Products of benz[a]anthracene photodegradation in the presence of known organic constituents of atmospheric aerosols

Article abstract of DOI:10.1021/es960559s

Products of benz[a]anthracene photodegradation were determined after irradiation of benz[a]anthracene in the presence of three common constituents of atmopsheric aerosols previously found to accelerate benz[a]anthracene photodegradation; 9,10-anthraquinone, 9-xanthone, and vanillin. Among the major tentatively identified products were benz[a]anthracene-7, 12-dione, phthalic acid, phthalic anhydride, and 1,2-benzenedicarboxaldehyde. Evidence for production of polycyclic dicarboxylic acids and dialdehydes was also obtained. Product distribution was strongly influenced by solvent effects and dissolved aerosol constituents. The results suggest that products of PAH photodegradation in atmospheric aerosols are likely to be quite complex and strongly influenced by the organic composition of PAH-containing aerosol particles.

Full text of DOI:10.1021/es960559s

Environ. Sci. Technol. 1997, 31, 1046-1053
Products of Benz[a]anthracene
TABLE 1. Elution Volumes and Solvent Ratios in Column
Separation
Photodegradation in the Presence of
Known Organic Constituents of
Atmospheric Aerosols
solvent ratio
MeCl2
fraction
hexane
acetone
elution vol (mL)
1
2
3
4
5
6
7
8
1.0
0.75
0.5
0.25
0
0
0
0
0
0
0
0
0
0
0.25
0
30
30
30
30
30
30
30
30
0.25
0.5
0.75
1.0
0.75
0.5
0
M YO S E O N J A N G A N D
S T E P H E N R . M C D O W *
Department of Environmental Sciences and Engineering,
University of North Carolina at Chapel Hill,
Chapel Hill, North Carolina 27599-7400
1.0
Experimental Section
Little is known about atmospheric reaction mechanisms
or products of polycyclic aromatic hydrocarbons associated
with aerosol particles. Products of benz[a]anthracene
photodegradation were determined after irradiation of benz-
[a]anthracene in the presence of three common constituents
of atmospheric aerosols previously found to accelerate
benz[a]anthracene photodegradation; 9,10-anthraquinone,
9-xanthone, and vanillin. Irradiation was carried out in
toluene, benzene, and benzene-d6, and products were analyzed
by GCMS, FTIR, and NMR. Among the major tentatively
identified products were benz[a]anthracene-7,12-dione,
phthalic acid, phthalic anhydride, and 1,2-benzenedicar-
boxaldehyde. Evidence for production of polycylic dicarboxylic
acids and dialdehydes was also obtained. Product
distribution was strongly influenced by solvent effects and
dissolved aerosol constituents. The results suggest that
products of PAH photodegradation in atmospheric aerosols
are likely to be quite complex and strongly influenced by
the organic composition of PAH-containing aerosol
particles.
Irradiation Experim ents. The photodegradation experi-
ments were carried out in a photochemical turntable reactor
(Ace Glassware, Vineland, NJ) equipped with a 450-W
medium-pressure mercury arc lamp in a borosilicate im-
mersion well (3, 4). Although this system does not exactly
emulate the solar irradiation spectrum, it does filter out high-
energy UV bands not encountered in the troposphere with
a cutoff greater than 300 nm. The entire apparatus was
submerged in a 70-L water bath, and the temperature was
maintained at 16 °C by pumping the water through an ice
bath at 75 mL/ min. Samples were irradiated in 13 mm × 100
mm fused quartz test tubes (Ace Glassware). The concentra-
tion of benz[a]anthracene and co-solutes was 1 × 10-3 M.
Each solution was irradiated for 5 h.
Benz[a]anthracene (99% Aldrich) was irradiated using
vanillin (99%, Aldrich), 9,10-anthraquinone (97%, Aldrich),
and 9-xanthone (99%, Aldrich) as co-solutes as well as in the
absence of a co-solute. Each of these solutions was irradiated
in three different solvents, optima grade toluene (Fisher T291-
4), benzene (EM Science BX0220-5), and benzene-d6 (Aldrich,
99.5 atom % D).
Product Analysis. After completion of the irradiation
experiments, solutions were concentrated under a stream of
nitrogen, and each solution was fractionated on a silica gel
column (1.5 cm i.d., l ) 30 cm, BIO-SIL A, 100-200 mesh,
supplied by Bio-Rad, activated at 140 °C for 24 h). The elution
solvent was changed progressively from n-hexane through
methylene chloride to acetone. Table 1 shows the solvent
ratios and elution volumes of the sequential fractions eluted.
After fractionation, the collected fractions were evaporated
under a stream of nitrogen and refrigerated until analysis.
Introduction
Results from outdoor chamber studies indicate that under
appropriate conditions PAHs associated with combustion
aerosols from a variety of sources decay rapidly in sunlight
(1). Laboratory studies have suggested that PAH photodeg-
radation rate is likely to be strongly influenced by the organic
composition of the aerosol (2) and that the mechanism of
PAH photodegradation in the presence of methoxyphenols,
abundant wood smoke constituents, may involve a phenoxy
radical intermediate (3). In a previous paper, we described
the photodegradation of benz[a]anthracene in the presence
of 15 different organic compounds identified in atmospheric
particulate matter or combustion sources (4). Decay of benz-
[a]anthracene was accelerated by certain methoxyphenols,
polycyclic aromatic ketones and quinones, substituted furans,
and substituted benzaldehydes. The results suggested that
in some cases singlet oxygen and in other cases free radicals
might be involved as reaction intermediates, but that further
study was required to understand the reaction mechanisms.
In this paper, we extend that work to identification of reaction
products by reporting the determination of photodecay
products of benz[a]anthracene in the presence of vanillin,
9-xanthone, and 9,10-anthraquinone, three of the constituents
previously observed to accelerate the photodegradation.
Irradiation products were analyzed by gas chromatogra-
phy/ mass spectrometry using a Hewlett-Packard 5890 gas
chromatograph interfaced to a Hewlett-Packard 5971A mass
selective detector. Chromatography was carried out on a
J&W, 30 m, 0.32 mm i.d., DB-5 column with a 0.25-µm film
thickness. A Hewlett-Packard 7673 autoinjector was used
for all injections. The temperature program was 120 °C for
2 min, 120 to 250 °C at 15 °C/ min, 250 to 300 °C at 8 °C/ min
and hold for 3 min at 300 °C. Mass spectra were obtained
by scanning from 50 to 550 amu at 1.5 scans/ s.
NMR spectra were obtained with a Varian XL-400 NMR
spectrometer using CDCl3 as the solvent. Chemical shifts
were referenced directly to tetramethylsilane as an internal
standard. One-dimensional proton spectra were obtained
with a pulse width of 30 deg and a line width of 5299 Hz at
ambient temperature. 1H-13C two-dimensional NMR spectra
were obtained at 90 deg with a spectral width of 2541.9 Hz.
An acquisition time of 0.118 s and a relaxation delay of 1.0
s were used. Infrared spectra were obtained using absorbance
* Corresponding author telephone: (919) 966-1024; fax: (919) 966-
7911; e-mail: smcdow@sophia.sph.unc.edu.
9
1 0 4 6 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 4, 1997
S0013-936X(96)00559-7 CCC: $14.00
1997 Am erican Chem ical Society

The FTIR spectrum also supports the identification of
TABLE 2. Tentative Product Identifications
compound I as a 7,12-benz[a]anthracenedione. The absorp-
tion bands at 1460 and 1589 cm-1
are typical of aromatic
I
II
7,12-benz[a]anthracenedione
phthalic acid
C-C ring stretch. The absorption band at 1666 cm-1 is
consistent with the carbonyl stretching for a quinone. The
absorption bands at 1278 and 1309 cm-1 are probably due to
bending vibration of carbon-carbon bonds between the
carbonyl carbon and adjacent carbon atoms. The small band
at 1774 cm-1 is probably an anhydride impurity.
III
phthalic anhydride
IV
V
1,2-benzenedicarboxaldehyde
2,3-napththalenedicarboxylic acid/anhydride
phenanthrene/anthracene dicarboxaldehyde
dialdehyde derivative
7,12-dihydrobenz[a]anthracene
unknown
10-benzyl-10-hydroxyanthracen-9-one
bibenzyl
benzyl alcohol
VI
VII
VIII
IX
X
XI
XII
XIII
Oxygenated aromatic compounds (II-VII) were found in
several of the irradiated solutions. Typical examples of mass
spectra of compounds III, IV, V, and VII are presented in
Figure 3. It was not possible to distinguish between phthalic
acid (II) and phthalic anhydride (III) by GCMS using the
splitless injector because phthalic acid is converted to phthalic
anhydride under the conditions used. Consequently, a mass
spectrum of phthalic acid was not obtained. However, both
compounds were identified as products after irradiation by
their NMR spectra. In Table 3, the identification of phthalic
acid and phthalic anhydride as products is based on their
NMR spectra.
At least one of these two compounds was found after
irradiation in all benz[a]anthracene solutions in benzene or
benzene-d6. Concentrations were highest in benzene-d6.
However, they were found in toluene solutions only when
vanillin was present and only at low concentrations. Com-
pound III eluted between the fourth and sixth fractions, and
compound II eluted between the sixth and eighth fractions.
This is consistent with expected elution order for an anhydride
and a dicarboxylic acid.
Typical NMR and FTIR spectra of compounds II and III
are provided in Figure 4. In this case, both spectra were
obtained after irradiation of benz[a]anthracene in the pres-
ence of 9-xanthone in deuterated benzene. In the NMR
spectra, both compounds are present as relatively minor
impurities in the presence of high concentrations of 9-xan-
thone. Only the aromatic region of each spectrum is shown,
and the phthalic acid and phthalic anhydride spectra each
appear as small clusters of peaks between much larger peaks
corresponding to 9-xanthone. In each spectrum, charac-
teristics of a complex disubstituted monoaromatic compound
can be observed, with aromatic protons between 7.9 and 8.1
ppm for compound II and between 7.6 and 8.0 ppm for
compound III.
In the FTIR spectrum of compound III, the absorption
bands at 1774 and 1851 cm-1 are consistent with symmetric
and asymmetric CdO stretch for a conjugated anhydride.
Bands also appear at 1147, 1240, and 1346 cm-1, correspond-
ing to C-O stretch, and at 1600-1450 cm-1, corresponding
to aromatic CdC stretch.
1,2-diphenylethanol
mode with a Mattson Polaris TM infrared spectrometer, and
samples were prepared in KBr pellets.
Results
After irradiation, a yellowish-orange precipitate was found
whenever benzene and benzene-d6 were used as solvents.
No precipitate formed after irradiation of toluene solutions.
The solid product would not dissolve in a small amount of
methylene chloride used for injection onto the fractionation
column, but it was possible to dissolve it in acetone for that
purpose. This suggests that the precipitate may consist of
highly oxidized compounds such as aromatic ketones, acids,
and anhydrides.
Tentative identifications of irradiation products are listed
in Table 2. Product I was tentatively identified as 7,12-benz-
[a]anthracenedione by matching its mass spectrum to the
NIST library spectrum. Other tentatively identified products
included a number of oxygenated monocyclic (II-IV) and
polycyclic (V-VII) aromatic compounds, dihydrobenz[a]-
anthracene (VIII), one unidentified product (IX), and several
compounds that are most likely solvent irradiation products
(X-XIII). Their structures are shown in Figure 1, and the
distribution of products among the irradiated solutions are
presented in Table 3. For compounds V and VI, other
structures are likely, and alternatives V′ and VI′ are also shown
in Figure 1.
7,12-Benz[a]anthracenedione (I) was the only product
identified in all 12 of the solutions irradiated. It was generally
found in the fourth, fifth, or sixth fraction of Table 1. Its
mass, FTIR, and NMR spectra are given in Figure 2. The
mass spectrum matches that of the NIST library spectrum of
7,12-benz[a]anthracenedione with a 98% fit. The base peak
at 258 amu corresponds to the molecular weight of a benz-
[a]anthracenedione. A prominent molecular ion peak is also
characteristic of aromatic ketones. Other prominent peaks
at 230 and 202 amu correspond to the loss of one and two
carbonyl groups, respectively. The most reactive sites of benz-
[a]anthracene are the 7 and 12 positions, so the reaction
product is most likely 7,12-benz[a]anthracenedione.
The NMR spectrum exhibits a complex splitting pattern
characteristic of protons on substituted aromatic hydrocar-
bons, with a number of peaks between 7.7 and 8.5 ppm. The
doublet between 9.7 and 9.8 ppm probably corresponds to
the proton at the 1 position, which is shifted downfield due
to its proximity to the carbonyl oxygen at the 12 position.
Compound IV was found when benz[a]anthracene was
irradiated in the presence of 9,10-anthraquinone or 9-xan-
thone with either benzene or benzene-d6 as a solvent. It was
not detected after irradiation of benz[a]anthracene in any of
the toluene solutions nor in the presence of vanillin in any
of the solvents used. It was detected when benz[a]anthracene
was irradiated without a co-solute in benzene, but not in
benzene-d6 or toluene.
Compound IV is tentatively identified as a 1,2-benzene-
dicarboxaldehyde, based on a molecular ion peak at 134 amu,
TABLE 3. Products Identified in Each Irradiation
solvent
co-solute
benzene-d6
I, II, III, V
I, II, III, IV
I, II, III, IV, VII, IX
I, II, III, VI, VII, VIII
benzene
toluene
I, IX, XII
I, X, XI, XII
I, IX, XI, XII, XIII
I, III, VII, VIII, IX, XI, XII
no co-solute
9,10-anthraquinone
9-xanthone
I, III, IV, IX
I, II, III, IV
I, III, IV, VII, IX
I, III, VI, VII, VIII, IX
vanillin
9
VOL. 31, NO. 4, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 1 0 4 7

FIGURE 1. Structures of tentatively identified products.
a base peak at 105 amu corresponding to the loss of one
aldehyde group, and prominent peaks at 77 and 51 amu
corresponding to the phenyl ion and C4H3+, both diagnostic
of aromatic systems. A 92% fit was observed for comparison
to the NIST library. It was not observed as a product in toluene
solutions or in the presence of vanillin, probably because of
its rapid photoreaction when a hydrogen source is present.
Compound V was only observed before elution over silica
gel and apparently did not survive the fractionation procedure.
It was only observed with certainty after irradiation of benz-
[a]anthracene only in benzene-d6. However, analysis before
fractionation was not generally carried out, so it may have
also been produced under other conditions. In this case, it
appeared to be present at much lower levels than phthalic
acid. Based on its base peak of m/e ) 126, molecular ion
peak at m/e ) 198, and a prominent peak at m/e ) 154, it is
tentatively identified as 1,2-naphthalic anhydride.
or benzene-d6. A complete mass spectrum of compound VI
could not be obtained because it appeared as only a trace
constituent in an unresolved portion of the total ion chro-
matogram. Its tentative identification is based on a well-
defined peak on the ion chromatogram at m/e ) 234, which
corresponds to the molecular weight of phenanthrene- or
anthracenedicarboxyaldehyde, both of which are possible
products of benz[a]anthracene decay.
Compound VII was found after irradiation of benz[a]-
anthracene in all solutions that contained vanillin as a co-
solvent and in the presence of 9-xanthone in benzene and
benzene-d6. A molecular ion peak of 260 amu in Figure 3D
corresponds to that expected for the ring-opening dialdehyde
product of benz[a]anthracene. A base peak at 231 amu and
another prominent peak at 202 amu corresponding to loss
of one and two aldehyde groups, respectively, further support
this interpretation. It can be concluded that irradiation of
benz[a]anthracene under these conditions appears to produce
several oxygenated ring-opening products.
Compound VI was only produced when benz[a]anthracene
was irradiated in the presence of vanillin in either benzene
9
1 0 4 8 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 4, 1997

FIGURE 2. Compound I spectra: (A) mass spectrum, (B) proton NMR spectrum, and (C) FTIR spectrum.
A variety of additional products were also identified.
Compound VIII was produced after irradiation of benz[a]-
anthracene in the presence of vanillin in each of the solvents
used, but only in amounts barely above the noise level. It
was not detected in any of the solutions that did not contain
vanillin. It was generally found in the first fraction of Table
1, making it the least polar product identified. It also eluted
earlier in the gas chromatogram than benz[a]anthracene. The
mass spectrum is relatively simple, with a molecular ion peak
at 230 amu and prominent M - 1 and M - 2 peaks. On the
basis of this mass spectrum and the GC and silica column
elution characteristics, compound VIII is tentatively identified
as 7,12-dihydrobenz[a]anthracene.
Compound IX generally appeared as a product of benz-
[a]anthracene irradiation in benzene and toluene except in
the presence of 9,10-anthraquinone. It was also detected
after benz[a]anthracene irradiation in the presence of 9-xan-
thone in benzene-d6. It eluted in the sixth fraction of Table
1. From its NMR spectrum, a complex splitting pattern was
observed between 6.8 and 7.8 ppm, indicating that compound
IX is a substituted aromatic compound. The only other
spectral features were two singlets at about 5 ppm, which
may indicate an olefinic or disubstituted aliphatic substituent.
No acid or hydroxyl functionality were indicated in the FTIR
spectrum, but from the band at 1737 cm-1 a carbonyl group
is evident. Compound IX appears to be a carbonyl-substituted
9
VOL. 31, NO. 4, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 1 0 4 9

FIGURE 3. Mass spectra: (A) compound III, (B) compound IV, (C) compound V, and (D) compound VII.
FIGURE 4. Proton NMR and FTIR spectra of (A) compound II and (B) compound III.
aromatic compound with a molecular weight greater than
259, but little more can be concluded.
compound X, which was observed only after irradiation of
the toluene solution containing benz[a]anthracene and 9,-
10-anthraquinone. Its mass spectrum showed a molecular
ion peak at m/e ) 300, with a base peak at m/e ) 209
corresponding to the loss of the benzyl radical.
One- and two-dimensional NMR spectra of compound X
are shown in Figure 5. In the one dimensional NMR of Figure
5A, a complex aromatic splitting pattern between 6.8 and 8.1
The remaining products from Table 2 are most likely
formed from solvent reactions rather than from benz[a]-
anthracene decay. Although they would not be expected to
form in atmospheric aerosols, they provide insight into the
benz[a]anthracene decay mechanism in the presence of other
aerosol constituents. The most interesting of these is
9
1 0 5 0 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 4, 1997

FIGURE 5. Compound X NMR spectra: (A) one-dimensional proton NMR spectrum and (B) high field two-dimensional COSY NMR spectrum
1
based on H-13C hetero nuclei correlation.
ppm and a singlet at 3.2 ppm corresponding to the methylene
hydrogens are observed. The peak at 6.1 ppm corresponds
to carbon 2′(6′), and the peak at 4.6 ppm is apparently an
impurity. Assignments for the two-dimensional NMR are
given in Figure 5B. Based on these spectra, compound X is
tentatively identified as 10-benzyl-10-hydroxyanthracen-9-
one. This structure is also supported by the FTIR spectrum,
which exhibits bands at 3600-3000 cm-1, corresponding to
O-H stretch and at 1660 cm-1 corresponding to CdO stretch.
Compound XI generally eluted in the second fraction of
Table 1, indicating that it was less polar than most of the
other products. It was tentatively identified as bibenzyl
because of its molecular ion peak at 182 amu, base peak at
91 amu corresponding to cleavage of the aliphatic C-C bond,
and prominent peak at 77 amu corresponding to the phenyl
ion. It matched the NIST library spectrum of bibenzyl with
a fit of 83%. Compound XII eluted in the fifth or sixth fraction
of Table 1, indicating that it was more polar than compound
XI. It was tentatively identified as benzyl alcohol because of
its molecular ion peak at 108 amu, phenyl ion peak at 77 amu
corresponding to cleavage of the C-C bond next to the oxygen
atom, and a prominent peak at 51 amu. Its mass spectrum
matched that of the NIST library spectrum for benzyl alcohol
with a 92% fit. Compound XIII was tentatively identified as
1,2-diphenylethanol, in agreement with an 80% fit when
matched to the NIST library spectrum of this compound.
9
VOL. 31, NO. 4, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 1 0 5 1

FIGURE 6. Photoreduction of 9,10-anthraquinone in toluene.
benzenedicarboxyaldehyde have been observed as products
of the reaction of ozone with naphthalene or alkylnaphtha-
lenes in solution (28-30). The observation of phthalic acid
and related compounds from benz[a]anthracene photodeg-
radation in the presence of organic aerosol constituents
suggests that other mechanisms besides ozone attack can
also lead to the production of phthalic acid and related
compounds from benz[a]anthracene and possibly from other
PAHs. Polycyclic aromatic dicarboxylic acids have also been
found in atmospheric particulate matter (27).
This compound exhibited a very small molecular ion peak at
198 amu and cleavage of the C-C bond next to oxygen, leading
to benzyl and hydroxybenzyl fragments.
Discussion
In spite of significant research efforts in the study of
atmospheric decay of PAHs, little is known concerning their
atmospheric decay products with the exception of a few
mutagenic nitro-subsitituted (5-10) and oxygenated (11-
13) PAHs produced in relatively low yield. It has proved
difficult to account for the bulk of benz[a]anthracene loss
after irradiation by UV light in laboratory experiments or by
sunlight in smog chamber studies. This is probably due at
least in part to difficulties in using GCMS to analyze highly
oxygenated or polymeric products likely to be produced by
photodegradation of PAHs.
Of the products observed in this work, 7,12-benz[a]-
anthracenedione stands out as a ubiquitous product of benz-
[a]anthracene photodegradation detected under all irradiation
conditions investigated. Previous smog chamber studies
suggest that significant decay of benz[a]anthracene in sunlight
can occur under typical atmospheric conditions (13). 7,12-
Benz[a]anthracenedione has also been found in particulate
matter under photochemical smog conditions (14). Com-
pound IX, another carbonyl-containing compound, was also
detected in many of the experiments. Formation of ketones
and quinones from irradiation of anthracene associated with
atmospheric particulate matter has been observed (11).
The tentative identification of phthalic acid, phthalic
anhydride, and 1,2-benzenedicarboxaldehyde are important
indications of benz[a]anthracene ring cleavage products.
There are many known examples of aromatic ring cleavage
involving free radical or singlet oxygen intermediates. It is
well known that radical addition to atmospheric monocyclic
aromatic compounds leads to ring-opening products, and a
variety of ring-opening mechanisms based on hydroxyl radical
addition have been proposed (15-20).
PAHs are also known to undergo aromatic ring cleavage
when exposed to ozone (21, 22), irradiated on silica surfaces
(23), exposed to singlet oxygen (24), and exposed to hydrogen
peroxide and UV light in advanced oxidation processes for
water treatment (25). Thus, the observation of ring cleavage
products observed in this work is consistent with this free
radical or singlet oxygen attack suggested by the reaction
kinetics described in a previous work (4).
Phthalic acid has also been observed in atmospheric
samples (14, 26, 27), and it has been suggested that the samples
are formed from atmospheric reactions of PAHs and ozone
(26). Photochemical formation of phthalic acid from naph-
thalene and substituted naphthalenes has also been observed
on silica surfaces (23) and in water in the presence of hydrogen
peroxide (25). Phthalic acid, phthalic anhydride, and 1,2-
Hydro-substituted PAHs, including 7,12-dihydrobenz[a]-
anthracene, have previously been identified in ambient air.
Lao et al. were the first to identify them and noted that they
accounted for approximately 15% of their total measured PAH
fraction (31). Cautreels and Van Cauwenberghe (27) also
identified several hydro-PAH derivatives in ambient air,
including dihydro derivatives of benzo[c]fluorene and benzo-
[c]phenanthrene, and reported a chromatographic peak
corresponding to a mixture of 7,12-dihydrobenz[a]anthracene
and dihydrochrysene.
The assignment of compound X as 10-benzyl-10-hydroxy-
anthracen-9-one can be made with more confidence than
assignments for any of the other products. It is most likely
the asymmetric radical recombination product of the reaction
scheme in Figure 6. A symmetric recombination product,
bibenzyl, was also tentatively assigned to compound XI, as
indicated in Table 3. The other symmetric product is unlikely
to form because of steric hindrance. These products are
consistent with a free radical reaction initiated by hydrogen
abstraction from the solvent by excited state 9,10-an-
thraquinone.
Compounds XI, XII, and XIII in Table 1 are solvent reaction
products that are most likely produced by hydrogen transfer
from toluene solvent to excited state triplet carbonyl com-
pounds. As described earlier, compound XI was tentatively
identified as bibenzyl. This was found in toluene in all cases
when a co-solute was present, but was not observed as a
reaction product when benz[a]anthracene was irradiated in
toluene in the absence of a co-solute. Since bibenzyl is
indicative of a hydrogen transfer reaction, the results suggest
that 9,10-anthraquinone, 9-xanthone, and vanillin all initiate
a hydrogen transfer reaction in toluene, which results in the
formation of bibenzyl. Bibenzyl formation was also observed
after irradiation in toluene of several other compounds that
appeared to accelerate benz[a]anthracene photodegradation
(4). Compound XII, benzyl alcohol, appeared to be produced
in the greatest abundance in the presence of 9-xanthone in
toluene. Its formation also probably involves hydrogen
abstraction from toluene to produce the benzyl radical. It is
possible that oxygen then reacts with the benzyl radical to
form benzyl alcohol. Compound XIII, tentatively identified
9
1 0 5 2 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 4, 1997

as 1,2-diphenylethanol, is probably formed through a similar
process.
However, it can be concluded that the organic composition
is likely to play an important role in influencing the reaction
mechanisms and products of PAH photodegradation in
atmospheric aerosol particles.
At least two mechanisms appear to be operative in these
irradation experiments. The first mechanism apparently
results in the production of phthalic acid (II) and other ring-
opening products (III-VII). At least some of these products
were observed whenever benzene or benzene-d6 was used as
the solvent, and their concentrations were highest in benzene-
d6. None of these products were observed in toluene
solutions, except in the presence of vanillin. It is possible
that this mechanism involves singlet oxygen addition to the
aromatic ring. This is a reasonable ring-opening reaction
process for unsubstituted aromatic compounds like benz-
[a]anthracene. Also singlet oxygen has a longer lifetime in
benzene-d6, and the singlet oxygen reaction is not expected
to be competitive in toluene because toluene is a strong
hydrogen donor. Similar products have been observed for
phenanthrene, anthracene, acenaphthylene, and 1-meth-
oxynaphthalene photodegradation in experiments in which
singlet oxygen was directly observed (32-35).
In the second mechanism, several solvent products (X-
XIII) containing a benzyl moiety are formed. At least one of
these products was observed whenever toluene was used as
the solvent. The mechanism probably involves hydrogen
abstraction from the solvent by carbonyl-substituted aromatic
compounds, leading to radical formation. It has been shown
that the benz[a]anthracene decay rate is related to bibenzyl
formation rate (4), but it is not clear whether benz[a]-
anthracene then decays by subsequent radical addition or
some other process. In any case, its decay does not appear
to lead to ring-opening products.
Literature Cited
(1) Kamens, R. M.; Bell, D. A.; Dietrich, A.; Perry, J. M.; Goodman,
R. G.; Claxton, L. D.; Tejada, S. Environ. Sci. Technol. 1985, 19,
63-69.
(2) McDow, S. R.; Sun, Q.; Vartiainen, M.; Hong, Y.; Yao, Y.; Fister,
T.; Yao, R.; Kamens, R. M. Environ. Sci. Technol. 1994, 28, 2147-
2153.
(3) Odum, J. R.; McDow, S. R.; Kamens, R. M. Environ. Sci. Technol.
1994, 28, 1285-1290.
(4) Jang, M.; McDow, S. R. Environ. Sci. Technol. 1995, 29, 2654-
2660.
(5) Pitts, J. N. Atmos. Environ. 1987, 21, 2531-2547.
(6) Arey, J.; Harger, W. P.; Helmig, D.; Atkinson, R. Mutat. Res. 1992,
281, 67-76.
(7) Helmig, D.; Lopez-Cancio, J.; Arey, J.; Harger, W. P.; Atkinson,
R. Environ. Sci. Technol. 1992, 26, 2207-2213.
(8) Nishioka, M. G.; Lewtas, J. Atmos. Environ. 1992, 26A, 2077-
2087.
(9) Nielsen, T. Environ. Sci. Technol. 1984, 18, 157-163.
(10) Kamens, R. M.; Fan., Z.; Yao, Y.; Vartiainen, M. Chemosphere
1994, 28, 1623-1632.
(11) Fox, M. A.; Olive, S. Science 1979, 205, 582-583.
(12) Pitts, J. N.; Lokensgard, D. M.; Ripley, P. S.; Van Cauwenberghe,
K. A.; Van Vaeck, L.; Shaffer, S. D.; Thill, A. J.; Belser, W. L. Science
1980, 210, 1347-1349.
(13) Kamens, R. M.; Karam, H.; Guo, J.; Perry, J.; Stockburger, L.
Environ. Sci. Technol. 1989, 23, 801-806.
(14) Rogge, W. F.; Mazurek, M. A.; Hildemann, L. M.; Cass, G. R.;
Simoneit, B. R. T. Atmos. Environ. 1993, 27A, 1309-1330.
(15) Dumdei, B. E.; Kenny, D. V.; Shepson, P. B.; Kleindienst, T. E.;
Nero, C. M.; Cupitt, L. T.; Claxton, L. D. Environ. Sci. Technol.
1988, 22, 1493.
Apparently, the first mechanism is most important in
benzene and benzene-d6, and the second mechanism is most
important in toluene. The observation of a precipate after
all irradiations in the presence of each of the co-solutes in
benzene-d6 and benzene but not in toluene also suggests
that different mechanisms are responsible for benz[a]-
anthracene decay in the different solvents.
(16) Atkinson, R. J. Phys. Chem. Ref. Data 1994, Monograph No. 2.
(17) Bartolotti, J.; Edney, E. O. Chem. Phys. Lett. 1995, 245, 119-122.
(18) Andino, J. M.; Smith, J. N.; Flagan, R. C.; Goddard, W. A.; Seinfeld,
J. H. J. Phys. Chem. 1996, 100, 10,967-10,980.
(19) Yu, J.; Jeffries, H. E.; Sexton, K. G. Atmos. Environ. In press.
(20) Yu, J.; Jeffries, H. E. Atmos. Environ. In press.
(21) Moriconi, E. J.; Rakoczy, B.; O’Connor W. F. J. Am. Chem. Soc.
1961, 83, 4618-4623.
Previous results from kinetic experiments in toluene were
also consistent with a mechanism involving hydrogen ab-
straction from the solvent (4). However, on the basis of those
experiments it was less certain whether singlet oxygen was
involved in the reaction mechanism. Since singlet oxygen
seems the most likely explanation for the ring-opening
products observed, the possibility of a singlet oxygen inter-
mediate in benzene and benzene-d6 should still be considered.
We previously concluded that aerosol organic composition
may strongly influence PAH photodegradation rates in
atmospheric particulate matter and that the mechanism of
decay apparently depends on the composition of light-
absorbing compounds present (2). The product results here
support these conclusions and further suggest that atmo-
spheric PAH decay processes may also depend on such
properties as hydrogen-donating facility and singlet oxygen
solubility, which are also likely to depend on aerosol organic
composition. Both hydrogen abstraction-initiated radical
reactions and singlet oxygen attack are possible mechanisms
of benz[a]anthracene decay in atmospheric particulate matter.
The model systems used in this work were useful for
determining possible reaction mechanisms that should be
explored in organic aerosol photochemistry. In real atmo-
spheric particles, the compounds investigated in this study
are minor components of a complex, largely unknown organic
mixture. The relative importance of the proposed mecha-
nisms depends on such properties as oxygen solubility and
singlet oxygen lifetime in the aerosol particle, hydrogen-
donating properties of its organic components, and light-
absorbing species present, all of which are difficult to
determine. Because the composition of the organic fraction
in most organic aerosols is incompletely understood, it is
possible that other mechanisms may be more important.
(22) Moriconi, E. J.; Taranko, L. B. J. Org. Chem. 1963, 28, 2526-
2529.
(23) Barbas, J. T.; Sigman, M. E.; Buchanan, A. C.; Chevis, E. A.
Photochem. Photobiol. 1993, 58, 155-158.
(24) Lee-Ruff, E.; Kazarians-Moghaddam, H.; Katz, M. Can. J. Chem.
1986, 64, 1297-1303.
(25) Tuhkanen, T. A.; Beltran, F. J. Chemosphere 1995, 30, 1463-
1475.
(26) Kawamura, K.; Ikushima, K. Environ. Sci. Technol. 1993, 27, 2227-
2235.
(27) Cautreels, W.; Van Cauwenberghe, K. Atmos. Environ. 1976, 10,
447-457.
(28) Sturrock, M. G.; Cravey, B. J.; Wing, V. A. Can. J. Chem. 1971, 49,
3047-3051.
(29) Legube, B.; Guyon, S.; Sugimitsu, H.; Dore, M. Water Res. 1986,
20, 197-208.
(30) Gaul, M. D.; Junk, G. A.; Svec, H. J. Environ. Sci. Technol. 1987,
21, 777-784.
(31) Lao, R. C.; Thomas, R. S.; Oja, H.; Dubois, L. Anal. Chem. 1973,
45, 908-915.
(32) Barbas, J. T.; Dabestani, R.; Sigman, M. E. J. Photochem. Photobiol.
A 1994, 80, 103-111.
(33) Dabestani, R.; Ellis, K. J.; Sigman, M. E. J. Photchem. Photobiol.
A 1995, 86, 231-239.
(34) Sigman, M. E.; Barbas, J. T.; Chevis, E. A.; Dabestani, R. New J.
Chem. 1996, 20, 243-248.
(35) Barbas, J. T.; Sigman, M. E.; Dabestani, R. Environ. Sci. Technol.
1996, 30, 1776-1780.
Received for review June 25, 1996. Revised manuscript re-
ceived November 19, 1996. Accepted December 4, 1996.X
ES960559S
X Abstract published in Advance ACS Abstracts, February 15, 1997.
9
VOL. 31, NO. 4, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 1 0 5 3