), 6.73 (t, J ) 7 Hz, 3 H), 4.57 (s, 2 H), 3.19 (q, J ) 7 Hz,
6 H), 1.29 (t, J ) 7 Hz, 9 H), 1.18 (m, 2 H), 0.85 (m, 4 H),
0.75 (t, J ) 7.4 Hz, 3 H). 11B NMR (DMSO, 128 MHz) δ
-28.77. Anal. Calcd for C42H50BNO: C, 84.69; H, 8.46; N,
2.31. Found: C, 84.71; H, 8.42; N, 2.31.
(18) The addition of 1% MeCN by volume to benzene is to increase
solubility.
(19) Murov, S. L.; Carmichael, I.; Hug, G. L. Handbook of Photochem-
istry, 2nd ed.; Marcel Dekker: New York, 1993; p 107-8.
(20) Tokumuru, K.; Ozaki, T.; Nosaka, H.; Saigusa, Y.; Itoh, M. J. Am.
Chem. Soc. 1991, 113, 4974 and references therein.
(21) (a) Moore, W. M.; Hammond, G. S.; Foss, R. P. J. Am. Chem.
Soc. 1961, 83, 2789. (b) Moore, W. M.; Ketchum, M. J. Am. Chem. Soc.
1962, 84, 1368.
(22) Gould, I. R.; Young, R. H.; Moody, R. E.; Farid, S. J. Phys. Chem.
1991, 95, 2068 and references therein.
(23) Fessenden, R. W.; Carton, P. M.; Shimamori, H.; Scaiano, J. C. J.
Phys. Chem. 1982, 86, 3803.
(24) (a) Wayner, D. D. M.; McPhee, D. J.; Griller, D. J. Am. Chem.
Soc. 1988, 110, 132. (b) Mann, C. K.; Barnes, K. K. Electrochemical
Reactions in Non-Aqueous Systems; Marcel Dekker: New York, 1970.
(25) Lan, J. Y.; Schuster, G. B. J. Am. Chem. Soc. 1985, 107, 6710.
(26) Mejiritski, A.; Polykarpov, A. Y.; Sarker, A. M.; Neckers, D. C.
Chem. Mater., in press.
Acknowledgment. This work was supported by the National
Science Foundation (DMR-9013109) and by the Office of Naval
Research (N00014-93-0772). The authors are most grateful for
this support. We also thank Ms. Elisabeth A. Morlino for her
technical assistance with the picosecond experiments.
References and Notes
(1) Contribution No. 247 from the Center for Photochemical Sciences.
(2) Eaton, D. F. Pure Appl. Chem. 1984, 56, 1191.
(3) Chatterjee, S.; Gottschalk, P.; Davis, P. D.; Schuster, G. B. J. Am.
Chem. Soc. 1988, 110, 2326.
(4) Chatterjee, S.; Davis, P. D.; Gottschalk, P.; Kurz, M. E.; Sauerwein,
B.; Yang, X.; Schuster, G. B. J. Am. Chem. Soc. 1990, 112, 6329.
(5) Schuster, G. B.; Yang, X.; Zou, C.; Sauerwein, B. J. Photochem.
Photobiol. A: Chem. 1992, 65, 191.
(6) Lan, J. Y.; Schuster G. B. J. Am. Chem. Soc. 1985, 107, 6710.
(7) Hassoon, S.; Neckers, D. C. J. Phys. Chem. 1995, 99, 9473.
(8) (a) Marcus, R. A. J. Phys. Chem. 1956, 24, 966. (b) Marcus, R.
A. Annu. ReV. Phys. Chem. 1964, 15, 155.
(9) Rehm, D.; Weller, A. Isr. J. Chem. 1970, 8, 259.
(10) Murphy, S.; Yang, X.; Schuster, G. B. J. Org. Chem. 1995, 60,
2411.
(27) Rodgers, M. A. J. In Primary Photo-Processes in Biology and
Medicine; Bensasson, R. V., Jori, G., Truscott, T. G., Eds.; Plenum: New
York, 1985; pp 1-23.
(28) Hubig, S. M.; Rodgers, M. A. J. In Handbook of Organic
Photochemistry; Scaiano, J. C., Ed.; CRC Press: Boca Raton, FL, 1989;
Vol. I, Chapter 13.
(29) Loganov, S. L.; Rodgers, M. A. J. J. Phys. Chem. 1992, 96, 2915.
(30) McCombs, D. A.; Menon, C. S.; Higgins, J. J. Polym. Sci., Part I
1971, 9, 1261.
(31) Itoh, T.; Hall, H. K. Jr. Macromolecules 1990, 23, 4879.
(32) Damico, R. J. Chem. Soc., Perkin Trans. 1964, 1971.
(33) Miller, S. B.; Jelus, B. L.; Smith, J. H.; Munson, B.; Brill, T. B. J.
Organomet. Chem. 1979, 170, 9.
JP960931Q