Hydrogen in oxide semiconductors

  • PDF / 461,253 Bytes
  • 9 Pages / 584.957 x 782.986 pts Page_size
  • 88 Downloads / 211 Views

DOWNLOAD

REPORT


Oxide semiconductors exhibit a range of physical properties and have potential optical, electronic, and energy applications. Transparent conducting oxides (TCOs) are currently used in products such as flat-panel displays. The prevailing n-type conductivity in these materials has historically been attributed to native defects such as oxygen vacancies. Recent calculations and experiments, however, have provided evidence that native defects are actually not responsible in majority of the cases. Hydrogen, on the other hand, does act as a shallow donor and can dramatically affect the electrical properties of oxides. In addition to contributing to n-type doping, hydrogen also passivates dangling bonds in cation vacancies and passivates acceptor dopants. Some oxides contain “hidden hydrogen,” perhaps H2 molecules, which dissociate at elevated temperatures. In this article, the many roles of hydrogen in zinc oxide, tin dioxide, titanium dioxide, indium (III) oxide, gallium (III) oxide, and strontium titanate are reviewed. The emphasis is on fundamental electronic, structural, and vibrational properties of hydrogen complexes, as determined by experiment and theory.

Address all correspondence to this author. e-mail: [email protected] DOI: 10.1557/jmr.2012.137

front contact in photovoltaic solar cells.9 The market for ITO is so large that demand for indium has caused the price of that scarce commodity to increase significantly. The high cost of indium has stimulated interest in alternative materials such as SnO2, gallium (III) oxide (Ga2O3), and zinc oxide (ZnO).10 The origin of n-type conductivity in TCOs is a subject of controversy.11 Thin films of In2O3 without intentional doping often show free-carrier concentrations in the 1020 cm 3 range.12 The prevalence of such high carrier concentrations has historically been explained by native defects; namely, oxygen vacancies and cation (indium) interstitials. However, the culpability of these “usual suspects” has been challenged by density functional theory (DFT) calculations. For a range of TCOs, these calculations indicate that the oxygen vacancy is a deep donor, whereas interstitials are too mobile to be stable at room temperature.13 As discussed further in this article, hydrogen is a common contaminant that can lead to n-type doping. A second possible source of n-type conductivity is surface defects. Photoemission spectroscopy experiments have shown that the surface Fermi level lies within the conduction band in In2O3.14 This means that donor defects on or near the surface contribute to a highly conducting accumulation layer. A combined theoretical and experimental study suggested that such surface defects might dominate the conduction properties of thin-film In2O3.15 Specifically, an oxygen vacancy on the surface was calculated to have a donor level nearly degenerate with the surface conduction band minimum. Surface conduction layers have also been measured in ZnO.16,17

2190

Ó Materials Research Society 2012

I. INTRODUCTION

Oxide semiconductors and insulators are currently fo