Single particle analysis of secondary organic aerosols formed from 1,4-cyclohexadiene ozonolysis using a laser-ionization single-particle aerosol mass spectrometer

Article abstract of DOI:10.1246/bcsj.81.120

Real-time analysis of secondary organic aerosol (SOA) particles formed from 1,4-cyclohexadiene (CHD) ozonolysis in a smog chamber was performed using a laser-ionization single-particle aerosol mass spectrometer (LISPA-MS). The instrument can be used to obtain both the size and chemical compositions of individual aerosol particles with a high time-resolution (≈2 s at the maximum). Both positive- and negative-ion mass spectra can be obtained by changing the voltage polarity of the instrument. The negative-ion spectra of the SOA particles provided important information about the chemical compositions of the SOA particles. In the negative-ion spectra, intense mass peaks were determined to correspond to ions with carboxyl and aldehyde groups. The signal intensities of the intense mass peaks from compounds with carboxyl groups were higher than those from compounds with aldehyde groups as a function of the particle size. The peaks suggest that the SOA particles contain more oxygenated organic compounds as the particle size increases, namely, the chemical compositions of the SOA particles vary as a function of the particle size. We demonstrated that the real-time single-particle analysis of SOA particles by using the LISPA-MS technique can be used to clarify the formation and transformation processes of SOA particles in smog chambers.

Full text of DOI:10.1246/bcsj.81.120

120 Bull. Chem. Soc. Jpn. Vol. 81, No. 1, 120–126 (2008)
ꢀ 2008 The Chemical Society of Japan
Single Particle Analysis of Secondary Organic Aerosols Formed
from 1,4-Cyclohexadiene Ozonolysis Using a Laser-Ionization
Single-Particle Aerosol Mass Spectrometer
y
yy
yyy
ꢀ1
Masahiro Narukawa,1; Yutaka Matsumi, Jun Matsumoto,1; Kenshi Takahashi,1;
yyyy
Akihiro Yabushita,2;
Kei Sato,3 and Takashi Imamura3
1Solar-Terrestrial Environment Laboratory, Nagoya University, 3-13 Honohara, Toyokawa 442-8507
2Horiba, Ltd., 2 Miyanohigashi, Kisshoin, Minami, Kyoto 601-8510
3National Institute for Environmental Studies, 16-2 Onogawa, Tsukuba 305-8506
Received March 27, 2007; E-mail: matsumi@stelab.nagoya-u.ac.jp
Real-time analysis of secondary organic aerosol (SOA) particles formed from 1,4-cyclohexadiene (CHD) ozonol-
ysis in a smog chamber was performed using a laser-ionization single-particle aerosol mass spectrometer (LISPA-MS).
The instrument can be used to obtain both the size and chemical compositions of individual aerosol particles with a high
time-resolution (ꢁ2 s at the maximum). Both positive- and negative-ion mass spectra can be obtained by changing the
voltage polarity of the instrument. The negative-ion spectra of the SOA particles provided important information about
the chemical compositions of the SOA particles. In the negative-ion spectra, intense mass peaks were determined to cor-
respond to ions with carboxyl and aldehyde groups. The signal intensities of the intense mass peaks from compounds
with carboxyl groups were higher than those from compounds with aldehyde groups as a function of the particle size.
The peaks suggest that the SOA particles contain more oxygenated organic compounds as the particle size increases,
namely, the chemical compositions of the SOA particles vary as a function of the particle size. We demonstrated that
the real-time single-particle analysis of SOA particles by using the LISPA-MS technique can be used to clarify the
formation and transformation processes of SOA particles in smog chambers.
Organic compounds account for a large, sometimes even
dominant, fraction of the particulate matter in the air. They
influence the chemical and physical properties of aerosol par-
ticles and thus have effects on the atmosphere and climate
through interaction with reactive trace gases, water vapor,
clouds, precipitation, and radiation. Several scientific papers
have reviewed organic aerosol chemistry and physics in the
atmosphere.1–6 The source of the main fraction of the organic
aerosols is of secondary origin;5 in other words, most organic
aerosol components are chemically formed in the atmosphere.
The chemistry of secondary organic aerosol (SOA) formation
and transformation processes has been studied using smog
chambers.4 Chemical compositions of SOA particles formed
in chambers have been studied using off-line techniques,
which have provided important information about the chemi-
cal compositions of SOA particles. However, the off-line tech-
niques are time-consuming and tend to have sampling arti-
facts.2 In addition, it is difficult to measure the size and chemi-
cal compositions of individual SOA particles simultaneously
and in real time by using the off-line techniques.
Real-time aerosol mass spectrometry has been developing
since the mid 1990s. Most of the instruments can determine
the size and count, and analyze the mass of individual particle
constituents online with high time-resolution. The chemistry of
SOA particles has been studied using such instruments. For
example, by using an aerosol time-of-flight mass spectrometer
(ATOFMS), real-time detection of oligomers in SOA particles
formed from photooxidation of 1,3,5-trimethylbenzene and
NOx has been reported.7 The role of organic peroxides in
SOA formation for reactions of monoterpenes with O3 has
been elucidated using a thermal desorption particle beam mass
spectrometer (TDPBMS).8 The direct polymerization of iso-
prene and ꢀ-pinene on acidic aerosols has been studied using
an aerosol mass spectrometer (AMS).9 Reviews of real-time
aerosol mass spectrometry are available in the literature.10,11
We have recently developed a laser-ionization single-par-
ticle aerosol mass spectrometer (LISPA-MS) for real-time
single-particle measurements.12–14 The LISPA-MS is capable
of high time-resolution measurements to determine the size
and chemical compositions of individual particles. Either posi-
tive- or negative-ion mass spectra can be obtained for individ-
ual particles by changing the voltage polarity of the instru-
ment. Apparent changes in the chemical compositions of Asian
y Present address: Heat and Fluid Dynamics Department,
Research Laboratory, IHI Corporation, 1 Shin-Nakahara-cho,
Isogo, Yokohama 235-8501
yy Present address: Integrated Research Institute, Tokyo Institute
of Technology, 4259 Nagatsuta, Midori-ku, Yokohama 226-
8503
yyy Present address: Pioneering Research Unit for Next Genera-
tion, Kyoto University, Gokasho, Uji, Kyoto 611-0011
yyyy Present address: Department of Molecular Engineering,
Graduate School of Engineering, Kyoto University, Kyoto
615-8510
Published on the web January 10, 2008; doi:10.1246/bcsj.81.120

M. Narukawa et al.
Bull. Chem. Soc. Jpn. Vol. 81, No. 1 (2008)
121
dust particles transported from the Asian continent to Japan
have been observed by using the LISPA-MS.12 By measuring
the standard aerosol particles of dicarboxylic acids and
SOA particles formed from cyclohexene ozonolysis using the
LISPA-MS, it has been found that the negative-ion mass spec-
tra provide information about the molecular weights of the
organic compounds in the particles.13
Here, we present an application of the LISPA-MS to the
study of SOA particles formed in a smog chamber. The size
and chemical compositions of the SOA particles formed from
1,4-cyclohexadiene (CHD) ozonolysis were analyzed using the
LISPA-MS. CHD is a hydrocarbon containing two symmetric
non-conjugated double bonds in the six-membered ring, and a
simple surrogate of ꢁ-terpinene that is emitted to the atmo-
sphere as a biogenic hydrocarbon. The purposes of this paper
were to better understand the formation and transformation
processes of single SOA particles by using real-time measure-
ments and to demonstrate the ability of the LISPA-MS.
-6
1.2x10
r2 = 0.95
1.0
0.8
0.6
0.4
0.2
0.0
0.0
0.5
1.0
1.5
PSL diameter / µm
Fig. 2. PMT signal intensities of YAG laser scattered light
as a function of polystyrene latex (PSL) particle diameter.
The vertical bars show two-standard deviations (2ꢂ) for
the averaged data. The correlation coefficient (r2 ¼ 0:95)
is calculated from the weighted regression line by the size
of the error bars.
Experimental
Laser-Ionization Single-Particle Aerosol Mass Spectrome-
ter (LISPA-MS).
Although the details of the LISPA-MS
have been described in the literature,13 we briefly describe the
LISPA-MS. A schematic diagram of the particle inlet and the
ion source region of the LISPA-MS is shown in Fig. 1a. Aerosol
particles are introduced into an aerodynamic lens through a flow-
limiting molybdenum orifice with a diameter of 300 mm. In the
reduced pressure region, particles are horizontally focused into a
narrow particle beam (ꢁ1 mm diameter) in the center of the gas
flow by the aerodynamic lens.15 After exiting the lens, the particle
beam is transmitted into the ion source region through differen-
tially pumped stages with two skimmers (Precision Instrument
Services).
For detecting the particles introduced into the ion source re-
gion, a continuous-wave laser beam, produced using a frequen-
cy-doubled Nd:YAG laser at 532 nm, crosses the particle beam
as shown in Fig. 1a. The scattered light from the individual
particles at 532 nm is detected by a photomultiplier tube (PMT,
Hamamatsu, 1P28A). The intensity of the scattered light is ap-
proximately proportional to the particle size and is utilized for
roughly determining the particle size. Particle size calibration of
the LISPA-MS is performed by sampling dry polystyrene latex
particles (Duke Scientific and JSP). Figure 2 shows the PMT
signal intensities of YAG laser scattered light as a function of
polystyrene latex particle diameter. The correlation coefficient
(r2 ¼ 0:95) in Fig. 2 is calculated from the weighted regression
line. The detection efficiency for particles smaller than the wave-
length of the detection laser light is very low due to Mie scatter-
ing. Thus, the lower limit of the detected particle diameter is
ꢁ200 nm. The signals from the PMT are amplified and discrimi-
nated, and those above a selectable threshold level trigger desorp-
tion/ionization laser pulses. A pulsed KrF excimer laser at 248 nm
(Lambda Physik, OPTex) is used as a light source for the desorp-
tion/ionization of particles.
(a)
Photomultiplier Tube
Aerodynamic Lens
KrF Excimer Laser
248 nm (pulsed)
Nd:YAG Laser
532 nm (CW)
(b)
A diagram of the ion source region and reflectron time-of-flight
mass spectrometer of the LISPA-MS is shown in Fig. 1b. The
resultant ions are accelerated from the ion source region to the
reflector assembly and are then detected by a microchannel plate
detector (MCP, Hamamatsu). By monitoring the ion current from
the MCP as a function of time, a complete mass spectrum is ob-
tained for the chemical compositions of a single particle. The
ion signals from the detector are amplified linearly, digitized,
and recorded as mass spectra with a high time-resolution (ꢁ2 s
at the maximum). Positive or negative-ion mass spectra are ob-
tained for individual particles by changing the voltage polarity
of the ion source region and the TOF-MS.
Turbo Molecular Pump
Gate Valve
Microchannel Plate
Deflector
Reflector
KrF Excimer Laser
Turbo Molecular Pump
Fig. 1. Diagrams of (a) the particle inlet and the ion source
region of the laser-ionization single-particle aerosol mass
spectrometer (LISPA-MS) and (b) the ion source region
and reflectron time-of-flight mass spectrometer of the
LISPA-MS.
Smog Chamber Experiments. Laboratory studies were con-
ducted by using an indoor evacuable and bakable photochemical
smog chamber at National Institute for Environmental Studies

122 Bull. Chem. Soc. Jpn. Vol. 81, No. 1 (2008)
Single Particle Analysis of Organic Aerosols
Table 1. Initial Conditions for 1,4-Cyclohexadiene (CHD)
Ozonolysis in the NIES Smog Chamber
3
2
1
0
6
CHD
O3
SOA
Run No.
P0/Torr
T0/K
[CHD]0/ppmv
[O3]0/ppmv
1
2
3
4
763
773
771
771
298
297
298
298
2.09
2.07
2.06
2.07
1.93
1.99
2.01
1.98
4
2
0
(NIES). The chamber was a stainless steel cylinder having a
volume of 6.065 m3, a length of 3500 mm and an inner diameter
of 1450 mm. The inner surface of the chamber was Teflon-coated
(tetrafluoroethylene–perfluoroalkyl vinyl ether copolymer). The
details of the NIES smog chamber have been described in the
literature.16
0
20
40
60
The NIES smog chamber was filled with 760 Torr of dry puri-
fied air prior to introducing the reactants. Ozone was introduced
into the chamber from an ozone generator (Iwasaki, OP-20W).
The reaction was started following the introduction of 1,4-cyclo-
hexadiene (CHD; 96% purity, Wako Pure Chemical Industries),
which was used without further purification. The initial conditions
for each run of the CHD ozonolysis are given in Table 1. The
reactant mixture was stirred well with a mixing fan for a few
minutes, after introducing CHD into the chamber. Each run was
performed in the dark under dry conditions (relative humidi-
ty < 1 ppmv) without any seed particles or hydroxyl radical scav-
engers and was typically terminated within ꢁ60 min. During each
run, the SOA particles formed in the chamber were sampled
through a stainless steel tube that was 0.635 cm (0.25 inches) in
diameter and ꢁ3 m in length and were directly introduced to the
LISPA-MS. The concentrations of the gaseous reactants were
measured every 6 min with a FT-IR spectrometer (Nicolet, Nexus
670) equipped with a liquid N2-cooled MCT detector. The spectral
resolution was 1.0 cmꢂ1. The spectrometer was coupled to a white
mirror system (base path 1700 mm), which was operated at 130
traverses, giving a total absorption path length of 221.5 m, includ-
ing the transfer path. The size distribution and the number concen-
trations of the SOA particles were measured on a scanning mobi-
lity particle sizer instrument (SMPS; TSI, Model 3934), which
consists of an electrostatic classifier (TSI, Model 3071A) and a
condensation particle counter (CPC; TSI, Model 3025A). The de-
tails of SOA formation studies by using the NIES smog chamber
are described in the literature.17,18
Reaction time / min.
Fig. 3. Temporal profile of the concentrations of 1,4-cyclo-
hexadiene (CHD) ( ), ozone ( ), and secondary organic
aerosol (SOA) mass ( ) in CHD ozonolysis (Run No. 3)
conducted in the NIES smog chamber.
in Fig. 3.
CHD has two double bonds; hence, one molecule of CHD
can react with two molecules of O3. The concentration of O3
should decrease more quickly than that of CHD. However,
as seen in Fig. 3, CHD was consumed in a one-to-one reaction
with O3. In cycloalkene ozonolyses, hydroxyl radical (OH)
formation in the gas-phase in excellent yields has been report-
ed.20 Without OH scavengers in cycloalkene ozonolyses, the
concentrations of cycloalkenes decreases more quickly than
those of O3,21,22 that is, a part of the cycloalkenes is consumed
by reactions with OH. CHD ozonolysis can form OH as well,
which undergoes subsequent reactions with CHD. Therefore,
CHD can react with both O3 and OH. The depletion rate of
CHD in our present study can be expressed as
d½CHDꢃ
ꢂ
¼ k1½CHDꢃ½O3ꢃ þ k2½CHDꢃ½OHꢃ;
ð1Þ
dt
where k1 and k2 are the rate constants for the gas-phase reac-
tions of CHD with O3 and OH, respectively. The values of k1
and k2 at 298 K have been reported to be 4:6 ꢄ 10ꢂ17 and
9:95 ꢄ 10ꢂ11 cm3 moleculeꢂ1 sꢂ1, respectively.23,24 Integrating
and rearranging Eq. 1, we derived the following equation:25
Results and Discussion
ꢀ
ꢁ
Z
Ozonolysis of 1,4-Cyclohexadiene (CHD). As a represen-
tative example, temporal profiles of the concentrations (ppmv)
of CHD and O3 from CHD ozonolysis for Run No. 3 are
shown in Fig. 3. The concentrations of the reactants decreased
immediately after mixing the reactants in the chamber. A tem-
poral profile of the mass concentrations (mg mꢂ3) of the SOA
particles for Run No. 3 is also shown in Fig. 3. The SOA mass
concentrations were calculated from the electrical mobility di-
ameter of the particles, assuming a spherical shape and a den-
sity of 1.4 g cmꢂ3, which was determined from the particulate
products of cyclohexene ozonolysis.19 The mass concentra-
tions were also corrected for wall loss.17 The SOA particles
grew to the detectable size (’200 nm) of the LISPA-MS over
a few minutes after the reactants were mixed. The SOA mass
concentrations remarkably increased within ꢁ15 min from
the start of the reaction, followed by a little increase, as shown
t
[CHD]t
ln
þ k1 ½O3ꢃdt ¼ ꢂk2½OHꢃavt;
ð2Þ
½CHDꢃ0
0
where [OH]av is the average concentration of OH radicals,
which react with CHD between t ¼ 0 and t. Using the values
of k1 and k2 and the measured concentrations of CHD and O3,
we could calculate the OH concentration by using Eq. 2.25,26
The average concentration of OH within the reaction time
was estimated to be ꢁ2 ꢄ 106 molecule cmꢂ3 (ꢁ0:1 pptv),
suggesting that the concentrations of OH is about six orders
of magnitude lower than those of O3. On the other hand, the
rate constant of the reaction of CHD with OH is six orders
of magnitude higher than that with O3. Thus, OH should
consume the same amount of CHD as consumed by O3.
As an example of the reactions in the CHD ozonolysis, a
reaction pathway that forms malonic acid (MW 104), the ion

M. Narukawa et al.
Bull. Chem. Soc. Jpn. Vol. 81, No. 1 (2008)
123
(a)
O3
radical-radical
reactions
radical-radical
reactions
O
O
O
O3
CHOO
CHO
COOH
COOH
, etc.
HOOC
COOH
radical-molecule
reactions
radical-molecule
reactions
C
A
B
(b)
A: 1,4-cyclohexadiene (M.W. 80)
B: Z-3-hexenedioic acid (M.W. 144)
C: malonic acid (M.W. 104)
D: benzene (M.W. 78)
OH
O2
HO2
H2O
A
D
Fig. 4. (a) Formation pathway for malonic acid (MW 104) during CHD ozonolysis. The double dagger z denotes an energy-rich
Criegee intermediate. (b) Formation pathway for benzene during reaction of CHD with OH.
peak at m=z ¼ 103, which corresponds to that which appeared
43
(a)
as a strong peak in the negative-ion mass spectra of the SOA
particles from CHD ozonolysis, as discussed later, is shown
in Fig. 4a. CHD has two symmetric non-conjugated double
bonds in the six-membered ring, so that O3 attacks a double
bond of CHD, leading to an energy-rich Criegee intermedi-
ate.24 Further reactions involving the intermediate afford Z-
3-hexenedioic acid (MW 144), which is a product of the first
ozonolysis. Ozone also attacks the double bond of Z-3-hexene-
dioic acid, followed by other reactions, to give malonic acid,
as shown in Fig. 4a. As a result, organic compounds with two
functional groups (aldehyde and/or carboxyl groups), such as
malonic acid, are expected to be the major products in CHD
ozonolysis. In addition, OH can attack CHD as discussed
above. As an example of the reactions of CHD with OH, a
reaction pathway that forms benzene is shown in Fig. 4b.
Benzene is a major gaseous product in the reactions of CHD
with OH.27 Thus, the products formed from the reactions of
CHD with O3 and OH should be a part of the SOA particles.
Mass Spectra of SOA Particles Measured by the LISPA-
MS. Figure 5 shows positive-ion (Run No. 2, detected at
30.7 min after the reaction start) and negative-ion (Run No. 1,
59.5 min) mass spectra of the SOA particles. In the positive-
ion mass spectra of the SOA particles formed during CHD
ozonolysis, the mass peaks at m=z ¼ 39 and 43 were dominant,
as shown in Fig. 5a. The ions at m=z ¼ 39 and 43 were as-
39
(b)
45
73
59
103
87
143
0
40
80
m/z
120
Fig. 5. Representative mass spectra obtained from each sin-
gle SOA particle formed from CHD ozonolysis: (a) posi-
tive ion (Run No. 2, detected at 30.7 min after the reaction
start) and (b) negative ion (Run No. 1, 59.5 min).
45
131
115
145
þ
signed to be C3H3þ, and C3H7 and/or CH3COþ, respective-
ly, which are obviously fragment ions. This result is consistent
with those from our previous study.13 In our previous study,13
the standard aerosol particles of dicarboxylic acids and SOA
particles formed from cyclohexene ozonolysis have been
analyzed using the LISPA-MS, and we have found that the
intense mass peaks in the positive-ion spectra of those particles
consist of fragment ions, and not molecular and molecular-
related ions.
In the negative-ion mass spectra of the SOA particles
formed from the CHD ozonolysis, a strong mass peak at m=z ¼
103 appeared, as shown in Fig. 5b. The ion at m=z ¼ 103 can
be formed by proton elimination from malonic acid (MW 104),
which is easily formed from CHD ozonolysis, as shown in
Fig. 4a. In our previous study,13 we have also found that
molecular-related ions ½M ꢂ Hꢃꢂ are observed with relatively
strong intensities in the negative-ion spectra of standard aero-
sol particles of dicarboxylic acids, including malonic acid. It
has also been found that the ions, which appeared as strong
mass peaks in the negative-ion spectra of SOA particles
0
50
100
m/z
150
Fig. 6. Representative negative-ion mass spectrum ob-
tained from a single SOA particle formed from cyclo-
hexene ozonolysis.
formed from cyclohexene ozonolysis, consist of the molecu-
lar-related ions ½M ꢂ Hꢃꢂ of the reaction products and
fragment ions,13 as shown in the negative-ion spectrum of
the SOA particles formed from the cyclohexene ozonolysis
in Fig. 6. Thus, the strong mass peak at m=z ¼ 103 in the
negative-ion spectra from CHD ozonolysis should correspond
to malonic acid. It is likely that strong mass peaks in the
negative-ion spectra from the CHD ozonolysis correspond to
molecular-related ions ½M ꢂ Hꢃꢂ of the reaction products

124 Bull. Chem. Soc. Jpn. Vol. 81, No. 1 (2008)
Single Particle Analysis of Organic Aerosols
Table 2. Intense Mass Peaks (m=z) in Negative-Ion Mass
Spectra and Corresponding Ions
45
45
(a) 0.3 µm
43
m=z
Ions
CH3COꢂ
HCOOꢂ
73
59
59
43
45
59
71
73
87
103
71
103
103
CH3COOꢂ
OHCCH2COꢂ and/or C3H7COꢂ
OHCCOOꢂ and/or C2H5COOꢂ
OHCCH2COOꢂ and/or C3H7COOꢂ
HOOCCH2COOꢂ
(b) 2.1 µm
73
and fragment ions as well. Therefore, the negative-ion mass
spectra provide important information for investigating the
SOA formation processes in CHD ozonolysis.
43
71
In the negative-ion mass spectra from cyclohexene ozonol-
ysis, the ion peak at m=z ¼ 145, which corresponds to adipic
acid (MW 146) that is formed via ozonolysis, was intense,
as shown in Fig. 6. On the other hand, in the negative-ion
spectra from CHD ozonolysis, the ion peak at m=z ¼ 103, cor-
responding to malonic acid was intense, whereas the ion peak
at m=z ¼ 143, corresponding to Z-3-hexenedioic acid (MW
144), which is formed via the first ozonolysis as shown in
Fig. 4a, was weak, as shown in Fig. 5b. The strong mass peak
at m=z ¼ 103 and the weak one at m=z ¼ 143 suggest that
the second ozonolysis in the CHD ozonolysis is rapid; namely,
the products of the first ozonolysis are quickly consumed by
O3 and/or OH.
The strong mass peaks at m=z ¼ 45, 59, 73, 87, and 103 in
the negative-ion mass spectra from CHD ozonolysis have also
been observed in the negative-ion spectra from cyclohexene
ozonolysis, as shown in Fig. 6.13 The compounds, which cor-
respond to those strong mass peaks in the negative-ion spectra
from cyclohexene ozonolysis, have been explained to be oxy-
genated organic compounds with aldehyde and carboxyl
groups.13 Thus, the strong peaks in the negative-ion spectra
from CHD ozonolysis can also be interpreted as being due to
oxygenated organic compounds with aldehyde and carboxyl
groups. Table 2 lists the intense mass peaks, which appeared
in the negative-ion spectra from the CHD ozonolysis, and
the corresponding ions. The ions corresponding to the peaks
at m=z ¼ 45 and 59 are HCOOꢂ and CH3COOꢂ, respectively,
and the ions corresponding to the peaks at m=z ¼ 73 and 87 are
OHCCOOꢂ and/or C2H5COOꢂ, and OHCCH2COOꢂ and/or
C3H7COOꢂ, respectively. Accordingly, it is suggested that
all of these ions are directly and/or indirectly formed via
proton elimination of the compounds with carboxyl groups.
On the other hand, the ion corresponding to the peak at
m=z ¼ 43 is CH3COꢂ, and the ions corresponding to the peak
at m=z ¼ 71 are OHCCH2COꢂ and/or C3H7COꢂ. Similarly,
it is suggested that these ions are directly and/or indirectly
formed via proton elimination of the compounds with aldehyde
groups.
40
60
80
100
m/z
120
140
Fig. 7. Averaged negative-ion mass spectra, in which
(a) the 40 smallest particles were averaged, and their mean
diameter was 0.3 mm, and (b) the 40 largest particles
were averaged, and their mean diameter was 2.1 mm. Data
obtained from all experiment (Run Nos. 1–4) were com-
bined and averaged.
and their mean diameter was 0.3 mm, and (b) the 40 largest
particles were averaged and their mean diameter was 2.1 mm.
Data obtained from all of the experiments (Run Nos. 1–4) were
combined and averaged. The signal intensity ratios in the range
of m=z from 45 to 43 and from 73 to 71 in the negative-ion
spectra generally changed with the particle size, as shown in
Fig. 7. It is thought that the peaks at m=z ¼ 43 and 71 corre-
spond to compounds with aldehyde groups, whereas the peaks
at m=z ¼ 45 and 73 correspond to compounds with carboxyl
groups, as discussed above. The carboxyl group can form from
the oxidation of an aldehyde group. Therefore, the differences
in the signal intensity ratios observed in the negative-ion
spectra imply that the SOA particles with relatively high signal
intensity ratios consist of the products formed from more pro-
gressed oxidation reactions than those with relatively low sig-
nal intensity ratios, whereas the absolute ionization response of
those compounds in the instrument have not been determined.
The signal intensity (peak area) ratios of the negative ions
(45/43 and 73/71) as a function of particle size were investi-
gated. Figure 8 shows plots of the (a) 45/43 and (b) 73/71 sig-
nal intensity ratios (mean ꢅ 1ꢂ) from the negative-ion mass
spectra against the diameters of the SOA particles. The ratios
obtained from all of the experiments (Run Nos. 1–4) were
combined and averaged every forty particles in order of the
particle size. Figure 8a shows that the mean values and one-
standard deviations of the 45/43 intensity ratios increased as
the particle size increased. The 45/43 ratios had a low value
of 1:9 ꢅ 0:8 for a particle size of 0.3 mm and a high value of
4:6 ꢅ 3:4 for a particle size of 2.1 mm, as shown in Fig. 8a.
The mean values of the 73/71 intensity ratios increased more
pronouncedly as the particle size increased, whereas the one-
standard deviations were almost constant, as shown in Fig. 8b.
The 73/71 intensity ratios had a low value of 1:8 ꢅ 0:8 for a
particle size of 0.3 mm and a high value of 3:1 ꢅ 0:8 for a
Size Dependence of Negative-Ion Mass Spectra of SOA
Particles. Distributions of the ion signal intensities in the
negative-ion mass spectra from the CHD ozonolysis changed
as a function of the particle size, which was roughly deter-
mined from the scattered light intensity of the SOA particles,
as shown in Fig. 2. Figure 7 shows averaged negative-ion
spectra, in which (a) the 40 smallest particles were averaged

M. Narukawa et al.
Bull. Chem. Soc. Jpn. Vol. 81, No. 1 (2008)
125
10
8
5
4
3
2
1
0
(a)
(b)
6
4
2
0
0.0 0.5 1.0 1.5 2.0 2.5
0.0 0.5 1.0 1.5 2.0 2.5
Diameter / µm
Diameter / µm
Fig. 8. Signal intensity (peak area) ratios (mean ꢅ 1ꢂ) of (a) m=z ¼ 45 to 43 and (b) m=z ¼ 73 to 71 in the negative-ion mass
spectra of SOA particles formed from the CHD ozonolysis as a function of SOA particle size (mean ꢅ 1ꢂ), namely, PMT signal
intensities. Data obtained from all experiment (Run Nos. 1–4) were combined and averaged every forty particles in order of the
particle size.
2
34, 2983.
B. J. Turpin, P. Saxena, E. Andrews, Atmos. Environ. 2000,
particle size of 2.1 mm. It is likely that as the particle size in-
creases, the signal intensities (m=z ¼ 45 and 73) of the ions of
the compounds with carboxyl groups are relatively higher than
those (m=z ¼ 43 and 71) of the ions of the compounds with
aldehyde groups. These results suggest that the SOA particles
contain more oxygenated organic compounds as the particle
sizes increase; namely, the chemical compositions of the
SOA particles vary as a function of the particle size.
3
4
Y. Rudich, Chem. Rev. 2003, 103, 5097.
J. H. Seinfeld, J. F. Pankow, Annu. Rev. Phys. Chem. 2003,
54, 121.
5
M. Kanakidou, J. H. Seinfeld, S. N. Pandis, I. Barnes, F. J.
Dentener, M. C. Facchini, R. Van Dingenen, B. Ervens, A. Nenes,
C. J. Nielsen, E. Swietlicki, J. P. Putaud, Y. Balkanski, S. Fuzzi,
J. Horth, G. K. Moortgat, R. Winterhalter, C. E. L. Myhre, K.
Tsigaridis, E. Vignati, E. G. Stephanou, J. Wilson, Atmos. Chem.
Phys. 2005, 5, 1053.
Conclusion
In this paper, we demonstrated the applicability of the
LISPA-MS technique to the study of the SOA particles formed
from CHD ozonolysis. The negative-ion mass spectra of the
SOA particles provided important information about the chem-
ical compositions of the SOA particles. Mass spectral analysis
of the SOA particles suggests that the second ozonolysis is
rapid, that is, the products of the first ozonolysis are quickly
consumed by O3 and/or OH. In the negative-ion spectra, the
intense mass peaks were explained to correspond to ions with
aldehyde and carboxyl groups, which form from the reactions
of CHD with O3 and OH. The distributions of intense mass
peaks in the negative-ion spectra varied as a function of the
particle size. In other words, the signal intensities of the peaks
(m=z ¼ 45 and 73) of the ions with carboxyl groups were
relatively higher than those (m=z ¼ 43 and 71) of the ions
with aldehyde groups as the particle size increased. This result
indicates that the chemical composition of the SOA particles
varies as a function of the particle size. It was demonstrated
that the real-time, single-particle analysis of the SOA particles
with high time-resolution by using the LISPA-MS technique
is useful to investigate the formation and transformation
processes of SOA particles in smog chambers.
6
S. Fuzzi, M. O. Andreae, B. J. Huebert, M. Kulmala, T. C.
Bond, M. Boy, S. J. Doherty, A. Guenther, M. Kanakidou, K.
Kawamura, V. M. Kerminen, U. Lohmann, L. M. Russell, U.
Poschl, Atmos. Chem. Phys. 2006, 6, 2017.
7
J. Dommen, M. R. Alfarra, J. Duplissy, K. Gaeggeler, A. Gascho,
D. S. Gross, M. E. Galli, M. Kalberer, A. S. H. Prevot,
¨
A. Metzger, U. Baltensperger, Anal. Chem. 2006, 78, 2130.
8
K. S. Docherty, W. Wu, Y. B. Lim, P. J. Ziemann, Environ.
Sci. Technol. 2005, 39, 4049.
J. Liggio, S.-M. Li, J. R. Brook, C. Mihele, Geophys. Res.
9
Lett. 2007, 34, L05814.
10 D. G. Nash, T. Baer, M. V. Johnston, Int. J. Mass. Spec-
trom. 2006, 258, 2.
11 D. M. Murphy, Mass Spectrom. Rev. 2007, 26, 150.
12 J. Matsumoto, K. Takahashi, Y. Matsumi, A. Yabushita, A.
Shimizu, I. Matsui, N. Sugimoto, Geophys. Res. Lett. 2006, 33,
L07816.
13 M. Narukawa, Y. Matsumi, J. Matsumoto, K. Takahashi,
A. Yabushita, K. Sato, T. Imamura, Anal. Sci. 2007, 23, 507.
14 M. Narukawa, Y. Matsumi, K. Takahashi, A. Yabushita,
Chem. Lett. 2007, 36, 904.
15 X. Zhang, K. A. Smith, D. R. Worsnop, J. Jimenez, J. T.
Jayne, C. E. Kolb, Aerosol Sci. Technol. 2002, 36, 617.
16 H. Akimoto, M. Hoshino, G. Inoue, F. Sakamaki, N.
Washida, M. Okuda, Environ. Sci. Technol. 1979, 13, 471.
17 K. Sato, B. Klotz, S. Hatakeyama, T. Imamura, Y.
Washizu, Y. Matsumi, N. Washida, Bull. Chem. Soc. Jpn. 2004,
77, 667.
This work was supported by SENTAN, Japan Science and
Technology Agency (JST).
References
1 M. C. Jacobson, H. C. Hansson, K. J. Noone, R. J.
Charlson, Rev. Geophys. 2000, 38, 267.
18 K. Sato, Chem. Lett. 2005, 34, 1584.
19 M. Kalberer, J. Yu, D. R. Cocker, R. C. Flagan, J. H.

126 Bull. Chem. Soc. Jpn. Vol. 81, No. 1 (2008)
Single Particle Analysis of Organic Aerosols
Seinfeld, Environ. Sci. Technol. 2000, 34, 4894.
20 J. D. Fenske, K. T. Kuwata, K. N. Houk, S. E. Paulson,
J. Phys. Chem. A 2000, 104, 7246.
21 S. Hatakeyama, T. Tanonaka, J. Weng, H. Bandow, H.
Takagi, H. Akimoto, Environ. Sci. Technol. 1985, 19, 935.
22 S. Hatakeyama, M. Ohno, J. Weng, H. Takagi, H.
Akimoto, Environ. Sci. Technol. 1987, 21, 52.
23 C. R. Greene, R. Atkinson, Int. J. Chem. Kinet. 1992,
24, 803.
24 J. G. Calvert, R. Atkinson, J. A. Kerr, S. Madronich, G. K.
Moortgat, T. J. Wallington, G. Yarwood, The Mechanisms of
Atmospheric Oxidation of the Alkenes, Oxford University Press,
New York, 2000.
25 H. Akimoto, F. Sakamaki, G. Inoue, M. Okuda, Environ.
Sci. Technol. 1980, 14, 93.
26 F. Sakamaki, H. Akimoto, M. Okuda, Environ. Sci.
Technol. 1981, 15, 665.
27 T. Ohta, Int. J. Chem. Kinet. 1984, 16, 1495.