J. C. Wenger et al.
mate, [HC(O)OCH2OC(O)H]. Although two of these products,
(2-oxoethoxy)methyl formate and methylene glycol diformate,
possess very similar FTIR spectra, as indicated by spectra d and
e in Figure 1, there are distinguishable differences in the ab-
sorption bands in the 1600–1800 cmꢁ1 and 1000–1200 cmꢁ1 re-
gions to enable satisfactory subtraction of the spectra. One of
the expected products, 1,3-dioxan-2-one, was not detected in
any of the FTIR spectra and it should be noted that in all ex-
periments, the residual spectrum, arising from subtraction of
all known reactants and products contained very few absorp-
tion features, thus indicating that all major products had been
accounted for.
sponse at longer reaction times, thereby indicating that it is
most likely a secondary product.
3. Discussion
The observed product distributions can be used to elucidate
the importance of various possible reaction pathways for the
photooxidation of 1,3-dioxane. Hydroxyl radicals or Cl atoms
can abstract a hydrogen atom from one of three possible reac-
tion sites in 1,3-dioxane. Three alkyl radicals may be produced
which react with O2 to produce the corresponding peroxy radi-
cals and, in the presence of NOx, react to form the correspond-
ing alkoxy radicals. The possible reaction pathways for these
three alkoxy radicals (I, II and III) in the presence of NOx are
shown in the reaction schemes given in Figures 4–6. These re-
action schemes are based on the current knowledge on the re-
activity of alkoxy radicals[14] and, in particular, on the observed
reactivity of cyclic alkoxy radicals produced during the atmos-
pheric oxidation of cyclic ethers.[2,5–9]
Representative concentration-time and product yield plots
are shown in Figures 2 and 3 respectively. The concentration of
HC(O)OCH2OC(O)H was determined using the FTIR product
Previous studies of the atmospheric oxidation of ethers indi-
cate that the reaction mainly proceeds by H-atom abstraction
from a carbon atom in the a-position to the O atoms.[2] The re-
action scheme for this pathway is shown in Figure 4. Hydrogen
atom abstraction from this site produces alkoxy radical I,
which can react with O2 or decompose via cleavage of the
CꢁC or CꢁO bond. Fission of the CꢁO bond is much less ener-
getically favourable than cleavage of the CꢁC bond, and is un-
likely to be important for this radical.[2] The expected product
of the reaction with O2 is 1,3-dioxan-4-one. However, there was
no evidence for the formation of this species in the FTIR spec-
tra, even at high partial pressures of O2, indicating that, like
the principal alkoxy radical produced during the photooxida-
tion of 1,4-dioxane,[5,6] cleavage of the CꢁC bond appears to
be the sole reaction pathway. The resulting ring-opened radical
is oxidised to the alkoxy radical Ia which can react with O2 to
form the major product, (2-oxoethoxy)methyl formate, under-
go 1,5-H shift isomerisation and further oxidation to produce
radical Ib, or decompose via CꢁC bond scission. The latter re-
action yields the minor product methylene glycol diformate
and formaldehyde. By analogy with the radical chemistry ob-
served in the photooxidation of cyclohexane,[15] radical Ib may
undergo further isomerisation to produce hydroxy(2-oxoethox-
y)methyl formate, though, the lack of an absorption band in
the O-H stretching region at ca. 3600 cmꢁ1 indicates that this
pathway is of minor importance. Alternatively, because it is of
the general structure RC(O)OCH(O·)R’, radical Ib may also un-
dergo an a-ester rearrangement to produce RC(O)OH and the
carbonyl radical R’(C·)O.[14,16] As shown in Figure 4, this reaction
pathway would produce formic acid, CO2 and glycolaldehyde.
The a-ester rearrangement of radical Ib is therefore a possible
source of the formic acid detected in these experiments. Al-
though the co-product glycolaldehyde was not detected in the
FTIR spectra, this compound is reactive towards both OH and
Cl and may also be photolysed under the conditions employed
in the experiments to produce formaldehyde and CO.[17,18]
Formaldehyde was observed in most experiments but only at
levels below the quantification limit - this compound is also
Figure 2. Concentration-time profile for the Cl atom initiated oxidation of
1,3-dioxane (124 ppmV) in purified air at 760ꢀ10 Torr and 298ꢀ2 K.
spectrum generated from the Cl atom initiated oxidation of
1,3,5-trioxane in the presence of NOx and assuming a 100%
yield.[8] The yield plots for HC(O)OCH2OCH2C(O)H, HC(O)-
OCH2OC(O)H and HCOOH are linear indicating that these com-
pounds are primary products of the reaction and are not re-
moved to any significant extent during the timescale of the ex-
periments. Their yields were determined from a linear least
squares fit of the data for reactant conversions <50%. As indi-
cated in Figure 3, the formation of CO shows a non-linear re-
Figure 3. Product yield plot for the Cl atom initiated oxidation of 1,3-diox-
ane (124 ppmV) in purified air at 760ꢀ10 Torr and 298ꢀ2 K.
3982
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2010, 11, 3980 – 3986