Photocatalytic reduction of nitroorganics over illuminated titanium dioxide: Electron transfer between excited-state TiO2 and nitroaromatics

Article abstract of DOI:10.1021/jp973224l

The present study investigates the steady-state photocatalytic reduction of methyl viologen and a suite of monosubstituted nitrobenzenes. Reduction was carried out in deoxygenated, illuminated aqueous slurries of titanium dioxide (Degussa P25) in the presence of a sacrificial electron donor, 2-propanol. Langmuir-Hinshelwood plots were obtained for the reduction of each compound and found to be linear, with an average correlation of 0.98 and with a standard deviation on the correlations of 0.02. The concentration independent rates for nitroaromatic reduction obtained from these plots were normalized against the rate of methyl viologen reduction and the ratio was used to solve for the rate constant of nitroaromatic reduction, assuming a bimolecular model. The assumptions behind this procedure were tested by the use of the Marcus expression. Using the reorganization energy for the reaction as the fitting variable, it was possible to fit the measured rates to the predicted rates with a reorganization energy of 138 kJ/mol.

Full text of DOI:10.1021/jp973224l

J. Phys. Chem. B 1998, 102, 2239-2244
2239
Photocatalytic Reduction of Nitroorganics over Illuminated Titanium Dioxide: Electron
Transfer between Excited-State TiO2 and Nitroaromatics
John L. Ferry*
Department of Chemistry and Biochemistry, UniVersity of Texas at Austin, A5300, Austin, Texas 78712
William H. Glaze
Department of EnVironmental Sciences and Engineering, UniVersity of North Carolina at Chapel Hill,
7400, Chapel Hill, North Carolina 27599
ReceiVed: October 2, 1997; In Final Form: December 31, 1997
The present study investigates the steady-state photocatalytic reduction of methyl viologen and a suite of
monosubstituted nitrobenzenes. Reduction was carried out in deoxygenated, illuminated aqueous slurries of
titanium dioxide (Degussa P25) in the presence of a sacrificial electron donor, 2-propanol. Langmuir-
Hinshelwood plots were obtained for the reduction of each compound and found to be linear, with an average
correlation of 0.98 and with a standard deviation on the correlations of 0.02. The concentration independent
rates for nitroaromatic reduction obtained from these plots were normalized against the rate of methyl viologen
reduction, and the ratio was used to solve for the rate constant of nitroaromatic reduction, assuming a
bimolecular model. The assumptions behind this procedure were tested by the use of the Marcus expression.
Using the reorganization energy for the reaction as the fitting variable, it was possible to fit the measured
rates to the predicted rates with a reorganization energy of 138 kJ/mol.
Introduction
SCHEME 1: Photocatalytic One-Electron Reduction of
Nitrobenzene and Concommitant Oxidation of
Although photocatalysts such as TiO2 offer a means for the
potential use of low-energy photons for the transformation of
redox-active substrates, practical application of these materials
is hindered by the lack of knowledge concerning the funda-
mental chemistry of electron-transfer processes at the photo-
excited TiO2 surface, especially with regard to kinetic models
for these processes.
2-Propanol
Suspensions of photocatalyst particles large enough so that
no quantization of energy levels is encountered are generally
opaque, and the few studies that have addressed the problem
of measuring the electron-transfer kinetics of large photocatalyst
particles have been forced to resort to techniques such as flash
photolysis with diffuse reflectance transient absorbance mea-
surements that are only applicable to substrates with very high
extinction coefficients (ꢀ ≈ 10 000 M-1 cm-1) and fluorescence
quenching measurements.1-3 Although this technique is capable
of elucidating the decay kinetics of reactive intermediates formed
during the flash (g10-8 s time scale), it is only recently that
practitioners have become able to observe the primary interfacial
electron-transfer reactions happening in the picosecond regime.4
However, these works are limited to monitoring compounds with
absorption spectra that fall into a specific “window”, the region
of the UV-visible spectrum where the substrate absorbs and
TiO2 does not. With the exception of textile dyes, there are
few compounds of environmental relevance that can be observed
using these techniques.
electron oxidation of PVA by valence band holes goes on to
reduce the particle surface directly7,8 rather than react with
MV2+
.
There are a number of organic compounds of environmental
significance that are easily reduced photocatalytically, such as
nitroaromatics and nitroaliphatics,9,10 haloalkenes,11,12 and ha-
loalkanes.13,14 The reduction of nitrobenzene and monosubsti-
tuted nitrobenzenes has been shown to progress through the
series RNO2 - RNO - RNHOH - RNH2,9 presumably through
an anion radical intermediate between each stable product. In
this series of reactions, the first step, the loss of RNO2, is
attributed to the one-electron reduction of the substrate by
conduction band electrons and is illustrated in Scheme 1.
In this study, the ratio of the second-order rate constant for
the photocatalytic one-electron reduction of methyl viologen
by TiO2 sols over the rate constant obtained for MV2+ reduction
Photocatalytic reduction of organic compounds is encouraged
by the addition of electron donors to the system. It has been
shown that methyl viologen (MV2+) is easily reduced by
conduction band electrons in the presence of the electron donor
poly(vinyl alcohol) (PVA).5,6 In this system, the corresponding
R-hydroxy carbon-centered radical that results from the one-
S1089-5647(97)03224-0 CCC: $15.00 © 1998 American Chemical Society
Published on Web 02/27/1998

2240 J. Phys. Chem. B, Vol. 102, No. 12, 1998
Ferry and Glaze
by Degussa P25 was measured. This ratio, in conjunction with
the known value for the rate of photocatalytic MV2+ reduction,
was used to determine the second-order rate constants for the
one-electron reduction of nitroaromatics by large particles of
TiO2. This procedure enables the modeling of photocatalytic
rates based on electrochemical properties of the substrate that
are independent of such considerations as reactor configuration
or light intensity.
min to 230 °C, ramp at 50 °C/min to 280 °C, and hold at 280
°C for 1 min. Carrier gas was He (99.999%, Sunox), and the
flow rate at 80 °C was 1 mL/min.
Analytical Standards. Standards were dissolved in MeOH
(GC2 grade) and spiked into an aqueous matrix that is a duplicate
of the experimental conditions (2.5 mL of 0.2 M alcohol, 0.1
wt % TiO2, 0.001 M LiClO4). Standards were extracted in an
identical manner to samples. Standards were prepared on the
day of the experiment, so they aged in a similar way to samples
before analysis. The concentration range of all calibration
curves covers at least two logs of substrate removal, with a
minimum of seven points.
Methyl Viologen Analysis. Reactor preparation and charg-
ing for methyl viologen reduction experiments were exactly the
same as for nitroaromatic reduction. However, rather than a
stock solution of nitroaromatic, a stock solution of aqueous
MV2+ was used to charge the reactor with substrate. Direct
analysis of MV+• proved difficult due to its sensitivity to oxygen,
so an indirect method using TNM was developed instead. The
reaction between TNM and one-electron reducing agents to
produce the nitroform anion (NTF) is well known and is utilized
here to measure indirectly the concentration of MV+•.15 NTF
is quite stable in air, and its concentration may be measured by
its UV absorbance at 350 nm (ꢀNTF ) 13 500 L mol-1 cm-1).15
The reaction/sampling sequence is illustrated in the following
expression:
Experimental Section
Illumination. The reactor assembly has been previously
described.9 Illumination was carried out at 350 nm, using a Xe
arc lamp (Osram) in conjunction with a monochromator (Photon
Technologies Incorporated) to provide the necessary photons,
with a bandwidth of 3 nm. Lamp output did not change over
the duration of the experiments, as determined by ferrioxalate
actinometry.
Materials. TiO2 powder (Degussa P25, 30 nm diameter
particles in aggregate form) was a gift from Dr. Mike Prairie at
Sandia National Laboratories. Methyl viologen dichloride
(MV2+, 98%), nitrobenzene (NB, 99+%), 4-nitrotoluene
(4NT, 99%), 3-nitrotoluene (3NT, 99%), 4-nitrophenol (4NP,
99+%), 3-nitrophenol (3NP, 99+%), 4-nitrobenzonitrile (4NBN,
97%), 3-nitrobenzonitrile (3NBN, 98%), 4-(R,R,R-trifluorom-
ethyl)nitrobenzene (4TFNT, 98%), 3-(R,R,R-trifluoromethyl)-
nitrobenzene (3TFNT, 99%), 4-chloronitrobenzene (4NCB,
99%), 3-chloronitrobenzene (3NCB, 98%), 4-aminotoluene
(4AT, 99%), 3-aminotoluene (3AT, 99%), 4-aminobenzonitrile
(4ABN, 98%), 3-aminobenzonitrile (3ABN, 98%), nitroso-
benzene (NSB, 98%), aniline (AN, 99%), tribromomethane
(99%), tetranitromethane (TNM, 99%), and LiClO3 (99.5%)
were from Aldrich Chemical Co. and used as supplied.
Methanol (GC2 grade), methyl-tert-butyl ether (MTBE, GC2
quality), and 2-propanol (IsOH, GC2 quality) were from Burdick
and Jackson and used as supplied. Water was NIST grade
supplied by a custom Dracor filtration system, arranged in the
order: activated carbon, anion-exchange resin, cation-exchange
resin, activated carbon, and macroreticular resin polisher.
e-cb + MV2+ f MV+• cannula8
MV+• + TNM f nitroform filter,UV-vis8 (1)
Rather than remove the reduced MV2+ solution with a
syringe, the samples were taken by forcing the reaction mixture
through a cannula into an oxygen-free 2 mL volumetric flask,
precharged with 200 µL of methanolic TNM (2.00 mM). When
the appropriate volume was obtained, the cannula tip was
removed from the suspension and sampling ceased. The
receiving flask was stored until the experiment was completed,
then the contents were filtered through a 0.2 µm nylon filter
into a quartz cuvette for analysis in a UV-vis spectrophotom-
eter.
Nitroaromatics. A TiO2 suspension (0.1 wt % Degussa P25,
0.001 M LiClO4) was degassed under vacuum (5 µm Hg) for 1
h. The stock suspension was stored in a glovebox under
nitrogen and used for several experiments. All other stock
solutions were treated similarly.
Results and Discussion
Langmuir-Hinshelwood Plots. All of the substrates used
in this study exhibited saturation kinetics at higher concentra-
tions; that is, they exhibited a change from apparent first-order
at low substrate concentration to zero-order at high concentra-
tion. This has been observed many times before11,12,16,17 and
does not necessarily reflect an actual change in mechanism;
rather it is assumed to be a result of surface saturation. The
most common procedure (in photocatalytic studies) for extract-
ing a concentration-independent rate constant from such data
is through the use of Langmuir-Hinshelwood kinetics.16,17 This
approach begins with the assumption that the substrate is at
equilibrium with respect to partitioning between the surface and
the solution phase. The extent of surface coverage, θ, is given
by the expression:
Before assembling the reactor, 50 mL of the suspension was
added to the test tube body, followed by sufficient IsOH to make
a 0.20 M solution. Nitroorganic substrates were added from
concentrated 2-propanolic stock solutions. The reactor was then
sealed and moved out of the glovebox, and the solution was
allowed to equilibrate with stirring under N2 for 30 min before
illumination.
Samples (2.5 mL) were removed from the top sample port
using an adjustable 4 mL syringe (Manostat). Samples were
immediately placed in 40 mL EPA vials (I-Chem) precharged
with 2.5 mL of chilled extraction solvent (MTBE with CHBr3
internal standard) and stored sealed at 4 °C until extraction.
Extraction was carried out on a vortex mixer for 40 s. The
organic layer was analyzed for remaining substrate by GC-ECD
analysis.
θ ) K[S]/(1 + K[S])
(2)
where K ) rate of adsorption/rate of desorption and [S] is the
substrate. The experimentally observed rate is proportional to
the extent of surface coverage, θ, according to the following:
GC-ECD Conditions (Hewlett-Packard 5890). The injector
port was set for splitless operation at 240 °C. The autoinjector
volume was set at 3 µL. The analytical column was a 30 m
DB-5 with a 0.25 µm film thickness. The temperature program
was as follows: isothermal for 4 min at 80 °C, ramp at 12 °C/
rate ) - d[S]/dt ) kLH
θ
(3)

Photocatalytic Reduction of Nitroorganics
J. Phys. Chem. B, Vol. 102, No. 12, 1998 2241
TABLE 1: Langmuir-Hinshelwood Plot Data Summary
concentration
range (µM)
kLH × 10 -8
K × 10 4
substrate
(s-1
)
(M-1
)
r2
NB
55.7-2.0
23.3-2.3
51.0-3.5
23.8-4.7
52.0-4.6
67.7-4.0
22.6-1.9
47.3-4.8
26.7-1.0
55.9-9.7
51.0-4.8
150.0-10.7
1.45
5.10
4.57
2.90
1.16
1.75
1.60
1.66
1.65
1.62
2.30
0.61
17.6
0.99
0.99
0.96
0.99
0.99
0.93
0.99
0.96
0.98
0.99
0.98
0.97
4TFNT
4NBN
4CNB
4NP
5.86
6.41
2.61
31.1
7.39
8.18
20.3
10.8
9.02
1.46
5.27
4NT
3TFNT
3NBN
3CNB
3NP
Figure 1. Hammett σ-effects on the rate of nitrobenzene reduction,
para-substituted nitrobenzenes (b), meta-substituted nitrobenzenes ([);
k determined from eq 10.
3NT
MV2+
Substituting in the value for θ from in eq 2 yields the following
expression:
TABLE 2: Bimolecular Rates for Nitrobenzene Reduction
As Determined by Eq 10
k
in (×107)
k
in (×107)
rate ) kLHK[S]/(1 + K[S])
The inverse of eqn 4 is
(4)
(5)
substrate
M-1 s-1
substrate
M-1 s-1
NB
2.9
7.5
9.0
5.7
2.3
3.5
3TFNT
3NBN
3CNB
3NP
3.2
3.3
3.3
3.2
4.5
4TFNT
4NBN
4CNB
4NP
1
1
1
)
+
rate
kLH
kLHK[S]
3NT
4NT
Photocatalytic reduction of nitroorganics has been shown to
generate products that may compete with starting material for
conduction band electrons. To avoid this complication, initial
rates (ln([S]/[S]o) g -1, where [S]o is the initial [S]) were used.
Since the reactions were pseudo-first-order in S, indicating that
the concentration of adsorbed S was significantly higher than
the concentration of conduction band electrons, the initial rate
(in units of s-1) was expressed as
particle, etc. We assumed that the reactor coefficient was the
same for all of the experiments in this study, since all of those
terms were held constant from experiment to experiment,
varying only the substrate. Taking the ratio of kLH for MV2+
over that of kLH for another electron acceptor (such as RNO2)
yielded the equality:
MV2+
in
MV2+
k
LH
k
rateo ) kobs ln([S]/[S]o)
(6)
)
(9)
RNO2
in
RNO2
k
k
LH
so that under initial conditions the final expression was
RNO2
in
This was rearranged to solve for k
:
[S]o
1
1
)
+
(7)
RNO2
LH
kobs kLH
K
kLH
k
MV2+
RNO2
in
k
) kin
(10)
MV2+
(
)
kLH
where kLH is the concentration-independent rate constant for the
reaction. Thus, for each substrate a plot of [S]o-1 vs (kobs[S]o)-1
should yield a straight line with a slope of (kLHK)-1 and a
y-intercept of kLH-1, with kLH in s-1 and K in M-1. When the
data from the substrates reduced in this study were plotted in
this manner, approximately linear correlations were found for
every one (R2 values > 0.93). The concentration-independent
rate kLH, the equilibrium constant K, the concentration range
on the Langmuir-Hinshelwood plot from which they were
determined, and the correlation coefficient for the individual
plot are reported for each substrate in Table 1. All of the
Langmuir-Hinshelwood plots for this study are available in
the Supporting Information.
with the core assumption being that RNO2 and MV2+ interact
with the TiO2 surface in approximately the same manner.
Gra¨tzel and co-workers found that although polymeric viologens
adsorb strongly to TiO2, monomeric viologens do not, instead
relying on outer- or hard sphere interactions to react.6 Henglein
and co-workers observed that TNM did not adsorb strongly
either.18,19 Although TNM was not obviously a good analogue
for nitroaromatics, it does possess nitro-groups in plenty, and
if nitro-groups caused adsorption it should adsorb. After the
nitro-group is eliminated as a sorber, it seemed unlikely that
any of the substrates chosen for this study had functionalities
that allowed for ready sorption onto TiO2 in water, with the
possible exception of nitrophenol, although the possibility of
π-complexation could not be ruled out. The solution of eq 10
for all the compounds tested is provided in Table 2. Table 2
reveals that although the rate constant displays a marked
dependence on the identity of the ring substituent in the para
position, there was essentially no change in rate for substituents
in the meta position, which is graphically demonstrated in a
Hammett σ-plot for the reduction rates, Figure 1.20 Such an
effect is consistent with a reduction mechanism where electron
addition occurred from the conduction band of the excited
particle to the π*-orbital of the nitroaromatic; the presence of
Solving for the Intrinsic Rate. Obtaining the value of kLH
for methyl viologen reduction allowed the normalizing of the
rest of the data against the known intrinsic rate of methyl
viologen reduction, which is 1.2 × 107 M-1 s-1 6
.
In any
photocatalytic system, the intrinsic rate of reaction (kin) for any
compound can be solved for by the following equation:
kin ) kLH (reactor coefficient)
(8)
The reactor coefficient is a descriptor that entails all the things
that make a reactor configuration unique, such as the steady-
state concentrations of holes and electrons, photon dose per

2242 J. Phys. Chem. B, Vol. 102, No. 12, 1998
Ferry and Glaze
SCHEME 2: Homogeneous Model for Hard Sphere
Electron Transfer
or oxidation of the reduced particulates were observed on the
time scale of our experiments. Nitrobenzene was chosen as a
representative substrate due to its central location on the
Hammett σ-plot (Figure 1). The similarity of photoreduced TiO2
and trapped conduction band electrons has been demonstrated
previously by comparisons of the transient absorption spectrum
of conduction band electrons with the absorption spectrum of
photoreduced sols, which are identical.21
According to Marcus theory, the free energy of activation
for the overall reaction, ∆Gq, is given by the following
expression:
Z1Z2e2f
λ
4
∆G° 2
λ
∆Gq )
+
1 +
(11)
(
)
Dr12
SCHEME 3: Photocatalytic Model for Hard Sphere
Electron Transfer, Assuming Photoexcitation after the
Adsorption Step
In this equation, the first term is the electrostatic term and is
used to solve for change in free energy upon forming the
precursor complex when Z1 and Z2 possess a charge. Z1 and
Z2 are the charges on the donor and acceptor, e is the electronic
charge, f is a measure of ionic strength effects, D is the dielectric
constant of the solvent, and r12 is the collision distance for the
donor and acceptor.22 In this study, one of the species is
assumed to be the surface, activated by a photon postcomplex-
ation, and the other was the substrate. Since the experimental
pH was 4.85, very close to the point of zero charge for TiO2,
and there was no charge on the organic molecule (Z2 is zero),
the electrostatic term was neglected. The second term is the
parabolic term, representing the free energy change associated
with ∆E for the reaction and the impact product formation has
on the surrounding solvent environment. When product is
formed, the energy requirement involved in adjusting the solvent
and product to their new configuration is λ (reorganization
energy), and ∆G°′ is the value of ∆G° adjusted for the ionic
strength effects of the medium, according to the following
expression:
a node in the meta position for this orbital blocked the
substituent from affecting the ring, rendering all meta-substituted
nitrobenzenes essentially equal to nitrobenzene in their reactiv-
ity.20
e2f
Dr12
Predicting the Intrinsic Rate. The intrinsic rate constants
predicted by eq 10 for the para-substituted substrates were
analyzed by using a slightly modified form of the Marcus
equation. Typically, the Marcus approach to electron transfer
involves the interaction of two hard spheres, one of which is
an electron donor and one of which is an electron acceptor,
Scheme 2.21 Upon collision, which occurs with rate constant
of kd, the two form a precursor complex. Although they often
drift apart again, with a rate constant of k-d (and an equilibrium
constant K equal to kd/k-d), occasionally it will happen that the
energy levels of the donor orbital and acceptor orbital will
fluctuate to within (RT of each other, allowing an electron
transfer to occur. In situations where this step is irreversible,
it will have a rate constant of kel. The new complex is referred
to as a successor complex, and the reaction is complete when
it breaks apart into two new species. This was applied to
photocatalytic systems by assuming the precursor complex was
equivalent to substrate adsorbed to the particle surface prior to
illumination, Scheme 3, and that the association constant K
obtained from the Langmuir-Hinshelwood expression was
equivalent to the association constant of the precursor complex
for the solution-phase case. Random photoexcitation of the
semiconductor surface was considered analogous to the random
thermal excitation experienced by the precursor complex. The
alternative model, that photoexcitation preceded the formation
of the precursor complex, was tested by attempting to react
photoreduced TiO2 in the dark with nitrobenzene under other-
wise identical reaction conditions; no reduction of nitrobenzene
∆G°′ ) ∆G° + (Z1 - Z2 - 1)
(12)
In this work, with a very low ionic strength and no charge on
the electron acceptor, the second term is neglected and ∆G° is
considered equivalent to ∆G°′. So, in the present study, a
careful analysis of the reaction conditions suggests that the
complex Marcus equation (eq 11) simplifies to
λ
4
∆G° 2
λ
∆Gq ) 1 +
(13)
(
)
∆Gq was related to kel through the Eyring equation:
kel ) øZe-∆Gq/RT
(14)
In this expression, ø is the transmission coefficient and Z the
universal frequency factor, which has a value of 6 × 1011s-1 22
.
The transmission coefficient is assumed to have a value of 1,
which is equivalent to saying that all activated complexes yield
reaction. When eqs 13 and 14 were combined, an expression
was obtained that related kel with ∆G° for the process and λ
(the reorganization energy)
2
kel ) øZe-(λ/4RT)(1+(∆G°/λ))
that could be expressed as
(15)

Photocatalytic Reduction of Nitroorganics
J. Phys. Chem. B, Vol. 102, No. 12, 1998 2243
k
log el ) -
1 +
(16)
λ
4RT
∆G° 2
λ
(
)
Z
The steady-state approximation for the donor-acceptor
complex was used to solve for k
RNO2
in
, the rate constant obtained
by normalizing kLH for the nitrated organics against kLH for
MV2+, assuming the irreversible case
kd
RNO2
in
k
)
(17)
kdk-d
kelkd
1 +
Figure 2. Marcus plot for para-substituted nitrobenzenes. Points
represent measured rates; the line is the predicted rate based on eq 19,
with λ of 138 kJ/mol.
where kd is the bimolecular rate constant for the formation of
the precursor complex and k-d the rate constant for the back
reaction. In this instance, since the particle has essentially no
diffusion coefficient, kd was assumed to have a value of 108
M-1 s-1. Since the reaction was a surface reaction, the ratio of
kd/k-d is equal to K, the association constant describing
adsorption at equilibrium, yielding the following expression
TABLE 3: Reduction Potentials and Calculated Free
Energy Change for the Reduction of para-Substituted
Nitrobenzenes by Conduction Band Electrons
substituent
Eo′ (V vs NHE)
∆Go′ (kJ/mol)
H
-0.485
-0.383
-0.322
-0.428
-0.492
-0.465
3.44
-3.23
-9.12
1.10
7.35
4.74
CF3
CN
Cl
OH
CH3
kd
RNO2
in
k
)
(18)
kd
1 +
kelK
In this work, the values for K were derived from the Langmuir-
Hinshelwood plots and are listed in Table 1. When eq 13 was
substituted for the ∆Gq expression in eq 14 and the resulting
equation substituted into eq 18 for kel, the final expression was
(neglecting the ionic terms)
than the rate of donor oxidation, was the rate-determining step
in the system. The reduction kinetics for all substrates were
found to obey Langmuir-Hinshelwood kinetics, a result
consistent with the hypothesis that the reaction occurred solely
on the surface but not proof of it. The observation that the
reduction rates of para-substituted nitrobenzenes were sensitive
to the identity of the substituent and the rates of meta-substituted
nitrobenzene reduction were not is consistent with the hypothesis
that the rate-controlling step of nitroaromatic reduction was a
one-electron transfer of the π*-orbital of the substrate. The ratio
of reduction rates between each para-substituted nitrobenzene
and methyl viologen could be used to estimate the rate constant
for the initial electron transfer step: this was confirmed by
modeling the reaction using the aqueous one-electron reduction
potentials for the substrate to calculate a theoretical rate constant
and comparing it to the rates derived from the data.
kd
2
RNO2
in
k
)
e[(λ/4)(1+(∆G°/λ)) ]/RT
(19)
kd
1 +
KZ
The oxidation potential for conduction band electrons at pH
4.85 was determined to be -0.416 V vs NHE,6 based on the
known pH dependence of the position of the conduction band
edge for TiO2:
E°′) -0.13 - 0.059 (pH)
(20)
b
The one-electron reduction potentials for all the para-substituted
nitrobenzenes in this study were available from the pulse
radiolysis literature.23,24 These values were used to calculate
∆G°′ for the reduction of the para-substituted nitrobenzenes
by conduction band electrons, according to the following:
Acknowledgment. We express gratitude to the National
Renewable Energy Laboratory and the UNC Department of
Environmental Sciences and Engineering for supporting this
work and Mr. Randall Goodman and Mr. Clifford Burgess for
their considerable assistance in designing and maintaining the
reactor.
∆G° ) -F(E°′(ArNO2) - E°′(TiO2))
(21)
The resulting ∆G°′ values are presented in Table 3.
Equation 19 was used to solve for the value of k
RNO2
in
Supporting Information Available: Langmuir-Hinshel-
wood plots for MV2+, NB, 4CNB, 3CNB, 4NP, 3NP, 4NBN,
3NBN, 4NT, and 3NT (5 pages). Ordering information is given
on any current masthead page.
for
each of the para-substituted nitrobenzenes, with the reorganiza-
tion energy λ used as the fitting variable. The minimal residual
for the data set was obtained with λ ) 138 kJ/mol. The
literature value for the reorganization energy for the one-electron
reduction of nitrobenzene is in close agreement at 125 kJ/mol.25
The measured and calculated rates were plotted together in
Figure 2 for illustrative purposes. The reorganization energy
is probably so large due to the significant change in the structure
of the substrate since it progresses from a neutral, moderately
polar molecule to a charged anion radical, although this
explanation is speculative.
References and Notes
(1) Fox, M. A.; Dulay, M. T. J. Photochem. Photobiol., A: Chem.
1996, 98, 91.
(2) Draper, R. B.; Fox, M. A. Langmuir 1990, 6, 1396.
(3) Kamat, P. J. Phys. Chem. 1989, 93, 859.
(4) Colombo, D. P.; Bowman, R. M. J. Phys. Chem. 1996, 100, 18445.
(5) Duonghong, D.; Ramsden, J.; Gra¨tzel, M. J. Am. Chem. Soc. 1982,
104, 2977.
(6) Moser, J.; Gra¨tzel, M. J. Am. Chem. Soc. 1983, 105, 6547.
(7) Bard, A. J. J. Phys. Chem. 1982, 86, 172.
(8) Bard, A. J. Science 1980, 207, 139.
(9) Ferry, J. L.; Glaze, W. H. Langmuir, in press.
(10) Mahdavi, F.; Bruton, T. C.; Li, Y. J. Org. Chem. 1993, 58, 744.
Conclusions
The observation that the rate of reduction differed between
substrates indicated that the rate of substrate reduction, rather

2244 J. Phys. Chem. B, Vol. 102, No. 12, 1998
Ferry and Glaze
(11) Glaze, W. H.; Kenneke, J. F.; Ferry, J. L. EnViron. Sci. Technol.
1993, 27, 177.
(18) , Bahneman, D. W.; Henglein, A.; Spahnel, L. Faraday Discuss.
Chem. Soc. 1984, 78, 151.
(12) Kenneke, J. F.; Ferry, J. L.; Glaze, W. H. The TiO2-Mediated
Photocatalytic Degradation of Chloroalkenes in Water. In Photocatalytic
Purification and Treatment of Water and Air; Ollis, D. F., Al-Ekabi, H.,
Eds.; Elsevier: New York, 1993.
(13) Bahneman, M. J.; Moenig, J.; Chapman, R. J. Phys. Chem. 1987,
91, 3782.
(14) Choi, W.; Hoffmann, M. R. EnViron. Sci. Technol. 1995, 27, 1646.
(15) Asmus, K. D.; Henglein, A.; Ebert, M.; Keene, J. P. Ber. Bunsen-
Ges. Phys. Chem. 1964, 68, 657-663.
(16) Pruden, A. L.; Ollis, D. F. J. Catal. 1983, 82, 404.
(17) Ollis, D. F.; Hsiao, C.; Budiman, L.; Lee, C. L. J. Catal. 1984, 88,
89.
(19) Henglein, A. Pure Appl. Chem. 1984, 56, 1215.
(20) Lowry, T. L.; Richardson, K. S. Mechanism and Theory in Organic
Chemistry, 3rd ed.; Harper and Row: New York, 1987.
(21) Howe, R. F.; Gra¨tzel, M. J. Phys. Chem. 1985, 89, 4495-4499.
(22) Eberson, L. Electron Transfer Reactions in Organic Chemistry;
Spinger-Verlag: New York, 1987.
(23) Meisel, D.; Neta, P. J. Am. Chem. Soc. 1975, 97, 5188.
(24) Jagannadham, V.; Steenken, S. J. J. Am. Chem. Soc. 1984, 106,
6542.
(25) Nelsen, S. F.; Blackstock, S. C.; Haller, K. J. Tetrahedron 1986,
42, 6101.