For a temperature of 297 ^ 2 K, a rate constant k \ 2 ^ 1
were carried out: (i) at a total pressure of 100 mbar at room
temperature (295 ^ 2 K, Ñow tube: 1.8 cm i.d. quartz glass)
and (ii) at a total pressure of 500 mbar in the temperature
range of 276È353 K using a Ñow tube surrounded with a
thermo-jacket (1.8 and 4.0 cm i.d. quartz glass). A schematic of
the experimental set-up is given in Fig. 1.
3
]
10~15 cm3 molecule~1 s~1 was estimated and the equi-
librium constant K \ k /k
is given as K \ 2.7 ^ 0.4
3
3 ~3
3
]
10~19 cm3 molecule~1.7 Also as a result of the recently
published study,7 the rate constants for the overall consump-
tion of OH/benzene adduct radicals by O , k
\ k
2
overall
4
]
K k , was determined to be 2.1 ^ 0.2 ] 10~16 cm3
For on-line FT-IR analysis, the main part of the reaction
gas was pumped continuously through a 2050 cm3 cell with a
White cell mirror system (optical path length 10 m). FT-IR
spectra were recorded with an instrumental resolution of 8
cm~1 using a MCT detector by coadding of 200È3000 scans
(Nicolet Magna 750). Benzene, phenol, nitrobenzene and
formic acid were identiÐed by characteristic band locations
and the concentrations were determined using calibrated
spectra of authentic samples. The three isomers of hexa-2,4-
dienedial and p-benzoquinone are solids. Because of their low
sublimation vapour pressure, a determination of the absorp-
tion cross section under the experimental conditions used here
was impossible. Reference spectra of these compounds were
recorded by Ñushing out the gas phase over a solid sample by
the carrier gas. For the cross section of the carbonyl absorp-
tion of these species the value for the structurally similar cis-
butenedial, obtained from this laboratory, of 7.9 ^ 0.4
3
5
molecule~1 s~1. This value is in line with former published
data from the same group.6
Nevertheless, the Ðrst stable products of the overall reaction
of O with the OH/benzene adduct radical are still uncertain.
2
Klotz et al.8 proposed the formation of benzene oxide/oxepin
as the primary product of the reaction of O with the OH/
2
benzene adduct radical, contrary to a theoretical prediction.9
The identiÐcation of benzene oxide/oxepin was the subject of
a former investigation from this laboratory.10 As a result of
this experimental study, performed under Ñow-tube conditions
without NO additions in the pressure range 13È100 mbar,
x
[
O ] \ 1È16 ] 1017 molecule cm~3, the formation of benzene
2
oxide/oxepin was found to be negligible, yield: \0.03 (FT-IR
analysis) and \0.01 (GC-MS analysis).10
]
10~19 cm2 molecule~1 (base 10) was used in each case.15
(
6)
For cis,trans- and trans,trans-hexa-2,4-dienedial, approx-
imated cross sections for the carbonyl absorption measured in
a heatable cell at 355È385 K with an instrumental resolution
of 1 cm~1 are reported in the literature to be 2.4 ] 10~18 cm2
molecule~1 each (base e).16 This value is approximately 30%
higher compared to that of cis-butenedial, representing a
rough estimate for the uncertainty of the concentrations of the
hexa-2,4-dienedials and p-benzoquinone in the present study.
To estimate the concentration of unidentiÐed carbonylic sub-
stances, an averaged integrated band intensity (IBI) for the
carbonyl absorption (1690È1780 cm~1) of 7.9 ] 10~18 cm
molecule~1 was taken, derived from methyl ethyl ketone
(7.8 ^ 0.3 ] 10~18 cm molecule~1), methyl vinyl ketone
(7.4 ^ 0.1 ] 10~18 cm molecule~1) and methacrolein
(8.4 ^ 0.1 ] 10~18 cm molecule~1). All product yields were
determined with reference to reacted benzene, yield \
[product]/*[C H ].
Furthermore, a small gas stream was pumped continuously
through a GC loop for MS analysis (HP 5890 with HP-MSD
5971) using an additional pump system. The heated sample
loop was directly connected with the outlet of the Ñow tube by
a heated transfer line. A 30 m, 0.25 mm i.d. column (HP 5MS)
for standard analysis and a 25 m, 0.32 mm i.d. column
(Chrompack Poraplot Q) for highly volatile compounds were
chosen for product separation.
IdentiÐed products from the reaction of OH radicals with
benzene resulting from chamber studies are glyoxal,11
nitrobenzene12 and phenol.12h14 The formation yields for
phenol are quite di†erent, 0.236 ^ 0.044,12 0.26 ^ 0.0513 and
0.50 ^ 0.05,14 indicating the inÑuence of experimental condi-
tions on the product yields. As a result of a Ñow-tube study
without NO additions at 100 mbar, phenol was found with a
x
yield of 0.23 ^ 0.0710 to be in line with the lower yields from
the chamber experiments. Besides CO, formic acid and p-
benzoquinone, unidentiÐed carbonylic substances (yield:
0
.59 ^ 0.11) represented the main product fraction under the
Ñow conditions used.10
Generally, knowledge concerning the main pathways
occurring in the OH radical initiated oxidation of benzene in
the atmosphere and the resulting product distributions is
sparse. The subject of this work is to determine the Ðrst stable
products from the reaction of OH radicals with benzene under
di†erent experimental conditions. To come close to atmo-
spheric conditions, the pressure range of the Ñow experiments
6
6
was extended to 500 mbar, allowing O concentrations of
2
8
] 1018 molecule cm~3. In a recently published kinetic study
of the reaction of O with the OH/benzene adduct7 a distinct
temperature dependence of the product distribution in the
2
In addition, for identiÐcation of produced carbonylic sub-
stances the derivatisation technique using PFBHA17 (O-(2,3,4,
5,6-pentaÑuorobenzyl)hydroxylamine hydrochloride) with
subsequent GC-MS analysis (HP 6890 with HP-MSD 5973)
was carried out. For this purpose, the reaction gas was
pumped through a cooled trap maintained at liquid argon
temperature. After disconnecting the trap from the Ñow tube,
range of atmospheric temperatures was discussed. The tem-
perature dependent measurements in the range of 276È353 K
within the present study should help to clarify this fact.
Furthermore, the subject of a second part was a product
study concerning the reaction of the OH/benzene adduct
radical with O . Because the reaction of the OH/benzene
3
adduct radical with O is relatively slow, at least under highly
2
polluted conditions the reaction of O can be in competition
3
with the O reaction. To the authorsÏ knowledge there are no
2
experimental studies in the literature concerning the kinetics
as well as the product distribution for the reaction of the OH/
benzene adduct radical with O .
3
Experimental
The gas-phase reaction of OH radicals with benzene was
investigated in di†erent He/O mixtures under low-Ñow con-
2
ditions using on-line FT-IR spectroscopy and GC-MS mea-
surements as analytical techniques. Two series of experiments
Fig. 1 Schematic of the experimental set-up.
Phys. Chem. Chem. Phys., 2001, 3, 4946È4956
4947