|
Index |
Doping Issues in Wide Band-Gap SemiconductorsExeter, United Kingdom21-23 March 2001 |
Next |
Lack of control over the conductivity of wide-band-gap semiconductors still presents a serious obstacle to optimum device performance. After describing the first-principles approach we have developed to address these problems, I will focus on three issues: acceptor doping of nitrides, conductivity control in ZnO, and the role of hydrogen in dopant engineering of semiconductors in general. P-type doping of nitrides still relies entirely on magnesium. Our recent investigations indicate that beryllium could be a promising alternative, provided compensation by beryllium interstitials can be brought under control. I will describe two strategies for achieving this, one being post-growth drift of interstitial Be (for which we have generated the required knowledge about its diffusion characteristics), the other the use of H as a codopant to suppress compensation by interstitials. The beauty of using H as a codopant is that it can be neutralized or removed from the p-type layer after growth, resulting in acceptor activation. The challenges involved in using other codopants, such as oxygen, will be discussed. When H is incorporated, acceptor-hydrogen complexes are formed, and their detection forms a powerful means of monitoring the activation process.
We have studied the vibrational modes of these complexes in detail, including anharmonic effects. In the process, we have identified a configuration of the Mg-H complex that is consistent with the recent IR spectroscopy results of Clerjaud et al. [1]. Turning to ZnO, we are confronted with the same problems that GaN faced about a decade ago: lack of understanding about unintentional n-type doping, and inability to achieve p-type. I will summarize our current thinking about the source of n-type conductivity, and the prospects for achieving p-type doping. Finally, since hydrogen plays an important role in many of these phenomena, I will discuss our present understanding of the electronic behavior of hydrogen in semiconductors and oxides in general. Our initially startling discovery that hydrogen acts as a donor in ZnO and in InN can actually be consistently explained in this broader context.
This work was supported in part by ONR (Contract No. N00014-99-C-0161) and by AFOSR (Contract No. F4920-00-C-0019).
[1] B. Clerjaud, D. Cote, A Lebkiri, C. Naud, J. M. Baranowski, K. Paula, D. Wasik, and T. Suski, Phys. Rev. B 61 (2000) 8238.