Analysis of airborne carboxylic acids and phenols as their pentafluorobenzyl derivatives: Gas chromatography/ion trap mass spectrometry with a novel chemical ionization reagent, PFBOH

Article abstract of DOI:10.1021/es970526s

The complex photochemical transformations of biogenic hydrocarbons such as isoprene and of anthropogenic hydrocarbons such as aromatics are an important source of carboxylic acids in the troposphere. The identification of unknown carboxylic acids can be d

Full text of DOI:10.1021/es970526s

Environ. Sci. Technol. 1998, 32, 299-309
acids, including formic and acetic acids, can contribute a
Analysis of Airborne Carboxylic
Acids and Phenols as Their
Pentafluorobenzyl Derivatives: Gas
Chromatography/Ion Trap Mass
Spectrometry with a Novel Chemical
Ionization Reagent, PFBOH
significant fraction (25-98%) of the natural acidity in
precipitation and cloudwater in many parts of the world
(3-5). These acids are emitted into the atmosphere from
various sources; in addition to primary emissions from
anthropogenic and biogenic sources, the photochemical
transformations of several precursors to carboxylic acids in
the atmosphere is also important (6-8). Laboratory and
field investigations indicate that atmospheric carboxylic acids
can be produced via oxidation of organic hydrocarbons and
aldehydes by oxidants (e.g., ozone) and various free radicals
•
•
(e.g., OH, HO2 ). For instance, significant sources of car-
boxylic acids in the ambient atmospheric environment are
aqueous-phase oxidation of aldehydes by hydroxyl radical
in cloudwater or rainwater (9-12) and the ozone-alkene
reaction (13, 14).
C H A O - J U N G C H I E N , †
M . J U D I T H C H A R L E S , ‡
K E N N E T H G . S E X T O N , † A N D
H A R V E Y E . J E F F R I E S * , †
Department of Environmental Sciences and Engineering,
University of North Carolina at Chapel Hill,
Chapel Hill, North Carolina 27599-7400, and Department of
Environmental Toxicology, University of California,
Davis, California 95616
In recent years, there has been considerable interest in
characterizing products arising from the photooxidation of
isoprene and aromatics in studies to elucidate chemical
reaction mechanisms affecting tropospheric ozone forma-
tion. Because of their large-scale emissions (isoprene has
the second largest flux of all hydrocarbons, and the aromatics
are abundant components in automobile emissions) and high
reactivity, isoprene and aromatics can have a significant
impact on ozone formation. The complexity of the atmo-
spheric oxidation of isoprene and aromatics, however, has
made the explicit description of their chemistry quite difficult.
Because there are multiple possibilities for carboxylic acid
formation in isoprene and aromatics oxidation, especially
for species like formic acid, determining the actual mecha-
nistic pathway is difficult. Although several chemical mech-
anisms have been postulated for the formation of carboxylic
acids (15-17), we lack sufficient knowledge regarding the
identity of the complement of carboxylic acids generated to
elucidate complete reaction pathways. Progress in under-
standing the oxidation chemistry has been deterred due to
the lack of specific and sensitive detection techniques capable
of identifying acid intermediates.
The analytical techniques previously employed to measure
airborne carboxylic acids include ion chromatography (IC)
by ion exchange or ion exclusion, liquid chromatography
(LC) (18-22), and gas chromatography (GC). Derivatization
methods in combination with LC and GC have also been
utilized (23). Although direct IC and LC methods are simple
and inexpensive to use, poor resolution and the necessity for
authentic standards for identification of compounds by
matching retention times are significant limitations to these
techniques. GC analysis of free carboxylic acids requires
either special column phases (24) or derivatization. Several
derivatization reagents, including methylformate (25) and
pentafluorobenzyl bromide (PFBBr), have been utilized with
GC (26-32). We were attracted to the employment of PFBBr
because of the success encountered with pentafluoroben-
zylhydroxyl amine (PFBHA), a structurally similar reagent,
with chemical ionization mass spectrometry (CIMS) to
identify airborne carbonyls (33-36). In addition, derivatives
of compounds with various functional groups, including
carboxylic acids, phenols (Scheme 1), amines, and mercap-
tans, can be formed allowing the identification of several
classes of molecules with excellent yields (37).
The complex photochemical transformations of biogenic
hydrocarbons such as isoprene and of anthropogenic
hydrocarbons such as aromatics are an important source
of carboxylic acids in the troposphere. The identification
of unknown carboxylic acids can be difficult, however, due
to the lack of standards as well as poor sensitivity and
lack of selectivity of existing techniques. In this study, we
describe the development of a method to analyze airborne
carboxylic acids using derivatization with
pentafluorobenzyl bromide (PFBBr) followed by capillary
gas chromatography/ion trap mass spectrometry (GC/
ITMS) analysis. In addition to the typical electron impact
ionization (EI) and methane chemical ionization (CI) mass
spectra of pentafluorobenzyl derivatives, we describe the
use of pentafluorobenzyl alcohol (PFBOH) as a novel
reagent gas for CIMS. Pentafluorobenzyl ions (m/z ) 181)
were produced at reduced pressure (1 × 10-5-2 × 10-5
Torr) from PFBOH gas, and the resulting CIMS were
characterized by the predominance of the [M + 181]+ ions.
The carboxylic acid products in the oxidation of isoprene
from indoor and outdoor smog chamber experiments
were analyzed using the described PFBBr GC/ITMS method.
The formation of methacrylic acid in an isoprene/O reaction
3
was demonstrated experimentally for the first time, and
acrylic acid, one of the U.S. EPA 189 Hazardous Air Pollutants,
is first reported here as an isoprene oxidation product.
In other outdoor toluene/NO chamber experiments, we
x
detected a series of carboxylic acids and various phenolic
compounds using this method.
Introduction
Carboxylic acids are ubiquitous throughout the troposphere
The resulting derivatives are stable and amenable to gas
chromatographic separation, after which mass spectrometry
detection greatly improves the selectivity and sensitivity
of the method. Recently, structural analyses of these
pentafluorobenzyl derivatives were described with GC/ MS
using electron-impact (EI) or electron capture negative ion-
(1, 2), and numerous studies have shown that some organic
* Corresponding author phone: (919)966-7312; fax: (919)966-7911,
e-mail: harvey@unc.edu.
† University of North Carolina at Chapel Hill.
‡ University of California.
9
S0013-936X(97)00526-9 CCC: $15.00
Published on Web 01/15/1998
1998 Am erican Chem ical Society
VOL. 32, NO. 2, 1998 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 2 9 9

SCHEME 1. Derivatization of Carboxylic Acids and Phenols to Respective Esters and Ethers
TABLE 1. PFBBr Derivatives of Selected Model Compounds and Their GC Properties
b
c
PFB derivative of
MWa
structure
RRT
RRF
Monoalkanoic Acids
form ic acid
226
240
254
268
282
296
HC(O)OCH2C6F5
-0.23
1.27
3.04
4.67
6.42
8.12
1.00
0.88
0.95
0.66
0.47
0.37
acetic acid
CH3C(O)OCH2C6F5
propionic acid
butanoic acid
valeric acid
CH3CH2C(O)OCH2C6F5
CH3(CH2)2C(O)OCH2C6F5
CH3(CH2)3C(O)OCH2C6F5
CH3(CH2)4C(O)OCH2C6F5
hexanoic acid
Alkenoic Acids
acrylic acid
252
266
266
266
CH2dCHC(O)OCH2C6F5
CH2dC(CH3)C(O)OCH2C6F5
CH2dCHCH2C(O)OCH2C6F5
CH3CHdCHC(O)OCH2C6F5
2.89
4.35
4.48
5.72
0.59
0.46
NDd
ND
m ethacrylic acid
vinylacetic acid
trans-crotonic acid
Hydroxy Acids
HOCH2C(O)OCH2C6F5
glycolic acid
256
5.15
0.40
Oxocarboxylic Acids
CH3C(O)C(O)OCH2C6F5
CH3C(O)(CH2)2C(O)OCH2C6F5
pyruvic acid
4-oxopentanoic acid
268
296
4.82
9.49
0.28
0.23
Aromatic Acids
C6H5C(O)OCH2C6F5
HOC6H4C(O)OCH2C6F5
benzoic acid
hydroxybenzoic acid
302
11.75
23.55
0.24
0.68
498e
Diacids
m aleic acid
succinic acid
476e
478e
C6F5CH2O(O)CCHdCHC(O)OCH2C6F5
C6F5CH2O(O)C(CH2)2C(O)OCH2C6F5
17.07
17.42
ND
0.28
Phenols
phenol
274
288
288
288
C6H5OCH2C6F5
8.64
9.74
10.07
10.24
0.29
0.19
0.20
0.12
o-cresol
CH3C6H4OCH2C6F5
CH3C6H4OCH2C6F5
CH3C6H4OCH2C6F5
m -cresol
p-cresol
salicyaldehyde
(2-hydroxybenzaldehyde)
2-nitrophenol
3-nitrophenol
nitrocresol
319
319
333
O2NC6H4OCH2C6F5
O2NC6H4OCH2C6F5
CH3C6H3(NO2)OCH2C6F5
15.40
16.44
16.06
0.59
0.44
0.32
a
b
c
Molecular weight of derivatives. Relative retention tim e (relative to PFBBr reagent peak). Response factor relative to form ic acid derivative.
d
e
All responses were derived from the reconstructed 181 ion m ass spectrum . Not determ ined. Diderivative.
chemical ionization (ECNICI) (38-40). The mass spectra of
the PFBBr derivatives using these methods, however, show
only specific m/z 181 ion and frequently do not provide
adequate information for the determination of molecular
weight or of compound structure.
pentafluorobenzyl alcohol (PFBOH) were also obtained from
Aldrich. Potassium carbonate (K2CO3) and anhydrous so-
dium sulfate (Na2SO4) were of reagent grade and were used
without further purification. All acids, phenols, and solvents
were analytical or higher grade quality.
In this study, we report for the first time the employment
of pentafluorobenzyl alcohol (PFBOH) and methane as
chemical ionization reagents to facilitate molecular weight
determinations of PFBBr derivatives of carboxylic acids and
phenols. We also demonstrate the power of the technique
by identifying airborne carboxylic acids and phenols pro-
duced from the oxidation of isoprene and toluene in indoor
and outdoor smog chamber experiments.
Preparation of Pentafluorobenzyl Esters and Ethers.
The derivatization procedure used to prepare the pentafluo-
robenzyl esters and ethers was based on methods described
by Brill et al. (37) and Glaceran et al. (38). Model compounds
(Table 1) were chosen for their relevance to atmospheric
systems and for structural variety.
A standard acid and phenol mixture (5 ng/ mL each) in
1 mL of acetone was placed into a 3-mL screw-top vial and
treated with 10 µL of 10% PFBBr in acetone, 50 µL of
18-crown-6 solution (4000 ng/ mL in acetone), and about 10
mg of powdered potassium carbonate (K2CO3). The vial was
stoppered and kept in a sonication bath at 40-60 °C for 2-3
h. Following completion of the reaction, the solution was
filtered and a 1-µL injection was introduced for GC/ MS
analysis without further purification or cleanup.
Experimental Section
Chem icals. PFBBr (also known as R-bromo-2,3,4,5,6-pen-
tafluorotoluene), which is a strong lachrymator and should
be handled with caution, was purchased from Aldrich
Chemical Co., Inc., Milwaukee, WI. The reagents 18-crown-6
(18c6), tri-n-octylphosphine oxide (TOPO), and 2,3,4,5,6-
9
3 0 0 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 32, NO. 2, 1998

TABLE 2. Relative Intensities of Characteristic EI Mass Spectra of PFBBr Derivatives of Carboxylic Acids and Phenols
% realtive intensity (RI) of characteristic ions
PFBBr ester of
M
m/z 197
m/z 181
m/z 161
M - 181
M - 197
M - 225
other
Monoalkanoic Acids
a
form ic acid
28
14
11
4
7
5
2
1
1
2
100
100
100
100
100
100
19
13
9
10
11
11
-
7
NAb
99
NA
NA
NA
38
acetic acid
propionic acid
butanoic acid
valeric acid
-
51
6
55
1
31
17
39
31
41 (29)
69 (78), 43 (43)
hexanoic acid
2
27
15
Alkenoic Acids
acrylic acid
15
29
4
3
66
86
21
9
10
13
2
-
-
-
-
100
100
100
100
NA
57
7
m ethacrylic acid
vinylacetic acid
crotonic acid
2
221 (48)
221 (55)
221 (8)
1
-
5
2
2
Hydroxy Acid
glycolic acid
2
4
100
13
3
6
NA
Keto Acids
pyruvic acid
4-oxopentanoic acid
-
-
-
-
80
56
4
8
-
-
-
77
100
7
43 (100)
51 (30)
Aromatic Acids
benzoic acid
hydroxyenzoic acid
27
13
1
-
46
100
9
8
-
5
100
6
49
-
Diacids
m aleic acid
succinic acid
-
-
-
6
100
100
4
-
-
1
15
-
5
13
101 (83)
Phenols
phenol
70
53
54
74
30
7
-
-
-
-
-
-
-
-
100
100
100
100
100
100
100
100
13
10
9
11
14
8
5
18
5
41
12
-
-
-
4
-
-
-
-
3
-
-
-
65 (22)
77 (35)
77 (25)
77 (50)
274 (25)
o-cresol
3
m -cresol
n-cresol
salicylaldehyde
2-nitrophenol
nitrocresol
3-nitrophenol
3
3
2
-
-
-
6
8
9
137 (21)
7
a
b
Not detectable or RI below 1%. Outside selected m ass range.
Sam ple Collection. Photochemical experiments were
conducted either in an indoor 0.25-m3 Teflon bag reactor
(TBR) contained in an aluminum foil-lined box or in the
UNC 150-m3 Teflon outdoor dual smog chamber located
near Pittsboro, NC.
Outdoor Smog Chamber Experiments. The UNC chamber
and its standard analytical instrumentations have previously
been described (41). The chamber uses natural sunlight as
its radiant energy source. For a typical daytime outdoor
chamber run, 2 ppmV of the selected hydrocarbon (isoprene
or toluene) and 0.67 ppmV of NO were injected into the
chamber. Injections were made approximately 0.5 h before
sunrise, and experiments typically ran until 1600 EDT. Air
samples were collected from beneath the chamber floor with
a Teflon sampling line no longer than 30 cm to minimize
sample loss. A chamber background air sample was drawn
30 min prior to the hydrocarbon injections. Throughout the
experiment, both batch and interval samples of 1 h were
taken in the same way as those described for indoor
experiments.
One isoprene/ O3 nighttime run was conducted on October
17, 1996, in which the initial conditions in the chamber were
4.8 ppmV of isoprene and ozone continuously injected at 0.3
ppm/ h. Air samples were taken for intervals of 1 h throughout
the 6-h experiment.
Sam ple Extraction and Preparation. After collecting
chamber air samples in aqueous phase, carboxylic acids
and/ or phenols were extracted into the organic phase using
a TOPO-methyl tert-butyl ether (MTBE) mixture solution.
The use of TOPO for facilitating liquid extraction for organic
compounds, e.g., acetic acid, acrylic acid, and phenols, has
been reported (42). Its low solubility in water and high
electron-donating ability make TOPO a suitable extractant
for polar molecules.
Indoor TBR Experiment. Two types of TBR experiments
were conducted: (1) organic compound mixed with NOx/ air
and irradiated with UV radiation and (2) organic compound
mixed with O3/ air in the dark. For first type of experiments,
the TBR was exposed to constant irradiation from a mixture
of black lamps and sun lamps at an intensity sufficient to
photolyze NO2 at a rate of about 4 × 10-4 molecule cm-3 s-1
.
In a typical experiment, after the TBR was flushed and filled
with clean air, about 10 ppmV of isoprene or toluene and 10
ppmV of nitrogen oxide (NO) were injected into the bag.
Hydrogen peroxide (H2O2), which provides a source of OH
radical by photolysis, was then delivered in the gas phase
into the bag by bubbling clean air through an impinger
containing 30% H2O2 solution. Once the initial condition
had been established, the lights were turned on for up to 3
h. Midget impingers with glass joints were used for collection
of the air samples. The impingers were wrapped in aluminum
foil to avoid continued photochemical reaction of the acids
during daylight sampling. Throughout the irradiation period,
a single sample was taken for the duration of the experiment
by pulling bag air into an impinger containing 10 mL of
deionized water at a flow rate of about 0.5 L/ min. The
collected samples were transferred into amber bottles with
Teflon-lined screw caps and stored in a refrigerator at 4 °C.
Analysis followed within 24 h. In isoprene/ ozone dark
experiments, about 0.8 ppmV of ozone was initially added
to the TBR, and then 10 ppmV of isoprene was injected. Air
samples were collected in the same way as described above.
To monitor any losses of the organic acids and phenols
in the entire analytical procedure, a known amount of
hexanoic acid (10 µL, 400 ng/ mL) was spiked into the collected
aqueous sample prior to extraction. This acid was chosen
9
VOL. 32, NO. 2, 1998 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 0 1

TABLE 3. Relative Intensities of Characteristic CH -CI Mass Spectra of PFBBr Derivatives of Carboxylic Acids and Phenols
4
% relative intensity (RI) of characteristic ions
PFBBr ester of
M + 181 M + 29 M + 1
M
M - 1 m/z 181 M - 181 M - 197 M - 225
m/z
Monoalkanoic Acids
a
form ic acid
-
-
-
4
-
-
-
-
-
-
-
-
2
5
3
5
6
5
4
2
2
100
100
100
100
100
100
-
-
-
1
22
45
NAb
12
NA
NA
NA
4
acetic acid
propionic acid
butanoic acid
valeric acid
3
23
434c (2)
448c (3)
462c (3)
6
3
45
3
4
3
8
57
3
hexanoic acid
-
14
59
4
Alkenoic Acids
acrylic acid
1
-
-
-
-
2
10
19
2
1
-
-
3
100
-
-
6
15
-
-
12
-
m ethacrylic acid
vinylacetic acid
crotonic acid
11
5
79
17
61
100
100
40
41
88
100
5
13
0
Hydroxy Acid
glycolic acid
-
-
1
2
-
100
-
2
-
Keto Acids
-
-
pyruvic acid
4-oxopentanoic acid
-
-
-
-
-
7
-
1
10
12
-
-
-
100
100
-
Aromatic Acids
benzoic acid
hydroxybenzoic acid
-
NA
-
-
36
19
11
1
-
100
-
3
42
57
-
1
-
100
Diacids
-
-
m aleic acid
succinic acid
NA
NA
-
-
15
16
1
1
100
100
-
-
2
29
-
-
Phenols
phenol
44
15
11
10
49
-
9
100
100
100
100
100
100
-
33
35
29
41
30
44
1
1
12
4
10
47
1
39
22
-
13
6
10
19
-
2
-
-
-
-
-
-
5
-
-
-
-
-
-
-
-
o-cresol
13
15
15
24
1
m -cresol
p-cresol
salicylaldehyde
2-nitrophenol
nitrocresol
3-nitrophenol
17
21
40
275d (63)
59
5
5
3
-
-
100
23
288e (84), 468f (26)
274e (4)
1
100
3
-
-
a
b
c
d
e
Not detectable or RI below 1%. Outside selected m ass range. M - H + 181. MH - CO. MH - NO2. f M + 181 - NO2.
as an internal standard because it would not form in the
atmospheric photochemical experiments and its PFBBr ester
did not coelute with other acids and phenols. The collected
aqueous samples were placed in 25-mL screw-capped culture
test tubes and were extracted twice with 2 mL of 0.5% TOPO
in MTBE after adding 2 drops of concentrated H2SO4. The
combined MTBE extract was then dried over anhydrous
sodium sulfate (Na2SO4) and evaporated with dry nitrogen
gas to dryness. The resulting residue was redissolved in 1
mL of acetone containing 1 ng/ mL of decafluorobiphenyl
(DFBP) as the internal standard. The solution was then
treated in a manner identical to that described for model
compounds.
Instrum entation. Spectrometric Analysis of the PFBBr
Derivatives. To demonstrate the detection ability and
selectivity of this technique, a Varian Saturn II GC/ ITMS was
employed. The system used an on-column injection port
and a RTX-5 60 m × 0.32 mm, 0.5 µm film thickness capillary
column. The following conditions were used: injection
temperature program, 60 °C for 0.2 min, 60-250 °C at 180
°C/ min, 250 °C for the rest of time; column temperature
program, 60 °C for 1 min, 60-290 °C at 8 °C/ min, 290 °C for
10 min; injection volume, 1 µL. Sample analytes were
analyzed in both EI and CI modes sequentially. The mass
scan range was 40-650, and a scan time of 1 scan/ s was
chosen.
a CH4-PFBOH gaseous mixture in the source of the mass
spectrometer was employed in this study. The ion trap
conditions under CH4-PFBOH CI mode were setup as
follows: automatic reaction control (ARC), 100 ms; CI
maximum ionization time, 2000 ms; and CI maximum
reaction time, 40 ms. The transfer line between the column
and the ion trap was kept at 280 °C, and the manifold (ion
source) was set at 175 °C. The reagent line and reagent flask,
containing the PFBOH, were both kept between 60 and
70 °C.
Results and Disscussion
Electron-Im pact Ionization (EI) Mass Spectral of PFBBr
Derivatives of Carboxylic Acids and Phenols. The EI mass
spectra of model compounds were evaluated to determine
whether molecular weight information of PFBBr derivatives
of unknown carboxylic acids and phenols could be obtained
by interpreting the EI mass spectra. Table 2 summarizes the
characteristic EI mass fragments of PFBBr derivatives of
standard compounds. Of the acids and phenols examined
for this study, the molecular ion (M+) in EI mass spectra is
present for most compounds except for the keto acids
(pyruvic and 4-oxopentanoic acids), the diacids (maleic and
succinic acids), and the nitrophenols. Although the mo-
lecular ions for phenolic compounds, e.g., phenol, cresols,
and salicylaldehyde, exhibit high relative intensities (RI )
∼74%), most of the remaining compounds show intensities
of less than 30%. The acid derivatives, primarily the alkanoic
and alkenoic acids, show a strong ion with m/z ) M - 197
resulting from a loss of C6F5CH2O from the derivative part
of the analyte. This ion, however, is often at an m/z lower
than 100, a region where many compounds exhibit fragment
ions. Thus, to provide for identification of a previously
Introduction of PFBOH. The PFBOH was introduced
directly into the ion trap cell using a two-hole graphite-vespel
ferrule in the transfer line, while the methane was introduced
through the CI inlet system to serve as both stabilization and
reagent gas. The amount of PFBOH and CH4 introduced
into the trap was monitored by recording m/z 181+ and 29+
ions, respectively. A ratio of 1:1 for m/z 29+ to m/z 181+ of
9
3 0 2 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 32, NO. 2, 1998

a unique feature of the CH4-CI mass spectra of PFBHA
derivatives of carbonyl compounds is the appearance of an
[M + 181]+ ion. It is only observed in spectra obtained on
an ion trap mass spectrometer due to the longer residence
time of fragment ions as compared to beam instruments,
and this promotes the secondary ion-molecule reactions.
In their studies, the protonated [M + H]+ ion was often the
base peak in the CH4-CI mass spectra, and the [M + 181]+
ion was present at low relative abundance (RI < 10%).
However, because the [M + 181]+ ion was the highest mass
and thus distinct from all other ions in the mass spectra, it
was easy to recognize. Together, the presence of the [M +
H]+ and [M + 181]+ ions served to confirm the molecular
weight of the derivative unambiguously. To determine if [M
+ H]+ and [M + 181]+ ions also result with respect to PFBBr
derivatives of carboxylic acids and phenols, CH4-CI mass
spectra of model compounds were examined, and the results
are listed in Table 3.
a
b
As it can be seen, the [M + H]+ ion is present as the base
peak in the CH4-CI mass spectra for most phenolic com-
pounds; however, the RI of this ion for acid derivatives is
considerably lower (<10%). As with the EI mass spectrum
of tested acid derivatives, the CH4-CI mass spectra are
dominated by the m/z 181 ion. Sufficient energy is thus still
imparted to the parent ion under CH4-CI conditions to induce
fragmentation of the molecules. It should be noted that an
[M + 181]+ ion is either absent or present at less than 11%
RI for the alkanoic, the alkenoic, the hydroxy, and the keto
acids. The phenols (phenol and cresols) meanwhile exhibit
higher [M + 181]+ relative ion intensities ranging from 10 to
49%. Again, as was the case with EI conditions, the abun-
dance of [M + 181]+, [M + H]+, and [M - H]+ ions for
carboxylic acids are low, and the significance of these ions
may be masked by background noise, making it difficult to
identify PFBBr derivatives of carboxylic acids for which
authentic standards are not available.
c
PFBOH Chem ical Ionization Mass Spectral Analyses.
Because the presence of an [M + 181]+ ion in the mass spectra
can assist in identifying the molecular mass of carboxylic
acids, it was thought that a means to facilitate the identifica-
tion and detection would be promotion of the formation of
this ion by introducing a source of the pentafluorobenzyl
cation (m/z 181+) directly into the ion trap mass spectrometer.
PFBOH was chosen as the precursor for the generation of
pentafluorobenzyl cation because under CI conditions it gives
the m/z 181 ion predominantly, comprising approximately
80% of the total ion current in the trap. In addition, the
vapor pressure of PFBOH liquid is sufficient to allow its
introduction in the gaseous phase to the mass spectrometer.
In an effort to evaluate PFBOH as a CI reagent gas, a gaseous
mixture of PFBOH and CH4 was employed to examine the
CI spectra of compounds, and the results from this work are
summarized in Table 4. As noted, the PFBOH/ CH4-CI mass
spectra of the examined compounds show increased RIs of
both the [M + 181]+ and [M + H]+ ions for all acid derivatives
except for glycolic and benzoic acids, which show no
formation of [M + 181]+ ion. While the [M + 181]+ ion
increases for all phenols and is the base peak for phenol,
cresols, and salicylaldehyde, the RI of the [M + H]+ ion
decreases in the mass spectra, indicating that the formation
of the [M + 181]+ ion is favored for these compounds.
FIGURE 1. Comparison of ion trap mass spectra obtained by EI (a),
CH -Cl (b), and PFBOH/CH -CI (c) for the PFBBr derivative of
4-oxopentanoic acid. (Note: All mass spectra were background
subtracted.)
4
4
unrecognized molecule, the mass spectra must exhibit either
a high intensity molecular ion (RI > 30%) or a moderate
intensity molecular ion (10-30%) along with a high [M -
197]+ ion. In cases where either the molecular ion or the [M
- 197]+ ion is absent or in low abundance, it would be very
difficult to identify carboxylic acids or phenolic compounds
for which authentic standards do not exist.
In a manner similar to that for the pentafluorobenzyl
derivatives of carbonyls (33, 34), the base peak in the EI mass
+
spectra is mostly a fragment ion at m/z 181 (C6F5CH2
,
pentafluorobenzonium ion). Although this ion is not very
informative regarding structural elucidation or molecular
weight determinations, it does provide an indication of the
presence of a carboxyl or a phenoxy moiety in the original
molecule. Our interpretation of the mass spectra also
indicate that carboxylic acid derivatives may be distinguished
from phenolic derivatives by the presence of an ion at m/z
M - 225, due to the loss of C6F5CH2CO2.
In light of our findings, we present the EI, CH4-CI, and
PFBOH/ CH4-CI ion trap mass spectra of 4-oxopentanoic acid
in Figure 1 as a typical example. In this case, the predominant
ions in the EI mass spectrum are at m/z 43 (M - 253), 99 (M
- 197), and 181. The molecular ion (M+) at m/z 296 is not
evident, and thus it would be difficult to ascertain that the
ion at m/z 99 is due to the loss of C6F5CH2O (m/z 197) from
the parent derivative. Although the analysis of the molecular
Methane-Chem ical Ionization (CH4-CI) Mass Spectral
Analyses. In previous research, Yu et al. (34) reported that
9
VOL. 32, NO. 2, 1998 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 0 3

TABLE 4. Relative Intensities of Characteristic PFBOH/CH -CI Mass Spectra of PFBBr Derivatives of Carboxylic Acids and Phenols
4
% relative intensity (RI) of characteristic ions
PFBBr ester of
M + 181
M + 29
M + 1
M
M - 1
M - 181
M - 197
other
Monoalkanoic Acids
a
propionic acid
butanoic acid
valeric acid
5
-
10
36
5
43
22
18
16
34
-
3
47
47
100
100
100
14
434c (29)
19
14
2
-
-
-
17
56
100
448c (32)
462c (36)
476c (6)
hexanoic acid
6
Alkenoic Acids
acrylic acid
2
-
-
-
-
11
100
48
17
21
14
5
-
-
-
-
80
432c (16)
m ethacrylic acid
vinylacetic acid
crotonic acid
38
23
60
100
73
45
446c (6)
4
100
100
401 (31), 221 (37)
401 (39)
95
7
Hydroxy Acid
glycolic acid
-
-
-
20
23
100
47
5
-
-
30
436c (6)
476 (11)
Keto Acid
4-oxopentanoic acid
54
14
100
Aromatic Acids
benzoic acid
hydroxybenzoic acid
-
-
-
100
100
46
64
14
7
-
4
85
45
NAb
Diacids
-
21
m aleic acid
succinic acid
NA
NA
-
-
100
100
4
3
-
-
5
58
101 (74)
Phenols
31
phenol
100
100
100
100
100
26
2
47
49
51
55
81
100
-
4
13
3
-
6
-
-
-
-
-
-
1
o-cresol
3
39
m -cresol
p-cresol
salicylaldehyde
2-nitrophenol
nitrocresol
3-nitrophenol
3
4
10
1
1
2
32
2
47
13
79
12
2
5
9
-
-
-
61
455 (22), 275 (54)
21
17
4
288d (29), 468e (100)
454e (78)
32
100
14
4
-
a
b
c
d
e
Not detectable or RI below 1%. Outside selected m ass range. M - H + 181. MH - NO2. M + 181 - NO2.
FIGURE 2. Reconstructed m/z 181 ion mass spectrum of a sample collected from an isoprene/O outdoor nighttime experiment. (IS, hexanoic
3
acid as an internal standard.)
weight of this analyte could be successfully conducted by
interpreting the CH4-CI mass spectrum, using PFBOH in
combination with CH4 as chemical ionization, the combina-
tion of an [M + 181]+ ion (m/z 477) and a high abundance
of protonated molecular ion (MH+, m/z 297) makes it easy
to determine the molecular weight of 4-oxopentanoate.
Carboxylic Acids Arising from the Oxidation of Isoprene.
To further establish the value of utilizing PFBOH along with
CH4 as chemical ionization reagents to identify airborne
carboxylic acids, we analyzed carboxylic acids resulting from
hydroxyl radical oxidation of isoprene in TBR experiments
and ozone oxidation in TBR and outdoor smog chamber
experiments. The indoor isoprene/ O3 experiments showed
the same carboxylic acid products as those in nighttime
outdoor chamber experiments; therefore, only the results
from the outdoor chamber experiments will be described.
9
3 0 4 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 32, NO. 2, 1998

FIGURE 3. Reconstructed m/z 181 ion mass spectrum of a sample collected from an isoprene/NO TBR experiment. (IS, hexanoic acid
x
as an internal standard.)
SCHEME 2. Formation of Possible Criegee Biradicals in Isoprene/O Chemistry
3
The reconstructed m/z ) 181 EI ion chromatograms of
the sample extracts collected from these isoprene experi-
ments are shown in Figures 2 and 3. By comparing the
retention times and the resulting EI and CI-CH4 mass spectra
of the carboxylic acid products with those of the model
compounds, we are able to identify formic, acetic, acrylic,
pyruvic, and methacrylic acids in the samples.
of methacrylic acid had been only proposed by model
prediction (7, 44), and we believe that its formation in this
system has not previously been demonstrated experimentally.
The production of acrylic acid has not previously been
predicted as a product of isoprene in daytime or nighttime
experiments. As indicated in Figures 2 and 3, we did detect
it in our samples.
While the production of formic and acetic acids has
received extensive laboratory study (6, 16, 43), the formation
It has been known that the electrophilic addition of ozone
to the unsaturated carbon-carbon bond of isoprene leads
9
VOL. 32, NO. 2, 1998 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 0 5

SCHEME 4. Possible Phenol and Benzoic Acid Formation
Pathways
a
b
c
carbonyls such as methyl vinyl ketone (MVK) and methac-
rolein (MACR), dicarbonyls such as glyoxal and methyl glyoxal
(MGLY), and hydroxycarbonyl products such as hydroxy
acetone (44, 45).
Jacob and Wofsy (7) have also proposed that biradicals
II and IV should react primarily with H2O in the same way
•
as CH2OO• to produce methacrylic acid and pyruvic acid:
O
•
•
C
COO + H2O
H
C
C OH + H2O
H2C
H2C
(3)
(4)
CH3
CH3
(II)
methacrylic acid
FIGURE 4. Mass spectra from EI (a), CH -CI (b), and PFBOH-CI (c)
4
for PFBBr derivative of an unknown compound (UN2) from an
O
C
O
C
O
C
isoprene/NO experiment.
x
•
•
H3C
COO + H2O
H
H3C
OH + H2O
SCHEME 3. Suggested Mechanism for the Formation of
(IV)
pyruvic acid
Acrylic Acid in Isoprene/O Reactions
3
It has been suggested that the biradicals III and V would not
react with water vapor to produce carboxylic acids, since the
radical carbons of these biradicals are fully substituted (47),
and instead would decompose to unreactive products (9,
44). To explain the formation of acrylic acid in our experi-
ments, a different mechanism is proposed as described below
and diagrammed in Scheme 3.
The substituted peroxy biradical III can undergo rear-
rangement leading to a more stable bisoxy biradical VI (48,
49), followed by a loss of the CH3 radical via unimolecular
decomposition and reaction with H2O to yield acrylic acid
(Scheme 3). The methyl hydroperoxide (MHP, CH3OOH)
formed from the reaction of the CH3OO radical with the HO2
radical has also been observed by Hatakeyama et al. (50) in
the reaction of ozone with olefins having a methyl group.
It should be noticed that ozonolysis produces OH radicals
(51, 52) and that both OH and O3 reactions occur in our
experimental systems. The yield of acrylic acid in isoprene/
NOx TBR experiments (Figure 4), where OH radical reactions
dominate, was very minor as compared to that in isoprene/
O3 reactions, suggesting that acrylic acid was formed
predominately via isoprene/ O3 chemical pathways. In
addition, the time series analysis in the isoprene/ O3 nighttime
to the formation of several possible energy-rich Criegee
biradicals (Scheme 2) (45, 46). The unsubstituted biradical
(I) either reacts with H2O or is collisionally stabilized by the
surrounding molecules to produce formic acid (HCOOH):
•CH2OO• 9M8 HCOOH + M
(1)
(2)
H O8 HCOOH + H2O
2
Reactions of the substituted biradicals (II-V) will form
9
3 0 6 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 32, NO. 2, 1998

TABLE 5. Proposed Structures of Unknown Carboxylic Acids from an Isoprene/NO TBR Experiment
x
FIGURE 5. Reconstructed m/z 181 ion mass spectrum of a sample collected from a toluene/NO outdoor chamber experiment. (IS, hexanoic
x
acid as an internal standard.)
outdoor chamber experiment showed that acrylic acid
appeared immediately at the beginning of the experiment,
and a linear increase in concentration as the reaction
proceeded was observed as well. These observations also
suggest that acrylic acid was formed as a direct product in
isoprene/ O3 reaction.
absence or low intensity of the molecular ion in the EI and
-CI makes it impossible to deduce the molecular weight
CH4
of the derivative. In contrast, the presence of [M + H]+ and
[M + 181]+ ions in the PFBOH/ CH4 mass spectra provides
for easy determination of the molecular weight of the
derivative, in this case 298, and for the formulation of possible
structures of carboxylic acids consistent with this molecular
weight. Possible structures consistent with MWs of other
unknown acids are given in Table 5. Various unidentified
carbonyl compounds including hydroxy carbonyls, unsatur-
ated carbonyls, and epoxy carbonyls have been suggested in
the oxidation of isoprene (34). Further oxidation of these
compounds may lead to numerous carboxylic acids as
For the unidentified compounds in Figure 3 whose
molecular weights could not be clearly determined either
from their EI or CH4-CI mass spectra, the sample was
subsequently analyzed using the new PFBOH/ CH4-CI method.
The EI, CH4-CI, and PFBOH/ CH4-CI mass spectra of the
PFBBr derivative for an unknown compound (UN2) is shown
in Figure 4 as an example of the power of this method. The
9
VOL. 32, NO. 2, 1998 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 0 7

SCHEME 5. (A) Reaction Sequence Suggested by Yu et al. for Photooxidation of Toluene; (B) Possible Reaction Sequence
Suggested in This Study for the Formation of Pyruvic Acid from OH-Initiated Oxidation of o-Cresol
suggested in Table 5. Because most authentic standards for
the acids proposed in Table 5 are not commercially available,
additional work is needed to verify their identities. In the
case where further determination was possible, we propose
that the compound with MW 266 is a C4 unsaturated
carboxylic acid. We have tentatively identified this as
cis-crotonic acid after elim inating the other possible
isomerssmethacrylic, trans-crotonic, and vinylacetic acidssby
comparing its retention time with authentic standards for
these.
Carboxylic Acids and Phenols Arising from the Photo-
oxidation of Toluene. A reconstructed ion chromatogram
of a sample obtained from the photooxidation of toluene in
the presence of NOx in an outdoor smog chamber is given
in Figure 5. We positively identified not only phenolic
compounds (including phenol, cresols, nitrophenols, and
nitrocresols) but also carboxylic acids (including formic,
acetic, pyruvic, and benzoic acids).
H-atom abstraction from the methyl group and OH radical
addition to the aromatic ring (53). The phenolic compounds
observed in our experiments result from the OH addition
pathway, while benzoic acid is believed to arise from the H
abstraction pathway. The formation of benzoic acid and
phenol may arise from the continued reaction of benzal-
dehyde, a photooxidation product of toluene (54), with OH
radical (Scheme 4).
Several recent studies have also suggested that the ring-
opening products from the photooxidation of aromatics
contain various carbonyl compounds (54-56). The reaction
of O3 with the unsaturated ring-opened intermediates can
lead to numerous carboxylic acids. The reaction of OH with
phenols can also yield keto acids, as suggested by Grosjean
(57), where the OH-initiated photooxidation of o-cresol yields
products including pyruvic acid. Since aromatic hydrocar-
bons comprise as much as 30% of the organics in urban
atmospheric environments (58), they could be potential
precursors of keto acids. A study by Yu et al. (35, 36) suggests
a formation pathway leading to ring products, diunsaturated
1,6-dicarbonyls, for oxidation of toluene. The postulated
Under atmospheric conditions, oxidation processes of
toluene are mainly initiated by the reaction of OH radicals.
Two major pathways are involved in the initial reactions:
9
3 0 8 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 32, NO. 2, 1998

(27) Davis, B. Anal. Chem. 1977, 49, 832-834.
(28) Kawahara, F. K. Environ. Sci. Technol. 1971, 5, 235-239.
(29) Cline, R. E.; Todd, G. D.; Ashley, D. L.; Grainger, J.; McCraw, J.
M.; Alley, C. C.; Hill, R. H., Jr. J. Chromatogr. Sci. 1990, 28, 167-
172.
(30) Lee, H.-B.; Peart, T. E. J. Chromatogr. 1990, 498, 367-379.
(31) Kroger, S. Analyst 1989, 114, 1647-1648.
(32) Jacobsson, S.; Larsson, A.; Arbin, A. J. Chromatogr. 1988, 447,
329-340.
(33) Le Lacheur, R. M.; Sonnenberg, L. B.; Singer, P. C.; Christman,
R. F.; Charles, M. J. Environ. Sci. Technol. 1993, 27, 2745-2753.
(34) Yu, J.; Jeffries, H. E.; Le Lacheur, R. M. Environ. Sci. Technol.
1995, 29, 1923-1932.
(35) Yu, J.; Jeffries, H. E.; Sexton, K. G. Atmos. Environ. 1997, 31,
2261-2280.
(36) Yu, J.; Jeffries, H. E.; Sexton, K. G. Atmos. Environ. 1997, 31,
2281-2287.
(37) Brill, J. H.; Bikshandarkoil, A. N.; McCormick, J. P. Appl. Spectrosc.
1991, 45, 1617-1620.
(38) Galceran, M. T.; Moyano, R.; Poza, J. M. J. Chromatogr. 1995,
710, 139-147.
(39) Sinkkonen, S.; Kolehmainen, E.; Paasivirta, J.; Hamalainen, S.;
Lahtipe, M. J. Chromatogr. 1995, 718, 391-396.
(40) Chalaux, N.; Bayona, J. M.; Albaiges, J. J. Chromatogr. 1994, 686,
275-281.
mechanism in photooxidation of toluene involves the
formation of 3-methyloxepin intermediate (Scheme 5A)
similar to the benzene oxidation mechanism originally
suggested by Barnes et al. (59, 60). By analogy with their
mechanism, the production of pyruvic acid from reaction of
o-cresol with OH radicals can be explained by the reaction
sequence depicted in Scheme 5B. Full evaluation of this
mechanism, however, awaits positive identification of the
production of the OH-substituted benzene oxide/ oxepin
intermediate and the diunsaturated oxo acids (XI).
In this work, we have demonstrated an advanced tech-
nique for analyzing airborne carboxylic acids and phenols
from various chamber experiments as their PFBBr derivatives
by using GC-ion trap mass spectrometry with the CI reagent-
PFBOH. The resulting CH4-PFBOH CIMS can be character-
ized by the appearance of the [M + 181]+ ions, facilitating
molecular mass determination of carboxylic acids. More
detailed mass spectrometry aspects are currently under
investigation and will be reported in a future paper.
Acknowledgments
(41) Jeffries, H. E.; Fox, D.; Kamens, R. M. Environ. Sci. Technol.
1976, 10, 1006-1011.
(42) Vairavamurthy, A. Anal. Chem. 1986, 58, 2684-2687.
(43) Atkinson R.; Baulch, D. L.; Cox, R. A.; Hampson, R. F.; Kerr, J.
A.; Troe, J. J. Phys. Chem. 1989, 18, 881-1097.
(44) Grosjean, D.; Willams, E. L., II; Grosjean, E. Environ. Sci. Technol.
1993, 27, 830-840.
This work was supported by the United States Environmental
Protection Agency through Grant R 824789-01-0 and Contract
68D50129. The authors would like to thank Dr. Jianzhen Yu
for her insight and helpful discussions and the Chromatog-
raphy Systems Division of Varian Associates, Inc., Walnut
Creek, CA, for making available a Saturn II gas chromatog-
raphy and ion trap mass spectrometer for this work.
(45) Kamens R. M.; Gery, M. W.; Jeffries, H. E.; Jackson, M.; Cole, E.
I. Int. J. Chem. Kinet. 1982, 14, 955-975.
(46) Paulson, S. E.; Flagan, R. C.; Seinfeld, J. H. Int. J. Chem. Kinet.
1992, 24, 103-125.
Literature Cited
(1) Chebbi, A.; Carlier P. Atmos. Environ. 1996, 30, 4233-4249.
(2) Talbot, R. W.; Andreae, M. O.; Berresheim, H.; Jacob, D. J.;
Beecher, K. M. J. Geophys. Res. 1990, 95, 16,799-16,811.
(3) Keene, W. C.; Galloway, J. N. J. Geophys. Res. 1986, 91, 14,-
466-14,474.
(4) Keene, W. C.; Galloway, J. N. Atmos. Environ. 1984, 18, 2491-
2497.
(5) Keene, W. C.; Galloway, J. N. Tellus 1988, 40B, 322-334.
(6) Madronich, S.; Chatfield, R. B.; Calvert, J. G. Geophys. Res. Lett.
1990, 17, 2361-2364.
(7) Jacob, D. J.; Wofsy, S. C. J. Geophys. Res. 1988, 93, 1477-1486.
(8) Andreae, M. O. J. Geophys. Res. 1987, 92, 6635-6641.
(9) Jacob, D. J. J. Geophys. Res. 1986, 91, 9807-9826.
(10) Chameides, W. L. J. Geophys. Res. 1984, 89, 4739-4755.
(11) Chameides, W. L.; Davis, D. D. Nature 1983, 304, 427-429.
(12) Gradel, T. E.; Weschler, C. J. Rev. Geophys. 1981, 19, 5209-5221.
(13) Khwaja, H. A. Atmos. Environ. 1995, 29, 127-139.
(14) Grosjean D. Environ. Sci. Technol. 1989, 23, 1506-1514.
(15) Atkinson, R. J. Phys. Chem. Ref. Data 2 Monogr. 1994, 1-216.
(16) Madronich, S.; Calvert, J. G. J. Geophys. Res. 1990, 95, 5697-
5715.
(17) Moortgat, G. K.; Veyret, B.; Lesclaux, R. Chem. Phys. Lett. 1989,
160, 443-447.
(18) Grosjean, D.; Grosjean, E.; Williams, E. L., II. Environ. Sci.
Technol. 1994, 28, 186-196.
(19) Lawrence, J. E.; Koutrakis, P. Environ. Sci. Technol. 1994, 28,
957-964.
(20) Gronberg, L.; Shen, Y.; Jonsson, J. A. J. Chromatogr. 1993, 655,
207-215.
(21) Schilling, M.; Klockow, D. Fresenius J. Anal. Chem. 1993, 346,
738-744.
(22) Grosjean, D.; Neste, A. V.; Parmar, S. S. J. Liq. Chromatogr. 1989,
12, 3007-3017.
(23) Frei, R. W.; Lawrence, J. F. Chemical derivatization in analytical
chemistry; Plenum Press: New York, 1995; Vol. 1.
(24) Betowski, L. D.; Kendall, D. S.; Pace, C. M.; Donnelly, J. R. Environ.
Sci. Technol. 1996, 30, 3558-3564.
(47) Hatakeyama, S.; Bandow, H.; Okuda, M.; Akimoto, H. J. Phys.
Chem. 1981, 85, 2249-2254.
(48) Horie, O.; Moortgat, G. K. Atmos. Environ. 1991, 25A, 1881-
1896.
(49) Hatakeyama, S.; Akimoto, H. Res. Chem. Intermed. 1994, 20,
503-524.
(50) Hatakeyama, S.; Lai, H.; Gao, S.; Murano, K. Chem. Lett. 1993,
1287-1291.
(51) Atkinson, R.; Aschmann, S. M.; Arey, J.; Shorees, B. J. Geophys.
Res. 1992, 97, 6065-6073.
(52) Chew, A. A.; Atkinson, R. J. Geophys. Res. 1996, 101, 28,649-
28,653.
(53) Finlayson-Pitts, B. J.; Pitts, J. N., Jr. Atmospheric chemistry:
Fundamentals and experimental techniques; John Wiley &
Sons: New York, 1986.
(54) Dumdei, B. E.; O’Brien, R. J. Nature 1984, 311, 248-250.
(55) Bandow, H.; Washida, N.; Akimoto, H. Bull. Chem. Soc. Jpn.
1985, 58, 2531-2540. (b) Bandow, H.; Washida, N. Bull. Chem.
Soc. Jpn. 1985, 58, 2541-2548. (c) Bandow, H.; Washida, N.
Bull. Chem. Soc. Jpn. 1985, 58, 2549-2555.
(56) Shepson, P. B.; Edney, E. O.; Corse, E. W. J. Phys. Chem. 1984,
88, 4122-4126.
(57) Grosjean, D. Atmos. Environ. 1984, 18, 1641-1652.
(58) Singh, H. B.; Salas, L. J.; Cantrell, B. K.; Redmond, R. M. Atmos.
Environ. 1985, 19, 1911-1919.
(59) Barnes I.; Klotz, B.; Becker, K. H. Aromatic hydrocarbon oxidation
mechanisms: New developments; Paper 96-RP130A.01, 89th
Annual Meeting of Air and Waste Management Association,
Nashville, TN, June 23, 1996.
(60) Klotz, B.; Barnes I.; Becker, K. H.; Golding B. T. J. Chem. Soc.,
Faraday Trans. 1997, 93, 1507-1516.
Received for review June 13, 1997. Revised manuscript re-
ceived October 14, 1997. Accepted October 20, 1997.X
ES970526S
(25) Sollinger, S.; Levsen, K.; Emmrich, M. J. Chromatogr. 1992, 609,
297-304.
(26) Chen, S.-H.; Wu, H. L. J. Chromatogr. 1990, 502, 257-264.
X Abstract published in Advance ACS Abstracts, December 1, 1997.
9
VOL. 32, NO. 2, 1998 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 3 0 9