Probing the role of surface energetics of electrons and their accumulation in photoreduction processes on TiO2

Article abstract of DOI:10.1002/chem.201402039

We address the role of the energetics of photogenerated electrons in the reduction of 4-nitrobenzaldehyde on TiO₂. This model molecule bears two functional groups featuring different reducibilities. Electrochemistry shows that reduction to 4-aminobenzyl alcohol occurs in entirely distinct potential ranges. Partial reduction of the -NO₂ group, affording 4-aminobenzaldehyde, takes place through surface states at potentials positive of the flatband potential (E_fb). Dark currents caused by reduction of the aldehyde group are observed only at potentials more negative than E _fb, and the process requires an electron accumulation regime. Photocatalysis with TiO₂ suspensions agrees with the electrochemical data. In particular, reduction of the nitro group is a relatively fast process (k=0.059 s^-1), whereas that of the aldehyde group is slower (k=0.001 s^-1) and requires electron photoaccumulation. Control of the photogenerated charge is a prospective means for achieving chemoselective reductions. Electron energetics: Parallel investigations on TiO₂ electrodes and illuminated suspensions show that partial reduction of O ₂NC₆H₄CHO to H₂NC₆H ₄CHO is mediated by intra-bandgap states (see figure). When this process is complete, a second reduction affording H₂NC ₆H₄CH₂OH occurs under an electron accumulation regime.

Full text of DOI:10.1002/chem.201402039

DOI: 10.1002/chem.201402039
Full Paper
&
Photocatalysis
Probing the Role of Surface Energetics of Electrons and their
Accumulation in Photoreduction Processes on TiO2
[a]
[b]
[c]
Alessandra Molinari,* Andrea Maldotti, and Rossano Amadelli*
Abstract: We address the role of the energetics of photo-
are observed only at potentials more negative than E , and
fb
generated electrons in the reduction of 4-nitrobenzaldehyde
the process requires an electron accumulation regime. Pho-
tocatalysis with TiO2 suspensions agrees with the electro-
chemical data. In particular, reduction of the nitro group is
on TiO . This model molecule bears two functional groups
2
featuring different reducibilities. Electrochemistry shows that
reduction to 4-aminobenzyl alcohol occurs in entirely dis-
ꢀ
1
a relatively fast process (k=0.059 s ), whereas that of the
ꢀ
1
tinct potential ranges. Partial reduction of the ꢀNO group,
aldehyde group is slower (k=0.001 s ) and requires elec-
tron photoaccumulation. Control of the photogenerated
charge is a prospective means for achieving chemoselective
reductions.
2
affording 4-aminobenzaldehyde, takes place through surface
states at potentials positive of the flatband potential (Efb).
Dark currents caused by reduction of the aldehyde group
Introduction
Electron accumulation represents a way of promoting con-
trol over band-edge positions (interface energetics). In this
regard, dark electrochemistry in the accumulation regime has
been the subject of several publications, focusing mainly on
Heterogeneous photocatalysis with titanium dioxide is an im-
portant research field because of its attractive applications in
[
1]
[9]
environment remediation, water splitting and hydrogen pro-
energy storage. The phenomenon has been characterized by
[
2]
[3]
[7]
duction, solar energy conversion, and the synthesis of
EPR spectroscopy, and, in detail, by spectroelectrochemistry
[
4]
[10]
chemical specialties. Most investigations on TiO -based pho-
and luminescence. Accumulation accompanied by the inter-
2
tocatalysis adopt an application-driven perspective, and in this
context, surface modification and doping represent important
calation of protons or cations is referred to as “electrochemical
[11]
doping”. In addition to energy storage, prospective addition-
[
5]
[12]
strategies. The storage of photogenerated electrons in semi-
al developments of electrochemical doping include a novel
[
6]
[13]
conductors modified with metal islands or in hybrid photo-
switching mechanism for sensor development. It has been
[
7]
catalysts has also long been employed for catalyzing impor-
tant processes.
pointed out that reductive doping of TiO2 is also achieved
upon illumination, and is accordingly referred to as “photodop-
[14]
A number of authors have approached the field by investi-
gating the relationships between fundamental aspects of
photon-initiated events and observed photocatalytic phenom-
ing”. Applications of anoxic photochemical electron accumu-
[15]
lation are seemingly sparser
than for the analogous dark
electrochemical doping. In particular, Kohtani et al. have very
recently reported the photohydrogenation of acetophenone
[
8]
ena, recognizing, for example, that the usefulness of photo-
catalysis depends critically on the energetic positions of the
band edges relative to the solution species. Understanding
and optimizing the semiconductor surface energetics may lead
to significant improvements in power conversion efficiency in
dye-sensitized cells and photocatalytic processes.
and its derivatives by illuminated TiO , confirming that the role
2
[15c]
of electrons is an interesting research topic.
In this paper, we examine the influence of surface energet-
ics, paying particular attention to the effect of photochemical
electron accumulation, in the multistep photoreduction of 4-ni-
trobenzaldehyde (O NC H CHO) on TiO . This compound has
2
6
4
2
[
a] Dr. A. Molinari
been chosen as a model molecule containing two functionali-
ties with different reducibilities. Several literature studies have
Dipartimento di Scienze Chimiche e Farmaceutiche
Via Fossato di Mortara 17, 44121 Ferrara (Italy)
Fax: (+39)0532240709
[16]
been devoted to the photoreduction of either nitroaromatics
[16e,17]
or aromatic aldehydes
on TiO , but seemingly none exam-
2
E-mail: alessandra.molinari@unife.it
ine specifically the factors governing the photocatalytic reduc-
tion of the two functional groups present simultaneously in
the same molecule. From a general point of view, we think this
is an interesting issue of direct concern in studies on the selec-
tivity of photocatalytic processes. Furthermore, selective reduc-
tion of this type of substrate by thermal catalysis is difficult
[
b] Prof. A. Maldotti
Dipartimento di Scienze Chimiche e Farmaceutiche
Via Fossato di Mortara 17, 44121 Ferrara (Italy)
[
c] Dr. R. Amadelli
ISOF CNR, UOS di Ferrara
c/o Dipartimento di Scienze Chimiche e Farmaceutiche
Via Fossato di Mortara 17, 44121 Ferrara (Italy)
E-mail: rossano.amadelli@unife.it
[18]
and not always successful, as often demonstrated.
Chem. Eur. J. 2014, 20, 7759 – 7765
7759
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
From results of electrochemical measurements we confirm
that “accumulation” is a broad term that refers generally to the
increase in the concentration of conduction-band electrons fol-
lowing the saturation of intra-bandgap trap states. In this con-
text, their degree of filling with electrons has a close relation-
ship with the main topic of this paper: until there are free
states, photoreactions are channeled toward the chemoselec-
tive reduction of O NC H CHO to H NC H CHO; then, their
2
6
4
2
6
4
complete filling will cause the accumulation of electrons in the
conduction band, eventually leading to complete reduction to
H NC H CH OH. On the basis of the results presented and dis-
2
6
4
2
cussed, we underline the prospective developments in the
field of photocatalysis.
Figure 1. Mott–Schottky plots for a TiO
dark and under illumination (l>360 nm, 100 mWcm ). Electrolyte: CH CN/
3
2
(P-25) film electrode both in the
ꢀ
2
2
0
-propanol (4:1, v/v) containing tetraethylammonium perchlorate (TEAP,
.2m).
Results and Discussion
In the following, we examine the behavior of TiO electrodes
2
3þ
þ
4þ
ꢀ
þ
ads
and suspensions. In all cases, the working medium is a mixture
of acetonitrile and 2-propanol containing O NC H CHO. We
Ti O ðH Þ ! Ti O þ e þ ðH Þ
ð1Þ
2
i
2
2
6
4
first present the results of electrochemical measurements in
the dark, with and without prior illumination on the electrode
at an open circuit. Then, we discuss the photocatalytic behav-
In the present case, protons released from the oxidation of 2-
propanol are first adsorbed and then intercalated as the nega-
tive field caused by electron accumulation increases. In any
case, as expected, no capacity change is observed in the pres-
ior in both the absence and presence of O . More specifically,
2
we evaluate the possibility of channeling the interface photo-
reactions toward the chemoselective reduction of
O NC H CHO. Moreover, ESR measurements provide informa-
ence of O NC H CHO, because its reduction consumes both
2 6 4
electrons and protons.
Generally, the role played by surface states (ss) is an impor-
tant feature of photo(electro)catalytic processes. We sought in-
sights into this aspect and on the effects of illumination by
using a variant of the potentiostatic method described by
2
6
4
tion on the nature of the radical intermediates, which is impor-
tant for the formulation of the reaction mechanism.
[23]
Wang et al. We recorded anodic linear sweep voltammetry
LSV) curves in the dark from a potential slightly more positive
(
Electrochemical characterization
than Efb to 0.0 V, immediately (2 s) after illumination for differ-
ent lengths of time at an open circuit. Curves of charge versus
potential are then obtained by integrating the LSV curves.
These are recorded as a function of potential, and can be con-
Electrochemistry is an important tool for studying physical and
[
19]
chemical phenomena occurring at interfaces. Moreover, this
technique can be conveniently employed for investigating sep-
arately the reduction and oxidation processes that occur simul-
taneously during photocatalytic experiments as a consequence
ꢀ
1
verted to a timescale knowing the scan rate (100 mVs ). In
the absence of electron acceptors, the re-oxidation charge is
[
20]
[24]
of primary charge separation.
equivalent to the total accumulated charge in the form of
3
+
Among several parameters, the flat band potential is a very
important one for the characterization of a semiconductor/
electrolyte interface as it correlates the band edges to the
redox potential in the electrolyte. The value of the flatband po-
Ti . Illumination causes a conspicuous accumulation of charge
(Figure 2), which is rather persistent after the light is switched
off. Typically, after 10 min illumination, the cyclic voltammetry
profile returns to its original value after cycling the electrode
in the dark from ꢀ1.85 to 0.5 V for 140 s.
tential (E ), from Mott–Schottky plots obtained from capaci-
fb
tance measurements in CH CN/TEAP, was in the range ꢀ2.0 to
Plots of charge versus potential (Figure 2) are expected to
provide information on the presence and distribution of ss
3
ꢀ
2.1 V versus SCE, in good agreement with that obtained pre-
[23]
viously through spectroscopic techniques for a polycrystalline
TiO electrode in neat CH CN containing the same supporting
across the bandgap. The potential variation of the charge as-
sociated with ss shows that they are widely distributed ener-
getically. Their filling reaches saturation (curve c) under contin-
ued illumination, that is, at E<ꢀ1.6 V, and the bands tend to
become flat, whereas at potentials above ꢀ0.7 V there are no
trapped electrons. For prolonged illumination, a shoulder ap-
pears (not shown) in the plots of Q versus E in the range ꢀ1
2
3
[
21]
electrolyte.
In the mixed solvent, it decreased slightly to
a value between ꢀ2.0 and ꢀ1.95 V (Figure 1). It is seen in the
figure that illumination does not change E , whereas the slope
fb
decreases significantly. This has frequently been observed in
the literature, and is attributed to an increase in the number of
donors or, more probably, to the intercalation of protons,
which compensates charge accumulation and leads to changes
in the surface characteristics, as shown in Equation (1), in
[
12]
to ꢀ0.8 V, which can be attributed to the filling of deeper ss.
At intermediate potentials, the charge changes rapidly, and the
derivatives of the curves (dQ/dE) show a peak (bottom of
+
[14,22]
[23]
which (H ) refers to intercalated protons.
Figure 2) corresponding to a maximum density of ss. The
i
Chem. Eur. J. 2014, 20, 7759 – 7765
www.chemeurj.org
7760
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 2. Top: Plots of the charge accumulated during illumination of a TiO
electrode at open circuit (l>360 nm, 100 mWcm ). The charge was ob-
2
ꢀ
2
Figure 3. Top: Steady-state current versus potential curves for the reduction
tained by integration of linear sweep voltammetry curves recorded from
of 1.1 mmol O
TiO in the dark. Curve 3 shows the reduction of 1.1 mmol H
a gold electrode. Electrolyte: CH CN/2-propanol (4:1, v/v) containing TEAP
(0.2m). Bottom: Partial currents for the photo-oxidation of 2-propanol and
for the dark reduction of O NC CHO (see text).
2
NC
6
H
4
CHO (curve 1) and 1.1 mmol H
2
NC
6
H
4
CHO (curve 2) on
ꢀ
1.85 to 0.0 V versus SCE following illumination: a) dark, b) 4 min irradiation,
and c) 10 min irradiation. Bottom: dQ/dE versus potential. Electrolyte:
CH CN/2-propanol (4:1, v/v) containing TEAP (0.2m).
2
2
NC CHO on
6 4
H
3
3
2
6 4
H
maximum is located at ꢀ1.6 V in the dark and shifts to ꢀ1.5 V
under illumination, probably on account of protons formed
during the photo-oxidation of 2-propanol. Indeed, there is no
positive shift of the maximum if potential is used to accumu-
late electrons in the dark.
ing the reduction of the aldehyde functional group at poten-
tials more positive than ꢀ2.0 V.
Considering, then, that H NC H CHO is a possible stable re-
2
6
4
duction intermediate, we studied its electrochemical reduction
on both Au and TiO2 electrodes. The results reported in
Figure 3 (curves 2, 3) show that the process starts at E<ꢀ2.0 V
Upon complete filling of the ss, electrons accumulate in the
conduction band. From Figure 2c, saturation of the ss corre-
sponds to a charge of 7.7 mC, and using a roughness factor of
on both electrodes, that is, in a potential range in which TiO
2
[
25]
13
ꢀ2
6
90, we calculate a trap density (N =Q/eA) of 8.7ꢁ10 cm ,
behaves as a metal. In other words, electron accumulation is
t
[23]
in agreement with the values reported by Wang et al. and
seemingly necessary for the reduction of ꢀCHO to ꢀCH OH.
2
[
26]
Kay et al. In principle, this density of states may cause a po-
From a comparison of the data of Figures 2 and 3, it appears
tential drop in the Helmholtz layer (DV =eN /C ) and unpin-
that ss can mediate the reduction of the ꢀNO group. In this
H
t
H
2
[
10,24]
ning of the band edges.
The steady-state reduction of O NC H CHO in the dark is
case, ss continue to trap electrons until complete reduction of
ꢀNO to NH has occurred, as proposed, for example, by Red-
2
6
4
2
2
[9a]
+
shown in Figure 3 (curve 1). Reduction starts at E<ꢀ1.1 V and
mond et al. for H reduction through surface states. Subse-
quently, electron accumulation takes place as shown by the ex-
perimental data of Figure 4. In this figure, in the presence of
added O NC H CHO (curve 1), the open-circuit photopotential
leads to H NC H CHO, as established by chemical analysis. For
2
6
4
comparison purposes, we also examined the reduction of
O NC H CHO on a gold electrode (not shown), and observed
2
6
4
2
6
4
two reduction peaks at about a ꢀ1.0 and ꢀ1.3 V, which are at-
first decreases to a negative value (a), and then increases. After
reaching a constant value (b), it becomes more negative again
(c) owing to electron accumulation. The results suggest that
the nitro group should be reduced completely before ꢀCHO
can, in turn, be reduced. No such positive drift of the photopo-
tential is seen for the same concentration of added
H NC H CHO and on the same timescale (curve 2).
tributed to consecutive reductions of the ꢀNO group. The first
2
reduction process is reversible, whereas the second is irrevers-
2ꢀ
ible, indicating that the NO2 species can be protonated by
[
27]
the alcohol, as reported previously by Bard. Both redox pro-
cesses are reversible in neat CH CN under otherwise identical
3
conditions. There is no evidence of any further process involv-
2
6
4
Chem. Eur. J. 2014, 20, 7759 – 7765
www.chemeurj.org
7761
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 4. Open-circuit potential versus time for a TiO
propanol (4:1, v/v) containing TEAP (0.2m). Curve 1: with added
NC CHO (0.4 mmol). Curve 2: in the presence of H NC CHO
0.4 mmol). Illumination at l>360 nm, 100 mWcm
2 3
electrode in CH CN/2-
Figure 5. Spectral changes obtained upon irradiation (l>350 nm, 25ꢁ18C)
ꢀ
1
2
of deaerated suspensions of TiO -P25 (3.5 gL ) in a 2-propanol/acetonitrile
O
(
2
6
H
4
2
6 4
H
ꢀ
5
ꢀ
2
2 6 4
mixture (1:4, v/v) containing O NC H CHO (7.8ꢁ10 m).
.
Figure 3 (bottom) shows that photo-oxidation of 2-propanol
is also likely to be influenced by ss. Their signature is evident
in the slow increase in photocurrent in the potential range
ꢀ
1.5 to ꢀ0.8 V, in the region where accumulated charge varies
significantly with potential (Figure 2). The stability of electrons
trapped in the surface states is probably responsible for the
slow photocurrent response in this potential interval. At more
positive potentials at which electron traps are empty, the cur-
rent increases markedly. More results on the photo-oxidation
of 2-propanol will be discussed in the next section.
In relation to the results displayed in Figure 3 (bottom), we
note further that the photocurrent and dark reduction of
O NC H CHO are superimposable and become equal at ap-
2
6
4
proximately ꢀ1.25 V. We can then infer that, according to the
[
28]
flux matching theory, the reduction of ꢀNO can take place
2
on electrodes at an open circuit and in TiO suspensions in the
2
same medium. In contrast, the reduction of the ꢀCHO func-
tional group is not predicted under normal photocatalytic con-
ditions.
Surface energetics and photocatalysis
Irradiation of deaerated suspensions of TiO -P25 in the same
2
mixture of CH CN/2-propanol containing O NC H CHO brings
3
2
6
4
about significant UV/Vis spectral changes, as illustrated by the
photoreduction course reported in Figure 5. This shows quali-
tatively that photoexcitation causes a rapid conversion of
O NC H CHO (l =264 nm) to H NC H CHO, as indicated by
2
6
4
max
2
6
4
the growth of a spectrum with two absorption maxima at 232
and 312 nm. These results are in agreement with the electro-
chemical data of Figures 3 and 4, from which we saw that re-
duction of the nitro group takes place at potentials more posi-
Figure 6. Time courses of the amounts of O
2
NC
6
H
4
CHO (*), H
2
NC
6
H
4
CHO (~),
-P25
and H
(
2
NC
6
H
4
CH
2
OH (^) during irradiation (l>350 nm, 25ꢁ18C) of TiO
2
ꢀ1
3.5 gL ) suspended in a 2-propanol/acetonitrile mixture (1:4, v/v). Initial
ꢀ5
concentration of O
2 6 4
NC H CHO: 7.8ꢁ10 m. Top: first 90 s; Bottom: up to
3000 s.
tive than E , and as anticipated (vide supra), occurs under
fb
normal photocatalytic conditions. On the other hand, it also
appears that reduction of the CHO group to give
H NC H CH OH (l =243 nm) does take place, and is a com-
experimental conditions, the nitro compound is consumed
almost completely after photoexcitation for 50 s. The decay
curve of O NC H CHO is characteristic of first-order kinetics,
2
6
4
2
max
paratively much slower process.
Figure 6 (top) shows the formation of H NC H CHO from
2
6
4
2
6
4
ꢀ
1
O NC H CHO quantitatively as a function of time. Under our
and the corresponding apparent rate constant is 0.059 s . At
2
6
4
Chem. Eur. J. 2014, 20, 7759 – 7765
www.chemeurj.org
7762 ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
the same time, the only photoreduction product is accumulat-
ed in the liquid phase in more than 90% yield.
protons. Pathways involving holes generally entail the forma-
tion of radicals that may, in turn, play an important role as re-
ducing species in the overall photocatalytic process. Reasoning
that the nature of radical intermediates is also important for
the formulation of a reaction mechanism, we performed EPR
measurements using the spin-trapping technique, which is
a particularly potent tool for the detection of short-lived radi-
These relatively high kinetics seem to imply an interaction of
the substrate with the surface. FTIR measurements confirm
that O NC H CHO interacts with the surface through the NO
2
2
6
4
group, as evidenced from the shift of the asymmetric vibration
ꢀ
1
frequency of the NꢀO bond from 1541 to 1529 cm , in ac-
[30]
cordance with previous data on nitrobenzene adsorption on
cal species.
Irradiation of powder dispersions of TiO -P25 suspended in
[
29]
TiO2.
2
The plot shown in Figure 6 (bottom) indicates that the
amount of H NC H CHO reaches a maximum value correspond-
a
deaerated
2-propanol/CH CN
mixture
containing
3
O NC H CHO and the spin trap a-phenyl N-tert-butyl nitrone
2
6
4
2
6
4
ing to 90% conversion of the starting compound; it is then re-
duced slowly to produce H NC H CH OH. The apparent first-
(PBN) leads to the detection of a triplet of doublets with hy-
perfine splitting constants a =13.9 G and a =2.2 G, which
2
6
4
2
N
H
order rate constant for the disappearance of H NC H CHO is
can be attributed to the paramagnetic adduct [PBN–
2
6
4
ꢀ
1
[31]
0.001 s , and the conversion yield to the alcohol is higher
(CH ) CHO]C (Figure 8). Then, the signal is derived from the
3
2
than 80% with respect to the starting nitro compound after
50 min photoexcitation.
On the basis of the electrochemical data of the previous sec-
tion, the reduction of the aldehyde group occurs only if elec-
trons can accumulate. This accumulation can take place after
the complete transformation of the nitro group (cf. Figure 4),
that is, H NC H CH OH formation is observed only after
2
6
4
2
O NC H CHO has been converted completely to H NC H CHO.
2
6
4
2
6
4
It has been reported previously that photocatalytic reduc-
tion of nitrobenzenes at TiO occurs even under aerobic condi-
2
[
16f]
tions.
On this basis, and in line with the above statements,
O can possibly be used as a competitive electron scavenger
2
to prevent charge accumulation on the semiconductor surface,
and consequently, to control the selectivity of the photocata-
lytic process. Indeed, if the photocatalytic experiment is carried
out under aerobic conditions, O NC H CHO reduction does not
Figure 8. EPR spin-trapping spectra obtained upon irradiation (10 s,
l>350 nm) of TiO -P25 suspended in deaerated 2-propanol/CH CN (1:4, v/v)
solutions containing PBN (5ꢁ10 m) and O
2
6
4
2
3
proceed beyond H NC H CHO, which can be accumulated in
2
6
4
ꢀ2
ꢀ4
2
6 4
NC H CHO (1ꢁ10 m).
the liquid phase in more than 90% yield (Figure 7). The appar-
ent rate constant of O NC H CHO reduction decreases to
2
6
4
ꢀ
1
0
.027 s because of the electron-scavenging ability of O2.
trapping of isopropoxy radicals (CH ) CHOC, formed by the re-
3 2
In the overall photocatalytic process, 2-propanol can work as
action of adsorbed 2-propanol with photogenerated holes as
a hole scavenger that can limit charge recombination, free
a large number of reductive equivalents, and act as a source of
shown by Eqs. (2) and (3).
þ
þ
ðCH Þ CHOH þ h ! ðCH Þ CHꢀOC þ H
ð2Þ
3
2
ads
3 2
The oxidation potential is a measure of the reducing power
of radicals. In the case of 2-propanol, to the best of our knowl-
edge, the oxidation potential of alkoxy radicals has not been
reported. That of hydroxyalkyl radicals (CH ) CCOH in CH CN/2-
3
2
3
[32]
propanol falls in the range ꢀ0.61 to ꢀ1.0 V versus SCE. Even
assuming that the oxidation potential of (CH ) CHOC radicals is
3
2
more negative, it is very unlikely that the difference is as much
as 1 V. From this reasoning, we infer that detected (CH ) CHOC
3
2
Figure 7. Time courses of the amounts of O
~) during irradiation (l>350 nm, 25ꢁ18C) of aerated suspensions of TiO
P25 (3.5 gL ) in a 2-propanol/acetonitrile mixture (1:4, v/v). Initial concen-
2 6 4 2 6 4
NC H NC H CHO
CHO (*) and H
radicals may only be able to inject electrons into deeper intra-
bandgap surface states, and not directly into the conduction
band. The photoreduction of the nitro group to an amino
(
2
-
ꢀ
1
ꢀ
5
tration of O
2 6 4
NC H CHO: 7.8ꢁ10 m.
Chem. Eur. J. 2014, 20, 7759 – 7765
www.chemeurj.org
7763
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
other hand, experimental conditions that favor electron accu-
mulation allow operation on molecules that are difficult to
reduce, which is also an important aspect. Future work will
consider the reduction of various structurally different multi-
functional organic species.
Experimental Section
TiO Degussa P25 (TiO -P25) was used throughout this work, be-
2
2
cause this commercial photocatalyst is well characterized and has
become a benchmark for comparative photocatalytic studies. Sol-
vents and reagents were purchased from Sigma and used without
further purification.
Titanium dioxide electrodes were prepared by spreading, on a tita-
Figure 9. Schematic energy diagrams illustrating the reduction of
NC group) or through accumulated
CHO through surface states (ꢀNO
electrons (ꢀCHO group).
nium foil, a paste obtained by mixing TiO (Degussa P-25, 3 g), bi-
2
O
2
6
H
4
2
distilled water (6 mL), acetylacetone (0.2 mL), and Triton X-100
(
0.2 mL), with subsequent calcination at 4508C for 30 min. The
2
electrodes had a geometric area of 1.3 cm .
group is not a limiting factor, because trapped electrons are
able to reduce O NC H CHO, as discussed earlier. Figure 9 is an
Cyclic voltammetry curves were obtained with an EG&G potentio-
stat using EG&G software. All measurements were made in deaer-
ated CH CN or 2-propanol/CH CN (1:4 v/v) using tetra-ethylammo-
nium perchlorate (TEAP, 0.2m) as supporting electrolyte. A conven-
tional three-compartment glass cell was employed, in which the
2
6
4
3
3
illustrative scheme that summarizes the findings in an energy
scale before and after electron accumulation.
For prospective applications in synthetic photocatalysis, it is
noteworthy that the photocatalyst is completely recyclable
after washing, drying in an oven, and calcination at 4008C for
working electrode was a Ti/TiO electrode (or Au sheet in some ex-
2
periments). O NC H CHO or H NC H CHO were added as necessary
2
6
4
2
6
4
for the experiments. Glassy carbon and Ag wire electrodes served
as the counter and reference electrodes, respectively. The potential
of the latter was found to be 0.025 V versus SCE through compari-
son of cycling voltammograms of ferrocene as an internal stan-
dard. All potentials are given versus SCE.
6
h. The absence of oxygen probably prevents the formation
of carboxylates that are able to poison the surface of the semi-
[
33]
conductor in aerated conditions. In addition, we have experi-
mental evidence that for initial concentrations of O NC H CHO
2
6
4
ꢀ
2
Mott–Schottky analysis was performed by cyclic voltammetry using
as high as 10 m, the process proceeds as described
ꢀ1
the relationship C=i/n, in which n is the scan rate (100 mVs ). Ex-
periments were performed in the potential range 0.0 to 1.0 V, in
which pure capacitive behavior is observed.
above, first accumulating H NC H CHO, and then yielding
2
6
4
H NC H CH OH as the final reduction product. From a quantita-
2
6
4
2
tive point of view, after a long irradiation time (about 20 h) the
photocatalytic yield of H NC H CH OH is over 80%. This result
Infrared spectra were obtained with a Nicolet 510P FTIR instrument
in KBr, fitted with a Spectra-Tech collector diffuse reflectance acces-
2
6
4
2
is encouraging, and opens up new synthetic routes to be veri-
fied with the use of other substrates.
ꢀ1
sory (range 4000 to 200 cm ). The sample was prepared using an
aliquot of TiO -P25, which was put in contact with a CH CN solu-
2
3
ꢀ4
tion of O NC H CHO (1ꢁ10 m). The suspension was stirred at
2
6
4
room temperature until the evaporation of the solvent was com-
plete. Then the powder impregnated with 4-nitrobenzaldehyde
was dried overnight in an oven.
Conclusion
We have examined the effect of the photoaccumulation of
ꢀ1
In a typical photocatalytic experiment, TiO -P25 (3.5 gL ) was sus-
electrons on TiO and its role within the context of the partial
2
2
pended in a mixture of 2-propanol and acetonitrile (1:4, v/v) con-
and total reduction of O NC H CHO, which was chosen as
2
6
4
ꢀ5
taining O NC H CHO (7.8ꢁ10 m). Then, the suspension was
2
6
4
a model organic compound bearing two reducible functional
purged with nitrogen for 20 min and subsequently irradiated at
groups. We have reported parallel investigations on TiO elec-
ꢀ2
2
wavelengths above 350 nm (25ꢁ18C, 15 mWcm ). The reaction
trodes and suspensions; the results are in agreement, and
course was followed by recording UV spectra of the irradiated solu-
tions and comparing them with UV spectra of pure substances.
The disappearance of O NC H CHO was evaluated on the basis of
show that partial reduction to H NC H CHO is probably medi-
2
6
4
ated by intra-bandgap states, and must be nearly complete
before the reduction of the aldehyde group occurs, affording
H NC H CH OH. The latter process occurs at potentials more
2
6
4
the absorbance decrease at 264 nm, and the formation of
H NC H CHO and H NC H CH OH was calculated from the absorb-
2
6
4
2
6
4
2
2
6
4
2
ance increases at 312 and 243 nm, respectively.
negative than the flatband potential, and then requires an
electron accumulation regime.
We also verified that direct light absorption by O NC H CHO was
2
6
4
negligible under the experimental conditions employed (l>
50 nm) and that no reaction occurred when photoexcitation was
We have demonstrated that control of the photogenerated
charge is a suitable means for selective reduction of a model
organic molecule, and we think that this is an interesting issue
of direct concern for studies on the selectivity of photocatalytic
processes. It is well known that selective reduction through
the use of chemical species is quite difficult to achieve. On the
3
carried out in the absence of the semiconductor. Other control ex-
periments indicated that no reduction products were formed
when the solution of O NC H CHO was kept in contact with the
2
6
4
semiconductor in the dark. The optimal amount of TiO -P25
2
ꢀ1
(3.5 gL ) used throughout this work was chosen on the basis of
Chem. Eur. J. 2014, 20, 7759 – 7765
www.chemeurj.org
7764 ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
measurements performed with a radiometer, which indicated that
this amount of semiconductor was able to absorb more than 90%
of the impinging radiation.
B 1999, 103, 7860–7868; d) G. Boschloo, D. Fitzmaurice, J. Electrochem.
Soc. 2000, 147, 1117–1123; e) H. G. Agrell, G. Boschloo, A. Hagfeldt, J.
Phys. Chem. B 2004, 108, 12388–12396.
[
[
10] T. S. Kang, D. Kim, K. J. Kim, J. Electrochem. Soc. 1998, 145, 1982–1986.
11] J. Idꢅgoras, T. Berger, J. A. Anta, J. Phys. Chem. C 2013, 117, 1561–1570.
[12] T. Berger, T. Lana-Villarreal, D. Monllor-Satoca, R. Gꢆmez, Electrochem.
Commun. 2006, 8, 1713–1718.
For recycling experiments, TiO -P25 used in a first photocatalytic
2
experiment was washed twice with aliquots of CH CN and then
3
with CH Cl . The photocatalyst was dried overnight in the oven,
2
2
and then calcined for 6 h at 4008C. After this treatment, it was
reused in a second photocatalytic experiment, performed as de-
scribed above.
[13] E. V. Milsom, H. A. Dash, T. A. Jenkins, M. Opallo, F. Marken, Bioelectro-
chem. 2007, 70, 221–227.
[
14] B. I. Lemon, J. T. Hupp, J. Phys. Chem. 1996, 100, 14578–14580.
15] a) E. Kikuchi, H. Sakamoto, J. Electrochem. Soc. 2000, 147, 4589–4593;
b) I. K. Levy, M. A. Brusa, M. E. Aguirre, G. Custo, E. San Romꢄn, M. I.
Litter, M. A. Grela, Phys. Chem. Chem. Phys. 2013, 15, 10335–10338; c) S.
Kohtani, Y. Kamoi, E. Yoshioka, H. Miyabe, Cat. Sci. & Technol., DOI:
10.1039/c3cy00879g and references therein.
[
EPR spin-trapping experiments were performed with a Bruker ER
2
00 MRD spectrometer equipped with a TE 201 resonator, at a mi-
crowave frequency of 9.4 GHz. The samples were constituted by
deaerated suspensions of TiO -P25 (20 mg) in 2-propanol/CH CN
2
3
(
(
1:4, v/v) solution (1 mL) containing a-phenyl N-tert-butyl nitrone
[16] a) J. L. Ferry, W. H. Glaze, Langmuir 1998, 14, 3551–3555; b) S. O. Flores,
O. Rios-Bernij, M. A. Valenzuela, I. Cordova, R. Gomez, R. Gutierrez, Top.
Catal. 2007, 44, 507–511; c) H. Kominami, S. Iwaasaki, T. Maeda, K. Ima-
mura, K. Hashimoto, Y. Kera, B. Ohtani, Chem. Lett. 2009, 38, 410–411;
d) A. Maldotti, R. Amadelli, L. Samiolo, A. Molinari, A. Penoni, S. Tollari, S.
Cenini, Chem. Commun. 2005, 1749–1751; e) R. Amadelli, L. Samiolo, A.
Maldotti, A. Molinari, D. Gazzoli, Int. J. Photoenergy 2011, 259453; f) K.
Imamura, S. Iwasaki, T. Maeda, K. Hashimoto, B. Ohtani, H. Kominami,
Phys. Chem. Chem. Phys. 2011, 13, 5114–5119; g) Y. Shiraishi, Y. Togawa,
D. Tsukamoto, S. Tanaka, T. Hirai, ACS Catal. 2012, 2, 2475–2481; h) A.
Hakki, R. Dillert, D. Bahnemann, Catal. Today 2009, 144, 154–159; i) V.
Brezovꢄ, A. Blazkova, I. Surina, B. Havlinova, J. Photochem. Photobiol. A
ꢀ
2
ꢀ4
PBN, 5ꢁ10 m) as the spin trap and O NC H CHO (1ꢁ10 m). The
2
6
4
obtained samples were put into a flat quartz cell and irradiated
l>350 nm) directly in the EPR cavity. Deaeration was carried out
(
by fluxing N and transferring the samples into the cell under a ni-
2
trogen-saturated atmosphere. Irradiations were performed with
a Helios Q400 Italquartz medium-pressure Hg lamp using a cutoff
filter (l>350 nm). No signals were obtained in the dark or during
irradiation of the solution in the absence of TiO2.
Keywords: electrochemistry
·
electron
energetics
·
1
997, 107, 233–237.
photocatalysis · reduction · titanium dioxide
[17] a) C. Joyce-Pruden, J. K. Pross, Y. Li, J. Org. Chem. 1992, 57, 5087–5091;
b) S. Kohtani, E. Yoshiota, K. Saito, A. Kudo, H. Miyabe, Catal. Commun.
2
010, 11, 1049–1053; c) Y. Matsushita, S. Kumada, K. Wakabayashi, K.
[
1] a) B. Ohtani, Trends Photochem. Photobiol. 1994, 3, 531–539; b) A. Fu-
jishima, X. Zhang, D. A. Tryk, Surf. Sci. Rep. 2008, 63, 515–582; c) M. L.
Hoffmann, S. T. Martin, W. Choi, D. W. Bahnemann, Chem. Rev. 1995, 95,
Sakeda, T. Ichimura, Chem. Lett. 2006, 35, 410–411.
[
[
18] A. Corma, P. Serna, Science 2006, 313, 332–334.
19] S. R. Morrison, Electrochemistry at Semiconductors and Oxidized Metal
Electrodes, Plenum Publishing Corporation, New York, NY, 1980.
20] a) T. Berger, D. Monllor-Satoca, M. Jankulovska, T. Lana-Villarreal, R.
Gꢆmez, ChemPhysChem 2012, 13, 2824–2875; b) R. Amadelli, L. Samio-
lo, in Photocatalysis and Water Purification (Ed.: P. Pichat), Wiley-VCH,
Weinheim, 2013, pp. 241–270.
69–96.
[
2] a) N. Serpone, R. F. Khairutdinov, in Application of nanoparticles in the
photocatalytic degradation of water pollutants (Eds.: V. P. Kamat, D.
Meisel), Stud. Surf. Sci. Catal. 1997, 103, 417–444; b) A. Mills, R. H.
Davies, D. Worsley, Chem. Soc. Rev. 1993, 22, 417–425; c) A. Hagfeldt,
M. Gratzel, Acc. Chem. Res. 2000, 33, 269–277.
[
[
[
21] G. Redmond, D. Fitzmaurice, J. Phys. Chem. 1993, 97, 1426–1430.
22] V. Spagnol, E. Sutter, C. Debiemme-Chouvy, H. Cachet, B. Baroux, Elec-
trochim. Acta 2009, 54, 1228–1232.
[
3] a) J. Nowotny, C. C. Sorrell, T. Bak, L. R. Sheppard, Solar Energy 2005, 78,
5
1
93–602; b) M. Kitano, K. Tsujimaru, M. Anpo, Top. Catal. 2008, 49, 4–
7.
[
23] H. Wang, J. He, G. Boschloo, H. Lindstrçm, A. Hagfeldt, S. E. Lindquist, J.
Phys. Chem. B 2001, 105, 2529–2533.
[
4] a) W. R. Duncan, O. V. Prezhdo, Annual Review of Physical Chemistry
007, 58, 143–184; b) A. Hagfeldt, G. Boschloo, L. C. Sun, L. Kloo, H.
2
[
24] F. Cao, G. Oskam, P. Searson, J. M. Stipkala, T. A. Heimer, F. Farzad, G.
Meyer, J. Phys. Chem. B 1995, 99, 11974–11980.
25] R. Amadelli, A. Maldotti, L. Samiolo, Catal. Today 2009, 144, 149–153.
26] A. Kay, R. Humphrey-Baker, M. Grꢃtzel, J. Phys. Chem. 1994, 98, 952–
Pettersson, Chem. Rev. 2010, 110, 6595–6663; c) Y. Shiraishi, T. Hirai, J.
Photochem. Photobiol. C 2008, 9, 157–170; d) M. Fagnoni, D. Dondi, D.
Ravelli, A. Albini, Chem. Rev. 2007, 107, 2725–2756.
[
[
[
5] a) N. Shaham-Waldmann, Y. Paz, in Photocatalysis and Water Purification
959.
(
Ed.: P. Pichat), Wiley-VCH, Weinheim, 2013, pp. 103–143; b) Z. Wang, C.
[27] W. H. Smith, A. J. Bard, J. Am. Chem. Soc. 1975, 97, 5203–5210.
[28] K. Rajeshwar, J. G. Ibanez, J. Chem. Educ. 1995, 72, 1044–1049.
[29] M. Boronat, P. Concepcion, A. Corma, S. Gonzales, F. Illas, P. Serna, J. Am.
Chem. Soc. 2007, 129, 16230–16237.
Chen, W. Ma, J. Zhao, in Photocatalysis and Water Purification (Ed.: P.
Pichat), Wiley-VCH, Weinheim, 2013, pp. 199–240.
6] K. Rajeshwar, W. Chanmanee, Electrochim. Acta 2012, 84, 96–102.
7] D. Zhao, C. Chen, C. Yu, W. Ma, J. Zhao, J. Phys. Chem. C 2009, 113,
[
[
[
30] A. Molinari, A. Maldotti, A. Bratovcic, G. Magnacca, Catal. Today 2013,
13160–13165.
2
06, 46–52.
[
8] a) G. Palmisano, E. I. Garcia-Lopez, G. Marcꢂ, V. Loddo, S. Yurdakal, V.
Augugliaro, L. Palmisano, Chem. Commun. 2010, 46, 7074–7089; b) A.
Maldotti, A. Molinari, Top. Curr. Chem. 2011, 303, 185–216; c) M. A. Hen-
derson, Surf. Sci. Rep. 2011, 66, 185–297; d) T. L. Thompson, J. T. Yates,
Chem. Rev. 2006, 106, 4428–4453; e) B. Liu, K. Nakata, X. Zhao, T. Ochiai,
T. Murakami, A. Fujishima, J. Phys. Chem. C 2011, 115, 16037–16042;
f) B. Liu, X. Wang, L. Wen, X. Zhao, Chem. Eur. J. 2013, 19, 10751–10759.
9] a) G. Redmond, D. Fitzmaurice, M. Grꢃtzel, J. Phys. Chem. 1993, 97,
[
[
31] G. R. Buettner, Free Radical Biol. Med. 1987, 3, 259–303.
32] T. Lund, D. D. M. Wayner, M. Johsson, A. G. Larsen, K. Daasbjerg, J. Am.
Chem. Soc. 2001, 123, 12590–12595.
[
33] J. T. Carneiro, J. A. Mouljin, G. Mul, J. Catal. 2010, 273, 199–210.
[
Received: February 5, 2014
Revised: March 14, 2014
6951–6954; b) L. Kavan, K. Kratochvilovꢄ, M. Grꢃtzel, J. Electroanal.
Chem. 1995, 394, 93–102; c) G. Boschloo, D. Fitzmaurice, J. Phys. Chem.
Published online on May 14, 2014
Chem. Eur. J. 2014, 20, 7759 – 7765
www.chemeurj.org
7765
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim