Reaction of triarylphosphine radical cations generated from photoinduced electron transfer in the presence of oxygen

Article abstract of DOI:10.1021/jo049032l

(Chemical Equation Presented) The 9,10-dicyanoanthracene (DCA)-sensitized photoreaction of triarylphosphines (1) was carried out in acetonitrile under aerobic conditions. Phosphine 1 was oxidized to the corresponding phosphine oxide with no appreciable side reactions. Product analysis and laser flash photolysis experiments suggest that the radical cation of 1 formed by the electron transfer from 1 to DCA in the singlet excited state ( ¹DCA*) reacts with O₂ to eventually afford the phosphine oxide.

Full text of DOI:10.1021/jo049032l

Reaction of Triarylphosphine Radical Cations Generated from
Photoinduced Electron Transfer in the Presence of Oxygen
Shinro Yasui,*,† Sachiko Tojo,‡ and Tetsuro Majima*,‡
Laboratory of Biology and Chemistry, Tezukayama University, 3-1 Gakuen-Minami,
Nara 631-8585, Japan, and Institute of Scientific and Industrial Research,
Osaka University, 8-1 Mihogaoka, Ibaraki, Osaka 567-0047, Japan
yasui@tezukayama-u.ac.jp; majima@sanken.osaka-u.ac.jp
Received June 9, 2004
The 9,10-dicyanoanthracene (DCA)-sensitized photoreaction of triarylphosphines (1) was carried
out in acetonitrile under aerobic conditions. Phosphine 1 was oxidized to the corresponding
phosphine oxide with no appreciable side reactions. Product analysis and laser flash photolysis
experiments suggest that the radical cation of 1 formed by the electron transfer from 1 to DCA in
the singlet excited state (1DCA*) reacts with O2 to eventually afford the phosphine oxide.
Introduction
phonite PhP(OR)2, diphenylphosphinite Ph2POR, and
triphenylphosphine Ph3P during the reaction with di-
azonium salts in the dark undergo an ionic reaction with
methanol or ethanol in the solvent.3 Meanwhile, the
radical cation resulting from phosphite (RO)3P in this
reaction undergoes, besides the ionic reaction, radical
coupling with the aryl radical simultaneously generated
in the system. The predominance of the ionic reaction
for aromatic trivalent phosphorus radical cations was
interpreted in terms of the delocalization of the spin into
the phenyl substituent(s), which makes the central
phosphorus atom less radical-like. However, since the
reaction was carried out in an alcoholic solvent, the rapid
ionic reaction with the alcohol could have obscured the
radical reaction. A question then remained unsolved as
to whether the aromatic trivalent phosphorus radical
cations can also react with radical species under certain
circumstances.
Trivalent phosphorus compound is known to be a
nucleophile, whereas it behaves as an electron donor
toward good electron acceptors either in the ground1-5
or excited state,1,6-9 producing the corresponding triva-
lent phosphorus radical cations. It is of great importance
to elucidate the reactivity of the radical cations not only
from a mechanistic point of view but also for potential
application in organic syntheses.10 Our previous work
showed that radical cations generated from phenylphos-
† Tezukayama University.
‡ Osaka University.
(1) Yasui, S. Rev. Heteroatom Chem. 1995, 12, 145-161 and
references therein.
(2) Yasui, S.; Shioji, K.; Tsujimoto, M.; Ohno, A. J. Chem. Soc.,
Perkin Trans. 2 1999, 855-862.
(3) (a) Yasui, S.; Shioji, K.; Ohno, A. Tetrahedron Lett. 1994, 35,
2695-2698. (b) Yasui, S.; Shioji, K.; Ohno, A. Heteroatom Chem. 1995,
6, 223-233.
(4) Yasui, S.; Itoh, K.; Tsujimoto, M.; Ohno, A. Bull. Chem. Soc. Jpn.
2002, 75, 1311-1318.
Photoexcitation of 9,10-dicyanoanthracene (DCA) to
DCA in the singlet excited state (1DCA*) in polar solvents
provides a powerful method to examine the reactivity of
the radical cations of various compounds11,12 because of
(5) Powell, R. D.; Hall, C. D. J. Am. Chem. Soc. 1969, 91, 5403-
5404.
(6) Nakamura, M.; Miki, M.; Majima, T. J. Chem. Soc., Perkin Trans.
2 2000, 1447-1452.
(7) Yasui, S.; Shioji, K.; Ohno, A.; Yoshihara, M. J. Org. Chem. 1995,
60, 2099-2105.
(11) (a) Nakamura, M.; Dohno, R.; Majima, T. J. Chem. Soc., Chem.
Commun. 1997, 1291-1292. (b) Nakamura, M.; Dohno, R.; Majima,
T. J. Org. Chem. 1998, 63, 6258-6265. (c) Nakamura, M.; Miki, M.;
Majima, T. Bull. Chem. Soc. Jpn. 1999, 72, 2103-2109.
(12) (a) Ganapathy, S.; Dockery, K. P.; Sopchik, A. E.; Bentrude,
W. G. J. Am. Chem. Soc. 1993, 115, 8863-8864. (b) Shukla, D.; Lu,
C.; Schepp, N. P.; Bentrude, W. G.; Johnston, L. J. J. Org. Chem. 2000,
65, 6167-6172.
(8) Yasui, S.; Tsujimoto, M.; Shioji, K.; Ohno, A. Chem. Ber./Recueil
1997, 130, 1699-1707.
(9) Pandey, G.; Hajra, S.; Ghorai, M. K. Tetrahedron Lett. 1994, 35,
7837-7840.
(10) (a) Jeon, G. S.; Bentrude, W. G. Tetrahedron Lett. 1998, 39,
927-930. (b) Hager, D. C.; Sopchik, A. E.; Bentrude, W. G. J. Org.
Chem. 2000, 65, 2778-2785.
10.1021/jo049032l CCC: $30.25 © 2005 American Chemical Society
Published on Web 01/19/2005
1276
J. Org. Chem. 2005, 70, 1276-1280

Reaction of Triarylphosphine Radical Cations with O2
TABLE 1. Decay Rate Constant of 1•+ Determined by
Transient Absorption Measurementa
SCHEME 1
b
radical cation
atmosphere
kdecay (105 s-1
)
1a•+
air
Ar
O2
air
Ar
air
Ar
air
Ar
4.3
1.4
52
7.2
2.2
2.2
1.8
6.9
3.1
1c•+
1d•+
1e•+
the excellent ability of 1DCA* as an one-electron oxidant
(E1/2 ) 1.69 V vs Ag/Ag+).6,13 Recently, we examined the
photoreaction of triarylphosphines (1) with DCA and
found that the radical cation of 1 (1•+) exhibits a cat-
ionic character.6 Thus, when the reaction is carried out
in aqueous acetonitrile under an argon atmosphere,
the resulting 1•+ undergoes an ionic reaction with water
to give the corresponding phosphine oxides. In this
reaction, water plays an important role acting as a
nucleophile.
a Laser flash photolysis of 1 with a 4-ns and 355-nm laser under
aerobic conditions: [1] ) 1.00 × 10-2 M, [BP] ) 1.00 × 10-1 M,
[DCA] ) 5.00 × 10-5 M in MeCN. b Pseudo-first-order rate
constant.
Now, our interest is to determine what reaction takes
place if the photoreaction is carried out in “dry” aceto-
nitrile under aerobic conditions. This system does not
contain a nucleophile, whereas molecular oxygen O2 as
a radical is available. Thus, there is the possibility that
1•+ undergoes radical coupling with O2. We examined
the photoreaction of 1 in the presence of DCA and bi-
phenyl (BP) in acetonitrile (MeCN) under aerobic condi-
tions, and found the almost quantitative conversion of 1
into the corresponding phosphine oxides (2) (Scheme 1).
The results from the product analysis as well as in the
laser flash photolysis demonstrate the occurrence of the
reaction of 1•+ with O2 under aerobic conditions. To the
best of our knowledge, only a few reports have been pre-
sented by one group on the reactions of 1•+ with O2.14
FIGURE 1. Relationship between the pseudo-first-order rate
constant of the decay 1c•+ (kobs) and concentration of O2.
TABLE 2. Time-Course of the Photooxidation of 1aa
yieldb (%)
time (s)
1a
2a
material balancec (%)
Results
0
20
40
60
90
100.0
86.0
74.0
61.9
48.0
0.0
13.9
26.5
37.5
54.0
100.0
99.9
100.5
99.4
102.0
Laser Flash Photolysis. During the irradiation of a
solution of 1, BP, and DCA in MeCN with a 355-nm laser,
a transient absorption spectrum with absorption maxima
at 370 and 525 nm was observed and assigned to 1•+ by
comparison with our previous observation.6 In the ab-
sence of BP, the absorbance of 1•+ was too small to be
monitored. The absorbance at 525 nm decreased accord-
ing to first-order kinetics with the rate constant as
summarized in Table 1. The decay rate was accelerated
under aerobic conditions. The pseudo-first-order decay
rate of 1c•+ was found to be roughly proportional to the
concentration of O2 as shown in Figure 1. No appreciable
acceleration in the decay was observed with the addition
of methanol (up to 0.48 mol dm-3).
Product Analysis. A solution of 1, BP, and DCA in
MeCN was irradiated using a xenon lamp (>390 nm with
a glass filter) under aerobic conditions. At appropriate
intervals, the reaction mixture was analyzed by GC,
showing that 1 was gradually oxidized to the correspond-
ing 2. For each reaction, the sum of the amounts of 1
and 2 accounted for more than 80% of the material
a [1a] ) 1.00 × 10-2 M, [BP] ) 1.00 × 10-1 M, [DCA] ) 5.00 ×
10-5 M; under aerobic conditions in MeCN; irradiated with visible
light (>390 nm) from a Xe lamp. b Determined by GC using benzyl
ether as the external standard. c Sum of the yields of 1a and 2a.
balance. No other product was detected by the GC. Table
2 gives the time-course of the photooxidation of 1a as a
typical example.
The amount of 1 consumed (in moles when normalized
to 1dm3 solution) was plotted versus the irradiation time
in Figure 2, which gave a linear line up to 30-90%
consumption of 1 for each reaction. Table 3 summarizes
the slopes of the linear lines obtained from the reactions
under the various conditions. Our previous work has
shown the intensity of the incident light from our xenon
lamp to be 1.76 × 10-6 einstein s-1cm-2 at 450 nm.8 The
lamp intensity at this wavelength was almost the same
as that at 420 nm where DCA was excited. Using this
intensity together with the values of the slopes in Figure
2, the apparent quantum yields for the conversion of 1
to 2 (Φcons) were calculated as shown in Table 3 (see the
Experimental Section).
(13) Gould, I. R.; Ege, D.; Moser, J. E.; Farid, S. J. Am. Chem. Soc.
1990, 112, 4290-4301.
(14) (a) Alfassi, Z. B.; Neta, P.; Beaver, B. J. Phys. Chem. A 1997,
101, 2153-2158. (b) Beaver, B.; Rawlings, D.; Neta, P. Alfassi, Z. B.;
Das, T. N. Heteroatom Chem. 1998, 9, 133-138.
J. Org. Chem, Vol. 70, No. 4, 2005 1277

Yasui et al.
SCHEME 2
Discussion
Photocatalytic Cycle. The photoreaction is certainly
initiated by the excitation of DCA to 1DCA*. The electron
transfer from BP to 1DCA* takes place to give the radical
cation of BP (BP•+) and the radical anion of DCA (DCA•-).
This process is followed by an electron transfer (ET) from
1 to BP•+ (i.e., hole transfer from BP•+ to 1), giving 1•+
and BP. Only a catalytic amount of DCA is necessary for
the conversion of 1 to 2 to be completed. That is, DCA
acts as a catalyst in the photooxidation reaction. Most
likely, O2 regenerates DCA by oxidizing DCA•- to drive
the catalytic cycle as shown in Scheme 2. Comparing the
potentials of the pertinent compounds, each process is
found to be feasible.15
FIGURE 2. Time-course of the amount of 1 consumed in the
DCA-sensitized photoreaction of 1 under the conditions given
in footnote a of Table 2, except for the concentration of 1f; [1f]
) 3.88 × 10-3 M. For simplicity, only selected plots are
indicated in this figure: 1a, O; 1b, b; 1c, 0; 1d, 9; 1e, 4; 1f,
2. The solid lines show the linear relationships between the
amount of 1 consumed and irradiation time.
TABLE 3. Photoreaction of 1 in the Presence of DCA
and BP in MeCNa
Reaction of 1•+ with O2. Taking into account the fact
that the MeCN solvent used in this study cannot be
absolutely “dry”,17 1•+ could react with water. The ionic
reaction of 1•+ with water is a major process for the decay
of 1•+ in aqueous MeCN under an argon atmosphere.6 The
reaction would initially afford the phosphoranyl radical
Ar3P•-OH which would easily react with O2 to give a
peroxy radical Ar3P(OH)-O-O• under the aerobic condi-
tions.18 This type of peroxy radical, namely, a radical
bearing a hydroxyl group on the phosphorus atom,
undergoes a chain reaction, where the hydroxyl radical
is the chain carrier.7 This is not the case in the present
photoreaction because no reaction occurred when the
reaction mixture was kept in the dark after irradiation.
As seen in Table 1, results of the laser flash photolysis
experiments show that the decay rate of 1•+ was faster
in air than in argon. Figure 1 indeed shows a linear
relationship of the decay rate constant of 1c•+ with the
concentration of O2, although only three data points are
available. This observation suggests that 1•+ bimolecu-
larly reacts with O2. This mechanism predicts a pseudo-
first-order kinetics for the decay of 1•+ under our condi-
tions, and this is indeed the case. Alternatively, 1•+ could
undergo radical coupling with the superoxide radical
anion O2•- generated by the electron transfer from DCA•-
to O2 (Scheme 2).19,20 However, if 1•+ decays according to
this mechanism, the decay would follow the second-order
slopeb
quantum
phosphine
[1]
(105 mol
yield
c
run
1
(102 M) atmosphere in 1 L s-1
)
(Φcons)
1
2
1a
2.00
1.00
0.341
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
0.388
air
air
air
air
air
air
O2
air
air
Ar
air
air
air
air
air
5.96
5.78
5.45
4.90
5.92
7.35
16.2
0.466
0.034
0.033
0.031
0.028
0.034
0.042
0.092
1a
3
1a
4
1ad
1ae
1af
1a
5
6
7
8
1ag
1ah
1a
0.0026
9
no reaction
no reaction
9.02
10
11
12
13
14
15
1b
1c
0.051
0.092
0.068
0.019
0.0095
16.2
1d
1e
12.0
3.26
1f i
1.67
a Carried out under the conditions given in footnote a of Table
2, unless otherwise indicated. b Slope of the plot of the amount of
1 consumed versus time; see the text. c Estimated by assuming
that the intensity of the incident light is 1.76 × 10-6 einstein s-1
cm-2 at 420 nm. d [DCA] ) 2.50 × 10-5 M. e [DCA] ) 1.00 × 10-4
M. f [BP] ) 5.10 × 10-2 M. g In the absence of BP. h In the absence
of DCA. i Poor solubility of 1f in MeCN did not allow the reaction
under standard conditions.
The Φcons value was nearly constant irrespective of
the initial concentration of 1 (runs 1-3). Differences in
the concentrations of BP and DCA exerted only a small
effect on Φcons (runs 4-6). No reaction took place in
the absence of DCA (run 9). Meanwhile, the conversion
of 1 to 2 was observed even in the absence of BP,
although the reaction was much slower than the reaction
in the presence of BP (run 8). The reaction under O2
atmosphere was faster than the reaction in air (run 7).
Practically, no reaction took place under the argon
atmosphere (run 10). With either an electron-withdraw-
ing substituent (p-Cl) or electron-releasing substituents
(p-Me and p-MeO) on the phenyl rings of 1, Φcons in-
creased (runs 11-13). Meanwhile, in the reactions of
tris(o-tolyl)phosphine (1e) and trimesitylphosphine (1f),
the Φcons values were smaller than those predicted from
the electronic effect of the methyl substituents (runs 14
and 15).
(15) The oxidation or reduction potentials of DCA, 1DCA*, and BP
were taken from ref 6. The reduction potential of O2 was reported in
ref 16. All values were adjusted to half-wave potentials vs Ag/Ag+ in
V.
(16) Fukuzumi, S.; Patz, M.; Suenobu, T.; Kuwahara, Y.; Itoh, S. J.
Am. Chem. Soc. 1999, 121, 1605-1606.
(17) Acetonitrile was refluxed over calcium hydride for at least 10
h and distilled. The solvent purified in this way still contains about
0.1% water. See, Yasui, S.; Shioji, K.; Ohno, A.; Yoshihara, M. Chem.
Lett. 1993, 1393-1396.
(18) Watts, G. B.; Ingold, K. U. J. Am. Chem. Soc. 1972, 94, 2528-
2529.
(19) (a) Steichen, D. S.; Foote, C. S. J. Am. Chem. Soc. 1981, 103,
1855-1857. (b) Mizuno, K.; Tamai, T.; Hashida, I.; Otsuji, Y. J. Org.
Chem. 1995, 60, 2935-2937. (c) Baciocchi, E.; Giacco, T. D.; Elisei, F.;
Gerini, M. F.; Guerra, M.; Lapi, A.; Liberali, P. J. Am. Chem. Soc. 2003,
125, 16444-16454.
(20) Somasundaram, N.; Srinvasan, C. J. Org. Chem. 1996, 61,
2895-2896.
1278 J. Org. Chem., Vol. 70, No. 4, 2005

Reaction of Triarylphosphine Radical Cations with O2
kinetics,19c which is opposite the observation. The mech-
anism is incompatible with the first-order kinetics even
SCHEME 3
SCHEME 4
•-
by assuming that a much higher concentration of O2
than that of 1•+ provides pseudo-first-order conditions.
Thus, taking into account the fact that the concentration
•-
of O2 is lower than the initial concentration of DCA (5
× 10-5 M), the pseudo-first-order rate constant is pre-
dicted to be <105 s-1 when the second-order rate constant
is diffusion-limited. This value is not consistent with the
observed pseudo-first-order rate constants, 105-106 s-1
(Table 1). The discussion here eliminates the possibility
•-
of radical coupling between 1•+ and O2
.
There is a possibility that the reaction of 1 with singlet
oxygen 1O2 affords 2, since it has been reported that 1O2
is generated in the DCA-sensitized photoreaction.21 Phos-
1
from 1e was slightly smaller than those from 1a-d
(Table 3). On the other hand, the decay rate of 1e•+
determined from the laser flash photolysis is comparable
to those of the other 1•+ (Table 1). The steric effect by
o-methyl groups in 1e may be operative in a step other
than the reaction of 1e•+ with O2 to lower the yield
of 2e.27
phine reacts with O2, when it is available, to give the
corresponding phosphine oxide.22-24 The pseudo-first-
1
order rate constant of the formation of O2 in oxygen-
saturated MeCN in the presence of DCA has been
reported to be 5.45 × 107 s-1 21b
.
Apparently, the rate
constant for the reaction in air is about one-fifth, i.e.,
about 1 × 107 s-1. This value is significantly smaller than
the pseudo-first-order rate constant for the electron-
Most likely, the reaction of 1•+ with O2 affords 3, which
well explains the formation of 2, although no specific
absorption was observed for the intermediacy of 3.
Another molecule of 1 may attack 3 to give 2 and 2•+.
The radical cation 2•+ eventually reacts with DCA•- and/
1
transfer quenching of DCA* by BP; this value is esti-
mated to be 6.1 × 107 s-1 under the aerobic conditions
where [O2] ) 8.1 × 10-3 M.25 This means that the DCA-
mediated formation of 1O2 cannot compete with the
electron-transfer quenching of 1DCA* by BP. In fact, we
have observed DCA•- during the laser flash photolysis.
In addition, crowded triarylphosphine, such as tris(o-
methoxy)phosphine, gives a considerable amount of aryl
triarylphosphinate (ArO-P(dO)Ar2) upon the reaction
with 1O2 in an aprotic solvent,24 whereas in the reactions
of the crowded arylphosphines, 1e and 1f, phosphinate
was not detected in the present study. Therefore, the
participation of 1O2 into the present reactions is very
unlikely.
•-
or O2 to give 2 (Scheme 3). A similar mechanism has
been presented for the γ-radiolysis of 1 in air-saturated
CH2Cl2, where the initially generated 1•+ eventually gives
2 via an intermediate 3.14 In that study, the sequential
hole transfer from 2•+ to 1 that produces 1•+, which
results in a chain mechanism, has been proposed to
explain the efficient formation of 2. On the other hand,
under the present experimental condition, the concentra-
tion of BP is much higher than 1, and therefore 2•+ is
quenched by BP and the positive charge developed on
BP disappears through a charge recombination with
DCA•- and/or O2•-. The reaction mechanism is sum-
marized as in Schemes 3 and 4.
In conclusion, 1•+ undergoes a radical coupling with
O2 under aerobic conditions, showing the reactivity as a
radical. In support of this mechanism, the yield of 2f from
1f was much smaller compared with the other cases
(Table 3). That is, the reactivity of 1f•+ with two o-methyl
groups at each aryl group is much lower than those of
the other 1•+. It has been shown that aryl groups with
bulky ortho substituents flatten the tetrahedral geometry
to a pyramidal geometry of 1•+, making the spin on the
Interestingly, Φcons was not dependent on the concen-
trations of DCA and BP in the range of the present
experimental conditions, indicating the rapid operation
1
of the catalytic cycle shown in Scheme 2. Thus, DCA*
and BP•+ are at the photostationary concentrations and
determined by the number of photons in the incident
light, at least during the early stage of the reaction.
central phosphorus atom more delocalized into the aryl
•+
groups.26 Such flattening may occur in 1f
and make
Comparison of the Reactivity of 1•+ with Those
of Other Radical Cations. We have shown that the cis-
stilbene radical cation generated during pulse radiolysis
in 1,2-dichloroethane undergoes either isomerization to
the trans isomer or radical coupling with O2.28 Dichotomy
in the reactivity of this radical cation has been inter-
preted in the term of its distonic structure; thus, the
charge-spin separation is responsible for the radical
character of the radical cation. It is usually accepted that
trivalent phosphorus radical cations are not distonic but
•+
1f
less reactive toward O2 than the other 1•+. Inter-
estingly, such a steric effect is seen even with only one
o-methyl group at each aryl group, since the yield of 2e
(21) (a) Dobrowolski, D. C.; Ogilby, P. R.; Foote, C. S. J. Phys. Chem.
1983, 87, 2261-2263. (b) Araki, Y.; Dobrowolski, D. C.; Goyne, T. E.;
Hanson, D. C.; Jiang, Z. Q.; Lee, K. J.; Foote, C. S. J. Am. Chem. Soc.
1984, 106, 4570-4575.
(22) (a) Tsuji, S.; Kondo, M.; Ishiguro, K.; Sawaki, Y. J. Org. Chem.
1993, 58, 5055-5059. (b) Akasaka, T.; Kita, I.; Haranaka, M.; Ando,
W. Quim. Nova 1993, 16, 325-327.
(23) Nahm, K.; Li, Y.; Evanseck, J. D.; Houk, K. N.; Foote, C. S. J.
Am. Chem. Soc. 1993, 115, 4879-4884.
(24) Gao, R.; Ho, D. G.; Dong, T.; Khuu, D.; Franco, N.; Sezer, O.;
Selke, M. Org. Lett. 2001, 3, 3719-3722.
(27) Although no experimental data are available, we are tempted
to speculate that the peroxy radical 3 attacks the phosphine 1 to give
the phosphoranyl radical with TBP geometry. This step would be
subjected to a steric effect by the ortho substituent since the C-P-C
angle becomes smaller.
(28) Tojo, S.; Morishima, K.; Ishida, A.; Majima, T.; Takamuku, S.
J. Org. Chem. 1995, 60, 4684-4685.
(25) In the absence of BP, 1DCA* is quenched by 1 with the first-
order rate constant of (1.2-1.7) × 108 s-1, which was independently
determined by the Stern-Volmer analysis.
(26) Culcasi, M.; Berchadsky, Y.; Gronchi, G.; Tordo, P. J. Org.
Chem. 1991, 56, 3537-3542.
J. Org. Chem, Vol. 70, No. 4, 2005 1279

Yasui et al.
conventional.29 It is suggested that the unpaired spin is
highly localized on the central phosphorus atom.
Experimental Section
Instruments. Laser flash photolysis measurements were
performed using the third harmonic generation (355 nm, 5 ns
full width at half-maximum, 6 mJ/pulse) from a Q-switched
Nd3+:YAG laser for the excitation operated with temporal
control by a delay generator. The reflected analyzing light from
a pulsed 450-W Xe-arc lamp was collected by a focusing lens
and directed through a grating monochromator to a photo-
multiplier tube. The transient signals were observed upon the
pulse excitation of DCA (ca. 5 × 10-5 M) in the presence of 0.1
M biphenyl in MeCN and recorded with a digitizer.
Materials. Phosphines 1a-e and biphenyl (BP) were com-
mercially available and purified by recrystallization from
ethanol. Trimesitylphosphine (1f) was prepared by a conven-
tional Grignard method based on a literature procedure.26
9,10-Dicyanoanthracene (DCA) was purchased and purified by
recrystallization from benzene.
The reactivity of 1•+ is compared with that of the amine
counterpart. Although many reports have been presented
on aliphatic and aromatic amine radical cations, there
is no report that amine radical cations undergo radical
coupling.30-32 One reason is that the amine radical
cations have a structure much more flattened than the
phosphine radical cations.30 That is, the spin in the amine
radical cations is more delocalized into the aryl rings.
An ab initio calculation for the triphenylamine radical
cation predicts that the unpaired electron is 59% localized
on the amine nitrogen atom,31 which can be compared
with the observation that about 85% of the unpaired
electron is localized on the phosphorus atom in 1•+ of the
single crystals.33
General Procedure. The solution of 1 (1.00 × 10-2 M), BP
(1.00 × 10-1 M), and DCA (5.00 × 10-5 M) in 2 mL of MeCN
was prepared in a square quartz cell (1 cm × 1 cm) in air and
irradiated with light from a xenon arc short lamp through a
sharp-cut filter (irradiation at λ > 390 nm). At intervals, a
50-µL aliquot was taken and diluted with 25 µL of MeCN
involving benzyl ether (for 1a,d,e) or hydroquinone dibenzyl
ether (for 1b,c,f) as an external standard, and the resulting
mixture was analyzed by GC.
Conclusions
The photooxidation of 1 was studied in the presence
of DCA as a sensitizer and BP as a co-sensitizer in “dry”
acetonitrile under aerobic conditions. Phosphine 1 con-
verted to the corresponding 2 without any appreciable
side reactions. Examining the substituent effects on the
total efficiency of the reaction and on the decay rate of
1•+, based on the product analysis and laser flash
photolysis, respectively, we showed that 1•+ undergoes
radical coupling with O2.
Laser Flash Photolysis. The solution of 1 (1.00 × 10-2
M), BP (1.00 × 10-1 M), and DCA (5.00 × 10-5 M) in MeCN
was photolyzed in air. The decay of the resulting absorption
was monitored at 525 nm.
Quantum yield Φcons. The intensity of the incident light
at 420 nm was assumed to be 1.76 × 10-6 einstein s-1 cm-2 on
the basis of the measurement at 450 nm in a previous study.8
Meanwhile, the area where the light was applied to the
reaction vessel (a square cell) was 2 cm2. The number of
photons being received by the reaction mixture was then 3.52
× 10-6 einstein s-1. Comparing this number with the amount
of 1 consumed per second (mol s-1) upon irradiation, Φcons was
then calculated.
(29) Smith, R. L.; Schweighofer, A.; Keck, H.; Kuchen, W.; Kent-
ta¨maa, H. I. J. Am. Chem. Soc. 1996, 118, 1408-1412.
(30) de Meijere, A.; Chaplinski, V.; Gerson, F.; Merstetter, P.;
Haselbach, E. J. Org. Chem. 1999, 64, 6951-6959.
(31) Pacansky, J.; Waltman, R. J.; Seki, H. Bull. Chem. Soc. Jpn.
1997, 70, 55-59.
(32) Zheng, Z.-R.; Evans, D. H.; Nelsen, S. F. J. Org. Chem. 2000,
65, 1793-1798.
(33) Berclaz, T.; Geoffroy, M. Mol. Phys. 1975, 30, 549-555.
JO049032L
1280 J. Org. Chem., Vol. 70, No. 4, 2005