Free Radical-Mediated Heterogeneous Photocatalytic Reduction of Metal Ions in UV-Irradiated Titanium Dioxide Suspensions

Article abstract of DOI:10.1021/jp036729m

This paper describes the indirect photocatalytic reduction of four metal ions: Cd²⁺, Zn²⁺, Mn²⁺, and Tl⁺ using formate radical anions (CO₂^.-) generated in UV-irradiated aqueous TiO₂ suspensions. Trends in the reactivity of these four metal ions are compared with corresponding rate constants reported in the literature for the homogeneous reactions of these metal ions with CO ₂^.- in aqueous media. In both cases, the reaction rates follow the same order: Cd²⁺ > Tl⁺ > Mn²⁺ > Zn²⁺. Using Tl⁺ as a model metal ion, a simple kinetics scheme is developed for the indirect reduction route, and the predictions from this model are shown to be in excellent agreement with experimental data.

Full text of DOI:10.1021/jp036729m

4784
J. Phys. Chem. B 2004, 108, 4784-4788
Free Radical-Mediated Heterogeneous Photocatalytic Reduction of Metal Ions in
UV-Irradiated Titanium Dioxide Suspensions
Sashikala Somasundaram, Yong Ming, C. R. Chenthamarakshan, Zoltan A. Schelly, and
Krishnan Rajeshwar*
Department of Chemistry and Biochemistry, The UniVersity of Texas at Arlington, Arlington, Texas 76019-0065
ReceiVed: September 12, 2003; In Final Form: February 17, 2004
This paper describes the indirect photocatalytic reduction of four metal ions: Cd2+, Zn2+, Mn2+, and Tl+
using formate radical anions (CO2•-) generated in UV-irradiated aqueous TiO2 suspensions. Trends in the
reactivity of these four metal ions are compared with corresponding rate constants reported in the literature
for the homogeneous reactions of these metal ions with CO2•- in aqueous media. In both cases, the reaction
rates follow the same order: Cd2+ > Tl+ > Mn2+ > Zn2+. Using Tl+ as a model metal ion, a simple kinetics
scheme is developed for the indirect reduction route, and the predictions from this model are shown to be in
excellent agreement with experimental data.
Introduction
Heterogeneous photocatalysis using UV-irradiated titanium
dioxide (TiO2) suspensions or films in aqueous media, is now
a mature field.1-6 The photocatalytic reduction of metal ions in
such media has both fundamental and practical interest.1,7 While
metal ions with rather positive standard reduction potentials [i.e.,
positive of 0 V versus the standard hydrogen electrode reference
(SHE), see ref 8] such as Cu2+, Cr6+, Au3+, Ag+, etc., have
been widely investigated, only a few studies exist on metal ions
that are much more difficult to be reduced, for example: Cd2+
,
Zn2+, Mn2+, Tl+.8 In this paper, we consider the photocatalytic
reduction of these four species using organic free radicals that
are photogenerated at the TiO2/water interface. Formate was
used as the reducing agent (i.e., free radical source) in this study.
Initial photoexcitation of TiO2 creates electron-hole pairs,
some of which escape recombination and diffuse to the oxide/
water interface.2,3 There they encounter electron acceptors (e.g.,
H+, O2) or electron donors (formate ions, OH-, water mol-
ecules). The photooxidation of formate (either directly by the
photogenerated holes or via the intermediacy of OH•) generates
the formate radical anion, CO2•-. While this species is well-
known to be strongly and quantitatively reducing toward a
variety of organic and inorganic compounds, there is some
uncertainty in the value for the standard reduction potential of
•-
CO2 on the SHE scale.9 Taking a value of -1.90 V,10
Figure 1. Relative disposition of the energy levels in TiO2 and solution
species. A pH of 7 was assumed for this interfacial situation; see
discussion in the text.
•-
however, Figure 1 shows that CO2 is thermodynamically
capable of reducing all four metal ions considered in this study,
namely Cd2+, Zn2+, Mn2+, and Tl+. On the other hand, the direct
reduction of Mn(II) and Zn(II) by the electrons photogenerated
in TiO2 is thermodynamically prohibited (Figure 1). [We assume
for this discussion that we are only dealing with thermalized
(i.e., not “hot”) electrons at the oxide/water interface]. By the
same token, the direct photocatalytic reduction of Cd(II) and
Tl(I) has only a modest driving force (Figure 1). Experimental
data are fully in agreement with these thermodynamic predic-
tions.8
generated on oxide semiconductor surfaces under UV illumina-
tion. Free radicals derived from formate or ethanol were shown
to enhance the photoreduction of p-nitrosodimethylaniline in
UV-irradiated ZnO aqueous suspensions.11 Photodeposition of
Pt, Ag, and Au was performed on a positively biased TiO2
electrode in an aqueous solution containing radical intermediates
•-
derived from alcohols.12 Nickel(II) was photoreduced by CO2
generated from the initial photooxidation of oxalate ions in UV-
irradiated TiO2 suspensions.13 In our own laboratory, the free
radical-mediated indirect photocatalytic reduction route has been
demonstrated for a variety of species: Ni(II),14 Pb(II),15 As(V),16
Zn(II),8,17,18 Cd(II),8,17,18 Mn(II),8,19 and Tl(I)8,20 in UV-irradiated
TiO2 suspensions.
As mentioned earlier, only a few studies exist on the indirect
photocatalytic reduction of metal ions using free radicals
* To whom correspondence should be addressed. E-mail: rajeshwar@
uta.edu.
10.1021/jp036729m CCC: $27.50 © 2004 American Chemical Society
Published on Web 03/24/2004

Photocatalytic Reduction of Metal Ions in UV-Irradiated TiO2
J. Phys. Chem. B, Vol. 108, No. 15, 2004 4785
Figure 3. Test of the model prediction (represented by eq 10) for three
initial concentrations of Tl(I). Other conditions are described in the
Experimental Section. The lines are least-squares fits and the error bars
are shown for cases where replicate experiments were performed.
Figure 2. Dependence of the initial rate of metal ion photoreduction
on the formate additive concentration in UV-irradiated TiO2 suspen-
sions. The initial concentration of the metal ion (added as the sulfate
salt) was 200 µM.
In the present study, we consider this indirect route in much
more detail than in the prior studies referenced above. Specif-
ically, we consider the importance of metal ion adsorption on
the TiO2 surface and also compare the heterogeneous photo-
reactivity trends for the four species [Cd(II), Zn(II), Mn(II), and
Tl(I)] with the corresponding rate constants reported in the
literature for the homogeneous reactions of these metal ions
with CO2•- in aqueous media. Finally, a simple kinetics scheme
is developed for the indirect reduction route, and the predictions
from this model, are shown to be in excellent agreement with
experimental data.
Experimental Section
Chemicals. Zinc sulfate (Aldrich, 99%), cadmium sulfate
(Aldrich, 98%), thallium sulfate (Johnson Matthey, 99.9%), and
manganese sulfate (Fisher, 99.2%) were used without further
purification. Sodium formate (99%) was from Alfa Aesar and
was used without further purification. The TiO2 (Degussa P-25)
used was predominantly anatase and had a specific surface area
of ∼60 m2/g. Deionized water (18 MΩ) was used in all cases
for preparing solutions or suspensions (unless otherwise speci-
fied).
Instrumentation and Procedures. In all the experiments the
TiO2 suspension dose was 2 g/L; the suspensions were agitated
by sparging ultrapure N2 through them. All measurements
pertain to the laboratory ambient temperature (25 ( 5 °C). The
photoreactor used was described previously.21 The light source
was a 400 W medium-pressure Hg arc lamp (Philips).
For obtaining the data presented in Figure 2, the initial
concentrations of the metal ions (Cd2+, Zn2+, Mn2+, and Tl+)
in all cases were 200 µM. Prior to UV irradiation, the metal
ion and sodium formate additive (ranging in concentrations from
1 to 30 mM) loaded TiO2 suspensions were equilibrated in the
dark for 30 min. Procedures for aliquot withdrawal and
determination of metal ion concentration have been described
previously.8,17,18 Manganese, cadmium, and thallium ion con-
centrations in solution were determined by flame atomic
absorption spectrometry (FAAS) at 279.5, 228.8, and 276.8 nm,
respectively. Zn(II) concentrations in solution were determined
by UV-Vis spectrophotometry at 492 nm after derivatizing with
4-(2-pyridylazo)-resorcinol.17,18
Figure 4. As in Figure 3 but for the set of experiments where Tl(I)
concentration was varied while holding the formate concentration
constant at the three values shown.
A more detailed set of experiments were performed on the
Tl(I) system for testing against the kinetics model. Thus, unlike
the procedure that led to Figure 2 where only a single metal
ion concentration (200 µM) was utilized, runs were done at three
initial concentrations of Tl(I) (200 µM, 400 µM, and 600 µM)
in the TiO2 suspensions containing variable amounts of formate
in the 1-12 mM range (Figure 3). Alternatively, the formate
concentrations were fixed at 5 mM, 10 mM, and 12 mM,
respectively, and the Tl(I) ion concentration was systematically
varied in the 100-800 µM range to generate the set of data in
Figure 4.
UV-Vis spectrophotometry employed a Hewlett-Packard
model HP 8452 diode array spectrometer. FAAS was performed
on a Perkin-Elmer 2380 atomic absorption spectrometer.
Results and Discussion
Trends in Photoreactivity and Dark Adsorption. Figure 2
contains plots of the initial reaction rate versus the formate
additive concentration ([D-]0) for the four metal ions considered
in this study. The initial rate was determined as usual from the

4786 J. Phys. Chem. B, Vol. 108, No. 15, 2004
Somasundaram et al.
TABLE 1: Rate Constantsa for the Homogeneous and
Heterogeneous Reactions between Metal Ions and Formate
Radicals
TABLE 2: Amount of Metal Ion Adsorbed on the TiO2
Surface with and without Formate Ions Presenta
amount adsorbed (µM)
homogeneous,b
heterogeneous,c
metal ionb
with formatec
without formated
metal ion
k (L mol-1 s-1
)
k′k′′ (L mol-1 s-1
)
Cd2+
Zn2+
Mn2+
Tl+
100
110
80
11.1
32.6
5.6
Cd2+
Zn2+
Mn2+
Tl+
5.1 × 106
< 2 × 104
< 2 × 105
3 × 106
3.2
0.14
0.15
0.55
50
20
a See Experimental Section for details on how adsorption was
quantified. b Initial concentration was 200 µM in all the cases. c Formate
concentration was 0.05 M except for Mn2+ where it was 0.15 M. d From
refs 18,19, 26.
a From ref 23. b Single electron transfer in all cases; see text. c See
text and eq 10 for details.
slopes of reactant conversion plots of metal ion concentration
versus time. In all four cases, the initial rate rapidly increases
with [D-]0 in a roughly linear fashion and then attains a plateau.
Figure 2 and Table 1. On the other hand, the situation with
Zn2+ shows that interfacial adsorption alone is not a determining
factor in the efficacy of its photocatalytic conversion in the
presence of the formate additive.
The Thallium(I) Model System: Kinetics Scheme and
Comparison with Experimental Data. Given that the photo-
catalytic (reductive) conversion of Tl(I) species involves a one-
electron-transfer situation, a simple reaction scheme can be
written for the indirect radical-mediated reaction route:
The rates were ordered as follows: Cd2+ . Tl+ > Mn2+
>
Zn2+. It must be noted that the photocatalytic conversion rates
are either zero (Mn2+, Zn2+) or negligible (Cd2+, Tl+) in the
absence of the formate additive.8 Thus the data in Figure 2
predominantly reflect trends in the radical-mediated indirect
conversion route. It is also worth noting that the ratio [D-]0/
[Mn+]0 (where [Mn+]0 ) initial metal ion concentration) far
exceeds unity in all the experiments pertinent to Figure 2. This
point is amplified in a subsequent section.
TiO2 hν(UV)8 e- + h+
(1)
(2)
k1
Since the band-edge positions of TiO2 in the solution and
some solution processes (e.g., acid-base equilibria) are de-
pendent on solution pH, considerable attention was focused on
this parameter. All the experiments in this study were performed
at the “natural” pH range of the oxide and metal ion/additive
laden suspensions, namely ∼6.5-7.0. The use of pH buffers
was avoided because of possible interference of the buffer agent
components with the oxide surface and/or the photocatalytic
process. The pH values of the solutions were also monitored
during the photocatalytic process; they did not change by more
than 0.2 units. This is within the experimental error; note also
that only the initial reaction rates were relevant for the purposes
of this study. Thus the reaction timespan did not exceed 5 min
in all the cases. The band-edge positions in Figure 1 are relevant
for a solution pH of ∼7.0, and it is also worth noting that since
the pKa of HCOO• is ∼2.3,22 almost all the radical species will
be in the dissociated state.
It is of interest to compare the reactivity trends in Figure 2
with the corresponding trends for the homogeneous reactions
of the metal ions with CO2•- radicals. Bimolecular rate constants
for the homogeneous (one-electron transfer) reactions are
available in the literature;23 Table 1 lists these values. The rate
constants broadly follow the same order: Cd2+ > Tl+ > Mn2+
> Zn2+, as observed in the heterogeneous case. The rate
constants for the heterogeneous reactions (see below) are also
listed in Table 1.
One could argue that initial adsorption of the metal ion (and
the formate additive) on the TiO2 surface could play an
important role in a heterogeneous photoreaction environment.
We have shown earlier8 that, in the absence of formate, the
proclivity of these metal ions to adsorb on the TiO2 surface is
quite low but varies among the four species considered. On the
other hand, induced adsorption (on the TiO2 surface) results
when chelating additives such as formate are present.17,18
Possible mechanistic factors underlying this interesting trend
have been discussed both by us17,18 and by other authors.24 Table
2 contains the dark adsorption data for the four metal ions. In
the presence of formate (the situation corresponding to the
experiments addressed in Figure 2), the adsorption affinity
follows the order: Zn2+ ≈ Cd2+ > Tl+ > Mn2+. Clearly, for
Cd2+, Tl+, and Mn2+, this trend parallels that seen earlier in
e- + h+ 9k 8 ∆ (heat)
2
D- + h+ 9k 8 D•- + H+
(3)
(4)
3
A+ + e- 9k 8 A
4
D
•- + M+ 9k 8 M + products
(5)
(6)
5
D
•- + X
9k 8 products
6
In the above scheme, the k’s are the rate constants, D- is an
electron donor (formate in our case), A+ is an electron acceptor
(e.g., H+), M+ is a metal ion (e.g., Tl+), and X is a D•-
scavenger (e.g., H•, OH•, or TiO2). It is reasonable to assume
that [X] . [D•-]. Thus step (6) can be treated as a pseudo-
first-order reaction, with a rate constant, k6 ) k6′[X].
Invoking the steady-state approximation for the formate
radical anion,
k3[D-][h+] ) k5[D•-][M+] + k6[D•-] )
[D•-]{k5[M+] + k6}
we obtain its concentration as
k3[D-][h+]
k5[M+] + k6
[D•-] )
(7)
Thus by assuming reaction 5 to be the rate-determining step,
the rate R for the reduction of the metal ion by the radical anion
can be written as
R5 ) k5[D•-][M+]
(8)
which, after substitution of eq 7, yields

Photocatalytic Reduction of Metal Ions in UV-Irradiated TiO2
J. Phys. Chem. B, Vol. 108, No. 15, 2004 4787
k3k5[D-][h+][M+]
TABLE 3: Values of k′k′′ (eq 10) from the Slopes of the
Plots in Figures 3 and 4
R5 )
(9)
k5[M+] + k6
parameter
k′k′′ (L mol-1 s-1
)
r2 a
[Tl+] ) 200 µM (Figure 3)
[Tl+] ) 400 µM (Figure 3)
[Tl+] ) 600 µM (Figure 3)
0.53
0.53
0.57
0.993
0.990
0.990
Considering the situation when k6 . k5[M+], eq 9 simplifies
to
avg: 0.54 ((0.02)
[HCOONa] ) 5 mM (Figure 4)
[HCOONa] ) 10 mM (Figure 4)
[HCOONa] ) 12 mM (Figure 4)
0.565
0.563
0.562
0.999
0.995
0.995
k3k5
R5 )
[D-][h+][M+]
k6
avg: 0.563 ((0.001)
which can be written as
R5 ) k′k′′[D-][M+]
with k′ ) k3[h+] and k′′ ) k5/k6.
Equation 10 affords a route to testing this simple kinetics
model against experimental data. Thus, the initial rate can be
plotted against the initial concentration of formate ([D-]0) with
[M+]0 held constant. Similar plots can be constructed from
experiments with [M+]0 varied while maintaining [D-]0 constant.
The two sets of plots should result in straight lines, and from
their slopes, the product k′k′′ can be computed. An important
test for the efficacy of the present model is that the values
obtained for k′k′′ from the two sets of experiments should be
the same, within experimental error.
Figure 3 displays plots of eq 10 generated from experiments
where the initial formate ion concentration was varied, while
maintaining the initial Tl+ concentration constant at three
different values. It must be noted, with respect to the Tl+ data
in Figure 2, that the formate ion concentrations (1-12 mM)
were varied within the “linear” rate regime, i.e., not in the
saturation region. As the counterpart for these sets of experi-
ments, the Tl+ ion concentration was varied for Figure 4. Most
of the data points in Figures 3 and 4 are averaged values
obtained from replicate (2-3) experiments so that the six plots
are built from a cumulative total of 60 runs in all. The lines in
Figures 3 and 4 are least-squares fits to the data points. Table
3 lists the six values of k′k′′ thus obtained from the slopes of
the corresponding plots in Figures 3 and 4. These values are
clearly in good accord with one another within the data scatter
and experimental error, which are further elaborated below.
a Correlation coefficient in the least-squares regression.
(10)
second electron is orders of magnitude higher than the first step,
e.g., 5.1 × 106 L mol-1 s-1 for Cd2+ f Cd+ versus 2 × 109 L
mol-1 s-1 for Cd+ f Cd0, ref 23. This is entirely consistent
with the relative stabilities of the metal ions in the two oxidation
states.
A simple kinetics scheme was developed and experimentally
tested for the first time for the free radical-mediated (indirect)
reduction route. Thallium(I) was used as the metal ion candidate
for reaction with the photogenerated formate radicals for this
purpose. In a historical sense, it is interesting to note that the
Tl+aq/Tl0 redox couple has been previously utilized for estima-
•-
tion of the reduction potentials of CO2 and the alcohol
radicals.9,10,25 These prior studies in homogeneous media utilized
pulse radiolysis for free radical generation. The excellent
agreement between our model predictions and experimental data
(Table 3, Figures 3 and 4) suggests that the simple scheme
represented by eqs 1-6 accounts for the essential aspects of
the indirect metal ion reduction route in UV-irradiated TiO2
suspensions and that thallium(I) is a good model for the
heterogeneous system as well.
Several points regarding our kinetics scheme deserve further
discussion. Thus, the role of the electron acceptor in scavenging
the photogenerated electrons in TiO2 (eq 4) is not specifically
addressed herein. The two possible sinks for electrons are H+
ions in solution (or at the interface) and the TiO2 surface itself.
Evidence for the participation of H+ ions was found in our
earlier study on the photocatalytic reduction of Mn2+ species.19
Charging of the TiO2 surface by the photogenerated electrons
and subsequent trapping of the latter (i.e., Ti(IV)-OH + e- f
Ti(III)-OH) can be directly monitored in chronopotentiometric
experiments under oxide illumination.26
General Discussion
Trends in the reactivity of four metal ions [Cd2+, Zn2+, Mn2+
,
While our first-generation kinetics scheme (eqs 1-10) is
undoubtedly simplified, the possible complicating role of other
solution processes such as radical dimerization (as pointed out
by a reviewer), can be discounted. For example, the dimerization
and Tl+] to undergo free radical-mediated heterogeneous
photocatalytic reduction in UV-irradiated TiO2 suspensions have
been presented in this study. These metal ions either have more
negative standard reduction potentials or lie very close to the
TiO2 conduction band edge (Figure 1) such that the direct
reduction route involving photogenerated electrons in TiO2 can
be neglected. The good correspondence of the present data with
the reactivity trends for these same four metal ions in homo-
geneous media (Table 1) suggests that the dominant role of TiO2
is as a source of the free radicals. Indeed, interfacial adsorption
does not appear to be an overriding factor, at least under the
experimental conditions utilized in this study (also see below).
It is also worth noting that the literature kinetics data in Table
1 pertain to single-electron-transfer processes. On the other hand,
the results in Figure 2 involve two-electron-transfer processes
with the exception of the Tl+ case. Nonetheless, it would appear,
from the similarity in trends in the two cases, that the rate-
determining step is the same in both the homogeneous and
heterogeneous processes. In this regard, it is worth noting that
the rate constants in homogeneous media for the transfer of the
of CO2 occurs only at pH values higher than ∼ 8.1.27 Other
•-
processes consuming the radicals, such as disproportionation,
are mechanistically accommodated by eq 6 in our scheme.
Under steady-state conditions, the electron and hole fluxes
on each (irradiated) TiO2 particle in the suspension must balance
one another.28 Furthermore, the assumption leading to eq 10
from eq 9 merits scrutiny. The condition k6 . k5[M+] is
consistent with the typically high rate of radical annihilation
steps (such as reaction 6) in the overall scheme. Indeed, this is
consistent with our experimental observation that a large excess
of the electron donor (i.e., the free radical source) relative to
the initial metal ion concentration is needed to sustain a
measurable photocatalytic reduction rate. This condition then
maximizes the encounter of a given metal ion with a free radical
in its immediate vicinity. In this sense, the rate constant k′′ ()
k5/k6) is a “branching ratio” in that it represents the competition
between reactions 5 and 6 for consumption of the photogener-

4788 J. Phys. Chem. B, Vol. 108, No. 15, 2004
Somasundaram et al.
ated free radical ions. It is worth noting that the opposite
situation where k5 [M+] . k6 (see eq 9) would have yielded a
final rate expression that does not contain a [M+] term. This
would clearly be in conflict with the experimental data (see
Figure 4).
Acknowledgment. This research was supported, in part, by
a grant from the U.S. Department of Energy, Office of Basic
Energy Sciences. Z.A.S. thanks The Welch Foundation for
support. We thank a reviewer for constructive criticisms of an
earlier version of this manuscript.
The postulate of TiO2 as a quencher of the D•- species (see
above) deserves comment. It is well-known that oxide electrodes
such as TiO2 act as hosts for electron injection from (unstable)
radicals leading to the so-called current-doubling effect.29 This
effect is likely to play a role in oxide suspensions as well, so
that TiO2 competes with the metal ions for electrons from the
photogenerated formate radical anions.
References and Notes
(1) Rajeshwar, K. J. Appl. Electrochem. 1995, 25, 1067.
(2) Linsebigler, A. L.; Lu, G.; Yates, J. T., Jr. Chem. ReV. 1995, 95,
735.
(3) Rajeshwar, K.; Ibanez, J. EnVironmental Electrochemistry; Aca-
demic Press: San Diego, 1997.
(4) Fujishima, A.; Rao, T. N.; Tryk, D. A. J. Photochem. Photobiol.
C: Photochem. ReV. 2000, 1, 1.
The plateaus seen in Figure 2 in the initial rates at formate
concentrations higher than 5-10 mM are explained as follows.
At formate concentrations lower than this regime, the photo-
catalytic reduction of 200 µM Mn+ ions is radical-limited. At
higher concentrations of formate, a proportionately higher
concentration of free radicals is generated so that the process is
no longer radical-limited and the conversion rate saturates.
Finally, this study has not specifically addressed the location
of the reaction zone involving the metal ions and the photo-
generated free radical ions. Nevertheless, initial concentrations
and initial rates were used in eq 10 for successfully testing the
model against the experimental data (Figures 3 and 4). This
suggests that interfacial adsorption (of either Tl+, formate ion,
or both) on the TiO2 surface exerts only a secondary effect on
the overall electron transfer rate. Close examination of the results
from analyzing the plots in Figures 3 and 4 (Table 3), however,
reveals that the data scatter (and the consequent slope uncer-
tainty) are distinctly higher in the cases where the formate
concentrations were systematically varied (Figure 3). This is
further borne out by the relative magnitudes of the regression
correlation coefficients in the two sets of cases (Table 3).
Recalling that formate ions provide the anchoring links for
adsorption of Tl+ (and other metal) ions on the TiO2 sur-
face,17,18,20,28 we attribute this additional data scatter to the
perturbation induced by the variable formate levels on the local
Tl+ ion concentrations at the TiO2/water interface.
(5) Blake, D. M. Bibliography of Work on the Heterogeneous Photo-
catalytic Removal of Hazardous Compounds from Water and Air. NREL/
TP-570-26797, National Renewable Energy Laboratory: Golden, CO, 1999.
Update No. 4 to October 2001, NREL/TP-510-31319.
(6) Diebold, U. Surf. Sci. Rep. 2003, 48, 53.
(7) Litter, M. I. Appl. Catal. B. EnViron. 1999, 23, 89.
(8) Rajeshwar, K.; Chenthamarakshan, C. R.; Ming, Y.; Sun, W. J.
Electroanal. Chem. 2002, 538-539, 173, and references therein.
(9) See, for example: Koppenol, W. H.; Rush, J. D. J. Phys. Chem.
1987, 91, 4429.
(10) Schwarz, H. A.; Dodson, R. W. J. Phys. Chem. 1989, 93, 409.
(11) Cunningham, J.; Zainal, H. J. Phys. Chem. 1972, 76, 2362.
(12) Baba, R.; Konda, R.; Fujishima, A.; Honda, K. Chem. Lett. 1986,
1307.
(13) Forouzan, F.; Richards, T. C.; Bard, A. J. J. Phys. Chem. 1996,
100, 18123.
(14) Lin, W.-Y.; Rajeshwar, K. J. Electrochem. Soc. 1997, 144, 2751.
(15) Chenthamarakshan, C. R.; Yang, H.; Savage, C. R.; Rajeshwar, K.
Res. Chem. Intermed. 1999, 25, 861.
(16) Yang, H.; Lin, W.-Y.; Rajeshwar, K. J. Photochem. Photobiol. A:
Chem. 1999, 123, 137.
(17) Chenthamarakshan, C. R.; Rajeshwar, K. Electrochem. Commun.
2000, 2, 527.
(18) Chenthamarakshan, C. R.; Yang, H.; Ming, Y.; Rajeshwar, K. J.
Electroanal. Chem. 2000, 494, 79.
(19) Ming, Y.; Chenthamarakshan, C. R.; Rajeshwar, K. J. Photochem.
Photobiol. A: Chem. 2002, 147, 199.
(20) Kajitvichyanukul, P.; Chenthamarakshan, C. R.; Rajeshwar, K.;
Qasim, S. R. J. Electroanal. Chem. 2002, 519, 25.
(21) Lin, W.-Y.; Rajeshwar, K. J. Electrochem. Soc. 1993, 140, 2477.
(22) von Sonntag, C.; Schuchmann, H.-P. Carbohydrates. Jonah, C. D.,
Rao, B. S. M., Eds.; Elsevier: Amsterdam, 2001.
(23) Neta, P.; Grodkowski, J.; Ross, A. B. J. Phys. Chem. Ref. Data
1996, 25, 709.
Notwithstanding such subtler effects, the overall kinetics
scheme and the simple rate expression in eq 10 do appear to
provide a satisfactory description for the free radical-mediated
reduction of metal ions in UV-irradiated TiO2 suspensions. At
this stage of our understanding of this interesting process, it is
not yet possible to further separate the contributions of the two
terms k′ and k′′ from the measured slopes. Efforts in this
direction are planned and fall beyond the scope of the present
study.
(24) Rajh, T.; Ostafin, A. E.; Micic, O. I.; Tiede, D. M.; Thurnauer, M.
C. J. Phys. Chem. 1996, 100, 4538.
(25) Butler, J.; Henglein, A. Radiat. Phys. Chem. 1980, 15, 603.
(26) Sun, W.; Chenthamarakshan, C. R.; Rajeshwar, K. J. Phys. Chem.
B 2002, 106, 11531.
(27) Fojtik, A.; Czapzki, G.; Henglein, A. J. Phys. Chem. 1970, 74,
3204.
(28) Kajitvichyanukul, P.; Chenthamarakshan, C. R.; Rajeshwar, K.;
Qasim, S. R. Adsorp. Sci. Technol. 2003, 21, 217.
(29) Morrison, S. R. The Chemical Physics of Surfaces; Plenum Press:
New York and London, 1977.