LETTERS

PUBLISHED ONLINE: 29 APRIL 2012 | DOI: 10.1038/NMAT3312

A red metallic oxide photocatalyst

Xiaoxiang Xu, Chamnan Randorn† , Paraskevi Efstathiou and John T. S. Irvine*

Light absorption across the bandgap in semiconductors is

exploited in many important applications such as photo- 0.6

Kubelka¬Munk (a.u.)
voltaics, light emitting diodes and photocatalytic conversion.
Metals differ from semiconductors in that there is no energy 0.5
gap separating occupied and unoccupied levels; however, it

Absorbance (a.u.)
is still possible to excite electrons between bands. This is 0.4
evidenced by materials with metallic properties that are also
strongly coloured. An important question is whether such 0.3 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
coloured metals could be used in light harvesting or similar Energy (eV)

applications. The high conductivity of a metal would preclude 0.2 Sr0.9NbO3

sufficient electric field being available to separate photocarri- Sr0.85NbO3
ers; however, the high carrier mobility in a metal might also fa- 0.1 Sr0.8NbO3
cilitate kinetic charge separation. Here we clearly demonstrate
for the first time the use of a red metallic oxide, Sr1−x NbO3 as 0
an effective photocatalyst. The material has been used under
200 300 400 500 600 700 800 900
visible light to photocatalyse the oxidation of methylene blue Wavelength (nm)
and both the oxidation and reduction of water assisted by
appropriate sacrificial elements. Figure 1 | Ultraviolet–visible absorbance spectra (converted from diffuse
If the stoichiometric cubic strontium niobate (SrNbO3 ) per- reflectance spectra) for Sr1−x NbO3 (x = 0.1, 0.15 and 0.2). Kubelka–Munk
ovskite could be prepared, it would be an archetypal d 1 oxide and transformation of the absorption curves is shown in the top inset. Typical
it might be expected to exhibit interesting electronic properties sintered pellets (Sr0.9 NbO3 ) are shown in the bottom inset.
such as metallicity. Electronic band structure calculations in this
and previous studies on this composition clearly indicate a metallic calculated from the slope of the curve. Thus, these materials are
structure, with the Fermi level located within the conduction band red metallic oxides.
(Supplementary Fig. S1)1 . Materials can only be strongly coloured if they strongly absorb
The stoichiometric SrNbO3 phase with cubic perovskite light, which must promote electrons to create electron–hole pairs.
structure has proved difficult to realize2 , possibly indicative of In a semiconductor or insulator, this is well understood as an
a charge disproportionation phenomenon distorting the cubic electron transition from the essentially occupied valence band to an
lattice; however, cubic perovskites are readily achieved in non- essentially unoccupied conduction band. In a metallic conductor,
stoichiometric compositions. The series Sr1−x NbO3−δ yields cubic the terms valence and conduction band are less easy to apply.
perovskites for 0.1 < x < 0.25 (refs 3,4), for example with unit It is reasonable to consider the partially occupied band as the
cell parameters of 4.019 Å (x = 0.2) and 4.027 Å (x = 0.1). A conduction band, but the valence band (VB) no longer equates
series of strongly coloured red materials with absorption energy with the highest occupied band. It is therefore better to describe
gaps around 1.9 eV (estimated from an absorption tail to 700 nm) the band below the conduction band (CB) as the highest fully
were synthesised for 0.1 ≤ x ≤ 0.2 in this study (Fig. 1). The occupied band and that above it as the lowest unoccupied band,
Fermi level of these materials might be expected to have a small termed as B−1 and B1 , respectively, in Fig. 2. Transitions that might
positive shift as Nb in the structure increases its formal oxidation be associated with photon absorption could be conceived as either
state compared with stoichiometric SrNbO3 , and a decrease in being from B−1 to CB or alternatively CB to B1 , that is, either
energy gap is observed as x increases. These materials were single involving transitions to or from the partially occupied conduction
phase according to X-ray powder diffraction analysis and electron band. In the case of the metallic oxides Sr1−x NbO3 (0.1 ≤ x ≤ 0.2),
microscopy examination. Thermogravimetric analysis showed that the absorption energy gaps (∼1.9 eV) observed in ultraviolet–visible
the phase was stable against oxidation in air up to 300 ◦ C and absorbance spectra are smaller than the indirect bandgap energy
confirmed that δ was generally small. (2.38 eV) predicted by theoretical calculations for B−1 to CB in
Not only are these materials strongly coloured, they have the related stoichiometric phase. As the Fermi level is calculated
high electronic conductivity, exhibiting a conductivity of to be some 1 eV above the conduction band edge1 and bandgaps
3 × 103 S cm−1 for x = 0.2 (Supplementary Table S1). Conductivity obtained from density functional theory (DFT) calculations are
decreased slightly on increasing the temperature from 82 K typically underestimated, especially for semiconductors; it seems
to ambient (Supplementary Fig. S2), consistent with previous unlikely that the transition is from an occupied level to the
studies that showed such an expected metallic signature conduction band (B−1 to CB). Therefore, it seems that the transition
with conductivity decreasing with temperature2,4,5 . A positive is most probably from the conduction band to a higher level
temperature coefficient of resistance (αt = 2.047 × 10−4 K−1 ) is unoccupied band (CB to B1 ). DFT calculations on the related

School of Chemistry, EaSTChem, University of St Andrews, St Andrews, Fife KY16 9ST, UK. † Present address: Department of Chemistry, Faculty of Science,
Chiang Mai University, Muang, Chiang Mai 50200, Thailand. *e-mail: [email protected].

NATURE MATERIALS | VOL 11 | JULY 2012 | www.nature.com/naturematerials 595

© 2012 Macmillan Publishers Limited. All rights reserved
LETTERS NATURE MATERIALS DOI: 10.1038/NMAT3312

the band calculations on the hypothetical SrNbO3 (ref. 1),

the created electrons would be located in the B1 band, with
the ‘holes’ remaining in the conduction band below (CB;
B1
Supplementary Fig. S1).
One of the most interesting properties of semiconductors where
charge carriers can be created by irradiation is photocatalysis of
reactions such as water splitting7–11 . The photocatalytic proper-
ties of these red metallic oxides were therefore investigated and
they were shown to be photocatalytically active using visible light
for oxidation and reduction water splitting processes (Fig. 3). In
both cases, sacrificial elements were used to facilitate the reverse
CB
processes; however, catalyst additives such as Pt (ref. 12), were
Energy

Ef not necessary to promote either oxidation or reduction processes.

The gradual decay of O2 production is due to the reduced Ag
that covers the surface of Srx NbO3 particles, therefore blocking
photons from reaching the surface. Direct water splitting with-
out the aid of sacrificial elements was not achieved, suggesting
the overpotentials for water splitting at this oxide were large
enough to hinder the process. The maximum difference in energy
between photogenerated holes and electrons is around 1.85 eV,
only about 0.6 eV more than the difference in potential between
water oxidation and reduction, which is 1.23 eV at 300 K. The
B–1 strontium niobate photocatalyst was also found to be active for
the oxidative process of methylene blue decolourization (Sup-
plementary Fig. S3).
One of the features of the high-temperature synthesis is
Figure 2 | Schematic of band structure for a metallic conductor. Bands that large particles are produced, the physical properties of
below and above the conduction band (CB) are labelled as B−1 and B1 . which will closely correlate with those of the bulk materials.
Possible transitions associated with photon absorption are marked Unfortunately, the small surface areas of the product will
by arrows. limit photocatalytic performance. This can be overcome by
ball milling the powders after synthesis and loading a suitable
‘SrNbO3 ’ suggest that this type of transition is associated with a co-catalyst such as Pt. The surface areas of the ball-milled powders
similar energy (∼1.8 eV at 0 point) to the absorption energy gaps increased considerably (from 0.2805 m2 g−1 to 11.5562 m2 g−1
observed (∼1.9 eV). These transitions seem to be between bands after ball milling) in accord with the large decrease in particle
with predominant Nb character and are associated with an intense size (Supplementary Fig. S4), which significantly enhanced
colour and sharp absorption (Fig. 1). photoactivity. The hydrogen production rate was increased from
Electron–hole pairs can be created by interband transitions 10.5 ± 0.5 µmol h−1 g−1 photocatalyst to 44.8 ± 5.9 µmol h−1 g−1
where electrons are excited from the conduction band into after ball milling and loading with 1 wt% Pt (Supplementary
the unoccupied energy levels in the higher band. However, Fig. S5). This is an increase by a factor of 4.3. Similar milling
one might assume that even when such electron–hole pairs are treatments of these phases without co-catalyst addition increased
created in a highly conducting, metallic material, the absence of performance by a factor of 3, indicating that both the improved
significant electric field would mean that recombination could surface area and the presence of co-catalyst contribute to
not be avoided. On the other hand, the high conductivity implies this enhancement.
high carrier mobility and this might allow carrier separation, As the effective oxidation state of Nb in strontium niobate
especially if the carriers could be rapidly transferred6 . From is close to four (as confirmed by thermogravimetric analysis
a b 14

40
12

10
30
8
H2 (µmol)

O2 (µmol)

20 6

4
10
Sr0.9NbO3 2 Sr0.9NbO3
Sr0.8NbO3 Sr0.8NbO3
0 0

0 5 10 15 20 0 1 2 3 4 5
Time (h) Time (h)

Figure 3 | Photocatalytic performance of as-prepared strontium niobates under visible light irradiation. a, Typical time courses of H2 evolution in aqueous
oxalic acid solution (0.025 M) with 0.2 g catalyst. b, Typical time courses of O2 evolution in aqueous silver nitrate solution (0.005 M) with 0.1 g catalyst.

596 NATURE MATERIALS | VOL 11 | JULY 2012 | www.nature.com/naturematerials

© 2012 Macmillan Publishers Limited. All rights reserved

NATURE MATERIALS DOI: 10.1038/NMAT3312 LETTERS
a results conclusively prove that this metallic conductor can be
Red– used as a photocatalyst.
The mechanism of operation of this metallic photocatalyst may
CB be considered as being between that of a photoelectrocatalyst
and that of a photocatalyst (Fig. 4). If we consider a normal
Red semiconductor photocatalyst, then once the two photocarriers
hν have been created, the photoreduction and photooxidation
Ox processes involve independent redox reactions of the photocarriers
(Fig. 4a). In a photoelectrochemical cell, photoelectrons or holes
VB are generated in a semiconductor photoelectrode, and these
Ox+ photocarriers drive the primary redox process. The induced
potential in the electrochemical circuit then drives the secondary
redox process at the metal counter electrode (Fig. 4b). In the
b case of the photoactive metallic oxide, it is reasonable to suggest
that photogenerated electrons drive the initial photoreduction
process, with the induced potential within the particle driving the
CB Red–
counter process (that is, the oxidation); this we term as a metallic
e– photocatalyst (Fig. 4c).

hν Methods
Ox Material synthesis. Strontium niobates Sr1−x NbO3 (0.1 ≤ x ≤ 0.2) were synthesised
Red
by calcination of homogenized Sr4 Nb2 O9 , Nb and Nb2 O5 mixtures in a flowing
Ar atmosphere at 1400 ◦ C for 20 h. Details of the synthesis can be found in the
VB Supplementary Information.
Ox+
Physical characterization. Crystal structures were examined by X-ray diffraction
analysis of powders on a SToe STADI/P powder diffractometer. Incident radiation
was generated using a CuKα1 source (λ = 1.54056 Å). The step size for data
collection was 0.02◦ with a collection time 1,000 s at each step. Vaseline was used
to mount the powders onto the holder. Diffuse reflectance spectra were collected
on an ultraviolet–visible spectrophotometer (JASCO-V550). The absorbance was
transformed by the Kubelka–Munk method. X-ray photoelectron spectroscopy was
performed on a Thermo Escalab 250 using a monochromatic Al Kα X-ray source.
All binding energies were referenced to the C 1s peak (284.6 eV) arising from
c adventitious carbon. The d.c. electronic conductivity values were measured on
sintered pellets by a d.c. four-probe technique according to a previously reported
Red– method12 . Particle size analysis was performed on Mastersizer 2000 Particle Size
B1 Analyser (Malvern Instruments).

Theoretical calculations. Theoretical calculations were carried out using the

density functional theory implemented in the Vienna ab initio simulation package
Red
(VASP; ref. 13). The Perdew, Burke and Ernzerhof (PBE) exchange–correlation
hν
Ox functional within the generalized gradient approximation14 and the projector
augmented-wave pseudopotential15 were adopted. A plane-wave basis set with
an energy cutoff of 500 eV was employed. Spin polarization was also taken
CB into account in this calculation. A cubic cell with a = 4.03 Å was used for the
Ox+ simulations. All geometry structures were fully relaxed until the forces on each
atom were less than 0.01 eV Å−1 . Static calculations were done with a 9 × 9 × 9
Monkhorst-Pack k-point grid16 .
Figure 4 | Schematic of possible photocatalytic-type processes to drive Photocatalytic reactions. Photocatalytic activity was evaluated by monitoring
solar redox processes. a, Photocatalytic processes occurring on a single the degradation of methylene blue as well as examining the H2 or O2 evolution
photocatalyst. b, Photoelectrochemical process involving the combination in various aqueous solutions (oxalic acid or silver nitrate solutions). Visible light
of a semiconductor photoelectrode and a metallic counter electrode. irradiation was generated by a 250 W iron-doped metal halide ultraviolet–visible
lamp with an ultraviolet cut-off filter (λ ≥ 420 nm; Borosilicate Coated Glass
c, Photocatalytic process occurring at the partially filled conduction band of
HM07, UQG (optic), Cambridge UK). For methylene blue degradation,
a metallic oxide particle with the counter process occurring at a higher 0.1 g sample powders were suspended in 200 ml of methylene blue aqueous
band. solution (1 × 10−4 M) under air bubbling. The suspension (0.5 g l−1 ) was
first kept in a dark condition for 2 h to reach absorption equilibrium before
and X-ray photoelectron spectroscopy, which also demonstrates irradiation. The photoactivity of the samples was then evaluated by monitoring
the intensity of the strongest absorption peak of methylene blue at 665 nm
that the surface and bulk states are similar; Supplementary on a Perkin Elmer Lambda 35 ultraviolet/visible spectrometer along with
Fig. S6), it might be expected to chemically reduce water; irradiation time. 2 ml aliquots were sampled at various time intervals and
however, this was not found to be a significant process under the initial pH value was kept at 7.0. The temperature during irradiation
the investigated conditions. No hydrogen was detected under was controlled at about 25 ◦ C by means of a water bath to remove infrared
control experimental conditions, even when sacrificial oxalate radiation and heat.
To measure H2 or O2 evolution, 0.1 g or 0.2 g sample powders were
was present and the light source was masked. Activity was suspended in 200 ml oxalic acid aqueous solution (0.025 M) or 200 ml silver nitrate
maintained over several photoreduction cycles with no detectable solution (0.005 M) under visible light irradiation. The suspensions were sealed in a
change in the colouration, morphology (by scanning electron home-made photoreactor purged with pure Ar as a protective gas. The evolved gases
microscopy), oxidation state (by X-ray photoelectron spectroscopy) under irradiation were then examined using an online gas chromatograph (Agilent
and crystal structure of the catalyst (by X-ray diffraction analysis; 3000 Micro Gas Chromatograph). Experiments were performed on samples with
or without Pt as a co-catalyst. Pt was loaded onto the specimen by an impregnation
Supplementary Figs S5, S7, S8 and S9). The number of moles method: the proper amount of H2 PtCl6 aqueous solution was impregnated into
of product significantly exceeded the number of moles of the sample powders and then evaporated at 90 ◦ C until dry. The powders were then
photoatalyst present, confirming photocatalytic behaviour. These heated at 180 ◦ C for 2 h for the decomposition of H2 PtCl6 .

NATURE MATERIALS | VOL 11 | JULY 2012 | www.nature.com/naturematerials 597

© 2012 Macmillan Publishers Limited. All rights reserved
LETTERS NATURE MATERIALS DOI: 10.1038/NMAT3312

Received 13 May 2011; accepted 23 March 2012; published online 11. Mao, S. S., Chen, X. B., Liu, L. & Yu, P. Y. Increasing solar absorption for
29 April 2012 photocatalysis with black hydrogenated titanium dioxide nanocrystals. Science
331, 746–750 (2011).
References 12. Konysheva, E. & Irvine, J. T. S. Evolution of conductivity, structure and
1. Shein, I. R., Kozhevnikov, V. L. & Ivanovskii, A. First-principles calculations thermochemical stability of lanthanum manganese iron nickelate perovskites.
of the elastic and electronic properties of the cubic perovskites SrMO3 J. Mater. Chem. 18, 5147–5154 (2008).
(M = Ti, V, Zr and Nb) in comparison with SrSnO3 . Solid State Sci. 10, 13. Kresse, G. & Furthmuller, J. Efficient iterative schemes for ab initio
217–225 (2008). total-energy calculations using a plane-wave basis set. Phys. Rev. B 54,
2. Peng, N. H., Irvine, J. T. S. & Fitzgerald, A. G. Synthesis and crystal structure 11169–11186 (1996).
of the distorted perovskite Sr0.97 NbO3 determined by high resolution powder 14. Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation
neutron diffraction. J. Mater. Chem. 8, 1033–1038 (1998). made simple. Phys. Rev. Lett. 77, 3865–3868 (1996).
3. Ridgley, D. & Ward, R. The preparation of a strontium–niobium bronze with 15. Kresse, G. & Joubert, D. From ultrasoft pseudopotentials to the projector
the perovskite structure. J. Am. Chem. Soc. 77, 6132–6136 (1955). augmented-wave method. Phys. Rev. B 59, 1758–1775 (1999).
4. Isawa, K., Sugiyama, J., Matsuura, K., Nozaki, A. & Yamauchi, H. Synthesis 16. Monkhorst, H. J. & Pack, J. D. Special points for Brillouin-zone integrations.
and transport-properties of Srx NbO3 (0.75 ≤ x ≤ 0.90). Phys. Rev. B 47, Phys. Rev. B 13, 5188–5192 (1976).
2849–2853 (1993).
5. Hessen, B., Sunshine, S. A., Siegrist, T. & Jimenez, R. Crystallization of reduced
Acknowledgements
We thank the Engineering and Physical Sciences Research Council (EPSRC) and National
strontium and barium niobate perovskites from borate fluxes. Mater. Res. Bull.
Science Foundation (NSF) for financial support and G. Liu (Institute of Metal Research,
26, 85–90 (1991).
Chinese Academy of Science) for assistance with reflectance measurements and X-ray
6. Frame, F. A. et al. Photocatalytic water oxidation with nonsensitized
photoelectron spectroscopy measurements, S. Ni (University of Science & Technology of
IrO2 nanocrystals under visible and UV light. J. Am. Chem. Soc. 133,
China) for theoretical calculations, H. Früchtl and M. Buehl (St Andrews), M. Hoffman
7264–7267 (2011).
(Caltech) and P. Robertson (Robert Gordon University) for helpful discussions.
7. Osterloh, F. E. Inorganic materials as catalysts for photochemical splitting of P.E. thanks Sasol UK and Energy Technology Partnership (ETP) for studentship support.
water. Chem. Mater. 20, 35–54 (2008).
8. Wang, X. C. et al. A metal-free polymeric photocatalyst for Author contributions
hydrogen production from water under visible light. Nature Mater. 8, X.X. performed the synthesis and collected data with C.R. and P.E., and J.T.S.I. analysed
76–80 (2009). the results and wrote the paper.
9. Zou, Z. G., Ye, J. H., Sayama, K. & Arakawa, H. Direct splitting of water under
visible light irradiation with an oxide semiconductor photocatalyst. Nature Additional information
414, 625–627 (2001). The authors declare no competing financial interests. Supplementary information
10. Maeda, K. et al. Photocatalyst releasing hydrogen from water–enhancing accompanies this paper on www.nature.com/naturematerials. Reprints and permissions
catalytic performance holds promise for hydrogen production by water information is available online at www.nature.com/reprints. Correspondence and
splitting in sunlight. Nature 440, 295 (2006). requests for materials should be addressed to J.T.S.I.

598 NATURE MATERIALS | VOL 11 | JULY 2012 | www.nature.com/naturematerials

© 2012 Macmillan Publishers Limited. All rights reserved