Reaction of arylphosphines with singlet oxygen: intra- vs intermolecular oxidation.

Article abstract of DOI:10.1021/ol010195v

[reaction--see text] The chemistry of singlet oxygen with all three isomers of tris(methoxyphenyl)phosphine has been studied. For the severely hindered ortho isomer, intramolecular rearrangement to form phenyl diphenyl phosphinate is preferred to formatio

Full text of DOI:10.1021/ol010195v

ORGANIC
LETTERS
2
001
Vol. 3, No. 23
719-3722
Reaction of Arylphosphines with Singlet
Oxygen: Intra- vs Intermolecular
Oxidation
3
†
Ruomei Gao, David G. Ho, Timothy Dong, Daisy Khuu, Nestor Franco,
Oemer Sezer, and Matthias Selke*
Department of Chemistry and Biochemistry, California State UniVersity,
Los Angeles, Los Angeles, California 90032
mselke@calstatela.edu
Received August 29, 2001
ABSTRACT
The chemistry of singlet oxygen with all three isomers of tris(methoxyphenyl)phosphine has been studied. For the severely hindered ortho
isomer, intramolecular rearrangement to form phenyl diphenyl phosphinate is preferred to formation of phosphine oxide at low concentration
in aprotic solvents. In protic solvents, no intramolecular reactivity is observed. A detailed kinetic analyses has been undertaken. There is no
physical quenching, regardless of solvent. Mechanistic implications of these findings are discussed.
The affinity of trivalent phosphorus for oxygen is well-
known, and the oxidation of phosphines is one of the best
known reactions in all of chemistry. The mechanism of
autoxidation of phosphines with triplet dioxygen was studied
studied to date, even though such phosphines are widely used
as ligands in organometallic chemistry. Chemical properties
of arylphosphines are influenced by the steric bulk of the
aryl groups, and the cone angle of a phosphine ligand has
1
4
a long time ago. However, there have been remarkably few
long been used as a measure of its bulk. This cone angle is
investigations concerning the reactivity of phosphines with
singlet oxygen, which is in stark contrast with the reaction
of singlet oxygen with organic sulfides. Trapping studies by
obtained by taking a space-filling model of the MP(R
3
)
group. The metal is the apex of the cone, and the entire ligand
4
,5
forms the actual cone. We reasoned that it would be
particularly interesting to study the photooxidation of
arylphosphines whose metal complexes have very large cone
angles (>180°). For such phosphines, the primary intermedi-
ate in the reaction with singlet oxygenspresumably the
phosphadioxiraneswould be expected to have a much longer
lifetime than that of phosphines with small cone angles, as
the approach of a second phosphine molecule to react with
the intermediate would be severely hindered. This might lead
2
Sawaki et al. indicate that the primary intermediate in the
photooxidation of triphenylphosphines and tributyl phosphite
is electrophilic, in contrast with the nucleophilic behavior
of the persulfoxide intermediate in the photooxidation of
organic sulfides. Ab initio calculations by Foote, Houk, and
3
co-workers suggest a cyclic phosphadioxirane intermediate
which would indeed be expected to behave as an electrophile.
No reactions of arylphosphines and singlet oxygen have been
2
to intramolecular reactivity of the intermediate or even loss
†
Research Intern, Biomedical Research Honors Class, Alhambra High
of the dioxygen molecule and re-formation of the starting
phosphine. The latter pathway would amount to indirect
School, Alhambra, CA 91801.
(
1) Buckler, S. A. J. Am. Chem. Soc. 1962, 84, 3093-3097.
(2) Tsuji, S.; Kondo, M.; Ishiguro, K.; Sawaki, Y. J. Org. Chem. 1993,
5
8, 5055-5058.
3) Nahm, K.; Li, Y.; Evanseck, J. D.; Houk, K.; Foote, C. S. J. Am.
Chem. Soc. 1993, 115, 4879-4884.
(4) Tolman, C. A. Chem. ReV. 1977, 77, 313.
(5) Crabtree, R. H. The Organometallic Chemistry of the Transition
Metals; John Wiley and Sons: New York, 1988; pp 72-74.
(
1
0.1021/ol010195v CCC: $20.00 © 2001 American Chemical Society
Published on Web 10/20/2001

6
physical quenching of singlet oxygen. Except for tri-
7
phenylphosphine and several trialkyl phosphites, no kinetic
Scheme 1. Photooxidation of Tris(o-methoxyphenyl)phosphine
(1) and Different Reactivity Pathways for the Peroxidic
Intermediate in This Reaction
data are known for reactions of phosphines with singlet
oxygen, and the possibility of physical quenching has never
been examined. We have therefore conducted detailed
product and kinetic studies of the reaction of the ortho, meta,
and para isomers of tris(orthomethoxyphenyl)phosphine with
singlet oxygen. The ortho-isomer has a particularly large cone
8
angle of 205°. All of these arylphosphines are unreactive
with triplet oxygen, thus allowing kinetic investigations of
their reactions with singlet oxygen.
Photooxidation of tris(orthomethoxyphenyl)phosphine (1)
in aprotic solvents leads to a mixture of the corresponding
phosphine oxide (2) and the orthomethoxyphenyl di(ortho-
9
methoxyphenyl) phosphinate (3). The product distribution
depends on the concentration of the starting phosphine 1:
The higher the starting concentration, the larger the ratio of
compound 2/3. When [1] > 0.5 M, more than 90% of the
total product is the phosphine oxide 2; conversely, when [1]
<
0.003 M, phosphinate 3 is more than 90% of the total
10
product formed. In contrast, in protic solvents or solvent
mixtures, only phosphine oxide 2 is obtained, regardless of
the concentration of 1. A large excess of protons relative to
the starting compound 1 is required to completely suppress
formation of the phosphinate 2. 2-Propanol also suppresses
formation of the insertion product but is less efficient than
MeOH. For example, in the absence of any protic solvent,
aprotic solvents, this intermediate may undergo three dif-
ferent reactions: intermolecular oxygen atom transfer to
another phosphine molecule; intramolecular rearrangement,
leading to the phosphinate observed for the photooxidation
of 1, and loss of dioxygen and regeneration of the starting
phosphine (indirect physical quenching). If Scheme 1 holds,
the ratio of phosphine oxide 2 vs insertion product 3 should
be proportional to the starting phosphine concentration (at
low conversion), regardless of whether physical quenching
occurs. Using the steady-state approach leads to eq 1:
photooxidation of a 0.013 M solution of 1 in CHCl
to equal amounts of products 2 and 3 (at 40% conversion of
). In contrast, treatment of the same solution of 1 with 2 M
MeOH results only in formation of phosphine oxide 2 (again,
at 40% conversion of 1). Conversely, addition of 2 M
3
leads
1
2-propanol instead of MeOH results in about 90% formation
of 2 and ca. 10% of 3 under otherwise identical conditions.
The other two isomers, tris(metamethoxyphenyl)phosphine
(4) and para(methoxyphenylphosphine) (5), show very dif-
ferent behavior: only the corresponding phosphine oxides
are obtained upon reaction with singlet dioxygen, regardless
of their concentrations or the nature of the solvent.
[
2]/[3] ) (2k [1])/ki
(1)
o
Kinetic Studies. Scheme 1 shows the different reaction
pathways of the possible peroxidic intermediate. The per-
oxidic intermediate must be formed upon reaction of a
phosphine with singlet dioxygen, as a primary product. In
Plots of [2]/[3] are indeed linear (Figure 1), with a slope
of 50 ( 6 in CHCl . Thus the rate ratio of intermolecular vs
intramolecular oxidation kox/k is 25 ( 3.
The total rate constant of singlet oxygen removal by 1
) has been measured by luminescence quenching experi-
ments in a variety of solvents (Figure 2). The values are
considerably smaller than those of arylphosphines without
the ortho substituents (see Table 1).
3
i
(k
T
(6) We use the term indirect physical quenching to describe deactivation
11
of singlet oxygen via formation of an unstable intermediate which then
loses a dioxygen molecule. Direct physical quenching refers to physical
deactivation without formation of any unstable adduct. This terminology
has been used for the reaction of singlet oxygen with organic sulfides. Foote,
rd
In the absence of direct physical quenching, the value of
k is the rate of formation of the peroxidic intermediate. To
T
C. S.; Peters, J. W. IUPAC Congr., 23 , Spec. Lec. 1971, 4, 129.
(
7) Nahm, K.; Foote, C. S. J. Am. Chem. Soc. 1989, 111, 909.
(8) Hirsivaara, L.; Guerricabeitia, L.; Haukka, M.; Suomalainen, P.;
determine whether all of the intermediate leads to formation
of either product 2 or 3 or whether loss of dioxygen from
the intermediate (indirect physical quenching) is a major
process, competition experiments have been carried out. 9,10-
Dimethylanthracene (DMA) was used as a singlet oxygen
Laitinen, R. H.; Pakkanen, T. A.; Pursiainen, J. Inorg. Chim. Acta 2000,
07, 47. Ab initio calculations by these authors indicate that smaller cone
3
angles which would make the phosphorous atom more exposed to
intermolecular attack correspond to higher energy conformers. Several X-ray
strucures described in this paper of Cr and W complexes containing the
ortho isomer do not indicate any unusual flattening of the phosphine ligand.
(
9) Sawaki et al. observed trace amounts of a similar insertion product2,
phenyl diphenyl phosinate, during the photooxidation of triphenylphosphine.
10) Sawaki et al. also noted that for triphenylphosphine a decrease in
(
(11) Time-resolved singlet oxygen luminescence quenching experiments
are conducted by exciting a solution containing the sensitizer and varying
amounts of substrate (quencher) with a short (a few ns) laser pulse and
monitoring the singlet oxygen decay at right angle.
the starting concentration leads to in increase in the amount of insertion
product. However, the maximum yield of insertion product that was obtained
from triphenylphosphine was only 1.2%. See ref 2.
3720
Org. Lett., Vol. 3, No. 23, 2001

Table 1. Kinetic Parameters for the Reaction of
Arylphosphines with Singlet Oxygen
k
T
(M- s-1 × 10-6)a
1
k
o
/k
i
in
CHCl
in
6
in CH
2 2
Cl /
in
CHCl
3
compound
3
C
H
6
MeOH
tris(o-methoxyphenyl)-
phosphine (1)
tris(m-methoxyphenyl)- 5.3 ( 0.5
phosphine (4)
2.8 ( 0.4
5.0 ( 0.2
9.2 ( 0.3
3.1 ( 0.2 25 ( 3
6.4 ( 0.2
tris(p-methoxyphenyl)- 14.9 ( 1.6 33.1 ( 4.3 27.4 ( 3.9
phosphine (5)
a
Average of four experiments; error is 1 standard deviation.
o q
formation (k ), and indirect physical quenching (k ) as follows
(at low conversion of 1):
Figure 1. Plot of product distribution for the photooxidation of 1
vs starting concentration of 1.
k (2k [1] + k )
T
o
i
kr )
(3)
k + k + k [1]
q
i
o
acceptor in these experiments. DMA is known to interact
with singlet oxygen by chemical reaction only, and we have
remeasured its reaction rates with singlet oxygen to be 2.5
Equation 3 can be used to determine whether there is
significant physical quenching at various concentrations of
7
-1 -1
7
-1 -1
1: if kq ) 0, then the ratio k /k ) 1 at very low
(
0.1 × 10 M
s
in CHCl
3
, 2.9 ( 0.2 × 10 M
s
in
r
T
7
-1 -1
concentrations of 1 where formation of phospine oxide 2 is
benzene, and 4.7 ( 0.6 × 10 M
s
2 2
in 75% CH Cl
/25%
12
insignificant. The ratio k /k will rise to a value of two at
MeOH, in excellent agreement with the literature. Loss of
DMA and compound 1 was monitored by NMR, and
conversion of 1 was kept at 20% or less. Relative reaction
r
T
high concentration of 1 where formation of the insertion
product 3 is insignificant. Competition experiments were
carried out at various starting concentrations of [1] in CDCl3,
1
3,14
rates were obtained from the equation of Higgins et al.
and a plot of the ratios k
concentrations of 1 is shown in Figure 3. Comparison with
the curve that is predicted by eq 3 if k ) 0 (based on a k /k
r T
/k against the corresponding starting
f
0
log{[1] /[1] ]
k (1)
r
)
0
(2)
q
o
i
f
kr(DMA)
log{[DMA] /[DMA] }
ratio of 25) shows that within limits of error there is no
significant physical quenching regardless of the phosphine
1
5
The relative reaction rate k
rates of insertion product formation (k
r
thus obtained is related to the
), phosphine oxide
concentration employed. Thus, loss of dioxygen from the
peroxidic intermediate is not a significant process. Competi-
tion experiments carried out under protic conditions (75%
i
CD
of k
2
Cl
2 3 r
/25% CD OD) resulted in values of k twice the rate
T
(data not shown), demonstrating that there is no physical
quenching under protic conditions. Kinetic parameters for
the reaction of compound 1 with singlet oxygen are sum-
marized in Table 1.
Kinetic parameters for meta and para isomers 4 and 5 have
also been obtained. These values are summarized in Table
1
T
. The values for singlet oxygen removal (k ) are consider-
ably larger than those for 1, undoubtedly because of the
smaller steric bulk of these molecules. In all cases the relative
(
12) Wilkinson, F.; Helman, W. P.; Ross, A. B. J. Phys. Chem. Ref.
Data 1995, 24, 663, and references therein.
13) Control experiments demonstrated that within 60 min no oxidation
within 1%) of the phosphines by the 9,10-dimethylanthracene endoperoxide
(
(
product occurs. Since the irradiation times during the competition experi-
ments were only 5-15 min, oxidation of the phosphine by DMA
endoperoxide did not interfere with the kinetic measurements. Compounds
2
and 3 are oxidized substantially over longer time periods by DMA
endoperoxide (20% during 36 h) while compound 1 is unreactive with DMA
endoperixe even over a 36 h time period, probably because of its steric
bulk.
(
14) Higgins, R.; Foote, C. S.; Cheng, H. AdV. Chem. Ser. 1968, 77,
02.
(15) This also rules out direct physical quenching prior to formation of
the peroxidic intermediate.
Figure 2. Singlet oxygen luminescence quenching by phosphines
, 4, and 5.
1
1
Org. Lett., Vol. 3, No. 23, 2001
3721

phosphadioxirane with water does indeed have an energy
minimum that is considerably lower (by 29.6 kcal/mol at
the RHF STO-G(*) level and by 19.9 kcal/mol at the RHF
3
6
-31G* level) than that calculated for the corresponding
phosphadioxirane. However, thus far, there has been no
experimental evidence for this reaction pathway. The inhibi-
tion of intramolecular oxidation observed during the pho-
tooxidation of 1 under protic conditions demonstrates that
addition of protons and an alkoxide anion does indeed lead
to a change in mechanism during the photooxidation of
phosphines and that a hydroperoxy phosphorane may be the
second intermediate under such conditions. Formation of a
pentacoordinate species is supported by the observation that
the bulkier 2-propanol is much less efficient in suppressing
formation of product 3 than MeOH. Protonation of the
phosphadioxirane may be reversible, as a large excess of
protons is required to completely suppress formation of the
insertion product. This result is in contrast with the observa-
tion by Sawaki et al. that addition of 10% MeOH does not
affect trapping of the intermediate in the photooxidation of
Figure 3. Plot of the rate ratio k
1
r
/k
. The solid line represents the curve predicted from eq 3 if k
and k /k values obtained
T
vs concentration of phosphine
q
)
0
o
i
) 25. The data points are based on the k
r
from the competition experiments at different starting concentrations
of 1.
2
tri-n-butyl phosphite. The intermediate in the photooxidation
of the phosphite may be much less susceptible to protonation
due to an electronic effect, i.e., because of the more electron-
poor character of the phosphite.
While an extensive amount of kinetic data has been
gathered for the reaction of singlet oxygen with organic
reaction rates obtained from competition experiments with
DMA are twice the values of k
T
, indicating that there is no
physical quenching. In dramatic contrast with the reaction
16
of organic sulfides with singlet oxygen, there is very little
variation of these numbers with solvent.
sulfides, this paper represents the first detailed kinetic study
The reactivity patterns of compound 1 in protic and aprotic
solvents and the kinetic results are consistent with the
mechanism outlined in Scheme 2. The observation that protic
1
of the reaction of O
2
with phosphines. In contrast to organic
sulfides, protic solvents only slightly accelerate the reaction
rate of phosphines with singlet oxygen, but they do lead to
a different product distribution (see above). Furthermore,
there is no significant amount of physical quenching of
singlet oxygen; instead of losing dioxygen, the peroxidic
intermediate undergoes intramolecular rearrangement when
intermolecular reactions are severely hindered. These results,
Scheme 2. Reactivity of Compound 1 with Singlet Oxygen
under Protic and Aprotic Conditions
2
together with the trapping experiments by Sawaki et al. that
demonstrated the electrophilic nature of the peroxidic
intermediate in the photooxidation of phosphines, lend strong
support to the formulation of this intermediate as a cyclic
phosphadioxirane. The absence of physical quenching and
the intramolecular rearrangement of the phosphadioxirane
derived from the ortho-substituted arylphosphine 1 suggests
that such phosphadioxiranes may have a lifetime sufficiently
long enough to be detected at very low temperatures or by
transient absorption spectroscopy. Experiments in this direc-
tion are in progress.
Acknowledgment is made to the donors of the Petroleum
Research Fund, administered by the American Chemical
Society, for partial support of this research. Support by the
Camille and Henry Dreyfus Foundation (Award number SU-
98-021) and the NIH-NIGMS MBRS program (Award
number GM 08101) is also gratefully acknowledged.
solvents inhibit formation of the insertion product 3 indicates
that under protic conditions protonation of the phospha-
dioxirane to a hydroperoxy species is taking place.
3
Foote, Houk, and co-workers have performed ab initio
calculations for the protonation of the phosphdioxirane
1
formed by reaction of PH
3
with O
2
. They have found that
Supporting Information Available: Detailed kinetic
analyses of the reaction of compound 1 with singlet oxygen,
including derivation of eqs 1 and 3. This material is available
free of charge via the Internet at http://pubs.acs.org.
the pentacoordinate species formed by reaction of the
(
16) (a) Jensen, F.; Greer, A.; Clennan, E. L. J. Am. Chem. Soc. 1998,
1
20, 4439. (b) Liang, J.-J.; Gu, C.-L.; Kacher, M. L.; Foote, C. S J. Am.
Chem. Soc. 1983, 105, 4717. (c) Gu, C.-L.; Foote, C. S.; Kacher, M. L. J.
Am. Chem. Soc. 1981, 103, 5949.
OL010195V
3722
Org. Lett., Vol. 3, No. 23, 2001