In this thesis, the physical origin of efficient charge separation in La, Rh co-doped SrTiO3
‘photocatalyst sheet’ electrodes and BiVO4 photoanodes, modified with a co-catalyst (cobalt-iron
Prussian blue), is investigated by means of time resolved in-situ optical spectroscopies and X-ray
photoelectron spectroscopy.
Chapter 1 provides the motivation for these studies and gives a brief introduction to the
field. In Chapter 2, the methods used to produce the materials and devices studied thereafter
are outlined. An overview of experimental methods is then given.
In Chapter 3, the results of spectroelectrochemistry, X-ray photoelectron spectroscopy and
density functional calculations (the latter obtained from collaboration) are combined to produce
a simple model of the band structure of doped SrTiO3 materials. It is found that Rh doping
introduces filled states above the SrTiO3 valence band, as well as a vacant mid-gap level, which
is thought to trap photogenerated electrons. Rh doping also drives a strong downward shift
in the position of the Fermi level of SrTiO3. La co-doping leads to the chemical reduction of
Rh4+ states, which in turn induces a Rh 4d orbital re-organisation. Remarkably, this effect
appears to enable the material to avoid an upward shift in the Fermi level that would otherwise
be expected from chemical reduction.
In chapter 4, the consequences of the change in electronic structure between Rh doped and
La, Rh co-doped SrTiO3 photocatalyst sheet electrodes are explored using time resolved in-situ
optical spectroscopies. Under operational conditions, photogenerated electrons in La,Rh doped
SrTiO3 persist for 10s of seconds. A similar effect can only be observed in Rh:SrTiO3 under a
negative applied potential. This is explained by competing effects of a strong potential drop,
driven by a deeply situated Fermi level, and recombination mediated by the mid-gap state in
Rh doped SrTiO3. In La, Rh co-doped SrTiO3, the mid-gap state is absent, enabling persistent
photogenerated electrons to accumulate across the entire potential window studied.
Finally, Chapter 5 examines the reasons why cobalt-iron Prussian blue modification improves
the efficiency of BiVO4 photoanodes, using transient and photoinduced absorption spectroscopies.
Holes in BiVO4 rapidly transfer to cobalt-iron Prussian blue, leading to hole accumulation at
potentials close to the flat-band potential of BiVO4. Multiply oxidised Prussian blue states
are found to turn over at a similar rate to unmodified BiVO4, despite an apparent 1 eV loss
in energetic driving force. Taken together, these results suggest that cobalt-iron Prussian
blue acts by combining an effective interface for charge separation with energy efficient water
oxidation catalysis. The chapter ends with a comparison to CoOx another well known and
effective co-catalyst, known to exhibit slow hole transfer and water oxidation kinetics despite an
ostensibly similar driving force for hole transfer.