Since the procedure requires irradiation, different to the classical
Mott–Schottky method, it is more correct to replace Efb by
the quasi-Fermi energy of electrons (nEf*). Bard et al. ( 9 )
measured the photocurrent generated in the presence of met-
hylviologen (MV^2 þ) and a reducing agent as function of the pH
value, whereas Roy recorded the photovoltage in the absence of
a reducing agent. In the latter method, the shape of the sigmoi-
dal voltage/pH curve depends on the potential of the reference
electrode, the [MV^2 þ]/[MVþ] ratio, the pH value, k, and on
nEf*. At the pH value of the inflection point (pHo), the quasi-
Fermi level is equal to the one electron reduction potential of
the methylviologen cation. From this the quasi-Fermi potential
of electrons at any pH can be calculated according to Eq. (11)
( 11 ) if the factorkis known. For titania, kis usually equal to
59 mV ( 12 ). Royet al.obtained it from
nE
f ð Þ¼pH E
MV^2 þ=þþkðÞ ðpHopH^11 Þthe slope of the voltage/current plot above the inflection point
whereas Bardet al.calculated it from the slope of the onset of
photocurrent versus pH( 10 ). However, in our hands, both met-
hods afforded only poorly reproducible results due to consider-
able voltage fluctuations and too low photocurrents. We have
determined thekvalue by an alternative method through mea-
suring the pH 0 value not only for one but also for a serious of
redox couples ( 13 ).
With the knowledge of the quasi-Fermi level of electrons, the
level of holes can be estimated by adding the band-gap energy
as obtained from diffuse reflectance spectra. This rough but help-
ful procedure is based on the assumption that both Fermi
levels are located very close to the corresponding band edges.
Since most of the employed powders represent highly doped
semiconductors, this seems a reliable approximation.
Unfortunately, titania can utilize only the very small UV part
(about 3%) of solar light and therefore practical applications like
cleaning of air require the use of UV lamps. However, also the
much larger visible part (l>400 nm) may induce photocatalysis
when titania is surface modified by dyes and transition or main
group elements. Whereas dye sensitization can be an efficient
method in photoelectrochemical devices operating in the absence
of oxygen, it is not suited for photocatalytic aerial oxidation
reactions since the photogenerated reactive oxygen compounds
in general attack not only the substrate but also the dye. Accord-
ingly, we thought that modification by oxidation-stable high-
valent transition metal complexes may be a promising method.
VISIBLE LIGHT PHOTOCATALYSIS 377