Reaction of diphenyldiazomethane with singlet oxygen studied by time-resolved IR spectroscopy

Article abstract of DOI:10.1021/jo800955w

(Chemical Equation Presented) The mechanism of the reaction of diphenyldiazomethane 4a with singlet oxygen has been investigated by nanosecond time-resolved UV-vis (LFP) and IR (step-scan) spectroscopy. The experiments were performed with fullerene (C₆₀) as photosensitizer for the generation of ¹O₂ in nonpolar solvents (toluene and CCl ₄). The UV-vis experiments allowed us to monitor the formation of benzophenone O-oxide 1a, while in the IR experiments the bleaching of 4a and the formation of benzophenone 7a and N₂O was observed. The kinetic data were evaluated using Monte Carlo simulation and DFT calculations. These methods allow us to present a consistent mechanistic scheme for the reaction of 4a with ¹O₂ and to explain why the elusive dioxadiazole 5a as key intermediate is not directly observed.

Full text of DOI:10.1021/jo800955w

Reaction of Diphenyldiazomethane with Singlet Oxygen Studied by
Time-Resolved IR Spectroscopy
Joel Torres-Alacan and Wolfram Sander*
Lehrstuhl fu¨r Organische Chemie II der Ruhr-UniVersita¨t, D-44780 Bochum, Germany
wolfram.sander@rub.de
ReceiVed May 2, 2008
The mechanism of the reaction of diphenyldiazomethane 4a with singlet oxygen has been investigated
by nanosecond time-resolved UV-vis (LFP) and IR (step-scan) spectroscopy. The experiments were
performed with fullerene (C60) as photosensitizer for the generation of 1O2 in nonpolar solvents (toluene
and CCl4). The UV-vis experiments allowed us to monitor the formation of benzophenone O-oxide 1a,
while in the IR experiments the bleaching of 4a and the formation of benzophenone 7a and N2O was
observed. The kinetic data were evaluated using Monte Carlo simulation and DFT calculations. These
1
methods allow us to present a consistent mechanistic scheme for the reaction of 4a with O2 and to
explain why the elusive dioxadiazole 5a as key intermediate is not directly observed.
SCHEME 1. Mechanism of the Synthesis of Carbonyl
Oxides 1 via the Singlet Route from Diazo Compounds 4
Introduction
Carbonyl oxides 1 are highly reactive, powerful oxidants that
are far less investigated than their isomeric dioxiranes 2.1-3 The
rearrangement to dioxiranes requires photochemical activation
and has been used for the synthesis of stable dioxiranes.4 The
thermal cyclization of 1 to 2 has never been observed.
There are several routes for the synthesis of carbonyl oxides
1: (i) the cycloreversion of 1,2,3-trioxolanes during the ozo-
nolysis of alkenes (Criegee reaction);5–7 (ii) the reaction of triplet
carbenes 3 with molecular (triplet) oxygen, the triplet route;8,9
and (iii) the reaction of diazo compounds 4 with singlet oxygen,
(1) Sander, W. W.; Patyk, A.; Bucher, G. J. Mol. Struct. 1990, 222, 21–31.
(2) Block, K.; Kappert, W.; Kirschfeld, A.; Muthusamy, S.; Schroeder, K.;
Sander, W.; Kraka, E.; Sosa, C.; Cremer, D. Peroxide Chem. 2000, 139–156.
(3) Sander, W.; Block, K.; Kappert, W.; Kirschfeld, A.; Muthusamy, S.;
Schroeder, K.; Sosa, C. P.; Kraka, E.; Cremer, D. J. Am. Chem. Soc. 2001, 123,
2618–2627.
(6) Criegee, R. Angew. Chem. 1975, 87, 765–771.
(4) Sander, W.; Schroeder, K.; Muthusamy, S.; Kirschfeld, A.; Kappert, W.;
Boese, R.; Kraka, E.; Sosa, C.; Cremer, D. J. Am. Chem. Soc. 1997, 119, 7265–
7270.
(7) Story, P. R.; Burgess, J. R. J. Am. Chem. Soc. 1967, 89, 5726–5727.
(8) Bartlett, P. D.; Traylor, T. G. J. Am. Chem. Soc. 1962, 84, 3408–3409.
(9) Casal, H. L.; Tanner, M.; Werstiuk, N. H.; Scaiano, J. C. J. Am. Chem.
Soc. 1985, 107, 4616–4620.
(5) Criegee, R.; Wenner, G. Liebigs Ann. Chem. 1949, 564, 9–15.
7118 J. Org. Chem. 2008, 73, 7118–7123
10.1021/jo800955w CCC: $40.75 2008 American Chemical Society
Published on Web 08/14/2008

Reaction of Diphenyldiazomethane with Singlet Oxygen
the singlet route.9-13 While there has been a lot of indirect
evidence collected that carbonyl oxides 1 are indeed intermedi-
ates in the Criegee reaction, so far it was not possible to detect
1 as an intermediate in these reactions by direct spectroscopic
methods. If 1 is indeed formed (which is generally accepted),
it is too short-lived to be directly detected under these conditions.
3
In contrast, the reaction of carbenes with O2 is very efficient
and makes it possible to detect 1 via laser flash photolysis (LFP)
in solution9 or via matrix isolation spectroscopy at cryogenic
temperatures in inert matrices.14-17 As expected, triplet carbenes
react much faster than singlet carbenes; however, in low
temperature matrices it was also possible to obtain carbonyl
oxides by oxygenation of singlet carbenes such as phenylchlo-
rocarbene.18 By reacting the sterically hindered dimesitylcarbene
with 3O2, a kinetically stabilized carbonyl oxide was synthesized
that at -78 °C was stable enough to be characterized via NMR
spectroscopy.19 The disadvantage of the triplet route is the
intermediacy of highly reactive carbenes 3, which frequently
results in unwanted side reactions.
FIGURE 1. Transient spectrum recorded 11 µs after the 532 nm
excitation (6-7 ns, ∼130 mJ/pulse) of a solution of C60 (0.0132 mmol/
L), 4a (1 mmol/L), and O2 (9.88 mmol/L) in toluene. The triangles
represent experimental points. Averaging over a 1 µs time window (10
data points) was performed to improve the signal-to-noise ratio. The
peak at 410 nm is assigned to carbonyl oxide 1a.
Results and Discussion
The singlet route also starts from diazo compounds 4 but
circumvents the formation of free carbenes 3. The key inter-
mediate of the singlet route is 1,2,3,4-dioxadiazole 5, a short-
lived heterocycle that decays via fast cycloreversions either to
a carbonyl oxide 1 and N2 or to a ketone 7 and N2O (Scheme
1). The dioxadiazole 5 is formed either via a concerted 1,3-
dipolar cycloaddition of 4 and 1O220 or stepwise with zwitterion
6 as an intermediate. Neither 5 nor 6 has been observed by direct
spectroscopic methods so far, and all experimental evidence
comes from measuring the ratio of N2O/N2 by gas chromatog-
raphy by Sawaki et al.11,21
LFP with UV-vis Detection. In the LFP experiments
reported by Scaiano et al., acetonitrile was used as solvent and
methylene blue with 587-nm excitation as triplet sensitizer for
the generation of 1O2.13 Since in our experiments shorter
wavelength excitation at 532 nm (second harmonic of a Nd:
YAG laser) was used where methylene blue has a very low
absorption coefficient, we used fullerene (C60) as sensitizer. In
toluene the absorption maxima of fullerene are at 457, 509, and
747 nm,22 and the lifetime of its triplet state is estimated to
>0.28 ms.22 Since the quantum yield of phosphorescence is
very low, the main mechanism of deactivation presumably is
triplet-triplet annihilation. Fullerene as sensitizer has the ad-
ditional advantages that in nonpolar solvents such as toluene
or CCl4 it is much more soluble than methylene blue and that
it is photochemically more stable.
Benzophenone O-oxide 1a (R ) Ph) is the prototype of a
carbonyl oxide and was investigated by classical trapping, matrix
isolation,16 and UV-vis nanosecond time-resolved spectroscopy
(LFP).13 In solution 1a was produced via both the triplet and
the singlet route as a relatively long-lived intermediate (>100
µs) with an absorption maximum at 410 nm.13 As a result of
the strong visible absorption at λ ) 410 nm, carbonyl oxide 1a
is readily detected in LFP experiments, while the other proposed
intermediates or products (Scheme 1) are not observable under
these conditions. Here, we describe a combined time-resolved
UV-vis and IR investigation of the reaction of diphenyldiaz-
Several blind experiments proved that C60 is indeed an
1
efficient sensitizer for the formation of O2 and that diphenyl-
diazomethane 4a is only bleached in the presence of both C60
and molecular oxygen. Thus, after excitation (532-nm laser
pulse) of a 1 mmol/L solution of 4a in toluene in the absence
of C60, no reaction or transient is observed in the absence or
presence of dissolved oxygen. Irradiation of an argon-purged
solution of C60 produces the expected long-lived triplet C60,
which in the presence of oxygen is rapidly quenched (see
Supporting Information, Figure S4).
1
omethane 4a with O2. The combination of time-resolved IR
spectroscopy with a time resolution of about 100 ns and UV-vis
spectroscopy with 10 ns time resolution provides detailed insight
into the singlet route for the synthesis of carbonyl oxides 1.
If a solution of 4a and C60 in toluene is excited in the absence
3
of oxygen, only the transient C60 is observed. In the presence
of oxygen and otherwise identical conditions, however, the
strong 410-nm absorption of 1a is easily detected. The UV-vis
absorption spectrum of 1a (Figure 1) formed under these
conditions is in excellent agreement with the data reported by
Scaiano et al.,13 which clearly show that C60 is a suitable
(10) Higley, D. P.; Murray, R. W. J. Am. Chem. Soc. 1976, 98, 4526–4533.
(11) Nojima, T.; Ishiguro, K.; Sawaki, Y. Chem. Lett. 1995, 545–546.
(12) Casal, H. L.; Sugamori, S. E.; Scaiano, J. C. J. Am. Chem. Soc. 1984,
106, 7623–7624.
(13) Scaiano, J. C.; McGimpsey, W. G.; Casal, H. L. J. Org. Chem. 1989,
54, 1612–1616.
(14) Bell, G. A.; Dunkin, I. R. J. Chem. Soc., Chem. Commun. 1983, 1213–
1215.
(15) Dunkin, I. R.; Bell, G. A. Tetrahedron 1985, 41, 339–347.
(16) Sander, W. Angew. Chem. 1986, 98, 255–256.
(17) Bucher, G.; Sander, W. Chem. Ber. 1992, 125, 1851–1859.
(18) Ganzer, I.; Hartmann, M.; Frenking, G. Chem. Org,c Germanium, Tin
Lead Compd. 2002, 2, 169–282.
1
sensitizer for the generation of O2 in LFP experiments. For
the formation of 1a, a pseudo-first-order rate constant of (2.2
( 0.3) × 105 s-1 is found (Table 1, Figure 2). The decay of 1a
is a second-order process (dimerization)13 and clearly much
slower than its formation. The formation of 1a and the bleaching
of 4a (see below) corresponds to rise and decay times on the
order of 2 µs, which is clearly much shorter than the lifetime
(19) Sander, W.; Marquardt, R.; Bucher, G.; Wandel, H. Pure Appl. Chem.
1996, 68, 353–356.
(20) Bethell, D.; McKeivor, R. J. Chem. Soc., Perkin Trans. 1977, 2, 327–
333.
(21) Nojima, T.; Ishiguro, K.; Sawaki, Y. J. Org. Chem. 1997, 62, 6911–
6917.
(22) Ebbesen, T. W.; Tanigaki, K.; Kuroshima, S. Chem. Phys. Lett. 1991,
181, 501–504.
J. Org. Chem. Vol. 73, No. 18, 2008 7119

Torres-Alacan and Sander
experiment
TABLE 1. Summary of Experimental Conditions and Kinetic Dataa
4a (mmol/L)
C60 (mM, mmol/L)
3O2 (mmol/L)
solvent
rate constants (s-1
)
∼1
∼2
0.0132
0.26
9.88
9.88
toluene
toluene
kgrowth(1a) ) (2.2 ( 0.3) × 105
kdecay(4a) ) (3.32 ( 0.09) × 105
kgrowth(N2O) ) (2.8 ( 0.4) × 105
kdecay(4a) ) (5.1 ( 0.4) × 105
TR-UV-vis
step-scan TR-FTIR
∼2.5
0.26
12.44
CCl4
step-scan TR-FTIR
kgrowth(N O) ) (4.2 ( 0.3) × 105
2
a
kgrowth(7a) ) (4.3 ( 0.3) × 105
b
kgrowth(7a) ) (7.2 ( 0.4) × 105
a Rate constant determined at 1665 cm-1 b Rate constant determined at 1278 cm-1 a First-order or pseudo-first-order behavior assumed. The
. .
statistical errors are shown with the kinetic data. In addition, there are systematic errors caused by uncertainties in the initial concentrations and different
experimental set ups in LFP and step scan experiments.
FIGURE 2. Transient trace recorded at 410 nm following 532 nm
excitation (6-7 ns, ∼130 mJ/pulse) of a solution of C60 (13.2 µmol/
L), 4a (1 mmol/L), and O2 (9.88 mmol/L) in toluene. Experimental
data were fit using a biexponential model. Exponetial rates k1 and k2
are indicated with standard errors in parenthesis. The dashed line
3
indicates the residue due to C60.
23
1
of O2 in toluene (29 µs) or CCl4 (700 µs). Thus, the decay
1
of O2 in these solvents does not influence our mechanistic
scheme.
Step-Scan FTIR Detection. IR experiments were performed
in toluene as solvent under similar conditions as the UV-vis
experiments described above, only the concentration of 4a was
increased to 2 mmol/L to optimize the IR detection (Table 1).
Again, 4a proved to be stable in the absence of oxygen and/or
C60. In the presence of both oxygen and C60, singlet oxygen is
produced and 4a is bleached (monitored at the Nt N stretching
vibration at 2041 cm-1) with a pseudo-first-order rate constant
of (3.32 ( 0.09) × 105 s-1. Because of strong absorptions of
toluene in large parts of the IR spectrum, N2O with absorption
at 2219 cm-1 was the only product observed. The rate for the
formation of N2O ((2.8 ( 0.4) × 105 s-1) is similar to that of
the decay of 4a and the formation of 1a.
FIGURE 3. (A) Difference spectra recorded 20 ns (red), 2 µs (green),
and 32 µs (blue) after 532-nm excitation (6-7 ns, ∼26 mJ/pulse) of a
solution of 4a (2.5 mmol/L), C60 (0.26 mmol/L), and O2 (12.44 mmol/
L) in CCl4. Besides statistical errors (noise) the spectra also show
artifacts presumably caused by fluctuations of the laser intensity and a
decrease of the optical quality of the sample/mirror set up during the
experiments. (B) IR spectrum of N2O in CCl4. (C) IR spectrum of 7a
in CCl4. (D) IR spectrum of 4a in CCl4. (E) Simulated difference spectra
based on Monte Carlo and B3LYP/6-31G(d) calculations. Arrows
indicate time evolution.
Additional experiments were performed in CCl4 as solvent,
which has the advantage of a much higher IR transmittance in
most parts of the spectrum but the disadvantage of a low
solubility of C60. However, since the initial experiments in
toluene proved that a concentration of 0.2 mg/mL of C60 is high
enough for the sensitization, CCl4 is well-suited for these
experiments. In CCl4 a barely visible absorption at 1665 cm-1
is assigned to the CdO stretching vibration of benzophenone
7a (Figure 3). A careful analysis of the kinetic traces at 1665
cm-1 clearly indicates an absorption band and not an artifact at
this position. Artifacts in the difference spectra result in a
background that is apparently statistic in frequency but not in
time (background at different times runs parallel). Presumably,
these artifacts are caused by fluctuations of the laser intensity
and a decrease of the optical quality of the sample/mirror set
up during the experiments. In addition, the 2219 cm-1 absorption
of N2O is also observed. Within the error limits the formation
of 7a and N2O shows the same kinetics as the decay of 4a
(Figure 4). Other transients and especially any band that could
be attributed to dioxadiazole 5a or triplet 4a could not be
detected.
Kinetic Simulation. The UV-vis and IR experiments clearly
show two major reaction channels of the elusive intermediate
5a, the formation of 1a observed in the UV-vis experiment
and N2 (not directly observed) on one hand, and the formation
of 7a and N2O, both observed in the IR experiments, on the
other hand. Although 1a and 7a/N2O could not be observed in
the same experiment, it is clear that both are formed simulta-
neously, as was deduced previously by measuring the N2/N2O
ratio by gas chromatography.11,21
(23) Salokhiddinov, K. I.; Byteva, I. M.; Gurinovich, G. P. Zh. Prikl.
Spektrosk. 1981, 34, 892–897.
7120 J. Org. Chem. Vol. 73, No. 18, 2008

Reaction of Diphenyldiazomethane with Singlet Oxygen
FIGURE 4. Transient traces after 532-nm excitation (6-7 ns, ∼26 mJ/pulse) monitored at (A) 1665, (B) 2041, (C) 1278, and (D) 2219 cm-1. (A)
and (C) show traces assigned to 7a, (B) is assigned to 4a, and (D) is assigned to N2O. Solid curves show least-squares fits to a single exponential.
SCHEME 2. Mechanistic Scheme Describing the Reaction
1
of 4a with O2
To describe the formation of 1a and 7a from 4a and singlet
oxygen, a simple mechanistic scheme with 5a as the presumed
key intermediate, k1 describing the formation of 5a, and k2 and
k3 describing the exit channels is used (Scheme 2). The
experimental kinetic data from the LFP and step scan FTIR
experiments assuming exponential decay and growth (first-order
FIGURE 5. Time profile at (A) 2041 and (B) 2219 cm-1 after 532 nm
excitation of a toluene solution containing 2 mmol/L of 4a and 0.26
mmol/L of C60 saturated with 3O2 (circles). In order to find the
or pseudo-first-order behavior assumed) are summarized in
normalizing factor for the experimental data, the last 16 experimental
Table 1.
data points were averaged. The dashed line (in blue) represents the
least-squares-fit to an exponential decay function with nonzero intercept.
The solid line (in red) represents the Monte Carlo simulation of 4a
based on the mechanistic scheme shown in Scheme 1. Parameters for
the MC simulation: n1 ) 280, n2 ) 1023, n3 ) 750. A 300 × 300 grid
The experimental kinetic data from the LFP experiments were
evaluated using Monte Carlo simulation. With the three rate
constants k1 ) 7.7 × 107 L mol-1 s-1, k2 ) 7.81 × 105 s-1, k3
) 5.73 × 105 s-1 obtained from the Monte Carlo simulation,
the decay and growth of reactant and products are nicely fitted
(Figure 5). The rate constant k1 is on the order of the value
obtained by quenching experiments in acetonitrile/dichlo-
romethane by Scaiano et al. (3.6 × 108 L mol-1 s-1). Since the
formation of 5a is the rate-limiting step (k1), the two subsequent
faster parallel reactions leading to 1a (k2) and 7a (k3),
respectively, cannot be accurately determined by modeling our
1
was initialized randomly with 4a and O2 in a ratio of 9:1 and 2200
cycles were scaled to 32 µs (ꢀ ) 6.875 × 107 cycles/s). The elementary
reactions were processed statistically, and four runs were averaged in
order to cancel out fluctuations.
experimental data. Therefore, the ratio k2/k3 was fixed to 1.36
during the Monte Carlo simulation to be consistent with the
experimental results.21
J. Org. Chem. Vol. 73, No. 18, 2008 7121

Torres-Alacan and Sander
and rapidly decomposes via two parallel reactions to form either
the desired 1a and N2 or benzophenone 7a and N2O. The kinetic
simulations reveal that the maximum concentration of 5a
building up is too low to be detectable in our experiments. The
kinetic data extracted from the LFP experiments are in good
agreement with the previously published scheme based on the
analysis of the N2/N2O ratio.21
Experimental Section
General Methods. C60 (99.9%) was obtained from a commercial
supplier and used without further purification; 4a was prepared
according to literature procedures.24
FIGURE 6. Monte Carlo simulation according to the mechanism
presented in Scheme 1. Parameters for the simulation: n1 ) 280, n2 )
1023, n3 ) 750, concentration scale coefficient R ) 360 L mol-1. The
Step-Scan Time-Resolved FTIR. A modified Bruker IFS 66v/S
FTIR spectrometer with step-scan option, a Nd:YAG laser, and
a sample cell connected to a continuous flow system was used
for the step-scan experiments. The flow cell was positioned
outside the spectrometer and the IR spectra measured in reflection
using a mirror unit in the spectrometer and a gold-coated mirror
on the back of the flow cell. A photovoltaic MCT detector
(risetime 25 ns) with internal preamplifier was used. The AC-
coupled output of the preamplifier is fed directly into a second
amplifier and then read out by a 200 MHz, 8 bit transient recorder
connected to a PC. Positioning of the interferometer mirror and
data acquisition was performed using OPUS software. The
solution was pumped trough the sample cell (two CaF2 windows,
0.4 mm Teflon spacer) by a peristaltic pump with low pulsation
(Ismatec IPC-N-4, ca. 2 mL/min flow) and transferred back into
a storage tank. The sample was excited using the second
harmonic (532 nm) of the Nd:YAG laser with a repetition rate
of 10 Hz. To avoid thermal effects and shockwaves the laser
energy was attenuated to 10-18 mJ/pulse by an external
attenuator. The laser beam was split by a variable dielectric
attenuator, and both beams crossed the sample overlapping the
IR profile (see Supporting Information for details). This reduced
the laser energy density in the sample and improved the signal-
to-noise ratio. During the measurements the spectrometer was
evacuated to 2 mbar pressure. In order to eliminate external
vibrations, the whole system was placed on a vibrationally
insulated table. All measurements were carried out with 6 cm-1
resolution in a spectral range between 0 and 3160 cm-1, resulting
in an interferogram containing 1060 points. The signal was
averaged 20 times per sampling position. In all experiments a
time range of 32 µs was recorded with a resolution of 20 ns.
After Fourier transformation the average of 20 successive spectra
was calculated to achieve a better signal-to-noise ratio, resulting
in an effective time resolution of 400 ns. The samples were
dissolved (0.26 mmol/L C60, ∼2 mmol/L 4a) in toluene and
carbon tetrachloride (spectroscopic grade) and purged with argon
or oxygen, respectively, for 60 min.
coefficient for scaling the time was set to ꢀ ) 6.875 × 107 cycles s-1
.
The elementary reactions were processed statistically, and 4 runs were
averaged in order to cancel out fluctuations.
Within the detection limits of our time-resolved IR experi-
ments, neither dioxadiazole 5a nor carbonyl oxide 1a could be
detected. We therefore calculated the IR frequencies and
absolute intensities of all species expected from our kinetic
scheme using DFT calculations. The kinetic data from the Monte
Carlo simulations were used to simulate the IR spectra of the
product mixture at various times after the initial laser pulse
(Figure 3). A comparison of the experimental difference spectra
measured in CCl4 after 20 ns, 2 µs, and 32 µs (Figure 3) with
the simulation based on the DFT calculation and Monte Carlo
simulation (Figure 3) reveals a very satisfying agreement.
1
For the simulations we assume that O2 reacts chemically
3
with 4a and physical quenching producing 4a does not take
place. Although we can not rigorously exclude physical quench-
ing, the formation of N2O and the bleaching of 4a in the
experiment nicely parallels that expected from the simulation.
This excludes that a long-lived triplet state of 4a is efficiently
formed, since this would result in a higher ratio of bleaching
of 4a compared to the formation of N2O. Even if physical
1
quenching of O2 occurs to a small extend, this would only
influence the yield of the products but not the pseudo-first-order
rate constants.
In the experiment the intensity of the CdO stretching
vibration is about twice that of the baseline noise. Thus, bands
with a lower intensity will be below the detection limit. This is
indeed the case for all IR bands of 1a and 5a, and therefore
these compounds cannot be detected in the IR experiments. The
Monte Carlo integration predicts the highest concentration of
5a to build up after 1.2 µs (Figure 6). This corresponds to an
estimated concentration of 32 µmol/L, which is approximately
seven times less than the detection limit.
Nanosecond Laser Flash Photolysis. A standard LFP setup was
used, consisting of a Nd:YAG laser, operated at 1 Hz and 532 nm
(100 mJ/pulse, 6-7 ns pulse duration), a pulse Xe arc lamp, a
(24) Smith, L. I.; Howard, K. L. Org. Synth. 1944, 24, 53–55.
Conclusion
(25) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.;
Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci,
B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada,
M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.;
Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian,
H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.;
Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski,
J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg,
J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.;
Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.;
Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.;
Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.;
Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill,
P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A.
Gaussian 03; Gaussian, Inc.: Pittsburgh, PA, 2003.
Carbonyl oxides 1 are highly potent oxidants that, however,
due to the lack of efficient syntheses, have not found many
synthetic applications so far. The singlet route is a promising
way starting from readily accessible precursors and avoiding
unwanted side reactions of highly reactive carbenes. Using
nanosecond time-resolved UV-vis and IR techniques in com-
bination with Monte Carlo simulation and DFT calculations
allowed us to gain mechanistic details for the synthesis of
benzopheneone O-oxide 1a via the singlet route. The rate-
determining step of the reaction sequence is the formation of
the elusive dioxadiazole 5a. This heterocycle is highly labile
7122 J. Org. Chem. Vol. 73, No. 18, 2008

Reaction of Diphenyldiazomethane with Singlet Oxygen
monochromator coupled to a photoelectron multiplier tube, and a
digital oscilloscope. The entire setup was controlled from a personal
computer using LabView software. In order to avoid depletion of
the precursor and product build-up, a flow cell was used. Toluene
solutions containing 13.2 µmol/L C60 and 1 mmol/L 4a were used
for the LFP experiments.
DFT Calculations. Calculations were performed with the
Gaussian 03 program package.25 Geometries and frequencies were
calculated at the B3LYP/6-31G(d) level of theory and scaled with
a factor of 0.961426 for better comparison with the experiment.
Monte Carlo Integration. The Monte Carlo integration of the
mechanism was accomplished according to the algorithm described
by Schaad.27 A program written using Visual Basic 6.0 was used
for the evaluation of data (the program is available from the authors
on request). A 300 × 300 grid was used, and the cells of the grid
were filled at 80% with the reactants 1O2 and 4a to avoid overflow
of the grid during the simulation. The number of molecules used
during the simulation was greater than 104, thus limiting the error
of the simulation to ∼1%.27
Dividing eq 1 by eq 2 results in
(
)
dc 4a
1
(
)
k c 4a c O
(
[
)
]
dt
1
2
)
]
(7)
(8)
(9)
n p 4a 1O
[
d 4a
[
]
1
2
dm
k1
n1pR2
ꢀ
R
)
From eq 8 we obtain for k1
k1 ) n1pRꢀ
Using the same procedure, eqs 10 and 11 are obtained for k2
and k3.
k2 ) n2pꢀ
(10)
k3 ) n3pꢀ
(11)
The values found for R and ꢀ were 360 L mol-1 and 6.875 ×
107 cycles/s, respectively, and p ) 1/(9 × 104), using the values n1
) 280, n2 ) 1023, n3 ) 750. The values for the rate constants are
then k1 ) 7.7 × 107 L mol-1 s-1, k2 ) 7.81 × 105 s-1, k3 ) 5.73
For the simulation of the step scan FTIR experiments we assumed
that the initial concentration of 1O2 equals the initial concentration
of C60 (0.26 mmol/L) and the initial concentration of 4a was 2
mmol/L (Table 1). The following equations have been used to map
the Monte Carlo simulation into the chemical system:
× 105 s-1
.
Simulation of Time-Resolved IR Spectra. Time-resolved IR
spectra were simulated by “mixing” the Monte Carlo and DFT
calculations.
The mapping was carried out using the following equation:
DFT
∆CMMC i × Si
(12)
(
)
( )
dc 4a
1
∑
t
(
)
-
) k c 4a c O
(1)
(2)
(3)
(4)
(5)
(6)
(
)
2
1
i
dt
where the sum runs over the species involved in the mechanism.
The factor ∆CtMMC(i) in the sum is the concentration (after mapping
Monte Carlo into the chemical system) of the component i at time
t minus the concentration of the component i at the initial time.
The superindex MMC stands for Mapped Monte Carlo. The factor
SiDFTinside the sum is the scaled calculated spectrum of component
i.
[
]
d 4a
) n p 4a 1O
[
]
-
[
]
2
1
dm
(
)
dc 1a
dt
(
)
) k2c 5a
[
]
d 1a
[
]
) n2p 5a
dm
Acknowledgment. J.T. is indebted to Dr. Christoph Kolano
and Dr. Goetz Bucher for their introduction to step-scan and
LFP, respectively. This work was financially supported by the
Deutsche Forschungsgemeinschaft and the Fonds der Chemis-
chen Industrie.
(
)
dc 7a
(
)
) k3c 5a
dt
[
]
d 7a
[
]
) n3p 5a
dm
Supporting Information Available: Time resolved IR
spectra recorded in toluene (S1); transient traces at 2041 and
2219 cm-1 corresponding to the experiment in toluene (S2);
Cartesian coordinates and total energies of 1a, 4a, 5a, 7a, and
N2O (S4-S8); set-up for step-scan time-resolved FTIR mea-
surements (S9); and transient traces recorded at 500 nm in
toluene containing C60 in the presence and absence of oxygen
(S10). This material is available free of charge via the Internet
at http://pubs.acs.org.
Where [x], c(x) represent the stochastic (fraction of molecules
in the grid) and real concentration respectively, and m and t the
“stochastic” and real time. The scaling coefficients are indicated
as R ([x] ) Rc(x)) and ꢀ (m ) ꢀt). The variable p is the grid
probability (e.g., a grid containing n rows and n columns will have
for every cell a constant probability p ) 1/n2).
(26) Scott, A. P.; Radom, L. J. Phys. Chem. 1996, 100, 16502–16513.
(27) Schaad, L. J. J. Am. Chem. Soc. 1963, 85, 3588–3592.
JO800955W
J. Org. Chem. Vol. 73, No. 18, 2008 7123