Photoelectrochemical cells for hydrogen production from solar energy - Solar Energy Sciences and Engineering Applications - page 301

decrease of TiO 2 bandgap (Wilke and Breuer, 1999). An impressive reduction on the
bandgap of TiO 2 photoanodes was achieved also by Khan et al. (2002) with carbon
incorporation into the TiO 2 -x lattice during heating in a natural gas flame (Khan et al.,
2002). Nevertheless, the studies concerning the doping effect of TiO 2 semiconducting
material do not provide clear conclusions, since the dopant that may have a positive
effect on the bandgap ( E g ) reduction, and thereby increasing the light absorption, has
a negative effect in the energy conversion efficiency (ECE). Moreover, the reduction
of E g should be followed by changing other relevant functional properties. Finally,
the procedures used to incorporate the dopants are often arbitrarily selected. Thus,
without a solid knowledge of the time and the temperature required to incorporate the
dopants, it is truly difficult to replicate the TiO 2 material and to obtain a homogeneous
distribution in the semiconductor (Bak et al., 2002).
The same holds for tin dioxide (SnO 2 ) semiconductor; it has also a large bandgap
in the range of 3.1-3.3 eV that makes this material able to absorb only the UV solar
spectrum (Grimes et al., 2008). Nevertheless, an n-type single crystal of SnO 2 doped
with Sb was investigated by Wrighton and co-workers for H 2 and O 2 production with
an applied bias of 0.5 V under UV light illumination (Wrighton et al., 1976). However,
comparing TiO 2 with SnO 2 , the latter requires a slightly higher potential to achieve
the photocurrent onset (Grimes et al., 2008).
Tungsten trioxide (WO 3 ) is an interesting semiconductor since it has an attractive
bandgap of 2.5-2.7 eV (Grimes et al., 2008; Santato et al., 2001; Butler, 1977). Theo-
retically, a bandgap of 2.7 eV allows use of 12% of the AM 1.0 solar spectrum, a very
high value compared to the barely 4% achieved with TiO 2 (Grimes et al., 2008; Butler,
1977). Although this material had been widely studied by Deb in 1972, it was Hodes
in 1976 that first recognized it as an active visible-light driven photoanode for water-
splitting (Hodes et al., 1976). This material shows good stability in water for pH < 4
and a favorable energy band edge for oxygen evolution (Butler, 1977). Nevertheless,
the minority carrier (hole) diffusion length plays a limiting role in the photoresponse of
tungsten trioxide photoanodes due to the indirect bandgap transition (Solarska et al.,
2012). Usually, WO 3 photoanode is used as thin films and can be found either in the
crystalline or in the amorphous forms. To obtain efficient WO 3 photoanodes, a highly
crystalline structure is desirable since it minimizes the imperfections and the surface
contaminations which may lead to a charge trapping and carrier recombination (Meda
et al., 2010). Recently, a nanostructured WO 3 photoanode has been described capa-
ble of producing a photocurrent of about 3 mA cm − 2 in 3 M CH 3 HSO 3 (AM 1.5 G)

(Solarska et al., 2012).
Alternatively, small bandgap materials can be considered as a starting point in the
research into single-photon systems, such as Fe 2 O 3 or BiVO 4 (Krol et al., 2008; Kay
et al., 2006; Luo et al., 2008; Long and Cai, 2008; Liang et al., 2008; Khan and
Akikusa, 1999). Particular attention has been given to hematite ( -Fe 2 O 3 ) and it has

actually been considered a material with great potential for PEC applications. Hematite
is one of the most abundant and inexpensive oxide semiconductors with an interesting
bandgap of 1.9-2.3 eV, is a non-toxic material and is stable in water (Satsangi et al.,
2010). As a drawback, pure-phase

α

-Fe 2 O 3 has intrinsically poor charge carrier trans-
portation, which limits its quantum efficiency. Moreover, it has poor oxygen evolution
reaction (OER) kinetics and the band edges are not well positioned to directly carry
out the reduction of water. Intensive research efforts have been conducted to improve

α