Characterization of polar compounds and oligomers in secondary organic aerosol using liquid chromatography coupled to mass spectrometry

Article abstract of DOI:10.1021/ac701852t

A generic method has been developed for the analysis of polar compounds and oligomers in secondary organic aerosol (SOA) formed during atmospheric simulation chamber experiments. The technique has been successfully applied to SOA formed in a variety of systems, ranging from ozonolysis of biogenic volatile organic compounds to aromatic photooxidation. An example application of the method is described for the SOA produced from the reaction of ozone with cis-3-hexenyl acetate, an important biogenic precursor. A range of solvents were tested as extraction media, and water was found to yield the highest recovery. Extracts were analyzed using reversed-phase liquid chromatography coupled to ion trap mass spectrometry. In order to determine correct molecular weight assignments and increase sensitivity for less polar species, a series of low-concentration mobile-phase additives were used (NaCl, LiBr, NH ₄OH). Lithium bromide produced better fragmentation patterns, with more structural information than in the other cases with no reduction in sensitivity. The main reaction products identified in the particle-phase were 3-acetoxypropanal, 3-acetoxypropanoic acid, and 3-acetoxypropane peroxoic acid and a series of dimers and trimers up to 500 Da. Structural identification of oligomers indicates the presence of linear polyesters possibly formed via esterfication reactions or decomposition of peroxyhemiacetals.

Full text of DOI:10.1021/ac701852t

Anal. Chem. 2008, 80, 474-480
Characterization of Polar Compounds and
Oligomers in Secondary Organic Aerosol Using
Liquid Chromatography Coupled to Mass
Spectrometry
Jacqueline F. Hamilton,*,† Alastair C. Lewis,† Trevor J. Carey,‡ and John C. Wenger‡
Department of Chemistry, University of York, Heslington, York, UK, YO10 5DD, and Department of Chemistry and
Environmental Research Institute, University College Cork, Cork, Ireland
atmosphere. The major precursors of SOA formation are biogenic
compounds such as terpenes and anthropogenic compounds such
as aromatic hydrocarbons.2 Since global emissions of biogenic
VOCs are ∼10 times larger than anthropogenic emissions,3 the
SOA produced from biogenic VOCs is expected to make the
largest contribution to the atmospheric aerosol burden.
A generic method has been developed for the analysis of
polar compounds and oligomers in secondary organic
aerosol (SOA) formed during atmospheric simulation
chamber experiments. The technique has been success-
fully applied to SOA formed in a variety of systems,
ranging from ozonolysis of biogenic volatile organic com-
pounds to aromatic photooxidation. An example applica-
tion of the method is described for the SOA produced
from the reaction of ozone with cis-3-hexenyl acetate, an
important biogenic precursor. A range of solvents were
tested as extraction media, and water was found to yield
the highest recovery. Extracts were analyzed using re-
versed-phase liquid chromatography coupled to ion trap
mass spectrometry. In order to determine correct molec-
ular weight assignments and increase sensitivity for less
polar species, a series of low-concentration mobile-phase
additives were used (NaCl, LiBr, NH4OH). Lithium bro-
mide produced better fragmentation patterns, with more
structural information than in the other cases with no
reduction in sensitivity. The main reaction products
identified in the particle-phase were 3-acetoxypropanal,
3-acetoxypropanoic acid, and 3-acetoxypropane peroxoic
acid and a series of dimers and trimers up to 500 Da.
Structural identification of oligomers indicates the pres-
ence of linear polyesters possibly formed via esterfication
reactions or decomposition of peroxyhemiacetals.
Recent evidence from atmospheric simulation chamber studies
has shown that the oxidation of gas-phase organic species may
lead ultimately to the formation of macromolecules (oligomers
and polymers up to 1000 Da) in aerosols.4 Oligomer formation is
now proposed as being a key mechanism by which VOCs are
incorporated into the particle phase. The sheer complexity of
organic compounds found in atmospheric aerosol is the main
practical barrier to relating condensed-phase organic species to
their parent gas-phase precursors, and a full structural evaluation
would require the use of a range of complementary techniques.
In order to simplify this complexity, controlled simulation chamber
experiments focusing on a single hydrocarbon allow direct links
between primary emitted organics and their conversion to SOA
to be investigated. A range of on-line and off-line techniques have
been used to investigate the polar and oligomeric content of SOA
formed during chamber experiments. On-line techniques such as
aerosol mass spectrometry, photoelectron resonance capture
ionization aerosol mass spectrometry, and thermal desorption
particle beam mass spectrometry provide real-time bulk properties
of SOA but give limited structural information.5-8 Off-line tech-
niques, where samples are collected onto a substrate and analyzed
by gas or liquid chromatography or mass spectrometry, provide
Atmospheric aerosols have recently received much attention
due to their effect on Earth’s climate and human health.1 The
environmental impact of atmospheric aerosols depends strongly
on the chemical composition, and in particular on the inorganic,
black carbon and organic carbon fractions. A significant amount
of the organic fraction is attributed to secondary organic aerosol
(SOA) formed as a result of gas-particle transfer of species during
the oxidation of volatile organic compounds (VOCs) in the
(2) Fuzzi, S.; Andreae, M. O.; Huebert, B. J.; Kulmala, M.; Bond, T. C.; Boy,
M.; Doherty, S. J.; Guenther, A.; Kanakidou, M.; Kawamura, K.; Kerminen,
V. M.; Lohmann, U.; Russell, L. M.; Poschl, U. Atmos. Chem. Phys. 2006,
6, 2017-2038.
(3) Guenther, A.; Hewitt, C. N.; Erickson, D.; Fall, R.; Geron, C.; Graedel, T.;
Harley, P.; Klinger, L.; Lerdau, M.; Mckay, W. A.; Pierce, T.; Scholes, B.;
Steinbrecher, R.; Tallamraju, R.; Taylor, J.; Zimmerman, P. J. Geophys. Res.
1995, 100 (D5), 8873-8892.
(4) Kalberer, M.; Paulsen, D.; Sax, M.; Steinbacher, M.; Dommen, J.; Prevot,
A. S. H.; Fisseha, R.; Weingartner, E.; Frankevich, V.; Zenobi, R.;
Baltensperger, U. Science 2004, 303 (5664), 1659-1662.
(5) Alfarra, M. R.; Paulsen, D.; Gysel, M.; Garforth, A. A.; Dommen, J.; Prevot,
A. S. H.; Worsnop, D. R.; Baltensperger, U.; Coe, H. Atmos. Chem. Phys.
2006, 6, 5279-5293.
* To whom correspondence should be addressed. E-mail: jfh2@york.ac.uk.
† University of York.
‡ University College Cork.
(1) Kanakidou, M.; Seinfeld, J. H.; Pandis, S. N.; Barnes, I.; Dentener, F. J.;
Facchini, M. C.; Van Dingenen, R.; Ervens, B.; Nenes, A.; Nielsen, C. J.;
Swietlicki, E.; Putaud, J. P.; Balkanski, Y.; Fuzzi, S.; Horth, J.; Moortgat, G.
K.; Winterhalter, R.; Myhre, C. E. L.; Tsigaridis, K.; Vignati, E.; Stephanou,
E. G.; Wilson, J. Atmos. Chem. Phys. 2005, 5, 1053-1123.
(6) Zahardis, J.; LaFranchi, B. W.; Petrucci, G. A. Atmos. Environ. 2006, 40
(9), 1661-1670.
(7) Ziemann, P. J. Faraday Discuss. 2005, 130, 469-490.
474 Analytical Chemistry, Vol. 80, No. 2, January 15, 2008
10.1021/ac701852t CCC: $40.75 © 2008 American Chemical Society
Published on Web 12/15/2007

detailed chemical speciation but at the expense of time resolution.
Gas chromatography is not particularly suited to the most polar
fraction of SOA, i.e., the water-soluble organic carbon, and may
cause thermal decomposition of oligomers, although some limita-
tions can be reduced using derivatization.9 Liquid chromatography
(LC) is becoming increasingly popular for the analysis of polar
compounds in aerosol and is routinely used for the analysis of
carboxylic acids. Tolocka et al.10 applied LC combined with matrix-
assisted laser desorption/ionization (MALDI), electrospray ioniza-
tion (ESI), and chemical ionization (CI) mass spectrometry for
the analysis of oligomers in SOA. CI was found to cause
decomposition of the parent oligomer, but MALDI and ESI
provided similar oligomer distributions. Gao et al.11 used LC
coupled to an ion trap mass spectrometer (LC-ITMS) to identify
oligomers up to 1600 Da in SOA formed during cyclohexene
ozonolysis and estimated that they accounted for ∼10% of the total
mass fraction. Fragmentation patterns for structural analysis were
performed in negative ionization mode, but a wider range of
species could be detected in positive ionization mode as [M +
Na]+ adducts. Hamilton et al.12 also investigated the formation of
oligomers in the cyclohexene system using LC-ITMS. The major
components of the SOA were dicarboxylic acids, which were
detected in both positive and negative modes. Higher sensitivity
was achieved for the [M + Na]+ adducts formed in the positive
mode, but due to poor fragmentation in the ion trap, these data
could not be used for structural analysis.
ing in complete domination of the [M + Li]+ adducts but did not
improve detection limits.17
This paper presents the development of a LC-MS technique
for the analysis of polar compounds and oligomers in secondary
organic aerosols. The extraction of oligomers and polar species
in a range of solvents was tested to ensure maximum efficiency.
Total mass distributions were obtained by directly introducing
extracts into the MS using a syringe pump. The influence of a
series of cationization agents in the mobile phase was investigated
in order to maximize sensitivity and improve identification. The
technique has been successfully applied to SOA formed in 10
systems, ranging from biogenic VOC ozonolysis to aromatic
photooxidation. The structural analysis of the SOA formed during
the ozonolysis of cis-3-hexenyl acetate in a simulation chamber
experiment is presented as an example. cis-3-Hexenyl acetate, also
known as leaf acetate, is a volatile compound emitted by green
foliage. Kirstine et al.18 analyzed the head space of grass clippings
(primarily Lolium perenee) shortly after it had been cut and found
that 40% of the grass emissions were cis-3-hexenyl acetate, with
the other related C6 oxygenated VOCs (hexenols and hexenals)
contributing 29%. The cutting or grazing of pastures induces the
production a range of hexenyl-type compounds, which act as
antibiotics at the site of the wound to prevent fungal growth and
bacterial infection.
EXPERIMENTAL SECTION
LC-MS/MS. LC-MS/MS analysis was performed using a
HCT Plus ion trap mass spectrometer (Bruker Daltonics GmbH,
Bremen, Germany) equipped with an Eclipse ODS-C18 column
with 5-µm particle size (Agilent, 4.6 mm × 150 mm). Samples (60
µL) were injected via an autosampler (Agilent 1100 series), and a
binary gradient elution was performed using (A) 0.2 µg mL-1
cationization agent (NaCl, LiBr, NH4OH) and 0.1% formic acid in
HPLC-MS grade water (100 to 40% over 40 min, hold for 10 min,
return to starting conditions) and (B) HPLC grade methanol, at
a flow rate of 0.6 mL min-1. ESI was carried out at 300 °C, with
a nebulizer pressure of 70 psi and nitrogen dry gas flow rate of
12 L min-1. The mass spectrometer was used in positive ion mode,
scanning from 50 to 600 Da. The automated MS2 function from
the Esquire software (Bruker Daltonics GmbH) was used to
fragment ions, where the two most abundant ions at each scan
were subjected to CID. After five consecutive scans, a mass was
rejected from auto-MS2 for 1 min before being reactivated to allow
minor components to be analyzed.
As oligomers grow larger, there may be some species with
low ionization potentials, which are unlikely to be identified in
positive mode electrospray ionization unless they form adducts.
The formation of adducts with metal ions, such as sodium and
potassium, can reduce sensitivity and can cause problems when
interpreting mass spectra as the operator is unsure of the correct
molecular weight, for example, due to uncertainty whether the
highest mass is due to the [M + H]+ or [M + Na]+ ion. In the
analysis of polymers, such as polyesters and polyethers, cationi-
sation agents, such as alkali buffers, can be added to the mobile
phase to increase the abundance of one specific ion [M + Cat]+,
which can lead to increased sensitivity.13,14 In the case of poly-
glycols, the molecular ions can be difficult to fragment in collision-
induced dissociation tandem MS, but lithium adducts readily
undergo fragmentation yielding structurally informative product
ion spectra.15,16 This technique has recently been applied to the
analysis of airborne isocyantes as dibutylamine derivatives, result-
Simulation Chamber Experiments. Experiments were car-
ried out at the European Photoreactor (EUPHORE), which
consists of two large atmospheric simulation chambers located
on the roof of a building in Valencia, Spain. Technical information
concerning the installation and its use for VOC oxidation experi-
ments has been previously reported in the literature.19-21 The
(8) Gross, D. S.; Galli, M. E.; Kalberer, M.; Prevot, A. S. H.; Dommen, J.; Alfarra,
M. R.; Duplissy, J.; Gaeggeler, K.; Gascho, A.; Metzger, A.; Baltensperger,
U. Anal. Chem. 2006, 78 (7), 2130-2137.
(9) Surratt, J. D.; Murphy, S. M.; Kroll, J. H.; Ng, N. L.; Hildebrandt, L.;
Sorooshian, A.; Szmigielski, R.; Vermeylen, R.; Maenhaut, W.; Claeys, M.;
Flagan, R. C.; Seinfeld, J. H. J. Phys. Chem. A 2006, 110 (31), 9665-9690.
(10) Tolocka, M. P.; Jang, M.; Ginter, J. M.; Cox, F. J.; Kamens, R. M.; Johnston,
M. V. Environ. Sci. Technol. 2004, 38 (5), 1428-1434.
(11) Gao, S.; Keywood, M.; Ng, N. L.; Surratt, J.; Varutbangkul, V.; Bahreini, R.;
Flagan, R. C.; Seinfeld, J. H. J. Phys. Chem. A 2004, 108 (46), 10147-10164.
(12) Hamilton, J. F.; Lewis, A. C.; Reynolds, J. C.; Carpenter, L. J.; Lubben, A.
Atmos. Chem. Phys. 2006, 6, 4973-4984.
(17) Karlsson, D.; Dahlin, J.; Marand, A.; Skarping, G.; Dalene, M. Anal. Chim.
Acta 2005, 534 (2), 263-269.
(18) Kirstine, W.; Galbally, I.; Ye, Y. R.; Hooper, M. J. Geophys. Res. 1998, 103
(D9), 10605-10619.
(13) Jackson, A. T.; Williams, J. P.; Scrivens, J. H. Rapid Commum. Mass Spectrom.
2006, 20 (18), 2717-2727.
(14) Jackson, A. T.; Slade, S. E.; Scrivens, J. H. Int. J. Mass Spectrom. 2004,
238 (3), 265-277.
(15) Chen, R.; Yu, X. L.; Li, L. J. Am. Soc. Mass Spectrom. 2002, 13 (7), 888-
897.
(16) Lattimer, R. P. J. Am. Soc.Mass Spectrom. 1992, 3 (3), 225-234.
(19) Volkamer, R.; Platt, U.; Wirtz, K. J. Phys. Chem A 2001, 105 (33), 7865-
7874.
(20) Wenger, J. C.; Le Calve, S.; Sidebottom, H. W.; Wirtz, K.; Reviejo, M. M.;
Franklin, J. A. Environ. Sci. Technol. 2004, 38 (3), 831-837.
(21) Becker, K. H. The European Photoreactor EUPHORE. Final Report of the
EU project EV5V-CT92-0059; Wuppertal, 1996.
Analytical Chemistry, Vol. 80, No. 2, January 15, 2008 475

Table 1. Initial Conditions and Aerosol Yields for
cis-3-Hexenyl Acetate SOA Experiment in EUPHORE
EUPHORE 03/05/06
initial cis-3-hexenyl
acetate concn
1614 ppbV (1.915 g)
ozone concn
1600 ppbV added over 60 min
SOA yielda
24%
maximum SOA concna
aerosol mass on filter
903 µg/m3
0.5926 mg
a Assuming a SOA density of 1.0 g cm-3
.
chambers are hemispheres made of fluorinated ethylenepropylene
film, with a volume of ∼200 m3, and are surrounded by a
retractable steel housing. The chamber is operated at ambient
temperature and approximately atmospheric pressure using dry
purified air, while two large fans ensure homogeneous mixing of
the chamber contents. The ozonolysis of cis-3-hexenyl acetate was
performed using starting concentrations of ∼1.6 ppmV. No radical
scavenger or seed aerosol was used, and the relative humidity
was ∼6%. The average temperature during the experiment was
19 °C. The decay of the hydrocarbon was monitored by FT-IR
spectroscopy, and the formation and evolution of SOA was
measured using a scanning mobility particle sizer (TSI 3022A
condensation particle counter and TSI 3081 differential mobility
analyzer). Samples of SOA were collected when the particle
concentration in the chamber had reached a maximum, ∼4 h after
the start of the reaction. Samples were collected on 47-mm-
diameter quartz-fiber filters using a high-volume sample pump
operating at 12 L min-1 for 1 h. Filters were prefired at 500 °C for
12 h prior to sample collection and were stored in foil below 4 °C
until analysis. Experimental details including the SOA yield and
total mass of aerosol collected are shown in Table 1.
Figure 1. Comparison of LC-MS chromatograms of the SOA
produced from the ozonlysis of cis-3-hexenyl acetate obtained using
a range of solvents for extraction. (A) water. (B) methanol. (C)
acetonitrile. (D) tetrahydrofuran.
Table 2. Relative Extraction Efficiencies (REE) and
Reproducibility (RSD) of SOA Componentsa
Filter Extraction. The filter was divided into four pieces, and
each piece was then placed into a vial with 5 mL of a high-purity
solvent (water, methanol, tetrahydrofuran, acetonitrile) purchased
from Fisher Chemicals. The sample vial was wrapped in foil to
avoid any possible degradation and left for 2 h. This was sonicated
for 1 h and the extraction filtered using a 0.45-mm PVDF syringe
filter (Whatman). Extracts were evaporated to dryness using a
V10 vacuum solvent evaporator (Biotage) at 30 °C, and the residue
was redissolved in 1 mL of a 50:50 solution of HPLC-MS grade
water and methanol.
water
methanol
RSD
acetonitrile
RSD
tetrahydrofuran
RSD
RSD
peak REE
(%)
REE (%) REE
(%)
REE
(%)
133
155
199
271
297
308
1.86
1.26
0.87
0.90
1.04
9.6
5.4
1.0
4.0
3.5
1.27 4.4
1.33 2.9
1.10 6.3
1.21 2.7
1.00 4.1
0.80 1.9
0.48 4.7
0.58 3.7
1.03 5.5
1.11 3.0
1.00 0.5
1.20 3.1
0.39
0.83
n.d.
0.79
0.95
0.62
11.9
14.8
n.d.
7.6
5.9
1.38 10.9
7.8
mean 1.22
5.75 1.12 3.71 0.90 3.40
0.60
9.60
a Peaks used correspond to those marked in Figure 1.
RESULTS
Extraction. The aerosol mass yield (Y) from the simulation
of ozonolysis of cis-hexenyl acetate presented here was 24%,
assuming spheric particles with an effective density of 1.0 g cm-3
.
relative mass of each quarter of filter paper, and the average was
taken. Good precision was obtained with relative standard devia-
tions in peak areas below 8% in most cases. For each peak, a mean
response of the four extraction solvents was calculated and the
relative extraction efficiency (REE) determined using the expres-
sion; REE ) As/Am, where As is the average area for a specific
solvent and Am is the average area of all solvents. Therefore, a
REE value above 1 indicates a better than average extraction and
below 1 indicates a poorer extraction. The REE and RSD values
for each extraction solvent are shown in Table 2. Tetrahydrofuran
was discounted due to high levels of impurities (see Figure 1d).
The extracts of the SOA formed were analyzed by LC-MS in
positive ion mode, and the base peak chromatograms (BPC) are
shown in Figure 1. The optimum extraction solvent was deter-
mined by picking five peaks, spread across the entire chromato-
gram, and comparing the average peak areas of three replicate
analyses for each of the solvents used. For the first peak, with a
retention time of ∼14 min, the [M + H]+ ) 133 Da and the [M
+ Na]+ ) 155 Da signals were detected and both areas were
determined. Since the concentrations are unknown, extraction
recoveries could not be used. Peak areas were corrected for the
476 Analytical Chemistry, Vol. 80, No. 2, January 15, 2008

Figure 2. Averaged mass spectra of the water-soluble extract of SOA produced from the ozonolysis of cis-3-hexenyl acetate. The four spectra
show the effect of a series of cationization agents. (A) uncatalyzed extract. (B) NaCl. (C) LiBr. (D) NH4OH.
Both methanol and water provided good extraction of oligomers
and polar compounds with good run-to-run reproducibility. Water
was chosen for further extractions due to lower levels of impuri-
ties. The use of the vacuum solvent evaporator significantly
reduced sample preparation time compared with conventional
rotary evaporation, with 5-mL water extracts evaporated to dryness
in 6 min at near-ambient temperatures.
Molecular Weight Distribution. To determine the mass
distribution of SOA components, the water extract was directly
injected into the ESI source of the ion trap MS using a syringe
pump. This allows a direct and quick characterization of the
molecular weight range of organic species in the sample as shown
in Figure 2. It is clear that the SOA from cis-3-hexenyl acetate is
dominated by species below 350 Da, but the mass spectrum does
extend to 500 Da, indicating the presence of oligomers. The peaks
at 328 and 330 Da are due to an artifact from the simulation
chamber (also present in blank filter samples) and should be
ignored. During previous analysis of SOA formed during cyclo-
hexene ozonolysis,12 it was found that fragmentation patterns
obtained without preseparation were generally from a mixture of
species with the same nominal mass and could not be used to
determine structures.
extract were mixed with 100 µL of a 2 µg mL-1 solution of
cationisation agent (LiBr, NaCl, NH4OH) in high-purity water and
injected using the syringe pump. The large excess of cationization
reagent results in the exclusive formation of [M + Cat]+, shifting
the position of peaks in the mass spectra as seen in Figure 2,
except for NH4OH, which had no effect. Comparison of the spectra
for the uncatalyzed extract and the NaCl extract indicates that
only sodium adducts are present in the original extract, even
without the addition of extra sodium, perhaps as a result of the
highly oxidized nature of SOA. This is confirmed by the [M +
Li]+ spectrum, where peaks are shifted down by 16 Da with
respect to the corresponding [M + Na]+ peaks, reflecting the
difference in cation mass (Li+ ) 7 Da, Na+ ) 23 Da). Using this
technique, the correct molecular weights can be assigned to the
dominant peaks in the spectra. For example, the three large peaks
below 200 Da ([M + Na]+ ) 139, 155, and 171 Da and [M + Li]+
) 123, 139, 155 Da) correspond to the neutral species with
molecular mass 116, 132, and 148 Da. Peak intensities are similar
in all four spectra, indicating that, for the highly oxidized species
found in SOA, cationisation agents can be used without any
detrimental effect on sensitivity.
Negative ionization could only be used for the uncatalyzed
extract and the NH4OH solution due to the complete domination
of adducts of Cl- and Br- in the other extracts. Negative ionization
was significantly less sensitive and the mass spectrum was
dominated by impurities in the methanol. This suggests that the
SOA formed in the ozonlysis of cis-3-hexenyl acetate does not
contain a significant acidic fraction, unlike the previous study of
The positive ionization mass spectrum is complex, and peaks
may correspond to the protonated molecular ion [M + H]+ or an
adduct, likely [M + Na]+. Often both peaks appear in the same
spectrum 22 Da apart, making the presence of the adduct obvious,
but there is no evidence of this in the SOA extract. To determine
the correct molecular weight assignments, 100-µL aliquots of the
Analytical Chemistry, Vol. 80, No. 2, January 15, 2008 477

Figure 3. LC-MS chromatograms (BPC) of water-soluble SOA components using a series of cationization agents (0.2 µg mL-1) in the aqueous
0.1% formic acid mobile phase. (A) uncatalyzed. (B) NaCl. (C) LiBr. (D) NH4OH.
cyclohexene SOA, which was dominated by dicarboxylic acids and
their heterogeneous reaction products.12
in the cis-3-hexenyl acetate SOA were identified from their Li+
adducts using LC-MS as 3-acetoxypropanal (116 Da), 3-acetoxy-
propanoic acid (132 Da. See Figure 4 for MS2 spectra.), and
3-acetoxypropane peroxoic acid (148 Da) and are shown in Table
3. These three species are among the gas-phase reaction products
expected from the ozonolysis of cis-3-hexenyl acetate22-23 and are
therefore present in the particle phase as a result of condensation.
The remaining, higher molecular weight, reaction products
detected in the SOA samples are likely due to reaction of the first-
generation products in the particle-phase.
Six of the most abundant heterogeneous reaction products
were identified, and possible structures are given in Table 3. The
oligomers identified contain ester and ether linkages, and all
contain the 3-acetoxypropanoic acid (APA) unit as indicated by
the common loss of 132 Da during fragmentation. The smallest
oligomer, 2-hydroxyethyl-3-acetoxypropanoate M ) 176 Da, is an
ester and appears to be a building block for larger oligomers. It
may be formed through an esterfication reaction of APA with 1,2-
ethanediol or from the decomposition of a peroxyhemiacetal24
(formed from reaction of 2-hydroperoxyethanol and 3-acetoxy-
propanal). The oligomers, M ) 232 Da and M ) 248 Da, are both
based on this building block and are structurally similar differing
only by one oxygen atom, reflecting the incorporation of either
3-hydroxylpropanal or 3-hydroxypropanoic acid formed from the
Criegee intermediate of the alkyl end of cis-3-hexenyl acetate.22-23
The larger oligomers, M ) 274, 290, and 348 Da, appear to be
formed from a mixture of monomers from both ends of the cis-
Chromatography and Structural Analysis. In order to obtain
daughter ion spectra for each individual compound, LC separation
was required. The chromatograms were similar for each of the
mobile phases used as shown in Figure 3, with a small reduction
in chromatographic efficiency and shift in retention time observed
when using the NaCl and LiBr solutions. Sensitivity was slightly
reduced for oligomers, but enhanced signals were seen for the
most polar species with all three cationization agents. The NH4OH
mobile phase again provided no improvement over the uncatalyzed
solution and is not considered further.
One of the major problems experienced in positive ionization
of highly oxidized species is the dominance of the [M + Na]+
adduct as shown in the direct injection of the uncatalyzed extract
above. In many cases, these adducts do not generate product
spectra when fragmented in the ion trap, due to complete
degradation. An example from the cis-3-hexenyl acetate extract is
shown in Figure 4 for [132 + Na]+ ) 155 Da, one of the major
small polar molecules. The sodium adduct is completely destroyed
and yields no fragments, only a noisy spectrum with no distinct
peaks. However, replacing the Na+ with Li+ completely reverses
this situation and yields highly informative product ion spectra
as shown in Figure 4 for the same species [132 + Li]+. In addition,
when fragmenting [M + Li]+, a heavier peak is often seen 18 Da
higher, indicating the formation of a water adduct, i.e., [M + Li
+ H2O]+ as shown in Figure 4 (m/z 157 Da).
A complete elucidation of the SOA composition is beyond the
scope of this paper. However, as seen in the averaged mass
spectrum in Figure 2, the SOA is dominated by a small number
of compounds and the structural elucidation of these species has
been performed. The three most abundant small reaction products
(22) Calvert, J. G.; Atkinson, R.; Kerr, J. A.; Madronich, S.; Moortgat, G. K.;
Wallington, T. J.; Yarwood, G. The Mechanisms of Atmospheric Oxidation of
the Alkenes; Oxford University Press: New York, 2000.
(23) Grosjean, E.; Grosjean, D. J. Atmos. Chem. 1997, 27 (3), 271-289.
(24) Durham, L. J.; Wurster, C. F.; Mosher, H. S. J. Am. Chem. Soc. 1958, 80
(2), 332-337.
478 Analytical Chemistry, Vol. 80, No. 2, January 15, 2008

Figure 4. Product ion spectra (MS2) for 3-acetoxypropanoic acid. Upper: uncatalyzed extract, compound present as [M + Na]+ ) 155 Da and
produces no useful fragmentation pattern. Lower: LiBr-catalyzed extract, compound present as [M + Li]+ ) 139 Da and produces a well-
defined fragmentation pattern showing losses of 42 (CH2CO), 60 (CH3COOH), and 72 (CH2CHCOOH) Da. A water adduct is also present [M
+ Li + H2O]+ ) 157 Da.
3-hexenyl acetate molecule. There are a number of molecules that
would produce the fragmentation patterns obtained, but the
proposed structures are those most likely based on known reaction
mechanisms, although the formation mechanisms are far from
being understood. Structural analysis of larger oligomers (M )
464 and 498 Da) indicates the presence of at least three ester
linkages, suggesting the formation of linear polyesters within the
SOA produced in this reaction.
The vapor pressures of 3-acetoxypropanal, 3-acetoxypropanoic
acid, and 3-acetoxypropane peroxoic acid were estimated using
the expanded Clausius-Clapeyron equation method in Jenkin et
al.,25 and the mean was 1.41 × 10-4 Torr, indicating these are
semivolatile species. Therefore, these species would be expected
to be partitioned between the gas and condensed phases. In
addition, if the APA undergoes reactive uptake, where it reacts
irreversibly in the aerosol phase, this will allow more APA to
partition from the gas phase. There is the possibility that these
small molecules are fragments of oligomer molecules that break
down during sample preparation. However, this has been mini-
mized by carrying out the workup at ambient temperatures.
The sample preparation method has also been tested on a
standard mixture containing a range of commercially available
esters, acids, and alcohols and no degradation of the analytes was
observed.
(25) Jenkin, M. E. Atmos. Chem. Phys. 2004, 4, 1741-1757.
Analytical Chemistry, Vol. 80, No. 2, January 15, 2008 479

Table 3. Polar Compounds and Oligomers Identified in cis-3-Hexenyl Acetate Ozonolysis SOA Using LC-MS with Li+
Cationisation
Fifteen additional oligomer structures were investigated, and
all contained the APA monomer unit indicating that only a limited
number of key chemical fragments are needed to describe much
of the composition. This is consistent with our previous study of
cyclohexene-ozone SOA where many of the oligomers contained
a common unit, adipic acid. If the formation of oligomers can be
explained using a limited subset of key intermediates, the
incorporation of heterogeneous reactions into detailed SOA
models, such as MCM-SOA,25-26 will be simplified and may
ultimately lead to reduced mechanisms for transport and climate
models. The data presented here are a first look at the composition
of SOA from this biogenic VOC, and further studies are required
at more atmospherically relevant conditions.
subset of agents were tested, and other ions, such as Ag+ or Co2+
,
may provide further enhancements such as increased sensitivity.
Understanding the chemical composition of SOA, in particular
heterogeneous reaction products, is a key mechanism for under-
standing their formation. The method presented has provided
information on both early-generation reaction products, which
have partitioned to the particle phase, and oligomers, including
linear polyesters, which may have formed on the surface or in
the bulk of the aerosol particle.
ACKNOWLEDGMENT
This work was supported by the European Commission
(project EUROCHAMP, contract RII3-CT-2004-505968), the Eu-
ropean Science Foundation (INTROP programme), and the NERC
APPRAISE project. The authors thank Esther Borra´s Garc´ıa and
Amalia Mun˜oz of Fundacio´n CEAM and Robert Healy, University
College Cork, for assistance with the simulation chamber experi-
ments.
CONCLUSIONS
The addition of LiBr as a cationization agent in the LC-MS
analysis has significantly increased the amount of structural
information that could be obtained on polar compounds and
oligomers in SOA. The use of cationization agents improved the
reliability of molecular weight assignments also. Only a small
Received for review September 3, 2007. Accepted October
22, 2007.
(26) Johnson, D.; Jenkin, M. E.; Wirtz, K.; Martin-Reviejo, M. Environ. Chem.
2005, 2 (1), 35-48.
AC701852T
480 Analytical Chemistry, Vol. 80, No. 2, January 15, 2008