Photochemical reaction of ozone and benzene: An infrared matrix isolation study

Article abstract of DOI:10.1021/ja984380o

The photolysis of benzene/ozone mixtures in an argon matrix at 12 K with UV light of λ ≥ 280 nm leads to the following products: phenol, 2,4-cyclohexadienone, benzene oxide, and butadienylketene (hexa-1,3,5-trien-1-one). The identification of butadienylketene as a product is based on deuterium isotopic shifts and agreement with density functional vibrational frequency calculations. We find an average phenol/ butadienylketene branching ratio of 4.3 during the course of photolysis. This is the first report in the literature of the observation of a ketene product from the reaction of oxygen atoms with benzene.

Full text of DOI:10.1021/ja984380o

J. Am. Chem. Soc. 1999, 121, 4271-4277
4271
Photochemical Reaction of Ozone and Benzene: An Infrared Matrix
Isolation Study
James K. Parker and Steven R. Davis*
Contribution from the Department of Chemistry, UniVersity of Mississippi, UniVersity, Mississippi 38677
ReceiVed December 18, 1998
Abstract: The photolysis of benzene/ozone mixtures in an argon matrix at 12 K with UV light of λ g 280 nm
leads to the following products: phenol, 2,4-cyclohexadienone, benzene oxide, and butadienylketene (hexa-
1
,3,5-trien-1-one). The identification of butadienylketene as a product is based on deuterium isotopic shifts
and agreement with density functional vibrational frequency calculations. We find an average phenol/
butadienylketene branching ratio of 4.3 during the course of photolysis. This is the first report in the literature
of the observation of a ketene product from the reaction of oxygen atoms with benzene.
Introduction
been studied with the following isotopomers: C6H6 + 16O, C6H6
1
8
16
18
+
O, C6D6 + O, and C6D6 + O. As a complement to the
3
The reaction of benzene with oxygen ( P) atoms has been
infrared studies we have also performed density functional ab
initio calculations. We have been able to identify four products
of the reaction.
the subject of several studies in the gas phase.1
-11
As a result,
much data are available for this reaction including the follow-
ing: specific rate constants, activation energies, the identity of
reaction products, as well as proposed mechanisms. It is well-
established that phenol is a primary product of this reaction.
Because multiple reaction channels for this system exist there
have been other products reported for this reaction. For example,
Experimental Procedure
6 6 6 6
Normal benzene C H (99.8%, Aldrich) and perdeuteriobenzene C D
(99.6 atom %, Aldrich) used in these experiments were freed of
relatively volatile impurities by three consecutive freeze-pump-thaw
cycles at liquid nitrogen temperature. The benzene was warmed to room
temperature and its vapors were expanded into an evacuated sample
chamber and diluted with argon (99.995%, Air Products) to an Ar:
3
Sloane studied this reaction under single collision conditions
in crossed molecular beams and reported two channels: forma-
tion of a long-lived complex, which rearranges to phenol, and
formation of CO and a C5H6 hydrocarbon, presumably 3-penten-
16
16
16
4
benzene mole ratio of 300:1. Normal ozone O- O- O and the
18 18
1
-yne. In a subsequent study, Sibener and co-workers also
isotopically pure 18O- O- O ozone were made by passing an electric
studied this reaction under single collision conditions in crossed
molecular beams and reported two distinct reaction channels:
one leading to phenol and the other to C6H5O and H. These
workers concluded that the channel leading to CO elimination,
if it occurs at all, is relatively minor. In another study under
multiple collision conditions, workers reported the formation
of phenol and an undefined polymeric species.
Because of the fundamental significance of this reaction to
the understanding of combustion processes in aromatic systems,
we have undertaken a matrix isolation study of this reaction in
solid argon. To our knowledge, this is the first matrix isolation
study of the benzene/oxygen atom reaction. It was our goal to
isolate, and characterize, via infrared spectroscopy, the primary
products and intermediates of this reaction. The reaction has
discharge, from a Tesla coil, through a sample of oxygen gas (99.996%,
18
Air Products; O isotope from Euriso-top, 98.01% enrichment)
contained in a glass U tube which was submerged in liquid nitrogen.
In this way the ozone condensed to the solid phase on the walls of the
glass tube as it formed. The solid ozone was freed of residual oxygen
molecules by pumping at 77 K. The purified ozone was then warmed
to room temperature, expanded into a separate evacuated sample
2
3
chamber, and diluted with argon to an Ar:O mole ratio of 300:1. The
IR spectrometer and accompanying vacuum chamber have been
described in detail elsewhere.12 The samples were co-deposited on a
gold mirrored matrix support at a combined rate of 8.4 mmol/h for a
total of 24.2 mmol. The matrix support was held at 12 K with a closed
cycle, single stage helium refrigerator (APD Cryogenics, model DE-
2
02). The temperature was measured at the cold plate with a gold/
calomel thermocouple. All infrared spectra were recorded using a
-
1
Nicolet 740 FTIR spectrometer at 1-cm resolution. After the initial
spectra were recorded, the samples were subjected to the photolyzing
radiation of a 200 W Hg-Xe arc lamp (lamp model UXL 200H Hg
Xe by Ushio and model A 1010 lamp housing by Photon Technology
International). The radiation was filtered through a 51 mm water cell
and a 580, 360, or 280 nm Hoya cutoff filter before coming in contact
with the matrix sample. The temperature of the matrix sample did not
rise above 14 K during any photolysis interval. All matrix spectra were
successively photolyzed at precise time intervals.
(
(
1) Grovenstein, E., Jr.; Mosher, A. J. J. Am. Chem. Soc. 1970, 92, 3810.
2) Bonanno, R. A.; Kim, P.; Lee, J.-H.; Timmons R. B. J. Chem. Phys.
1
972, 57, 1377.
(3) Sloane, T. M. J. Chem. Phys. 1977, 67, 2267.
(4) Sibener, S. J.; Buss, R. J.; Casavecchia, P.; Hirooka, T.; Lee, Y. T.
J. Chem. Phys. 1980, 72, 4341.
5) Nicovich, J. M.; Gump, C. A.; Ravishankara, A. R. J. Phys. Chem.
982, 86, 1684.
(
1
(
(
6) Rotzoll, G. Int. J. Chem. Kinet. 1985, 17, 637.
7) Ure n˜ a, A. G.; Hoffmann, S. M. A.; Smith, D. J.; Grice, R. J. Chem.
Soc., Faraday Trans. 2 1986, 82, 1537.
8) Tappe, M.; Schliephake, V.; Wagner, H. Gg. Z. Phys. Chem. 1989,
62, 129.
The ab initio calculations were performed using the Gaussian 94
suite of programs13 running on either a Cray C98/8512 or a Silicon
Graphics Power Challenge computer. The density functional meth-
ods1 of Becke, Lee, Yang, and Parr, commonly known as B3LYP
(
1
(9) Leidreither, H. I.; Wagner, H. Gg. Z. Phys. Chem. 1989, 165, 1.
10) Sol, V. M.; van Drunen, M. A.; Louw, R.; Mulder, P. J. Chem.
4,15
(
Soc., Perkin Trans. 2 1990, 937.
11) Ko, T.; Adusei, G. Y.; Fontijn, A. J. Phys. Chem. 1991, 95, 8745.
(
(12) Liu, L.; Davis, S. R. J. Phys. Chem. 1992, 96, 9719.
1
0.1021/ja984380o CCC: $18.00 © 1999 American Chemical Society
Published on Web 05/05/1999

4272 J. Am. Chem. Soc., Vol. 121, No. 17, 1999
Parker and DaVis
and BLYP, were used to treat electron correlation using the 6-311G-
(
Discussion.) Initially, both products grow in at approximately
the same rate. However, at long photolysis times the rate of
decay of ozone is small and the kinetic behavior of these two
products is quite different. The ketene product is unstable toward
the photolyzing radiation while the phenol band maintains a
positive slope during the course of photolysis. This difference
in kinetic behavior allowed us to assign the bands in Table 1.
The third species produced in this reaction is benzene oxide.
We were able to identify this species by comparison of the
product bands with the gas-phase FTIR spectrum.24 Only the
three most intense bands were observed (see Table 1). All three
bands exhibit the same growth kinetics: maximum intensity at
t ) 10 min, followed by rapid exponential decay. This
observation is in agreement with benzene oxide’s known
photoinstability: upon irradiation with photons of 320 nm < λ
d,p) and 6-311G(3df,3pd) basis sets,16-19 respectively. Analytic
gradients were used in the geometry optimizations with all electrons
correlated. All stationary points were verified to be true minima by
calculation of normal vibrational modes using the analytic second
derivative method in the harmonic approximation. B3LYP/6-311G-
2
0,21
(
d,p) vibrational modes were scaled
point calculations were carried out on the optimized B3LYP/6-
11G(d,p) geometries using the 6-311G(3df,3pd) basis set. Zero-point
by the factor 0.96325. Single-
3
energies were added to the electronic potential energies of each species
to obtain their relative energies at 0 K.
Results
Our infrared spectra of matrix-isolated ozone22 and benzene23
are in good agreement with literature data for the separate
species. The matrix spectra of co-deposited benzene and ozone
in argon revealed no additional bands.
<
480 nm, benzene oxide isomerizes to phenol and a ketene.25
Figure 3B shows growth kinetics of product bands which
After 30 min of irradiation at λ g 580 nm no new features
appeared. Irradiation at λ g 360 nm revealed several new bands
cannot be assigned to phenol on the basis of their absorption
frequencies, nor to the ketene on the basis of their kinetic
-1
in the 2100 cm region of the spectrum (Table 1). Other bands
were also observed and their locations and assignments are given
in Table 1. This table lists all product band absorptions and
relative intensities resulting from λ g 280 nm photolysis for
the four isotopomeric reactions. In each reaction, the strongest
-1
behavior. The band at 1728 cm is most likely due to a carbonyl
stretch. The growth kinetics of this band suggest that it is a
primary product which is unstable toward the photolyzing
-1
-1
radiation. The bands at 1683, 1202, and 903 cm (the 903 cm
band is not plotted in Figure 3 for clarity) have the same growth
kinetics and are most likely due to the same species. Moreover,
this species seems to be a secondary product as its concentration
increases greatly at long photolysis times. Finally, there is a
-
1
product band falls in the 2100 cm region. Figure 1 (parts
-
1
A-D) shows spectra recorded in the 2070 to 2150 cm range
after a total deposition of about 24 mmol of the gas mixtures.
The only bands present in two of the spectra, before photolysis,
-1
second unknown species at 774 cm with growth kinetics that
-
1
near 2110 cm are due to the ozone precursor. In the spectra
where the 18O isotope was used the ozone bands have moved
out of this region and the product bands have shifted to lower
wavenumber.
suggest it to be a primary product which is very unstable toward
the radiation.
Discussion
Gas-phase studies of the reaction of oxygen atoms with
benzene have concluded that phenol is the major product of
reaction. We find phenol as a product as well in the argon matrix
The primary processes in ozone photolysis26 in the 280-
11 nm range are as follows:
6
(
Table 1 and Figure 2). These bands are in good agreement
1
1
1
O ( A) f O ( ∆ ) + O( D)
λ e 310.0 nm
λ e 411.0 nm
(1)
3
2
g
with the infrared spectrum of phenol reported in the literature.
Figure 2 shows the infrared spectra of product bands (standard
3
-
g
1
O ( Σ ) + O( D)
isotopes) in the 715 to 840 cm-1 region.
2
(
2)
3)
Finally, we have identified three species produced in this
reaction besides phenol. Figure 3 shows the growth kinetics of
product bands when the matrix is subjected to radiation of λ g
1
+
g
3
O ( Σ ) + O( P)
λ e 463.0 nm
2
(
2
80 nm. In Figure 3A is plotted the growth kinetics of phenol
and a ketene product as well as the decay kinetics of ozone.
The product bands near 2120 cm-1 are due to a ketene, see
1
3
O ( ∆ ) + O( P) λ e 611.0 nm
(4)
2
g
(
When radiation is filtered through a 280 nm cutoff filter
process 1 should dominate because it is spin allowed. Therefore,
(
13) Gaussian 94, Revision E.3, Frisch, M. J.; Trucks, G. W.; Schlegel,
1
H. B.; Gill, P. M. W.; Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.;
Keith T.; Petersson, G. A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham,
M. A.; Zakrzewski, V. G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.;
Stefanov, B. B.; Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala,
P. Y.; Chen, W.; Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts,
R.; Martin, R. L.; Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart,
J. P.; Head-Gordon, M.; Gonzalez, C.; Pople, J. A. Gaussian, Inc.:
Pittsburgh, PA, 1995.
most of the oxygen atoms produced initially will be D. Argon
27
1
is known to be an effective quencher of O( D), therefore many
3
of these atoms will go on to react as O( P). The minimum
translational energy of the oxygen atoms formed in processes
27
1 to 4 can be calculated from spectroscopic and thermochemi-
28
1
cal data. The minimum translational energy of the O ( D) atom
formed in process 1 is 39.4 kcal/mol and in process 2 is 39.3
(
(
(
14) Becke, A. D. Phys. ReV. A 1988, 38, 3098.
3
15) Lee, C.; Yang, W.; Parr, R. G. Phys ReV. B 1988, 37, 785.
kcal/mol. O( P) is produced in process 3 with near zero
16) Dithcfield, R.; Hehre, W. J.; Pople, J. A. J. Chem. Phys. 1971, 54,
translational energy. Processes 2 and 3 are significant only when
the 360 nm filter was used. Process 4 was not significant under
our experimental conditions as we could not observe ozone
photolysis with radiation filtered at λ g 580 nm and an exposure
7
24.
(
(
(
17) Hariharan, P. C.; Pople, J. A. Theo. Chim. Acta 1973, 28, 213.
18) McLean, A. D.; Chandler, G. S. J. Chem. Phys. 1980, 72, 5639.
19) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys.
1
980, 72, 650.
20) This scaling factor was empirically derived in our laboratory by
(
(24) Klotz, B.; Barnes, I.; Becker, K. H.; Golding, B. T. J. Chem. Soc.,
Faraday Trans. 1997, 93, 1507.
(25) Jerina, D. M.; Witkop, B.; McIntosh, C. L.; Chapman, O. L. J. Am.
Chem. Soc. 1974, 96, 5578.
recording the infrared spectrum of dimethylketene in solid argon and
dividing the asymmetric ketene stretching frequency by the B3LYP/6-311G-
d,p) infrared frequency of this mode.
(
(
(
(
21) Scott, A. P.; Radom, L. J. Phys. Chem. 1996, 100, 16502.
22) Andrews, L.; Spiker, R. C., Jr. J. Phys. Chem. 1972, 76, 3208.
23) Brown, K. G.; Person, W. B. Spectrochim. Acta, Part A 1978, 34A,
(26) DeMore, W. B.; Raper, O. F. J. Chem. Phys. 1966, 44, 1780.
(27) Okabe, H. Photochemistry of Small Molecules; John Wiley and
Sons: New York, 1978.
1
17.
(28) Wagman, D. D. J. Phys. Chem. Ref. Data 1982, 11, Suppl. 2, 1.

Photochemical Reaction of Ozone and Benzene
Table 1. Absorptions (cm-1) Observed in the Reaction of Oxygen Atoms with Benzene at λ g 280 nm
6
/benzene-H /Ar /benzene-H /Ar /benzene-D /Ar /benzene-D /Ar
16O 18O 16O 18O
J. Am. Chem. Soc., Vol. 121, No. 17, 1999 4273
3
6
3
6
3
6
3
O
3
58% 50% 53% 69%
rel
O
3
rel
O
3
rel
O
3
rel
photolyzed photolyzed
intensity
intensity
photolyzed
intensity
photolyzed
intensity
assignment
3
633.2
10.1
3620.2
10.4
2682.6
6.7
2666.0
6.8 phenol
2
665.1
a
a
a
a
2
2
2
2
133.4
124.7
121.4
116.3
2105.3
2097.5
2096.8
2094.5
2129.5
2125.8
2121.2
2118.7
2117.2
2102.6
2099.8
2098.0
2094.6
2091.7
1
00
100
100
}
}
}
}
1
00
BDK
2093.4
2
2
2
090.4
089.6
084.9
}
1
728.6
1.6
1694.8
2.4
1.2
1718.6
1586.1
2.5
2.5
1674.1
0.5 2,4-cyclohexadienone
0.3 unknown 1
1
1
693.4
693.8-min
1
1
683.2
679.0
0.9
2.1
1672.0
1650.7
1
623.7
1586.5
1585.8
1
583.5
4
.2 phenol
1
1
585.0
583.6
}
}
}
1
610.4
6.0
18.2
0.5
1609.7
9.0
1573.5
1571.0
1569.2
1567.7
1565.3
1
1
1
572.3
570.4
569.3
7
.6
11.2 phenol
a
a
1
1
603.8
601.9
1603.5
1601.1
22.7
1399.5
1398.1
1399.8
1397.4
1394.5
26.9
2
4.3 phenol
1
397.1
1
392.4
1
544.7
?
a
1
1
504.6
501.6
1504.5
1500.1
1356.5
0.4
2.1
1354.3
1.1 phenol
a
12.6
6.3
11.9
5.5
1
1
472.8
471.2
1471.5
1469.9
1294.4
1293.0
1294.0
1292.7
1.9 phenol
}
1
290.8
1
344.3a
328.5
5.8
1.4
1343.0
5.9
1.2
1270.7
1191.1
4.8
1268.1
17.9 phenol
11.2 phenol
1341.0
1
1327.6
16.3
1176.6
1173.3
1
183.1
1
1
268.1
256.9
5.0
1.9
?
?
1
1
262.7
255.9
1250.8
1249.1
1101.0
3.2
1.5
phenol
16.2
6.7
}
1248.4
1
248.1
0.2
2.4
benzene oxide
unknown 1
1
1
202.2
199.7
177.0a
1200.0
1173.0
2.7
13.6
1.3
1
20.2
2.4
1062.0
1060.3
1058.9
1057.7
2.0 phenol
1170.1
1
1
170.4
168.8
1168.3
phenol
phenol
1
1
160.1
158.7
1.4
1158.8
1157.5
2.1
960.6
959.6
0.3
a
1
1
152.8
8.7
1150.5
1148.4
12.6
896.1
895.3
a
149.9
904.5
16.5
14.5 phenol
}
8
94.5
a
1
1
081.4
1.6
1.4
benzene oxide
a
079.8
a
1
074.0
1073.8
1070.7
1.6
1.9
878.0
877.2
2.9
1.9
866.6
754.6
0.9 phenol
?
a
1
071.1
1
1
027.2
025.5
1
000.6
0.9
755.4
1.7 phenol
7
54.7

4274 J. Am. Chem. Soc., Vol. 121, No. 17, 1999
Parker and DaVis
Table 1. continued
16O
/benzene-H
58%
photolyzed
/Ar
18O
/benzene-H
50%
photolyzed
/Ar
16O
/benzene-D
53%
photolyzed
18O
/benzene-D
69%
photolyzed
3
6
3
6
3
6
/Ar
3
6
/Ar
O
3
rel
intensity
O
3
rel
intensity
O
3
rel
intensity
O
3
rel
intensity
assignment
9
9
87.9
42.3
0.9
1.6
?
?
942.2
1.0
940.0
9
8
8
02.7
99.9
98.3
904.9
903.1
900.0
2.8
1.8
}
3.3
1.9
unknown 1
phenol
}
898.3
8
8
82.4
80.7
882.2
880.5
752.0
750.6
2.6
745.7
744.5
2.6
0.5
8
26.0
1.4
3.8
624.6
586.9
0.8
0.4
624.4
phenol
phenol
a
8
8
17.3
11.8
804.0
802.0
4.1
a
7
7
7
84.0
73.7
52.3
1.0
1.7
783.2
769.7
752.7
0.5
1.1
?
unknown 2
phenol
15.6
13.9
551.5
550.4
6.1
1.2
551.2
550.2
5.6
751.9
7
7
38.2a
24.2
4.2
1.0
5.1
737.7
724.2
2.3
0.9
7.6
benzene oxide
BDK
a
6
6
88.9
87.4
689.2
687.2
504.0
phenol
a
5
28.6
04.9a
0.4
2.4
phenol
5
503.8
1.8
1.2
0.3
0.5
phenol
5
4
4
02.5
94.0
85.3
?
?
?
a
a indicates band was observed at λ > 360 nm and t ) 30 min. The 360 nm filter was used only in the reaction of standard isotopes.
time of 30 min. A recent study29 of the photochemistry of ozone
in argon matrices suggests that once an O( D) atom escapes
1B.) Since there is reasonable agreement with the reduced mass
ratio in the diatomic approximation, we conclude that the
measured shifts are consistent with the ketene functional group.
In order for a molecule with a ketene functional group to
form from the reaction of benzene and oxygen atoms, the ben-
zene ring must open. Since no evidence for smaller fragments
is present in the spectra, we will only consider ketene molecules
with the empirical formula C6H6O as possible products. Al-
though the reaction is exothermic, and consequently the initial
product should be formed in an excited vibrational state, con-
densed phases favor collisional deactivation over fragmentation.
Two possible structures of the ketene product are
1
the matrix cage where it is formed it moves easily through the
argon lattice because of a short-range attractive interaction
1
3
between argon and O( D). The O( P) atom experiences a long-
range repulsive interaction with the argon lattice. Molecules such
as benzene, however, are unable to diffuse through an argon
1
matrix at 12 K. In this study we have assumed that O( D) can
diffuse through the matrix (and to a lesser degree triplet atomic
oxygen) before reacting with benzene.
Referring to Table 1 and Figure 1, we find that the most
-
1
intense product bands occur in the 2100 cm region of the
spectrum. Considering the reaction of the normal isotopomers,
four intense product bands at 2133.4, 2124.7 and 2121.4, and
-
1
2
116.3 cm are easily seen in Figure 1A. Upon substitution
18
-1
with O these bands shift to 2105.3, 2096.8, and 2094.5 cm .
In compounds of the elements C, H, and O, the only functional
30
group with an intense infrared absorption in this spectral region
is the ketene group, R2CdCdO. For a given infrared mode,
the square of the frequency ratio gives the reduced mass ratio
in the diatomic approximation. The asymmetric stretch of a
ketene group can be modeled as a vibration of the central carbon
atom against stationary oxygen and carbon atoms on either side;
therefore, µ18/µ16 ) 1.020. Taking the square of the ratio (ν16/
ν18), for the bands listed above, yields the following: (2133.4/
We can readily decide between these two structures on the basis
of the isotopic shifts in their infrared spectra. Referring to Table
-1
1
and Figure 1A, there is an intense ketene band at 2124.7 cm .
Upon substituting all hydrogen atoms with deuterium (Figure
-1
1
C) this band shifts to 2118.7 cm , a shift of -6.0 wavenum-
2
2
2
2
105.3) ) 1.027, (2124.7/2096.8) ) 1.027, and (2121.4/
bers. Table 2 lists the B3LYP/6-311G(d,p) calculated frequen-
2
-1
094.5) ) 1.026. (The 2116.3 cm mode is rather poorly
cies for structures 1 and 2. For structure 1, the perhydrogen
defined and we were unable to assign it to a band in Figure
-1
compound has an asymmetric stretch at 2129.0 cm while the
perdeuterium compound has the corresponding mode at 2123.5
(
(
29) Benderskii, A. V.; Wight, C. A. J. Chem. Phys. 1994, 101, 292.
30) Daimay, L.-V. The Handbook of Infrared and Raman Characteristic
-
1
cm . The deuterium isotope has lowered the frequency of this
-
1
Frequencies of Organic Molecules; Academic Press: Boston, 1991.
mode by 5.5 cm . This calculation is in good agreement with

Photochemical Reaction of Ozone and Benzene
J. Am. Chem. Soc., Vol. 121, No. 17, 1999 4275
Figure 1. Infrared spectra of product bands from 280 nm photolysis in the 2070-2150 cm region for the four isotopomeric reactions: (A) 16O,
-
1
1
18
1
16
2
18
2
H, (B) O, H, (C) O, H, and (D) O, H (a) spectra before photolysis and after (b) 10 min, (c) 30 min, (d) 60 min, and (e) 120 min of Hg arc
photolysis.
-1
Figure 2. Infrared spectra in the 715-840 cm range for the reaction
16
of O and benzene-H
6
: (a) spectrum before photolysis and (b) 10 min,
(c) 30 min, (d) 60 min, and (e) 120 min after photolysis.
the experimental observation. In structure 2, the perhydrogen
compound exhibits the ketene asymmetric stretch at 2129.2
-
1
cm . Substituting all hydrogens with deuterium we find that
-
1
the mode splits, and shifts to 2133.0 and 2096.5 cm ,
-
1
corresponding to frequency changes of +3.8 and -36.5 cm .
The reason for the splitting of this mode is that the asymmetric
ketene stretch has become strongly coupled with a C-D
symmetric stretching mode in this isotopomer, causing the band
to split and shift to higher and lower frequencies. However,
this behavior is not observed experimentally, and consequently
structure 1 is assigned as the structure of the ketene product.
For the rest of the paper we will refer to 1 as butadienylketene
Figure 3. Photolysis time dependence of product band intensities from
the 16O /benzene-H photolysis reaction: (A) phenol and BDK product
3
6
(BDK). (The IUPAC name for BDK is hexa-1,3,5-trien-1-one).
bands and (B) 1,3-cyclohexadienone and unidentified product bands.

4276 J. Am. Chem. Soc., Vol. 121, No. 17, 1999
Parker and DaVis
Table 2. B3LYP/6-311G(d,p) Frequencies (cm-1) of Ketene
Asymmetric Stretch for Structures 1 and 2
Scheme 1
freq
1
2
16
1
12 16
2
structure
( C, O, H)
( C, O, H)
shift
1
2
2129.0
2129.2
2123.5
2133.0
-5.5
+3.8
-32.7
2096.5
Table 3. B3LYP/6-311G(3df,3pd) // B3LYP/6-311G(d,p) Relative
Energies and Frequencies of Products from the Benzene/Oxygen
Atom Reaction
species
E)-BDK-t,t
rel energy (kcal/mol) asym str freq (cm-1)
(
(
(
(
(
(
(
(
-65.4
-63.7
-61.7
-60.0
-64.0
-62.0
-60.5
-58.0
-81.9
-100.8
0
2129.0
2122.9
2127.9
2123.2
2129.5
2117.1
2128.4
2120.4
NA
E)-BDK-c,t
E)-BDK-t,c
E)-BDK-c,c
Z)-BDK-t,t
Z)-BDK-c,t
Z)-BDK-t,c
Z)-BDK-c,c
with the letters t or c. The first letter refers to the orientation of
the C-C single bond adjacent to the ketene group; the second
letter refers to the remaining C-C single bond. In Table 3, the
average relative energy of the BDK isomers (as compared to
2
,4-cyclohexadienone
phenol
benzene + O( P)
NA
NA
3
3
the energy of the unreacted benzene and oxygen P atom) is
To our knowledge, butadienylketene has never been reported
as a product of the benzene/oxygen atom reaction. However,
the reaction has precedent. Capponi et al.31 report that photo-
lytically generated 2,4-cyclohexadienone (CHD) in water un-
dergoes secondary photolysis to form butadienylketene. This
observation provides insight into the mechanism of the reaction.
Our postulated mechanism is given as
-61.9 kcal/mol with a standard deviation of (2.4 kcal/mol.
Because the relative energies of the isomers are closely grouped,
it is not unreasonable to assume that most or all of these isomers
have been trapped in the matrix reaction. There are four
discernible absorptions in the ketene region for products with
standard isotopes (Table 1). It is likely that each of these four
absorptions is due to one or more of the conformational isomers
of butadienylketene. The range of deviation among the calcu-
lated frequencies is 12.4 wavenumbers. Experimentally, we find
18
a range of 17.1 wavenumbers. For products with the O and
2
H isotopes, eight absorptions are observed in the ketene region.
These observations lend support to the possibility that all eight
conformational isomers of butadienylketene are trapped in the
matrix.
To investigate the possibility that the multiple ketene absorp-
tions are due to the same species trapped in different lattice
sites, each matrix was annealed to 27 K for 1 min and then an
infrared spectrum was taken. There was no change in the peak
positions, supporting each absorption being due to a unique
conformation.
The first step in the mechanism is the reaction with triplet state
oxygen atoms to form a benzene-oxygen biradical complex.
Initially, this complex must be formed in the triplet state to
conserve spin. This structure rapidly decays to cyclohexadienone
by hydrogen atom migration and intersystem crossing. The
cyclohexadineone now absorbs a photon and the carbonylcar-
bon-methylene carbon single bond cleaves homolytically to
form butadienylketene. This mechanism is further supported by
the data in Figure 3 which show that cyclohexadienone is
Further examination of Table 3 reveals that the relative energy
of cyclohexadienone is approximately midway between the
average relative energies of the butadienylketene isomers and
phenol. According to our postulated mechanism, the BDK
product forms from photolysis of the intermediate cycohexa-
dienone. We have observed a product absorption in the carbonyl
-
1
region of the spectrum at 1728.6 cm . The reduced mass ratio
25
of the carbonyl functional group is µ /µ ) 1.050 in the
unstable toward the photolyzing radiation. Jerina and McIntosh
18 16
diatomic approximation. From Table 1 the ratio of the squares
have postulated a similar mechanism for the photorearrangement
1
6
1
18
1
2
of benzene oxide in solution at 77 K. Klotz24 and co-workers
of the O, H and O, H modes is (1728.6/1693.8) ) 1.042.
1
6
2
18
2
Also, the ratio of the squares of the O, H and O, H modes
is (1718.6/1674.1) ) 1.054. Since there is reasonable agreement
have also referred to this mechanism in their investigation of
the gas-phase photorearrangement of benzene oxide.
2
of the observed reduced mass ratio with the approximated ratio,
and because this band falls within the region of the spectrum
characteristic of carbonyls, we assign this band to 2,4-cyclo-
hexadienone.
There are eight possible conformational isomers of butadi-
enylketene. Their computed relative energies and ketene asym-
metric stretching frequencies are given in Table 3. The E/Z label
has its usual meaning and refers to the orientation of the
molecule about the central C-C double bond. The orientation
of each C-C single bond is designated as either trans or cis
A summary of the matrix reactions is given in Scheme 1. It
is noted that the possibility exists that some of the phenol
1
product may have been formed by insertion of O( D) into a
C-H bond of benzene. The branching ratio, [phenol]/[BDK],
can be determined from Beer’s law:
(
31) Capponi, M.; Gut, I.; Wirz, J. Angew. Chem., Int. Ed. Engl. 1986,
2
5, 344.

Photochemical Reaction of Ozone and Benzene
C]b ) A/I
J. Am. Chem. Soc., Vol. 121, No. 17, 1999 4277
[
where A is the integrated absorbance, I is the band intensity,
and b is the path length. The experimentally determined intensity
-
1
of the phenol ν19 mode (1504.6 and 1501.6 cm in argon
matrix) is given as 36 km/mol by Brownlee32 and co-workers.
Our calculated BLYP/6-311G(3df,3pd) intensity of the asym-
metric stretching mode of butadienylketene is 1221 km/mol.
3
3
34
Porezag and Nascimento have reported that this density
functional method allows for accurate calculation of infrared
intensities. Indeed, we find a calculated intensity of 35.70 km/
mol for the phenol ν19 mode at this level; the resulting error in
the calculation is only 0.8%.
Figure 4 plots the phenol/BDK branching ratio as a function
of photolysis time. The value of the ratio lies in the range 4.1
to 4.5 from t ) 10 to 440 min. During this period the average
branching ratio is 4.3. Photolysis at times greater than 440 min
causes the ratio to increase in a seemingly linear fashion. This
can be explained by the observation that phenol is stable toward
photolyzing radiation of λ g 280 nm while BDK is not (Figure
Figure 4. Plot of the phenol/BDK concentration ratio as a function of
photolysis time.
2
,4-cyclohexadienone, benzene oxide, and butadienylketene.
Butadienylketene is formed by photolysis of cyclohexadienone.
We find an average branching ratio for the formation of phenol
and BDK of 4.3. The average relative energy of a BDK molecule
is calculated to be -61.9 ( 2.4 kcal/mol. The discovery of this
photochemical reaction has possible implications in the atmo-
spheric chemistry of benzene.
3
). The photolysis of BDK must lead to the formation of a
product with secondary growth kinetics. We now postulate that
unknown 1 is the product of the photolytic decay of BDK.
Conclusions
Acknowledgment. Partial support from the National Science
Foundation through NSF EPSCOR (OSR-9452857), NSF (CHE-
512473) as an equipment grant, and the Office of Naval
Research (N00614-93-1-0079) is gratefully acknowledged. We
also thank the Mississippi Center for Supercomputing Research
for computer time.
The photochemical reaction of benzene and ozone in a solid
argon matrix at 12 K yields the following products: phenol,
9
(
32) Brownlee, R. T. C.; Cameron, D. G.; Topsom R. D.; Katritzky, A.
R.; Sparrow, A. J. J. Mol. Struct. 1973, 16, 365.
33) Porezag, D.; Pederson, M. R. Proc. Sci. Technol. At. Eng. Mater.
996, 483.
34) Nascimento, I. C.; DeOliveira, A. E.; Bruns, R. E. Spectrochim.
Acta, Part A 1998, 54A(6), 831.
(
1
(
JA984380O