Products and mechanism of the reaction of OH radicals with 2,2,4-trimethylpentane in the presence of NO

Article abstract of DOI:10.1021/es011134v

Alkanes are important constituents of gasoline fuel and vehicle exhaust, with branched alkanes comprising a significant fraction of the total alkanes observed in urban areas. The products and mechanism of the OH radical-initiated reaction products and mec

Full text of DOI:10.1021/es011134v

Environ. Sci. Technol. 2002, 36, 625-632
organic compounds measured in ambient air in urban areas
2, 3). In the troposphere, alkanes present in the gas phase
Products and Mechanism of the
Reaction of OH Radicals with
(
react mainly with the OH radical (4) to form alkyl radicals,
•
which react rapidly with O
4):
2
to form alkyl peroxy (RO
2
) radicals
(
2
,2,4-Trimethylpentane in the
•
Presence of NO
OH + RH f H O + R
(1)
(2)
2
•
•
S A R A M . A S C H M A N N ,
J A N E T A R E Y, * , A N D
R O G E R A T K I N S O N * ,
R + O f RO
2
2
†
† , ‡
In urban areas in the presence of sufficiently high concen-
Air Pollution Research Center, University of California,
Riverside, California 92521
trations of NO, alkyl peroxy radicals react with NO to form
the alkyl nitrate, RONO , or the corresponding alkoxy radical
2
2
plus NO (4):
RO2• + NO M
RO2• + NO f RO + NO2
9
8 RONO2
(3a)
(3b)
Alkanes are important constituents of gasoline fuel and
vehicle exhaust, with branched alkanes comprising a
significant fraction of the total alkanes observed in urban
areas. Products of the gas-phase reactions of OH radicals
with 2,2,4-trimethylpentane and 2,2,4-trimethylpentane-d18 in
the presence of NO at 298 ( 2 K and atmospheric
•
2
The alkoxy radicals formed in reaction 3b then react with O ,
decompose by C-C bond scission, or isomerize through a
six-membered ring transition state (4). The isomerization
pressure of air have been investigated using gas chroma-
tography with flame ionization detection (GC-FID),
4
reaction, which can occur for gC alkanes, forms a 1,4-
hydroxyalkyl radical that adds O , giving a 1,4-hydroxyalkyl
2
combined gas chromatography-mass spectrometry (GC-
MS), and in situ atmospheric pressure ionization tandem
mass spectrometry (API-MS). Acetone, 2-methylpropanal,
and 4-hydroxy-4-methyl-2-pentanone were identified and
quantified by GC-FID from 2,2,4-trimethylpentane with molar
formation yields of 54 ( 7%, 26 ( 3%, and 5.1 ( 0.6%,
respectively; upper limits to the formation yields of
acetaldehyde, 2,2-dimethylpropanal, and 4,4-dimethyl-2-
pentanone were obtained. Additional products observed
from 2,2,4-trimethylpentane by API-MS and API-MS/MS
analyses using positive and negative ion modes were
hydroxy products of molecular weight 130 and 144, a product
of molecular weight 128 (attributed to a C8-carbonyl),
and hydroxynitrates of molecular weight 135, 177, and 191
peroxy radical. The 1,4-hydroxyalkyl peroxy radical then
reacts further by reactions analogous to reactions 3a and 3b
and leads to the formation of hydroxycarbonyls and hy-
droxynitrates (4-8).
While rate constants for the reactions of the OH radical
with a large number of alkanes have been measured (4), the
products and mechanisms of the OH radical-initiated reac-
tions of alkanes have received less attention. Most of the
more recent and comprehensive studies of the products and
mechanisms of the OH radical reactions of alkanes have
involved n-alkanes and cycloalkanes (4-10), with the excep-
tion of a recent study that included 3,4-diethylhexane (10).
However, branched alkanes are important, accounting for
∼
30-50% of the alkanes other than methane in urban areas
(
1, 2). Accordingly, in this work we have investigated the
products and mechanism of the OH radical-initiated reaction
of 2,2,4-trimethylpentane (isooctane) in the presence of NO.
In addition to the use of gas chromatography for product
analyses, we have used atmospheric pressure ionization
tandem mass spectrometry to identify reaction products not
amenable to gas chromatography. 2,2,4-Trimethylpentane-
d18 was also used in the experiments with API-MS analyses
to aid in the interpretation of the API-MS spectra.
(attributed to HOC4H8ONO2, HOC7H14ONO2, and HOC8H16-
ONO2, respectively). Formation of HOC8H16ONO2 and HOC7H14-
ONO2 is consistent with the observation of products of
molecular weight 207 (HOC8D16ONO2) and 191 (HOC7D14-
ONO2), respectively, in the API-MS analyses of the 2,2,4-
trimethylpentane-d18 reaction (-OD groups rapidly exchange
to -OH groups under our experimental conditions). These
product data allow the reaction pathways to be delineated
to a reasonable extent, and the reaction mechanism is
discussed.
Experimental Methods
Experiments were carried out at 298 ( 2 K and 740 Torr total
pressure of air (at ∼5% relative humidity, RH) in a 7900-L
Teflon chamber with analysis by gas chromatography with
flame ionization detection (GC-FID) and combined gas
chromatography-mass spectrometry (GC-MS), with ir-
radiation provided by two parallel banks of blacklamps, and
in a 7300-L Teflon chamber interfaced to a PE SCIEX API III
MS/ MS direct air sampling, atmospheric pressure ionization
tandem mass spectrometer (API-MS), again with irradiation
provided by two parallel banks of blacklamps. Both chambers
are fitted with Teflon-coated fans to ensure rapid mixing of
reactants during their introduction into the chamber.
Hydroxyl radicals were generated in the presence of NO
by the photolysis of methyl nitrite in air at wavelengths >300
nm (6-10), and NO was added to the reactant mixtures to
Introduction
Alkanes are important constituents of gasoline fuel and
vehicle exhaust (1) and account for ∼50% of the non-methane
*
Address correspondence to either author. J.A. phone: (909)-
7
87-3502; fax: (909)787-5004; e-mail: janet.arey@ucr.edu. R.A.
phone: (909)787-4191; fax: (909)787-5004; e-mail: roger.atkinson@
ucr.edu.
†
Also at Interdepartmental Graduate Program in Environmental
Toxicology and Department of Environmental Sciences, University
of California, Riverside.
‡
Also at Department of Chemistry, University of California,
Riverside.
3 3
suppress the formation of O and hence of NO radicals.
1
0.1021/es011134v CCC: $22.00

2002 Am erican Chem ical Society
VOL. 36, NO. 4, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6 2 5
Published on Web 01/03/2002

-
or [PFBOH‚O
2
]-), where M
Analyses by GC-FID and GC-MS. For the reactions
carried out in the 7900-L Teflon chamber (at ∼5% RH), the
(M) and the reagent ion (NO
2
can also be PFBOH when PFBOH was added to the sampled
airstream:
-3
initial reactant concentrations (in molecule cm units) were
1
4
14
as follows: CH
3
ONO, (2.2-2.4) × 10 ; NO, (1.9-2.2) × 10 ;
13
and 2,2,4-trimethylpentane, (2.24-3.02) × 10 . Irradiations
-
-
[
PFBOH‚O ] + M f [PFBOH‚O ‚M]
2
2
were carried out at 20% of the maximum light intensity for
1
0-45 min, resulting in up to 23% reaction of the initially
-
-
NO2 + M f [NO ‚M]
present 2,2,4-trimethylpentane. The concentrations of 2,2,4-
trimethylpentane and reaction products were measured
during the experiments by GC-FID. Gas samples of 100 cm3
volume were collected from the chamber onto Tenax-TA
solid adsorbent, with subsequent thermal desorption at ∼250
2
Previous work in this laboratory (8, 10) indicated that the use
-
-
of NO2 or [PFBOH‚O ] reagent ions allows primarily hydroxy
2
compounds to be detected (for example, hydroxycarbonyls
-
°
C onto a DB-1701 megabore column in a Hewlett-Packard
and hydroxynitrates). When NO2 was used as the reagent
(
HP) 5710 GC, initially held at -40 °C and then temperature
ion, quantification of hydroxycarbonyls was carried out (8)
by adding a measured amount of 5-hydroxy-2-pentanone to
the chamber after the irradiation as an internal standard
-1
programmed to 200 °C at 8 °C min . In addition, gas samples
were collected onto Tenax-TA solid adsorbent for GC-MS
analyses, with thermal desorption onto a 60-m HP-5 fused
silica capillary column in an HP 5890 GC interfaced to a HP
-
and assuming that the intensities of the [NO ‚M] ion peaks
2
were proportional to the concentrations of the hydroxycar-
5
970 mass selective detector operating in the scanning mode.
GC-FID response factors were determined as described
previously (11). NO and initial NO concentrations were
measured using a Thermo Environmental Instruments, Inc.,
model 42 NO-NO -NO chemiluminescence analyzer.
bonyls, M, present in the chamber (8).
The initial concentrations of CH ONO, NO, and 2,2,4-
3
2
trimethylpentane (or 2,2,4-trimethylpentane-d ) were ∼(2.4-
18
4.8) × 1013 molecule cm-3 each, and irradiations were carried
2
x
out at the maximum light intensity for 1-3 min (resulting
in up to ∼15% reaction of the initially present 2,2,4-
trimethylpentane).
Analyses by API-MS. In the experiments with API-MS
analyses, the chamber contents were sampled through a 25
-
1
mm diameter × 75 cm length Pyrex tube at ∼20 L min
Chem icals. The chemicals used and their stated purities
were as follows: 4,4-dimethyl-2-pentanone (99%), 2,2-
dimethylpropanal (97%), 4-hydroxy-4-methyl-2-pentanone
(99%), 2-methylpropanal (99+%), octanal (99%), penta-
fluorobenzyl alcohol (98%), and 2,2,4-trimethylpentane
(99%), Aldrich Chemical Co.; 5-hydroxy-2-pentanone, TCI
America; 2,2,4-trimethylpentane-d18 (98% atom D), C/ D/ N
Isotopes, Inc.; 3-octyl nitrate, Fluorochem, Inc.; and NO
(g99.0%), Matheson Gas Products. Methyl nitrite was
prepared as described previously (12) and stored at 77 K
under vacuum.
directly into the API mass spectrometer source. The operation
of the API-MS in the MS (scanning) and MS/ MS [with
collision-activated dissociation (CAD)] modes has been
described elsewhere (6-10). Use of the MS/ MS mode with
CAD allows the “product ion” or “precursor ion” spectrum
of a given ion peak observed in the MS scanning mode to be
obtained (6-10). Both positive and negative ion modes were
used in this work.
In the positive ion mode, protonated water clusters,
+
H
3
O (H
2
O)
n
, formed from a corona discharge in the chamber
diluent air (at ∼5% RH) are the reagent ions, and a range of
oxygenated species can be observed in this mode of operation
Results
(
6-9). Ions are drawn by an electric potential from the ion
source through the sampling orifice into the mass-analyzing
first quadrupole or third quadrupole. Neutral molecules and
particles are prevented from entering the orifice by a flow
of high-purity nitrogen (“curtain” gas), and as a result of the
declustering action of the curtain gas on the hydrated ions,
the ions that are mass-analyzed are mainly protonated
molecules ([M + H] ) and their protonated homo- and
heterodimers. Product peaks were identified based on the
observation of homo- or heterodimers (for example, [(MP1)
2
GC-FID and GC-MS Analyses. GC-FID and GC-MS analyses
of irradiated CH ONO-NO-2,2,4-trimethylpentane-air mix-
3
tures showed the formation of acetone, 2-methylpropanal,
and 4-hydroxy-4-methyl-2-pentanone. As discussed below,
consideration of the potential reactions involved indicated
that a number of other product species that can be observed
by gas chromatography could be formed, including acet-
aldehyde, 2,2-dimethyl-propanal, and 4,4-dimethyl-2-pen-
tanone. However, GC-FID analyses showed no evidence for
the formation of these three potential products, and upper
limits to their concentrations were obtained. Secondary
reactions of acetone, 2-methylpropanal, and 4-hydroxy-4-
methyl-2-pentanone (and of the potential products acet-
aldehyde, 2,2-dimethylpropanal, and 4,4-dimethyl-2-pen-
tanone) with the OH radical during these experiments were
taken into account as described previously (13) using rate
+
+
+
+
+
2
H] , [(MP2) + H] , and [MP1 + MP2 + H] , where P1 and
P2 are products) in the API-MS/ MS mode in precursor ion
spectra of the [M
consistency of the API-MS/ MS product ion spectrum of a
homo- or heterodimer ion with the precursor ion spectra
+
P
+ H] ion peaks and confirmed by
(
9).
In the negative ion mode, negative ions were generated
by the negative corona around the discharge needle. The
-
12
constants for reactions of the OH radical (in units of 10
-
-
cm3 molecule
-1
-1
superoxide ion (O
2
), its hydrates [O
2
(H
2
O)
n
] , and O
2
clusters
s ) of 2,2,4-trimethylpentane, 3.57 (4);
-
[O
2
(O
2
)
n
] are the major reagent negative ions in the chamber
acetaldehyde, 15.8 (14), acetone, 0.219 (14); 2-methylpro-
panal, 26.3 (14); 4-hydroxy-4-methyl-2-pentanone, 4.0 (15);
2,2-dimethylpropanal, 26.5 (14); and 4,4-dimethyl-2-pen-
tanone, 1.47 [estimated (16)]. The multiplicative factors F to
take into account the secondary reactions of OH radicals
with the products increase with the rate constant ratio k(OH
+ product)/ k(OH + 2,2,4-trimethylpentane) and with the
extent of reaction (13), and the maximum values of F were
1.01 for acetone, 1.74 for acetaldehyde, 2.38 for 2-methyl-
propanal and 2,2-dimethylpropanal, 1.16 for 4-hydroxy-4-
methyl-2-pentanone, and 1.06 for 4,4-dimethyl-2-pentanone.
The formation yields (or upper limits thereof) of these
products obtained by least-squares analyses of the data
-
-
3
and NO ,
pure air. Other reagent ions, for example, NO
are formed through reactions between the primary reagent
ions and the neutral molecules such as NO . Instrument
2
2
tuning and operation were designed to induce adduct
formation. Two species served as reagent ions in this work;
-
NO
2
present in the irradiated CH
3
ONO-NO-2,2,4-tri-
-
methylpentane-air mixtures and [PFBOH‚O
pentafluorobenzyl alcohol, C CH
2
] (where PFBOH
)
6
F
5
2
OH) generated by
addition of pentafluorobenzyl alcohol to the sampled air-
stream from the chamber by diffusion through a pinhole in
a covered, heated vial containing PFBOH (8). Analytes were
then detected as adducts formed between the neutral analyte
6
2 6
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 4, 2002

TABLE 1. Products Observed and Their Molar Formation Yields, from the OH Radical-Initiated Reactions of 2,2,4-Trimethylpentane
(TMP) and 2,2,4-Trimethylpentane-d18 (TMP-d18) in the Presence of NO
API-MS
TMP negative ion mode
positive ion mode
product
GC-FIDa
TMP
TMP-d18
PFBOH addition
NO2- adduct
acetone
0.58 ( 0.13b
b
0
.53 ( 0.07c
obsd
obsd
nd
nd
nd
0
.54 ( 0.07
b
2
4
-m ethylpropanal
0.28 ( 0.05
0
0
b
.25 ( 0.03c
obsd
obsd
nd
.26 ( 0.03
-hydroxy-4-m ethyl-2-pentanone (MW 116)
0.051 ( 0.006
obsd
obsd
obsd
nd
obsd
obsd
obsd
nd
obsd
nd
obsd
obsd
no db sd, yield ∼0.08
d
MW 128 product (CH3)3CCH2CH(CH3)CHO
MW 130 product HOCH2C(CH3)2CH2C(O)CH3
MW 135 product C4-hydroxyalkyl nitrate
MW 144 product (CH3)2C(OH)CH2C(CH3)2CHO
MW 177 product C7-hydroxyalkyl nitrate
d
obsd, yield ∼0.03
d
obsd
d
obsd
obsd
obsd
obsd
obsde
obsde
obsd (weak)
obsd
obsd
obsd, yield ∼0.03
obsd
obsd
d
d
MW 191 product C8-hydroxyalkyl nitrate
potential product
acetaldehyde
<0.04
2
4
,2-dim ethylpropanal
,4-dim ethyl-2-pentanone
<0.013
<0.004
a
Indicated errors are two least-squares standard deviations com bined with estim ated overall uncertainties in the GC-FID response factors for
b
c
2
,2,4-trim ethylpentane and products of (5% each. obsd ) observed. nd ) not detected. Independent determ inations. Weighted average.
Attributed, see text and Table 2. e Observed as dim ers with acetone-d
d
6
.
Positive Ion Mode. API-MS/ MS product ion and precur-
sor ion spectra were obtained for ion peaks observed in the
API-MS analyses of the products of the reactions of OH
radicals with 2,2,4-trimethylpentane and 2,2,4-trimethyl-
pentane-d18. As noted above, product peaks were observed
as protonated molecules and homo- and/ or heterodimers.
Using the API-MS/ MS mode, products were identified based
on the consistency of the API-MS/ MS product ion spectrum
of a protonated homo- or hetero-dimer with the precursor
+
ion spectra of the [M
P
+ H] ion peaks (9). The products
observed from the OH radical-initiated reaction of 2,2,4-
trimethylpentane in the presence of NO and their suggested
structures or empirical formulas (see Discussion for reaction
schemes leading to these products) are shown in Table 1.
The positive ion API-MS and API-MS/ MS analyses indicated
the formation of products of molecular weight 58 (assumed
to be acetone, quantified by GC-FID), 72 (assumed to be
2
-methylpropanal, quantified by GC-FID), 116 (assumed to
be 4-hydroxy-4-methyl-2-pentanone, quantified by GC-FID),
28, 130, 144, 177, and 191. As expected from their odd
1
molecular weights, API-MS/ MS product ion analyses of the
+
[
1
(
M + H] ion peaks of the two products of molecular weight
FIGURE 1. Plots of the amounts of acetone, 2-methylpropanal, and
-hydroxy-4-methyl-2-pentanone formed, corrected for reaction with
77 and 191 showed the presence of fragment ions at 46 u
4
+
NO
2
), and these are therefore attributed to organic nitrates.
OH radicals (see text), against the amounts of 2,2,4-trimethylpentane
reacted. Quantification by GC-FID.
As discussed below, additional evidence of these organic
nitrates as well as of a C -hydroxyalkylnitrate of molecular
4
(
Figure 1) are given in Table 1. While no obvious GC peaks
were observed in the GC-FID analyses that could be attributed
to C -carbonyls or to C -nitrates, GC-MS analysis of an
irradiated CH ONO-NO-2,2,4-trimethylpentane-air mix-
weight 135 was obtained in the negative ion mode.
The low reactivity of 2,2,4-trimethylpentane-d18 (17)
limited the extent of reaction achieved and the amount
of deuterated products formed. However, as noted in
Table 1, the API-MS and API-MS/ MS analyses of an
8
8
3
ture showed the presence of two closely eluting peaks with
high-mass ions at m/ z 128 [and more intense fragments at
m/ z 127, 109 (-HCO), and 43] and of three peaks with fragment
ions at 46 u (interpreted as organic nitrates). However, no
isomer-specific identifications could be made, and no obvious
corresponding peaks were observed in the GC-FID analyses
and hence no quantifiations could be made.
irradiated CH ONO-NO-2,2,4-trimethylpentane-d18-air mix-
3
ture showed evidence for each of the corresponding deu-
terated products and provided additional structural infor-
mation. It has previously been observed that an alcohol -OD
group, if present, will rapidly exchange to an -OH group (6,
7). Therefore, the fact that the protonated molecule of the
following products were fully deuterated (i.e., no exchange-
able hydrogens were present) is consistent with their
API-MS Analyses. A series of CH ONO-NO-2,2,4-
3
trimethylpentane (or 2,2,4-trimethylpentane-d18)-air ir-
radiations were carried out with analyses using in situ API-
MS. In these analyses, the API-MS was operated in the
positive ion mode, or in negative ion mode with NO
structures being carbonyls or aldehydes: acetone-d
6
, 2-meth-
ylpropanal-d , and a C 16O aldehyde/ carbonyl (corre-
8
8
D
-
2
or
sponding to the molecular weight 128 product in Table 1).
The API-MS/ MS product ion spectra of the 145 u ion peak
-
[PFBOH‚O
2
]
ions as the ionizing species.
VOL. 36, NO. 4, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6 2 7

-
-
Using NO
2
as the ionizing agent, [NO ‚M] adduct ions
2
were observed at 162, 181, 223, and 237 u, corresponding to
products of molecular weight 116, 135, 177, and 191. A
number of weaker ion peaks were also observed, including
peaks corresponding to molecular weight 130 and 144
products. Using 5-hydroxy-2-pentanone as an internal
standard added after irradiation of a CH ONO-NO-2,2,4-
3
trimethylpentane-air mixture allowed approximate quan-
tification of presumed hydroxycarbonyl products of molec-
ular weight 116, 130, and 144 with formation yields of ∼8%,
∼
3%, and ∼3%, respectively, after correction for secondary
reactions using the measured (15) or estimated (16) OH
radical reaction rate constants for the products of these
molecular weights predicted from the reaction mechanisms
[
corrections for secondary reactions were only significant
for the molecular weight 144 product attributed to (CH C-
OH)CH C(CH CHO, with an estimated multiplicative cor-
3 2
)
(
2
3 2
)
rection factor of 1.76]. Thus, the negative ion mode analyses
confirmed the formation of the molecular weight 116, 130,
1
44, 177, and 191 products observed in the positive ion mode
analyses and suggested that the molecular weight 116, 130,
and 144 products are hydroxycarbonyls (with the molecular
weight 116 product being 4-hydroxy-4-methyl-2-pentanone)
and that the molecular weight 177 and 191 products are
hydroxynitrates. Additionally, the high sensitivity of the
FIGURE 2. API-MS/MS CAD “precursor ion” spectrum (negative
-
2 3
] reagent ion in an irradiated CH ONO-
ion mode) of the [PFBOH‚O
NO-2,2,4-trimethylpentane-air mixture. The peaks at 365, 407, and
21 u are tentatively identified as adducts of hydroxynitrates of
molecular weight 135, 177, and 191 u, respectively. The peaks at
46, 360, and 374 u are tentatively identified as adducts of
4
negative ion mode allowed detection of a C
of molecular weight 135.
4
-hydroxynitrate
3
hydroxycarbonyls of molecular weight 116 (4-hydroxy-4-methyl-2-
pentanone), 130, and 144 u, respectively.
Discussion
+
(
D
[M + H] of the C
8
D
16O product) showed losses of HDO and
The product data given in Table 1 allow the reaction pathways
to be delineated to a reasonable extent. OH radicals can react
with 2,2,4-trimethylpentane at four positions; the three
2
O (but not H
2
O), confirming that no -OH group was present
in this compound.
Consistent with their assignment as substituted alco-
hols, the following hydroxycarbonyls/ aldehydes were
observed as protonated molecules exhibiting an ex-
equivalent CH
secondary CH
group at the 4-position (52%), and the two equivalent CH
3
groups attached to the 2-carbon (11%), the
group at the 3-position (30%), the tertiary CH
2
3
change of -OD/ -OH: 4-hydroxy-4-methyl-2-pentanone-d11
,
groups attached to the 4-carbon (7%), where the numbers
in parentheses are the percentages of the four reaction
pathways calculated using the estimation method of Kwok
and Atkinson (16). These calculated percentages may be
subject to significant uncertainties, especially because the
total rate constant for 2,2,4-trimethylpentane is overpredicted
by 30%. In particular, it is likely that the fraction of the overall
reaction proceeding by H-atom abstraction from the tertiary
C-H bond at the 4-position is overestimated (due to steric
hindrance), while the fractions of the overall reaction
proceeding by H-atom abstraction from the other C-H bonds
are underestimated.
C
7
D
13O(OH), and C 15O(OH), corresponding to the mo-
8
D
lecular weight 116, 130, and 144 products in Table 1,
respectively. Consistent with their assignment as hydroxy-
nitrates, the deuterated products corresponding to the C
7
-
and C -hydroxyalkyl nitrates also exhibited -OD/ -OH ex-
8
change. API-MS/ MS product ion spectra were obtained for
the deuterated products corresponding to the molecular
weight 144 hydroxycarbonyl and the molecular weight 177
and 191 hydroxynitrate products listed in Table 1, and in
each case the CAD product spectrum showed losses including
H
2
O, indicative of -OD/ -OH exchange.
Negative Ion Mode. As noted in the Experimental Section
After H-atom abstraction from these various C-H bonds,
above, the negative ion mode is very sensitive to hydroxy
compounds, and the hydroxycarbonyls and hydroxynitrates
observed in this mode are listed in Table 1. PFBOH forms
2
the resulting alkyl radicals add O to form alkyl peroxy radicals,
which under our experimental conditions then react with
•
NO to form the corresponding alkoxy radical (C
NO or the alkyl nitrate C
8
H
17O ) plus
adducts with O
‚M] ions that can be readily fragmented by MS/ MS with
CAD to give [PFBOH‚O
precursor [PFBOH‚O ‚M] ion and the 230 u [PFBOH‚O
product ion is the molecular weight of the product M (8).
2
and certain species M, forming [PFBOH‚
2
8
H
17ONO
2
(of molecular weight 175),
-
O
2
as shown in reactions 1-3 in the Introduction. It should be
noted that the lack of authentic standards for the four
expected C H17ONO nitrates precluded their identification
8 2
by gas chromatographic or GC-MS analyses, and under our
experimental conditions, API-MS is rather insensitive to alkyl
-
2
] . Thus the difference in mass of the
-
-
2
2
]
API-MS and API-MS/ MS analyses of irradiated CH ONO-
3
NO-2,2,4-trimethylpentane-air mixtures were carried out
nitrates (8).
As noted in the Introduction, the four C
radicals can react with O
•
in the negative ion mode with PFBOH added to the sampled
8
H
17O alkoxy
gas stream. API-MS/ MS precursor ion spectra of the 230 u
2
, unimolecularly decompose by
-
[
(
PFBOH‚O
2
]
ion showed ion peaks at 346, 360, 365, 374
C-C bond scission, or isomerize through a six-membered
ring transition state to form a 1,4-hydroxyalkyl radical (4, 14,
18), with not all of these reaction pathways being possible
for all four alkoxy radicals initially formed from 2,2,4-
-
weak), 407, 421, and 428 ([(PFBOH)
2
‚O
2
] ) (Figure 2).
Reflecting the high sensitivity for hydroxynitrates and the
significant yield of 4-hydroxy-4-methyl-2-pentanone, it was
possible to obtain API-MS/ MS CAD product ion spectra of
the ion peaks at 346, 365, 407, and 421 u. The product ion
trimethylpentane. Schemes 1-4 show the potential reactions
•
of the four C
8
H
17O alkoxy radicals leading to first-generation
spectra were consistent with these ion peaks being [PFBOH‚
2
products. In these reaction schemes, detailed reactions are
-
•
•
•
O
‚M] ions, confirming molecular weight 116 (4-hydroxy-
omitted for CH
3
C HCH
3
, (CH
3
)
3
C , (CH
3
)
2
CHC H
2
, and (CH
3
)
2
C-
•
4
-methyl-2-pentanone), 135, 177, and 191 hydroxy-substi-
(OH)C H
2
radicals, which are known to react in the atmo-
3 2
C(O)CH , CH C(O)CH + HCHO, (CH ) -
tuted products.
sphere to form CH
3
3
3
3
6
2 8
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 4, 2002

SCHEME 1
SCHEME 2
2-methylpropanal, and 4-hydroxy-4-methyl-2-pentanone
observed and quantified by GC-FID analyses must originate.
•
O CH
2
C(CH
3
)
2
CH
2
CH(CH
3
)
2
Radical (11% Estim ated
Yield of Precursor Peroxy Radical). Scheme 1 shows that
this alkoxy radical can react with O , decompose, or isomerize
with predicted reaction rates at 298 K and atmospheric
2
4
6
6
-1
pressure of air of 4.9 × 10 , 1.1 × 10 , and 5 × 10
s ,
respectively (4, 18, 21). Because isom erization of the
•
HOCH
2
C(CH
3
)
2
CH
2
C(O )(CH
3
)
2
radical is predicted to domi-
nate over decomposition by a factor of ∼20 (4, 18, 21, 22),
•
the dominant predicted product from the O CH
2
C(CH
3
)
2
-
CH
CH
2
CH(CH
3
)
2
radical is the molecular weight 144 compound
C(CH CHO (consistent with our API-MS
(
3
)
2
C(OH)CH
2
3 2
)
3 3
CHCHO, and CH C(O)CH + HCHO, respectively (4, 14, 18),
analyses, including API-MS analyses of the 2,2,4-trimethyl-
pentane-d18 reaction where a corresponding product of
molecular weight 159 was observed). In addition, acetone +
and decomposition or isomerization of 1,2-hydroxyalkoxy
radicals is assumed to dominate over their reaction with O
2
(
4, 14, 18-20). The rates of the decomposition, isomerization,
2
(
-methylpropanal, the molecular weight 130 compound
CH C(OH)CH CH(CH )CHO, and the C -hydroxynitrate
HOCH C(CH CH C(ONO )(CH (products consistent with
2
and reaction with O of the various alkoxy and hydroxyalkoxy
3
)
2
2
3
8
radicals involved in Schemes 1-4 were calculated using the
estimation methods described by Atkinson (4, 18), as revised
by Aschmann and Atkinson (21). The calculated rates of these
reactions are not shown in Schemes 1-4, rather the relative
importance of the various reactions of a given alkoxy radical
are denoted by the arrows: a dashed arrow indicates that
the reaction pathway is estimated to account for e1% of the
overall reaction rate of the alkoxy radical, and a large,
boldface, arrow indicates that the reaction pathway is
2
3
)
2
2
2
3 2
)
our GC and API-MS analyses) are predicted to be formed
from more minor pathways.
•
(CH
3
)
3
CCH(O )CH(CH
3
)
2
Radical (30% Estim ated Yield
of Precursor Peroxy Radical). Scheme 2 shows that this
alkoxy radical can react with O
2
, decompose to (CH
3
)
3
CCHO
•
•
+ CH
3
C HCH
3
, or decompose to (CH
3
)
2
3 3
CHCHO + (CH ) C
with predicted reaction rates at 298 K and atmospheric
4
5
7
-1
estimated to dominate over the other(s) by a factor of g10.
pressure of air of 4.1 × 10 , 5.4 × 10 , and 2.4 × 10 s
,
•
The reactions of the four initially formed C
8
H
17O radicals
respectively (4, 18, 21). The formation of acetone and
are discussed below and compared with the products
observed. Considering the product yields (Table 1) and the
yields of the precursor alkoxy radicals, the boxes shown in
Schemes 2 and 3 indicate how all, or most, of the acetone,
2-methylpropanal as coproducts is therefore anticipated to
•
be essentially the only process for the (CH
(CH
ylpropanal of 26 ( 3% is consistent with this expectation
3
)
3
CCH(O )CH-
3 2
) radical, and our measured formation yield of 2-meth-
VOL. 36, NO. 4, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6 2 9

SCHEME 3
(
with most or all of the 2-methylpropanal arising from this
consistent with predictions, although the sum of the forma-
tion yields of acetone plus acetone and of 4-hydroxy-4-
methyl-2-pentanone (19 ( 5%) is significantly lower than
the predicted importance of H-atom abstraction from the
4-position (16).
alkoxy radical decomposition). Furthermore, our lack of
observation of 2,2-dimethylpropanal (with a formation yield
of <1.3%) confirms the prediction that decomposition of the
•
•
(
CH
HCH
acetone observed, 26 ( 3% is presumably associated with
3
)
3
3
CCH(O )CH(CH
3
)
2
radical to (CH
3
)
3
CCHO + CH
3
C -
is of no importance. Of the 54 ( 7% molar yield of
•
(
CH
3
)
3
CCH
2
CH(CH
3
)CH
2
O Radical (7% Estim ated Yield
of Precursor Peroxy Radical). Scheme 4 shows that this
alkoxy radical can decompose or react with O , and the
reaction rates of these two processes are predicted to be
2
-methylpropanal, leaving 28 ( 8% acetone (molar) to be
2
formed via other pathways.
•
(
CH
3
)
3
CCH
2
C(O )(CH
3
)
2
Radical (52% Estim ated Yield
4
-1
essentially identical at (5-6) x 10
s
(4, 18, 21). The
of Precursor Peroxy Radical). Scheme 3 shows that this
alkoxy radical can decompose or isomerize (but not react
with O
•
(
CH
3
)
3
CCH
2
CH(O )CH
3
radical formed after the decom-
, decompose or isomerize
position reaction can react with O
2
2
because of the lack of an abstractable R-H atom).
4
4
5
-1
with estimated rates of 4.1 × 10 , 6.9 × 10 , and 6 × 10 s
,
The decomposition and isomerization rates are calculated
5
-1
•
respectively (4, 18, 21). Neither 4,4-dimethyl-2-pentanone
to be equal at 6 × 10 s (4, 18, 21). The O CH
CH C(OH)(CH radical formed after the isomerization
reaction cannot isomerize and is calculated (4, 18, 21) to
dominantly decompose rather than react with O
anticipated major products from the (CH
2 3 2
C(CH ) -
(
(
which would be the product of the O
a product of the decomposition reaction), with formation
2 3
reaction) nor CH CHO
2
3 2
)
yields of <0.4% and <4%, respectively, were observed,
2
. The
•
•
indicating that reaction of the (CH
with O
this alkoxy radical probably is not dominant. The O CH
CH CH CH(OH)CH radical formed after isomerization of
the (CH
3
)
3
CCH
2
CH(O )CH
3
radical
3
)
3
CCH
2
C(O )(CH
3
)
2
2
is of no importance and that the decomposition of
radical are therefore acetone (with two molecules of acetone
formed per molecule of 2,2,4-trimethylpentane reacting) or
•
2
C-
(
3
)
2
2
3
•
4
-hydroxy-4-m ethyl-2-pentanone [(CH
CH ] (plus 2HCHO) from its isomerization reaction. The
reactions of the various hydroxyalkylperoxy radicals formed
3 2 2
) C(OH)CH C(O)-
3 3 2 3
) CCH CH(O )CH radical is predicted to primarily
3
isomerize. Scheme 4 therefore predicts that the major
•
after the initial isomerization are predicted to form C
and C
The predicted dominant decomposition of the (CH
radical [formed by decomposition of the (CH CCH
CH
formation of its O
Table 1). Our API-MS and GC data are therefore qualitatively
8
-, C
7
-,
products formed from the (CH
are the molecular weight 128 aldehyde (CH
CHO and the m olecular weight 130 hydroxycarbonyl
HOCH C(CH CH C(O)CH , consistent with our API-MS
data. In addition, form ation of the C -hydroxynitrate
3
)
3
CCH
2
CH(CH
3
)CH
2
O radical
-hydroxynitrates, all observed in the API-MS analyses.
3
)
3
CCH
2 3
CH(CH )-
4
•
3
)
3
CCH
2
O
•
)
3
2
C(O )-
2
3
)
2
2
3
3
(
3
)
2
radical] is consistent with the observed lack of
reaction product 2,2-dimethylpropanal
7
2
CH
3
CH(OH)CH
2
C(CH
3
)
2
CH
2
ONO
CCH
2
is anticipated following
•
(
isomerization of the (CH
3
)
3
2
CH(O )CH
3
radical.
6
3 0
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 4, 2002

SCHEME 4
TABLE 2. Initially Formed Peroxy Radicals from OH Radical-Initiated Reaction of 2,2,4-Trimethylpentane in the Presence of NO
and Observed Products Arising from Their Subsequent Reactions
peroxy radical
estd yield (%)
products formed and yield (%)
observed by
•
O OCH2C(CH3)2CH2CH(CH3)2
11
(CH3)2C(OH)CH2C(CH3)2CHO (MW 144) (∼3%)
HOC8H16ONO2
API-MS
API-MS
•
(
CH3)3CCH(OO )CH(CH3)2
CH3)3CCH2C(OO )(CH3)2
30
52
CH3C(O)CH3 + (CH3)2CHCHO (26 ( 3%)
2 CH3C(O)CH3 (14 ( 4%)
GC, API-MS
GC, API-MS
GC, API-MS
API-MS
•
(
(
CH3)2C(OH)CH2C(O)CH3 (MW 116) (5.1 ( 0.6%)
HOC8H16ONO2
HOC7H14ONO2
HOC4H8ONO2
(CH3)3CCH2CH(CH3)CHO (MW 128)
HOCH2C(CH3)2CH2C(O)CH3 (MW 130) (∼3%)
HOC7H14ONO2
API-MS
API-MS
(
CH3)3CCH2CH(CH3)CH2OO•
7
API-MS
API-MS
API-MS
Table 2 shows the initially formed peroxy radicals, their
predicted formation yields, and the assignment of observed
products. The products quantified by GC-FID and those for
which approximate [to within a factor of ∼2 (8)] formation
yields were obtained by API-MS analyses account for 51 (
Acknowledgments
The authors thank the U.S. Environmental Protection Agency,
National Center for Environmental Research and Quality
Assurance (Grant R-825252-01-0) and the California Air
Resources Board (Contract 99-330) for supporting this
research. The statements and conclusions are those of the
authors and not necessarily those of the funding Agencies.
1
0% of the overall products and reaction pathways, with the
remainder including alkyl nitrates, hydroxynitrates, and the
aldehyde (CH CCH CH(CH )CHO. In agreement with pre-
vious studies of the gC alkanes (4-10), our product data
3
)
3
2
3
4
Literature Cited
(1) Hoekman, S. K. Environ. Sci. Technol. 1992, 26, 1206.
show the importance of alkoxy radical isomerization reactions
leading to hydroxycarbonyls and hydroxynitrates.
VOL. 36, NO. 4, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6 3 1

(
(
2) Seila, R. L.; Lonneman, W. A.; Meeks, S. A. Determination of C
2
(12) Taylor, W. D.; Allston, T. D.; Moscato, M. J.; Fazekas, G. B.;
Kozlowski, R.; Takacs, G. A. Int. J. Chem. Kinet. 1980, 12, 231.
(13) Atkinson, R.; Aschmann, S. M.; Carter, W. P. L.; Winer, A. M.;
to C12 Ambient Air Hydrocarbons in 39 U.S. Cities, from 1984
through 1986; U.S. Environmental Protection Agency Report
EPA/ 600/ 3-89/ 058; U.S. EPPA: Research Triangle Park, NC,
September 1989.
Pitts, J. N., Jr. J. Phys. Chem. 1982, 86, 4563.
(
(
14) Atkinson, R. J. Phys. Chem. Ref. Data 1994, Monograph 2, 1.
15) Atkinson, R.; Aschmann, S. M. Environ. Sci. Technol. 1995, 29,
3) Lurmann, F. W.; Main, H. H. Analysis of the Ambient VOC Data
Collected in the Southern California Air Quality Study; Final
Report to California Air Resources Board Contract No. A832-
5
28.
(
(
16) Kwok, E. S. C.; Atkinson, R. Atmos. Environ. 1995, 29, 1685.
17) Atkinson, R. J. Phys. Chem. Ref. Data 1989, Monograph 1, 1.
1
30; Sacramento, CA, February 1992.
(
(
4) Atkinson, R. J. Phys. Chem. Ref. Data 1997, 26, 215.
5) Eberhard, J.; M u¨ ller, C.; Stocker, D. W.; Kerr, J. A. Environ. Sci.
Technol. 1995, 29, 232.
6) Atkinson, R.; Kwok, E. S. C.; Arey, J.; Aschmann, S. M. Faraday
Discuss. 1995, 100, 23.
(18) Atkinson, R. Int. J. Chem. Kinet. 1997, 29, 99.
(19) Kwok, E. S. C.; Atkinson, R.; Arey, J. Environ. Sci. Technol. 1996,
30, 1048.
(
(
(
(
(
20) Tuazon, E. C.; Aschmann, S. M.; Arey, J.; Atkinson, R. Environ.
Sci. Technol. 1998, 32, 2106.
7) Kwok, E. S. C.; Arey, J.; Atkinson, R. J. Phys. Chem. 1996, 100,
2
14.
(
(
21) Aschmann, S. M.; Atkinson, R. Int. J. Chem. Kinet. 1999, 31, 501.
22) Bethel, H. L.; Atkinson, R.; Arey, J. Int. J. Chem. Kinet. 2001, 33,
8) Arey, J.; Aschmann, S. M.; Kwok, E. S. C.; Atkinson, R. J. Phys.
Chem. A 2001, 105, 1020.
9) Aschmann, S. M.; Chew, A. A.; Arey, J.; Atkinson, R. J. Phys.
Chem. A 1997, 101, 8042.
3
10.
(
(
10) Aschmann, S. M.; Arey, J.; Atkinson, R. J. Phys. Chem. A 2001,
05, 7598..
11) Atkinson, R.; Tuazon, E. C.; Aschmann, S. M. Environ. Sci.
Technol. 1995, 29, 1674.
Received for review July 13, 2001. Revised manuscript re-
ceived October 23, 2001. Accepted October 29, 2001.
1
ES011134V
6
3 2
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 4, 2002