« Previous Up Next »Contents
Previous: 1 The Negative Bias Temperature Instability Top: 1 The Negative Bias Temperature Instability Next: 1.3 Operating conditions causing bias temperature instability

1.2 Historical Background and Relevance

The bias temperature instability (BTI) in MOS devices has been known since the middle of the 1960s [1, 2, 3, 4] and has been attributed to defects located at the Si–SiO $_\mathrm {2}$ interface since the beginning of the 1970s [5]. Since that time BTI has been a vigorously discussed topic in the semiconductor community and numerous papers focusing on measurement techniques, modeling attempts and technology impact circulate in various scientific journals and conference proceeding. The increasing importance of the topic may be illustrated schematically by searching via ‘Google Scholar’ for publications containing variants of the keywords ‘bias temperature stress’, ‘silicon’ and ‘oxide’ itemized by the past decades of years. In Fig. 1.1 the result of such a simple keyword search is depicted as ‘number of Google Scholar Hits’ graded by five year intervals since the 1960s.

Figure 1.1: ‘Google Scholar hits’ of BTI related publications since the 1960s to the present date (2010). The number of hits (column bars) found by ‘Google Scholar’ increase roughly exponentially from the 1960s to 2010 (solid line). Within the past five years between 2005 and 2010 ‘Google Scholar’ lists more than 1200 hits for the selected keywords.

Naturally, the exact number of hits will depend considerably on the selected keywords making their absolute values somewhat arbitrary. Nevertheless, the overall roughly exponentially increasing trend of BTI related publications is likely to be reproduced independently of the subjective choice of the keywords.

The first attempt to describe the BTI phenomenon by a micro-physical degradation model is due to Jeppson and Svensson in 1977 [6] who assumed the surface trap growth as being caused by hydrogen release from Si–H bonds located at the interface. The dynamics of degradation were assumed to be diffusion controlled at low fields and tunneling limited at high fields. Since that time a lot of progress has been made to get the BTI under control. Although, numerous papers have been published on the topic, the issue is still far from being closed. There are a lot of open questions left which have to be resolved in order to understand discrepancies in experimental data and eliminate inconsistencies between suggested models. Since the BTI problem is going to become even more relevant in the future due to continuous processing and scaling changes, a universally valid model including process influences is urgently needed. Besides that a fundamental understanding of BTI may also lead to a better understanding of related degradation mechanisms like hot carrier injection (HCI) or TDDB.