Atmospheric fate of alkoxy radicals: Branching ratio of evolution pathways for 1-propoxy, 2-propoxy, 2-butoxy and 3-pentoxy radicals

Article abstract of DOI:10.1039/b307761f

As the last step of VOC oxidation in the atmosphere, the evolution of alkoxy radicals determines the nature and the concentration of the secondary compounds formed. Branching ratios between decomposition and reaction with O₂ of 1-propoxy, 2-propoxy, 2-butoxy, and 3-pentoxy radicals were measured at room temperature and 1 atm in a simulation chamber using FTIR spectroscopy as an analytical device. The ratio varied depending on the leaving alkyl group and the class of alkoxy. No additional decomposition due to excited radicals was observed. The results could be used directly for tropospheric simulation purposes. Formaldehyde might be a photolytic source of HO_x through the production of H and HCO radicals and acetaldehyde is the key precursor of the toxic NO_x reservoir, peroxy-acetyl nitrate. In the lower troposphere, 1-propoxy and 2-propoxy radicals react mainly with O₂ while decomposition is an important reaction pathway for 2-butoxy and 3-pentoxy. Consequently, C₁ and C₂ aldehyde production from the two longer chain alkoxys will occur very close to the area of initial VOC oxidation, while for the alkoxys exhibiting a minor decomposition pathway, the formaldehyde or acetaldehyde production will take place after oxidation of all the intermediate secondary compounds, far from the emission area of the primary compound.

Full text of DOI:10.1039/b307761f

View Article Online / Journal Homepage / Table of Contents for this issue
Atmospheric fate of alkoxy radicals: branching ratio of evolution
pathways for 1-propoxy, 2-propoxy, 2-butoxy and 3-pentoxy radicals
N. Meunier,* J. F. Doussin, E. Chevallier, R. Durand-Jolibois, B. Picquet-Varrault and P. Carlier
L.I.S.A., Universit e´ paris 7, paris 12 et CNRS (UMR7583), 61 avenue du g e´ n e´ ral de Gaulle,
F94010, Cr e´ teil cedex, France. E-mail: meunier@lisa.univ-paris12.fr
Received 8th July 2003, Accepted 17th September 2003
First published as an Advance Article on the web 2nd October 2003
As the last step of VOC oxidation in the atmosphere, the evolution of alkoxy radicals determines the nature and
the concentration of the secondary compounds formed. Branching ratios between decomposition and
reaction with O
temperature and atmospheric pressure in a simulation chamber using FTIR spectroscopy as an analytical
2
of 1-propoxy, 2-propoxy, 2-butoxy and 3-pentoxy radicals have been measured at room
1
6
16
18
18
device. kdec/k
O
is equal to (3.8 ꢀ 0.4) ꢁ 10 , (2.9 ꢀ 0.3) ꢁ 10 , (2.9 ꢀ 0.4) ꢁ 10 and (4.1 ꢀ 0.8) ꢁ 10
2
ꢂ3
molecules cm for each, respectively. This ratio has been observed to vary, depending on the leaving alkyl
group and the class of alkoxy. No additional decomposition due to excited radicals has been observed.
6
-membered cyclic transition state. The d-hydroxy alkyl
I Introduction
formed evolves to a d-hydroxy peroxy one. This in turn pro-
duces a d-hydroxy alkoxy radical with the same number of car-
bons as the first alkoxy radical. This d-hydroxy alkoxy radical
then reacts according to the three pathways described to
produce d-hydroxy-carbonyl compounds.
The three pathways give different products showing different
reactivity and toxicity. The relative importance of the kinetics
of each determines which products will be present in the atmo-
sphere and in what quantities. Consequently, knowledge of the
branching ratio is of immense importance.
The atmospheric photooxidation of Volatile Organic Com-
pounds (VOC) in the troposphere leads to the formation of
secondary compounds and ozone which may be harmful to
the environment. Their impact from the medical, ecological
and climatic perspective makes the investigation of these pro-
cesses an area of concern for society as a whole.
The general mechanism of VOC degradation is relatively
well known and has been the subject of some reviews. In this
process, the last step is the reaction of alkoxy radicals (RO ).
1
–3
ꢃ
The evolution of these radicals determines the nature and the
concentration of the secondary compounds formed, which
can greatly influence the chemistry of the atmosphere. As a
result, a good knowledge of alkoxy radical fate is a priority
for this research field.
Depending on the nature of the alkyl group, the alkoxy radi-
cals can undergo b-C–C bond cleavage (reaction 1), reaction
Even if this overall mechanism is well known, only a few
experiments have measured the branching ratio of the various
4
–9
10
pathways of alkoxy radical evolution.
A recent review
ꢂ15
3
cm molecule
ꢂ1 ꢂ1
s
for the reaction with O and proposes a simple structure activ-
confirms a value of k ¼ (8 ꢀ 2) ꢁ 10
2
2
1
1
ity relationship (SAR) for estimating the decomposition rate.
In alkoxy chemistry studies, the chemical system chosen to
generate the studied radical is a key parameter of the experi-
ment. Some systems (alkane + OH. . .) lead to alkoxy radicals
under realistic conditions (i.e. excitation level, secondary
reactions. . .) but are not very specific. Some others (alkyl
nitrite photolysis, ROOR photolysis. . .) produce only the stu-
died radical but may produce artefacts like excess internal
energy which can increase the dissociation rate. Indeed, the
generated chemistry must be simple and sufficiently well
known to allow good sensitivity for the analysis of the alkoxy
chemistry. All products have to be identified and their IR spec-
with O
-5-H shift isomerization via a 6-member cyclic transition state
reaction 3).
2
(reaction 2) and, for alkoxy radicals with a d-carbon,
1
(
0
ꢃ
ꢃ
0
RR CHO ! R þ R CHO
ð1Þ
ð2Þ
0
ꢃ
0
ꢃ
RR CHO þ O2 ! RR CO þ HO2
4
tra calibrated. Libuda et al. studied the kinetics of the 2-
butoxy radical generated by the photolysis of the 2-butoxy
iodide. The alkyl radical produced evolves to the alkoxy radi-
ð3Þ
The decomposition (reaction 1) produces a smaller carbonyl
compound and an alkyl radical. The alkyl radical can react
with O to produce a peroxy radical and this can produce
2
cal via the RO + NO reaction. The chemistry from alkyl to
alkoxy radical is tropospherically relevant. In contrast, the
chemistry of iodo compounds is not well known and may be
2
another smaller alkoxy which reacts according to the three
given pathways.
5
6
complex. Carter et al. and Cox et al. have studied the 2-
butoxy and the 1-butoxy radicals in environmental simulation
chambers. In these experiments where the chemistry of n-
butane was used as an alkoxy source, the generation of alkoxy
radicals was realistic but not very specific (both 1-butoxy and
2-butoxy radicals were generated). Some studies have gener-
ated alkoxy radicals by the photolysis of alkyl nitrites
The reaction with O
This produces one HO
and an OH radical via interconversion of NO and NO
2
(reaction 2) occurs via a-H abstraction.
radical which mainly produces ozone
. In this
2
2
case, a carbonyl compound with the original number of
carbons is formed.
The third pathway (reaction 3) can only occur with long
chain alkoxy radicals. It is a 1-5-H shift isomerization via a
1
2–15
(RONO)
but in some cases an additional decomposition
4
834
Phys. Chem. Chem. Phys., 2003, 5, 4834–4839
This journal is # The Owner Societies 2003
DOI: 10.1039/b307761f

View Article Online
4
,16
was encountered.
in alkoxy radicals, causing them to be less stable and to decom-
This is due to an excess of internal energy
1
pose rapidly. However, all of these studies used relatively
7
high energy radiation (l < 360 nm) sources. In view of Heick-
len’s review, it has been established that the photolysis of
1
6
alkyl nitrite at 254 nm produces mainly excited alkoxy radicals
(
> 80%) but at 360 nm, a maximum of 5% are excited.
Moreover, the theoretical chemistry has been studied espe-
1
1
cially by M e´ reau et al., who have established a structure reac-
tivity relationship for the decomposition pathway. These
relationships have been parameterised, using quantum chemis-
try ab initio BAC-MP4 and density functional theory calcu-
lated activation energies and reaction enthalpies on a set of
linear and branched C
processes. Other studies
1
to C
5
alkoxy radical decomposition
have measured the rate constant
of the decomposition of various alkoxy species.
1
8–20
2
1
Atkinson has also proposed a relationship to evaluate the
rate constants for alkoxy radical reactions for atmospheric
purposes. The literature data concerning the rates of these
reactions was evaluated, and predictive schemes proposed.
The main drawback of these results is that they are based on
too few experiments. The ratio between the rate constants of
the different alkoxy radicals are often reproduced but the abso-
lute value may be different from the atmospherically relevant
value. Consequently, recent absolute experimental studies have
Fig. 2 Spectrum of irradiation lamps.
that the cleavage to give NO and alkoxy radicals is by far the
dominant photolysis channel.
2
6
Alkyl nitrites have been prepared by adding a 50% solution
of H SO to a mixture of NaNO and the appropriate alco-
2
27
4
2
1
4,22
been carried out. Deng et al.
O of the 2-butoxy and the 3-pentoxy radical, while Fittschen
have studied the reaction with
hol, i.e. 1-propanol, 2-propanol, 2-butanol and 3-pentanol
to synthesise 1-propyl nitrite, 2-propyl nitrite, 2-butyl nitrite
and 3-pentyl nitrite, respectively.
2
1
et al. have studied this reaction with ethoxy, 2-propoxy and
5
the 1-propoxy radicals. Contrary to Atkinson’s recommenda-
Reaction 2 produces a HO radical which reacts with NO to
2
give an OH radical and NO . Therefore, in order to minimise
2
2
1
ꢂ15
3
tions
molecule
molecule
in which k
s
2
is taken as being 9.5 ꢁ 10
cm
cm
ꢂ1 ꢂ1
ꢂ1 ꢂ1
ꢂ15
3
for primary alkoxy radicals and 8 ꢁ 10
the chemistry due to the OH radical and in particular its reac-
tion with monitored products, an OH-scavenger was added to
the chamber when carrying out experiments. CO has been the
s
for secondary radicals, these studies show that
this rate constant cannot be assumed to be independent of
molecular size, even for radicals with very similar structure.
2
8
subject of many studies. From these, it is well known that
the products of its reaction with OH are CO and H radical.
2
2
Deng et al. further indicated the need for kinetic studies on
a much more diverse set of alkoxy radicals.
2
H reacts to form HO
CO cause problems for data analysis. The reaction rate of
2
and then OH. Neither the CO nor the
In this study, we have investigated the branching ratios of
the evolution of 1-propoxy, 2-propoxy, 2-butoxy and 3-pent-
oxy radicals (Fig. 1) by measuring the formation of specific
products from each pathway in a simulation chamber. We gen-
erated the alkoxy radical by the photolysis of alkyl nitrite at
2
CO with OH is not very fast, but this problem was resolved
by using a high concentration of CO. Cyclohexane reacts more
quickly with OH but the main product is cyclohexanone,
2
8
29
and both of these may complicate the FTIR data analysis.
Therefore, only a few experiments have been performed with
this scavenger. CO and cyclohexane were added in such con-
centrations that the reaction of OH with other compounds
was negligible. This has been verified using well published rate
constants for the reaction with OH of CO, cyclohexane and
studied products.
around 420 nm. These four radicals can react either with O
reaction 2) or decompose (reaction 1). We have determined
the k /k ratio from the yield of the products and the concen-
tration of O
2
(
1
2
2
.
Initial concentrations of nitrite were in the ppm range in a
synthetic mixture of N /O . The percentage of O , between
II Experimental
2
2
2
0
.5 and 60% in a total pressure of one bar, was adjusted to
Experiments have been performed in an evacuable environ-
mental chamber. This Pyrex reactor of 977 litres is surrounded
by 40 UV-fluorescent tubes (Philips TL03) centred on 420 nm
maintain the two reaction pathways compatible with the detec-
tion limit of the two monitored compounds. In fact, various
partial pressures of O were used to vary the reaction rate of
2
(
Fig. 2). This irradiation device has been chosen in order to
reaction with O versus decomposition. All experiments have
2
give the minimum energy to the system when photolysing alkyl
nitrite and to work with small reactant losses. The reactor con-
tains a multiple reflection optical system interfaced to an FTIR
spectrometer (Bomem DA8-ME). Typically, the pathlength
was 156 metres. Details of the environmental chamber have
been conducted at 298 ꢀ 2 K with continuous irradiation
(
tra approx. every five minutes with a resolution of 0.5 cm
approx. two hours) and monitored by acquiring infrared spec-
.
ꢂ1
Alcohol used for nitrite synthesis and all products were
obtained from commercial sources with a purity higher than
2
3
been described previously by Doussin et al.
Alkoxy radicals have been generated by the photolysis of the
9
9%. 1-propanol, 2-propanol, 2-butanol and 3-pentanol, reac-
tants for nitrite synthesis, were purchased from Prolabo (pur-
ity > 99%). Propanal, acetaldehyde and formaldehyde were
obtained from Fluka (purity > 99%) and acetone, 2-butanone
and 3-pentanone from Aldrich (purity > 99%).
appropriate alkyl nitrite via the reaction: RONO + hn (ꢄ420
ꢃ
nm) ! RO + NO. This well known alkoxy radical source
1
4,20,22,24,25
has been used in several studies
as it has been shown
1
and 2-propoxy radical reaction, were synthesised following the
-propyl nitrate and 2-propyl nitrate, products of 1-propoxy
3
0
protocol of Desseigne by slowly adding 0.1 mL of 1-propa-
nol or 2-propanol, respectively, to a mixture containing 0.2
mL of H SO , 0.2 mL of HNO and a grain of urea, cooled
2
4
3
ꢅ
Fig. 1 Studied alkoxy radicals.
at 0 C. After a few minutes, the supernatant layer was
Phys. Chem. Chem. Phys., 2003, 5, 4834–4839
4835

View Article Online
separated and washed with a solution of 2% Na
to remove residual acid.
2 3
CO in water
To calibrate their infrared spectra, a known pressure of
compound was introduced in a 0.55 litre Pyrex bulb and
flushed into the chamber with nitrogen. Formaldehyde has
been calibrated in taking samples from the simulation chamber
and analysing it after derivatization with 2,4-dinitrophenylhy-
drazine (2,4-DNPH) followed by HPLC separation and UV
detection. The infrared Integrated Band Intensities (IBI) given
in Table 1 were used.
III Results
1
-propoxy radical experiments
Six experiments were performed on this radical by photolysing
the 1-propyl nitrite in the environmental chamber described
above with different partial pressures of O (0.5, 2 and 3%)
2
in N , including approx. 2% of CO as OH-scavenger with a
2
total pressure of 1.0 bar.
With the help of sequential infrared spectra acquired during
the photolysis, concentrations of reactants and products versus
time were followed. In addition to H O, NO , and O , the for-
Fig. 3 Concentration versus time for the photolysis of 1-propyl
nitrite in 0.5% of O
tion start.)
2
2
3
2 2
in N at atmospheric pressure. (Time 0 at irradia-
mation of specific marker carbonyl compounds (formaldehyde
and propanal for decomposition and reaction with O , respec-
2
tively) was monitored. 1-propyl nitrate from the reaction of
alkoxy with NO was also observed (Fig. 3).
2
environmental chamber with varying amounts of O (0.5 and
2
By plotting the formation of products versus the normalised
loss of nitrite (Fig. 4), it appears that the formaldehyde and the
propanal, whose plots have initial non-zero slopes, are primary
products (see yields in Table 2). The plots are linear confirming
that these carbonyl compounds do not further react, due to the
presence of OH-scavenger and that their photolysis rates are
negligible under the conditions used. Hence, the yield of each
product has been determined by linear regression.
1
.2%) in N
as scavenger, with a total pressure of 1 bar.
Aside from H O, NO , and O , the formation of marker
carbonyl compounds (formaldehyde and acetone for decom-
position and reaction with O , respectively) has been followed.
2
, with approx. 2% of CO or 27 ppm of cyclohexane
2
2
3
2
Formaldehyde was chosen has an indirect marker for decom-
position as acetaldehyde sensitivity of the analysis was not
high enough to use this compound. It has been verified, by
If reactions 1 and 2 are applied to the 1-propoxy radical:
4
0
numerical simulation using FACSIMILE software, that the
coproduct of acetaldehyde (i.e. methyl radical) gives exclu-
sively formaldehyde in our conditions. Indeed, from calcula-
tion we verified that methyl nitrate, methyl peroxy nitrate
and methyl hydroperoxide concentrations were always very
low. The presence of 2-propyl nitrate from the reaction of
ꢃ
ꢃ
CH3CH2CH2O ! CH3CH2 þ HCHO k1
ð1Þ
ð2Þ
ꢃ
CH3CH2CH2O þ O2 ! CH3CH2CHO þ HO2 k2
the following relationships are obtained:
k1
k2½O2ꢆ
Y1
Y2
k1
Y1 ¼
; Y2 ¼
then
¼
2
alkoxy with NO was also noted.
k2½O2ꢆ þ k1
k2½O2ꢆ þ k1
k2½O2ꢆ
By plotting the formation of products versus the normalised
loss of nitrite, it is apparent that these carbonyl compounds are
primary products (see yields in Table 2).
1 2 2
where Y and Y are the yields of reactions 1 and 2. With [O ]
and the ratio of the yield of the product from each pathway
being known, the k
(
1
/k
2
ratio has been calculated to be
1
6
ꢂ3
3.8 ꢀ 0.4) ꢁ 10 molecules cm (see Table 3).
2
-propoxy radical experiments
This radical has been studied over the course of seven
experiments involving photolysis of 2-propyl nitrite in the
Table 1 Infrared Integrated Band Intensities (IBI) of main monitored
products (uncertainty taken as twice s)
ꢂ1
ꢂ1
Product
Band/cm
IBI/cm molecule
Sources
ꢂ17
Ethyl nitrite
3087–2784
3030–2810
3036–2859
3018–2839
3015–2810
3125–2585
3084–2600
3067–2611
1282–1158
3075–2783
1819–1586
(1.4 ꢀ 0.1) ꢁ 10
This work
This work
This work
This work
This work
This work
This work
This work
This work
This work
This work
ꢂ17
ꢂ17
ꢂ17
ꢂ17
ꢂ18
1
2
2
3
-propyl nitrite
-propyl nitrite
-butyl nitrite
-pentyl nitrite
(2.28 ꢀ 0.02) ꢁ 10
(1.61 ꢀ 0.08) ꢁ 10
(2.56 ꢀ 0.1) ꢁ 10
(3 ꢀ 0.4) ꢁ 10
Formaldehyde
Acetaldehyde
Propanal
(3.8 ꢀ 0.2) ꢁ 10
ꢂ17
ꢂ17
ꢂ17
ꢂ17
ꢂ17
(1.85 ꢀ 0.04) ꢁ 10
(2.81 ꢀ 0.06) ꢁ 10
(1.11 ꢀ 0.05) ꢁ 10
(1.68 ꢀ 0.03) ꢁ 10
(1.97 ꢀ 0.06) ꢁ 10
Acetone
2
3
-butanone
-pentanone
Fig. 4 Normalised concentration of products versus normalised loss
of nitrite for the fate of 1-propyl nitrite under 0.5, 2 and 3% of O .
2
4
836
Phys. Chem. Chem. Phys., 2003, 5, 4834–4839

View Article Online
Table 2 Yield of the main primary product issue of each pathway at
0% of O pressure (uncertainty taken as twice s)
decomposition pathway) has been detected. Consequently, an
1
5
ꢂ3
2
2
upper limit for the ratio k /k of 5 ꢁ 10 molecules cm
1
2
was determined.
1
-propoxy
2-propoxy
2-butoxy
0.72(0.03)
0.64(0.05)
3-pentoxy
Formaldehyde
Acetaldehyde
Propanal
<0.02
<0.01
a
IV Discussion
1
0.42(0.01)
0.49(0.08)
Acetone
Butan-2-one
Pentan-3-one
1
Thermalisation of alkoxy radicals
As explained above, one of the main issues, regarding the
study of RO chemistry, is to evaluate whether or not the che-
ꢃ
a
One 2-butoxy radical produces two acetaldehyde molecules, one directly
and one via the ethyl radical, which changes exclusively into acetaldehyde
mical system used leads to the formation of chemically acti-
vated alkoxy radicals. In this study, nitrite was photolysed
using less energetic radiation than that used by other groups
2
in the conditions used with 20% of O .
ꢃ
who have encountered additional decomposition of RO radi-
1
6
1
6
cals produced from nitrites. In addition, Heicklen has shown
that if nitrite is photolysed at 254 nm, more than 85% of
alkoxy radicals produced are excited, but at 360 nm this falls
to less than 5%. In this work nitrite was photolysed at around
The ratio k
molecules cm (see Table 3).
1
/k
2
has been found to be (2.9 ꢀ 0.3) ꢁ 10
ꢂ3
2
-butoxy radical experiments
4
20 nm (Fig. 2).
Nevertheless, it was necessary to check the atmospheric rele-
A further nine experiments were carried out to investigate the
evolution of the 2-butoxy radical. These involved photolysis of
vance of our data. This was done by verifying that all the
excited RO radicals produced (if any) were thermalized before
decomposition. Hence, the data was first interpreted following
the protocol described by Wallington et al., assuming a sim-
ple mechanism shown in Fig. 5. In this process, the photolysis
of nitrite gives two types of RO radicals; those with internal
excitation energies above the barrier for C–C bond scission
that decompose promptly and those with energies below the
barrier that are collisionally equilibrated to give thermalized
RO radicals. Thermalized RO radicals can either undergo
thermal decomposition or reaction with O . If Y is the molar
yield of the product from the reaction with O2 and a the
fraction of alkoxy radicals which become thermalized, then:
2
pressures of O (40 and 60%) in N with approx. 2% of CO as
-butyl nitrite in the simulation chamber with different partial
ꢃ
2
2
OH-scavenger with a total pressure of 1 bar.
As well as H O, NO , and O , the formation of acetalde-
17
2
2
3
hyde and butan-2-one was traced throughout the experiments.
These are formed as primary compounds from decomposition
and reaction with O , respectively (see yields in Table 2).
2
By plotting the increase of products versus the normalised
loss of nitrite, the branching ratio between the decomposition
ꢃ
ꢃ
ꢃ
reaction and the reaction with O
(
2
, k
2.9 ꢀ 0.4) ꢁ 10 molecules cm (see Table 3). This value has
been derived by taking into account that the decomposition of
-butoxy radicals produces two molecules of acetaldehyde:
one directly, the other via the ethyl radical. Numerical simula-
1 2
/k , was found to be
2
1
8
ꢂ3
2
ꢀ
ꢁ
k2½O2ꢆ
1
1 k1 1
1
Y ¼ a
so
¼
þ
4
0
tions using FACSIMILE software have been performed to
check that ethyl radical evolves exclusively to acetaldehyde
via ethyl peroxy radical and ethoxy radical, in the experimental
conditions used. These calculations shown that the contribu-
tion of all the other C organic compounds (mainly C H OOH)
k ½O ꢆ þ k
Y
a k O2
a
2
2
1
2
The intercept of the plot of 1/Y versus 1/[O
/a is greater than 1 then an additional decomposition is
assumed.
From Fig. 6 and Table 4, it can be seen that a is not signifi-
2
] gives 1/a. If
1
2
2
5
was always smaller than 5%. Nevertheless, this small artefact
was taken into account in the uncertainty evaluation.
cantly different from unity for all the studied alkoxy radicals.
From this it can be concluded that the produced alkoxy radi-
cals are thermalized and that the branching ratios derived from
this study are relevant to the atmosphere.
3
-pentoxy radical experiments
Seven experiments were carried out on the 3-pentoxy radical,
during which the relevant nitrite was photolysed in the envir-
This behaviour of alkoxy radicals is not surprising as it has
been shown previously by Wallington et al. that an energy of
1
7
ꢂ1
onmental chamber with varying partial pressures of O
0 and 60%) with approximately 2% of CO as an OH-scaven-
ger made up to 1 bar with N
As well as monitoring H O, NO
specific carbonyl compounds (propanal and pentan-3-one for
2
(10,
approximately 40 kcal mol is available at 360 nm (the lower
limit of our photolysis spectrum, see Fig. 2) and from molecu-
2
3
1,32
17
Wallington et al. evaluated
2
.
lar beam studies in literature,
ꢂ1
2
2
, and O
3
, the formation of
that only approx. 7 kcal mol ends up in the vibrational mode
of alkoxy radicals, i.e. typically half the amount required to
overcome the dissociation barrier of the alkoxy.
decomposition and reaction with O
lowed throughout the experiments.
2
, respectively) was fol-
By plotting the formation of products versus the normalised
loss of nitrite, it is apparent that these carbonyl compounds are
primary products (see yields in Table 2).
Reactivity of alkoxy radicals
1
8
For the studied alkoxy radicals, the branching ratio has been
calculated in tropospherically relevant conditions. For atmo-
spheric interpretation purposes, only these ratios are relevant
as reactions 1 and 2 are so fast that they are never the limiting
step.
The ratio k
molecules cm (see Table 3).
1
/k
2
has been found to be (4.1 ꢀ 0.8) ꢁ 10
ꢂ3
Ethoxy radical experiments
The same protocol was used to study the ethoxy radical.
Despite this, no primary formaldehyde (i.e. from the
For 1-propoxy and 2-propoxy, the present work is, to our
knowledge, the first direct measurement of k /k . The existing
1 2
1 2
Table 3 Rate constant ratio k /k of the four alkoxys studied (uncertainty taken as twice s)
1
-propoxy
2-propoxy
2-butoxy
3-pentoxy
3
/ cm molecule
ꢂ1
16
16
18
18
k
1
/k
2
(3.8 ꢀ 0.4) ꢁ 10
(2.9 ꢀ 0.3) ꢁ 10
(2.9 ꢀ 0.4) ꢁ 10
(4.1 ꢀ 0.8) ꢁ 10
Phys. Chem. Chem. Phys., 2003, 5, 4834–4839
4837

View Article Online
Table 4 Fraction of excited alkoxy radicals being thermalized
1
-propoxy
2-propoxy
2-butoxy
3-pentoxy
þ0:02
0.98
a
0.9 ꢀ 0.1
0.9 ꢀ 0.1
0.9 ꢀ 0.1
ꢂ0:06
On the other hand, a comparison of 2-propoxy and 3-pent-
Fig. 5 Mechanism of reaction, if the photolysis of 1-propyl nitrite
produces one excited 1-propoxy radical (example).
oxy, where the class remains the same but the leaving alkyl
group is growing from methyl in 2-propoxy to ethyl in 3-pent-
oxy shows that the branching ratio is again multiplied by a
factor of 100.
These observations confirm the conclusions of M e´ reau
available data are measurements of k1 or k2 sepa-
11
et al. which state that the smaller the ionisation potential
of the leaving group, the faster the decomposition rate.
4
,13,15,19,33
rately.
For 1-propoxy and 2-propoxy, k /k values published here
falls within the range calculated from various previously pub-
1
2
lished k and k values. Nevertheless, for these compounds, the
1
2
V Conclusion
direct determination method used in the present work bring a
better precision in the evaluation of these k /k ratios.
1
2
The branching ratios between decomposition (1) and reaction
with oxygen (2) for four alkoxy radicals (1-propoxy, 2-pro-
poxy, 2-butoxy and 3-pentoxy) have been measured under sev-
eral partial pressures of O2 in an evacuable simulation
chamber at 298 ꢀ 2 K. Photolysis of alkyl nitrite at around
The branching ratios between decomposition and reaction
with oxygen for 2-butoxy and 3-pentoxy have been the subject
4
–6,34
of previous direct determinations.
Table 5 that the data from this study are consistent with both
It can be seen from
4
–6,35
these previous values
lished values for separate measurements of k and k .
and the ranges derived from pub-
4
20 nm has been used as the alkoxy radical source and no che-
1
2
mical activation of the produced radical has been observed.
A strong dependence of this branching ratio on the leaving
alkyl group has been identified. As there is very little variation
For the 2-butoxy radical, the calculated values for k
1
/k
/k
2
range over an order of magnitude. This is principally due to
3
5
value determined by Hein et al. From the k
the low k
1
1
2
in the rate constants for the reaction with O
is mainly due to changes in the decomposition rate. The values
found for k /k are in good agreement with previously pub-
2
, this dependence
value determined in this study it can be suspected that the
above rate constant may be underestimated.
The four alkoxy radicals shown in Fig. 1 were chosen, in
order to study the change in reactivity with the variation of
the alkoxy radical structure using a common procedure. From
the results obtained it is possible to comment on the variation
1
2
lished ratios i.e. for the 2-butoxy and 3-pentoxy radicals. The
tendencies found support the conclusions made by M e´ reau
et al. from ab initio calculations. Studies are in progress to
11
extend these observations to other branched alkoxys and to
derive a topological relationship between the fate of alkoxy
radicals and their structure.
The results presented here can be used directly for tropo-
spheric simulation purposes as great care has been taken
with the atmospheric relevance of the experimental condi-
tions used (ambient temperature, atmospheric pressure, low
concentrations).
1
6
3
of the k /k ratio. This ratio is close to 3 ꢁ 10 cm
1
ꢂ1
2
molecule for the 1-propoxy and the 2-propoxy radical but
multiplies by a factor of 100 for 2-butoxy and 3-pentoxy. As
ꢂ15
the values published for k
2
only range from 7 ꢁ 10
ꢂ1 ꢂ1
cm molecule s , this huge increase in k
to
is
ꢂ15
3
1
2 ꢁ 10
1 2
/k
probably mainly due to a change in the decomposition rate.
1
1
The tendencies found by M e´ reau et al. from their quantum
calculations show that the extent of the decomposition path-
way depends on the structure of the leaving alkyl group and
on the class of alkoxy radical involved. In this work, similar
trends were observed.
If neither the leaving alkyl group nor the class changes, then
the branching ratio remains almost unchanged as is the case
with the 2-butoxy and 3-pentoxy radicals.
But, on comparing 1-propoxy and 3-pentoxy radicals, where
the leaving alkyl group remains the same (ethyl) but the class
of alkoxy radical changes from primary to secondary, the
branching ratio increases by two orders of magnitude.
In the atmosphere, the four alkoxys studied here are pro-
duced either directly from the oxidation of C , C and C5
3
4
alkanes or from the decomposition of longer chain alkoxys
produced from the oxidation of larger VOCs. Their atmo-
spheric importance relies on the extent of C
formation and on the subsequent reactions of these products.
Indeed, formaldehyde may be a photolytic source of HO
through the production of H and HCO radicals and acetalde-
1 2
or C aldehyde
x
hyde is the key precursor of the toxic NO
acetyl nitrate (PAN).
x
reservoir, peroxy-
In the lower troposphere, 1-propoxy and 2-propoxy radicals
react mainly with O2 while decomposition is an important
reaction pathway for 2-butoxy and 3-pentoxy. Consequently,
C1 and C2 aldehyde production from the two longer chain
alkoxys will occur very close to the area of initial VOC oxida-
tion whereas for the alkoxys exhibiting a minor decomposition
pathway, the formaldehyde or acetaldehyde production will
take place after the oxidation of all the intermediate secondary
compounds, far from the emission area of the primary
compound.
Acknowledgements
This work has been carried out with the support of CNRS
through the PNCA program. Professor Friedhelm Zabel is
gratefully thanked for sharing spectroscopic information.
ProfessorJ.C. RayezandProfessorM.T. Rayezarealso thanked
2 2
Fig. 6 Plots of 1/Y for the product of reaction with O versus 1/O
for the four alkoxy radicals.
4
838
Phys. Chem. Chem. Phys., 2003, 5, 4834–4839

View Article Online
Table 5 Comparison between our values and those from literature
ꢂ1
3
/cm molecule
ꢂ1 ꢂ1
3
range / cm molecule
ꢂ1 a
3
/ cm molecule
ꢂ1
k
1
/s
k
2
s
k
1
/k
2
k
1
/k
2
2
2
2
11c
37
þ7
ꢂ15 36
ꢂ15 15
ꢂ15 21b
16
16
(3.8 ꢀ 0.4) ꢁ 10 this work
1
2
-propoxy
-propoxy
2.5 ꢁ 10
(10 ) ꢁ 10
(2.3–7.4) ꢁ 10
ꢂ5
þ5
5
.9 ꢁ 10
.9 ꢁ 10
(11 ) ꢁ 10
ꢂ4
38c
2
9.5 ꢁ 10
ꢂ15 10b
8
ꢁ 10
2
19
þ10
ꢂ15 13
16
16
(2.9 ꢀ 0.3) ꢁ 10 this work
4 ꢁ 10
(7.8 ) ꢁ 10
(1.4–6.1) ꢁ 10
ꢂ5
2
11c
þ0:7
(6.8 ) ꢁ 10
ꢂ0:5
ꢂ15 33,36
1
.3 ꢁ 10
2
2
37
þ1:4
ꢂ15 15
(6.6 ) ꢁ 10
ꢂ1:3
1
2
.1 ꢁ 10
.9 ꢁ 10
38c
ꢂ15 10,21b
8 ꢁ 10
3
35
ꢂ15 14
ꢂ15 35
ꢂ15 10,21b
18
18
(2.9 ꢀ 0.4) ꢁ 10 this work
2
-butoxy
(3.5 ꢀ 2) ꢁ 10
(12 ꢀ 4) ꢁ 10
(0.29–5.4) ꢁ 10
4
4
4
4
4
4
11c
37
18 5
3
2
1
.5 ꢁ 10
.1 ꢁ 10
.6 ꢁ 10
(6.5 ꢀ 2) ꢁ 10
3.15 ꢁ 10
18 4
8 ꢁ 10
(2.3 ꢀ 0.5) ꢁ 10
38c
11c
37
18 6
(2.6 ꢀ 0.35) ꢁ 10
ꢂ15 22
18
18
(4.1 ꢀ 0.8) ꢁ 10 this work
3
-pentoxy
3.4 ꢁ 10
(12 ꢀ 6) ꢁ 10
(1.4–4.7) ꢁ 10
ꢂ15 39
18 34
2
.1 ꢁ 10
.6 ꢁ 10
(7.2 ꢀ 3.5) ꢁ 10
(3.6 ꢀ 0.6) ꢁ 10
38c
ꢂ15 10,21b
1
8 ꢁ 10
a
b
c
Range calculated from two first column values. Review. Theoretical calculation.
for helpful discussions. MsMargaret O’Connor is also gratefully
thanked for her help in the preparation of this manuscript.
19 P. Devolder, C. Fittschen, A. Frenkel, H. Hippler, G.
Poskrebyshev, F. Striebel and B. Viskolcz, Phys. Chem. Chem.
Phys., 1999, 1, 675–681.
2
0
C. Fittschen, H. Hippler and B. Viskolcz, Phys. Chem. Chem.
Phys., 2000, 2, 1677–1683.
References
2
2
1
2
R. Atkinson, Int. J. Chem. Kinet., 1997, 29, 99–111.
W. Deng, A. J. Davis, L. Zhang, D. R. Katz and T. S. Dibble,
J. Phys. Chem. A, 2001, 105, 8985–8990.
J. F. Doussin, D. Ritz, R. Durand-Jolibois, A. Monod and P.
Carlier, Analusis, 1997, 2, 236–242.
A. C. Baldwin and D. M. Golden, Chem. Phys. Lett., 1978, 60(1),
1
2
R. Atkinson, Atmospheric Environment, 2000, 34, 2063–2101.
W. P. L. Carter, Reactions of Alkoxy Radicals, in Documentation of
the SAPRC-99 chemical mechanism for VOC reactivity assessment,
California Air Resources Board, Riverside, CA, 1999, pp. 88–131.
R. Atkinson, E. S. C. Kwok, J. Arey and S. M. Aschmann,
Faraday Discuss., 1995, 100, 23–37.
H. G. Libuda, O. Shestakov, J. Theloke and F. Zabel, Phys.
Chem. Chem. Phys., 2002, 4, 2579–2586.
W. P. L. Carter, A. C. Lloyd, J. L. Sprung and J. N. Pitts, Jr., Int.
J. Chem. Kinet., 1979, 11, 45–101.
R. A. Cox, K. F. Patrick and S. A. Chant, Environ. Sci. Technol.,
23
24
25
26
27
3
4
5
6
7
8
1
08–111.
C. Fittschen, B. Delcroix, N. Gomez and P. Devolder, J. Chim.
Phys., 1998, 95, 2129–2142.
M. I. Christie and P. M. Hetherington, J. Photochem., 1976–1977,
6(4), 285–295.
W. D. Taylor, T. D. Allston, M. J. Moscato, G. B. Fazenkas, R.
Koslowski and G. A. Takacs, Int. J. Chem. Kinet., 1980, 12(4),
2
1981, 15(5), 587–592.
31–240.
NIST, Chemical Kinetics Database on the Web, http://kinetics.
nist.gov.
H. Niki, P. D. Maker, C. M. Savage and L. P. Breitenbach,
J. Phys. Chem., 1981, 85, 2698–2700.
J. Theloke, H. G. Libuda and F. Zabel, The Atmospheric Fate of
Isobutoxy Radicals, in Proceedings of EUROTRAC Symposium
2
2
8
9
S. M. Aschmann, A. A. Chew, J. Arey and R. Atkinson, J. Phys.
Chem. A, 1997, 101, 8042–8048.
G. Desseigne, Mem. Poudres, 1948, 30, 58.
F. Lahmani, C. Lardeux and D. Solgadi, Chem. Phys. Lett., 1983,
98, Boston, ed. P. M. Borrel and P. Borrel, 1999, WIT Press,
Southampton, pp. 169–172.
3
3
0
1
9
0
1
2
3
4
5
F. Zabel, The atmospheric fate of long-chain alkoxy radicals, in
EC/eurotrac-2 Joint Workshop, 2000, GPP22.
P. Devolder, J. Photochem. Photobiol. A: Chem., 2003, 157,
1
02(6), 523–528.
U. Br u¨ hlmann, M. Dubs and J. R. Huber, J. Chem. Phys., 1987,
6(3), 1249–1257.
32
33
34
1
1
1
1
1
1
8
137–147.
C. Mund, C. Fockenberg and R. Zellner, Phys. Chem. Chem.
Phys., 1999, 1, 2037.
O. Shestakov, H. G. Libuda and F. Zabel, The Atmospheric Fate
of 3-Pentoxy Radicals, in EC/eurotrac-2 joint workshop, ed.
R. Vogt and G. Axelsdottir, Ford Forschungszentrum, Aachen,
R. M e´ reau, M. T. Rayez, F. Caralp and J. C. Rayez, Phys. Chem.
Chem. Phys., 2000, 2, 3765–3772.
O. J. Nielsen, H. W. Sidebottom, M. Donlon and J. Treacy, Int.
J. Chem. Kinet., 1991, 23, 1095–1109.
R. J. Balla, H. H. Nelson and J. R. McDonald, Chem. Phys.,
1
999, pp. 97–100.
1985, 99, 323–335.
35
36
37
38
39
40
H. Hein, A. Hoffmann and R. Zellner, Ber. Bunsen-Ges. Phys.
Chem., 1998, 102(12), 1840–1849.
C. Mund, C. Fockenberg and R. Zellner, Ber. Bunsen-Ges. Phys.
Chem., 1998, 102(5), 709–715.
J. Peeters, G. Fantechi and L. Vereecken, J. Atmos. Chem., 2003,
submitted.
H. Somnitz and R. Zellner, Phys. Chem. Chem. Phys., 2000, 2,
W. Deng, C. Wang, D. R. Katz, G. R. Gawinski, A. J. Davis and
T. S. Dibble, Chem. Phys. Lett., 2000, 330, 541–546.
C. Fittschen, A. Frenzel, K. Imrik and P. Devolder, Int. J. Chem.
Kinet., 1999, 31, 860–866.
1
1
6
7
J. Heicklen, Adv. Photochem., 1988, 14, 177–272.
T. J. Wallington, M. D. Hurley, J. M. Fracheboud, J. J. Orlando,
G. S. Tyndall, J. Sehested, T. E. Møgelberg and O. J. Nielsen,
J. Phys. Chem., 1996, 100, 18 116–18 122.
4
319–4325.
H. Hein, H. Somnitz, A. Hoffmann and R. Zellner, Z. Phys.
Chem., 2000, 214(4), 449–471.
FACSIMILE (v 3.0), AEA Technology, UK.
18
F. Caralp, P. Devolder, C. Fittschen, N. Gomez, H. Hippler, R.
Mereau, M. T. Rayez, F. Striebel and B. Viskolcz, Phys. Chem.
Chem. Phys., 1999, 1, 2935–2944.
Phys. Chem. Chem. Phys., 2003, 5, 4834–4839
4839