Photoreaction of valerophenone in aqueous solution

Article abstract of DOI:10.1021/jp981130l

Kinetics and products of the photoreaction of the phenyl ketone valerophenone were investigated as a function of temperature, pH, and wavelength in aqueous solution. Under these conditions (<10^-4M), the photoreactions are pseudo-first-order with respect to valerophenone concentration. Type II quantum yields for photoreaction were close to unity throughout the 290-330 nm spectral region and in the temperature range from 10 to 40 °C. The quantum yields for the photoproducts were 0.65 ± 0.04 for cleavage to acetophenone and propene and an overall yield of 0.32 ± 0.03 for cyclization to two cyclobutanols at 20 °C. A small amount of 1-phenylcyclopentanol (2% yield) also was formed. These photoreactions were quenchable by additions of the triplet quenchers sorbic alcohol or sorbic acid, and Stern-Volmer plots were linear up to at least 80% quenching of the photoreactions. On the basis of quenching studies with steady-state irradiations, the triplet lifetime of valerophenone at 20 °C was estimated to be 52 ns, 7 times longer than that observed in hydrocarbon solvents. Since the triplet lifetime is controlled by intramolecular hydrogen abstraction, these results indicate that the rate constant for H abstraction is significantly lowered in aqueous media. The slower H abstraction in aqueous solution is attributed to stabilization of the excited π,π* state by water and vibronic mixing and slight inversion of the reactive n,π* triplet and the unreactive π,π* triplet states. This interpretation also is supported by changes in the UV absorption spectra of phenyl ketones in water compared to organic solvents. Red shifts, compared to the polar organic solvent acetonitrile, were observed in the π-π* transitions of valerophenone and acetophenone, reflecting stabilization of the excited π,π* state by water. Other results indicated that the quantum yields for valerophenone photoreaction are pH-independent from pH 9 to pH 2 but decrease significantly below pH 2. The decrease at low pH is attributed to quenching of triplet reactivity via protonation of the excited triplet state. The use of valerophenone as a convenient actinometer for studies in water is discussed; its half-lives during midday exposure to summer sunlight in temperate latitudes are <30 min.

Full text of DOI:10.1021/jp981130l

5716
J. Phys. Chem. A 1998, 102, 5716-5723
Photoreaction of Valerophenone in Aqueous Solution
R. G. Zepp,*,† M. M. Gumz,‡ W. L. Miller,§ and H. Gao|
Ecosystems Research DiVision, National Exposure Research Laboratory, U.S. EnVironmental Protection
Agency, Athens, Georgia 30605; TAI Corporation, 960 College Station Road, Athens, Georgia 30605, and
ORISE, P.O. Box 117, Oak Ridge, Tennessee 37831
ReceiVed: February 12, 1998; In Final Form: May 19, 1998
Kinetics and products of the photoreaction of the phenyl ketone valerophenone were investigated as a function
of temperature, pH, and wavelength in aqueous solution. Under these conditions (<10-4 M), the photoreactions
are pseudo-first-order with respect to valerophenone concentration. Type II quantum yields for photoreaction
were close to unity throughout the 290-330 nm spectral region and in the temperature range from 10 to 40
°C. The quantum yields for the photoproducts were 0.65 ( 0.04 for cleavage to acetophenone and propene
and an overall yield of 0.32 ( 0.03 for cyclization to two cyclobutanols at 20 °C. A small amount of
1-phenylcyclopentanol (∼2% yield) also was formed. These photoreactions were quenchable by additions
of the triplet quenchers sorbic alcohol or sorbic acid, and Stern-Volmer plots were linear up to at least 80%
quenching of the photoreactions. On the basis of quenching studies with steady-state irradiations, the triplet
lifetime of valerophenone at 20 °C was estimated to be 52 ns, ∼7 times longer than that observed in hydrocarbon
solvents. Since the triplet lifetime is controlled by intramolecular hydrogen abstraction, these results indicate
that the rate constant for H abstraction is significantly lowered in aqueous media. The slower H abstraction
in aqueous solution is attributed to stabilization of the excited π,π* state by water and vibronic mixing and
slight inversion of the reactive n,π* triplet and the unreactive π,π* triplet states. This interpretation also is
supported by changes in the UV absorption spectra of phenyl ketones in water compared to organic solvents.
Red shifts, compared to the polar organic solvent acetonitrile, were observed in the π-π* transitions of
valerophenone and acetophenone, reflecting stabilization of the excited π,π* state by water. Other results
indicated that the quantum yields for valerophenone photoreaction are pH-independent from pH 9 to pH 2
but decrease significantly below pH 2. The decrease at low pH is attributed to quenching of triplet reactivity
via protonation of the excited triplet state. The use of valerophenone as a convenient actinometer for studies
in water is discussed; its half-lives during midday exposure to summer sunlight in temperate latitudes are
<30 min.
Introduction
Investigations of the photoreactions of aromatic alkyl ketones
over the past several decades have provided a wealth of
information concerning structural and solvent effects on triplet
reactivities1-9 and reactivities of biradical intermediates.2,3,5,9-15
Aromatic ketones with hydrogen on γ- and δ-carbons on their
alkyl side chains photoreact via their triplet states to produce
1,4- or 1,5-biradicals, respectively, that disproportionate back
to starting ketone, cleave, or cyclize,9-15 as illustrated by the
valerophenone photoreactions shown below. Most studies of
these reactions have been conducted in organic solvents, with
in water compared to organic solvents. First, the high polarity
a few in mixtures of polar organic solvents with water such as
of water can have important effects on the relative energies of
wet acetonitrile. Investigations of hydrophobic ketones in
the lowest-lying n,π* and π,π* excited states of aromatic
ketones. Solvent- and substituent-induced changes in the
relative levels and mixing of reactive n,π* triplets and unreactive
aqueous media have been limited by their low water solubility.
Recent advances in analytical methodologies, however, have
made it possible to conveniently study organic photoreactions
in aqueous solutions.
Previous studies have indicated that the kinetics and products
of aromatic ketone photoreactions may be altered significantly
π,π* triplets of aromatic alkyl ketones are well-known to affect
their triplet reactivities.5-7 The π,π* triplets are stabilized in
high dielectric and hydrogen-bonding organic solvents.16,17
Moreover, kinetics and products of the photoreactions of
aromatic alkyl ketones with γ- and δ-carbon-hydrogen bonds
can be altered by interactions of polar, hydrogen-bonding
solvents with their 1,4- and 1,5-hydroxy biradical inter-
mediates.9-15 Polar and basic solvents such as water slow the
disproportionation of biradical intermediates to starting ketone,
† U.S. Environmental Protection Agency.
‡ TAI Corp.
§ National Research Council Associate; current address, Department of
Oceanography, Dalhousie University, Halifax, Nova Scotia.
| ORISE.
S1089-5639(98)01130-X CCC: $15.00 © 1998 American Chemical Society
Published on Web 06/20/1998

Photoreaction of Valerophenone
J. Phys. Chem. A, Vol. 102, No. 28, 1998 5717
thus enhancing biradical lifetimes and overall quantum efficien-
cies for photoreaction compared to nonpolar solvents.5,6,13,14
Moreover, it has been suggested that the longer biradical
lifetimes in hydrogen-bonding solvents reflect increased distance
between radical sites, which decreases triplet to singlet inter-
system crossing rates,12 although subsequent studies by other
groups appear to have discounted this interesting possibil-
ity.11,14,15 Because biradical reactions and excited state reactivity
are sensitive to conformational effects,8,9,12,15 hydrophobic
interactions between water and aromatic ketones that alter
conformations of the alkyl moieties might have significant
effects on the kinetics and product distributions. Finally, acid-
base reactions of triplet ketones potentially may be an important
consideration in aqueous media. Triplet aromatic ketones are
105-106 times more basic than corresponding ground state
ketones, and protonation of triplets can influence aromatic
ketone photophysics and photochemistry under strongly acidic
conditions.18-20
and pH. The results of these studies are then compared to
previous observations in organic solvents, and the use of
valerophenone for UV actinometry in aqueous systems is
discussed.
Experimental Section
Chemicals. Valerophenone (99+%) and acetophenone were
obtained from Aldrich Chemical Co. They were further purified
by flash chromatography through silica gel followed by distil-
lation under reduced pressure. Sorbic alcohol (2,4-hexadien-
1-ol) was purchased from Aldrich, purified by distillation, and
stored under nitrogen at 4 °C to slow oxidation. Reagent grade
sorbic acid from Aldrich was used as received. Potassium
ferrioxalate was either synthesized in this laboratory from
K2C2O4 and FeCl3 as described by Hatchard and Parker or
obtained from Alfa Products.34 The results were identical for
both the synthesized and commercial products. The Fe(II)
analytical reagents, 1,10-phenanthroline and ferrozine, were
obtained from Alfa Products and Aldrich Chemicals, respec-
tively. A standard solution of 1000 ppm of Fe in 2% HNO3
was obtained from Fisher Scientific. Reagent grade hydroxy-
lamine hydrochloride from Matheson Scientific was used, and
the source of ammonium acetate was Baker Chemical. Water
was treated by a Barnstead Nanopure model D-7331 purification
system. Acetonitrile (HPLC grade) from Burdick and Jackson
and n-hexane (Sigma, 99+% pure) were used as received. All
other chemicals were of reagent grade quality and used as
received. Generally, phosphate buffer (5.0 mM) was used for
the pH region 4.6-8.0. Addition of dilute HCl solution was
used to adjust the pH below this range, and addition of dilute
NaOH solution was employed above this range.
Our interest in factors that affect the photoreactions of ketones
in water is prompted by several considerations. First, aromatic
ketone derivatives are increasingly being used as photoreactive
probes that selectively generate free radical sites in aqueous
biochemical studies.21 Furthermore, carbonyl compounds likely
are involved in photoreactions that affect carbon, nutrient, sulfur,
and metal cycles in freshwater and marine environments and
in atmospheric condensed phases.22,23 Photoreactions enhance
the biological availability of persistent natural organic polymers
such as humic substances, cleaving them into lower molecular
weight organic compounds that can be readily assimilated by
microorganisms.22,24,25 These aqueous photoreactions also
produce gases such as carbon dioxide,24d carbon monoxide,24c
and carbonyl sulfide,22d with concurrent reduction in UV
absorbance that permits deeper penetration of harmful solar
UV-B radiation (290-315 nm) into aquatic environments.26
Reactive oxygen species produced on absorption of solar UV
radiation have been shown to mediate these reactions.27
Preparation of Solutions for Kinetics Studies. Valerophe-
none solutions were prepared using two different methods for
comparison. Stock solutions of valerophenone were prepared
in pure water as previously described.33 The aqueous solubility
of valerophenone was found to be 8.1 × 10-4 M at 25 °C. This
solution was diluted to concentrations <7.0 × 10-4 M for the
measurements of UV-visible absorption spectra and for the
Although structural elucidation of the photoreactive chro-
mophores in humic substances are scarce, a few studies suggest
that carbonyl compounds are involved. The lignocellulose
component of organic matter in freshwater and coastal marine
environments has been shown to be photoreactive.22d,28 Past
research has demonstrated that type II cleavage of ketone
moieties incorporated within synthetic organic polymers en-
hances their environmental degradability by sunlight.29 Pho-
tochemical degradation of marine organic matter to produce
various alkenes has been proposed to involve type II reaction
of UV-absorbing carbonyl compounds.30 A recent study further
demonstrated that aromatic ketones likely are involved in
photosensitized and free radical oxidations in condensed phases
of the atmosphere (clouds, fog, and rain droplets).23 In addition
to their role in environmental reactions, photoreactions of phenyl
alkyl ketones such as valerophenone may provide a convenient
actinometer for measuring irradiance fluxes in the UV-B region
(290-315 nm). Finally, synthetic contaminants that are intro-
duced into lakes, rivers, and oceans by human activities have
carbonyl constituents, for example, p,p′-dichlorobenzophenone,
and studies of their photoreactions can provide useful informa-
tion for mathematical models of their environmental fate and
transport.31 Photoreactions of halogenated phenyl ketones can
be used to probe the excited state behavior of haloaromatic
compounds.32
photochemical kinetics studies. Alternatively, 5.0 × 10-6
M
aqueous solutions of valerophenone that contained 0.05 vol %
acetonitrile were prepared from 0.010 M stock solutions of
valerophenone in acetonitrile. After addition of the acetonitrile
solution to the water, the resulting solution was stirred magneti-
cally for at least 2 h to ensure complete dissolution of the
valerophenone. The photoreaction of valerophenone occurred
at the same rate in pure water and in the solution containing
0.05% CH3CN. Removal of oxygen by bubbling the solution
with nitrogen or argon had no effect on the rate constant within
experimental error, although bubbling did reduce the initial
concentration of valerophenone by volatilization. Therefore,
most of the photochemical studies were conducted using non-
degassed water containing trace amounts of CH3CN. The
solutions were freshly prepared each day. Aliquots (3.00 mL)
of the 5.0 × 10-6 M solutions of valerophenone were added to
either 100 × 13 mm Pyrex test tubes (for the temperature
studies) or 1.00 cm path length quartz cells (for the quantum
yield studies) and tightly capped. Acidic solutions (0.100 N
sulfuric acid) of potassium ferrioxalate were stored in the dark.
All glassware was soaked in dilute hydrochloric acid before
use to remove any traces of iron.
Irradiation Conditions. Temperature and pH studies were
conducted by parallel irradiation of solutions in a rotating
turntable photoreactor with a high-pressure quartz mercury vapor
lamp.33 Wavelengths below 300 nm were filtered out by a
In this paper, we examine the kinetics and products of
photoreactions of the phenyl alkyl ketone valerophenone in
dilute aqueous solution as a function of temperature, wavelength,

5718 J. Phys. Chem. A, Vol. 102, No. 28, 1998
Zepp et al.
borosilicate sleeve in the immersion well surrounding the lamp.
To isolate the 313 nm line for some of the experiments, the
borosilicate filter was supplemented by a 1.0 cm thick solution
of 0.002 M potassium chromate in 3% aqueous potassium
carbonate. The photoreactor was immersed in an insulated water
bath; a GCA Precision Scientific Group RDC 40 cooling system
maintained the temperature of the photoreactor within (0.2 °C
by pumping thermostated coolant through copper tubing in the
water bath. Quenching studies (313 nm) also were conducted
in this photoreactor at 20 and 30 °C, as were studies that
compared the kinetics of air-saturated and nitrogen-saturated
solutions. The latter were conducted using completely filled,
gastight tubes with Teflon-lined septum caps. Quantum yield
studies of valerophenone at various wavelengths were conducted
using the Spectral Energy Monochromatic Illumination System,
which consists of a 1000-W xenon lamp, double monochroma-
tor, and sample compartment equipped with a magnetic stirrer.
The band-pass of the monochromator was set at (1.0 nm. To
determine light intensity, well-stirred aqueous solutions of acidic
potassium ferrioxalate (0.15 M in 0.1 N H2SO4) were used as
actinometer. At each wavelength, three replicate samples of
both valerophenone and potassium ferrioxalate solutions were
irradiated with stirring, alternating valerophenone and ferriox-
alate solutions.
columns was acetophenone (m/z 120, 105, 77, 51, 43, 39). Two
other products that were isomers of valerophenone also ap-
peared. Although the retention times were similar, it was
possible to achieve baseline separation on both capillary
columns. Their mass spectra were virtually identical (m/z 162,
134, 133, 120, 105, 91, 78, 77, 51, 39), and their retention times
were strongly affected by the nature of the GC column. On
the nonpolar column, these products appeared prior to vale-
rophenone, and on the polar column, they emerged after the
valerophenone peak. Irradiation of valerophenone in acetone
produced two isomers with the same retention times and mass
spectra as those produced in water. Earlier studies demonstrated
that trans-1-phenyl-2-methylcyclobutanol was the major cyclic
photoproduct in acetone and the other isomer was the corre-
sponding cis isomer.35 The trans isomer also was the major
cyclobutanol formed in water. In contrast to retention times
observed on packed columns such as QF-1/Carbowax, the trans
isomer emerged after the cis isomer on the capillary columns.
The concentrations and yields of the cyclobutanols were
computed from peak integrations of data obtained with flame
ionization detection, assuming that the cyclobutanols had the
same response as valerophenone. The yield of acetophenone
was directly determined from the HPLC chromatographic data.
In addition to these major products, another minor product (∼2%
yield, assuming the same response as valerophenone) was
observed on the DB-5 column. The mass spectrum (m/z 144,
129, 115, 91, 77) exactly matched the spectrum of 1-phenyl-
cyclopentanol that is included in the database of the NIST Mass
Spectrometry Data Center.
Equipment. For the kinetics studies valerophenone and
acetophenone were analyzed by direct injections of the aqueous
solutions into an HPLC system consisting of a Waters Associates
high-performance liquid chromatograph model 600 pump, a
variable wavelength UV-visible detector (set at 245 nm), a
Hewlett-Packard ChemStation for data acquisition and integra-
tion, and a Whatman Partisil ODS-3 column (70-30 or 75-
25% acetonitrile/water as mobile phase). Electronic absorption
spectra were measured using a Shimadzu 265 UV-visible
spectrophotometer. Measurements of pH were made using an
Orion Ross 8102 combination electrode calibrated using NIST
buffers. Photoproduct mass spectra were obtained using a
Hewlett-Packard Series II model 5890 gas chromatograph
equipped with an HP 5972 Series mass selective detector with
limited mass range.
Kinetics Procedures. The molar absorptivities of the iron-
(II)-ferrozine and iron(II)-phenanthroline complexes used to
analyze Fe(II) in the ferrioxalate actinometers were determined
experimentally under the same conditions used to analyze the
ferrioxalate actinometers. The molar absorptivity of the Fe-
(II)- ferrozine complex [1.08 × 10-4 M ferrozine, (1-10) ×
10-6 M Fe(II)] at pH 4.55 (0.020 M ammonium acetate buffer)
was determined experimentally to be 27 300 L mol-1 cm-1 at
562 nm, in close agreement with previous studies.36 The molar
absorptivity of the Fe(II)-phenanthroline complex was 11 200
L mol-1 cm-1 at 510 nm, as previously reported.34
Photoproduct Identification. An aqueous solution (50 mL)
of valerophenone (4.00 × 10-4 M) was irradiated using
borosilicate-filtered radiation from a mercury lamp until reaction
had reached the first half-life. For comparison, a solution of
valerophenone (4.00 × 10-4 M) in acetone was irradiated under
the same conditions. The concentrations of valerophenone and
acetophenone from the aqueous photoreaction were determined
using HPLC analysis. The solution was transferred to a
separatory funnel and extracted three times using methylene
chloride. The extract was concentrated, dried over anhydrous
sodium sulfate, and analyzed by GC/mass spectrometry using
nonpolar and polar capillary columns to separate the products.
Two capillary columns were used to separate the photoprod-
ucts: (1) a Chiraldex (â-cyclodextrin B-DM) column (30 m ×
0.25 mm) purchased from Astec Co.; (2) a DB-5 column (30
m × 0.25 mm) that was purchased from J&W Co. The polar
Chiraldex column was used isothermally (150 °C) with carrier
gas flow at 1 mL min-1, and the nonpolar DB-5 column, also
with carrier gas flow at 1 mL min-1, was programmed with an
initial isothermal period at 55 °C (1.0 min) followed by a steady
increase in temperature at a rate of 2 °C min-1. Three major
products were observed on both columns, in addition to
unreacted valerophenone. Under the conditions used for the
analysis, low molecular weight gases, for example, propene,
were not detectable. The product that eluted first on both
Two methods were used in the analysis of the potassium
ferrioxalate actinometer (0.15 M). In the first, exposed potas-
sium ferrioxalate solution (100 µL) was added to 3.00 mL of
1.0 × 10-4 M phenanthroline and allowed to react in the dark
for 1 h. The absorbance of the solution was then measured at
510 nm. In the second method, exposed potassium ferrioxalate
(60 µL) and 1 M ammonium acetate buffer (150 µL) were added
to 3.00 mL of 1.0 × 10-4 M ferrozine. The reaction was
complete within 5 min, after which time the absorbance of the
solution was measured at 562 nm. The absorbances of
unexposed blanks were determined using the same procedure.
All analyses were performed in triplicate. The number of
einsteins per liter that entered the cell per unit time (Iλ) was
calculated using the equation
Iλ ) (Abs/ꢀFe)(Vf/Vi) (t)/ΦFe(II)
(1)
where Abs is the difference between absorbances of blank and
exposed potassium ferrioxalate; ꢀFe the experimentally deter-
mined molar extinction coefficient for Fe(II) complex; Vf the
final volume, either 3.21 or 3.10 mL; Vi the volume of solution
taken, either 60 or 100 µL; t the exposure time in h; and ΦFe(II)
the quantum yield for Fe(II) photoproduction (from Hatchard
and Parker34).

Photoreaction of Valerophenone
J. Phys. Chem. A, Vol. 102, No. 28, 1998 5719
TABLE 1: Comparison of Yields of Valerophenone Type II
Photoproducts in Water (20 °C)a and in Organic Solventsb
trans:cis ratio
yield, %
for
solvent
water
acetonitrile
tert-BuOH
acetone
acetophenone cyclobutanols cyclobutanols
reference
65 ( 4
85
90
d
d
32 ( 3c
15
10
d
2.4
2
this study
37
37
35, this study
this study
37, 38
37
2
2e
d
18
13
5.4f
4.2
5
benzene
hexane
82
87
a Valerophenone concentration 4.0 × 10-4 M for cyclobutanol
measurements; acetophenone yield determined at valerophenone con-
centrations ranging from 5.0 × 10-6 to 4.0 × 10-4 M; average yield (
standard deviation for five replicates reported. b Valerophenone con-
centration 0.10 M except where noted. c 1-Phenyl-1-cyclopentanol also
formed in ∼2% yield. d Not determined. e 5% valerophenone in acetone;
determined by elution chromatography in ref 35 and gas chromatog-
raphy in this study. f 5.0 × 10-4 M valerophenone in acetone.
Figure 1. Effects of changing concentration on valerophenone pho-
tokinetics in aqueous solution. Slopes of first-order plots were linear
out to 2 half-lives (r2 > 0.95) with no change within experimental error
over a 20-fold variation in initial concentration.
quantum yield for reaction as well as the light flux. The
conditions are usually quite different in aqueous media. Due
to solubility limitations, the concentrations of such ketones are
typically sufficiently low that only a small fraction of the
incident radiation is absorbed (i.e., with light path lengths <10
cm). Under these conditions the fraction of light absorbed by
the ketone, and hence its light absorption rate, is approximately
proportional to its concentration. Thus, the overall photoreaction
rate of the ketone and the rate of production of nonphotoreactive
products can be described by pseudo-first-order kinetic expres-
sions.43,44 The first-order rate constant is proportional to the
reaction quantum yield and to the cross-product of the irradiance
and the molar absorptivity of the ketone integrated over all the
effective wavelengths of the light source. Moreover, because
the photoproducts are fully exposed to the radiation, their
buildup occurs only if they are much less photoreactive than
their precursor.
Valerophenone and acetophenone were analyzed by HPLC
as noted above, using injections of 100 µL of the reaction
solution. Concentrations were determined by comparing peak
areas in the reaction mixtures to calibration standard curves.
All chromatographic data were acquired and analyzed by
computer using the HP ChemStation software.
Lotus 123 (v. 5) was used to conduct exponential regressions
for the valerophenone or acetophenone concentrations versus
time data. Pseudo-first-order rate constants were calculated from
slopes of the regressions.
Results
Photoproducts in Water. Type II cleavage to acetophenone
and propene and cyclization to two isomeric 1-phenyl-2-
methylcyclobutan-1-ols were the dominant pathways for the
photoreaction of valerophenone in water. The cyclic product,
1-phenylcyclopentanol, also formed with a low yield (∼2%).
As discussed below, the total quantum yields for these photo-
reactions are close to unity. The photoreaction can be followed
by HPLC using reverse phase columns with a UV detector (245
nm). Acetophenone is the major detectable product. The
cyclobutanols absorb UV radiation more weakly than vale-
rophenone or acetophenone and are difficult to detect at sub-
micromolar concentrations. In these studies the cyclobutanols
and the cyclopentanol were identified by combined GC/MS and/
or by comparisons of their retention times to authentic samples
that were prepared by valerophenone photolysis in acetone.35
At 20 °C cleavage accounted for 65 ( 3% of the photoreaction
and cyclization for 32 ( 3% (mean ( standard deviation of
results for five separate experiments). Although cleavage was,
as expected, the predominant reaction, the ratio of cleavage to
cyclization was considerably lower than that observed in organic
solvents (Table 1). The difference is particularly pronounced
in the case of comparisons with polar solvents, for example,
alcohols, where cleavage usually accounts for 90% or more of
the type II reaction and cyclization is a very minor reaction
pathway. In aprotic solvents the trans:cis ratio of cyclobutanol
products is in the 4-5 range; this ratio drops to 2.6 in water,
about the same as observed in polar organic solvents (Table 1).
Quantum Yields. Studies of aromatic ketone photoreactions
involving steady state irradiations in organic solvents often have
been conducted under optically opaque conditions with low
conversions to photoproducts. Under these conditions the
photoreaction kinetics are zero-order with respect to ketone
concentration and the photoreaction rate is proportional to the
The solubility limit of valerophenone in water is 8.1 × 10-4
M, and even at its maximum solubility, the quantum yields could
not be determined under conditions in which the radiation was
completely absorbed. With continuous irradiations slopes of
log plots of valerophenone concentration versus time were linear
(r2 values typically >0.95) with the same slopes at initial
concentrations ranging from 5.0 × 10-6 to 1.0 × 10-4 M (Figure
1). An initial concentration of 1.0 × 10-5 M was used for most
of the studies reported here. The quantum yield at wavelength
λ (ΦII,λ) was computed from the first-order rate constants [kobs
-
(λ)] using the equation43,44
kobs(λ)
ΦII,λ
)
(2)
2.303(Iλ)(ꢀλ)(l)
where Iλ is defined by eq 1, ꢀλ is the molar absorptivity of
valerophenone at wavelength λ, and l is the light path length.
Ferrioxalate actinometry was used to measure Iλ. Quantum
yields were found to be close to unity in the 290-330 nm range
(Table 2). The quantum yield for acetophenone formation (0.65
( 0.03) remained approximately constant out to >2 half-lives,
and acetophenone was shown to be stable in this wavelength
region during the period used for the irradiations. A series of
experiments at various temperatures in the 10-20 °C range with
borosilicate-filtered polychromatic radiation (λ > 300 nm) or
313 nm radiation from a mercury lamp at constant light intensity
indicated that the first-order rate constant and thus the average
ΦII were temperature-independent (Table 2).
Variations in pH had no detectable effect on ΦII over a broad
pH range (from pH 9 to 2). At pH <2, however, the rate

5720 J. Phys. Chem. A, Vol. 102, No. 28, 1998
Zepp et al.
TABLE 2: Molar Absorptivities and Quantum Yields for
Total Norrish Type II Reaction (Cleavage and Cyclization)
of Valerophenone (20 °C)
A
molar absorptivity quantum yields
wavelength, nm
temp, °C
(ꢀλ) (M-1 cm-1
)
[ΦII(λ)]a
290
300
310
313
320
330
313
313
25 ( 1
755
287
117
101
71.8
40.2
101
101
101
d
0.97 ( 0.03b
0.98 ( 0.03b
0.98 ( 0.04b
0.99 ( 0.03b
0.98 ( 0.04b
0.97 ( 0.05b
0.97 ( 0.04b
0.99 ( 0.03b
0.98 ( 0.03b
0.98 ( 0.05c
0.99 ( 0.05c
0.98 ( 0.05c
0.98 ( 0.05c
25 ( 1
25 ( 1
25 ( 1
25 ( 1
25 ( 1
10.0 ( 0.2
20.0 ( 0.2
30.0 ( 0.2
10.0 ( 0.2
20.0 ( 0.2
30.0 ( 0.2
40.0 ( 0.2
313
B
polychromatic
polychromatic
polychromatic
polychromatic
d
d
d
a Mean of triplicate experiments ((standard deviation); band-pass
b
was (1.0 nm for studies at 25 °C. Computed using eq 1. c Average
value computed by multiplying quantum yield at 313 nm by ratio of
rate constant at temperature of experiment to that at 20 °C. d Not
applicable.
Figure 3. UV absorption spectra of valerophenone in aqueous solution
compared to spectra in organic solvents and an acetonitrile/water
mixture: (A) comparison of spectra in hexane, acetonitrile, 25%
acetonitrile/water, and water, illustrating red shift in π-π* transitions;
(B) long-wavelength part of absorption spectra compared, illustrating
blue shift of n-π* transition and red shift of π-π* transitions in going
from the polar organic solvent, CH3CN, to water.
Figure 2. pH effects on first-order rate constants for photoreaction of
valerophenone (1.0 × 10-5 M) and on triplet-triplet absorbances of
acetophenone and propiophenone.18 The valerophenone results are
represented by a line fit to the data using a polynomial expression in
Sigma Plot. The pH effects on the triplet absorbances of acetophenone
and propiophenone were previously measured by Shizuka and Kimura.18
All data were normalized to results at neutral pH.
constants dropped sharply with increasing acidity (Figure 2).
To compare pH effects on valerophenone triplets with those
previously determined for acetophenone and propiophenone
triplet-triplet absorbance in 4:1 H2O/CH3CN,18 the data for all
three ketones were normalized to results obtained at pH 7
(Figure 2). This comparison shows that the titration plots for
triplet absorbance reduction with acetophenone and propiophe-
none were quite similar to those observed for reduction of the
valerophenone photoreaction rate constants in water.
Figure 4. Comparison of Stern-Volmer quenching plots of vale-
rophenone photoreaction in water at 20 °C using sorbic alcohol or sorbic
acid as triplet quenchers. Quenching effect is the same for both dienes.
The quenching data for sorbic acid were corrected for light screening
effects; maximum correction was <10%.
of the triplet quenchers sorbic alcohol (2,4-hexadien-1-ol) or
sorbic acid (2,4-hexadienoic acid) to the system. As shown in
Figure 4, Stern-Volmer quenching plots of valerophenone
photoreaction in water are linear out to >80% quenching. In
Figure 4, (ΦII)o and ΦII are the quantum yields without and
with added quencher, respectively; [Q] is the quencher con-
centration; and kqτ is the quenching constant. At higher
concentrations of sorbic acid, the rate constant also was reduced
somewhat by light screening by the quencher. After correction
for the light screening effect, the quenching constant for sorbic
acid was the same as that for sorbic alcohol within experimental
error. The value of the quenching constant in water in
temperature-dependent, ranging from 152 M-1 at 20.0 °C to
138 M-1 at 30.0 °C. Therefore, when added quencher is present,
the temperature must be well-defined to ensure reproducible,
accurate results. Stern-Volmer quenching constants in water
UV Absorption Spectra. Unsubstituted phenyl ketones
absorb UV radiation in the UV-B (280-315 nm) and UV-C
(200-280 nm) spectral region (Figure 3). There is a pro-
nounced solvent effect on the spectrum in this region, with a
blue shift of the weakly absorbing n-π* transition in the UV-B
region and a red shift of the π-π* (1La) transition (237 to 245
nm λmax) in changing from organic solvents to water. The shifts
occur even in changing from polar solvents, such as acetonitrile,
to water. Interestingly, the shift in going from nonpolar hexane
to polar acetonitrile (+1.1 kcal mol-1) is considerably lower
than the shift in going to water (+4.2 kcal mol-1).
Quenching Studies. The type II quantum yield for vale-
rophenone can be reduced (and half-life increased) by addition

Photoreaction of Valerophenone
J. Phys. Chem. A, Vol. 102, No. 28, 1998 5721
benzene ring of valerophenone.6 Interestingly, the triplet
lifetime of valerophenone is prolonged to the same extent by
p-methyl substitution6 as by a solvent change from hydrocarbon
to water (Table 3); ∼7-fold increases were observed in both
cases. Spectroscopic evidence indicates that p-methyl substitu-
tion inverts the reactive n,π* and unreactive π,π* triplets of
phenyl ketones, thus reducing their triplet reactivities.6 By
analogy, the spectral shift and triplet lifetime increase observed
for valerophenone in water support the proposition that these
triplets have inverted in response to solvent perturbations.
Wagner et al.6 have argued that the lower triplet reactivity likely
results from a combination of vibronic mixing of the n,π* and
π,π* states which leads to a lowest energy, unreactive π,π*
mixed state and a slightly more energetic reactive n,π* mixed
triplet. It was further proposed that these two states may be at
an equilibrium in which most of the triplets populate the
unreactive state, thus reducing overall triplet reactivity. In
further support of this idea, Scaiano’s group7 provided temper-
ature studies of the triplet lifetimes of methoxyl-substituted
phenyl ketones that indicated that their photoreactions proceed
from reactive upper n,π* triplets that are thermally populated
from lower energy, less reactive π,π* triplets. Detailed studies
of temperature effects on the triplet lifetime of valerophenone
were not conducted, but it was shown that the quenching
constant dropped ∼10% to 138 ( 4 in going from 20 to 30 °C,
corresponding to an ≈37% increase in triplet reactivity (Table
3). The increase in reactivity in this temperature range is larger
than that observed in organic solvents for valerophenone,7,8
possibly reflecting a larger activation energy required to populate
reactive upper n,π* triplets in aqueous medium.6,7 Although
valerophenone triplets are longer lived in water than in organic
solvents, oxygen quenching of the triplets still is not experi-
mentally detectable in air-saturated water, where O2 concentra-
tions are only ∼0.2 mM.
TABLE 3: Quenching Constants and Triplet Lifetimes for
Valerophenone in Water and Organic Solvents
kq x 10-9
,
solvent
water
kqτ, M-1
M-1 s-1
τ x 109, s
reference
this study
152 ( 5a
2.9b
3.6
7.2
11
2.3
5.5
5.0
52
38
8.7
3.6
14
138 ( 4
63
wet acetonitrile
acetonitrile
tert-BuOH
methanol
39
37
37
39
37
7
38
40
90
36
16
benzene
7.2
5.6
7.2
hexane
78
11
37
a Mean of triplicate experiments (( standard deviation). b Quenching
constant for sorbic alcohol in water.33,40 Value at 30 °C computed by
assuming temperature changes in kq are inversely proportional to
changes in viscosity of water (from ref 42).
are considerably higher than such constants that have been
observed for valerophenone in organic solvents (Table 3).
Discussion
As discussed in the Introduction, solvent and pH effects on
the photoreaction kinetics and products of phenyl alkyl ketone
depend on several factors that relate to the nature of the lowest
triplet as well as interactions of the biradical intermediate with
the solvent. In the following discussion, we first consider the
relationship between spectroscopic data, quenching results, and
past studies of phenyl ketones to infer the effects of water on
the triplet reactivity of valerophenone. The effects of water on
the biradical reactivity are then discussed, taking into account
the quantum yields and photoproduct distributions. Next, the
mechanism for reduction of the quantum yield with increasing
acidity is considered. Finally, we consider the potential use of
valerophenone as an aqueous actinometer for the ultraviolet
spectral region.
Biradical Reactions in Water. As expected from earlier
studies, the quantum yield for type II cleavage and cyclization
of valerophenone jumps to nearly unity in water (Table 2),
sharply higher than in nonpolar organic solvents. The quantum
yield in water is essentially temperature- and wavelength-
independent when the n-π* absorption band is irradiated (Table
2). Previous studies have established that the photoreactions
of aromatic alkyl ketones with γ- and δ-carbon-hydrogen bonds
can be altered by interactions of their 1,4- and 1,5-hydroxy
biradical intermediates with polar, hydrogen-bonding solvents.9-15
Abstraction of hydrogen from the γ-carbon to produce the 1,4-
biradical is strongly favored over abstraction of the δ-hydrogen
to form the 1,5-biradical (Scheme 1). Polar and basic solvents
such as water slow the disproportionation of biradical intermedi-
ates to starting ketone, thus enhancing biradical lifetimes and
overall quantum efficiencies for photoreaction compared to
nonpolar solvents.5,6,13,14 On the basis of previous observations
that the cyclization efficiency decreases in changing from
hydrocarbon solvents to alcohols,37,38 we anticipated little or
no cyclization in water. Instead, cyclization accounted for
almost one-third of the reaction of the 1,4-hydroxy biradicals
in water (Table 1). In addition to the cyclobutanols, we detected
a small amount (∼2% yield) of 1-phenylcyclopentanol, a product
from cyclization of the 1,5-biradical formed from δ-hydrogen
abstraction.15,38 On the other hand, the 2.4:1 trans to cis ratio
of cyclobutanols formed in water (Table 1) parallels the usual
reduced selectivity that is observed in alcohols.37 We assume
that the higher yield of cyclic products may reflect the great
sensitivity of the cleavage reaction to conformational fac-
tors.8,9,12,15 Cleavage of a 1,4-biradical can occur efficiently
Triplet Reactivity in Water. Previously determined spec-
troscopic results have substantiated that phenyl alkyl ketones
have lowest n,π* and π,π* excited states that are close in energy
and vibronically coupled.6,16,17 Stabilization of the π,π* triplet
states occurs on addition of electron-releasing substituents such
as p-methoxyl groups to the aromatic ring or on an increase in
solvent dielectric constant and hydrogen-bonding ability.6 Such
stabilization is evidenced by red shifts of the π-π* electronic
transitions, which are generally located in the 235-300 nm
spectral region. Shifts in the λmax of the second benzene
transition (the 1La band) in the UV absorption spectrum are most
readily apparent. Water has a high dielectric constant and it is
1
an excellent H-bonding solvent, so the La band of valerophe-
none shifts significantly to longer wavelengths in aqueous media
compared to organic solvents. In comparison, the λmax for the
1La band is shifted only slightly in going from nonpolar hexane
to polar acetonitrile (Figure 3). The same is true for “wet”
acetonitrile: addition of 2% up to 10% water hardly affected
the UV absorption spectrum. Likewise, the triplet lifetime of
valerophenone in acetonitrile and “wet” acetonitrile is similar
to its lifetime in hydrocarbon media (Table 3). However, the
spectrum in a 4:1 water/acetonitrile mixture was almost the same
as in water (Figure 3). In water the S-1La transition is lowered
∼8 nm, corresponding to 4.2 kcal mol-1, compared to hydro-
carbon solvents (Figure 3A). At the same time the n-π*
transition is increased ∼2 kcal mol-1, because the lowest n,π*
state is less polar than the ground state (Figure 3B).
The solvent shift observed in water is remarkably similar to
the shift observed when a p-methyl substituent is added to the

5722 J. Phys. Chem. A, Vol. 102, No. 28, 1998
Zepp et al.
SCHEME 1
only in a conformation in which the two p orbitals are parallel
to the carbon-carbon σ bond being broken. Hydrophobic
interactions between water and the alkyl chain of the biradical
may inhibit the assumption of such a conformation.
comparison shows that the reduction of the valerophenone
photoreaction rates in water with decreasing pH was very similar
to titration plots for triplet absorbance reduction with acetophe-
none and propiophenone in 4:1 H2O/CH3CN. The UV absorp-
tion spectra of phenyl ketones are almost the same in 4:1 H2O/
CH3CN as in water (Figure 3), suggesting, as discussed above,
that the triplet reactivity is about the same in both solvents.
Assuming that the solvent difference has minimal influence on
the results, these results indicate that acid-base equilibrium is
achieved in the case of valerophenone, despite its much shorter
excited state lifetime. Using the midpoint of the valerophenone
titration curve, we estimate that the triplet pKa of valerophenone
is ∼0.9, within 0.3 unit of the pKa values reported for the
acetophenone and propiophenone triplets.18 The pKa for the
propiophenone triplet is ∼0.15 unit higher than that for
acetophenone,18 likely due to an inductive effect caused by
additional substitution on the alkyl side chain. The slightly
higher value for valerophenone also likely is attributable to such
an inductive effect.
Although it has been reported that oxygen can affect the yield
of type II photoproducts from valerophenone through scavenging
of the biradical intermediate in organic solvents and “wet”
acetonitrile,41 these effects are observed at oxygen concentrations
that are well above that found in air-saturated water. Oxygen
is at least an order of magnitude less soluble in water than in
organic solvents and, consequently, oxygen in air-saturated water
(concentration is ∼0.2 mM) has no detectable quenching effect
(<5%).
pH Effects on Valerophenone Photoreactions. Previous
studies of the acid-base reactions of excited organic compounds
have employed methods involving assumptions that the proton
transfer reactions are sufficiently rapid that acid-base equilib-
rium can be achieved during the lifetime of the excited state.
In the case of phenyl ketones, acid-base equilibria were shown
to be established within the triplet state lifetimes of acetophe-
none, propiophenone, benzyl phenyl ketone, and benzophe-
none.18,19 pKa values of protonated triplet ketones were found
to be 5-6 orders of magnitude larger than those for the
corresponding ground states. pH effects on the triplets were
quantified by measuring changes in the triplet absorbance and
lifetime, both of which are reduced with increasing acidity.19
The lifetimes of the phenyl alkyl ketones, acetophenone and
propiophenone, are 1-1.2 µs in circumneutral water, ∼20-24
times longer than that of valerophenone. The shorter lifetime
of the latter is attributable to its rapid γ-hydrogen abstraction,
a reaction that is not available in the case of acetophenone and
propiophenone. Thus, valerophenone triplets have considerably
less time to reach acid-base equilibrium. To compare pH
effects on valerophenone triplets with those on acetophenone
and propiophenone triplets, we compared the previously deter-
mined pH dependence for the triplet absorbance data for
acetophenone and propiophenone with the kinetic data for
valerophenone photoreaction. To facilitate this comparison, the
data for all three ketones were normalized to results obtained
at pH 7 and a line depicting a polynomial fit to the kinetic data
for the valerophenone results was obtained (Figure 2). The
Valerophenone as an Ultraviolet Actinometer. Valerophe-
none has long been used as a UV actinometer for laboratory
studies in organic solvents.3,6,14 The present study indicates that
it is also a convenient, readily analyzed, optically thin actinom-
eter for studies in weakly absorbing aqueous solutions. Organic
solvents cannot be used for such studies due to refractive index
effects that amplify the light flux entering the reaction cell. This
is particularly true with multidirectional light sources such as
sunlight impinging on curved surfaces such as test tubes. The
photoreaction rates in dilute aqueous solution are unaffected
by the photoproducts that form, and the lifetime of valerophe-
none is sufficiently brief that air-saturated solutions can be used
without significant quenching effects.
Using data and equations discussed earlier, the number of
einsteins per liter that enters the cell per unit time (Iλ) can be
computed from the first-order rate constant for valerophenone
photoreaction. For example, at 313 nm in a reactor with a 1.0
cm path length, a commonly used system for photochemical
studies, Iλ ) 4.30 × 10-3kobs einstein L-1 h-1, where kobs is the
pseudo-first-order rate constant expressed in h-1. The photolysis
rates and half-lives (t1/2) of dilute, aqueous valerophenone under

Photoreaction of Valerophenone
J. Phys. Chem. A, Vol. 102, No. 28, 1998 5723
(9) Wagner, P. J. Acc. Chem. Res. 1983, 16, 461.
(10) Zimmt, M. B.; Doubleday, C. E., Jr.; Turro, N. J. Chem. Phys.
Lett. 1987, 134, 549.
(11) Doubleday, C. E., Jr.; Turro, N. J.; Wang, J. F. Acc. Chem. Res.
1989, 22, 199.
(12) (a) Scaiano, J. C. Tetrahedron 1982, 38, 819-824. (b) Scaiano, J.
C. Acc. Chem. Res. 1982, 15, 252-258.
(13) Wagner, P. J.; Kelso, P. A.; Zepp, R. G. J. Am. Chem. Soc. 1972,
94, 7480.
(14) Caldwell, R. A.; Dhawan, S. N.; Majima, T. J. Am. Chem. Soc.
1984, 106, 6454.
(15) Wagner, P. J. Acc. Chem. Res. 1989, 22, 83-91.
(16) Lamola, A. A. J. Chem. Phys. 1967, 47, 4810.
(17) Lutz, H.; Bre´ke´ret, E.; Lindquist, L. J. Phys. Chem. 1973, 77, 1758.
(18) Shizuka, H.; Kimura, E. Can. J. Chem. 1984, 62, 2041-2046.
(19) Rayner, D. M.; Wyatt, P. A. H. J. Chem. Soc., Faraday Trans. 2
1975, 70, 945-954.
Figure 5. Half-lives computed for valerophenone photoreaction in
sunlight on clear days at latitude 40° N. The half-lives were computed
using a mathematical model24c that uses quantum yields, absorption
coefficients, and simulated solar spectral irradiance to compute pho-
tolysis rate constants.
(20) Anastasio, C.; Faust, B. C.; Rao, C. J. EnViron. Sci. Technol. 1997,
31, 218.
(21) Dorman, G.; Prestwich, G. D. Biochemistry 1994, 33, 5661.
(22) (a) Zafiriou, O. C.; Joussot-Dubien, J.; Zepp, R. G.; Zika, R. G.
EnViron. Sci. Technol. 1984, 18, 358A. (b) Mopper, K.; Stahovec, W. L.
Mar. Chem. 1986, 19, 305. (c) Kieber, D. J.; Mopper, K. Mar. Chem. 1987,
21, 135. (d) Zepp, R. G.; Callaghan, T. V.; Erickson, D. J. Ambio 1995,
24, 181.
polychromatic light sources can be computed using the equa-
tions44,45
(23) Anastasio, C.; Faust, B. C.; Rao, C. J. EnViron. Sci. Technol. 1997,
rate ) [V] 2.303j-1E ꢀ ΦII,λ dλ ) kobs[V]
(3)
(4)
31, 218.
∫
o,λ
λ
(24) (a) Stewart, A. J.; Wetzel, R. G. Arch. Hydrobiol. 1981, 92, 265.
(b) Amador, J. A.; Alexander, M.; Zika, R. G. Appl. EnViron. Microbiol.
1989, 55, 2843. (c) Valentine, R. L.; Zepp, R. G. EnViron. Sci. Technol.
1993, 27, 409. (d) Miller, W. L.; Zepp, R. G. Geophys. Res. Lett. 1995, 22,
417. (e) Wetzel, R. G.; Hatcher, P. G.; Bianchi, T. S. Limnol. Oceanogr.
1995, 40, 1369.
t1/2 ) 0.693/kobs
where kobs is the pseudo-first-order rate constant for the
photoreaction; [V] the valerophenone concentration; Eo,λ the
scalar irradiance at wavelength λ; and j a unit conversion factor
that, for example equals 6.02 × 1020 when the irradiance is
expressed in units of photons cm-2 s-1. The rates in sunlight
also can be readily estimated using eq 3 along with mathematical
models that simulate the solar spectral irradiance reaching the
Earth’s surface, the quantum yield data, and molar absorptivities
(Table 2).24c The computed seasonal changes for this actinom-
eter are illustrated in Figure 5. Half-lives of the valerophenone
actinometer with no added quencher are <30 min in full, midday
sunlight (July, latitude 40° N) or ∼1 order of magnitude shorter
than that of the p-nitroanisole actinometer.45 Addition of sorbic
acid or sorbic alcohol can be used to adjust the half-life of the
valerophenone actinometer. Our computations using eq 3 and
simulated solar spectral irradiance indicate that the broad-band
irradiance measured by the valerophenone actinometer in
sunlight during midday, midsummer at latitude 40° N corre-
sponds to a peak response at 315 nm with a band-pass of (10
nm.
(25) (a) Kieber, D. J.; McDaniel, J.; Mopper, K. Nature 1989, 341, 637.
(b) Bushaw, K. L.; Zepp, R. G.; Tarr, M. A.; Schulz-Jander, D.;
Bourbonniere, R. A.; Hodson, R. E.; Miller, W. L.; Bronk, D. A.; Moran,
M. A. Nature 1996, 381, 404. (c) Miller, W. L.; Moran, M. A. Limnol.
Oceanogr. 1997, 42, 1317. (d) Moran, M. A.; Zepp, R. G. Limnol. Oceanogr.
1997, 42, 1307.
(26) (a) Degrandpre, M. D.; Vodacek, A.; Nelson, R. F.; Bruce, B. J.;
Blough, N. V. J. Geophys. Res. 1996, 101 (C10), 22727. (b) Morris, D. P.;
Hargreaves, B. R. Limnol. Oceanogr. 1997, 42, 239.
(27) Blough, N. V.; Zepp, R. G. In ReactiVe Oxygen Species in
Chemistry, Foote, C. S., Valentine, J. S., Eds.; Chapman and Hall: New
York, 1995; p 280.
(28) (a) Berinstain, A. B.; Whittlesey, M. K.; Scaiano, J. C. Photo-
chemistry of Lignocellulosic Materials; ACS Symposium Series 531;
American Chemical Society: Washington, DC, 1993; pp 111-121.
(29) (a) Guillet, J. E.; Regulski, T. W.; McAneney, T. B. EnViron. Sci.
Technol. 1974, 8, 923. (b) Guillet, J. E. Polymer Photophysics and
Photochemistry; Cambridge: New York, 1985; p 261.
(30) Reimer, D.; Zika, R. G. J. Geophys. Res. (D), submitted for
publication.
(31) Zepp, R. G.; Ritmiller, L. In Aquatic Chemistry; Huang, C. P.,
O’Melia, C. R., Morgan, J. J., Eds.; Advances in Chemistry Series, American
Chemical Society: Washington, DC, 1995.
(32) Wagner, P. J.; Sedon, J. H.; Gudmundsdottir, A. J. Am. Chem. Soc.
1996, 118, 746.
(33) Zepp, R. G.; Schlotzhauer, P. F.; Sink, R. M. EnViron. Sci. Technol.
Acknowledgment. We thank G. Chapman for his assistance
with identification of the photoproducts using gas chromatog-
raphy/mass spectrometry. M.M.G., W.L.M., and H.G. thank TAI
Inc., the National Research Council, and the Oak Ridge Institute
for Science and Education (ORISE), respectively, for their
support.
1985, 19, 74.
(34) (a) Hatchard, C. G.; Parker, C. A. Proc. R. Soc. London, Ser. A
1956, 235, 518. (b) Hatchard, C. G.; Parker, C. A. Trans. Faraday Soc.
1961, 57, 1093.
(35) Lacount, R. B.; Griffith, C. E. Tetrahedron Lett. 1965, 1549.
(36) (a) Stookey, L. L. Anal. Chem. 1970, 42, 779-781. (b) Zepp, R.
G.; Faust, B. C.; Hoigne´, J. EnViron. Sci. Technol. 1992, 26, 313.
(37) Wagner, P. J. J. Am. Chem. Soc. 1967, 89, 5898.
(38) Wagner, P. J.; Kelso, P. A.; Kemppainen, A. E.; McGrath, J. M.;
Schott, J. N.; Zepp, R. G. J. Am. Chem. Soc. 1972, 94, 7506.
(39) Small, R. D., Jr.; Scaiano, J. C. J. Phys. Chem. 1977, 81, 2126.
(40) Yip, R. W.; Riddell, W. D.; Szabo, A. G. Can. J. Chem. 1970, 48,
987.
References and Notes
(1) Norrish, R. G. W. Trans. Faraday Soc. 1937, 33, 1521.
(2) Yang, N. C.; Yang, D.-H. J. Am. Chem. Soc. 1958, 80, 2913.
(3) Wagner, P. J.; Hammond, G. S. J. Am. Chem. Soc. 1966, 88, 1245.
(4) Gagosian, R. B.; Dalton, J. C.; Turro, N. J. J. Am. Chem. Soc. 1970,
92, 4752.
(41) Small, R. D., Jr.; Scaiano, J. C. J. Am. Chem. Soc. 1978, 100, 4512-
4519.
(5) Wagner, P. J. Acc. Chem. Res. 1971, 4, 168.
(6) Wagner, P. J.; Kemppainen, A. E.; Schott, H. N. J. Am. Chem.
Soc. 1973, 95, 5604.
(7) Encina, M. V.; Lissi, E. A.; Lemp, E.; Zanocco, A.; Scaiano, J. C.
J. Am. Chem. Soc. 1983, 105, 1856.
(8) Lewis, F. D.; Johnson, R. W.; Johnson, D. E. J. Am Chem. Soc.
1974, 96, 6090.
(42) Viswanath, D. S.; Natarajan, G. Data Book on the Viscosity of
Liquids, Hemisphere Publishing: New York, 1989; p 714.
(43) Zepp, R. G. EnViron. Sci. Technol. 1978, 12, 327.
(44) Zepp, R. G.; Cline, D. M. EnViron. Sci. Technol. 1977, 11, 359.
(b) Zepp, R. G. In The Handbook of EnVironmental Chemistry; Hutzinger,
O., Ed.; Springer-Verlag: New York, 1982; Vol. 2, Part B, p 19.
(45) Dulin, D.; Mill, T. EnViron. Sci. Technol. 1982, 16, 815.