Dielectric Reliability Scaling Trends: Models and Mechanisms

Paul E. Nicollian
Texas Instruments

We review the effects of voltage and thickness scaling on dielectric reliability breakdown models and mechanisms. The understanding and progress in exploiting breakdown physics for improving reliability margins are discussed.

Paul E. Nicollian

Paul E. Nicollian received the B.S. degree in physics from Penn State University (1983) and the M.S. degree in physics from the University of Texas, Dallas (1990). He was employed by Mostek Corp. in 1984. He joined Texas Instruments in 1985, where he is currently a Senior Member of the Technical Staff in the Silicon Technology Development center. His research interests include the reliability physics of ultrathin oxide, nitrided oxide, and high-k gate dielectrics. His contributions include a proof that TDDB is voltage rather than field driven, which helped end the controversy on breakdown models. In the same experiment, he demonstrated that the optimistic anode hole injection model remains a plausible breakdown mechanism at low voltages. He has also shown that the dominant SILC mechanism in ultrathin oxides is due to interface traps below 5V stress, eliminating much of the confusion in analyzing ultrathin oxide stress data. In the area of dielectric technology, he has demonstrated that plasma nitridation overcomes the reliability limitations of oxides, thus enabling a significant extension of the reliability scaling limit of SiO2 based films. He has co-authored 15 publications, holds 2 US patents, and is a recipient of the 2000 IRPS Best Paper Award. He serves on the IRPS Technical Program Committee, and was co-chairman of Device Dielectrics in 2001. Mr. Nicollian is a Senior Member of the IEEE.