Kinetic and mechanistic investigations of multielectron transfer reactions induced by stored electrons in TiO2 nanoparticles: A stopped flow study

Article abstract of DOI:10.1021/jp108958w

The kinetics and the mechanism of various multielectron transfer reactions initiated by stored electrons in TiO₂ nanoparticles have been investigated employing the stopped flow technique. Moreover, the optical properties of the stored electrons in the TiO₂ nanoparticles have been studied in detail following the UV (A) photolysis of deaerated aqueous suspensions of TiO₂ nanoparticles in the presence of methanol. The reduction of common electron acceptors that are often present in photocatalytic systems such as O₂, H₂O₂, and NO ₃^- has been investigated. The experimental results clearly show that the stored electrons reduce O₂ and H₂O ₂ to water by multielectron transfer processes. Moreover, NO ₃^- is reduced via the transfer of eight electrons evincing the formation of ammonia. On the other hand, the reduction of toxic metal ions, such as Cu(II), has been studied mixing their respective anoxic aqueous solutions with those containing the electrons stored in the TiO₂ particles. A two-electron transfer is found to occur, indicating the reduction of the copper metal ion into its non toxic metallic form. Other metal ions, such as Zn(II) and Mn(II), could not be reduced by TiO₂ electrons, which is readily explained on the bases of their respective redox potentials. The underlying reaction mechanisms are discussed in detail.

Full text of DOI:10.1021/jp108958w

ARTICLE
pubs.acs.org/JPCA
Kinetic and Mechanistic Investigations of Multielectron Transfer
Reactions Induced by Stored Electrons in TiO2 Nanoparticles:
A Stopped Flow Study
Hanan H. Mohamed,†,‡ Cecilia B. Mendive,†,§ Ralf Dillert,† and Detlef W. Bahnemann*,†
†Institut f€ur Technische Chemie, Leibniz Universit€at Hannover, Callinstrasse 3, D-30167 Hannover, Germany
‡Chemistry Department, Faculty of Science, Helwan University, Ain Helwan, Cairo, Egypt
§Departamento de Química, Facultad de Ciencias Exactas y Naturales, Universidad Nacional de Mar del Plata, Dean Funes 3350,
7600 Mar del Plata, Argentina
ABSTRACT: The kinetics and the mechanism of various
multielectron transfer reactions initiated by stored electrons
in TiO2 nanoparticles have been investigated employing the
stopped flow technique. Moreover, the optical properties of the
stored electrons in the TiO2 nanoparticles have been studied in
detail following the UV (A) photolysis of deaerated aqueous
suspensions of TiO2 nanoparticles in the presence of methanol.
The reduction of common electron acceptors that are often
present in photocatalytic systems such as O2, H2O2, and NO3- has been investigated. The experimental results clearly show that the
stored electrons reduce O2 and H2O2 to water by multielectron transfer processes. Moreover, NO3- is reduced via the transfer of
eight electrons evincing the formation of ammonia. On the other hand, the reduction of toxic metal ions, such as Cu(II), has been
studied mixing their respective anoxic aqueous solutions with those containing the electrons stored in the TiO2 particles. A two-
electron transfer is found to occur, indicating the reduction of the copper metal ion into its non toxic metallic form. Other metal ions,
such as Zn(II) and Mn(II), could not be reduced by TiO2 electrons, which is readily explained on the bases of their respective redox
potentials. The underlying reaction mechanisms are discussed in detail.
’ INTRODUCTION
nanocrystalline TiO2 layers under negative applied potential is
strongest at 400 nm and gradually decreases to zero at 600 nm,
which is superimposed with the visible absorption spectrum of
conduction band electrons that is observed when applying a
more negative potential. The authors determined the extinction
coefficient of an electron trapped in a surface state to be 1900 M-1
cm-1 at λmax = 400 nm, more than three times higher than that of
a conduction band electron at the same wavelength.
A fundamental understanding of the dynamics of the charge
carrier transfer at TiO2 nanoparticles is of crucial importance for the
understanding and industrialization of photocatalytic reactions as
well as for the rational design of photocatalytic systems. During the
last decades, much effort was devoted to the study of the kinetics and
the mechanistic details of the interfacial electron transfer processes at
the TiO2/water interface using the laser flash photolysis technique.
Gr€atzel and co-workers6 dealt with the transfer of electron and hole
from the conduction and valence band of ultrafine TiO2 and CdS
particles to reactants in solution. Bahnemann et al.7 investigated the
kinetics and mechanistic details of the processes occurring upon band
gap irradiation in transparent suspensions of nanosized TiO2
particles in the absence as well as in the presence of electron and
hole scavengers. Recently, Gao et al.8 analyzed the kinetics of the
Photocatalytic applications employing UV irradiated TiO2
nanoparticles for the decontamination of organic compounds
as well as for the detoxification of hazardous metal ions in
wastewater have become a mature field of study during the past
decades.1,2 Upon bandgap excitation of semiconducting TiO2
particles, electrons are promoted to their conduction band with
the simultaneous generation of holes in the valence band, both of
which will be consumed in the photocatalytic reactions. In the
absence of suitable adsorbed hole scavengers and in the presence
of molecular oxygen (i.e., an electron scavenger), the remaining
holes oxidize surface water to produce adsorbed hydroxyl
radicals, which can subsequently induce further oxidation reac-
tions. However, in the presence of an adsorbed hole scavenger
such as methanol and in the absence of molecular oxygen the fate
of the photogenerated electrons is still not clear. Bahnemann
et al.3 reported that excess electrons are trapped close to the
surface of colloidal TiO2 particles forming Ti3þ, which is
characterized by broad absorption spectra with λmax = 650 nm.
O’Regan et al.4 observed near-infrared absorbance spectra when
TiO2 membranes were biased at a negative potential. The
authors attributed these absorbance spectra to free conduction
band electrons with only a small fraction of the electrons being
trapped at surface sites. Boschloo and Fitzmaurice5 reported that
the optical absorption of trapped electrons produced in
Received: September 19, 2010
Revised:
January 3, 2011
Published: March 02, 2011
r
2011 American Chemical Society
2139
dx.doi.org/10.1021/jp108958w J. Phys. Chem. A 2011, 115, 2139–2147
|

The Journal of Physical Chemistry A
ARTICLE
Scheme 1. Schematic Illustration of (A) the Method Employed to Store Electrons in TiO2 Nanoparticles and (B) the Stopped
Flow Experiment for the Multielectron Transfer Reactions
reactions of excess electrons in TiO2 produced by radiolysis
employing steady state and pulse radiolysis techniques.
Analytical Methods. The UV-vis spectra over a range of
200-800 nm were recorded employing a Varian Cary 100 Scan
UV-vis system. Transmission electron microscopy (TEM) was
conducted at 200 kV with a JEOL JEM-2100F-UHR field-emission
instrument equipped with a Gatan GIF 2001 energy filter and a 1k-
CCD camera. X-ray diffraction (XRD) data were acquired on a
Bruker AXS D4 Endeavor X diffractometer using Cu KR1/2, λR1 =
154.060 pm, λR2 = 154.439 pm radiation. The concentration of
dissolved oxygen in water was determined using a Microprocessor
Oximeter oxi 2000 in combination with a Trioximatic EO 200
electrode. The Gas Chromatographic (GC) measurements were
carried out using a Shimadzu GC (Shimadzu 8A) equipped with a
molecular sieve 5 Å packed column for analysis of inorganic gases.
Argon was used as the carrier gas and a thermal conductivity detector
(TCD) was employed for the gas detection.
The present work reports the results of an ongoing study
concerning the kinetics and mechanism of a variety of reactions of
stored electrons in TiO2 nanoparticles employing the stopped flow
technique. The basic concept of this work is rather straightforward,
the electrons are generated on the TiO2 nanoparticles by UV
illumination in the presence of a hole scavenger, stored on them
in the absence of molecular oxygen and subsequently used for the
reduction of various oxidants such as O2, H2O2, and NO3-. The
reduction of Cu2þ has also been studied following not only the
decay of the transient absorbance of the trapped electrons but also
by the detection of the respective surface plasmon absorbance band
of the resulting metal clusters.
Storing of Electrons on TiO2 Nanoparticles. The electrons
were stored on the TiO2 nanoparticles by illuminating a deaerated
(Argon purged) freshly prepared transparent colloidal suspension
of 3 g/L TiO2 (3.8 ꢀ 10-2 M in terms of TiO2 molecules) for 4 h
in the presence of 0.02 M methanol as a hole scavenger. The UV
(A) illumination was performed using a high-pressure Xe-lamp
(OSRAM HBO-500W) placed inside a quartz jacket and
equipped with a cooling tube (cf. Scheme 1 (A)). The employed
’ EXPERIMENTAL SECTION
Materials. All chemicals were of the highest purity available
and were used as received without further purification: TiCl4
(99.9%), methanol (99.99%), NaNO3 (99.995%), and H2O2
(30%). They were all purchased from Sigma-Aldrich Company.
All metal salts were chlorides (CuCl2, MnCl2, ZnCl2). Nessler’s
reagent was used as received. All solutions were prepared with
deionized water from a SARTORIUS ARIUM 611 apparatus
(resistivity = 18.2 Ω cm).
Preparation of TiO2 Nanoparticles. For transient absorp-
tion measurements in the stopped flow studies it is important
that the light scattering of the particles is negligible, which can be
achieved using transparent colloidal suspensions of small particle
size. The transparent TiO2 nanoparticles (2-3 nm particle size)
have been prepared following the recipe reported by Kormann
et al.9 A 3 mL quantity of titanium tetrachloride, TiCl4, prechilled
to -20 °C, was added slowly to 900 mL of vigorously stirred
deionized water at 0 °C. After being continuously stirred for 1 h,
the resulting transparent colloidal suspension was dialyzed using
deionized water to bring the pH to 2.3. The colloid was dried by
vacuum evaporation at 20 mbar at 27 °C. Off-white shining
crystals were obtained, which can be resuspended in pure water
to obtain perfectly transparent colloidal TiO2 suspensions.
UV (A) light intensity was 2.6 ꢀ 10-3 J cm-2 s-1
.
Determination of the Extinction Coefficient of the Elec-
trons Stored on TiO2 Nanoparticles. The extinction coeffi-
cient of the stored electrons in TiO2 at 600 nm has been
determined through the titration of the stored electrons with
benzoquinone (BQ) in the strict absence of molecular oxygen.
The stored electron absorbance at 600 nm decreases linearly with
increasing benzoquinone concentration. The extinction coeffi-
cient of the stored TiO2 electrons, that is, ε600nm = 600 M-1 cm-1
,
has been obtained from the slope according to the stoichiometry
of the reaction. This value agrees rather well with the values of
470 M-1 cm-1 and 800 M-1 cm-1 determined in acidic13 and
alkaline14 solutions, respectively.
Stopped Flow Investigations. Stopped flow experiments
were performed using a SX.17MV-R Rapid Mixing Spectro-
photometer (Applied Photophysics, United Kingdom) working
in the UV-vis range (from 200-700 nm) with 0.2 cm optical
2140
dx.doi.org/10.1021/jp108958w |J. Phys. Chem. A 2011, 115, 2139–2147

The Journal of Physical Chemistry A
ARTICLE
Figure 1. (a) HRTEM micrograph of the as-prepared TiO2, (b) UV-vis absorbance of an aqueous suspension of the as-prepared TiO2 nanoparticles
(4 g/L), and (c) XRD pattern of the as-prepared TiO2 (thick line) and pure anatase (thin line).
Figure 2. (a) Build up of the TiO2 electron spectrum recorded at different periods of UV irradiation of a 3.8 ꢀ 10-2 M deaerated aqueous TiO2
suspension in the presence of 0.02 M methanol at pH 2.3. (b) Absorbance at 600 nm as a function of the employed irradiation time. Reprinted with
permission from J. Photochem. Photobiol. A 2011, 217, 271-274. Copyright 2010 Elsevier B.V.
path and 1 ms dead time. In a typical stopped flow experiment
(cf. Scheme 1B), the suspension of the nanosized TiO2 particles
loaded with electrons was filled in one of the stopped flow
syringes and an aqueous solution of the electron acceptor (metal
ions, O2 (H2O), H2O2, N2 (H2O), etc.) was filled into the other
syringe. The solutions were injected into a mixing chamber (1:1
v/v), and the resulting reactant mixture reached the optical cell
where the change in absorbance with time was measured. The
kinetics of the electron transfer reactions was studied by follow-
ing the decay of the absorbance of the TiO2 electrons at 600 nm
or the build-up of the absorption signals of the products, for
example, the plasmonic bands of nanosized copper metal parti-
cles with time.
non-light-scattering, particles. Due to the band gap energy of
TiO2 (Eg = 3.2 V for anatase),10 the absorption rises rapidly at
wavelengths shorter than 390 nm. The crystalline character of
the nanoparticles was investigated by XRD (cf. Figure 1c).
The samples exhibit a broad diffraction pattern characteristic
of the crystal phase of pure anatase of 3 nm particle size (as
calculated from the XRD line broadening). The size of the
TiO2 particles was also determined using TEM measurements
to be 2-3 nm, which is consistent with the XRD measure-
ments (cf. Figure 1a).
Storing of Electrons on TiO2 Nanoparticles. The UV (A)
photolysis of a system containing TiO2 (3.8 ꢀ 10-2 M) and
methanol as hole scavenger in the absence of molecular oxygen
results in the formation of a transparent blue suspension,
indicating the generation of the stored electrons in the TiO2
particles. Figure 2 shows the absorption spectra of the aqueous
TiO2 suspension observed after different UV irradiation periods.
The blue coloration is characterized by a broad absorption band
in the visible region with a maximum observed after 25 min of
illumination and extending approximately from 510 to 700 nm.
’ RESULTS AND DISCUSSION
Characterization of TiO2 Nanoparticles. The absorption
spectrum of 4 g/L TiO2 nanoparticles resuspended in pure
water is shown in Figure 1b. The absence of any light absorp-
tion above 400 nm evinces the presence of very small, that is,
2141
dx.doi.org/10.1021/jp108958w |J. Phys. Chem. A 2011, 115, 2139–2147

The Journal of Physical Chemistry A
ARTICLE
In comparison to other works,13 such a maximum is not observed
in systems where the electrons are first produced by γ irradiation
in the solvent of TiO2 suspensions and then transferred to the
conduction band of the semiconductor (Figure 4 in ref 13).
Although those spectra do not show a clear maximum, they
remain approximately constant above 700 nm. We suggest that
the spectral differences between Figure 2 and Figure 4 in ref 13
evince the two different methods employed for the production of
stored electrons, that is, band gap excitation of the semiconduc-
tor to directly produce electrons in the conduction band (this
work) and the indirect production of stored electrons employed
in ref 13.
in the absence of methanol, due to the stored TiO2 electrons
(data not shown).
TiO2 f TiO2ðe-CB þ hVþB
Þ
ð1Þ
ð2Þ
e-CB þ TiO2 f e-ðTiðIIIÞÞ
•
hþVB þ CH3OHads f CH2OH þ Hþ
ð3Þ
•CH2OH þ TiO2 f e-CB þ CH2O þ Hþ
hþVB þ e-CB f heatðrecombinationÞ
hþVB þ Cl- f Cl•
ð4Þ
ð5Þ
The photonic efficiency ζ of the storing of the TiO2 electrons
is calculated from the ratio of the rate of the buildup of the stored
electrons (R) and the incident photon flux (J0):11
ð6Þ
R V
2Cl• f Cl2
ð7aÞ
ð7bÞ
3
ζð%Þ ¼
100
3
J0 A
3
Cl• þ Cl- f Cl•2-
with
33
33
33
I λ
3
J0 ¼
NA h c
2Cl•2- f Cl-3 þ Cl-
Cl-3 f Cl2 þ Cl-
ð7cÞ
ð7dÞ
3 3
I is the light intensity (2.6 ꢀ 10-3 J s-1 cm-2), which was
calculated based on the UV light meter measurement, λ is the
wavelength at the maximum emission (350 nm), NA is Avogadro’s
number, h is Planck’s constant, c is the light velocity, V is the
volume (6 ꢀ 10-2 l), and A represents the irradiated surface area
(12.56 cm2). The average photonic efficiency measured under
the above conditions was 3.8%.
Determination of the Number of Electrons per TiO2
Nanoparticle. Assuming that the TiO2 nanoparticles have a
spherical shape, the volume of one TiO2 particle with a diameter
of 3 nm is calculated to be 1.41 ꢀ 10-20 cm3. Considering that
the mass density of anatase TiO2 is 3.894 g cm-3, the average
weight of one 3 nm TiO2 nanoparticle is readily calculated to be
5.43 ꢀ 10-20 g. In our system of 3 g/L TiO2, the number of TiO2
particles will consequently be 5.53 ꢀ 1019 particles/L.
The formation rate of the stored electrons (R) is calculated
from the slope of the absorbance at 600 nm as a function of the
illumination time (Figure 2b) to be 6 ꢀ 10-8 M s-1
.
According to the mechanisms proposed in the literature,3 it is
suggested that the TiO2 electrons (et-) are trapped as Ti(III)
states. The shift of the absorption maximum of the stored TiO2
electrons to longer wavelengths upon longer periods of illumina-
tion may be attributed to the storing of the excess electrons in the
conduction band of TiO2 after the available surface traps have
been filled11 or by the trapping of the electrons in the bulk rather
than in surface traps, that is, assuming that bulk electron traps
exhibit slightly red-shifted absorption spectra. The mechanism of
electron storing can be summarized as follows (eqs 1-7a-d).
Upon bandgap illumination electrons are promoted from the
valence to the conduction band leaving holes behind in the
valence band (eq 1). While the electrons are trapped at Ti(III)
sites (eq 2) the valence band holes migrate to the surface of the
particles where they react with the adsorbed methanol molecules
producing R-hydroxyl methyl radicals (eq 3) that can directly
inject electrons to the conduction band of the TiO2 nanoparticle
(eq 4). The process described by eqs 3 and 4 is known as current
doubling.12 The mobile and partially trapped holes can therefore
recombine with excess electrons in the conduction band releas-
ing heat (eq 5) and react with adsorbed chloride ions producing
chlorine atoms (eq 6) and finally chlorine gas. This gas can either
be produced by a termination reaction between two chlorine
atoms (eq 7a) or involve the production of other radical
intermediates with a high oxidative power such as Cl2•-33
(eqs 7b-7d). Indirect indications of a loss of chloride ions by
conversion to chlorine are (i) the decrease in the chloride ion
concentration measured after illumination and (ii) the blue
coloration obtained by illuminating an aqueous TiO2 suspension
The concentration of stored TiO2 electrons was determined
from the absorbance measurements (cf. Figure 2) using the
extinction coefficient ε600 nm = 600 M-1 cm-1 obtained from the
BQ titration experiments. The concentration of stored electrons
ranges from 0.5 ꢀ 10-3 to 0.8 ꢀ 10-3 M (30 - 48 ꢀ 1019
electrons/L), corresponding to an average of about 6-9 elec-
trons per TiO2 particle.
Stopped Flow Investigations. As a blank experiment, the
absorption spectrum of the stored electrons on TiO2 was also
measured using the stopped flow instrument. The blue colloidal
suspension of the stored TiO2 electrons at pH 2.3 was mixed with a
deaerated aqueous HCl solution at the same pH in the stopped flow
apparatus and the time evolution of the absorbance was recorded.
Figure 3 shows the transient absorbance spectrum (380-660 nm)
recorded 5 ms after mixing with 20 nm resolution; the typical
absorption spectrum of stored TiO2 electrons recorded with the
spectrometer and prepared under the same conditions, that is, after
1:1 dilution of the stored TiO2 electron suspension with deaerated
aqueous HCl solution; and the last spectrum shown in Figure 2
normalized to the minimum at 390 nm to demonstrate that the
shape of the spectrum is unchanged at the same wavelength range.
The normalization was done to compare the spectrum shapes
because they were recorded after different times of mixing the
colloids with the solution at both equipments, that is, the stopped
flow and the spectrometer. As depicted in Figure 2, a broad
absorption maximum around 600 nm is as well observed in Figure 3.
The inset of Figure 3 shows the transient absorption signal at 600
2142
dx.doi.org/10.1021/jp108958w |J. Phys. Chem. A 2011, 115, 2139–2147

The Journal of Physical Chemistry A
ARTICLE
Figure 5. Time profiles of the decay of the e-
absorbance at 600 nm
TiO2
([e-TiO ] = 6.9 ꢀ 10-4 M) upon mixing with Ar-purged H2O2 solutions
2
Figure 3. Transient absorption spectra measured after 5 ms of mixing a
3.8 ꢀ 10-2 M deaerated TiO2 electron suspension at pH 2.3 (dots
connected by straight lines to help the eye); steady state absorbance of
the same system but scaled with a factor 5 due to the larger optical path
of these type of measurements as compared to the small cuvette in the
stopped flow equipment (solid line); and the last spectrum shown in
Figure 2, which was normalized to the minimum at 390 nm (dotted line).
Inset shows transient absorbance vs time signal after mixing of TiO2
electrons with deaerated water at pH 2.3 (HCl). Reprinted with permis-
sion from J. Photochem. Photobiol. A 2011, 217, 271-274. Copyright
2010 Elsevier B.V.
at pH 2.3 (HCl). Reprinted with permission from J. Photochem.
Photobiol. A 2011, 217, 271-274. Copyright 2010 Elsevier B.V.
fading rapidly, with the rate depending on the concentration of
O2 or H2O2, respectively. The kinetic curves were analyzed
according to a double exponential decay fitting. The observed
first order rate constants were found to increase linearly with
increasing the concentration of the electron acceptors (O2,
H2O2) and the respective second order rate constants were
obtained from the slope of the linear dependencies. The values of
the resulting second order rate constants are summarized in
Table 1.
The rate constant of the first initial decay (kIobs = 2.0 ꢀ
104 M-1 s-1) and that of the second decay (kII = 5.7 ꢀ
obs
103 M-1 s-1) for the reaction of stored TiO2 electrons with
molecular oxygen were found to be similar to the values of the
rate constants of the two slower processes kIII (7.3 ( 1.5 ꢀ
103 M-1 s-1) and kIV (1.0 ( 0.2 ꢀ 103 M-1 s-1) reported by
Gao et al.8 who studied the reaction of electrons formed by
radiation chemical processes in small TiO2 particles and who
suggested that these slow processes are due to the reaction of
excess TiO2 electrons with molecular oxygen in the bulk.
Bahnemann et al.7 determined a reaction rate constant of
7.6 ꢀ 107 M-1 s-1 for the reaction of trapped electrons
produced by laser flash photolysis of platinized TiO2 nanopar-
ticles with molecular oxygen producing the superoxide radical
which is considerably higher than the present value. However, it
is important to note that the experimental conditions differ
considerably from those employed in the laser flash photolysis
study and the present stopped flow experiments. While in the
former just one electron/hole pair was generated at each TiO2
nanoparticle during each single laser pulse, in the present study
the TiO2 particles contained between six and nine electrons
each. Consequently, the observed differences can be attributed
to multielectron transfer processes that are suggested to occur in
the present system. Moreover, while the laser flash experiments
were carried out employing a colloidal TiO2 suspension in
equilibrium with adsorbed and dissolved O2 molecules, there
is no molecular oxygen present on the surface of the TiO2
particles here. Hence, all observed electron transfer processes
clearly reflect reactions with dissolved electron acceptors. The
present results can therefore be regarded as an experimental
proof for the assumption of Gao et al.8 that the “slow processes”
are indeed reflecting the electron transfer to dissolved O2
molecules.
Figure 4. Time profiles of the decay of the e-
absorbance at 600 nm
TiO2
([e-TiO ] = 5.7 ꢀ 10-4 M) upon mixing with different concentrations of
dissolve2d O2 saturated aqueous solutions at pH 2.3 (HCl). Solid gray
lines over the points show the double exponential fits. Reprinted with
permission from J. Photochem. Photobiol. A 2011, 217, 271-274.
Copyright 2010 Elsevier B.V.
nm (mixing occurred at t = 0 s), which remains unchanged during
the entire measurement time of 1 s.
TiO2
1. Reaction of e- with O2 and H2O2. Figure 4 shows the
transient absorption signals observed upon mixing aqueous
stored TiO2 electron suspensions ([e-TiO ] = 5.7 ꢀ 10-4 M)
2
at pH 2.3 with aqueous solutions containing different concentra-
tions of O2 (0.65 ꢀ 10-4 M, air-saturated (1.25 ꢀ 10-4 M), O2
saturated (6.25 ꢀ 10-4 M), and argon saturated H2O). A similar
kinetic experiment for the reduction of hydrogen peroxide by the
stored TiO2 electrons was also performed. Figure 5 shows the
transient absorption signals observed upon mixing the blue
suspension of TiO2 electrons ([e-TiO ] = 6.9 ꢀ 10-4 M) at pH
2
2.3 with different concentrations of aqueous H2O2 solutions, that
is, 1.56 ꢀ 10-4, 3.12 ꢀ 10-4, and 6.25 ꢀ 10-4 M. It is observed
that in the argon flushed solution the absorbance signals decrease
very slowly with a rate of 8.9 ꢀ 10-7 M s-1, while in solutions
containing dissolved O2 or H2O2 the electron absorbance is
The rate constant of the first initial decay (kIobs = 2.7 ꢀ
105 M-1 s-1) and that of the second decay (kII = 3.2 ꢀ
obs
104 M-1 s-1) for the reaction of the stored TiO2 electrons with
hydrogen peroxide are found to be in very good agreement with
2143
dx.doi.org/10.1021/jp108958w |J. Phys. Chem. A 2011, 115, 2139–2147

The Journal of Physical Chemistry A
ARTICLE
Table 1. Summary of Rate Constants for the Reactions of Stored Electrons in TiO2 Colloidal Particles with Different Electron
Acceptors (Reprinted in Part with Permission from J. Photochem. Photobiol. A 2011, 217, 271-274. Copyright 2010 Elsevier B.V.)
electron acceptors
O2
reaction pathway
[e-
number of transferred electrons
k/M-1 s-1
TiO2]/M
e-
e-
þ O2 þ Hþ f HO2
5.7 ꢀ 10-4
1.1
2 ꢀ 104
•
TiO2
TiO2
þ HO2 þ Hþ f H2O2
2
5.6 ꢀ 103
•
H2O2 þ e-
f OH- þOH•
total (3.1)
2 e-TiO2 þTHiO22O2f 2 OH-
TiO2
H2O2
H2O2 þ e-
TiO2
f OH- þOH•
6.9 ꢀ 10-4
5.5 ꢀ 10-4
5.4 ꢀ 10-4
1.1
2.7 ꢀ 105
3.2 ꢀ 104
2 e-
þ H2O2 f 2 OH-
0.5
total (1.6)
4.6
NO3
2NO3 þ 12Hþ þ 10e-
f N2 þ 6 H2O (17)
1.7 ꢀ 104
1.6 ꢀ 103
-
-
TiO2
þ 5Hþ f N2H5þ (18)
TiO2
N2 þ 4e-
3.3
N2H5þ þT2iOe-2
þ 3Hþ f 2NH4þ (19)
total (7.9)
1.7
Cu2þ
Cu2þ þ 2e-
f Cu0
9.4 ꢀ 104
TiO2
nCu0 f Cu0
k1 = 0.6 s-1
n
OH- þ Hþ f H2O
ð11Þ
the respective kI and kII values obtained by Gao et al.8 for small
TiO2 particles. In the latter study, the fast and slow decay rate
constants were suggested to correspond to the reaction of stored
electrons with the adsorbed H2O2 molecules and H2O2 mole-
cules in the bulk, respectively.
•
•OH þ CH3OH f CH2OH
ð12Þ
ð13Þ
ð14Þ
•CH2OH þ TiO2 f TiO2ðe-Þ þ Hþ þ CH2O
The number of consumed electrons can also be determined
from the kinetic curves shown in Figures 4 and 5 for the case that
the initial concentration of the electron acceptor is lower than the
initial concentration of the stored electrons. Thereby, it is found
that in the presence of molecular oxygen an average of 3.1
electrons are transferred per O2 molecule, while in the presence
of H2O2, an average of 1.6 electrons are transferred to each H2O2
molecule. Safrany et al.13 estimated a similar value of 3.3 elec-
trons consumed during the reduction of one oxygen molecule
employing radiation induced TiO2 electrons. The authors attrib-
uted this to the occurrence of two parallel pathways, that is, the
two and the four electron reduction of O2 to water, which can be
explained as follows: in acidic medium molecular oxygen reacts
2e-TiO þ H2O2 f 2OH-
2
2. Reaction of e-
with Nitrate Ions. The photocatalytic
TiO2
reduction of nitrate has recently received special attention in view
of pollution control.16-21 Therefore, it is of particular interest to
study the kinetics of these reduction reactions by the stored TiO2
electrons.
There are two ways of reducing nitrate and nitrite ions, that is,
the dissimilative and the assimilative process, yielding N2 gas and
ammonia, respectively:22
•
with one electron producing HO2 radicals (eq 8), which are
known to disproportionate to O2 and H2O2 (eq 9). Thus, oxygen
effectively reacts with two electrons to produce H2O2. However,
no H2O2 could be detected here after mixing suspensions of
stored e-
with acidic aqueous solutions containing O2 using a
TiO2
sensitive enzymatic analytical method.15 Presumably, H2O2
remains on the TiO2 surface where it rapidly reacts with another
electron producing an •OH radical and a OH- ion (eq 10), which
is neutralized in acidic solution (eq 11). The •OH radical reacts
with methanol by hydrogen abstraction resulting in the forma-
tion of a •CH2OH radical (eq 12), which can in turn transfer an
electron to TiO2 (eq 13). This pathway results in the reduction of
one oxygen molecule to water consuming a net amount of only
two electrons. On the other hand, when hydrogen peroxide
reacts with two electrons (eq 14), the reduction of oxygen to
water requires a net amount of four electrons.
According to the scheme above, five or eight electrons are
required to reduce nitrate to N2 or ammonia, respectively.
Moreover, also the one-electron reduction of nitrate ions is
feasible in photocatalytic systems:
NO3- þ eT-iO þ Hþ f NO2 þ OH-
2
23
Eredox ¼ 0:79 Vðvs NHEÞ
ð15Þ
Hþ
O2 þ e- f O•2-
s
f
HO•
ð8Þ
ð9Þ
TiO2
2
In aqueous solutions, NO2 will, however, disproportionate
according to the following equation:
2HO•2 f O2 þ H2O2 ðdisproportionationÞ
2NO2 þ H2O f NO-3 þ NO2- þ 2Hþ
ð16Þ
H2O2 þ e- f OH- þ HO•
TiO2
Thus, eqs 15 and 16 cannot be easily distinguished from the two
electron reduction of nitrate forming nitrite.
14
Eredox ¼ 0:797 Vðvs NHEÞ
ð10Þ
2144
dx.doi.org/10.1021/jp108958w |J. Phys. Chem. A 2011, 115, 2139–2147

The Journal of Physical Chemistry A
ARTICLE
Figure 6. Time profiles of the decay of e-
absorbance at 600 nm
TiO2
([e-TiO ] = 5.5 ꢀ 10-4 M) upon mixing with deareated (argon purged)
Figure 7. Steady state UV-vis measurements, while Cu0 nanoparticles
grow as indicated from the surface plasmon absorbance with maximum
at 570 nm after bringing in contact TiO2 electrons ([e-TiO ] =5.5ꢀ 10-4 M)
with a Cu2þ solution ([Cu2þ] = 2.5 ꢀ 10-4 M)2 (a) stored TiO2
electrons before mixing (b) immediately after mixing, (c) after 2 min and
(d) after 4 min of mixing with the Cu2þ solution.
2
aqueous NaNO3 solutions at pH 2.3, solid lines show the double
exponential fits. Inset shows time profile of the decay of e-
absorbance
TiO2
at 600 nm ([e-TiO ] = 2.5 ꢀ 10-3 M) upon mixing with N2 saturated
2
aqueous solutions at pH 2.3.
In the presence of excess electrons two nitrate ions can rather
react by a 10 electron transfer step to form N2 via
stoichiometry of the assimilative pathway of reduction of nitrate
ions to ammonia.
According to the observed results and the estimated number of
electrons for each decay step (see Table 1), the mechanism can
be described as follows: in the presence of excess electrons
initially a five electron transfer process occurs leading to the
reduction of nitrate to N2 (eq 17), the latter can be further
reduced to ammonia through eqs 18 and 19. To confirm the
2NO-3 þ 12Hþ þ 10eT-iO f N2 þ 6H2O
2
24
Eredox ¼ 1:25 Vðvs NHEÞ
ð17Þ
N2 in turn can react by a four electron transfer reaction to
produce N2H5þ ions
reaction of N2 gas with excess e-TiO , a stopped flow experiment
2
was performed mixing an aqueous TiO2 electron suspension
N2 þ 4e- þ 5Hþ f N2Hþ5
(2.5 ꢀ 10-3 M e-TiO ) with a N2 saturated aqueous solution
TiO2
2
([N2 aq] = 3.35 ꢀ 10-4 M)25 at pH 2.3. The e-
absorbance
24
TiO2
Eredox ¼ - 0:23 Vðvs NHEÞ
N2H5þ will further react with two electrons to produce 2NH4
ions
ð18Þ
decays according to a first order rate law (cf. inset of Figure 6).
From the resulting first order rate constant of 5.9 s-1 a second
order rate constant k2 = 1.8 ꢀ 104 M-1 s-1 is calculated. This rate
constant is faster than that obtained for the slow process in the
reduction of nitrate which is expected to be due to the reduction
of molecular nitrogen to ammonia (eqs 18 and 19). This can be
readily understood as the steady state concentration of N2 gas
that is reduced during the reduction of nitrate ions which will be
much lower than the N2 concentration employed in the experi-
þ
N2Hþ5 þ 2e- þ 3Hþ f 2NH4þ
TiO2
24
Eredox ¼ 1:27 Vðvs NHEÞ
ð19Þ
In the present kinetic study, stopped flow experiments were
performed by mixing the stored electrons in TiO2 and aqueous
nitrate solution at acidic pH (2.3).
ment of e-
and molecular nitrogen (e-
þ N ).
TiO2
TiO2
2
It is calculated that ∼6 electrons were consumed per N2
molecule in agreement with the theoretical number of electrons
required for the reduction of N2 gas to ammonia.
The transient decays of the absorbance of the TiO2 electrons
([e-TiO ] = 5.5 ꢀ 10-4 M) upon mixing with various concentra-
tions of2sodium nitrate in aqueous solution are shown in Figure 6.
The kinetic curves were analyzed according to double exponen-
tial decays (cf. solid lines in Figure 6). The observed first order
rate constants of both decay processes are found to increase
linearly with increasing nitrate ion concentrations. The respec-
tive second order rate constants were obtained from the slope of
the linear dependencies and are summarized in Table 1. The
values of the second order rate constants obtained here differ
considerably from those reported by Gao et al.8 This is readily
explained by the fact that these authors observed the one electron
reduction of nitrate ions while in the present study a multi-
electron transfer is occurring (the concentration of electrons
employed in their experiments are considerably lower than the
electron concentration employed here).
The formation of ammonia was furthermore confirmed by
performing separate experiments, that is, by the 1:1 mixing of
aqueous suspensions of TiO2 electrons and NaNO3 solutions
(under the conditions identical to those of the kinetic experi-
ments) in an isolated reactor and measuring the evolved ammo-
nia by Nessler’s method.26 The pH of the resulting mixed
solutions were adjusted to 9 by adding NaOH, then the solutions
were cooled to 4 °C. Nessler’s reagent (0.09 M solution of
potassium tetraiodomercurate(II) (K2[HgI4]) in 2.5 M potas-
sium hydroxide) was then added slowly to the cooled solutions.
Strong yellow-brown colorations were observed indicating the
presence of ammonia. In addition, GC measurements of the
headspace above the mixed solutions indicated the absence of
any measurable amounts of N2 gas.
It is in fact estimated that an overall of about 8 electrons are
3. Reaction of e-
with Zn2þand Mn2þ. It was observed
TiO2
-
consumed per NO3 ion which is in consistence with the
that after mixing the aqueous suspension of stored TiO2
2145
dx.doi.org/10.1021/jp108958w |J. Phys. Chem. A 2011, 115, 2139–2147

The Journal of Physical Chemistry A
ARTICLE
electrons (2 ꢀ 10-3 M) with an acidic solution containing either
Zn2þ (0.5 ꢀ 10-3 M ZnCl2) or Mn2þ (0.5 ꢀ 10-3 M MnCl2),
the absorbance signals remain stable even over 200 s, indicating
that eqs 20 and 21 did not occur.
Zn2þ þ 2e-TiO f Zn0
2
24
Eredox ¼ - 0:76 Vðvs NHEÞ
ð20Þ
ð21Þ
Mn2þ þ 2e-TiO f Mn0
2
24
Eredox ¼ - 1:18 Vðvs NHEÞ
Figure 8. Time profiles of the decay of e-
absorbance at 600 nm
TiO2
([e-TiO ] = 5.5 ꢀ 10-4 M) upon mixing with deareated (Ar purged)
2
This can be attributed to the fact that the values of the redox
potentials of Zn2þ and Mn2þ are considerably more negative
than the potential of the stored electrons on the TiO2 (-0.5 V vs
NHE).27
aqueous CuCl2 solutions of different concentrations at pH 2.3, solid
lines show the first order fits.
4. Reaction of e-
with Cu2þ. Effluents from electroplat-
The kinetics of the reaction between TiO2 electrons and
Cu2þ ions were studied by performing stopped flow mixing
experiments of TiO2 electron suspensions with various con-
centrations of CuCl2 in aqueous solutions in the absence of
oxygen (Figure 8). A first order rate law was observed at all
Cu2þ concentrations. The second order rate constant ob-
tained from the slope of this linear dependency is k2 = 9.4 ꢀ
104 M-1 s-1. This value is very similar to the value obtained by
Gao et al.8 (1.3 ( 0.3 ꢀ 105 M-1 s-1), however, the authors
attributed this rate constant to the one electron reduction of
Cu2þ to Cuþ.
Following the fast initial decay, a slow build up of a transient
absorbance is observed at 570 nm (not shown). The build-up of
this signal can be fitted by a first order rate kinetics. The second
order rate constant of the TiO2 electron absorbance decay and
the first order rate constant of the Cu0 particles growth are
summarized in Table 1.
TiO2
ing, electrical and leather industry, and machinery contain often
high concentrations of copper ions. This places Cu2þ on the list
of strategically important metals. Moreover, environmental
pollution caused by copper contamination must not be
neglected.28 The detailed understanding of the kinetics and the
mechanism of the photocatalytic reduction of Cu2þ ions is thus
of crucial importance to develop respective systems for the
removal of these metal ions from waste streams. It has been
reported that Cu2þ can be reduced to Cu0 by TiO2 electrons.29,30
Herrmann et al.31 suggested that in the presence of TiO2
suspensions Cu2þ ions are not reduced to Cu0 but to Cuþ even
in the absence of molecular oxygen. In the present study, aqueous
solution of Cu2þ ions were mixed with stored TiO2 electrons and
the reaction products were detected by steady state absorption
spectroscopy as well as by following the reaction kinetics by
stopped flow measurements.
It was calculated from the kinetic decays that an average
between 1.9 and 1.5 electrons are consumed per Cu2þ ion which
is lower than the theoretical value of 2 because the absorbance of
Figure 7 shows the steady state UV-vis spectra recorded
immediately and after various times of mixing of TiO2 electrons
and Cu2þ ions from CuCl2 aqueous solution. Because the
reduction of Cu2þ ions depends strongly on the presence of
molecular oxygen, great care was taken to exclude molecular
oxygen as far as possible during the manipulation of the solutions.
As the broad absorption of TiO2 electrons decays, the surface
plasmon absorption of colloidal copper with a maximum at 570
nm32 develops. The blue suspension of TiO2 electrons turned to
the ruby-red color of colloidal copper after mixing with an
aqueous solution of Cu2þ ions. A two-electron transfer reaction
is thus assumed to take place resulting in the formation of copper
atoms (eq 22), which form colloidal copper nanoparticles
(eq 23)
0
the products (Cun ) also contribute to the absorbance at the
same wavelength where e-
absorb.
TiO2
’ CONCLUSIONS
This work focuses on the analysis of the dynamics of multi-
electron transfer reactions in photocatalytic systems utilizing the
stopped flow technique to study the reactions of stored electrons
in TiO2 nanoparticles with various electron acceptors. The TiO2
conduction band electrons have been photogenerated on nano-
sized TiO2 particles upon UV (A) illumination in the presence of
a hole scavenger (methanol) and stored in the absence of
molecular oxygen. The results obtained from this study can be
summarized as follows: (1) the stored TiO2 electrons reduce O2
and H2O2 to water by multielectron transfer processes; (2) the
stored TiO2 electrons are able to reduce nitrate ions to ammonia
through the transfer of eight electrons with the intermediate
formation of molecular nitrogen; (3) Cu(II) ions are reduced
through a two-electron transfer reaction to form Cu(0) nano-
particles which are detected by their typical surface plasmon
resonance band at 570 nm. Neither the reduction of Zn2þ nor
that of Mn2þ can be induced using stored TiO2 electrons in these
systems, attributing it to the thermodynamic infeasibility of the
respective reduction processes.
24
Cu2þ þ 2eT-iO f Cu0 Eredox ¼ 0:34 Vðvs NHEÞ
ð22Þ
2
nCu0 f Cu0n
ð23Þ
It was difficult to determine the size of the deposited copper
particles by means of transmission microscopy since the particles
were very sensitive toward molecular oxygen and completely
dissolved by exposure to air within a few minutes. Moreover, no
theoretical correlation between the size of the copper particles
and the position of their surface plasmon band could be found in
the literature. Hence, the particle size of the Cun0 could not be
estimated from the UV-vis spectra (cf. Figure 7).
2146
dx.doi.org/10.1021/jp108958w |J. Phys. Chem. A 2011, 115, 2139–2147

The Journal of Physical Chemistry A
ARTICLE
’ AUTHOR INFORMATION
(31) Herrmann, J.-M.; Disdier, J.; Pichat, P. J. Phys. Chem. 1986,
90, 6028–6034.
(32) Henglein, A. J. Phys. Chem. B 2000, 104, 1206–1211.
(33) Nadtochenko, V; Kiwi, J. Inorg. Chem. 1998, 37, 5233–5238.
Corresponding Author
*E-mail: bahnemann@iftc.uni-hannover.de.
’ ACKNOWLEDGMENT
H.H.M. gratefully acknowledges the Egyptian Ministry of
Higher Education for granting her a doctoral scholarship.
’ REFERENCES
(1) Hoffmann, M. R.; Martin, S. T.; Choi, W.; Bahnemann, D. W.
Chem. Rev. 1995, 95, 69–96.
(2) Litter, M. I. Appl. Catal., B 1999, 23, 89–114.
(3) Bahnemann, D. W.; Henglein, A.; Lilie, J.; Spanhel, L. J. Phys.
Chem. 1984, 88, 709–711.
(4) O'Regan, B.; Gr€atzel, M.; Fitzmaurice, D. Chem. Phys. Lett. 1991,
183, 89–93.
(5) Boschloo, G.; Fitzmaurice, D. J. Phys. Chem. B 1999, 103,
2228–2231.
(6) Gr€atzel, M.; Frank, A. J. J. Phys. Chem. 1982, 86, 2964–2967.
(7) Bahnemann, D. W.; Hilgendorff, M.; Memming, R. J. Phys. Chem.
B 1997, 101, 4265–4275.
(8) Gao, R.; Safrany, A.; Rabani, J. Radiat. Phys. Chem. 2003,
67, 25–39.
(9) Kormann, C.; Bahnemann, D. W.; Hoffmann, M. R. J. Phys.
Chem. 1988, 92, 5196–5201.
(10) Rajeshwar, K. J. Appl. Electrochem. 1995, 25, 1067–1082.
(11) Serpone, N.; Terzian, R.; Lawless, D.; Kennepohl, P.; Sauve, G.
J. Photochem. Photobiol., A 1993, 73, 11–16.
(12) Wang, C.; Bahnemann, D. W.; Dohrmann, J. K. Water Sci.
Technol. 2001, 44, 279–286.
(13) Safrany, A.; Gao, R.; Rabani, J. J. Phys. Chem. B 2000,
104, 5848–5853.
(14) K€olle, U.; Moser, J.; Gr€atzel, M. Inorg. Chem. 1985,
24, 2253–2258.
(15) Shiraishi, F.; Nakasako, T.; Hua, Z. J. Phys. Chem. A 2003,
107, 11072–11081.
(16) Stanbury, D. M. Adv. Inorg. Chem. 1989, 33, 69–138.
(17) Ranjit, K. T.; Viswanathan, B. J. Photochem. Photobiol., A 1997,
108, 73–78.
(18) Ranjit, K. T.; Viswanathan, B.; Varadarajan, T. K. J. Mater. Sci.
Lett. 1996, 15, 874–877.
(19) Ranjit, K. T.; Varadarajan, T. K.; Viswanathan, B. J. Phys. Chem.
A 1995, 89, 67–68.
(20) Zhang, F.; Pi, Y.; Cui, J.; Yang, Y.; Zhang, X; Guan, N. J. Phys.
Chem. C 2007, 111, 3756–3761.
(21) Halmann, M.; Zuckerman, K. J. Chem. Soc., Chem. Commun.
1986, 455–462.
(22) Vicente, F.; Garcia- Jareno, J. J.; Tamarit, R.; Cervilla, A.;
Domenech, A. Electrochem. Acta 1995, 40, 1121–1126.
(23) Douglas, B.; McDanial, D. H.; Alexander, J. J. Concepts and
Model of Inorganic Chemistry, 2nd ed.; Wiley: New York, 1983.
(24) (a) Shriver, D. F.; Atkins, P. W.; Langford, C. H. Inorganic
Chemistry; Oxford University Press: New York, 1990. (b) Bard, A. J.;
Parsons, R.; Jordan, J. Standard Potential in Aqueous Solution; Dekker:
New York, 1985.
(25) Booth, G.; McDuel, B.; Sears, J. World of Science; Oxford
University Press: New York, 1999.
(26) Langea, B.; Vejd^elek, Z. J. Photometrische Analyse; Verlag
Chemie: Weinheim, 1980.
(27) Gr€atzel, M. Nature 2001, 414, 338–344.
(28) Salomons,W., F€orstner, U., Mader, P., Eds. Heavy Metals,
Problems and Solutions; Springer: Berlin, 1995; p 386.
(29) Reiche, H.; Dunn, W. W.; Bard, A. J. J. Phys. Chem. 1979,
83, 2248–2251.
(30) Ward, M. D.; Bard, A. J. J. Phys. Chem. 1982, 86, 3599–3805.
2147
dx.doi.org/10.1021/jp108958w |J. Phys. Chem. A 2011, 115, 2139–2147