Degradation mechanism of t-butyl methyl ether (MTBE) in atmospheric droplets

Article abstract of DOI:10.1016/S0045-6535(03)00547-2

Information about the degradation of MTBE in atmospheric water droplets (rain, clouds, fog) was studied. These water droplets contained hydrogen peroxide and iron ions, which are a source of the powerful oxidizing radical OH°, particularly under solar irradiation (photo-Fenton reaction). MTBE was selected because of its current use as an oxygenated additive in gasoline. MTBE was not stable in the atmosphere. Acetone was about 15 times more stable than MTBE in atmospheric conditions. The formation of acetic and formic acid contributed to the acidification of the water droplets.

Full text of DOI:10.1016/S0045-6535(03)00547-2

Chemosphere 53 (2003) 469–477
www.elsevier.com/locate/chemosphere
Degradation mechanism of t-butyl methyl ether (MTBE)
in atmospheric droplets
a,
Chantal Guillard , Nathalie Charton , Pierre Pichat
b
b
*
a
Laboratoire d’Application de la Chimie aꢀ l’Environnement (LACE), CNRS––Universit eꢁ Claude Bernard LYON 1,
eꢀ me
Unit eꢁ Mixte de Recherche 5634, B a^ timent J. Raulin-3 eꢁ tage, 43, Boulevard du 11 novembre 1918, 69622 Villeurbanne Cedex, France
Laboratoire ‘‘Photocatalyse, Catalyse et Environnement’’, CNRS UMR 5621 ‘‘IFoS’’, Ecole Centrale de Lyon,
b
BP 163, 69131 Ecully C eꢁ dex, France
Received 16 December 2002; received in revised form 9 May 2003; accepted 19 May 2003
Abstract
The aim of this study was to obtain information about the degradation of t-butyl methyl ether (MTBE; (CH
3 3
) C–O–
CH ) in atmospheric water droplets (rain, clouds, fog). These water droplets contain hydrogen peroxide and iron ions,
3
which are a source of the powerful oxidising radical OH°, particularly under solar irradiation (photo-Fenton reaction).
MTBE was chosen for this work because of its current use as an oxygenated additive in gasoline.
In this study we found that MTBE is not stable in the atmosphere. More than 15 intermediate products were
identified, five of which were quantified (t-butyl formate (TBF), methyl acetate (MA), t-butyl alcohol (TBA), acetone
(
AC), formaldehyde). The evaluation of the disappearance kinetic of the main intermediate compounds shows the
following activity pattern kðTBAÞ > kðMTBEÞ > kðTBFÞ; k >ðACÞ. Acetone was found to be about 15 times more stable than
MTBE in atmospheric conditions. The degradation pathways are discussed on the basis of these identifications and on
the degradation of the main intermediate products in similar conditions to MTBE.
Ó 2003 Elsevier Ltd. All rights reserved.
Keywords: MTBE; Degradation; Photo-Fenton; Mechanism; Atmosphere
1
. Introduction
ted by the major petroleum companies. t-butyl methyl
ether (MTBE) is the most common oxygenated fuel ad-
ditive used in reformulated gasoline because of its low
cost, high solubility, ease of production, and transfer and
blending characteristics. From 1995 to 2000, the pro-
duction of MTBE quadrupled. However, due to vehicular
emissions, underground petroleum storage leakage and
its high volatility and hydrosolubility, MTBE is begin-
ning to appear in ground and surface water and in the
atmosphere near gasoline station, raising serious prob-
lems (LaGrone, 1991; Kelly et al., 1993; Squillace et al.,
1997; Grosjean et al., 1998; Vainiotalo et al., 1998). Ac-
cording to Henryꢀs law, for 4 ppbv present in the atmo-
sphere, the concentration of atmospheric droplets will be
A few years ago, additives including mainly organo-
3 4 2 5
lead compounds such as (CH ) Pb and (C H )Pb were
used in gasoline to maintain the octane rating, but the
longevity of the catalytic convectors was affected by ex-
posure to lead additives. The implementation of legisla-
tions on fuels aiming at the suppression of lead and the
reduction in the level of aromatics and of the more vol-
atile fractions has led to changes in fuel formulations.
Therefore, the use of oxygenated fuel as an octane
x
booster and a reducing agent of CO, NO and hydro-
carbon emissions from vehicle exhausts, has been adop-
ꢀ
1
ꢀ1
around 0.8 mg l at 293 K and 5 mg l at 273 K. H
nþ
2
O
2
*
and Fe can also be present at relatively high concen-
trations in fog and clouds. The average concentration of
Corresponding author. Tel.: +33-4-72-44-62-15; fax: +33-4-
2-44-84-38.
E-mail address: chantal_guillard@yahoo.fr (C. Guillard).
7
0
2 2
H O and dissolved iron in atmospheric droplets was near
045-6535/$ - see front matter Ó 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0045-6535(03)00547-2

470
C. Guillard et al. / Chemosphere 53 (2003) 469–477
4
0 and 200 ppb respectively (Hoigne et al., 1994). The
purchased from Air Liquide and sulphuric acid was a
Merck product. All these products were employed
without further purification. The water used for solution
and eluent preparations was doubly distilled water
(Milli-Q System).
addition of both these compounds, known as Fenton
reactants, forms hydroxyl radicals. In the presence of UV
light, at k < 350 nm, where 4% of solar energy is efficient,
the amount of OH° increases and ferrous ions become
catalysts. Thus, MTBE should not be stable in the at-
mosphere but should be degraded by the OH° formed by
2.2. Photoreactor
nþ
the photo-Fenton reaction (UV/H
2
O
2
/Fe ).
In previous studies of general gasoline additives
The aqueous reaction was carried out in a 90 ml
Pyrex cylindrical flask with a bottom optical window
(
Wallington et al., 1988a,b; Dagaut et al., 1989), the OH°
2
reaction rate constants are measured at different tem-
peratures using the flash photolysis technique. He found
that the reaction rate constant of formation of aliphatic
ether increases with the length of the carbon chain. At 298
K, the rate constant of OH° with MTBE is found to be
whose surface area was approximately 11 cm and which
transmitted wavelengths greater than approximately 290
nm. This flask was directly connected to a gas chro-
matograph with a gas valve in order to analyse the
atmosphere above the liquid. The irradiation was per-
formed by a Philips HPK 125 W high pressure mercury
lamp. A 2.2 cm-thick circulating-water cell was placed
above the lamp to absorb the IR part of the emission. A
United Detector Technology model 21A powermeter
was used to measure the radiant flux; the value was
found to be 50 ꢁ 2 mW cmꢀ . The corresponding num-
ber of photons per second was estimated to be
ꢀ
12
3
ꢀ1 ꢀ1
s . This result agrees
3
.09 ꢁ 0.15ꢂ 10 cm molecule
with the relative data found by several others (Cox and
Goldstone, 1982; Atkinson, 1989; Nelson et al., 1990).
Since 1982 and until now, an important number of
publication treat of the degradation of MTBE in atmo-
sphere or in water droplets either by using OH° radicals
2
formed by the photolysis of methyl nitrite (CH
3
ONO)
1
8
ꢀ1
ꢀ1
(
Cox and Goldstone, 1982; Japar et al., 1990; Smith et al.,
991; Tuazon et al., 1991) which seem to be the major
1.185 ꢂ 10 photons s
W . Experiments were carried
out at 295 K. This temperature was controlled by cir-
culating water in the jacket surrounding the reactor.
1
source of hydroxyl radicals in fog (Anastasio and
McGregor, 2001), or by different oxidation or photo-
oxidation processes such as O
Leitner et al., 1994; Liang et al., 1999; Acero et al., 2001),
UV (Miyake and Shibamoto, 1999), UV/H (Cater
3
, O
3
/H
2
O
2
(Karpel Vel
2.3. Procedures
2
O
2
Low concentrations in reactants and products with
low concentrations were used to study the degradation
of the MTBE close to atmospheric conditions. Five
micromoles per litre of MTBE was degraded by using
et al., 2000; Stefan et al., 2000). t-butyl formate (TBF)
and t-butyl alcohol (TBA) were found the main oxida-
tion products of MTBE, the other compounds formed
were generally formaldehyde, acetone (AC), methyl
acetate (MA), 2-methoxy-2-methyl propionaldehyde.
However, no publications have studied what happens to
ꢀ
1
a solution containing 5 lmol l of FeSO
4 2
Æ 7H O as a
Fe(II) source and H 30% v/v in water ([H
2
2
O
2
O
2
] ¼ 5
ꢀ1
lmol l ). To study the production and subsequent de-
struction of the intermediate products, higher concen-
MTBE in atmospheric droplets containing H
2
O
2
and
ꢀ
1
iron ions while the presence of iron ions favors the for-
mation of OH° in presence of solar irradiation, that is to
say at wavelength situated between 290 and 340 nm.
Therefore, the aims of this study were (i) to determine
trations were used: [MTBE] ¼ 1000 lmol l , [Fe(II)] ¼
ꢀ
1
ꢀ1
100 lmol l , [H
2
O
2
] ¼ 400 lmol l . The pollutant was
introduced into the photoreactor containing the ferrous
solution. The pH was adjusted to 3 (i.e. within the pH
range of tropospheric water) with sulphuric acid. The
suspension was magnetically stirred for 30 min in the
2 2
the effect of dissolved iron, H O and solar light on the
disappearance of MTBE and of its intermediate products
in the atmosphere, and (ii) to identify and quantify the in-
termediate products formed during the destruction of
MTBE in order to get an insight into the degradation
pathways and to understand its role in atmospheric
chemistry.
2 2
dark before the addition of H O and the irradiation.
This length of time was more than that needed for the
pollutant to reach equilibrium between the gaseous and
the aqueous phases. The atmosphere was analysed by
sampling it at different irradiation times. The amount of
2 2
H O was kept stable throughout the degradation.
2
. Experimental methods
2.4. Analyses
2
.1. Material
2.4.1. GC/FID analyses
To measure the kinetics of MTBE removal, the gas-
eous phase was directly injected into a Intersmat/Delsi
IGC 120FL gas chromatograph equipped with a Vocol
column (length: 105 m; 0.53 mm i.d.; film thickness:
MTBE (99.8%) and most of the chemical products
were supplied by Aldrich; 2-methylpropene was ob-
tained from Fluka. Methane, ethane and isobutane were

C. Guillard et al. / Chemosphere 53 (2003) 469–477
471
3
lm), and a flame ionisation detector (FID); nitrogen
optical Pyrex window to filter out the wavelengths
shorter than 290 nm.
ꢀ1
was the carrier gas (flow rate: 5 ml min ); the oven
temperature was maintained at 413 K for 30 min. The
identification of light intermediate products was carried
out with a GS-Q column (length: 30 m; 0.543 mm and
2
2
.4.5. Methods for concentrating the gaseous products
.4.5.1. Solid phase microextraction. Solid phase
0
.324 mm i.d.).
microextraction (SPME) is a solvent-free extraction
technique which is significantly more rapid, simple, in-
expensive and selective than the commonly employed
extraction methods (Pawliszyn, 1997). Analytes are ad-
sorbed directly from the sample onto a fused-silica fibre
which is coated with an appropriate stationary phase.
Before each use, this phase is first conditioned at a
temperature just below the maximal temperature to
which it can be submitted. Different SPME fibres have
been used: carbowax divinylbenzene phase (65 lm) for
polar compounds, polydimethylsiloxane (PDMS) phase
2
.4.2. GC/MS analyses
Identification of the intermediate products was car-
ried out using a gas chromatograph HP 5890 Series II
equipped with three different columns and connected to
a mass spectrometer HP 5971A operated in electron
ionisation (EI) mode. The characteristics of the different
columns were: CP Wax 57 CB (length: 50 m; i.d.: 0.25
ꢀ1
mm; film thickness: 0.2 lm; flow rate: 0.73 ml min ),
Vocol (length: 60 m; i.d.: 0.25 mm; film thickness: 1.5
ꢀ
1
lm; flow rate: 0.9 ml min ), CP Sil 5 CB (length: 25 m;
i.d.: 0.25 mm; film thickness: 1.2 lm; flow rate: 0.47
(
100 lm) for volatile compounds, PDMS (30 lm) for
nonpolar compounds, polyacrylate (85 lm) for volatile
and nonpolar compounds, and PDMS/divinylbenzene
ꢀ1
ml min ). If not indicated otherwise, we used the fol-
lowing temperatures: oven ¼ 313 K; injector ¼ 483 K;
detector ¼ 553 K. The identity of the intermediate
products was confirmed by comparing the retention time
and mass spectra with those of commercially available
compounds.
(
PDMS/DVB) (65 lm) for very polar compounds.
2
.4.5.2. Active carbon cartridge. Active carbon is poorly
selective and adsorbs most pollutants. An ORBO 32
Supelco) active carbon cartridge was placed at the exit
(
of the reactor. A helium gas stream swept the gaseous
phase off the reactor. Active carbon was then poured
into a flask containing 1 ml of dichloromethane and the
filtrate was analysed by GC-MS (columns CP Wax 57
CB and Vocol).
2
.4.3. HPLC analyses
Carbonyl products were analysed after derivatisation
with 2,4-DNPH (2,4-dinitrophenylhydrazine) to form
hydrazones, and subsequent concentration with Sep-Pak
C18 cartridges and extraction with acetonitrile. The
hydrazones were separated by a HPLC system with a
Varian 9065 polychrom diode-array HPLC detector and
a LC Star Varian System. A Spherisorb ODS2 column
2
.4.5.3. Hydroxymethyl piperidine cartridge. A piperidine
cartridge was placed at the output of the reactor. Deri-
vatisation of aldehydes leads to the formation of alde-
hyde oxazolidine. Adsorbed products were extracted by
(
length: 25 cm; 4.6 i.d.; 5 lm film thickness) was used.
ꢀ1
The mobile phase (flow rate: 1.5 ml min ) was com-
posed of three solvents: acetonitrile ¼ solvent A, tetra-
hydrofuran ¼ solvent B and water ¼ solvent C. These
solvents were flown with the following time program: 0–
1
ml of toluene placed under ultrasonication for 1 h.
2
.4.5.4. Trapping in liquid nitrogen. The gaseous phase
was swept off the reactor by a helium gas stream and
trapped into liquid nitrogen. The trap had a W-shape
and a volume of 15 ml. It contained small glass beads to
maximize the contact area. The trap was then brought to
room temperature and the gaseous phase sampled with a
gas syringe and injected into the GC-MS apparatus.
1
1
min: A 30% B 10% C 60%; 1–10 min: from A 30% B
0% C 60% to A 60% B 40%.
Products were identified by comparing the retention
time and the UV spectra obtained to those of the pure
compounds.
Carboxylic acids were analysed with a LDC/Milton
Roy System which comprised a Constametric 3000 Iso-
cratic pump, a Spectro-Monitor UV detector set at 210
nm, a Sarasep CAR-H column (length: 30 cm; i.d.; 7.8
mm) and a guard column. The eluent was sulphuric acid
3. Results and discussion
3.1. Disappearance of MTBE
ꢂ 10ꢀ mol l (flow rate: 0.7 ml min ).
1
ꢀ1
ꢀ1
5
Fig. 1 shows the degradation of MTBE by photo-
Fenton for concentrations close to those of tropospheric
2
.4.4. Quadripolar mass analyser
¼ 5 lmol lꢀ for MTBE, Fe and
1
2þ
Some identifications were performed by using of a
water droplets (C
). The H
0
quadripolar mass analyser RIBER QMM 17. MTBE
photochemistry experiments were carried out in a 10 ml
quartz cell. The aqueous suspension was irradiated by a
Philips HPK-125 W lamp. The cell was fitted with an
H
2
O
2
2
O
2
concentration, measured by an UV
spectrophotometer, was maintained constant. A sigmoid
curve was observed. The first period can be described
by Eqs. (1) and (2):

472
C. Guillard et al. / Chemosphere 53 (2003) 469–477
experiments are reported in Table 1. In conclusion, the
0
.4
.35
.3
.25
.2
.15
.1
3
2
1
0
0
0
0
0
0.028min-1
analysis of gaseous and aqueous phases of the reactor
enabled us to detect 13 organic intermediate products,
whose identifications were confirmed by comparing the
mass spectrum with that reported in the literature or
found for commercial standards.
0
0
0
50
100
t (min)
150
0
3.3. Kinetic variations in TBF, AC, MA, TBA and
formaldehyde amounts
.05
0
0
20
40
60
80
100
120
140
In the case of the five most abundant intermediates,
GC/FID analysis allowed us to follow their production
and subsequent destruction as a function of the illumi-
nation time, in a suspension initially containing 1000
t (min)
Fig. 1. Degradation of MTBE as a function of irradiation time.
¼ 5 lmol lꢀ for MTBE, Fe and H
1
2þ
C
0
2 2
O .
ꢀ1
lmol l of MTBE. Their evolution is shown in Fig. 2a.
The other intermediate products had concentrations
too low to be quantified. Methane and ethane reach
amonts of 1.15 and 0.2 lmol respectively after 1 h and
10 min and 2 h and 45 min of illumination. They ac-
count for about 5% of initial MTBE. At the beginning of
the reaction about 50% of MTBE degradation produced
the TBF. During the reaction, the five major products
accounted for 60–80% of initial MTBE (Fig. 2b).
Most of these products do not reach important
amounts. The main intermediate, TBF, reaches its high-
est amount (20.5 lmol) after about 3 h of illumination
and then its concentration decreases. Its degradation
appears to be very slow since this compound is detected
even after 500 min of illumination.
2
þ
3þ
ꢀ
Fe þ H
2
O
2
! Fe þ OH þ OH°
k ¼ 76 l molꢀ
1
ꢀ1
s
ðSafarzadeh-Amiri et al:; 1997Þ
ð1Þ
Fe þ OHꢀ ! FeðOHÞ2þ
3
þ
ð2Þ
After this period, the increase of the rate is due to the
photoreduction of ferric ions (Eq. (3))
FeðOHÞ þ ! Fe2þ þ OH°
2
ð3Þ
The disappearance of MTBE can be described by the
equations d½MTBEꢃ=dt ¼ kOH°½OH°ꢃ ½MTBEꢃ .
t
t
From the insert of Fig. 1, after the induction period
the first apparent rate constant was determined to be
The fact that the pollutant and the intermediate
products progressively disappear, whereas substantial
amounts of AC remains in the solution, indicates that
AC is more stable than MTBE. Thus, we have to take
into account in atmospheric pollution models the for-
mation of AC from MTBE, unintentionally released in
the atmosphere during gas transfers.
ꢀ1
0
.028 min . Therefore, we can considerate that a con-
stant OH° radical concentration has been found. Taking
into account the rate constant of OH° radical onto
9
ꢀ1 ꢀ1
s )
MTBE cited by Buxton (kOH° ¼ 1.6 ꢂ 10 l mol
(
Buxton et al., 1988), a OH° radical concentration of
ꢀ
7
ꢀ1
2
.9 ꢂ 10 lmol l will be formed under our conditions.
This result agrees with the concentration of OH°
radicals inside the atmospheric droplets, showing that
our experimental conditions are close to those found in
the atmosphere. However, our study does not take into
3.4. Photodegradation of major products
The photodegradation of the main intermediate
products of MTBE is interesting for two reasons.
Firstly, some compounds may be more stable and/or
more toxic than the initial pollutant. Thus, they can also
be found in atmospheric water droplets and it is there-
fore important to know their phototransformations.
Secondly, the knowledge of their photooxidation mech-
anism may help us to propose a more detailed MTBE
degradation scheme.
account the night period when NO
most numerous species.
3
° radicals are the
3
.2. Identification of degradation intermediate products
Without concentrating the products in the gaseous
phase, only MTBE and the major degradation products,
identified by their retention times, as TBF, MA, AC and
TBA, are detected by on-line GC/FID analysis. TBF is
the only one confirmed by GC/MS analysis. Therefore,
to confirm the formation of MA, AC and TBA, and to
identify other intermediate products, different concen-
tration methods, adsorbent cartridges and SPME fibres
were used. Several additional products of the degrada-
tion of MTBE have been identified. Results from these
We studied the photodegradation of AC, TBF and
TBA under the same conditions as those used for the
degradation of MTBE. Initial concentrations of pollu-
ꢀ1
tants were 1000 lmol l . No method for concentrating
products was used because only the main products were
studied in this experiment. The reactivity pattern was
found to be the following:

Table 1
Sump up of the intermediate products of MTBE identified by using different methods
Techniques Column
Intermediate products
TBF
AC
TBA
Methyl Acetic
acetate acid
Formic Formal- Acetal- Pro-
acid
Iso- Meth-
butene ane
Ethane
Iso-
butane
Tertiobutyl
peroxide
dehyde
dehyde
panal
GC/MS
Gas valve
Gas
syringe
CPWax
57CB
X
X
X (a,b)
X (a,b)
X (c,d,e) X (c,d,e) X (d)
X (a,b)
X (a,b)
X (a,b) X (a,b)
SPME
Vocol
CPSil
X (a,b) X (a,b) X (a,b) X (b)
X (c,d)
X (b)
5CB
Active
carbon
CPWax
57CB
Vocol
X
X
X
X
X
X
X
X
X
X
Piperidine
CPSil
5CB
X
X
N
2
trap
Vocol
/
X
MS
HPLC/UV CAR-H
X
X
X
X
X
HPLC/
diode
array
Nova-Pak
C18
X
X
X
X
Sperisorb
ODS2
X
X
GC/FID
Vocol
GS-Q
X
X
X
X
X
X
X
X
X
X
X
X
(a) PDMS 100 lm, (b) Carbowax divinylbenzene, (c) polyacrylate, (d) PDMS/DVB 65 lm and (e) PDMS 30 lm.

474
C. Guillard et al. / Chemosphere 53 (2003) 469–477
80
MTBE
t-Butylformate TBF
methyl acetate MA
acetone AC
70
ter- Butylformate TBF
Acetone AC
6
4
2
0
0
0
0
6
5
0
0
ter-Butyl Alcohol TBA
t-Butyl Alcohol TBA
formaldehyde
40
3
2
1
0
0
0
0
0
100
200
300
400
500
(
a)
t (min)
0
100
200
300
T (min)
400
500
(a)
8
0
1
00
70
ter-Butyl Alcohol TBA
Acetone AC
6
5
0
0
8
6
4
2
0
0
0
0
0
40
3
2
1
0
0
0
0
0
20
40
60
80
100
t (min)
120
140
160
180
(b)
0
100
200
300
t (min)
400
500
(
b)
Fig. 3. Evolution of main intermediate compounds of (a) TBF
and (b) AC by photo-Fenton.
Fig. 2. (a) Evolution of intermediate products of MTBE as a
function of time and (b) calculated percentage of organic car-
bon coming from the main intermediate products.
Fig. 3a shows that TBF is not quantitatively con-
verted to TBA. Acetone from TBA only accounts for
about one-half of the original TBF, that is to say the
formation of other compounds. MA and 2-methylpro-
pene might account for the rest of the TBF. Acetone
becomes stable only when TBA has totally disappeared.
The data shown in Fig. 3b indicates that AC is a product
of TBA oxidation and perhaps comes from TBA alone.
The reaction appears to be nearly quantitative with only
a slight side reaction.
ꢀ
1
ꢀ1
ꢀ1
k
ðTBAÞ0:028 min > kðMTBEÞ0:007 min > kðTBFÞ0:03 min
ꢀ
1
> kðACÞ0:002 min
This reactivity order is in accordance with the OH°
radical reaction. Indeed, the reaction rate constants of
9
OH° radical with MTBE and AC, respectively 1.6 ꢂ 10
8
ꢀ1 ꢀ1
s according to the Buxton values
and 1.1 ꢂ 10 l mol
(
1988), are in the same ratio (about 14) as those obtained
The photo-Fenton reaction of AC gives mainly acetic
acid by OH° attack on methyl groups. Besides these
main intermediate products, isobutane, 2-methylpro-
pene, formaldehyde, formic acid, methane and ethane
also account, to a smaller extent, in TBF and TBA de-
gradation. These last two compounds have also been
detected in the degradation of AC.
under our conditions. The rate constants of OH° reaction
with TBA and TBF are not listed in Buxtonꢀs data table.
However, the reactivity order observed fit well with the
order found for OH° reactions with alcohols, ethers and
esters. For example, the rate constant for butan-1-ol,
diethylether and ethyl acetate are listed below:
k
4
OH°ðbutan-1-olÞ > kOH°ðdiethyletherÞ > kOH°ðethyl acetateÞ
9
8
8
ꢀ1 ꢀ1
:2 ꢂ 10
9:5 ꢂ 10
4 ꢂ 10 in lmol
s
3.5. Degradation scheme
The smallest reactivity of AC with an OH° radical can
explain the relative stability of this compound when
MTBE is totally degraded. It can also be concluded that
the release of MTBE into the atmosphere contribute to
the formation of AC which is found as one of the main
atmospheric pollutant. The average atmospheric con-
centration of AC over one year is about 6.1 ppbv (La-
Grone, 1991). Moreover, near some industrial sites, the
AC concentration reaches 94 ppbv (Kelly et al., 1993).
MTBE contains methyl groups from which a hy-
drogen atom can be abstracted by an OH° radical. The
abstraction is much more likely to occur from the meth-
oxy group than from one of the methyl groups because
the bond dissociation energy for a C–H bond whose
carbon atom is attached to an ether oxygen atom is
lowered by 12.6 ꢁ 8.4 kJ molꢀ with respect to a C–H
bond in an alkane, the decrease being the greatest for a
1
ꢀ
1
CH group, e.g., 21 kJ mol between dimethyl ether and
3

C. Guillard et al. / Chemosphere 53 (2003) 469–477
475
ethane (Mc Millen and Golden, 1982). Therefore the
reaction that initiates the MTBE degradation is mainly
the H-abstraction on the methoxy group (Eq. (4))
ferrous iron in acidic conditions following Eqs. (10)
and (11).
2
þ
ðCH
3
Þ3C–O–CH
½ðCH Þ3C–O–CH
Þ3C–O–CH
ðCH
2
O
2
° þ Fe
ðCH
3
Þ3C–O–CH
3
þ OH° ! ðCH
3
Þ3C–O–CH
2
° þ H
2
O
O–O–Feꢃ2þ
!
3
2
ð10Þ
ð11Þ
ð4Þ
½
ðCH
3
2
O–O–Feꢃ þ þ Hþ
2
The resulting alkyl radical easily adds O
peroxo-radical (Eq. (5)). To a lesser extent the radical
2
to form the
3þ
!
3
Þ3C–O–CH
2
O–OH þ Fe
transforms into formaldehyde and t-butyl radical (Eq.
(6)). This radical is at the origin of the production of
isobutane and/or 2-methylpropene (Eqs. (7) and (8)).
(
ii) The peroxo-radical can form a tetroxyde as pro-
posed by von Sonntag (1991) with other organic
compounds and form TBF whatever the pathways
suggested (Scheme 1). Since the hemiacetal was not
detected, pathways (a) and (d) seem less likely to
take place, though under acidic conditions hemiac-
etal should be rapidly hydrolysed. Pathway (b) can
occur; however, it leads only to TBF, unlike path-
way (c) through which are produced not only
TBF, but also formaldehyde and the radical t-but-
oxy. This radical could be at the origin of the forma-
tion of t-butyl peroxyde, TBA, AC and isobutane
°
ðCH
ðCH
ðCH
ðCH
3
Þ C–O–CH þ O
2
! ðCH
3
Þ3C–O–CH
2
OO°
ð5Þ
ð6Þ
ð7Þ
ð8Þ
3
2
°
2
3
3
3
Þ C–O–CH ! ðCH
3
3
Þ3C° þ HCHO
Þ3C° þ MTBE ! ðCH
3
2
Þ3CH þ MTBE°
Þ3C° ! ðCH
3
Þ2C@CH
þ H°
Two pathways are proposed to form the TBF which is
the main intermediate:
(
Eqs. (12)–(15)). Therefore, the pathway (c) seems
to be more credible than the other ones. The data
shown in Fig. 2a in combination with that in Fig.
(
i) The peroxo-radical can react directly with MTBE by
abstraction of an H atom to form an hydroperoxide
Eq. (9) (Faust and Allen, 1992).
3
agree with this hypothesis, as a matter of fact these
data suggest that TBA and TBF are formed in par-
allel from MTBE.
ðCH
3
Þ C–O–CH
2
O
2
° þ ðCH
3
Þ3C–O–CH
2
–O–OH
3
3
!
ðCH
3
Þ C–O–CH
3
The t-butyl peroxyde could be formed by the cou-
þ ðCH
3
Þ C–O–CH
2
°
ð9Þ
3
pling of two t-butoxy radical [Eq. (12)]. This radical
could also abstract a hydrogen atom from a MTBE
molecule to form TBA [Eq. (13)] or decompose into AC
and a methyl radical [Eq. (14)]
Then, the hydroperoxide, in acidic conditions, forms
TBF (Behar et al., 1973) which can be easily hydro-
lysed into TBA (Church et al., 1999). In oxidant
conditions t-butoxy radical can be obtained. Hy-
droperoxide can also result from the reaction with
2
ðCH
3
Þ3C–O° ! ðCH
3
Þ3C–O–O–CðCH
3
Þ3
ð12Þ
2
(CH
3
)
3
C O CH2
O
O°
(CH ) C
O
CH2
O
O
O
O
CH2
O
C(CH3)3
3 3
(
a) - O
2
(b) - H
2
O
2
(c) - O
2
(d) - O
2
O
O
H
O
2
(CH
3
)
3
C
O
C
2 (CH ) C
O
C
+
+
2 H°
3 3
2 (CH ) C
O CH
2
O°
(
3 3
CH ) C
O C
3 3
H
H
+
and/or
O
(
CH
3
)
3
C
O
CH OH
+ 2 H
C
3 3
2 (CH ) C
O°
2
H
O
(
CH
3
)
3
C
O
+
CH OH
(CH
)
C
O
O
CH° OH
2 (CH
)
C
O
C
+ 2 H°
3 3
(CH ) C
2
O CH2 O C(CH3)3
O O
CH
2
3
3
3
3
H
and/or
2
2
- 2 HO
2
°
O
C
H
O
H
(
CH
3
)
3
C
O
O
+
2 H
C
+ 2 (CH
) C
O°
3 3
2
3 3
(CH ) C
O C
H
Scheme 1. Degradation of MTBE following the von Sonntag mechanism (von Sonntag, 1991).

4
76
C. Guillard et al. / Chemosphere 53 (2003) 469–477
H-atom abstraction. Due to the electronegativity of the
ðCH
3
Þ C–O° þ ðCH
3
Þ C–O–CH
3
3
3
!
ðCH
3
Þ C–OH þ ðCH
3
Þ3C–O–CH
2
°
ð13Þ
oxygen atom the main abstraction occurs at the methoxy
side of MTBE. However as found by Stefan et al. (2000),
the abstraction of an H-atom of the t-butyl group also
contributes to the MTBE degradation mechanism due to
the presence of nine abstractable hydrogens in the
t-butyl group. Then, the radicals formed by H-atom
abstraction react either with an oxygen molecule or de-
compose. The main intermediate formed is TBF which
produces TBA, formic acid and AC. This last compound
is slowly degraded into acetic acid. To a lesser extent,
hydrocarbons such as isobutane, butene, methane and
ethane are detected. Therefore, even if the use of MTBE
3
Eq. (14) shows the formation of AC from a t-butoxy
radical as suggested by several authors (Cox and
Goldstone, 1982; Batt et al., 1989)
ðCH
3
Þ3C–O° ! CH
3
COCH
3
þ CH
3
°
ð14Þ
Although, most of the intermediate products can be
explained by the von Sonntag mechanism (von Sonntag,
1
991) (Scheme 1, pathway c), other pathways can also
explain the formation of these intermediate compounds.
Indeed, TBA is the main product of the TBF degrada-
tion. It could come from TBF hydrolysis in acidic con-
ditions according to Eq. (15). However, acid hydrolysis
is far too slow to yield the rate of conversion of TBF to
TBA, as a matter of fact the half-life of TBF at pH 3 is
x
as a fuel additive reduces the CO, NO and hydrocarbon
emissions harmful to the environment, this additive
nevertheless represents a risk when it is emitted into the
troposphere. Indeed AC, a molecule which is very stable
in atmospheric conditions, is formed in water droplets.
Moreover, the formation of acetic and formic acid
contributes to the acidification of the water droplets.
Therefore it is recommended that this risk be avoided by
decreasing MTBE emissions from adapted devices.
2
.2 days (Church et al., 1999).
þ
ðCH
Þ3C–O–CHO þ H
3 3
O
!
ðCH
3
Þ3C–OH þ H–COOH þ Hþ
ð15Þ
The t-butyl radical can be obtained from TBA by
H-abstraction. Therefore, t-butyl peroxyde, AC and
isobutane can also originate from TBF following Eqs.
Acknowledgements
(
12)–(14). The hypothesis is in agreement with the de-
The financial support to N.C. from the French
Ministry of Research and Technology is gratefully ac-
knowledged.
tection of these different intermediate products from the
degradation of TBF.
Acetone can also come from H-abstraction on t-butyl
group as proposed by Smith (13) and summed up in Eqs.
References
(
17)–(21). This pathway also explains the formation of
MA from MTBE as detected in our experimental con-
ditions.
Acero, J.L., Haderlein, S.B., Schmidt, T.C., Suter, M.J.-F., von
Gunten, U., 2001. MTBE oxidation by conventional ozo-
nation and the combination ozone: hydrogen peroxide:
Efficiency of the processes and bromate formation. Environ.
Sci. Technol. 35, 4252–4259.
ðCH
3
Þ C–O–CH
3
þ OH°
3
!
°CH
2
ðCH
3
Þ C–O–CH
3
þ H
2
O
ð16Þ
ð17Þ
2
Anastasio, C., McGregor, K.G., 2001. Chemistry of fog waters
in Californiaꢀs Central Walley: 1. In situ photoformation of
hydroxyl radical and singlet molecular oxygen. Atmos.
Environ. 35, 1079–1089.
2
°CH
!
2
ðCH
3
Þ C–O–CH
3
þ O
2
2
2°OOCH
2
ðCH
3
Þ C–O–CH
3
2
°
OCH
2
ðCH
3
Þ2C–O–CH
3
! ðCH
3
Þ2C°–O–CH
3
þHCHO
ð18Þ
Atkinson, R., 1989. Kinetics of hydroxyl radical reactions.
J. Phys. Chem. Ref. Data, Monograph 1, 76–81.
Batt, L., Hisham, M.W.M., Mackay, M., 1989. Decomposition
of the t-Butoxy Radical: II. Studies over the Temperature
Range 303–393 K. Int. J. Chem. Kinet. 21, 535–545.
Behar, D., Samuni, A., Fessenden, R.W., 1973. Reactions of
Iron (II) and Titanium (III) with organic radicals. J. Phys.
Chem. 7, 2055–2059.
2
ðCH
3
Þ2C°–O–CH
3
þ O
2
! 2ðCH
3
Þ2CO°–O–CH
3
ð19Þ
ð20Þ
ðCH
3
Þ2CO°–O–CH
3
! ðCH
3
Þ2CO þ CH
3
O°
&
Buxton, G.V., Greenstock, C.L., Helman, W.P., Ross, A.B.,
1988. Critical review of rate constants for reactions of
hydrated electrons, hydrogen atoms and hydroxyl radicals
CH
3
Cð¼ OÞ–O–CH
3
þ CH
3
°
ð21Þ
ꢀ
(
.OH/.O ) in aqueous solution. Phys. Chem. Ref. Data 17,
13–780.
The formation of methane and ethane can easily be
accounted for by the reactions of methyl radicals with
each other or with hydrogen atoms. Acetic acid is pro-
duced from AC as proposed by Stefan et al. (1996).
To sum up, our research showed that the degradation
of MTBE in the atmospheric water droplets occurs by
5
Cater, S.R., Stefan, M.I., Bolton, J.R., 2000. UV: H
2
O
2
treatment of methyl tert-butyl ether in contaminated waters.
Environ. Sci. Technol. 34, 659–662.
Church, C.D., Pankow, J.F., Tratnyek, P.G., 1999. Hydrolysis
of tert-butyl formate: kinetics, products, and implications

C. Guillard et al. / Chemosphere 53 (2003) 469–477
477
for the environmental impact of methyl tert-butyl ether.
Environ. Toxicol. Chem. 18 (12), 2789–2796.
for the reactions of hydroxyl radicals and chlorine atoms
with a series of aliphatic alcohols and ethers at 298 K. Int. J.
Chem. Kinetics 22, 1111–1126.
Cox, R.A., Goldstone, A., 1982. Proceedings of the 2nd
European Symposium on the Physico Chemical Behavior
of Atmospheric Pollutants. Riedel, Dorecht, Holland. p. 112.
Dagaut, P., Liu, R., Wallington, T.J., Kurylo, M.J., 1989. The
gas phase reactivity of aliphatic polyethers towards OH
radicals: measurements and predictions. Int. J. Chem.
Kinetics 21, 1173–1180.
Pawliszyn, J. (Ed.), 1997. Solid Phase Microextraction, Theory
and Practice. Wiley-VCH, New York.
Safarzadeh-Amiri, A., Bolton, J.R., Cater, S.R., 1997. Ferriox-
alate-mediated photodegradation of organic pollutants in
contaminated water. Water Res. 31 (4), 787–798.
Smith, D.F., Kleindienst, T.E., Hudgens, E.E., Mc Iver, C.D.,
Bufalini, J.J., 1991. The photooxidation of methyl tertiary
butyl ether. Int. J. Chem. Kinetics 23, 907–924.
Faust, B.C., Allen, J.M., 1992. Aqueous-phase photochemical
sources of peroxyl radicals and singlet molecular oxygen in
clouds and fog. J. Geophys. Res. 97, 12913–12926.
Grosjean, E., Rasmussen, R.A., Grosjean, D., 1998. Ambiant
levels of gas phase pollutants in Porto Alegre, Brazil.
Atmos. Environ. 32 (20), 3371–3379.
Squillace, P.J., Pankow, J.F., Korte, N.E., Zogorski, J.S., 1997.
Review of the environmental behavior and fate of methyl
tert-butylether. Environ. Toxicol. Chem. 16 (9), 1836–1844.
Stefan, M.I., Hoy, A.R., Bolton, J.R., 1996. Kinetics and
mechanism of the degradation and mineralization of
acetone in dilute aqueous solution sensitized by the UV
photolysis of hydrogen peroxide. Environ. Sci. Technol. 30
(7), 2382–2390.
Hoigne, J., Sedlak, D.L., B u€ hler, R.E., 1994. EUROTRACK
Conference Proceedings. Leipzig, Germany.
Japar, S.M., Wallington, T.J., Richert, J.F.O., Ball, J.C., 1990.
The atmospheric chemistry of oxygenated fuel additives:
t-butyl alcohol, dimethyl ether, and methyl t-butyl ether.
Int. J. Chem. Kinetics 22, 1257–1269.
Stefan, M.T., Mack, J., Bolton, J.R., 2000. Degradation
pathways during the treatment of methyl tert-butyl ether
by the UV/H
Karpel Vel Leitner, K., Papailhou, A.L., Crou ꢁe , J.P., Peyret, J.,
Dor ꢁe , M., 1994. Oxidation of methyl tert-butyl ether
2 2
O process. Environ. Sci. Technol. 34, 650–
658.
Tuazon, E.C., Carter, W.P.L., Aschmann, S.M., Atkinson, R.,
1991. Products of the gas-phase reaction of methyl tert-
(
MTBE) and ethyl tert-butyl ether (ETBE) by ozone and
combined ozone: hydrogen peroxide. Ozone Sci. Technol.
6, 41–54.
1
butyl ether with the OH radical in the presence of NO
J. Chem. Kinetics 23 (11), 1003–1015.
x
. Int.
Kelly, T.J., Callatran, P.J., Piell, J., Evans, G.F., 1993. Method
development and field measurements for polar volatile
organic compounds in ambient air. Environ. Sci. Technol.
Vainiotalo, S., Peltonen, Y., Pf €a ffli, P., 1998. MTBE concen-
trations in ambient air in the vicinity of service stations.
Atmos. Environ. 32 (20), 3503–3509.
27, 1146–1153.
LaGrone, F.S., 1991. Potential community exposure to toxic
chemicals using volatile organics measurements for assessing
community air quality. Environ. Sci. Technol. 25, 366–368.
Liang, S., Palencia, L.S., Yates, R.S., Davis, M.K., Bruno,
J.-M., Wolfe, R.L., 1999. Oxidation of MTBE by ozone and
peroxone processes. J. AWWA 91 (6), 104–114.
von Sonntag, C., Schuchmann, H.P., 1991. The elucidation of
peroxyl radical reactions in aqueous solution with the help
of radiation-chemical methods. Angew. Chem., Int. Ed.,
Engl. 30, 1229–1253.
Wallington, T.J., Liu, R., Dagaut, P., Kurylo, M.L., 1988a. The
gas phase reactions of hydroxyl radicals with a series of
aliphatic ethers over the temperature range 248–440 K. Int.
J. Chem. Kinetics 20, 41–49.
Mc Millen, D.F., Golden, D.M., 1982. Hydrocarbon bond
dissociation energies. Ann. Rev. Phys. Chem. 33, 493–532.
Miyake, T., Shibamoto, T., 1999. Formation of formaldehyde
from methyl tert-butyl ether (MTBE) upon UV irradiation.
Bull. Environ. Contam. Toxicol. 62, 416–419.
Wallington, T.J., Dagaut, P., Liu, R., Kurylo, M.J., 1988b.
Gas-phase reactions of hydroxyl radicals with the fuel
additives methyl tert-butyl ether and tert-butyl alcohol over
the temperature range 240–440 K. Environ. Sci. Technol.
22, 842–844.
Nelson, L., Rattigan, O., Neavyn, R., Sidebotton, H., Treacy,
J., Nielsen, O., 1990. Absolute and relative rate constants