Role of quinone intermediates as electron shuttles in fenton and photoassisted fenton oxidations of aromatic compounds

Article abstract of DOI:10.1021/es9610646

Fenton and related reactions are potentially useful oxidation processes for destroying toxic organic compounds in water. In these reactions, H₂O₂ is combined with Fe(II) or Fe-(III) in the presence or absence of light to generate hyd

Full text of DOI:10.1021/es9610646

Environ. Sci. Technol. 1997, 31, 2399-2406
core thermal and photochemical inorganic reactions involved
in HO production and decay are well established (eqs 1-13,
Table 1) (10). Little is known, however, about the influence
of the target compounds or their degradation products on
the reactivity of iron. Organic compounds may act as ligands
Role of Quinone Intermediates as
Electron Shuttles in Fenton and
Photoassisted Fenton Oxidations of
Aromatic Compounds
•
(
L) or redox agents. Possible reactions include (i) photolysis
of Fe(III)-L complexes (5, 11); (ii) formation of reactive high-
valent iron-oxo or iron-peroxo complexes stablized by L
+
(
1
12); (iii) reduction of Fe3 by neutral organic molecules (13,
R U Z H O N G C H E N A N D
J O S E P H J . P I G N A T E L L O *
4) or radicals (10); and (iv) oxidation of Fe2 by radicals or
+
carbocations (10). An understanding of these reactions is
important for the successful application of Fenton-type
remediation technologies.
Department of Soil and Water, The Connecticut
Agricultural Experimental Station, P.O. Box 1106,
New Haven, Connecticut 06504
We have undertaken a study of the Fenton degradation of
phenol as a model for aromatic pollutants. Previous studies
•
show that HO initially adds to the aromatic ring to give a
substituted hydroxycyclohexadienyl radical (I, Scheme 1) (10).
Depending on pH and availability of suitable oxidants or
reductants, I may react further by: acid-catalyzed elimination
of water to form a radical cation (path A); formation of a
dioxygen radical adduct (path B); oxidation to a cyclohexa-
dienyl cation (path C); orswhen paths B and C are
unavailablescoupling and disproportionation (path D) (15,
Fenton and related reactions are potentially useful oxidation
processes for destroying toxic organic compounds in
water. In these reactions, H2O2 is combined with Fe(II) or Fe-
(
III) in the presence or absence of light to generate hydroxyl
•
radicals (HO ). A relatively neglected area of research is
the influence of organic species on the reactivity of iron,
and hence on the rate or course of the reaction. This study
examined the oxidation of phenol (2 mM) by Fenton systems
1
6).
The impetus for this study was 2-fold. First, it was reported
3
+
2+
in the dark (Fe /H2O2, Fe /H2O2, and mixed
by Hamilton et al. (17) and later by Litvintsev et al. (18) that
catalytic amounts of 1,4-hydroquinone (1,4-HQ) or 1,2-
hydroquinone (catechol, 1,2-HQ)sboth intermediates in
Fe /Fe /H2O2 systems) and under UV/visible light (Fe3+/
3+
2+
H2O2/hν). Reactions were conducted at an initial pH of
3
+
phenol degradationsgreatly increased the rate of the Fe
degradation of substituted benzenes. They proposed
that the active oxidizing agent is a ternary HQ-Fe-H
/
2.8, with H2O2 in excess and iron in catalytic concentrations.
2 2
H O
In all cases, the reactions display autocatalysis. In dark
2
O
2
reactions, the lag phase decreases (a) with increasing
total iron concentration, [Fe]T; (b) as initial [Fe2+] increases
complex. However, since 1,2- and 1,4-HQ are known to
reduce Fe3 rapidly at low pH (13, 14), it seemed more likely
to us that rate acceleration was due to cut off production of
+
when [Fe]T is held constant; (c) in proportion to the amount
of hydroquinone (1,4- or 1,2-) added; and (d) in proportion
to the amount of quinones (1,4-benzoquinone or 5-hydroxy-
2+
•
Fe , which then generates HO through the normal Fenton
pathway (eq 4). Nevertheless, the observations suggested at
least the possibility of a novel role for organics in the pathway,
one that might involve an alternative oxidant. Second, it has
1
,4-naphthoquinone [juglone]) added. The hydroquinones
3+
2+
reduce Fe to Fe . An important result is the finding
that quinones serve as electron-transfer catalysts between
•
been shown in studies where HO was generated by radiolysis
•
dihydroxycyclohexadienyl radicalsthe HO adduct of
(15) that quinone intermediates of phenol degradation are
able to oxidize I (Scheme 1, quinone ) OX in path C). Since
the reduction product of quinonesa semiquinone rad-
icalsmay reduce Fe3 to regenerate the quinone, this sug-
gested that quinone intermediates may act as electron-transfer
catalysts between I and iron ions.
phenolsand Fe3+ by way of a semiquinone radical. This
conclusion is supported by simulations with a kinetic
model employing known and proposed steps. In irradiated
solutions, the lag phase decreases with decreasing
wavelength (480-300 nm; 10 nm band-pass). A cause of
light initiation, especially above ∼410 nm where photolysis
of Fe3 and/or H2O2 is negligible, is direct photolysis of quinones
+
Experimental Section
+
Chem icals. Water was distilled and then further purified in
a Barnstead Nanopure system. Chemicals were used as
received except where noted. The following were from
Aldrich: 1,4-benzoquinone (1,4-BQ; 98%), 1,2-HQ (99+%),
•
to HO and semiquinone radicals, which subsequently
reduce Fe3 and re-form the quinone. Thus, quinones play
an important catalytic role in Fenton oxidation of aromatic
compounds.
+
1
6
,4-HQ (99+%), resorcinol (99+%), NaClO
4 4 2
(99%), Fe(ClO ) ‚-
H
2
O (98%), Fe(ClO ‚xH O (11.8% Fe by weight), and
4
)
3
2
FerroZine reagent [3-(2-pyridyl)-5,6-diphenyl-1,2,4-triazine-
p,p′-disulfonic acid, monosodium salt; 97%]. 1,4-BQ was
resublimed under vacuum. Juglone (5-hydroxyl-1,4-naph-
thoquinone, 98+%) was from Lancaster. Hydrogen peroxide
Introduction
•
Hydroxyl (HO ) is a reactive, nonselective radical (1, 2)
underlying the chemistry of many advanced oxidation
processes (AOPs) for degrading toxic organic compounds in
water. Recently there has been a renewed interest in the use
(
2 2 4 3
30%), K C O , FeCl , and 1,10-phenanthroline were from
Fisher Scientific. Methanol and acetonitrile (HPLC grade)
were from J. T. Baker.
Phenol was purchased from Sigma (99+%, no preserva-
2
tives) and redistilled under reduced pressure of N . Repro-
ducible results were obtained with this material. The lag
period in phenol degradation was found, however, to be
sensitive to the source of phenol due to the presence of
•
of Fenton-type AOPs, which generate HO from hydrogen
peroxide and an iron compound (3-6). A number of variants
of the Fenton reaction have been studied, including those
2
+
3+
using Fe or Fe , in the presence or absence of complexing
ligands and in the presence or absence of light (3, 5-9). The
impurities with a quinone or hydroquinone structure.
A
*
Corresponding author telephone: (203) 789-7237; fax: (203) 789-
7
232; e-mail: jpignat@caes.state.ct.us.
sample from Sigma containing 0.1% H PO as preservative to
3
2
S0013-936X(96)01064-4 CCC: $14.00

1997 Am erican Chem ical Society
VOL. 31, NO. 8, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2 3 9 9

c
k
k
k
k
k
k
k
k
k
b
c
k
29
o
p
o
p
k
k
k
p
k
14
13,
k
k
28
k
27
k
15
k
k
26
k
k
k
k
25
e
e
e
k
k
k
d
2
c
K
k
k
k
k
k
k
10
k
k
24
b
k
K
k
b
k
k
k
k
k
a
e
2
4 0 0
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 8, 1997

SCHEME 1
suppress air oxidation showed a decreasing lag occurring
over several months of use.
column and detected with a Hewlett-Packard 1050 diode array
UV/ vis detector. The mobile phase was 50% acetonitrile/
50% water at 1.0 mL/ min for the analyses of phenol
Procedures. All experiments were conducted in aerated
solution at 0.1 M ionic strength (NaClO
4
) and initial pH of
(monitored at 210 nm and/ or 270 nm; retention time, t
min) and 1,4-BQ (monitored at 250 nm; t , 4.2 min) in dark
and photoassisted Fenton experiments. In the 1,4-BQ
quantum yield experiments, the mobile phase was 30%
acetonitrile/ 70%water with 0.08% trifluoroacetic acid at 1.0
mL/ min for the analyses of 1,4-HQ (monitored at 220 nm
R
, 4.7
2
.80 adjusted with HClO
4
. All solutions were made fresh daily.
R
3
+
2+
Stock solutions of Fe or Fe at 10-20 mM concentration
were prepared in 0.05 or 0.1 M HClO . Stock solutions of
2 2
were prepared by diluting 30% H O . The final reaction
4
H
2
O
2
solution was usually made by mixing equal volumes of
solutions A and B, each adjusted to pH 2.8: solution A
contained iron at twice target concentrations; solution B
and/ or 290 nm; t
R
, 4.0 min) and 1,4-BQ (monitored at 250
nm; t , 6.4 min). HPLC analysis was usually done within 3
R
contained H
2
O
2
and phenol at twice target concentrations.
h, but never longer than 6 h, after sampling.
Other chemicals, if needed, were added as quickly as possible
to the mixture. For the juglone experiment, chemicals were
Ferrous ion was determined by a modified FerroZine
method (19). A methanolic solution of 2 mM FerroZine
3
+
added to pH-adjusted Fe solutions in the order phenol,
, and juglone.
Dark experiments were carried out in Teflon-lined septum
reagent and 18 mM NH
4
F was mixed with an equal volume
2 2
H O
of reaction solution and the absorbance at 562 nm was
3+
-
measured. The strong complexation of Fe by F was used
screw-cap 40 mL vials or 250 mL Erlenmeyer flasks in a water
bath at 25 ( 0.1 °C in a room illuminated by dim red light.
The solution was stirred with a Teflon-coated magnetic stir
bar. The solution was sampled through the septum by syringe
with a Teflon needle (Aldrich); typically, 1 mL was withdrawn
and mixed with an equal volume of methanol, which
quenched the reaction. The actual volume of liquid with-
drawn was determined by weight.
3+
to suppress any further reduction of Fe after the reaction.
When the measurement is carried out within 2 min, the
minimum time required for manipulation, a correction of
2+
only e3% in [Fe ] is needed. Hydrogen peroxide concen-
tration was measured by iodometry (20). The light intensity
was determined by chemical actinometry (21).
Kinetic Modeling. Reaction rates were simulated with
GEAR (PC version 1.11, Project Seraphim, Department of
Chemistry, University of Wisconsin, Madison, WI 53706),
which finds a numerical solution to the set of differential
equations defined by the users’ multistep reaction scheme
Photolyses were carried out in a 3-mL Teflon-lined septum
screw-cap quartz cell. The light source was a L201 tunable
1
50W xenon illuminator from Photon Technology Interna-
tional (South Brunswick, NJ). The band-pass was set at 10
nm. The reaction solution was prepared as above, and 3 mL
was transferred into the quartz cell containing a small Teflon-
coated magnetic stir bar. Typically, 50 µL of sample was
withdrawn by syringe and mixed with 350 µL of methanol.
Analyses. Phenol, 1,4-BQ, and 1,4-HQ were analyzed by
HPLC on a 5 µm, 25 cm × 5 mm Spherisorb ODS-2 C-18
(
22).
Results and Discussion
Dark Reactions. Figure 1 shows phenol loss (A), 1,4-BQ
formation (B), and Fe2 formation (C) in a series of experi-
+
2 2
ments carried out at 2 mM phenol, 4 mM H O , and different
VOL. 31, NO. 8, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2 4 0 1

FIGURE 1. Dark reactions. Effect of the ratio [Fe3+
mM). Initial concentrations: 2 mM phenol; 0.05 mM [Fe]
Lines represent simulations.
]
i
/[Fe2+
T
;4mMH
i
] (mM/
2
O
2
.
FIGURE 2. Dark reactions. (A, B) Effect of added hydroquinone (1,4-
HQ). (C) Effect of added catechol (1,2-HQ) on phenol loss. Initial
concentrations: 2 mM phenol; 0.05 mM [Fe3
+
i 2 2
] ; 4 mM H O . Lines
initial Fe3+:Fe2+ ratios, where total iron concentration [Fe]
T
represent simulations.
was kept constant at 0.05 mM. In all cases, the loss of phenol
displays autocatalysis; that is, there is a lag period followed
by a fast, first-order decomposition. In general, the pH
remained constant throughout the lag phase and dropped
slightly (to pH ∼2.7) during the fast reaction phase.
state is due to its consumption in the Fenton reaction (eq 4);
the subsequent rapid recovery in [Fe ] is due to production
2+
of reducing organic intermediates such as 1,2- and 1,4-HQ.
This accounts for the decreasing length of the lag period with
decreasing Fe3 :Fe ratio. The involvement of reducing
+
2+
Several important results of Figure 1 are to be noted. (i)
The lag period declines as the initial concentration of ferrous
intermediates such as hydroquinones is confirmed by the
2
+
2+
3+
ion ([Fe
(
]
i
) increases but still exists even when [Fe
]
i
) [Fe]
T
.
results of Figure 2, which shows that amendment of Fe /
ii) The rate of reaction after the lag period is nearly
independent of the initial Fe :Fe ratio. (iii) When Fe is
2 2
H O systems with 1,4-HQ or 1,2-HQ greatly reduces the lag
3
+
2+
period in relation to the amount added. The rate of the fast
phase becomes nearly independent of the amount of HQ
added as the stoichiometry of the reduction (HQ:Fe ) 1:2) is
approached and then exceeded.
2
+
present initially in the ferric state, [Fe ] exists at undetectable
levels throughout most of the lag period but then increases
rapidly as phenol enters its reaction phase, reaching steady
state at ∼0.6[Fe]
T
+
. (iv) When Fe is present initially in the
The phenol lag phase persists until reducing intermediates
build up to sufficient levels. Inspection of the rate constants
2
ferrous state, [Fe ] drops by more than an order of magnitude
within one minute but then recovers to the same steady state
concentration as when Fe was present initially in the ferric
state.
•
in Table 1 shows that HO
will preferentially react with phenol
over H O or Fe2+ until [phenol] declines below ∼0.005 mM.
2
2
•
The initial product of HO
attack is dihydroxycyclohexadienyl
•
These results can be explained in the following way. When
radical (DHCD , I, R ) OH) which reacts mainly by dehydra-
tion and peroxidation pathways in the presence of O (paths
2
3
+
i
[
Fe]
T
) [Fe
]
, the lag phase of phenol actually represents a
3
+
slow rate-limiting reduction of Fe by H
to Fe which sustains the Fenton reaction (eq 4). Progression
from the lag phase to the reaction phase is due to the buildup
2
O
2
(eq 3, Table 1)
A and B, respectively; Scheme 1). Hydroxylated products are
among several formed in these pathways. For benzene, path
B gives only about 50% hydroxylated products (16). Path C
leads exclusively to hydroxylation but requires a suitable
oxidant.
2
+
2
+
of [Fe ] through additional, more efficient reactions. One
3
+
such reaction is the reduction of Fe by hydroquinones, 1,4-
and 1,2-HQ, which are major initial products of phenol
degradation (Scheme 1), producing 1,4- and 1,2-BQ, respec-
tively (13, 14). These reactions take place by two one-electron
transfer steps with a semiquinone radical as intermediate
Subsequent experiments showed that quinones play a role
as oxidants in path C, driving the reaction toward hydroxy-
3
+
lation. Addition of 1,4-BQ reduces the lag period of Fe
/
+
H
2
O
2
reactions (Figure 3) and Fe2 / H
2
O
2
reactions (Figure 4)
3
+
(
eqs 22 and 23). Reduction of Fe by this route is much
in proportion to the amount of 1,4-BQ added. Furthermore,
3
+
faster than reduction of Fe by H
transition from the lag phase to the reaction phase. In Figure
B it can be seen that the maximum rate of 1,4-BQ formation
2
O
2
and accounts for the
0.05 mM 5-hydroxy-1,4-naphthoquinone (juglone) had the
same effect as 0.05 mM 1,4-BQ in the Fe3 / H
+
O
system (Figure
2
2
1
3). These results are consistent with the involvement of
•
coincides with that of phenol decomposition. 1,2-BQ was
not monitored due to its instability in water (23) and lack of
a reference standard. The final yield of 1,4-BQ is relatively
low because it is only one of several products formed (see
quinones in the oxidation of DHCD in path C. One-electron
•
oxidation of DHCD by 1,4-BQ affords a semiquinone radical
+
and dihydroxycyclohexadienyl cation (DHCD ) (eq 17, Table
+
1). Subsequently, DHCD loses a proton to give a dihy-
•
Scheme 1) and because it is consumed by reaction with HO .
droxybenzene. The semiquinone radical and two of the three
possible dihydroxybenzeness1,2- and 1,4-HQ but not the
1,3-isomer (resorcinol)sare capable of reducing Fe3 (13, 14).
This results in regeneration of the corresponding quinone.
Depletion of H
2 2
O toward the end of the reaction accounts
+
for the leveling off in [1,4-BQ]. The early rapid decline in
2
+
[
Fe ] observed when all Fe is present initially in the ferrous
2
4 0 2
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 8, 1997

FIGURE 5. Effect of substituting reaction 25 for reaction 17 on
simulations of the data previously shown in Figures 3A and C. Dashed
lines correspond to simulations with k25 ) 0.001k17. Solid lines
correspond to simulations with k25 ) k17
.
used the p- isomer values but showed that the quality of fit
is independent of whichever is used. For reaction 24, we
used the rate constant for the p isomer, since that of the o
isomer is unknown. For reaction 16 the branching ratio
between hydroxylation (k16a) and ring-opening (k16b) was
allowed to vary around 50:50, the approximate ratio reported
for benzene (16). In summary, a total of 28 reactions were
FIGURE 3. Dark reactions. Effect of added 1,4-benzoquinone (open
3+
symbols) or juglone (closed symbol) on reactions starting with Fe
.
3+
Initial concentrations: 2 mM phenol; 0.05 mM [Fe
Lines represent simulations.
i 2 2
] ; 4 mM H O .
used to simulate 10 sets of data with five unknowns (k16a, k16b
,
k
19, k21, k23f).
2
+
Simulations of data on [phenol] and [Fe ] are shown in
3
+
2+
Figures 1, 2, and 3 for systems of variable initial Fe :Fe
ratio, added hydroquinone, and added 1,4-BQ, respectively.
The values of the five unknowns resulting from the simulations
are in Table 1. While an exact fit of the data was not obtained,
considering the complexity of the reaction mechanism the
kinetic model does a good job of predicting the curve shape
and the position of the curves relative to each other.
The results of a sensitivity analysis are depicted in graphical
forms in the Supporting Information of this paper. Briefly,
this analysis shows that any combination of reasonable values
for the unknowns predicts qualitatitively the general curve
shape and position of the curves relative to one another. The
simulations are sensitive to k16 () k16a + k16b), as well as the
hydroxylation to ring-opening branching ratio, k16a:k16b. The
best value of k16 is 25 times greater than that reported for
anisole, while k16a:k16b (60:40) is close to the 50:50 ratio for
benzene (16). The values of the other unknowns seem
reasonable; in any case, the simulations are relatively
insensitive to 1 order of magnitude variations in them.
Several other important aspects of the simulations are to
be noted. First, the model consistently overestimates the
rate of the fast reaction phase. This is because the model
FIGURE 4. Dark reactions. Effect of added 1,4-benzoquinone on
2+
reactions starting with Fe . Initial concentrations: 2 mM phenol;
2+
0.05 mM [Fe
i 2 2
] ; 4 mM H O .
Thus, the quinones serve as catalysts to shuttle electrons from
•
3+
DHCD to Fe
.
Quinone should have no effect on a
complex as
mechanism involving a ternary HQ-Fe-H
2
O
2
•
the active oxidant, as proposed by the Hamilton (17) and
Litvintsev (18) groups.
assumes HO reacts only with phenol. We attribute the
overestimation to competition from degradation intermedi-
ates whose concentrations build up as phenol concentration
declines.
Kinetic Modeling. To test the proposed redox chain
mechanism, we simulated phenol decomposition using a
kinetic model that incorporates the reactions and their rate
constants (when available) listed in Table 1 (2, 10, 13-15,
Second, we have evaluated the contribution of reaction
•
25 to oxidation of DHCD .
2
4-29) except the photolysis reactions 1 and 2. In several of
the reactions, the rate constant depends on the substitution
pattern (o, m, p) in the benzene ring. The m position is
ordinarily the least reactive, and we have assumed this to be
true in cases where it has not been established. Values of k14
(
2), k18 (27), and k20 (28) are reported as composite values,
Although Fe3 has a greater reduction potential than 1,4-BQ
[E° ) 0.77 V (1) and 0.31 V (30), respectively], previous results
(15, 31), when taken together, show the order of reactivity
+
since the isomer-specific values either are unknown, or else
the isomer distribution could not be determined experimen-
tally. For k15, an average of the o and p values was used (15).
For reactions 22 and 23, the k22f and k23r and the ratio k22r/ k23f
are known for both o and p isomers (13, 14); we arbitrarily
3
-
toward cyclohexadienyl radicals to be Fe(CN)
Cu > Fe . When reaction 25 is added to the kinetic model
6
∼ 1,4-BQ >
2
+
3+
VOL. 31, NO. 8, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2 4 0 3

encompassing reactions 3-24, it has a negligible effect on
the simulations as long as k25 e 0.1k17. When k25 is between
TABLE 2. Rate Constants for the Fast Reaction Phase of
Phenol Oxidation by Photoassisted and Dark Fenton Reactions
0
.1k17 and k17, the effect of reaction 25 is merely to smooth
3+
2 2
and by Direct Photolysis of H O or Fe
out the transition from the lag period to the fast reaction
period in the phenol curves. On the other hand, when
reaction 25 is substituted for reaction 17, the kinetic model
fails completely to simulate even the qualitative features of
the experimental curves and fails to register the dependence
on quinone concentration, regardless of the value assigned
to k25 within the range 0.001k17 to k17. This is shown in Figures
(kobs/I0)
6
b
3
-2
wavelength
(nm)
I0 ×10
kobs × 10
× 10
-
1
reaction
(Einstein/L s)
(s
)
(L/Einstein)
Fe3+/H
O
/hνa
300
320
3.46
4.01
5.66
6.78
6.76
8.45
2.1
2.0
1.8
1.7
1.6
1.8
6.1
5.0
3.2
2.5
2.4
2.1
2
2
3
4
4
4
50
00
50
80
5
A and B, which are to be compared with Figures 3A and C,
respectively. Simulations to be compared with the results in
Figures 1 and 2 are in the Supporting Information. The failure
of simulations based on reaction 25 lends further support for
the hypothesized role of electron-transfer catalysis by quino-
nes.
b
H
2
O
2
+
/hν
300
300
3.46
3.46
0.016
0.0017
0.045
0.0049
Fe /hνc
3
Fe3+/H
O
2
a
dark
1.5
2
a
Photoassisted Reactions. Figure 6A shows the reaction
2 2
For photoassisted and dark Fenton reactions, 4 m M [H O ] and
and 0.05 mM Fe3 in the
+
0.05 m M [Fe3+] were used; k were calculated from Figure 6A assum ing
obs
of 2 mM phenol with 4 mM H
2
O
2
first-order kinetics. b For direct photolysis of 4 m M H
2 2
O . kobs was
dark and irradiated at different wavelengths from 300 to 480
nm. The lag, which is present under all conditions, declines
in light compared to dark and with decreasing wavelength.
Figure 6B shows that irradiation at 300 nm of phenol solutions
calculated from corresponding curve in Figure 6B and divided by 2 to
c
perm it direct com parison with photoassisted Fenton reactions. For
direct photolysis of 0.05 m M Fe3
+
.
k
obs was calculated from corre-
sponding curve in Figure 6B and divided by 10 to perm it direct
com parison with photoassisted Fenton reactions.
3
+
containing either H
lag in phenol decomposition and that the kobs values (Table
) are 2-3 orders of magnitude smaller than when H and
Fe are present together. The observed first-order rate
constant normalized to light intensity (kobs/ I ) for the fast
/ hν reaction is only slightly
2 2
O alone or Fe alone gives no obvious
2
2
O
2
We suggest that initiation also takes place by direct
photolysis of 1,2- and 1,4-BQ intermediates, resulting in
3
+
products that can reduce Fe3 . Figure 1B, where [Fe ]i )
+
3+
o
3
+
reaction phase in the Fe / H
2
O
2
[Fe] , shows that during the lag phase in the dark, thermal
T
wavelength-dependent, declining about 3-fold as the wave-
length increases from 300 to 480 nm (Table 2). The relative
reactions cause a gradual increase in [1,4-BQ] from initially
very low levels. 1,4-BQ has three major absorption bands in
the UV/ visible region: λmax ) 248 nm (Log ꢀ ) 4.32), 298 nm
(2.48), and 426 nm (1.33). According to Ononye et al. (33),
photolysis of 1,4-BQ alone in water gives semiquinone radical
and hydroxyl radical (eq 26),
insensitivity of kobs/ I
o
to wavelength suggests that phenol
oxidation is controlled by the Fenton and other thermal
2
+
reactions governing the steady state [Fe ] and that light
initiates the reaction.
Initiation becomes more efficient as the wavelength
decreases. Photolysis of H
2
O
2
(eq 1, Table 1) is possible below
its absorption edge of 350 nm. Photoreduction of Fe3
+
2
+
(
actually FeOH ) (eq 2) is possible below its absorption edge
at ∼410 nm, although it has a very low quantum yield above
50 nm (32). Nevertheless, it can be seen in Figure 6A that
3
wavelengths well above 400 nm are also capable of initiating
the reaction.
with subsequent reactions producing 1,4-HQ and hydroxy-
FIGURE 6. Irradiated reactions. (A) Effect of wavelength on phenol loss under photo-Fenton conditions. The dark reaction is also shown.
Dashed line is simulation of the dark reaction; dotted line is simulation of the photoassisted reaction including eq 26 (k26 ) ΦI El ) 0.003
. (B) Phenol loss during photolysis
0
-
1
-1
3+
i
M
min , l ) 1 cm, the light pass length). Initial concentrations: 2 mM phenol; 0.05 mM [Fe
]
2 2
;4mMH O
3+
of 0.49 mM Fe alone (O) or 8mM H O alone (0) at 300 nm. Initial concentration: 2 mM phenol.
2 2
2
4 0 4
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 8, 1997

SCHEME 2. Quinone (BQ) and Hydroquinone (HQ) Roles in
TABLE 3. Quantum Yields for the Conversion of 2.48 mM
,4-Benzoquinone as a Function of Wavelength in Irradiated
Aqueous Solution (0.1 M NaClO , pH 2.8)
Fenton Reactionsa
1
4
6
I0 × 10
Ls)
E
final observed
wavelength (Einstein/ (L/mol photolysis [hydroquinone quantum
(nm)
cm) time (min)
(mM)
yield, Φ
3
3
3
4
4
4
00
20
50
00
50
80
3.46
4.01
5.66
6.78
6.76
8.45
300
179
18
17
15
4
8
10
25
25
25
25
0.20
0.27
0.18
0.17
0.18
0.07
0.28
0.34
0.40
0.34
0.40
0.40
1
,4-benzoquinone. A similar reaction sequence presumably
occurs for 1,2-BQ. No complexation occurred between 1,4-
3
+
BQ and Fe by UV/ vis.
We observed 1,4-HQ accumulation in irradiated solutions
of 1,4-BQ alone at pH 2.8 and 0.1 M NaClO ; Table 3 lists the
4
quantum yield for 1,4-BQ conversion (based on 1,4-HQ
formation) as a function of wavelength from 300 to 480 nm.
The quantum yield varies between 0.28 and 0.40 almost
independently of λ . Quantum yields of 0.3 (34) and 0.47 (33)
using polychromatic light have been reported. Note in Table
a
Legend: SQ, sem iquinone; DHCD+ and DHCD , dihydroxycyclo-
•
hexadienyl cation and radical, respectively.
these pages in the microfilm edition of this volume of the
journal. Photocopies of the Supporting Information from
this paper or microfiche (105 × 148 mm, 24× reduction,
negatives) may be obtained from Microforms Office, American
Chemical Society, 1155 16th St. NW, Washington, DC 20036.
Full bibliographic citation (journal, title of article, names of
authors, inclusive pagination, volume number, and issue
number) and prepayment, check or money order for $21.00
for photocopy ($23.00 foreign) or $12.00 for microfiche ($13.00
foreign), are required. Canadian residents should add 7%
GST. Supporting Information is also available via the World
Wide Web at URL http:/ / www.chemcenter.org. Users should
select Electronic Publications and then Environmental Science
and Technology under Electronic Editions. Detailed instruc-
tions for using this service, along with a description of the file
formats, are available at this site. To download the Supporting
Information, enter the journal subscription number from your
mailing label. For additional information on electronic access,
send electronic mail to si-help@acs.org or phone (202)872-
3
that significant amounts of 1,4-HQ are generated at all
wavelengths.
In photoassisted Fenton reactions, the products of quinone
•
photolysis (eq 26) represent 2 equiv of HO because semi-
3
+
2+
quinone reduces Fe to Fe (eq 23). Moreover, the quinone
is regenerated. Thus, quinones act as photocatalysts in the
presence of Fe3 and H
period even above wavelengths where Fe and H O are not
2 2
+
2
O
2
and cause a reduction in the lag
3
+
photolyzed. Computer simulation incorporating reaction 26
into the kinetic scheme of Table 1 (but leaving out eqs 1-2
-
1
to simulate photolysis at >410 nm and taking ꢀ ) 17 L mol
-
1
cm ) shows that the lag phase is indeed reduced compared
to the dark reaction but is still longer than the experimental
lag phases for the photoassisted reaction. This suggests that
other reactions contribute to reduction of the lag period;
possibilities include photolysis of a Fe(III)-peroxide complex
(
5) and photolysis of Fe(III) complexes of organic acids (11).
3
+
Between 400 and 300 nm, photolyses of Fe and H
and 2) play an increasingly dominant role.
2 2
O (eqs
1
6
333.
Experimental and simulation work have demonstrated that
•
quinone intermediates shuttle electrons from the HO radical
adduct of the starting aromatic compound to Fe3 , thus
facilitating the degradation of the starting aromatic com-
pound. In light, quinones serve as photosensitizers via their
+
Literature Cited
(1) Strukul, G., Eds. Catalytic Oxidations with Hydrogen Peroxide as
Oxidant; Kluwer Academic: Boston, 1992.
(2) Buxton, G. V.; Greenstock, C. L.; Helman, W. P.; Ross, A. B. J.
Phys. Chem. Ref. Data 1988, 17, 513.
3) Safarzadeh-Amiri, A.; Bolton, J. R.; Cater, S. R. J. Adv. Oxid.
Technol. 1996, 1, 18.
(4) Pignatello, J. J. Environ. Sci. Technol. 1992, 26, 944.
(5) Sun, Y.; Pignatello, J. J. Environ. Sci. Technol. 1993, 27, 304.
6) Ruppert, G.; Bauer, R.; Heisler, G. J. Photochem. Photobiol. A:
Chem. 1993, 73, 75.
7) Zepp, R. G.; Faust, B.; Hoigne, J. Environ. Sci. Technol. 1992, 26,
3
+
•
conversion to semiquinone, which reduces Fe , and HO .
The role of quinone is summarized in Scheme 2. Phenol
degradation can be rationalized according to this scheme
without having to invoke the ternary complex (17, 18). In
waste treatment applications it is desirable to use iron in
catalytic concentration in order to hold down reagent costs
and minimize iron oxide sludge production. For the reaction
to be catalytic in iron, pathways must be available for the
regeneration of Fe(II) from Fe(III). Such pathways are
provided by reactions 22 and 23, as well as by reaction 2 in
light. In terms of effective degradation, hydroxylation and
peroxidation pathways (paths B and C, Scheme 1) are more
productive with respect to mineralization than dehydration
(
(
(
3
13.
(8) Faust, B. C.; Hoigne, J. Atmos. Environ. 1990, 24A, 79.
(9) Sun, Y.; Pignatello, J. J. J. Agric. Food Chem. 1993, 41, 308.
10) Walling, C. Acc. Chem. Res. 1975, 8, 125.
11) Balzani, V.; Carassiti, V. In Photochemistry of Coordination
Compounds; Academic Press: London, 1970; Chapter 10, pp
(
(
1
45-192.
(
path A). Therefore, the presence of co-oxidants, such as
(
12) Bielski, B. H. J. Free Radical Res. Commun. 1991, 12-13, 469.
(13) Baxendale, J. H.; Hardy, H. R.; Sutcliffe, L. H. Trans. Faraday Soc.
1951, 47, 963.
14) Mentasti, E.; Pelizzetti, E.; Saini, G. J. Chem. Soc. Dalton Trans.
1973, 19, 2609.
15) Raghavan, N. V.; Steenken, S. J. Am. Chem. Soc. 1980, 102, 3495.
16) Fang, X.; Pan, X.; Rahmann, A.; Schuchmann, H.-P.; Von Sonntag,
C. Chem. Eur. J. 1995, 1, 423.
oxygen and quinones, plays a crucial role in steering oxidation
into the desired pathways.
(
Acknowledgments
This work was done under a grant from the National Science
Foundation (BES 9414594). We thank Tina Arounsak and
Susan Devlin for technical assistance.
(
(
(
(
17) Hamilton, G. A.; Hanifin, J. W.; Friedman, J. P. J. Am. Chem. Soc.
1
966, 88, 5269.
Supporting Information Available
Seven figures plus their captions showing the results of the
simulation sensitivity analysis (9 pp) will appear following
18) Litvintsev, I. Y.; Mitnik, Y. V.; Mikhailyuk, A. I.; Timofeev, S. V.;
Sapunov, V. N. Kinet. Catal. 1993, 34, 71.
(19) Stookey, L. L. Anal. Chem. 1970, 42, 779.
VOL. 31, NO. 8, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2 4 0 5

(
(
20) Kolthoff, I. M.; Sandell, E. B.; Meehan, E. J.; Buckenstein, S. In
(29) Patel, K. B.; Willson, R. L. J. Chem. Soc., Faraday Trans. 1, 1973,
69, 814.
Quantitative Chemical Analysis, 4th ed.; Macmillan: New York,
1
969; Chapter 44, pp 842-860.
(30) Wardman, P. J. Phys. Chem. Ref. Data 1989, 18, 1637.
(31) Eberhardt, M. K.; Martinez, M. I. J. Phys. Chem. 1975, 79, 1917.
(32) Benkelberg, H.; Warneck, P. J. Phys. Chem. 1995, 99, 5214.
(33) Ononye, A. I.; McIntosh, A. R.; Bolton, J. R. J. Phys. Chem. 1986,
90, 6266.
21) Murov, S. L.; Carmichael, I.; Hug, G. L. In Handbook of
Photochemistry, 2nd ed.; Dekker: New York, 1993; Chapter 13,
pp 299-305.
(
(
22) Stabler, R. N.; Chesnick, J. Int. J. Chem. Kinet. 1978, 10, 461.
23) Musso, H. In Oxidative Coupling of Phenols; Taylor, W. I.,
Battersby, A. R., Eds.; Marcel Dekker: New York, 1967; Chapter
(34) Kurien, K. C.; Robins, P. A. J. Chem. Soc. B 1970, 855.
1
, pp 1-94.
(
(
24) Walling, C.; Goosen, A. J. Am. Chem. Soc. 1973, 95, 2987.
25) Bielski, B. H. J.; Cabelli, D. E.; Arudi, R. L.; Ross, A. B. J. Phys.
Chem. Ref. Data 1985, 14, 1041.
Received for review December 23, 1996. Revised manuscript
received March 27, 1997. Accepted April 8, 1997.
X
(
(
26) Rush, J. D.; Bielski, B. H. J. J. Phys. Chem. 1985, 89, 5062.
27) O’Neill, P.; Steenken, S.; Schulte-Frohlinde, D. J. Phys. Chem.
ES9610646
1
975, 79, 2773.
28) Foti, M.; Ingold, K. U.; Lusztyk, J. J. Am. Chem. Soc. 1994, 116,
440.
(
X
9
Abstract published in Advance ACS Abstracts, June 15, 1997.
2
4 0 6
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 8, 1997