Effect of aerosol chemical composition on the photodegradation of nitro- polycyclic aromatic hydrocarbons

Article abstract of DOI:10.1021/es990566r

The photodegradation of four ring nitro-PAHs in the presence of organic aerosol constituents has been investigated in model systems to determine their degradation mechanism under ambient air conditions. Light-induced radical chain reactions initiated by organic aerosol constituents, e.g., oxy- PAHs, is most likely to be the dominant degradation pathway for particle- associated nitro-PAH. The photodegradation rates were investigated in a chemical model system simulating the liquid film on particles from diesel exhaust and wood stove stack gases. Relative rates were obtained in pure solutions (direct photolysis) and in the presence of different classes of organic aerosol components comprising PAHs, hydroxy-PAHs, substituted phenols, benzaldehydes, and oxy-PAHs (including polycyclic aromatic quinones) (indirect photolysis). Members of the four latter compound classes are demonstrated to accelerate the photodegradation rate. The mechanisms proposed to explain these effects include both radical chain reactions and photoinduced hydrogen abstraction by nitro-PAHs. PAHs that are known sensitizers of singlet O₂ did not accelerate the decay. Certain oxy-PAHs strongly accelerate the nitro-PAH decay, whereas others have no effect. This difference is related to the nature of the excited state of the oxy-PAH and the ability to initiate radical chain reactions. The relevance of the model system and the environmental implications of the results are discussed, and the results are compared with relative degradation rates from smog chamber studies with the same compounds. The photodegradation of four ring nitro-PAHs in the presence of organic aerosol constituents has been investigated in model systems to determine their degradation mechanism under ambient air conditions. Light-induced radical chain reactions initiated by organic aerosol constituents, e.g., oxy-PAHs, is most likely to be the dominant degradation pathway for particle-associated nitro-PAH. The photodegradation rates were investigated in a chemical model system simulating the liquid film on particles from diesel exhaust and wood stove stack gases. Relative rates were obtained in pure solutions (direct photolysis) and in the presence of different classes of organic aerosol components comprising PAHs, hydroxy-PAHs, substituted phenols, benzaldehydes, and oxy-PAHs (including polycyclic aromatic quinones) (indirect photolysis). Members of the four latter compound classes are demonstrated to accelerate the photodegradation rate. The mechanisms proposed to explain these effects include both radical chain reactions and photoinduced hydrogen abstraction by nitro-PAHs. PAHs that are known sensitizers of singlet O₂ did not accelerate the decay. Certain oxy-PAHs strongly accelerate the nitro-PAH decay, whereas others have no effect. This difference is related to the nature of the excited state of the oxy-PAH and the ability to initiate radical chain reactions. The relevance of the model system and the environmental implications of the results are discussed, and the results are compared with relative degradation rates from smog chamber studies with the same compounds.

Full text of DOI:10.1021/es990566r

Environ. Sci. Technol. 2000, 34, 789-797
(
13) to condense on airborne particles immediately, so that
Effect of Aerosol Chemical
Composition on the
none of these compounds were detected in the gas phase.
To study the fate of the nitro-PAHs consisting of four or
more rings it is therefore only relevant to take into account
reactions in condensed phases. The presence of four ring
nitro-PAHs is important from an environmental health
perspective. They are among the most abundant particle
associated nitro-PAHs in ambient air (8, 12, 14). Furthermore,
the four ring nitro-PAHs appear to be more mutagenic than
the others (2). Smog chamber studies (1) have revealed that
when sorbed to diesel exhaust and wood smoke particles
photolysis is the major degradation pathway for the nitrof-
luoranthenes and -pyrenes. However, it was observed that
the photolysis rates were independent of molecular con-
formation (1). This is surprising since studies of the pho-
tochemistry of pure nitro-PAHs in solution or adsorbed in
pure state have shown that the torsion angle of the nitro
group relative to the aromatic plane governs the photostability
Photodegradation of Nitro-polycyclic
Aromatic Hydrocarbons
A N D E R S F E I L B E R G * A N D
T O R B E N N I E L S E N
PBK-313, Risø National Laboratory, PO Box 49,
DK-4000 Roskilde, Denmark
The photodegradation of four ring nitro-PAHs in the
presence of organic aerosol constituents has been
investigated in model systems to determine their degradation
mechanism under ambient air conditions. Light-induced
radical chain reactions initiated by organic aerosol
constituents, e.g., oxy-PAHs, is most likely to be the dominant
degradation pathway for particle-associated nitro-PAH.
The photodegradation rates were investigated in a chemical
model system simulating the liquid film on particles from
diesel exhaust and wood stove stack gases. Relative rates
were obtained in pure solutions (direct photolysis) and
in the presence of different classes of organic aerosol
components comprising PAHs, hydroxy-PAHs, substituted
phenols, benzaldehydes, and oxy-PAHs (including
polycyclic aromatic quinones) (indirect photolysis).
Members of the four latter compound classes are
demonstrated to accelerate the photodegradation rate.
The mechanisms proposed to explain these effects include
both radical chain reactions and photoinduced hydrogen
abstraction by nitro-PAHs. PAHs that are known sensitizers
of singlet O2 did not accelerate the decay. Certain oxy-
PAHs strongly accelerate the nitro-PAH decay, whereas
others have no effect. This difference is related to the nature
of the excited state of the oxy-PAH and the ability to
initiate radical chain reactions. The relevance of the model
system and the environmental implications of the results
are discussed, and the results are compared with relative
degradation rates from smog chamber studies with the
same compounds.
(
15-18). Therefore, it is necessary to study the mechanism
of photodegradation of nitro-PAHs in model systems simu-
lating diesel soot and wood smoke particles. The photo-
degradation of unsubstituted PAHs on diesel soot and wood
smoke (and possibly other organic aerosols) seems to follow
a complex mechanism, that is likely to include other chemical
aerosol components (19-21).
Combustion generated aerosols are among the major
contributors to urban fine particles (22). Several of these
particles (including diesel soot and wood smoke) as well as
ambient soot particles are believed to consist of an elemental
carbon core coated by an organic liquidlike layer (20, 21).
Diffusion in the organic liquid film (23) will allow electroni-
cally excited molecules to participate in bimolecular reactions
with other aerosol constituents.
This work describes studies of the photochemistry of
representative particle-associated nitro-PAH in a chemical
model system consisting of an organic solvent and known
constituents of organic aerosols such as diesel exhaust
particles and wood smoke particles. The compound classes
studied include substituted phenols, PAHs, hydroxy-PAHs,
oxy-PAHs, and substituted benzaldehydes.
Experimental Section
Irradiations were performed with a turntable photoreactor
using a 450 W medium-pressure mercury lamp (Ace Hanovia)
as light source (20). The lamp was placed in a quartz
immersion well with cooling water. The immersion well was
surrounded by a Pyrex sleeve to filter out high-energy UV
bands (λ > ∼290 nm). Although this does not resemble the
spectral distribution of the actinic flux, it does reproduce the
wavelengths normally encountered in the atmosphere. The
turntable-reactor technique has been introduced to the study
of environmental photochemistry by Hoign e´ and co-workers
Introduction
Nitropolycyclic aromatic hydrocarbons (nitro-PAHs) com-
prise a group of potent mutagenic and carcinogenic com-
pounds that potentially pose a threat to human health (1-
(
see for example refs 24 and 25).
The photoreactor was positioned in a water bath with
6
). Nitro-PAHs are present in ambient air both as a result of
constant water circulation. The circulation and temperature
in the water bath was maintained with a combined circulation
and temperature control system (HETO). All experiments
were performed with a temperature of 20 ( 1 °C. Samples
of nitro-PAHs with and without cosolutes were irradiated in
parallel in 13 × 100 mm quartz reaction tubes in cyclohexane.
The reaction tubes were sealed but contained sufficient
headspace to keep the solutions saturated with air during
the experiment.
incomplete combustion, primarily by diesel engines (7), and
as a result of atmospheric processes (8-10). The phase
distribution of nitro-PAHs in the atmosphere has only been
systematically investigated in a few cases. 1- and 2-Ni-
tronaphthalene are mainly present in the gas phase (11),
whereas the heavier nitrofluoranthenes and -pyrenes exist
almost exclusively in the particle phase (12, 13). In fact,
2
-nitrofluoranthene and 2-nitropyrene formed in the gas
phase by OH initiated reactions were observed by Fan et al.
The reproducibility of the irradiation conditions within
experiments was determined by comparing the 1-nitropyrene
decay rate constant in parallel samples. The variation was
*
Corresponding author phone: (+45) 46 77 42 01; fax (+45) 46
7
7 42 02; e-mail: anders.feilberg@risoe.dk.
1
0.1021/es990566r CCC: $19.00

2000 Am erican Chem ical Society
VOL. 34, NO. 5, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7 8 9
Published on Web 01/22/2000

less than 10%. Some variation in the 1-nitropyrene rate
constant was observed between different experiments.
However, absolute decay rates were only compared within
each experiment to determine the effect of the cosolutes.
Furthermore, the decay rate constant for 1-nitropyrene in
the presence of phenol relative to the rate constant without
phenol varied ∼5% for three different experiments, even
though the effect of phenol is not very great (a factor of ∼1.3).
This shows that the relative effect of the cosolutes can be
compared between experiments. A few test experiments were
performed with a monochromaticlight source (Bausch&Lomb).
This setup consists of a SP200 Hg light source with the
emission focused into a monochromator to produce light
with the desired wavelength. A wavelength of 366 nm was
used to selectively produce excited nitro-PAHs and avoid
photolysis of the cosolutes. Samples were irradiated in pairs
in quartz cuvettes. One sample contained only the nitro-
PAH, in the other a cosolute was added. To test that the
photon flux into the two cuvettes were the same, two tests
were performed. First, two solutions of the classic ferrioxalate
actinometer (26) were irradiated, and second, identical
solutions of 1-NP were irradiated. The production of Fe2
Semivolatile organic compounds (SVOC) appear to have
higher affinity for aerosols containing organic matter and
will preferentially condense on these as opposed to inorganic
aerosols (32, 33). Low-volatility nitro-PAHs formed in the
gas phase by reactions of the parent PAHs are therefore
expected mainly to be absorbed into the organic fraction of
ambient particles. Likewise, the heavier nitro-PAHs from
combustion sources are expected to be associated with the
organic fraction of soot particles.
Based on the above discussion, an organic solvent with
relevant constituents of combustion aerosols seems a relevant
system for studying the photodecomposition of nitro-PAHs
associated with urban particulate matter. This methodology
has previously been applied to study the photodegradation
of unsubstituted PAHs (19-21, 34, 35).
Cyclohexane was used as solvent because it is transmissive
and inert. In addition the organic fraction of diesel exhaust
particles consists largely of branched and linear alkanes (36),
although as a whole it is more polar than cyclohexane. It
should be noted that ambient organic aerosols such as diesel
exhaust might be more viscous than cyclohexane. This
physical parameter is probably also important, but in the
present study, we only intend to address the effect of chemical
composition. The cosolutes chosen are all major components
of combustion derived and/ or ambient particulate matter.
In the experiments the concentration of nitro-PAHs were
always much lower than the concentration of cosolutes,
which seems the most realistic with respect to ambient air.
Most of the experiments were performed with 1-nitro-
pyrene (1-NP) and 3-nitrofluoranthene (3-NF) as represen-
tative nitro-PAHs, but 2-nitrofluoranthene (2-NF) and 2-ni-
tropyrene (2-NP) were also studied.
The cosolutes chosen in this investigation are listed in
Table 1 together with ambient sources. Known constituents
of wood smoke or diesel particles were primarily chosen,
since these types of particles were used in previous smog
chamber experiments with nitro-PAHs (1). However, several
of these compound classes are also found on other combus-
tion derived aerosols, e.g., from gasoline vehicles (36). Wood
smoke particles are composed of compound derived from
combustion of biopolymers. Particularly the occurrence of
methoxyphenols (37) arising from combustion of lignin is
interesting from a chemical viewpoint (20, 21). The organic
phase of diesel particles is believed to be nonpolar, but
substituted phenols and hydroxy-PAHs have been detected
in diesel standard reference material (38). PAHs, substituted
benzaldehydes, and oxy-PAHs are found on diesel particles
and wood smoke particles as well as on particles from gasoline
vehicles (36, 37). It should be noted that there are still large
uncertainties regarding the chemical composition of diesel
soot and wood smoke particles (36, 37) as well as other
combustion derived particles.
+
(
complexed with 1,10-phenanthrolin) and the loss of 1-NP
were spectrophotometrically determined. The difference
between the two cuvettes was less than 5%.
-
4
The concentrations of nitro-PAHs were ∼10 M, and the
-
3
-3
cosolute concentrations were about 1 × 10 -5 × 10 M.
In experiments with polar cosolutes, small amounts of
dichloromethane or ethyl acetate were added to all samples
in equal amounts in order to dissolve the cosolutes. Samples
were removed with a pipet and transferred to brown
chromatography vials. The samples were analyzed by HPLC
using a Shimadzu HPLC (LC10-AD) equipped with diode
array UV-vis detection (SPD-M10A). The column was packed
2
with Nucleosil 50-5 C18. Methanol containing 10 or 20% H O
was used as isocratic eluent. The response of the analytical
2
system was a linear function of concentration (r > 0.99).
2
-Nitrofluoranthene (2-NF) was synthesized following the
procedure of Zielinska et al. (27): Fluoranthene was nitrated
with freshly prepared N (28) in CCl at 25 °C. 2-NF was
2
O
5
4
isolated from dinitrofluoranthenes and unreacted fluoran-
thene by preparative HPLC fractionation using a Nucleosil
5
0-5 column and a Gilson (model 201) fraction collector
connected to the analytical HPLC described above. All other
chemicals were commercially available and were used as
received.
Results and Discussion
Relevance of the Model System . As discussed by McDow et
al. (20, 21) there is increasing evidence that at least for some
aerosols formed by combustion of organic matter, e.g., diesel
and wood smoke particles, the organic fraction is liquid or
liquidlike. This evidence includes that the impaction of wood
and diesel soot on a metal foil produced a viscous liquid
slick that retains no or little particle integrity (20). Further-
more, transmission electron micrographs of ambient soot
particles showed that the soot was coated by a liquid film
Photolysis in the Absence of Cosolutes. 1-NP, 2-NP, 2-NF,
and 3-NF were irradiated in cyclohexane solutions in order
to investigate the photostability of the individual compounds.
For comparison 9-nitroanthracene (9-NA) and 6-nitrobenzo-
[a]pyrene (6-NBaP) were also irradiated. In the absence of
cosolutes, the photodegradation of nitro-PAHs is strongly
dependent on the orientation of the nitro-group. In 9-NA
and 6-NBaP the nitro-group adapts an approximately
perpendicular orientation relative to the aromatic plane (15,
39, 40), and as expected both compounds decayed very fast
with less than 1% remaining after 15 min. As for the other
nitro-PAHs, 1-NP decayed moderately (see Figure 1), whereas
2-NF, 2-NP, and 3NF were stable toward photolysis within
the course of the experiment. In 1-NP the nitro-group is
twisted out of the aromatic plane due to a peri-proton (39).
In 3-NF the nitro-group is probably twisted slightly but much
less than in 1-NP (39) and in 2-NP and 2-NF the nitro-group
lies in the same plane as the aromatic rings (39). It has been
(
29). Depth profiles of ambient particulate matter have
revealed that the surface of particles collected in polluted
urban air was dominated by organic carbon (30). The
observation by several authors that absorption is probably
the most important sorption mechanism for ambient air (31-
3
3) also suggests that an organic liquidlike or amorphous
surface layer is involved. Diffusion in the organic fraction of
diesel soot and wood smoke particles has been demonstrated
by modeling the dynamic behavior of gas-particle partitioning
of PAHs (23). The diffusion is retarded by viscosity, tortuosity,
and adsorption (23) but will nevertheless allow electronically
excited molecules to undergo bimolecular reactions.
7
9 0
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 5, 2000

TABLE 1. Tested Aerosol Constituents, Their Sources and Effects on Nitro-PAH Photodegradation
compound
sourcesa
effect on nitro-PAH decay
proposed mechanism
benzene-1,4-diol (hydroquinone)
W
B
accelerated
accelerated
slightly acc.
no effect
H-abstraction
H-abstraction
H-abstraction
4
4
4
4
4
1
2
2
3
4
4
4
2
2
1
9
9
9
9
7
-chlorophenol
-hydroxybenzoic acid
-hydroxybenzaldehyde
-nitrophenol
D, P
D, P
D, P
D, P
D, P
D, P
G
W
W
W
W
no effect
-hydroxybiphenyl
accelerated
accelerated
accelerated
accelerated
slightly acc.
acceleratedb
acceleratedb
acceleratedb
acceleratedb
acceleratedb
highly acc.
highly acc.c
highly acc.
acceleratedc
no effect
H-abstraction
H-abstr/rad. react.
H-abstr/rad. react.
radical react.
radical react.
H-abstraction
H-abstraction
H-abstraction
H-abstraction
H-abstraction
radical react.
radical react.
radical react.
radical react.
-hydroxynaphthalene (1-naphthol)
-hydroxynaphthalene (2-naphthol)
,4,6-trim ethylbenzaldehyde
,4,5-trim ethoxybenzaldehyde
-hydroxy-3-m ethoxybenzaldehyde (vanillin)
-hydroxy-3-m ethoxybenzoic acid (vanillic acid)
-hydroxy-3-m ethoxybenzyllic acid (hom ovanillic acid)
,6-dim ethoxyphenol (syringol)
-m ethoxyphenol (guaiacol)
0H-anthracene-9-one (anthrone)
H-xanthen-9-one (xanthone)
,10-anthracenedione (anthraquinone)
,10-phenanthrenedione (9,10-phenanthrenequinone)
H-fluoren-9-one (fluorenone)
W
W
D, G, NG
D, G, NG
W, D, G, NG, B, P
W, D, G, NG, P
D, G, NG, B
W, D, G, NG, B
D, G, W, B, NG
D, G, W, B, NG
D, G, W, B, NG
D, G, W
H-benz(de)anthracene-7-one (benzanthrone)
no effect
no effect
no effect
no effect
anthracene
pyrene
chrysene
perylene
no effect
a
D: diesel vehicles (36, 38), G: gasoline vehicles (36), W: wood sm oke (37), P: photochem ical form ation (17, 41, 45, 64), B: boilers (65), NG:
natural gas hom e appliances (66). b 3-NF only, no effect on 1-NP. Solvent dependent.
c
) 1 × 10 mol l-1. 2-NF and 2-NP were stable over the course
-4
FIGURE 1. Photodegradation of 3-NF and 1-NP in cyclohexane solution. C
of the experiment, analogous to 3-NF.
0
known for many years that a conformation with the nitro-
group out-of-plane relative to the aromatic rings allows
photochemical nitro-nitrite rearrangement of the nitro-group
followed by further reaction steps, as seen in reaction 1 (15).
with previous findings that the photorearrangement of nitro-
PAHs is independent of the presence of O (16, 42, 43).
2
Probably, singlet oxygen is relatively unreactive toward nitro-
PAHs because the aromatic moiety is deactivated by the
electron withdrawing nitro-group.
Ar-NO + hν f Ar-ONO f products
(1)
In the case of 1-NP only one product was observed. This
product was identified as 1-hydroxypyrene (OH-PY) by
comparing with the retention time and UV-vis spectrum of
an authentic standard, in agreement with previous studies
2
The observed 1-NP photolysis is therefore ascribed to
rearrangement of the nitro-group and subsequent loss of
NO and the photostability of the other nitro-PAHs stems
from their approximately coplanar structure, which prevents
such rearrangement. Our observations are in excellent
agreement with those of other studies (1, 15, 16, 18).
Since 2-NF, 2-NP, and 3-NF are not decaying, our results
suggest that photoautooxidation (reaction with singlet
(
16, 42). 1-Hydroxypyrene is formed via reaction 1 followed
by loss of NO and hydrogen abstraction by the formed
pyrenyloxy radical from the solvent.
The Effect of Phenolic Cosolutes. The photodegradation
of 1-NP and 3-NF was more or less accelerated in the presence
of various phenolic aerosol constituents. The strongest effect
was observed for 1- and 2-hydroxynaphthalene, representing
hydroxy-PAHs present in diesel exhaust (44) and formed in
the atmosphere (45, 46) (see Figure 2A). o-Methoxyphenols,
oxygen), which has been observed for PAHs (41), is insig-
1
nificant for nitro-PAHs, although O
2
sensitization by triplet
state nitro-PAHs should be energetically feasible. This agrees
VOL. 34, NO. 5, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7 9 1

TABLE 2. Percentage Remaining upon Irradiation for 60 min
with and without Phenolic Cosolutes
cosolute
1NP
3NF
2NF
2NP
hydroquinone
phenol
none
24%
66%
72%
44%
63%
95%
75%
98%
100%
76%
100%
100%
nitro-PAH molecule leading to unstable radical intermediates
instead of decay to the ground state or (2) photodissociation
of the phenolic O-H bond giving rise to phenoxy radicals
and hydrogen atoms. These will subsequently add O
2
to
produce HO , which in turn may react with stable molecules
2
such as ground state nitro-PAHs. Triplet state 1-NP has
previously been observed to undergo photoreduction by
hydrogen abstraction (but not electron transfer) in a nearly
diffusion controlled reaction (42). Similarly, 1- and 2-ni-
tronaphthalene have both been observed to undergo pho-
toreduction via hydrogen abstraction (47). Furthermore,
Odum et al. (48) proposed that PAHs upon excitation are
able to abstract phenolic hydrogen from substituted meth-
oxyphenols. On the other hand, 2-hydroxynaphthalene has
been used as a photoinitiator of PAH degradation (49) due
to O-H photodissociation and subsequent radical produc-
tion.
As a control, experiments with a monochromatic light
source was performed in which 1-NP and 3-NF was irradiated
in the presence of 2-hydroxynaphthalene with light of 366
nm. At this wavelength, 2-hydroxynaphthalene does not
absorb light, whereas both 1-NP and 3-NF absorb consider-
ably. The decays were observed to be accelerated under these
conditions. For example, the 1-NP decay rate constant was
∼1.8 times higher in the presence of 2-hydroxynaphthalene.
Thus, the effect is weaker than in the experiments with the
full irradiation spectrum (Figure 2A), where the 1-NP decay
rate constant is ∼3 times higher in the presence of 2-hy-
droxynaphthalene, but it indicates that H-abstraction do
occur. However, it is likely that O-H photodissociation
initiates radical reactions (49) so that with the full light
spectrum two mechanisms occur simultaneously. The pho-
todissociation mechanism may also explain why the effects
of 1- and 2-hydroxynaphthalene are comparable (Figure 2A)
despite the O-H bond energy is higher in 2-hydroxynaph-
thalene (50).
FIGURE 2. A. Photodegradation of 3-NF and 1-NP in the presence
of 1- and 2-hydroxynaphthalene (1- and 2-HN). B. Photodegradation
of 3-NF and 1-NP in the presence of methoxyphenols represented
On the other hand, phenol itself accelerated the degrada-
tion of both 3-NF and 1-NP as shown in Figure 3 and Table
2
, a fact which cannot be ascribed to photodissociation since
-
4
by guaiacol (GU) and homovanillic acid (HVA). C
0
) 1 × 10 mol
phenol does not absorb light above ∼290 nm. In addition,
the effect of o-methoxyphenols cannot be explained by
photodissociation because this would lead to degradation of
both 1-NP and 3-NF. As mentioned previously this was not
the case in the experiments. It can be speculated that the
difference is due to steric effects; the methoxy-group is in
the ortho-position relative to the hydroxy-group, and as
mentioned previously, in 1-NP the nitro-group is twisted
out of the plane. Consequently it may be difficult for the
nitro-group in 1-NP to approach the phenolic hydrogen. A
compound like vanillin may possibly also initiate radical chain
reactions via hydrogen abstraction by the aldehyde group
-
1
l
.
which are important constituents of wood smoke aerosols,
were also tested with respect to their influence on nitro-PAH
decay. The methoxyphenols tested were vanillin, vanillic acid,
homovanillic acid, syringol, and guaiacol (see Table 1). These
cosolutes accelerated the decay of 3-NF as expected but
surprisingly had no effect on 1-NP. Examples are presented
in Figure 2B. Certain para-substituted phenols (e.g., hyd-
roquinone and 4-chlorophenol) were also able to accelerate
photodegradation, while para-substituted phenols having
an electron attracting substituent (e.g., 4-nitrophenol and
(
see later). However, this can be ruled out since it would lead
4
1
-hydroxybenzaldehyde) had no or little effect (see also Table
). Results for phenol and hydroquinone are presented in
to degradation of both 1-NP and 3-NF.
Experiments with 2-NF and 2-NP indicate that the triplet
states of these compounds are less reactive toward H-donors.
There is no effect of phenol and the effect of hydroquinone
is less pronounced than in the case of 1-NP and 3-NF (see
Table 2). Not all para-substituted phenols have an effect on
the photodegradation although they absorb light at higher
wavelengths. For example, 4-nitrophenol and hydroxyben-
zaldehyde had very little effect indicating that these com-
Table 2 as the percentage of the initial concentration
remaining after 60 min irradiation. In experiments with 2-NF
and 2-NP, photodegradation was affected by hydroquinone
but not by phenol itself.
The effect of certain phenolic compounds on the pho-
todegradation of nitro-PAHs can be explained either by (1)
abstraction of the phenolic hydrogen by an excited triplet
7
9 2
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 5, 2000

) 1 × 10 mol l-1. Without cosolutes, 3-NF, 2-NF, and
-4
FIGURE 3. Photodegradation of 3-NF, 2-NF and 2-NP in the presence of phenol. C
-NP were stable over the course of the experiment.
0
2
FIGURE 4. Hammet plot for 3-NF. Only phenols with no or very little absorption above 300 nm (black circles) have been included in the
regression (full line). Results for phenols with significant absorption above 300 nm (open diamonds) are also shown.
pounds are not efficient H-donors and do not photodissociate
to a significant degree. To further investigate the mechanism
for the effect of phenol on the nitro-PAH photolysis it was
attempted to obtain Hammet plots for the photodegradation
of 3-NF and 1-NP. This was done by plotting the logarithm
affect the unimolecular photolysis of 1-NP in addition to
being H-donors so that the overall mechanism becomes more
complicated than for 3-NF. A reaction sequence involving
hydrogen abstraction by excited nitro-PAHs is proposed
below:
(
base e) to the photolysis rate constant for each nitro-PAH/
1
3
substituted phenol sample relative to that of unsubstituted
phenol (ln(ksubst/ kphenol)) versus the Hammet σ parameter. σ
is a relative measure of a substituents electron donor-
acceptor properties. A good correlation for 3-NF may be
obtained, but only if those substituted phenols that absorb
light above ∼300 nm are excluded from the plot, as seen in
Figure 4. This is a strong indication that H-abstraction by
triplet state 3-NF from phenolic compounds do occur since
σ correlates very well with the O-H bond strength (51). For
example, De Lucas and De Netto-Ferrera (52) obtained good
Hammet correlations for H-abstraction by triplet state oxy-
PAH from a series of para-substituted phenols. However, it
is also evident from Figure 4 that the situation is more
complex if the phenols absorb light in the same wavelength
region as the nitro-PAHs. In that case, both light shielding
and O-H photodissociation may obscure the picture. We
could not obtain a reasonable Hammet correlation for 1-NP
even when the light-absorbing phenols were excluded. The
reason for this is not clear. It is possible that the phenols will
Ar-NO + hν f Ar-NO * f Ar-NO *
(2)
2
2
2
3
•
•
Ar-NO * + Ph-OH f Ar-NO H + Ph-O f products
2
2
(
3)
Interestingly, only phenolic hydrogen donors are able to
accelerate the photodegradation. Other potential hydrogen
donors such as toluene or 2-propanol with relatively weak
C-H bonds do not have any effect. This suggests that
hydrogen abstraction (reaction 3) may proceed via a partially
hydrogen-bonded exciplex.
In the case of 1-NP, a 1-hydroxypyrene peak was observed
both in the control samples and in samples containing
H-donors. In all experiments with efficient H-donors, several
early eluting peaks were observed to build up with time in
the HPLC chromatograms, both in experiments with 1-NP
and 3-NF. These were not seen in the control samples without
cosolutes. The early eluting peaks presumably results from
formation of polar reaction products, perhaps via ring-
VOL. 34, NO. 5, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7 9 3

not related to the nitro-PAH photodegradation. 1O
2
is
undoubtedly formed in the presence of the sensitizers and
the lack of effect on the nitro-PAHs can be explained by the
1
following: (a) the lifetime of O
(
2
in cyclohexane (τ ) 23 µs
53)) is too short for it to be formed by energy transfer, diffuse,
encounter, and react with another molecule (nitro-PAH), (b)
1
2
O does not react with nitro-PAHs to a significant degree as
1
previously discussed, or (c) the reaction of
sensitizer itself is very efficient, so that O
O
2
with the
1
2
does not escape
the solvent cage after its formation. It has been observed
that the decay of benz[a]anthracene was inhibited by
benzanthrone in toluene (τ)29 µs (53)) due to energy transfer
1
but was accelerated in benzene-d
is much longer (19). The lifetime of O
is expected to be comparable to the lifetime in toluene or
6
where the lifetime of O
2
1
2
in organic aerosols
1
cyclohexane. It thus seems that reactions of sensitized O
with nitro-PAHs are unlikely to play an important role.
2
Some PAHs and oxy-PAHs may also sensitize triplet nitro-
PAHs (42) if they possess sufficiently high triplet energies
(
e.g., chrysene and anthraquinone), but this process alone
does not lead to accelerated decay, since both PAHs and
oxy-PAHs should have an effect.
Radical Form ation by Carbonyl Com pounds. As noted
above oxy-PAHs with an (nπ*) triplet state configuration
strongly accelerates the nitro-PAH decay (Figures 5-7). It is
well-known that oxy-PAHs with this triplet state configuration
efficiently abstracts hydrogen from donors, e.g., cyclohexane,
whereas oxy-PAHs with a (ππ*) triplet state configuration
usually does not abstract hydrogen (54). The only reasonable
explanation for the effect of oxy-PAHs is that hydrogen
abstraction by an oxy-PAH triggers radical chain reactions
(see reactions 4 and 5 below, where ArCO denotes an oxy-
PAH), which ultimately lead to nitro-PAH degradation.
FIGURE 5. Photodegradation of 1-nitropyrene in the presence of
anthraquinone (AQ), chrysene (CHR), fluorenone, anthrone, and 2,4,6-
-
4
-1
trimethylbenzaldehyde (TMB). C
0
) 1 × 10 mol l
.
TABLE 3. Photochemical Classification of Oxy-PAHs
excited-state charactera O ( O2)b
1
compound
Reactive Oxy-PAHs
anthraquinone (cyclohexane)
anthraquinone (MeOH)
anthrone
nπ*
nπ*
nπ*
nπ*
nπ*
0.2
0.2
0.25
0.33
1
3
c
c
xanthone
ArCO + hν f ArCO* f ArCO*
(4)
9
,10-phenanthrenequinone
3
•
•
ArCO* + RH f ArCOH + R f radical chain reactions
Unreactive Oxy-PAHs
(
5)
benzanthrone
xanthone (MeOH)
fluorenone
ππ*
ππ*
ππ*
0.33
0.7
The hydrogen abstraction reaction initially produce ketyl
•
•
radicals (ArCOH ) and alkyl radicals (R ), which in the presence
of O will react further to give reactive products including
HO (49, 55, 56) and possibly OH radicals (49). Other products
a
Indicates the nature of the lowest excited triplet state. b 1O
2
quantum
yield, from Gorm an et al. (53). c Solvent dependent.
2
2
include alcohols and hydroperoxides (49).
opening reactions, since the corresponding UV spectra do
not resemble those of pyrene or fluoranthene derivatives.
To summarize, several different phenolic compounds,
including hydroxy-PAHs and methoxyphenols, are able to
accelerate the photodegradation of nitro-PAHs. This effect
is partly ascribed to reaction 3, but for compounds which are
able to undergo phenolic O-H photodissociation upon
irradiation, for example 2-hydroxynaphthalene (49), radical
reactions initiated by this photodissociation are also im-
portant and perhaps dominating.
A further confirmation of the proposed mechanism was
achieved by irradiating 3-NF in the presence of xanthone
both in cyclohexane and in methanol. As can be seen in
Figure 6, the nitro-PAH photodegradation in cyclohexane
was very fast, whereas in methanol there was no effect of
xanthone. The reason for this solvent effect is that the triplet
state of xanthone is inverted from the reactive (nπ*) state in
nonpolar cyclohexane (ꢀ(dielectric constant) ) 2.02) to the
unreactive (ππ*) state in polar methanol (ꢀ ) 32.66) (54). The
same experiment was performed in ethyl acetate (ꢀ ) 6.05)
with the same result as in cyclohexane. As a control the
experiments were performed with anthraquinone for which
solvent induced state inversion does not occur (57) and as
expected the nitro-PAH decay was accelerated in all three
solvents.
Influence of Photosensitizers. As discussed above, nitro-
PAHs do not appear to undergo photooxidation. However,
a number of potential sensitizers of reactive singlet oxygen
1
(
2
O ) are present on combustion aerosols, most notably PAHs
and oxy-PAHs. It was therefore important to investigate if
1
O
2
formation could influence the photodegradation of nitro-
Another experiment to verify the mechanism was per-
PAHs. Experiments were performed in which 1-NP and 3-NF
were irradiated in the presence and absence of PAHs and
4
formed in which 1-NP and 3-NF was irradiated in CCl in the
presence of anthraquinone. The results of this experiment
are presented in Table 4. No effect of anthraquinone was
oxy-PAHs. All of these compounds except perylene are
1
efficient O
2
sensitizers (53). Examples of the decays are shown
observed in CCl
anthraquinone itself decayed slower in CCl
4
for either 1-NP or 3-NF. Additionally,
(91% remaining
in Figure 5, and the experiments are summarized in Table
. As can be seen PAHs had no observable effect on nitro-
4
1
after 90 min.) compared to experiments in cyclohexane where
normally less than ∼60% remained. Clearly, the ability of the
oxy-PAH to abstract hydrogen from the surroundings de-
termines its effect on nitro-PAH photodegradation but also
seems to govern its own photostability. 2,4,6-Trimethylben-
PAH decays. Furthermore, only those oxy-PAHs with the
lowest triplet state possessing (nπ*) character had an effect
3
(
see Table 3). The effect of ( nπ*) oxy-PAHs is discussed in
1
2
a subsequent section. The ability to sensitize O is clearly
7
9 4
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 5, 2000

FIGURE 6. Photodegradation of 3-NF in the presence of xanthone
in methanol (MeOH) and cyclohexane and in the presence of 2,4,6-
-
4
-1
trimethylbenzaldehyde (TMB). C
0
) 1 × 10 mol l
.
TABLE 4. Percentage Remaining after 90 mina
irr conditions
1NP
3NF
AQ
CCl4
CCl4 + AQ
26%
38%
96%
97%
91%
a
4
Irradiation in CCl with and without AQ.
zaldehyde was similarly able to accelerate the degradation
of 1-NP and 3-NF (see Figures 5 and 6). This is also ascribed
to radical formation following hydrogen abstraction by the
triplet state aldehyde. 3,4,6-Trimethoxybenzaldehyde had
only a small effect, which is not unexpected since strongly
electron donating substituents (methoxy-groups) on a car-
bonyl can lead to triplet state inversion or at least mixing of
the (nπ*) and (ππ*) triplet states resulting in a lowered
reactivity (54). Alkyl-substituted benzaldehydes are significant
constituents of exhaust particles from gasoline vehicles and
are present on diesel particles (36), whereas methoxy-
substituted benzaldehydes are constituents of wood smoke
FIGURE 7. Comparison of the effect of anthraquinone on the
-
4
-1
photodegradation of different nitro-PAHs. A. CAQ ) 6 × 10 mol l
,
-
4
-1
-3
-1
-4
C
0
) 1 × 10 mol l . B. CAQ ) 2 × 10 mol l , C ) 1 × 10
0
-1
mol l
.
hydrogen atom acceptors (nπ*) were observed to decay
rapidly compared to the unreactive oxy-PAHs (ππ*).
(
37).
The decay rates of 1-NP and 3-NF in the presence of
reactive carbonyl compounds were comparable, although
-NF did appear to decay slightly faster. In an experiment
The results obtained in this work allow a straightforward
classification of oxy-PAHs with respect to (a) the radical
formation potential and thus the effect on other aerosol
constituents and (b) the photostability of the individual oxy-
PAHs (see Table 3).
3
in which samples of 1-NP, 2-NF, and 3-NF were irradiated
in an excess of anthraquinone, 2-NF decayed a little faster
than 1-NP and 3-NF (see Figure 7A). In a similar experiment
with 2-NF and 2-NP, 2-NP was observed to decay faster than
Com parison with Sm og Cham ber Experim ents. To assess
the environmental significance of the types of mechanisms
proposed to occur in the organic phase of combustion
aerosols, it is natural to compare our results with the detailed
smog chamber study of nitro-PAH photostabilities by Fan et
al. (1). In that study it was found that, associated with diesel
particles, the photolysis rate constants of 1-NP, 2-NP, 2-NF,
and 3-NF were not very different (knitro-PAH ) (0.022-0.035)
× kNO2). The same was observed when the compounds were
associated with wood smoke particles, but the decay was
significantly faster (knitro-PAH ) (0.045-0.050) × kNO2).
When the relative nitro-PAH photolysis rates in pure
solution (see Figure 1) are compared with smog chamber
experiments with the same compounds associated with wood
smoke or diesel particles (1), it is clear that simple unimo-
lecular photolysis is an unimportant pathway for nitro-PAHs
2
-NF (see Figure 7B). These experiments suggest that the
reactivity of radical species toward nitro-PAHs decrease in
the order 2-NP > 2-NF > 3-NF ∼ 1-NP, although the
differences are minor.
The effect of oxy-PAHs and alkyl-benzaldehydes is not
specific for nitro-PAHs but is a general degradation pathway
for all organic compounds associated with aerosols that
contain oxy-PAHs. Anthraquinone, phenanthrenequinone,
xanthone, and 2,4-dimethylbenzaldehyde have previously
been observed to accelerate the photodegradation of benz-
[a]anthracene (19).
The photostability of the individual oxy-PAHs is also
strongly dependent on the ability to abstract hydrogen. The
VOL. 34, NO. 5, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7 9 5

on these types of particles, although for 1-NP it may
contribute to the overall decay.
pendent on the chemical composition of the aerosols.
Members of several different compound classes are able to
accelerate the nitro-PAH decay rates significantly.
A further level of complexity is introduced when it is taken
into account that the chemical composition will change with
time due to chemical reactions and gas-particle conversions.
We propose that an important factor governing the overall
aerosol chemistry is the presence of reactive carbonyl
compounds such as certain oxy-PAHs, which can trigger
radical chain reactions and ultimately change the chemical
composition. We have demonstrated this to be the case for
nitro-PAH degradation, and consequently, the photodeg-
radation of nitro-PAHs may be faster on aged organic aerosols
than on freshly emitted aerosols.
Since we found that the nitro-PAHs were degraded more
rapidly in the presence of phenolic aerosol constituents,
reaction 3 or radical production via phenolic photodisso-
ciation could be responsible for the reactivities. However,
hydrogen abstraction (reaction 3) would probably be more
important for 3-NF than for 1-NP on wood smoke particles
since some of the potential H-donors are o-methoxyphenols,
which as mentioned previously have no effect on 1-NP
photodegradation. Wood smoke particles, however, also
contain other phenols such as hydroquinone (37) which
accelerates both 1-NP and 3-NF photodegradation strongly.
Nevertheless, if reactions 2 and 3 were the dominant process
on wood smoke particles, 3-NF would be expected to decay
somewhat faster than 1-NP, which was not the case in the
smog chamber experiments (1). Of course, it cannot be ruled
out that this reduced reactivity of triplet 1-NP is compensated
for by monomolecular photolysis.
Radical reactions on organic aerosols may also change
the hygroscopic properties of the aerosols through the
production of polar compounds.
Adsorbed on coal fly ash 1-NP has been reported to be
stable toward photolysis (61). The same has been observed
for unsubstituted PAHs (62), especially for the carbonaceous
fraction of the coal ash. It was speculated by these authors
that this was due to light shielding, but energy transfer from
the excited molecule to the graphitic surface to which it is
attached also provides a plausible explanation. Coal fly ash
is a complex inhomogeneous mixture characterized by a
relatively high mineral content (61, 62) and a low content of
organic carbon (e.g., less than 1% (63)), whereas diesel soot
and wood smoke consist almost exclusively of organic and
elemental carbon (21).
In contrast to 1-NP and 3-NF the photodegradation of
2
-NF and 2-NP was not accelerated in the presence of phenol
and was accelerated to a lesser extent in the presence of
hydroquinone (see Figure 3 and Table 2). Based on the above
discussion it seems clear that the effect of phenolic aerosol
constituents alone cannot explain the smog chamber ob-
servations, although it is possible that radical production via
phenolic O-H photodissociation may contribute to the decay
on ambient aerosols, especially in the presence of hydroxy-
PAHs.
As mentioned previously 2-NF and 2-NP (as well as 1-NP
and 3-NF) decayed rapidly in the presence of the radical
precursor, anthraquinone (see Figure 7A,B). This and the
reduced effect of phenolic compounds indicate that radical
chain reactions are the main degradation pathways for 2-NF
and 2-NP. This is strongly supported by the fact that 2-NP
decayed slightly faster than 2-NF both in our experiments
However, elemental carbon in the atmosphere is probably
under most conditions coated with organic carbon and water
(21, 29, 30) and condensation of organic vapors is therefore
not expected to occur directly to graphite surfaces in the
atmosphere. Nevertheless, because of diffusion through the
organic coating a certain fraction of a given SVOC will be
adsorbed onto elemental carbon with time (23). This fraction
will probably be less susceptible to direct photolysis but may
still be subject to radical reactions such as those outlined in
this paper.
(
in the presence of anthraquinone, see Figure 7B) and on
diesel particles, where the relative rates were k2-NP ) (0.030-
0
.035) × kNO2 and k2-NF ) (0.022-0.025) × kNO2 (1).
We propose that radical chain reactions initiated by
reactions 4 and 5 is the dominant mechanism for nitro-PAH
photodegradation on both diesel particles and wood smoke
particles. In addition, hydrogen abstraction by the excited
triplet state (reactions 2 and 3) may contribute to the
degradation of 1-NP and 3-NF. It is feasible to explain the
faster decay on wood smoke observed for all nitro-PAHs by
a higher concentration of radical precursors or more efficient
radical formation in the more polar organic phase of wood
smoke. Oxy-PAHs and substituted benzaldehydes are present
on both diesel particles (36) and wood smoke particles (37,
Acknowledgments
The authors are most grateful to Prof. Richard M. Kamens,
University of North Carolina at Chapel Hill (U.S.A.), for
lending us the photochemical turntable setup applied in this
work. This work has been supported by a grant from the
Danish Research Academy and by the Danish Natural Science
Research Council.
Literature Cited
5
8). A number of other aromatic carbonyl compounds are
(1) Fan, Z.; Kamens, R.; Hu, J.; Zhang, J.; McDow, S. R. Environ. Sci.
Technol. 1996, 30, 1358-1364.
also present on wood smoke particles, e.g., substituted
acetophenones and phenylacetones (37). In addition to oxy-
PAHs emitted by a variety of combustion sources (see Table
(
(
2) Tokiwa, H.; Ohnishi, Y. Crit. Rev. Toxicol. 1986, 17, 23-60.
3) Monographs on the Evaluation of the Carcinogenic Risk of
Chemicals to Humans. Vol. 46; IARC International Agency for
Research on Cancer; Lyon, 1989.
1
), these are also expected to be produced by gas and particle
phase reactions of the parent PAHs (35, 41, 59). Therefore,
oxy-PAHs are expected to be present in the same environ-
ments as the ubiquitous PAHs.
The occurrence of the radical precursors and the radical
yields are, however, highly uncertain and subject to large
variations.
Environm ental Im plications. In ambient polluted air
semivolatile organic compounds (SVOCs) have been observed
to be absorbed (rather than adsorbed) in organic combustion
aerosols (32). These aerosols could be due to diesel vehicles,
wood burning, gasoline vehicles, and other combustion
sources (60). In addition, secondary organic aerosol repre-
sents an important sorption medium for organic compounds
in the atmosphere (32). In this work, we have demonstrated
that the fate of particulate nitro-PAHs will be highly de-
(
4) M o¨ ller, L.; Lax, I.; Eriksson, L. C. Environ. Health Perspect. 1993,
101, 309-315.
(5) Li, H.; Banner, C. D.; Mason, G. G.; Westerholm, R. N.; Rafter,
J. J. Atmos. Environ. 1996, 30, 3537-3543.
(
(
6) Rosenkranz, H. S. Mutat. Res. 1996, 367, 65-72.
7) Schuetzle, D.; Riley, T. L.; Prater, T. J.; Harvey, T. M.; Hunt, D.
F. Anal. Chem. 1982, 54, 265-271.
8) Nielsen, T.; Ramdahl, T.; Bjørseth, A. Environ. Health Perspect.
1983, 47, 103-114.
(
(9) Nielsen, T.; Seitz, B.; Ramdahl, T. Atmos. Environ. 1984, 18,
159-2165.
10) Ramdahl, T.; Zielinska, B.; Atkinson, R.; Arey, J.; Winer, A. M.;
Pitts, J. N., Jr. Nature 1986, 321, 425-427.
11) Feilberg, A.; Kamens, R.; Strommen, M. R.; Nielsen, T. Atmos.
Environ. 1999, 33, 1231-1243.
(12) Arey, J.; Zielinska, B.; Atkinson, R.; Winer, A. M. Atmos. Environ.
1987, 21, 1437-1444.
2
(
(
7
9 6
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 5, 2000

(
(
13) Fan, Z.; Chen, D.; Birla, P.; Kamens, R. M. Atmos. Environ. 1995,
(42) Scheerer, R.; Henglein, A. Ber. Bunsen-Gesell. Phys. Chem. 1977,
81, 1234-1239.
(43) Hamanoue, K.; Nakayama, T. K. K.; Yamanaka, S.; Ushida, K.
J. Chem. Soc., Faraday Trans. 1992, 88, 3145-3151.
(44) Paputa-Peck, M.; Hampton, C.; Marano, R.; Schuetzle, D.; Riley,
T. L.; Prater, T. J.; Skewes, L. M.; Ruehle, P. H.; Bosch, L. C.;
Duncan, W. P. Anal. Chem. 1983, 55, 1946-1954.
2
9, 1171-1181.
14) Ciccioli, P.; Cecinato, A.; Brancaleoni, E.; Frattoni, M.; Zacchei,
P.; Miguel, A. H.; Vasconcellos, P. C. J. Geophys. Res. 1996,
1
01(d14), 19567-19581.
(
(
(
(
15) Chapman, O. L.; Heckert, D. C.; Reasoner, J. W.; Thackaberry,
S. P. J. Am. Chem. Soc. 1966, 88, 5550-5554.
16) van den Braken-vanLeersum, A. M.; Tintel, C.; van’t Zelfde, M.;
Lugtenburg, J. Recl. Trav. Chim. Pays-Bas 1987, 106, 120-128.
17) Finlayson-Pitts, B. J.; Pitts, J. N., Jr. Chemistry of the Upper and
Lower Atmosphere, Academic Press: San Diego, 2000.
(
45) Atkinson, R.; Arey, J.; Zielinska, B.; Aschmann, S. M. Environ.
Sci. Technol. 1987, 21, 1014-1022.
(
46) Sasaki, J.; Aschmann, S. M.; Kwok, E. S. C.; Atkinson, R.; Arey,
J. Environ. Sci. Technol. 1997, 31, 3173-3179.
18) Holloway, M. P.; Biaglow, M. C.; McCoy, E. C.; Anders, M.;
Rosenkranz, H. S.; Howard, P. C. Mutat. Res. 1987, 187, 199-
(
(
47) D o¨ pp, D. Springer-Verlag: Berlin, 1975; pp 49-82.
48) Odum, J. R.; McDow, S. R.; Kamens, R. M. Environ. Sci. Technol.
2
07.
19) Jang, M.; McDow, S. R. Environ. Sci. Technol. 1995, 29, 2654-
660.
1
994, 28, 1285-1290.
49) Payne, J. R.; Phillips, C. R. Environ. Sci. Technol. 1985, 19, 569-
79.
50) Beilstein Crossfire Database http:/ / www.midas.ac.uk/ cross-
fire/ : 1998.
51) Bordwell, F. G.; Zhang, X.; Satish, A. V.; Cheng, J.-P. J. Am. Chem.
Soc. 1994, 116, 6605-6610.
52) de Lucas, N. C.; Netto-Ferreira, J. C. J. Photochem. Photobiol.
A 1998, 116, 203-208.
53) Gilbert, A.; Baggott, J. Essentials of Molecular Photochemistry,
Blackwell Science, Ltd.: Oxford, 1991.
54) Gorman, A. A.; Rodgers, M. A. J. In CRC Handbook of Organic
Photochemistry. Vol. 1-2; Scaiano, J. C. Ed.; CRC Press: Boca
Raton, 1989; pp 229-247.
(
(
(
(
(
(
(
(
2
5
20) McDow, S. R.; Sun, Q.; Vartianen, M.; Hong, Y.; Yao, Y.; Fister,
T.; Yao, R.; Kamens, R. M. Environ. Sci. Technol. 1994, 28, 2147-
2
153.
(
21) McDow, S. R.; Jang, M.; Hong, Y.; Kamens, R. M. J. Geophys. Res.
996, 101, 19593-19600.
1
(
(
22) Gray, H. A.; Cass, G. R. Atmos. Environ. 1998, 32, 3805-3825.
23) Strommen, M. R.; Kamens, R. M. Environ. Sci. Technol. 1997,
3
1, 2983-2990.
24) Haag, W. R.; Hoigne, J. Environ. Sci. Technol. 1986, 20, 341-
48.
25) Ahel, M.; Scully, F. E. Jr.; Hoigne, J.; Giger, W. Chemosphere
994, 28, 1361-1380.
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
3
1
26) Murov, S. L.; Carmichael, I.; Hug, G. L. Handbook of Photo-
chemistry 2nd ed.; Marcel Dekker: New York, 1993.
27) Zielinska, B.; Arey, J.; Atkinson, R.; Ramdahl, T.; Winer, A. M.;
Pitts, J. N., Jr. J. Am. Chem. Soc. 1986, 108, 4126-4132.
28) Kamens, R.; Guo, J.; Guo, M.; McDow, S. R. Atmos. Environ.
(55) K u´ ti, Z.; G a´ l, D. Ber. Bunsen-Ges. Phys. Chem. 1992, 96, 1843-
1847.
(56) Rontani, J. F.; Raphel, D.; Aubert, C. J. Photochem. Photobiol.
A: Chem. 1993, 72, 189-193.
(57) Hamanoue, K.; Nakayama, T.; Sasaki, H.; Ikenaga, K.; Ibuki, K.
Bull. Chem. Soc. Jpn. 1992, 65, 3141-3148.
(58) Kamens, R. M.; Karam, H.; Guo, J.; Perry, J. M.; Stockburger, L.
Environ. Sci. Technol. 1989, 23, 801-806.
(59) Bunce, N. J.; Liu, L.; Zhu, J.; Lane, D. A. Environ. Sci. Technol.
1997, 31, 2252-2259.
(60) Schauer, J. J.; Rogge, W. F.; Hildemann, L. M.; Hildemann, L.
M.; Mazurek, M. A.; Cass, G. R. Atmos. Environ. 1996, 30, 3837-
3855.
1
990, 24A, 1161-1173.
29) Venkataraman, C.; Friedlander, S. K. Environ. Sci. Technol. 1994,
8, 563-572.
30) Faude, F.; Goschnick, J. Fresnius’ J. Anal. Chem 1997, 358, 667-
2.
2
7
31) Bidleman, T. F.; Billings, W. N.; Foreman, W. T. Environ. Sci.
Technol. 1986, 20, 1038-1043.
32) Liang, C.; Pankow, J. F.; Odum, J. R.; Seinfeld, J. H. Environ. Sci.
Technol. 1997, 31, 3086-3092.
33) Goss, K.-U.; Schwarzenbach, R. P. Environ. Sci. Technol. 1998,
(
(
(
(
(
(
61) Holder, P. S.; Wehry, E. L.; Mamantov, G. Polycyclic Aromat.
Comp. 1994, 4, 135-139.
62) Behymer, T. D.; Hites, R. A. Environ. Sci. Technol. 1988, 22,
3
2, 2, 2025-2032.
34) McDow, S. R.; Vartiainen, M.; Sun, Q.; Hong, Y.; Yao, Y.; Kamens,
R. M. Atmos. Environ. 1995, 29, 791-797.
1
311-1319.
63) Mastral, A. M.; Call e´ n, M. S.; Garcia, T. Environ. Sci. Technol.
999, 33, 3177-3184.
35) Jang, M.; McDow, S. R. Environ. Sci. Technol. 1997, 31, 1046-
1
1
053.
64) Chien, C.; Charles, M. J.; Sexton, K. G.; Jeffries, H. E. Environ.
Sci. Technol. 1998, 32, 299-309.
65) Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.;
Simoneit, B. R. T. Environ. Sci. Technol. 1997, 31, 2731-2737.
66) Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.;
Simoneit, B. R. T. Environ. Sci. Technol. 1993, 27, 2736-2744.
36) Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.;
Simoneit, B. R. T. Environ. Sci. Technol. 1993, 27, 636-651.
37) Rogge, W. F.; Hildemann, L. M.; Mazurek, M. A.; Cass, G. R.;
Simoneit, B. R. T. Environ. Sci. Technol. 1998, 32, 13-22.
38) Wenclawiak, B. W.; Jensen, T. E.; Richert, J. F. O. Fresenius’ J.
Anal. Chem. 1993, 346, 808-812.
(
(
(
39) Lin, S.; Jih, Y.; Fu, P. P. J. Org. Chem. 1996, 61, 5271-5273.
40) Trotter, J. Acta Crystallogr. 1959, 12, 237-242.
41) Lane, D. A. In Organic Chemistry of the Atmosphere; Hansen,
L. D.; Eatough, D. J., Eds.; CRC Press: Boca Raton, 1991; pp
Received for review May 18, 1999. Revised manuscript re-
ceived November 8, 1999. Accepted December 6, 1999.
1
55-236.
ES990566R
VOL. 34, NO. 5, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7 9 7