• Issue 2  |
  • January 2010

Bridging the Gap for More Solar Energy

GMC samples

Schematic illustration of non-compensated
anion-cation codoping. Upper panel: Atomic
structure of a non-compensated pair in
anatase TiO2. Lower panel: Formation of
an intermediate band in the gap region via
non-compensated anion-cation codoping.

Converting the abundant energy of the sun into a form convenient for human consumption is the ultimate dream for sustainable generation of environmentally clean energy.  Since the seminal discovery by Fujishima and Honda in the early 1970s (1), titanium oxide (TiO2), an inexpensive white pigment widely used in daily life, has been considered as the most promising photocatalyst for solar energy utilization and environmental cleanup (2). The optimal photocatalyst requires high efficiency while it is also stable against corrosion in the electrolyte. Robustly stable in harsh conditions, TiO2 can absorb photons that in turn generate charge carriers. These charge carriers can be sufficiently energetic to break apart water molecules and produce hydrogen gas. However, TiO2 also has a major drawback. Because its intrinsic band gap is too wide it absorbs only ultraviolet (UV) light that is a small fraction of the sunlight. This results in less than 1% efficiency for converting the energy in sunlight to chemical energy in hydrogen. Thus, reducing the band gap of TiO2 has been recognized as the main avenue for increasing the energy conversion efficiency. A commonly used approach for altering the properties of conventional semiconductors is by doping the host material with foreign impurities (3). Numerous attempts at doping TiO2 including single-element doping and two-element codoping have failed to produce a breakthrough.  This longstanding obstacle is inherently tied to poorly understood fundamental limitations that render thermodynamic solubility of most dopants extremely low. As a result, most of the dopant atoms reside at undesirable interstitial sites, which not only compromise the effectiveness of band gap narrowing but also generate numerous electron-hole recombination centers.

In a forthcoming paper in Physical Review Letters (4), a collaborative team of researchers from Oak Ridge National Laboratory, the University of Tennessee, and Argonne National Laboratory reports on a conceptually novel doping scheme, termed non-compensated anion-cation codoping, for narrowing the band gap of TiO2. An optimal non-compensated anion-cation pair was identified based on density functional theory calculations and validated by experimental studies. A known variant of simultaneous two-element doping uses compensated dopants, i.e. the number of electrons introduced by an n-type dopant equals the number of holes introduced by a p-type dopants, and thus they compensate each other as they form a pair in the host semiconductor. In the non-compensated anion-cation codoping scheme, the researchers intentionally introduce anion-cation dopants that cannot compensate each other. This new approach embodies two key ingredients: (1) the Columbic attraction within the anion-cation dopant pair enhances both the thermodynamic and kinetic solubility, and (2) the non-compensated nature ensures the creation of intermediate bands in the gap region of the host semiconductor, effectively narrowing the band gap. Additionally, the position and the intensity of the intermediate band can be tuned by choosing different combinations of the non-compensated anion-cation pairs and changing the doping concentration.

Following this guiding principle, the researchers have first predicted that Cr-N is an optimal non-compensated anion-cation pair for narrowing the band gap of TiO2 with enhanced thermodynamic and kinetic solubilities. The calculated density of states clearly shows the appearance of new electronic bands in the gap region. In particular, strong hybridization produces new levels that are substantially broader than the weak localized impurity levels found from single element doping.  As a consequence, the intrinsic band gap is reduced, making possible absorption of visible light. Compared with compensated anion-cation codoping, the distinctive merit of non-compensated anion-cation codoping is that the non-compensated nature ensures the formation of extended intermediate bands in the gap region.

To test the idea of the non-compensated anion-cation codoping, the researchers synthesized Cr-N codoped anatase TiO2 nanocrystals using a wet chemical technique and performed a series of delicate experiments to characterize the samples. X-ray photoelectron spectroscopy measurements show that codoping increases the concentration of both Cr and N above that for equivalent single element doping. The optical absorption spectra show an unprecedented magnitude of both absorbance and red shift in the visible light region for the Cr-N codoped TiO2 samples in comparison with single element dopeding. A direct evidence for band gap narrowing was obtained by scanning tunneling spectroscopy (STS) measurements. The STS spectra show that doping with a single element cannot substantially reduce the TiO2 bandgap. In contrast, striking band gap narrowing is observed in Cr-N codoped TiO2 samples that is in good agreement with theoretical prediction. The measurements using low temperature electron paramagnetic resonance (EPR) spectroscopy further demonstrate enhanced photoactivity manifested in efficient electron-hole generation by Cr-N codoped TiO2 irradiated with sub-bandgap light.

Given the potential of the non-compensated anion-cation codoping in controllably tailoring the band gap of semiconductors, this concept may also prove to be instrumental in developing a wide range of advanced materials for catalysis, optoelectronics, spintronics, and clean energy, such as high-efficiency intermediate band solar cells (5) and diluted magnetic semiconductors (6). In a broader context this finding will have deep fundamental and practical importance that may impact condensed matter physics, materials chemistry, and basic energy sciences in facilitating environmentally friendly applications of renewable energy.

(1)   Fujishima and K. Honda, Nature 238, 37 (1972).
(2)   A. L. Linsebigler, G. Lu, and J. T. Yates Jr., Chem. Rev. 95, 735 (1995).
(3)   S. B. Zhang, J. Phys. Condens. Matter 14, R881 (2002).
(4)   Wenguang Zhu, Xiaofeng Qiu, Violeta Iancu, Xing-Qiu Chen, Hui Pan, Wei Wang, Nada M. Dimitrijevic, Tijana Rajh, Harry M. Meyer III, M. Parans Paranthaman, G. M. Stocks, Hanno H. Weitering, Baohua Gu, Gyula Eres, and Zhenyu Zhang, Phys. Rev. Lett. (in press) (2009).
(5)     A. Luque and A. Marti, Phys. Rev. Lett. 78, 5014 (1997).
(6)     S. C. Erwin and I. Zutic, Nat. Mater. 3, 410 (2004).

This work was supported by the Division of Materials Science and Engineering, Office of Basic Energy Sciences, Department of Energy, and in part by the Laboratory Directed Research and Development Program of ORNL.  The calculations were performed at the National Energy Research Scientific Computing Center of DOE.  The EPR experiments were performed at Argonne National Laboratory under DOE BES contract No. DE-AC02-06CH11357.