Академический Документы
Профессиональный Документы
Культура Документы
Invited review
a r t i c l e i n f o a b s t r a c t
Article history: Topics, in alphabetical order from Activity, Band structure and Crystallinity to X-ray photoelectron
Received 22 November 2010 spectroscopy, Yield and Z-scheme photocatalysis, related to photocatalysis and photoelectrochem-
Received in revised form 1 February 2011 ical reaction are discussed with interpretation of what we know and what we do not know in a scientic
Accepted 3 February 2011
sense.
2011 Elsevier B.V. All rights reserved.
Keywords:
Photocatalysis
Photoelectrochemical reaction
Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
2. Topics in alphabetical order . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
2.1. Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
2.2. Band structure (see also Junction, Thermodynamics and Visible light) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
2.3. Crystallinity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
2.4. Doping (see also Visible light-induced photocatalytic activity) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
2.5. Energy conversion (see also Gibbs energy and Thermodynamics) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
2.6. First-order kinetics (see also LangmuirHinshelwood mechanism) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
2.7. Gibbs energy (see also Band structure, Energy conversion and Thermodynamics) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
2.8. Hydroxyl radical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
2.9. Identication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
2.10. Junction (see also Band structure). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
2.11. KubelkaMunk function (see also Optical band gap, Quantum efciency and Wavelength dependence) . . . . . . . . . . . . . . . . . . . . . . . . . . 164
2.12. LangmuirHinshelwood mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
2.13. Mutual recombination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
2.14. Nanostructure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
2.15. Optical band gap (see also Band structure and KubelkaMunk function) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
2.16. Particle size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
2.17. Quantum efciency (see also Mutual recombination and Wavelength dependence) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
2.18. Rate-determining step (see also First-order kinetics) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
2.19. Synergetic effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
2.20. Thermodynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
2.21. Ultraviolet light in solar radiation (see also Energy conversion and Visible-light induced photocatalysis) . . . . . . . . . . . . . . . . . . . . . . . . . . 172
2.22. Visible light-induced photocatalysis (see also Band structure and Doping). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
2.23. Wavelength dependence (see also, Optical density and Quantum efciency) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
2.24. X-ray photoelectron spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
2.25. Yield (see also Identication and Quantum efciency) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
2.26. Zscheme photocatalysis (see also Gibbs energy and Thermodynamics) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
3. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
Bunsho Ohtani started his research work on photocatal- on a catalyst accounts for the catalytic reaction. The reaction rate
ysis in 1981 when he was a Ph. D. course student in Kyoto per active site can be estimated and should be catalytic activity.
University. Since then he has been studying photocatalysis Sometimes the term turnover frequency, i.e., number of turnovers
and relating topics for more than 25 years and published
more than 170 original papers and two single-author
per unit time of reaction, is used to show how many times one
books. After getting Ph. D. degree in Kyoto University in active site produces a reaction product(s) within unit time. On the
1985, he became an assistance professor in the univer- other hand, there are no active sites, in the same meaning used for
sity. In 1996, he was promoted to an associate professor
thermal catalysis, i.e., rate of catalytic reaction is predominantly
in Division of Chemistry, Graduate School of Science,
Hokkaido University. After collaboration work with Pro- governed by the number or density of active sites, on a photocat-
fessor Kohei Uosaki for about 2 and half years, he got a full alyst [6], and the reaction rate strongly depends on various factors
professor position in Catalysis Research Center, Hokkaido
such as the intensity of irradiated light which initiates a photocat-
University in 1998. For education of graduate students,
Professor Ohtanis laboratory belongs to Graduate School of Environmental Science alytic reaction. Considering that the dark side of a photocatalyst
since 1999. or suspension does not work for the photocatalytic reaction, the
use of the term active site is inappropriate, and a relation-
ship of photocatalytic activities with active sites therefore cannot
1. Introduction be expected [7].
In kinetic analysis of general chemical reactions, a rate con-
It seems unnecessary to explain the importance of photocatal- stant is estimated and compared. Considering that photoexcited
ysis in both fundamental and application studies. Since results electrons (e ) and positive holes (h+ ) induce a redox reaction,
of many studies on photocatalysis have been reported, it seems it may be possible to estimate the rate constant of these active
rather difcult, at least for the author, to make a complete review species [8]. Since e and h+ recombine with each other (see
by introducing all or a large part of the reported studies on pho- Mutual recombination), the overall photocatalytic reaction rate
tocatalysis, while, of course, successful reviews [13] have been depends also on this recombination rate. Assuming that k(redox)
published by talented researchers in this eld. This critical review and k(recombination) are rate constants of reactions by e and
has been written in order to make clear what we know and what h+ and their recombination, respectively, i.e., the simplest kinetic
we do not know about photocatalysis, not to compile studies on model, the ratio k(redox)/k(recombination) should be a measure
photocatalysis, based on the authors experience in studies on of intrinsic photocatalytic activity [9]. However, unfortunately, we
photocatalysis for more than twenty-ve years, with apology to have no way to estimate k(recombination), since the recombination
the readers for limiting topics to so-called semiconductor pho- does not produce any chemical species to be detected for estimation
tocatalysis (see Thermodynamics). The topics are presented in of its rate [10].
alphabetical order with one gure for each topic. A small part of the
topics overlap with those discussed in the authors recent review
[4], but those topics have been revised with new gures. Known: Rate of photocatalytic reactions under given conditions,
i.e., relative photocatalytic activity and general empirical trends.
2. Topics in alphabetical order Unknown: Intrinsic photocatalytic activity, overall kinetic equa-
tion, and true correlation between physical or structural properties
Absorbance, see KubelkaMunk function and Optical band and photocatalytic reaction rate.
gap.
Action spectrum, see Doping and Wavelength dependence.
Activation energy, see First-order kinetics and Thermodynam- Adsorption, see LangmuirHinshelwood mechanism.
ics. Amorphous, see Crystallinity.
Anatase, see Crystallinity and Synergetic effect.
2.1. Activity Arrhenius plot, see Thermodynamics.
Average, see Particle size.
The term activity is often used in papers on photocatalysis as
photocatalytic activity. Although the author does not know who
rst started using this term in the eld of photocatalysis, people
involved in the eld of catalysis were using this term before the
1980s, when photocatalysis studies had begun to be accelerated
by the famous work of the so-called HondaFujishima effect on
photoelectrochemical water splitting using a single-crystal titania
electrode [5]. Most authors, including the present author, use the
term photocatalytic activity, but in almost all cases the meaning
is the same as that of absolute or relative reaction rate. One rea-
son why we like to use the term photocatalytic activity may be
that it is possible to make readers think of photocatalytic reac-
tion rate as one of the properties or abilities of a photocatalyst,
i.e., photocatalysts have individual activity, while reaction rate
is controlled by the activity under given reaction conditions. In
Fig. 1. Difference in concepts of catalytic and photocatalytic reactions: A catalyst
the eld of catalysis, catalytic activity has been used to show a contains active sites of which a substrate is converted into a product, while no active
property or performance of a catalyst, since an active site (Fig. 1) sites are present on a photocatalyst.
B. Ohtani / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 11 (2010) 157178 159
Sharpness of the peak indicated higher crystallinity of a photocat- 300 400 500 600 700
alyst. Since the width of an XRD peak reects the size of a particle, 0.025 1
i.e., the depth of crystallites measured in the direction vertical to a
absorption (extinction)
[16,4], peak sharpness shows the size of crystallites. In this sense,
crystallinity is used to show how crystallites grow to be larger-
sized particles. Another usage of the term crystallinity is to show 0.6
0.015
perfectness of crystals, i.e., higher crystallinity means lesser den-
sity of crystalline defects. As mentioned above, sharpness of XRD S-TiO 2
peaks might be a measure of crystallite size. Assuming that larger 0.01 0.4
crystallites possess smaller density of crystalline defects, sharpness
of XRD peaks can also be a relative measure of crystallinity.
0.005 0.2
Known: Relative extent of crystal growth. P25
Unknown: Methods for determination of precise composition of
0 0
crystals and actual structure of an amorphous phase.
300 400 500 600 700
wavelength/nm
Defect, see Doping, Hydroxyl radical, Nanostructure and
Visible light-induced photocatalysis. Fig. 4. Diffuse reectance (photoabsorption) and action spectra for acetic acid
decomposition in aerated aqueous solutions of sulfur-doped titania (S-TiO2 ) and
2.4. Doping (see also Visible light-induced photocatalytic Degussa (Evonik) P25.
activity)
4 4
3
3
[A]/mmol dm
ln([A0]/[A])
2 2
(slope) = k
0 0
0 10 20 30 40
time/min
a sufficient condition
necessary condition
b compound B compound A
nc5
Fig. 8. Transfer of positive holes from the bulk of a photocatalyst through the inter-
face to aqueous solution phase. Surface-adsorbed and free (hydrated) hydroxyl nc2
radicals and their derivatives can be a possible form of positive holes.
nc1
tive hole passing through this water layer into a solution may be a
hydroxyl radical or its protonated or deprotonated species (Fig. 8).
CH3 CH2 CH2 NH2 + 2 h+ + H2 O CH3 CH2 CHO + NH3 + 2H+ Interface, see Junction and Hydroxyl radical.
Intermediate, see Hydroxyl radical, Identication, Thermo-
(I-2) dynamics and Quantum efciency.
164 B. Ohtani / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 11 (2010) 157178
2.10. Junction (see also Band structure). and to check liberated whether hydrogen contains electrons orig-
inating from titania. As indirect evidence, i.e., a supporting fact,
In the early stage of studies on photocatalysis in the 1980s, a Nakabayashi et al. reported that the observed isotope distribution
photocatalyst particle was often considered to be a short-circuited in hydrogen liberated by a photocatalytic reaction with platinized
photoelectrochemical cell [33]. Titanium(IV) oxide (titania) par- titania suspended in a mixture of water and heavy water (D2 O) was
ticles loaded with platinum (Pt) were often employed as a consistent with the ratio calculated on the basis of the assumption
photocatalyst working under deaerated conditions, the same as in of hydrogen generation at the Pt surface, not on the titania surface
ordinary photoelectrochemical measurements; titania and Pt work [34]. Combining this with the fact that electrons in titania, not in
as an anode and a cathode, respectively, which are short-circuited Pt, are excited by photoirradiation, migration of e from titania to
with each other. It had been assumed that photoexcited electrons Pt has been elucidated. Similarly, small Pt deposits may enhance
(e ) in titania migrate to Pt through a junction (border) between visible light-induced photocatalytic reaction of tungsten(VI) oxide
titania and Pt. This junction had been discussed and concluded to be (WO3 ) by catalyzing multi-electron reduction of oxygen (O2 ) on
an ohmic one, not a Schottky one, considering the possible n-type their surfaces [35]. In this case, the CB bottom of WO3 is reported
character of titania electrodes prepared by heat treatment at a high to be more anodic than the standard electrode potential of O2 /O2
temperature under vacuum or hydrogen atmosphere (Fig. 10(a) and and thereby one-electron reduction of O2 by e in WO3 is thermo-
(b)). There must be such a contact between titania and Pt. Then does dynamically impossible (see Gibbs energy). The results showing
e actually migrate from titania and Pt and reduce protons to lib- enhanced photocatalytic activity of WO3 suggest the migration of
erate hydrogen in photocatalytic reaction? At present, we have no e in WO3 to Pt and reduction of O2 to hydrogen peroxide or water
technique to directly observe such migration through the junction in multi-electron processes, though no direct evidence has been
obtained and there has been no discussion on the junction between
Pt and WO3 .
It had also been believed that there should be a Schottky junction
at an n-type semiconductor (n-SC)electrolyte solution interface;
a more cathodic position of a Fermi level of n-SCs results in trans-
fer of electrons in the donor levels of n-SC to the electrolyte, and
a depletion (space charge) layer appears (Fig. 10(c) and (d)). This
is the reason why n-SC electrodes show rectifying properties, i.e.,
negligible dark anodic current even under anodically polarized con-
ditions due to the so-called Schottky barrier inhibiting electron
injection from an electrolyte to an n-SC electrode in the dark and
under photoirradiation (Fig. 10(e) and (f)); electron transfer occurs
from an n-SC electrode to an electrolyte under cathodic polarization
since no barrier exists for such electron ow, while under photoir-
radiation and anodic polarization positive holes (h+ ) can oxidize
redox species in the electrolyte and photoexcited electrons (e )
ow as an anodic current [36]. The thickness of the depletion layer,
governing slope of the electric eld, depends on the concentration
of donors (impurities). For particles of titania that are not treated
for crystalline defect (donor) introduction, the donor concentration
seems to be low, and the depletion layer thickness might be larger
than the particle size, resulting in a small potential slope in the
particles.
a 7 0.06
2x10 kr = 13 cm3 ps-1
absorbance
0.04
7 0.02
1.5x10
0
7
1/r
1x10 -0.02
-10 0 10 20 30 40 50 60
delay/ps
6
5x10
Fig. 13. An example of picosecond-time-region decay of photoabsorption (620 nm)
of trapped electrons in Degussa (Evonik) P25 particles after excitation by a ca. 100 fs
pump pulse (310 nm). The curve was analyzed by a second-order rate law (Eq. (M))
0 with a baseline component (BL), and a second-order rate constant (kr ) was obtained
0 2 4 6 8 10 to be 13 cm1 ps1 .
1/C
[42] and similarity of adsorption equilibrium constants estimated
b 8x10
6 using photocatalytic reaction rate and by dark adsorption experi-
ments are only required conditions; the observed reaction rate is
C 1 1 consistent with kinetics of a substrate undergoing Langmuir-type
= xC x
6x10
6 r kS kKS adsorption and does not exclude the possibility of other reaction
kinetics.
a second-order recombination process cannot be reproduced in an erty (properties) determining the activity of a given reaction system
ordinary photoirradiation process in which lower light intensity [48]: standardized photocatalytic activities for ve kinds of reac-
induces single-electron photoexcitation and mutual recombination tions were fairly well reproduced by a linear combination of six
occurs obeying the rst-order rate law [45]. kinds of physical and structural properties of photocatalysts, i.e.,
specic surface area, density of crystalline defects, primary parti-
Known: Behavior of recombination of e and h+ depending on their cle size, secondary particle size and existence of anatase and rutile
density. phases. It was suggested that high levels of photocatalytic activ-
Unknown: Absolute rate constant of mutual recombination ity of DAPs could not be reproduced by the correlation equations
occurring during photocatalytic reaction under ordinary photoir- derived in the above-mentioned multivariable analysis [49], i.e.,
radiation systems. another property, such as shape may affect the photocatalytic
activities.
2.14. Nanostructure
Known: Wide variety of photocatalysts with characteristic mor-
There have been many reports on preparation of photocatalysts phology can be obtained by changing reaction conditions,
with nanometer-sized structures, e.g., nanoparticles, nanoplates, especially under hydrothermal conditions.
nanocubes, nanorods, nanotubes or nanowires, as well as their Unknown: How the nano- or other size-range structures them-
photocatalytic activities [46], and the number of such reports is selves affect photocatalytic activities.
still increasing. A possible reason for such an explosive increase
in studies on nanostructured photocatalysts is popularization of Necessary condition, see First-order kinetics, Gibbs energy
low-priced pressure-tolerant Teon bottles for hydrothermal reac- and Identication.
tions. By using this type of apparatus, a variety of inorganic
compounds can be prepared and morphology can be changed 2.15. Optical band gap (see also Band structure and
depending on the reaction conditions. Scanning or transmission KubelkaMunk function)
electron microscopic images of those nanostructured photocat-
alysts (even non-photocatalysts) are attractive and interesting. Since the ability of a photocatalyst is essentially governed by
However, considering that we do not know what structural param- its band position, i.e., conduction-band (CB) bottom and valence-
eters of photocatalysts govern the photocatalytic activity and how band (VB) top, as described in Band structure, estimation of the
they govern the photocatalytic activity, there seems to be no potential of the CB bottom and VB top is important to predict the
assured reason why nanostructured materials, rather than ordinary possibility for driving a photocatalytic reaction from the thermo-
non-structured ones, should be employed. dynamic point of view. A general way for such estimation involves
Recently, decahedral-shaped anatase-titania particles (DAPs) measurements of at-band potential (see Band structure and
have been prepared by controlled gas-phase reaction of tita- Junction) and band gap corresponding to the potential of the CB
nium(IV) chloride and oxygen at 1473 K (Fig. 14) [47]. The bottom and difference in potential between the CB bottom and VB
photocatalytic activity of DAPs was reported to be much higher than top, respectively, enabling estimation of the potential of the VB top
the photocatalytic activities of commercial titania particles, e.g., from them. Optical band gap can be estimated using the following
Degussa (Evonik) P25, presumably due to relatively large specic equation:
surface area to adsorb a large amount of substrate(s) and high crys- n
(h Eg )
tallinity, i.e., less crystalline defects to reduce e h+ recombination. or ()1/n h Eg , (O)
Then, how does the decahedral shape itself affect the photocatalytic h
activity? where , h, , Eg and n are absorption coefcient, Planck constant,
In a recent study by the authors group, photocatalytic activi- oscillation frequency, optical band gap and constant relating to a
ties and physical and structural properties of 35 commercial titania mode of transition. The constant n is 1/2, 3/2 or 2 for allowed direct
powders were statistically analyzed to nd the predominant prop- transition, forbidden direct transition or indirect transition, respec-
tively, but the second forbidden direct transition has often been
neglected. It is described in textbooks of semiconductor physics
(or electronics) that by checking linearity of plots of ()1/n against
h using n as 2 or 1/2, the mode of transition of a given crystal and
optical band gap are estimated, the latter of which is estimated by
an x-intercept of the straight line [50].
Fig. 15 shows an example of the above-mentioned analysis
for well-crystallized rutile titania (TIO-3) calcined at 1173 K in
air for 2 h [51]. KubelkaMunk function (KM) is used, as usual,
instead of in Eq. (O), assuming that coefcient of scattering, s (see
KubelkaMunk function), is constant throughout a wavelength
range for measurement. Straight lines could be drawn, as shown in
Fig. 15(a), based on the assumption of indirect and direct allowed
transitions, and their x-intercepts, 3.00 and 3.14 eV correspond-
ing to band gaps for indirect and direct transition, respectively,
are obtained. Then, which is better, indirect and direct transi-
tion? An approximate range of the straight part of these plots was
0.14 eV and 0.20 eV for indirect and direct transitions, respectively,
suggesting direct transition for this rutile titania sample [52].
However, can we conclude that the transition of the rutile sam-
Fig. 14. A representative SEM image of decahedral anatase-titania particles pre-
ple is direct and the band gap is 3.14 eV? At least for the author,
pared by controlled gas-phase reaction of titanium(IV) chloride and oxygen at
1473 K. Most particles expose two square (001) facets and eight trapezoidal (101)
the answer is no. The range of the linear part of plots in Fig. 15(a)
facets. seems to have a large extent of facultativity, and there seem to be
168 B. Ohtani / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 11 (2010) 157178
400 K
(KMh)2
3 L= , (P)
cos
where L, K, , , and are size of particles in the direction vertical to
2
the corresponding lattice plane, a constant [54], wavelength of X-
200
ray, corrected full width at half maximum (FWHM) of the XRD peak,
1 0.20 and angle of diffraction, respectively. Since K has been introduced
eV in the derivation of this equation [16] to be 0.891, K has its effective
0 0 digits. The effective digits of the resulting size of particle L should
2.50 2.75 3.00 3.25 3.50 be one if 0.9 is used for K. In order to show the size L with three-
h /eV digit accuracy, a value of 0.891 should be used. Another point
to be noted is the correction of FWHM. Two kinds of correction are
b 100 10 required. One is correction for broadening due to K2 radiation, and
the other is for broadening due to the optical path in the diffrac-
tometer. Generally speaking, the former and the latter corrections
80 8
are made by assuming a radiation intensity ratio of K1 and K2 and
by using FWHM of a standard large crystalline sample. There are at
%absorption
K-M function
60 6 least three ways for the latter correction, but there seems to have
been no discussion on the best way. The simplest way, subtraction
of FWHM of the standard, has often been employed. To the authors
40 4 knowledge, both corrections appreciably affect the size of particles;
a large error is expected without such corrections. Therefore, when
the size of particles is demonstrated, the methods used for FWHM
20 2
corrections should be described [55].
As described above, the size of particles in the direction vertical
0 0 to the corresponding lattice plane can be estimated using Scher-
2.50 2.75 3.00 3.25 3.50 rers equation. In comparison with the data obtained in another
h /eV way, e.g., transmission or scanning electron microscopy, difference
due to this may be observed. Furthermore, broadening of FWHM of
Fig. 15. (a) Two kinds of plot for determination of band-gap energy (Eg ) of titania XRD peaks is also induced by distortion of the crystalline lattice. The
(JRC-TIO-3 calcined at 1173 K). Assumptions of indirect (left) and direct allowed WilliamsonHall equation includes this, as well as the effect of par-
transitions give band gaps of 3.00 and 3.14 eV, respectively. (b) Plots of %absorption
(%extinction) and KubelkaMunk function of the same sample. X-intercepts of lines
ticle size [56]. In other words, Scherrers equation neglects the effect
tting to the spectra in the absorption edge region are 2.91 and 3.07 eV, respectively. of crystal lattice distortion. For samples that are expected to have
a large degree of distortion, analyses using the WilliamsonHall
equation should be carried out.
no rules to discriminate a mode of transition on the basis of range Then, can we know particle size from the above-mentioned
of the linear part. An interesting point is that it is also possible to analysis? A problem is still hidden. Distribution of size and its effect
estimate band gaps from absorption spectra. Fig. 15(b) shows such on evaluation of average size should be considered. Fig. 16 shows
examples; straight lines can also be drawn for plots of %absorp- hypothetic fractions of spherical particles of 1, 10, 25 and 100 nm
tion and KM function to give x-intercepts of 2.91 eV and 3.07 eV, in diameter calculated on the basis of number, surface area and vol-
respectively. Again, there are no xed ways to nd reliable and ume (weight). Even if number-based fractions are all the same, 25%,
reasonable data on band-gap energy [53]. for the four kinds of particles, surface area-based and volume-based
Another problem in the band-gap energy estimation is inu- fractions are 93% and 98%, respectively. Therefore the average size
ence of impurity or surface electronic states on absorption spectra; of particles calculated on the basis of these fractions is 34, 94 or
spectral curves at the absorption edge region may be markedly 98 nm when number, surface area or volume fraction is employed
modied to give smaller band-gap energy, and it is very difcult [57]. (A notable point is that the difference in average size between
to extract a part of the absorption due to band-gap excitation from the mode of surface area-based and volume-based ones is smaller
that by impurities (dopants) or surface states. Therefore, especially
for doped samples, discrimination of narrowing of the band gap and
overlapping of absorption by bulk or surface modication from shift
of absorption to longer wavelengths is difcult (see Doping and
Visible light-induced photocatalysis).
than 5%.) When particle size is measured by electron microscopic molecule, four silver metal atoms, or four protons, and when the
images of SEM or TEM, particles are categorized by size and counted O2 yield is used, the quantum efciency is calculated to be:
(making a histogram) from images in ordinary procedures [58]. The
average size may be a number-based one unless volume fraction is 4n(O2 )
, (Q2)
considered. n(photon)
Then, what is the mode of average size of particles, with wide
where n is the number of molecules or photons [63]. Thus, for the
distribution of size, obtained by Scherrers equation; average in
calculation of quantum efciency of heterogeneous photocatalytic
number, surface area or volume based one. There seems to be no
reactions, it is reasonable to consider the efciency of utilization
clear answer. Assuming a spherical shape of titania particles with
of electrons-positive holes assuming that an electron and positive
density of 4 g cm3 , a product of specic surface in the unit of m2 g1
hole pair is produced by absorption of a photon. However, since
and diameter in the unit of nm is calculated to be 1500 [59]. In the
neither a photoexcited electron nor a positive hole appears in stoi-
authors experience, many commercial titania samples obey this
chiometry, the above-mentioned consideration may not always be
correlation when (primary) particle size is measured using Scher-
straightforward. For example, acetic acid dissolved in air-saturated
rers equation, suggesting that average particle size obtained using
water is decomposed into carbon dioxide by an appropriate sus-
Scherrers equation is similar to the surface area-based or volume-
pended photocatalyst with the following proposed stoichiometry:
based average.
Another topic related to particle size of a photocatalyst is CH3 COOH + 2O2 2CO2 + 2H2 O (Q3)
quantum-size effect. This frequently used term means that when
the size (radius) of solid particles becomes smaller than their Bohr How many photons are required for this reaction? Assuming that
radius, the bottom of the conduction band and the top of the valence only O2 is reduced by photoexcited electrons in this reaction and
band shift in negative (high electronic energy) and positive (low that reduction of an O2 molecule requires four electrons, this reac-
electronic energy) directions, respectively, resulting in expansion tion is an eight-electron process. However, since the photocatalytic
of the band gap. The Bohr radius for anatase particles and that for reaction of acetic acid may include a radical chain mechanism or at
rutile particles have been estimated to be 2.5 and 0.3 nm, respec- least addition of O2 to intermediate radicals, an acetic acid molecule
tively [60]. Preparation of crystalline titania particles of such small could be decomposed by less than 8 electrons. It is impossible to cal-
size seems to be difcult, and titania particles claimed in papers culate intrinsic quantum efciency, i.e., efciency of utilization of
to show a quantum-size effect might be larger than these sizes. A electron-positive hole pairs, only from the product yield. Therefore,
blue (shorter wavelength) shift, if observed, of the absorption edge quantum efciency is reported with the description that a given
of those samples might be due to the amorphous part of titania, reaction is assumed to proceed through a proposed multi-electron
not to the quantum-size effect. At least for a titania photocata- process (e.g., 8 for acetic acid decomposition or 4 for carbon dioxide
lyst, use of quantum-size effect for interpreting the results seems liberation in Eq. (Q3)).
inappropriate. Another problem for determination of quantum efciency is the
difculty in determining the number of absorbed photons. Unlike
Known: Several methods to determine particle size. measurement for homogeneous solutions, solid materials scatter
Unknown: Mode of average size, number, surface-area or volume incident photons to reduce the light intensity arriving at a detec-
(or others)-based one. tor in a spectrophotometer. In the wavelength region in which
only some of the photons are absorbed, i.e., around the band-edge
wavelength, it is difcult to measure the photoabsorption ef-
P25, see Synergetic effect
ciency. (This problem is discussed in KubelkaMunk function.)
Photocatalysis, see Thermodynamics and Band structure.
Therefore, apparent quantum efciency (photonic efciency) has
Positive hole, see Hydroxyl radical.
often been used instead of quantum efciency, and apparent quan-
tum efciency is calculated by the number of incident photons
2.17. Quantum efciency (see also Mutual recombination and rather than the number of photons used for quantum efciency
Wavelength dependence) calculation. Since quantum efciency is dened as efciency of
electron-positive hole utilization, apparent quantum efciency is
The term quantum efciency or quantum yield was origi- a product of efciencies of photoabsorption and electron-positive
nally dened as a ratio of number of products (or consumed starting hole utilization as shown in Fig. 17(a). Of course, both quantum
material) to that of absorbed photons in photoreaction in homo- efciency and apparent quantum efciency depend on the irradi-
geneous phase, i.e., in solutions or gas phase, assuming that one ation wavelength and sometimes on the irradiation intensity (see
photon induces reaction or change in one molecule (second law of Fig. 17(b)), and thereby the data should be shown with wavelength
photochemistry) [61] since a multi-photon process and subsequent and preferably with intensity. In this sense, such measurement
multi-electron transfer can be neglected in ordinary photoirradia- must be performed by monochromatic irradiation; irradiation with
tion conditions with relatively low photon ux. Confusion might sharp-cut optical lters is inappropriate (see Wavelength depen-
arise when this concept is applied to photocatalysis, in which a dence).
multi-photon process and at the same time radical chain reaction
may be included, especially in reaction in the presence of oxygen. Known: Practical method for determination of apparent quantum
For example, photocatalytic silver metal deposition accom- efciency dened as the ratio of number of electrons or holes used
panied by molecular oxygen (O2 ) liberation proceeds with the in photocatalytic reaction to number of photons incident on a reac-
following stoichiometry [62]: tion system, assuming the number of electrons or holes required
to produce one molecule of product(s).
4Ag+ + 2H2 O 4Ag + O2 + 4H+ (Q1)
Unknown: Intrinsic quantum efciency, i.e., the fraction of
electron-positive hole pairs that escape from their mutual recom-
For the determination of quantum efciency, it is necessary to make
bination.
an assumption on how many photons are required for the reaction.
One of the possible and the most frequently employed assump-
tions is that four photons are required for liberation of one oxygen Quantum-size effect, see Particle size.
170 B. Ohtani / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 11 (2010) 157178
Fig. 18. (a) Model interpreting a rate-determining step in a 4-step series reaction.
Step 3 with largest activation energy determines the overall rate. (b) Simplied
scheme of photocatalytic reaction with a photocatalyst pc. It is clear that this
process is not a series reaction since a deactivation step is included.
Fig. 17. (a) Schematic representation of relations among photoabsorption, quan- Known: Inappropriateness of discussion on rates of reduction by
tum and apparent quantum efciencies. (b) Examples of dependence of apparent photoexcited electrons and oxidation by positive holes.
quantum efciency on incident-light intensity at several wavelengths; (upper) Unknown: Intrinsic rate of recombination governing the overall
dehydrogenation of methanol and (lower) oxidative decomposition of acetic acid. reaction rate.
Merck titania was used as a photocatalyst.
2.18. Rate-determining step (see also First-order kinetics) Recombination, see Mutual recombination.
Redox mediator, see Z-scheme photocatalysis.
Assuming that a certain reaction proceeds through a series of Rutile, see Synergetic effect.
steps without any branching reactions, the rate must be the same as Scherrer equation, see Particle size.
the rate of the slowest step, the rate-determining step, i.e., the over- Second-order kinetics, see Mutual recombination.
all activation energy is that of the rate-determining step (Fig. 18(a)). Semiconductor, see Band structure, Energy conversion, Junc-
This original denition cannot be directly applied to photocatalysis. tion, Optical band gap and Thermodynamics.
A plausible reason is that reactions by photoexcited electrons and Solar radiation, see Ultraviolet light in solar radiation.
positive holes occur in parallel, not in series (Fig. 18(b)). Consid- Stoichiometry, see Quantum efciency, Visible light-induced
ering the requirement of photocatalysis for the same numbers of photocatalysis and Yield.
electrons and positive holes to be used, it seems possible to com-
pare the rates of electron and positive-hole reactions. However, 2.19. Synergetic effect
it seems that overall reaction rate must also be inuenced by the
rate of recombination of e h+ included in the process. In ordinary The term synergetic effect in photocatalysis could be dened
photochemistry in homogeneous phase, steady (stationary)-state as follows: when more than two kinds of photocatalysts are used
approximation is used to analyze kinetics, assuming a forward reac- as a mixture, the overall photocatalytic activity exceeds the sum of
tion to give products and a backward reaction, i.e., deexcitation. In activities of each photocatalyst. When a certain component alone
such analyses, there may not be a rate-determining step. is not a photocatalyst, and its mixture with a photocatalyst shows
B. Ohtani / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 11 (2010) 157178 171
energy
G > 0
absorption absorption
synergetic
4 effect HOMO HOMO
Gh < 0 Gh < 0
atom/molecule
3
e e
2 G > 0 G < 0
1.2 It should be noted that the calculation has been made on the
1.6
basis of energy of light, not on the basis of number of photons. In
1.2
discussing energy conversion efciency (see Energy conversion),
0.8 0.8
e.g., for solar cells, in which efciency of conversion of solar energy
0.4 is signicant, content of light of a certain wavelength region should
0.4 0 be discussed on the basis of energy. However, photocatalytic reac-
200 400 600 800
tions except for energy conversion systems of, e.g., photosplitting
wavelength/nm
of water into hydrogen and oxygen, do not convert energy. Thus,
0
500 1000 1500 2000 2500 3000 3500 4000 the energy conversion efciency cannot be discussed and, instead,
wavelength/nm quantum efciency (described below) becomes signicant. Under
such conditions, the content of light of a certain wavelength region
Fig. 21. A solar spectrum drawn on the basis of data of solar radiation at the ground should be calculated or measured on the basis of photon numbers.
surface in the temperate zone (AM1.5) reported by the American Society for Test-
ing and Materials (ASTM, http://www.astm.org/ ASTM). An inserted spectrum is a
Energy of a photon in the ultraviolet light region is larger than that
magnied drawing for a part of ultraviolet and visible light. in the visible light and infrared light regions and thereby the con-
tent calculated on the basis of number of photons must be much
smaller than 35%.
above-mentioned band structure on or in a solid can be a pho-
tocatalyst (Fig. 20), and even when a bulk material is used, the
photoabsorption and resultant photocatalytic reaction may pro- Known: Content of ultraviolet light in solar radiation based on
ceed at a localized site when, for example, photocatalysts are energy.
photoirradiated at a wavelength near the band gap. Therefore, the Unknown: Content of ultraviolet light in solar radiation based on
interpretation using a band model is not always adequate for under- photon number.
standing photocatalysis. In this sense, the term heterogeneous
photocatalytic reaction (photocatalysis) seems better than semi-
2.22. Visible light-induced photocatalysis (see also Band
conductor photocatalytic reaction based on the electronic band
structure and Doping).
structure.
Since solar radiation includes light of wavelengths from 280 to
Known: Thermodynamic explanation using a band model of semi- 4000 nm, use of photocatalysts that absorb light in visible and near-
conducting materials. infrared regions is desired. However, light of longer wavelength has
Unknown: General description of thermodynamics of photocatal- smaller energy, leading to a decrease in potential for redox reac-
ysis dened for materials including molecular or highly dispersed tions, and thereby reactions driven by visible-light irradiation are
compounds. limited to those satisfying the thermodynamic and kinetic require-
ments. Many studies have been performed to design and develop
Time-resolved spectroscopy, see Mutual recombination photocatalysts that work under visible-light irradiation. Strategies
Turnover number, see Yield that have usually been employed in such studies are modication
Ultravioletvisible spectroscopy, see Optical density. (doping) of titania to give visible-light absorption or use of colored
mixed metal oxide and nitride. Although it is expected that such
2.21. Ultraviolet light in solar radiation (see also Energy doping of crystalline and use of mixed metal oxide/nitride may
conversion and Visible-light induced photocatalysis) induce formation of lattice defects, which enhance electron-hole
recombination, resulting in lower photocatalytic activity (see Dop-
Since titania, a representative potential photocatalyst, can ing), discussion of the strategies is not a purpose of this section.
absorb only ultraviolet light, at a wavelength shorter than ca. Some problems in studies on visible light-photocatalytic activity
390 nm for anatase titania (see Optical band gap), many studies to are discussed here.
extend this absorption range to the visible light region have been First, the boundary wavelength between ultraviolet light and
performed in order to utilize solar radiation as an energy source visible light should be dened. The meaning of the term visible
for photocatalysis. For the practical application of photocatalysis light is light that can be seen, and the limiting wavelength dif-
using solar radiation and indoor light, such an extension of wave- fers among individuals. Many studies demonstrating that visible
length range is very important. In the introductory parts of reports light induced, as expected, a photocatalytic reaction used the condi-
on related subjects, the content of ultraviolet light in solar radiation tion of photoirradiation through an optical cut-off lter, L-42 or its
is often stated to be ca. 35% without showing a reference, presum- equivalent, and the irradiation wavelength under such conditions
ably due to repeated re-quotations. The author could not nd the used to be described as >420 nm. However, this is inadequate
original paper showing calculation of this content. because this lter transmits light of wavelength >ca. 390 nm. Actu-
Using data on solar radiation at the ground surface in the tem- ally, in the authors experience, appreciable photocatalytic activity
perate zone (AM1.5; Fig. 21) [66] reported by the American Society of not only rutile but also anatase titania photocatalysts could be
for Testing and Materials (ASTM, http://www.astm.org/ ASTM), observed by irradiation through the lter. When this optical lter is
accumulated intensity of ultraviolet light below wavelengths of used for irradiation, at least comparison of the photocatalytic activ-
380, 390, 400 and 410 nm corresponds to 3.24, 3.87, 4.63 and 5.81%, ity with that of a representative titania photocatalyst, such as P25
respectively, of total energy of light of wavelength up to 4000 nm (Degussa (Evonik)) or ST-01 (Ishihara Sangyo), is necessary. Other-
(1001.0 W m2 ) [67]. Considering that the boundary wavelength wise, optical cut-off lters of longer transmission limits should be
of ultraviolet light and visible light is ambiguous (since the term used. One of the possible and smart ways is to dene visible light
visible means that one can see the light and the boundary is per- as light that negligibly drives photocatalytic reaction by titania and
ceived to be different among individuals), the above-mentioned to use an appropriate optical lter to realize this, considering the
content, 35% (or 36%), seems reasonable. The rst paper, if it history of studies on photocatalysis.
B. Ohtani / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 11 (2010) 157178 173
Fig. 22. Possible mechanism of dye-sensitized photoinduced oxidative decompo- Since photocatalytic reaction is induced by photoabsorption as
sition of dye in the presence of oxygen. Components (1)(4) are necessary conditions
a general understanding, something, a photocatalyst, must absorb
for dye-sensitized reaction and they are also necessary conditions for ordinary
photocatalytic reaction. light to drive the photocatalytic reaction. Therefore, photocatalytic
activity depends strongly on the wavelength of incident light,
and showing resemblance of an action spectrum [72], wavelength
Showing that a certain reaction proceeds under the above- dependence of photocatalytic activity, with an absorption spectrum
mentioned visible-light irradiation conditions is not proof of visible of a photocatalyst candidate is essential to prove photocatalysis
light-induced photocatalytic activity unless an appropriate com- (assuming constant, wavelength-independent, true quantum ef-
pound is used for the photocatalytic activity test. Dyes have ciency, as discussed in Quantum efciency).
relatively large photoabsorption (extinction) coefcients (This is An action spectrum is a plot of apparent quantum efciency,
the reason why they are used as dyes, i.e., coloring agents.) and not (true) quantum efciency, against wavelength of light used for
therefore measurements of their concentrations in solutions are apparent quantum efciency measurement (see Fig. 17(a)). There-
easy even if the concentrations are very low. However, the fact fore, it is clear that monochromatic light irradiation is required to
that dyes absorb visible light indicates that a photoreaction might record an action spectrum. A grating-type monochromator with a
be induced by visible-light photoabsorption (dye sensitization) light source such as a xenon arc lamp is usually used. Interference-
as well as by photoabsorption of a photocatalyst. Although this type optical lters, transmitting at only a certain wavelength
problem had been pointed out earlier [68], there have been no region, can also be used for monochromatic irradiation, though
clear experimental results showing the self-photodecomposition wavelengths of possible irradiation are limited. Because of possi-
of dyes, presumably due to the difculty in determining the mech- ble dependence of apparent quantum efciency on light intensity
anism of this photoreaction. Actually, determination of the extent [73], it is preferable to adjust the light intensity at each wave-
to which incident photons are absorbed by a dye and photocatalyst length [74]. Wavelength-selective irradiation can be carried out
is difcult. A paper has been published to show that methylene blue using optical lters that transmit light of wavelength longer than a
(MB), the most frequently employed dye, as well as, presumably, certain limit, i.e., cut-off lters or sharp-cut lters. By using sev-
other kinds of organic dyes, is inappropriate as a model compound, eral cut-off lters with different cut-off wavelengths [75], a plot of
particularly for testing visible light-induced photocatalytic activity apparent quantum efciency against the cut-off wavelengths can
[19,69] There are at least three reasons for its inappropriateness. be obtained, and this pseudo action spectrum seems like a true
One is that the dye molecules absorb photons, especially in the vis- action spectrum obtained by the above-described monochromatic
ible light range, and thus-photoexcited electrons may be injected irradiation. However, those are completely different; a pseudo
into photocatalyst particles as has been suggested by similarity action spectrum is an integrated (from longer to shorter wave-
of the action spectrum similar to the photoabsorption spectrum lengths) form of a true action spectrum if it is assumed that light
of the dye [70,71] (Fig. 22). Another reason is that the absolute intensity is constant in the whole range of irradiation, because of
molar amount of dye contained in the reaction system can be much the difference in irradiation wavelength region (Fig. 23). Conse-
smaller than that of a solid photocatalyst. The concentration of dye quently, the corresponding action spectrum should be estimated by
in the solution should be relatively low since the absorption coef- differentiation of the pseudo action spectrum, i.e., a pseudo action
cient is large, as described in the rst part of this section (see spectrum is an integrated form, from a longer wavelength side, of
Yield). These two facts are closely related to the problem of how corresponding action spectrum. A horizontal part in a pseudo action
we can prove a given reaction to be photocatalytic, as discussed spectrum (Fig. 23), if present, shows that apparent quantum ef-
in Thermodynamics in this review. The third reason is that the ciency at the wavelength is negligible even if an appreciable value
mechanism of dye degradation is so complicated that efciency is seen in the pseudo action spectrum.
of the photocatalytic reaction, e.g., quantum efciency, cannot be
measured. Measuring the consumption (decrease) of a dye during
photoirradiation requires only a spectrophotometer, but the use of Known: Importance of discussion with wavelength dependence,
dyes as model compounds is inappropriate, and if dyes are used, i.e., an action spectrum.
care must be taken in the analysis of experimental results. Unknown: (partly) Difference in an action spectrum and a pseudo
Then, what experimental results are needed to show that a action spectrum.
given photocatalyst can drive the photocatalytic reaction under the
conditions of visible light irradiation? In the authors experience,
it is preferable to show an action spectrum, not pseudo action X-ray diffraction, see Crystallinity.
174 B. Ohtani / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 11 (2010) 157178
apparent quantum
action spectrum
efficiency
pseudo
action spectrum
wavelength/nm
Fig. 23. Examples of action (solid line) and pseudo action (dotted line) spectra for
photoinduced reaction by an ordinary semiconductor photocatalyst (shorter wave-
length) and an organic dye (longer wavelength peak). A pseudo action spectrum
taken by cut-off lters corresponds to integration of the true action spectrum
Fig. 25. An example of calculation of turnover number in photocatalytic conversion
from the longer-wavelength side.
of A into B. %Yield cannot show turnover of the photocatalyst.
3. Conclusions
tion and Grant-in-Aid for Scientic Research (A) (General) from even if the molar absorption (extinction) coefcient is unknown and thereby
the Ministry of Education, Culture, Sports, Science and Technology absolute concentration cannot be determined. Moreover, taking the logarithm
means a reciprocal value can also be used, e.g., [A0 ]/[A], where [A0 ] is the initial
(MEXT) of Japan. concentration of a substrate A (see Fig. 6).
[23] Under these conditions, rate constant k obtained by the rst-order kinetic
analysis does not reect the reactivity of a photocatalyst; k contains diffusion
References constant of a substrate and surface area of a photocatalyst (strictly speaking,
area of diffusion layer on the photoirradiated active surface).
[1] M.R. Hoffmann, S.T. Martin, W.Y. Choi, Chem. Rev. 95 (1995) 69. [24] It should be noted that the overall rate obeys the rst-order rate law in regard
[2] The following reviews on photocatalysis have been published in this journal; to surface concentration of a substrate but not concentration in the bulk.
(a) S. Ananda, M. Yoon, this journal 4 (2003) 5; [25] In the lower concentration region of a Langmuir adsorption isotherm, similar
(b) T. He, J.N. Yao, ibid 4 (2003) 125; linear dependence is also observed. However, if a given reaction condition lies
(c) D. Chatterjee, S. Dasgupta, ibid 6 (2005) 186; in this region, a Henry-type adsorption isotherm, not a Langmuir-type one,
(d) D.G. Shchukin, D.V. Sviridov, ibid 7 (2006) 23; should be used, since the most signicant characteristic of a Langmuir isotherm
(e) V. Augugliaro, M. Litter, L. Palmisano, J. Soria, ibid 7 (2006) 127; is saturation at a high concentration.
(f) U.I. Gaya, A.H. Abdullah, ibid 9 (2008) 1; [26] Backward electron (hole) transfer can be avoided thermodynamically only
(g) K. Rajeshwar, M.E. Osugi, W. Chanmanee, C.R. Chenthamarakshan, M.V.B. when the CB bottom and VB top are more positive and negative than standard
Zanoni, P. Kajitvichyanukul, R. Krishnan-Ayer, ibid 9 (2008) 171; electrode potentials of a reductant and an oxidant, respectively.
(h) Y. Shiraishi, T. Hirai, ibid 9 (2008) 157. [27] A probable rst report on photocatalytic liberation of hydroxyl radical: C.D.
[3] (a) A. Fujishima, X.T. Zhang, D.A. Tryk, Surf. Sci. Rep. 63 (2008) 515; Jaeger, A.J. Bard, J. Phys. Chem. 83 (1979) 3146.
(b) A. Fujishima, X. Zhang, D.A. Tryk, Int. J. Hydrogen Energy 32 (2007) 2664. [28] Recent papers on photocatalytic production of hydroxyl radical:
[4] B. Ohtani, Chem. Lett. 37 (2008) 216. (a) T. Hirakawa, Y. Nosaka, Langmuir 18 (2002) 3247;
[5] A. Fujishima, K. Honda, Nature 238 (1972) 37. (b) Y. Nosaka, S. Komori, K. Yawata, T. Hirakawa, A.Y. Nosaka, Phys. Chem. Chem.
The HondaFujishima effect is a well-known chemical phenomenon closely Phys. 5 (2003) 4731;
related to photocatalysis, and the paper published in Nature in 1972 had (c) B. Tryba, M. Toyoda, A.W. Morawski, R. Nonaka, M. Inagaki, Appl. Catal. B:
undoubtedly promoted research on photocatalysis but was not be an origin Environ. 71 (2007) 163;
of heterogeneous photocatalysis in the bibliographic sense [3(a)]. (d) Y. Murakami, K. Endo, I. Ohta, Y. Nosaka, J. Phys. Chem. 111 (2007) 11339;
[6] Sometimes the term active site is used for a photocatalytic reaction system (e) T. Hirakawa, K. Yawata, Y. Nosaka, Appl. Catal. A: Gen. 325 (2007) 105;
with dispersed chemical species, e.g., metal complexes or atomically adsorbed (f) S. Sroiraya, W. Triampo, N.P. Morales, D. Triampo, J. Ceram. Proc. Res. 9 (2008)
species, on support materials. Even in this case, a photocatalytic reaction occurs 146;
only when the species absorb light, and species not irradiated therefore can- (g) C.Y. Chang, Y.H. Hsieh, L.L. Hsieh, K.S. Yao, T.C. Cheng, J. Hazard. Mater. 166
not be active sites. A possible mechanism of photoinduced reaction is that (2009) 897, and references therein.
photoirradiation produces stable active sites which work as reaction cen- [29] For example: M. Fleischmann, K. Korinek, D. Pletcher, J. Electroanal. Chem. 31
ters of thermal catalytic reactions, though this is different from the ordinary (1971) 39.
mechanism of photocatalysis by photoexcited electrons and positive holes. [30] S.-I. Nishimoto, B. Ohtani, T. Yoshikawa, T. Kagiya, J. Am. Chem. Soc. 105 (1983)
[7] Therefore, concepts of catalyst and photocatalyst are different and it can be 7180;
stated, at least, that a photocatalyst cannot be a photoirradiated catalyst or (a) B. Ohtani, S. Tsuru, S.-I. Nishimoto, T. Kagiya, K. Izawa, J. Org. Chem. 55 (1990)
photocatalytic reaction cannot be induced by irradiation of a catalyst. How- 5551;
ever, many papers describe that the author(s) prepared a catalyst(s) and used (b) B. Ohtani, B. Pal, S. Ikeda, Catal. Surv. Asia 7 (2003) 165.
it for photocatalytic reactions. This opened up a way for photocatalytic organic syntheses, e.g., cyclic amino
[8] Of course, overall rate of photocatalytic reaction must depend on photoir- acid synthesis.
radiation intensity and may be expressed with the rate constant as well as [31] In addition to this accidental coincidence of a peak position, authentic sam-
concentration of reaction substrate(s) and photoirradiation intensity. ples sometimes contains impurities giving chromatographic peaks, which are
[9] In a recent paper of a collaboration work with the author, discussion of the often misunderstood as peaks of authentic compounds. For example, commer-
ratio k(redox)/k(recombination) was presented: S.-Y. Murakami, H. Kominami, cial aqueous solutions of formaldehyde (formalin) generally contains methanol
Y. Kera, S. Ikeda, H. Noguchi, K. Uosaki, B. Ohtani, Res. Chem. Interm. 33 (2007) as a stabilizer. In gas-chromatography analysis using an FID (ame ionization
285. detector), formaldehyde is not detected (no peak) but methanol gives a peak,
[10] It has been suggested that second-order rate constant of decay of photoin- which is often assigned to formaldehyde in error.
duced absorption due to trapped electrons in titanium(IV) oxide photocatalysts [32] At least it is necessary to isolate and identify the product(s) when a given
can be a relative, but not absolute, measure of recombination rate constant, process is claimed to be applicable to practical synthetic process.
k(recombination). See Mutual recombination. [33] For example: A.J. Bard, J. Phys. Chem. 86 (1982) 172.
[11] In femto-second pump-probe transition photoabsorption measurements, tita- [34] (a) S. Nakabayashi, A. Fujishima, K. Honda, Chem. Phys. Lett. 102 (1983) 464;
nia photocatalysts gave visible-light photoabsorption of trapped e within (b) R. Baba, S. Nakabayashi, A. Fujishima, K. Honda, J. Phys. Chem. 89 (1985)
ca. 100 fs pump pulse without showing photoabsorption of e in the CB. See 1902.
Mutual recombination. [35] R. Abe, H. Takami, N. Murakami, B. Ohtani, J. Am. Chem. Soc. 130 (2008)
[12] Since the band structure of semiconductors, as well as insulators, consists of 7780.
a lled valence band (VB) and vacant conduction band (CB), their Fermi level [36] The height of the Schottky barrier, inducing a rectifying property, depends on
must be located between the VB and CB. the difference between Fermi level of an n-SC and standard electrode potential
[13] B. Ohtani, Y. Ogawa, S.-I. Nishimoto, J. Phys. Chem. B 101 (1997) 3746. of an electrolyte solution but not on the donor concentration and electrode
[14] B. Ohtani, Y. Azuma, D. Li, T. Ihara, R. Abe, Trans. Mater. Res. Soc. Jpn. 32 (2007) potential. This might be a reason why a rectifying property is observed for
401. not only strongly reduced n-SCs with relatively high donor density but also
[15] B. Ohtani, O.O. Prieto-Mahaney, D. Li, R. Abe, J. Photochem. Photobiol. A: Chem. particulate lms with low donor density.
216 (2010) 179. [37] The rst law of photochemistry, Only the light which is absorbed by a molecule
[16] H.P. Klug, L.E. Alexander, X-Ray Diffraction Procedures for Polycrystalline and can be effective in producing photochemical change in the molecule was for-
Amorphous Materials, 2nd ed., Wiley, New York, 1974, p. 687. mulated by Grotthus and Draper in the 19th century (J.G. Calvert, J.N. Pitts, Jr,
[17] R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki, Y. Taga, Science 293 (2001) 269. Photochemistry, Wiley, New York, 1966, p. 19).
[18] (a) D. Mitoraj, R. Beranek, H. Kisch, Photochem. Photobiol. Sci. 9 (2010) 31; [38] For turbid samples with light scattering, log(I0 /I) may be called optical den-
(b) D. Mitoraj, H. Kisch, Chem. Eur. J. 16 (2010) 261. sity.
[19] X. Yan, T. Ohno, K. Nishijima, R. Abe, B. Ohtani, Chem. Phys. Lett. 429 (2006) [39] Another point to check is adequate use of substrate concentration in analysis.
606. Since a Langmuir isotherm is derived on the basis of the adsorption equilibrium
This paper claimed inappropriate use of organic dyes as test compounds for between species adsorbed and desorbed in solution, the isotherm is a func-
visible-light sensitive photocatalysts. Citation of this paper was not expected tion of concentration of the adsorbate (molecules to be adsorbed on surfaces)
at all, since the authors using methylene blue (MB) as a model compound for in solution, not concentration of the adsorbate in feed. Therefore, the actual
photocatalytic reaction do not want to refer to this, and, on the other hand, concentration, which must be reduced from that in feed due to appreciable
those who do not use MB need not to refer to this. However, there has been an adsorption, must be measured at least before photoirradiation.
appreciable number of citations, and, to the authors surprise, approximately [40] In some reports on photocatalytic reaction, it has been stated that the
half of the citations of this paper were for reasonable use of MB, indicating that time-course curve obeys the rst-order rate law and that the rate of
authors of those papers did not read the paper. reaction changes following Langmuir-type adsorption behavior (so-called
[20] In other words, it is impossible to discuss energy conversion efciency for LangmuirHinshelwood mechanism) when the substrate concentration is
reaction of negative Gibbs energy change as shown in Fig. 5(b). changed, though these two facts are incompatible. A possible situation is that a
[21] It is thought that there should be at least ca. 200 mV overpotential in both double reciprocal plot of rate and substrate concentration (Fig. 8(a)) is linear but
reduction by e and oxidation by h+ , and this shifts the limiting wavelength for that the line passes through the origin, suggesting that the rate is proportional
water splitting by ca. 250 nm. to the substrate concentration.
[22] In this analysis of rst-order kinetics, any value can be plotted against time of [41] As an example of papers showing the coincidence of equilibrium adsorption
reaction, unless the value is proportional to the molar amount of a substrate constants obtained from the photocatalytic reaction rate and adsorption in the
or product; e.g., absorbance of a compound at a given wavelength can be used dark, see: F. Amano, K. Nogami, B. Ohtani, Langmuir 26 (2010) 7174.
B. Ohtani / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 11 (2010) 157178 177
[42] It should be noted that the concept of Langmuirian adsorption does not con- [65] Papers showing activation energy of photocatalytic reactions in the period
tain a meaning of surface and the isotherm is derived even assuming an 20002004. Figures and compounds in square brackets and parentheses,
equilibrium of reaction, production of associates of surface adsorption site respectively, show activation energy in the unit of kJ mol1 and substrate;
and adsorbate in a homogeneous phase, as no interaction between surface (a) A.V. Vorontsov, I.V. Stoyanova, D.V. Kozlov, V.I. Simagina, E.N. Savinov, J.
adsorption sites are presumed. Such interactions are of signicant character- Catal. 189 (2000) 360 [10.9 (acetone)];
istics of surfaces. (b) Y.M. Xu, Chem. J. Chin. Univ. Chin. 21 (2000) 1539 [4.24.6 (acetophenone)];
[43] (a) D.P. Colombo Jr., R.M. Bowman, J. Phys. Chem. 99 (1995) 11752; (c) W.Y. Su, X.Z. Fu, K.M. Wei, Chem. J. Chin. Univ. Chin. 22 (2001) 272 [13.7 (bro-
(b) D.P. Colombo Jr., R.M. Bowman, J. Phys. Chem. 100 (1996) 18445; momethane)];
(c) D.P. Colombo Jr., K.A. Roussel, J. Saeh, D.E. Skinner, J.J. Cavaleri, R.M. Bowman, (d) J. Lea, A.A. Adesina, J. Chem. Tech. Biotech. 76 (2001) 803 [7.83 (nitrophe-
Chem. Phys. Lett. 232 (1995) 207. nol)];
[44] B. Ohtani, H. Kominami, R.M. Bowman, P. Colombo Jr., H. Noguchi, K. Uosaki, (e) O.E. Kartal, M. Erol, H. Oguz, Chem. Eng. Technol. 24 (2001) 645 [16.2 (phe-
Chem. Lett. 27 (1998) 579. nol)];
[45] An example of a recent report: Y. Tamaki, K. Hara, R. Katoh, M. Tachiya, A. (f) A.N. Okte, M.S. Resat, Y. Inel, J. Catal. 198 (2001) 172178 [17.1 (1,3-
Furube, J. Phys. Chem. C 113 (2009) 11741. dihydroxy-5-methoxybenzene)];
[46] A recent review on nanostructured titania photocatalysts: X. Chen, S.S. Mao, (g) H. Tada, F. Suzuki, S. Yoneda, S. Ito, H. Kobayashi, Phys. Chem. Chem. Phys.
Chem. Rev. 107 (2007) 2891. 3 (2001) 13761382 [19.7/29.4 (bis(2-dipyridil)disulde)];
[47] (a) F. Amano, O.O. Prieto-Mahaney, Y. Terada, T. Yasumoto, T. Shibayama, B. (h) N.C. Lee, W.Y. Choi, J. Phys. Chem. B 106 (2002) 11818 [18.7 (soot)];
Ohtani, Chem. Mater. 21 (2009) 2601; (i) W.Q. Cui, L.R. Feng, C.H. Xu, S.J. Lu, F. Qiu, Chin. J. Catal. 24 (2003) 937 [8.46
(b) B. Ohtani, F. Amano, T. Yasumoto, O.O. Prieto-Mahaney, S. Uchida, T. (methanol)];
Shibayama, Y. Terada, Top. Catal. 53 (2010) 455. (j) D.V. Kozlov, A.A. Panchenko, D.V. Bavykin, E.N. Savinov, P.G. Smirniotis, Russ.
[48] O.O. Prieto-Mahaney, N. Murakami, R. Abe, B. Ohtani, Chem. Lett. 38 (2009) Chem. Bull. 52 (2003) 1100 [6.3 0.4 (benzene)];
238. (k) A. Mills, G. Hill, S. Bhopal, I.P. Parkin, S.A. ONeill, J. Photochem. Photobiol.
[49] B. Ohtani, O.O. Prieto-Mahaney, F. Amano, N. Murakami, R. Abe, J. Adv. Oxid. A: Chem. 160 (2003) 185 [19 (stearic acid)];
Technol. 13 (2010) 247. (l) R. Al-Rasheed, D.J. Cardin, Chemosphere 51 (2003) 925 [17 0.6 (fumic
[50] Correctly speaking, those descriptions suggest only that it is consistent to acid)];
obtain a linear correlation in a plot assuming n to be 1/2 or 2 for direct or (m) J.C. Garcia, K. Takashima, J. Photochem. Photobiol. A: Chem. 155 (2003)
indirect transition, respectively, and there have been no suggestions for a way 215222 [24.8 (imazaquin)];
to determine the mode of transition. (n) A.E.H. Machado, J.A. de Miranda, R.F. de Freitas, E.T.F.M. Duarte, L.F. Ferreira,
[51] This titania sample was supplied as a Japanese reference catalyst (JRC) TIO-3 Y.D.T. Albuquerque, R. Ruggiero, C. Sattler, L. de Oliveira, J. Photochem. Photo-
from the Catalysis Society of Japan. After calcination at 1173 K for 2 h in air, the biol. A: Chem. 155 (2003) 231 [7.910.5 (organic matter)];
specic surface area was ca. 5 m2 g1 . (o) S. Parra, S.E. Stanca, I. Guasaquillo, K.R. Thampi, Appl. Catal. B: Environ. 51
[52] Mode of transition, direct or indirect, of rutile (anatase) titania crystals seems (2004) 107 [10.9 (atrazine)].
to be still under discussion. Examples of band calculation are: [66] URL: http://rredc.nrel.gov/solar/spectra/am1.5/.
(a) K.-H. Hellwege, O. Madelung (Eds.), LandoltBrnstein Numerical Data and [67] The value 1000 mW m2 is usually used in Japan as the total energy
Functional Relationships in Science and Technology, New Series (Semiconduc- of AM1.5 radiation based on JIS C8911 (URL: http://www.jisc.go.jp/app/
tors), vol. 17, Springer-Verlag, Berlin, 1984, p. 414 [indirect]; pager?id=57836), while the total energy calculated using the table of ASTM
(b) S.D. Mo, W.Y. Ching, Phys. Rev. B 51 (1995) 13023 [indirect]. This suggests is 1001.0 mW m2 . (Total energy claimed in the authors recent review [4],
difculty in discrimination of the mode of transition from results of calculation. 958 mW m2 , as well as the contents of ultraviolet light should be corrected.)
(Band-gap energy estimated by calculation is often unreliable.). The difference is presumably attributable to the difference in instruments used
[53] It seems to be known in the eld of photocatalysis that band-gap energies for the measurements.
of anatase and rutile titania are 3.2 (387 nm) and 3.0 eV (413 nm), respec- [68] A. Mills, J. Wang, J. Photochem. Photobiol. A: Chem. 127 (1999) 123.
tively, though the author does not know by whom and how those data were [69] Watanabe et al. reported similar action spectrum analysis of photoinduced
obtained. degradation of Rhodamine B with a cadmium sulde suspension and pointed
[54] This coefcient is often called shape factor, but the author does not know the out a similar dye-sensitization mechanism:
reason for this. T. Watanabe, T. Takizawa, K. Honda, J. Phys. Chem. 81 (1977) 1845;
[55] FWHM is generally evaluated by measuring width in diffraction angle at the Photocatalytic reaction of MB in aerated titania suspensions was reported in
height of half maximum as the term FWHM means, but there is another way 1937 by a Japanese photochemist: M. Horio, Nihon Gakujutsu Kyokai Hokoku
for evaluation: an XRD peak is reproduced by a Gaussian curve or a triangle 12 (1937) 204 (in Japanese). As far as the author knows, this is the rst report
and then FWHM is measured. on titania photocatalysis.
[56] The original paper on this plot is: G.K. Williamson, W.H. Hall, Acta Metall. 1 [70] The fact that irradiation of a dye solution in the absence of a photocatalyst
(1953) 22. decomposes the dye negligibly has often been described in papers to support
The author has not checked discussion, if any, on this plot following the original negligible photoinduced reaction by photoexcited dye molecules. However,
paper. the photoinduced electron injection requires an acceptor, such as titania, and
[57] In the eld of polymer science, average molecular weight is calculated mainly thereby there are no ideal control experiments to exclude the possibility of
in two modes: number average and weight average. photoinduced electron injection (see Fig. 22).
[58] A problem of this method is that particles not detected in SEM or TEM images [71] When organic dyes themselves are pollutants to be decomposed, visible light-
are not counted. Sometimes noble-metal deposits on solid surface are so small induced, but not photocatalytic, reaction can be a useful technique: X. Chen, Z.
(<1 nm) that they cannot be seen. Zheng, X. Ke, E. Jaatinen, T. Xie, D. Wang, C. Guo, J. Zhao, H. Zhu, Green Chem.
[59] It is clear that measured specic surface area reects surfaces of all particles 12 (2010) 414.
and there is no dependence of sensitivity on particle size. [72] This is based on the assumption of a linear light-intensity dependence of reac-
[60] Y. Nosaka, A. Nosaka, Nyumon Hikarishokubai, Tokyo Tosho, Tokyo, 2004, p. 59 tion rate.
(Japanese). [73] T. Torimoto, Y. Aburakawa, Y. Kawahara, S. Ikeda, B. Ohtani, Chem. Phys. Lett.
[61] J.G. Calvert, J.N. Pitts Jr, Photochemistry, Wiley, New York, 1966, p. 20. 392 (2004) 220, This paper showed that the rate of photocatalytic reaction in
[62] This stoichiometry was claried for the rst time by the authors group and the presence of molecular oxygen, i.e., photocatalytic oxidative decomposition,
reported as: S.-I. Nishimoto, B. Ohtani, H. Kajiwara, T. Kagiya, J. Chem. Soc., may strongly depend on the intensity of light irradiation, while the rate of
Faraday Trans. 1 (79) (1983) 2685. reaction in the absence of oxygen seems to be almost independent of the light
Before publication in this journal, we submitted a part of the experimental intensity (see Fig. 17(b)).
results to Chemistry Letters as the rst paper in the authors career in the eld [74] Strictly speaking, when apparent quantum efciency is discussed, the light
of photocatalysis, but the paper was rejected presumably due to a reviewers intensity should be adjusted to be the same in number of photons, not in energy.
comment that the content had been reported in a Russian journal without [75] In the Japanese Industrial Standards (JIS B7113, discontinued), limiting trans-
showing bibliographic data. (We could not nd this.) To submit the results as a mission wavelength is dened as the center of wavelengths giving 72% and 5%
full paper, additional experiments were carried out to support the stoichiom- transmission. For example, the limiting transmission wavelength for an Asahi
etry. For example, an oxygen-isotope experiment was performed and showed Technoglass L-42 cut-off lter is 420 nm based on the standard, and this lter
the origin of molecular oxygen to be water. As a result, we were able to have the transmits ca. 60% at 420 nm. The number 42 indicates only that 420 nm is the
paper published, and it has been cited more than 100 times, including recent limiting transmission wavelength dened in JIS, and practically irradiation at
citations. This is an example of rejection of a submitted paper not always being >390 nm is made using this lter. Statements such as visible-light irradiation
disadvantageous for the authors. at >420 nm was performed using an optical sharp cut lter seem misleading
[63] Oxygen can be liberated through the formation of hydrogen peroxide by a when L-42 has been used.
two-positive-hole process followed by disproportionation, a dark process, into [76] Sayama et al. reported for the rst time photocatalytic water splitting into
oxygen and water. Overall, however, production of one molecule of oxygen hydrogen and oxygen by an aqueous suspension of two kinds of photocatalysts
requires four positive holes also in this case. under visible-light irradiation: K. Sayama, K. Mukasa, R. Abe, Y. Abe, H. Arakawa,
[64] A recent paper from the authors group showed that P25 and an articial mix- Chem. Commun. 2416 (2001).
ture of isolated pure anatase and rutile (from P25) with amorphous material [77] (a) R. Abe, K. Sayama, H. Sugiura, J. Phys. Chem. B 109 (2005) 16052;
exhibited photocatalytic activities, in several reaction systems, which were (b) R. Abe, T. Takata, H. Sugiura, K. Domen, Chem. Commun. (2005) 3829;
at the level intermediate between those of pure isolated anatase and rutile, (c) H. Higashi, R. Abe, A. Ishikawa, T. Tanaka, B. Ohtani, K. Domen, Chem. Lett.
suggesting no synergetic effect [15]. 37 (2008) 138;
178 B. Ohtani / Journal of Photochemistry and Photobiology C: Photochemistry Reviews 11 (2010) 157178
(d) M. Higashi, R. Abe, K. Domen, Chem. Mater. 21 (2009) 1543; Junction): Y. Sasaki, H. Nemoto, K. Saito, A. Kudo, J. Phys. Chem. C 113 (2009)
(e) R. Abe, K. Shinmei, K. Hara, B. Ohtani, Chem. Commun. (2009) 3577; 17536.
(f) A. Kudo, H. Kato, I. Tsuji, Chem. Lett. 33 (2004) 1534; [82] A.J. Bard, J. Photochem. 10 (1979) 59.
(g) H. Kato, Y. Sasaki, A. Iwase, A. Kudo, Bull. Chem. Soc. Jpn. 80 (2007) 2457. [83] During the course of writing this review, the author found that value of total
[78] It seems to be fallen Z for the author. energy of solar radiation, and hence contents of ultraviolet light, reported
[79] For example, see: R.E. Blankenship, Molecular Mechanisms of Photosynthesis, in the recent review [4] was incorrect due to simple miscalculation (see
Wiley, New York, 2002, doi:10.1002/9780470758472.ch1 (Chapter 1). Ultraviolet light in solar radiation), suggesting we, at least I, have to think
[80] If Gibbs energy change of the overall reaction is negative, i.e., energy releasing, twice.
use of two kinds of photocatalyst, requiring two photons for overall photocat- [84] On the other hand, a hypothesis can be negated logically when at least one
alytic reaction, does not make sense. necessary condition does not hold true.
[81] Recently Kudo et al. reported that a certain combination of two photocata-
lysts under controlled conditions works even without a redox mediator (see