A Reexamination of the Ozone-Triphenyl Phosphite System. The Origin of Triphenyl Phosphate at Low Temperatures

Article abstract of DOI:10.1021/jo982339y

The reaction of ozone with triphenyl phosphite (P) at -78°C affords a labile 1:1 complex (PO3) together with small amounts of triphenyl phosphate (PO) (Q. E. Thompson, J. Am. Chem. Soc. 1961, 83, 845). In this work we found that the amount of PO present initially after complete ozonation of P in toluene was 12 ± 2% at -78°C and 11 ± 2% at -95°C. Partial ozonation of solutions of P in toluene at -78°C gave mixtures of P, PO, and PO3 whose composition changed with time as a result of the reaction of P with PO3 to give additional PO. Between -25 and -60 °C, the rate constant of the latter reaction is given by the expression log k (M^-1 s^-1) = (8.64 ± 0.04) - (11.44 ± 0.74) kcal/AT. This reaction at -78°C is too slow to account for the PO formed during the ozonation, which is proposed to arise instead by competitive reactions of an intermediate. The solubility of PO in toluene at -78°C was measured as 0.06 M, and that of PO3 about 6 times greater.

Full text of DOI:10.1021/jo982339y

J . Org. Chem. 1999, 64, 5783-5786
5783
A Reexa m in a tion of th e Ozon e-Tr ip h en yl P h osp h ite System . Th e
Or igin of Tr ip h en yl P h osp h a te a t Low Tem p er a tu r es
,
†
‡
G. David Mendenhall* and Duane B. Priddy
Department of Chemistry, Michigan Technological University, Houghton, Michigan 49931, and Styrenics
Research, Dow Chemical Company, Midland, Michigan 48667
Received November 30, 1998
The reaction of ozone with triphenyl phosphite (P) at -78 °C affords a labile 1:1 complex (PO3)
together with small amounts of triphenyl phosphate (PO) (Q. E. Thompson, J . Am. Chem. Soc.
1
961, 83, 845). In this work we found that the amount of PO present initially after complete
ozonation of P in toluene was 12 ( 2% at -78 °C and 11 ( 2% at -95 °C. Partial ozonation of
solutions of P in toluene at -78 °C gave mixtures of P, PO, and PO3 whose composition changed
with time as a result of the reaction of P with PO3 to give additional PO. Between -25 and -60
-
1
-1
°
-
C, the rate constant of the latter reaction is given by the expression log k (M
s
) ) (8.64 ( 0.04)
(11.44 ( 0.74) kcal/RT. This reaction at -78 °C is too slow to account for the PO formed during
the ozonation, which is proposed to arise instead by competitive reactions of an intermediate. The
solubility of PO in toluene at -78 °C was measured as 0.06 M, and that of PO3 about 6 times
greater.
In tr od u ction
Sch em e 1
In 1961 Thompson reported the observation of labile
intermediates in the low-temperature ozonation of triph-
enyl phosphite, which had oxidizing powers and decom-
posed above -30 °C with evolution of oxygen and
1
formation of triphenyl phosphate. This system subse-
quently attracted attention as a source of singlet molec-
ular oxygen2
-5
and because of an unusual “singlet
oxygenoid” reaction involving bimolecular insertion of two
Exp er im en ta l Section
6
-8
oxygen atoms into organic molecules (Scheme 1).
The
system is interesting from a synthetic standpoint because
the slow, direct reaction with oxidizable substrates (A )
sulfides, phosphines, phosphites, reactive olefins) via
reaction 3 potentially can give a higher yield of oxidized
product (AO ) than the much faster reaction of singlet
2
oxygen (reaction 4) because of losses by the competing
fast decay of singlet to triplet molecular oxygen (reaction
Triphenyl phosphite was vacuum distilled from Na and
divided into screw-cap vials of 2-mL capacity, which were
discarded after 1-2 openings to avoid accumulation of hydro-
lyzed products. The ozonide was prepared by passing oxygen
from a cylinder through a commercial Welsbach ozonizer and
bubbling the effluent into solutions of the phosphite at -78
°C (acetone/dry ice for all preparations unless otherwise stated)
or -95 °C (toluene/liquid N ). For solubility studies, the excess
2
5
). Since the rate of the bimolecular reaction 3 increases
ozone was removed with a stream of nitrogen. The ozone-
oxygen stream passed into one arm of a glass T-joint with the
vertical opening attached to a steel needle inserted in a 5-mm
NMR tube (hood) and the second arm to a short piece of plastic
tubing that could be constricted with a screw-clamp to control
the rate of gas flow into the NMR tube. The concentrations of
P, PO, and PO3 were determined from the initial concentration
of P, and the integrated ratios of the three products by NMR
in a Varian 400 MHz instrument. A total sample (solvent and
solute) weight of 0.60 ( 0.02 g was maintained for the kinetic
experiments, except for that shown in Figure 3 (0.56 g).
Density changes of a 0.1 M solution of P in toluene were
measured by filling a 5-mm NMR tube with the solution and
noting the change in length of the liquid column at a series of
temperatures between 25 and -78 °C. Correction factors of
with reactant concentrations, the limiting solubility of
PO3 at low temperature was also of interest. Finally
triphenyl phosphite itself is an oxidizable substrate and
can reduce its ozonide, so the measurement of the rate
constant of that reaction was desired, with the objective
of optimizing the overall system for synthetic applica-
tions.
†
Michigan Technological University.
Dow Chemical Company.
‡
(
1) Thompson, Q. E. J . Am. Chem. Soc. 1961, 83, 845-851.
(2) Review: Mendenhall, G. D. In Advances in Oxygenated Processes;
Baumstark, A. L., Ed.; J AI Press: Greenwich, CT, 1990; Vol. 2, pp 203-
31.
3) Wasserman, E.; Murray, R. W.; Kaplan, M. L.; Yager, W. A. J .
Am. Chem. Soc. 1968, 90, 4160-4161.
4) Murray, R. W.; Kaplan, M. L. J . Am. Chem. Soc. 1969, 91, 5358-
364.
2
1
.04, 1.06, and 1.08 were determined for -25, -40, and -60
(
°
C, respectively, and applied to derive corrected low-temper-
(
ature concentrations. Temperature calibration of the NMR
cavity was carried out in the usual way with methanol, which
showed a temperature of -42 °C at an instrument setting of
5
2
1
(
(
5) Falk, H.; Leimhofer, J . Monatsh. Chem. 1995, 126, 85-90.
6) Bartlett, P. D.; Mendenhall, G. D. J . Am. Chem. Soc. 1970, 92,
-
40 °C. An error of (2 °C was assumed in our measurements.
10.
(
(
Satisfactory, narrow resonances were acquired without de-
coupling or lock. Shimming was achieved by optimizing the
intensity of the FID. Integration of the 31P NMR spectra of
7) Koch, E. Tetrahedron 1970, 26, 3503-1319.
8) Dussault, P. H.; Woller, K. R. J . Org. Chem. 1997, 62, 1556-
559.
1
0.1021/jo982339y CCC: $18.00 © 1999 American Chemical Society
Published on Web 07/23/1999

5
784 J . Org. Chem., Vol. 64, No. 16, 1999
Mendenhall and Priddy
solutions with known amounts of P and PO indicated an error
of about (2%.
We used the same device with a T-joint to produce a high
signals at high vertical expansion. Analysis of the solution by
GC showed no change within error in the ratio of areas of PO
and toluene.
concentration of ozone in about 0.6 mL of CFCl
3
in a tared
5
-mm NMR tube that was immersed for brief periods in liquid
Resu lts
nitrogen. The inlet tube was then removed, and the contents
of the NMR tube frozen completely in liquid nitrogen. A small
Rea ction of P w ith P O3. The reduction of PO3 by P
amount of CFCl
3
containing 8.0 µL (31 µmol) of triphenyl
to give PO is assumed to proceed by the following steps:
phosphite was added, and the tube was capped loosely and
warmed in a dry ice/acetone bath. The contents were then
mixed with the aid of a steel needle previously cooled briefly
in liquid nitrogen. After mixing, an excess of ozone was
indicated by a light blue color. Subsequent integration of the
P + PO3 f PO + PO2
P + PO2 f 2PO
rate constant k2
k3 . k2
3
1
P signals at -60 °C revealed a composition of 12.73% PO
The species PO2 may be similar to the initial reactive
product of singlet molecular oxygen and phosphites,
which reacts rapidly with phosphite to give phosphate
and 87.27% PO3. The contents of the tube were then weighed
by difference (1.13 g), from which we estimated a total
phosphorus concentration of 32 mM.
9
In a second experiment carried out similarly, triphenyl
phosphite (8.0 µL) in 1.29 g of solution in CFCl was ozonized
3
as the isolated product. A second, kinetically equivalent
possibility is that the rate-limiting step is insertion of P
into PO3 to give a labile five-membered heterocycle. Since
we did not observe any new P signals that could be
ascribed to these intermediates, we assume that the
intermediate in either case is reduced rapidly by a second
molecule of P to PO even at low temperatures and is
present only in low concentrations. Applying the steady-
state assumption to PO2, we obtain the relations
at -78 °C in an NMR tube, and 2.0 min elapsed until the
appearance of a blue color in the solution. Subsequent analysis
by NMR at -60 °C showed the presence of 11.75% PO and
31
8
8.25% PO3.
Solu bility of P a n d P O3 in Tolu en e a t -78 °C. A solution
of triphenyl phosphate (0.49 g) in 6.28 g of toluene in a serum-
capped test tube was placed in a freezer at -78 °C. Portions
of the supernatant were removed at intervals for GC analysis
(
HP Model 5890, J &W Scientific DB-WAX/0.25 µm poly-
(ethylene glycol) film, 0.32 mm × 30 m capillary column,
dPO/dt ) 3k [P][PO3]
(1)
(2)
(3)
2
injection 325 °C, oven 300 °C, FID detector 325 °C). The weight
fractions of PO initially and after 19 and 23 days were 0.072,
-dPO3/dt ) k [P][PO3]
2
0
.022 ( 0.005, and 0.022 ( 0.005, respectively. A second trial
with 0.67 g of PO and 1.98 g of toluene (wt. fraction PO 0.253)
-dP/dt ) 2k [P][PO3]
2
after 1 and 4 days at -78 °C similarly gave values of 0.045 (
0
.012 and 0.027 ( 0.004, respectively. Additional values could
From modification of the usual second-order equation,
we obtain also
not be obtained because the liquid phase became indistinct.
Solid triphenyl phosphite ozonide was prepared by dilution
of a solution in CFCl
solution was prepared from 2.04 g of P and a total of 30 mL of
CFCl at -78 °C, by addition of a solution of P in CFCl to
magnetically stirred ozonized solvent at -78 °C in a 50 mL,
three-necked flask equipped with an addition funnel, a glass
gas-inlet tube, and a U-shaped drying tube. The ozonized
3
with petroleum ether. The ozonide
ln([PO3] /[P] ) + ln([P] /[PO3] ) )
o
o
t
t
3
3
([P] - 2[PO3] )k t (4)
o o 2
At the lowest temperature studied, -60 °C, the reac-
tion was too slow to follow to completion and initial rates
of decay were measured (Figure 1; Table 1).
The slopes of the data in Figure 1 were analyzed
according to the approximation
solution, purged of excess ozone with N
2
, was then added under
positive N pressure to 200 mL of low-boiling petroleum ether
2
at -78 °C. After standing at -78 °C for 1 day (chest freezer),
the supernatant was quickly poured off while playing a stream
of N into the flask. The flask was allowed to stand at -78 °C
2
with a paper towel inserted into the flask to blot up residual
solvent. Exposure to air in these operations was minimized,
and containers were closed with serum caps except when
transferring materials; serum caps at -78 °C were carefully
warmed by hand until they could be removed.
∆[PO3]/∆t ) -k [P] [PO3]
m
(5)
2
m
where the subscript refers to the mean value during the
experiment. Similar expressions of course defined the
changes in P and in PO. The slopes of the plots for P,
PO, and PO3 were experimentally in the ratio -2:3:-1,
A solution of PO3 was prepared as described above by
inverse addition from P (2.63 g) and a total of 6.0 mL of
toluene. The milky liquid was allowed to stand overnight at
2
and the three derived values of k were within a few
percent of each other. The average of the three values
generated one point for the Arrhenius calculation.
At -40 and -25 °C, the reaction of P with PO3 could
be followed conveniently to completion, and the data were
fitted to eq 4 over ranges of concentrations where the
successive NMR integrations of the smallest component
did not fluctuate wildly. Plots of data obtained at -25
-
78 °C. The towel was removed from the flask containing solid
PO3, and the solution of PO3 in toluene was poured into the
flask. The slurry was agitated, stoppered with a serum cap,
and allowed to stand at -78 °C with intermittent agitation.
After 10 days, an attempt to remove an aliquot of supernatant
with a precooled pipet revealed a milky suspension.
A glass-fritted funnel and receiver were cooled to -78 °C in
the chest-freezer, and a portion of the slurry containing PO3
was added to the funnel. After standing overnight at -78 °C,
about half of the slurry had passed through the filter. The clear
filtrate was removed from the freezer, placed in a sample vial,
and allowed to warm to room temperature. Analysis by GC
revealed a weight fraction of PO ) 0.15 ( 0.02.
°
C are shown in Figure 2 (see Table 1). The reaction at
25 °C was fast, so that reactant concentrations de-
-
creased rapidly to levels whose NMR signals could not
measured accurately. At the encouragement of a referee,
we repeated the reaction at -25 °C with a more frequent
rate of data acquisition and with manual instead of
A control experiment to determine the reactivity of triphenyl
phosphate toward ozone was conducted by ozonizing 0.10 M
PO in toluene at -78 °C for 40 min, followed by flushing with
(
9) (a) Bolduc, P. R.; Goe, G. L. J . Org. Chem. 1974, 39, 3178-3179.
b) Tsuji, S.; Kondo, M.; Ishiquro, K.; Sawak, Y. J . Org. Chem. 1993,
58, 5055-5059.
3
1
N
2
until colorless. The P NMR spectrum at 25 °C showed
(
neither broadening of the signal from PO nor any additional

Ozone-Triphenyl Phosphite System
J . Org. Chem., Vol. 64, No. 16, 1999 5785
F igu r e 1. Initial changes of (top to bottom) PO3, P, and PO
in toluene at -60 °C (Table 1, entry 7).
F igu r e 3. Second-order plot of the reaction of P with PO3
according to eq 7, corresponding to entry 4 in Table 1.
Ta ble 1. Rea ction of Tr ip h en yl P h osp h ite w ith
Tr ip h en yl P h osp h ite Ozon id e
2
scattered (r ) 0.68). A second experiment fortuitously
showed initial concentrations of [P] ) 2[PO3] within
error, and eq 4 no longer applied. From this unique
condition we derive a different set of equations:
[
P]o, mM
[PO3]o, mM
T, °C
-25
k2, M- s-1
1
r2
1
2
3
4
5
6
7
8
49.2
14.2
61.4
38.5
21.4
65.2
50.1
119
7.27
50.5
10.8
19.2
28.9
17.6
23.3
27.7
0.0400
0.0360
0.0329
0.0405
0.0071
0.00879
0.919
0.988
0.914
1
/[P] - 1/[P ] ) k t
(6)
(7)
o
2
0.96
-40
-60
0.978
0.904
0.90-0.98
0.97-0.99
3
/([Ptot] - [PO]) ) 2k t + constant
2
a
0.000796
0.00086
a
) 0.029 M-1
Application of eq 6 to the data led to k
2
a
s-1, but the plot was scattered (r ) 0.87). The use of eq
2
Average value derived from initial decays of P, PO, and PO3.
7
was more satisfactory, since the term Ptot, the total
concentration of P compounds, is constant, and the
concentration of PO increased with time, so that the
precision of its NMR integral did not deteriorate as the
reaction proceeded. Unfortunately the data acquisition
was accidentally stopped when the reaction was about
6
0% complete, but the plot (Figure 3) was satisfactory
2
-1 -1
(
r ) 0.96) and gave k
Regression analysis of the Arrhenius plot of the reliable
data for this reaction (Figure 4) led to activation param-
2
) 0.0405 M
s .
-
1
-1
eters of E
8
generalized value of 10
a
) 11.44 ( 0.74 kcal/mol and log A (M
s ) )
.64 ( 0.04. The value of the A factor lies within a
8
.5(0.5
-1 -1
M
s
proposed for the
somewhat analogous second-order reaction of radicals
with molecules.10
Solu bilities. Ozonation of concentrated solutions of
triphenyl phosphite in a number of solvents gives opaque
solutions due to the presence of solid material, which we
suspected was PO. The solubility of PO in toluene was
measured by sampling the supernatant above a mixture
of the two, which revealed that equilibration at -78 °C
required several days. When solid PO3 was slurried in a
concentrated solution of P ozonized in toluene (containing
PO3 and PO) at -78 °C, the liquid phase after filtration
at that temperature and warming contained PO derived
both from the initial synthesis of PO3 and from its
thermal decomposition. The difference between the ob-
F igu r e 2. Second-order plots of the reaction of PO3 with P
at -25 °C. The curves from top to bottom represent entries 1,
3
, and 2, respectively, in Table 1.
automatic phasing of the signals before integration. The
first experiment led to a rate constant (0.034 M-1
s
-1
) in
(10) Benson, S. W. Thermochemical Kinetics, 2nd ed.; Wiley-Inter-
agreement with the others, but the plot was badly
science: New York, NY, 1976; p 156.

5
786 J . Org. Chem., Vol. 64, No. 16, 1999
Mendenhall and Priddy
ity and intensity of the spectrum made it attractive for
automated data collection and integration.
The kinetic data in Figure 4 give an extrapolated value
-
5
-1 -1
of k
2
) 7.1 × 10
M
s
at -78 °C. In addition, two
in which the times
experiments were carried out in CFCl
3
required to react solutions of P with ozone to completion
were recorded. The side reaction between P and PO3 in
these experiments competed with the conversion of P to
PO3 and was at a maximum when the two were present
in equal concentrations. On this basis we can estimate
-
7
conservatively that an upper limit of 2 × 10 and 8 ×
-
6
1
0
M PO would have been formed during the ap-
proximate reaction times of 20 s and 2.0 min, respec-
tively, in the experiments carried out by ozonizing P in
CFCl
nation. Since we observed concentrations of about 4 ×
M phosphate in each case, the reaction between P
3
at low temperatures by inverse and normal ozo-
-
3
1
0
and PO3 cannot account for more than 0.2% of the
observed PO.
A reactive, open-chain adduct of P and ozone might
undergo reduction by a second P in competition with
cyclization to PO3. This explanation seems unlikely
because observed fractions of PO were identical within
error (0.12 ( 0.01) when ozone was added to P or P was
added to a solution containing excess ozone.
F igu r e 4. Arrhenius plot of data from Table 1.
served PO in that solution after warming and the
solubility limit of PO was calculated as the solubility of
PO3, which was 5.8-fold greater.
The solubility experiments were carried out in contain-
ers exposed to the atmosphere. The explosive character
of pure PO3 made it difficult to handle the compounds
at low temperature with complete exclusion of moisture.
We would argue that the low temperatures and low
Discu ssion
The small amounts of PO formed during the initial
ozonation can be explained most readily by stepwise
formation of PO3 through a noncyclic adduct of P and
ozone that partitions between pathways of scission and
cyclization:
3
solubility of water in toluene and in CFCl would tend
to minimize the potential effects of residual water. We
showed earlier that the rate of decomposition of PO3 in
toluene did not change when the solution was deliber-
ately exposed to moist air.12
Thompson inferred from the position of the 31P NMR
signal that the phosphorus atom in triphenyl phosphite
1
ozonide was pentacovalent, and Bartlett and Chu de-
termined with freezing-point methods that the ozonide
was monomeric.13 The greater solubility of the pentaco-
valent PO3 over PO in a nonpolar solvent is consistent
with the presence of the strong P -O dipole in the
phosphate.
It is puzzling that the ratio of PO to PO3 is experi-
mentally almost independent of temperature, because the
cyclization of an open-chain intermediate should entail
both ring strain energy and loss of O-O rotational modes
in the transition state. Perhaps intersystem crossing of
+
-
The saturated limit of PO in toluene at -78 °C is 0.06
( 0.01 M, so that ozonation of triphenyl phosphite in any
concentration in excess of 0.4 M in that solvent will tend
to precipitate PO. Our experimental solubility of PO3
corresponds to a saturated concentration of 0.37 ( 0.05
M at -78 °C. Higher concentrations might be achieved
if the solid-solution equilibration in toluene is slow, as
it is for PO.
3
a singlet to a triplet biradical, (PhO) POOO, determines
the extent of the minor pathway to PO. Alternatively,
formation of PO from P and ozone may occur through an
independent, “ballistic” transition state as suggested
recently by Carpenter.11
Thompson reported the reduction of PO3 by P and
1
noted that it proceeded rapidly only above about -40 °C.
Koch, however, reported that the same reaction took
place at -90 °C and proposed that it gave PO2 that
regenerated PO3 by disproportionation at higher tem-
Ack n ow led gm en t. We thank Mr. J erry Lutz for
assistance with the NMR experiments, Dr. Kevin Sik-
kima for some informative preliminary observations,
and the Dow Chemical Company for financial support.
The NMR spectrometer was obtained under NSF equip-
ment grant CHE-9512445.
7
peratures. In our hands, between -25 and -78 °C the
3
1
P NMR spectrum of samples of P incompletely ozonized
at -78 °C revealed only narrow signals of P, PO, and
PO3, and we have no evidence for these additional
reactions or species. Although NMR is generally not the
preferred method for kinetic measurements, the simplic-
J O982339Y
(
12) Mendenhall, G. D.; Kessick, R. F.; Dobrzelewski, M. J . Photo-
(
0.
11) Carpenter, B. K. Angew. Chem., Int. Ed. Engl. 1998, 37, 3340-
chem. 1984, 25, 227-234.
(13) Bartlett, P. D.; Chu, H.-K. J . Org. Chem. 1980, 45, 3000-3004.
5