Abstract
Varying the voltage ramp rate during conventional I-V testing allows one to distinguish between metals that react with the surface of a dielectric or barrier (Al), metals that react and can be injected into the dielectric or barrier (Cu), and metals that behave as inert electrodes (Au). By performing experiments over a wide range of ramp rates, one can distinguish between intrinsic breakdown driven by energetic electrons and holes and breakdown that is catalyzed by injected metal ions. The magnitude of the slopes of the I-V traces indicates whether breakdown is intrinsic or catalyzed by metal injection. A mass transfer model describing the drift of copper ions through the dielectric was able to reproduce the broad features of the experimental data. Predictions of the model, including that the slope of the I-V curve should be steeper for metal ion injection, that the breakdown field strengths for all metallizations should converge at very high ramp rates, and that d(ln(t fail))dR≈-1, were confirmed experimentally. Breakdown was shown to be controlled by processes occurring at the anode and differences in the breakdown field strength for the different metals appear to be related to the formation of an interfacial oxide layer between the metal and dielectric.
Original language | English |
---|---|
Article number | 6058641 |
Pages (from-to) | 4354-4360 |
Number of pages | 7 |
Journal | IEEE Transactions on Electron Devices |
Volume | 58 |
Issue number | 12 |
DOIs | |
State | Published - Dec 2011 |
Externally published | Yes |
Bibliographical note
Funding Information:Manuscript received June 15, 2011; revised August 6, 2011 and September 1, 2011; accepted September 6, 2011. Date of publication October 21, 2011; date of current version November 23, 2011. This work was supported in part by the Semiconductor Research Corporation and in part by the Rensselaer Polytechnic Institute. The review of this paper was arranged by Editor B. Kaczer.
Keywords
- Breakdown
- low-k dielectrics
- metal injection
- reliability