Analytical modeling of stress-induced leakage currents in 5.1-9.6-nm-thick silicon-dioxide films based on two-step inelastic trap-assisted tunneling

Markus Lenski, Tetsuo Endoh, Fujio Masuoka

Research output: Contribution to journalArticlepeer-review

7 Citations (Scopus)

Abstract

The thickness dependence of stress-induced leakage currents (SILCs) has been investigated for silicon-dioxide films with thicknesses between 5.1 and 9.6 nm. Assuming a two-step trap-assisted tunneling process accompanied by an energy relaxation process of trapped electrons, a set of analytical equations is given, which describes quantitatively the SILC dependence on oxide electric field with trap site location, trapped sheet charge density, and trap state energy as characteristic trap site parameters. Applying this model to the SILC data of 5.1-9.6-nm-thick silicon-dioxide films, the best agreement between experimental and calculated I-V data is achieved by a constant trap state energy of 1.93 eV relative to the silicon-dioxide conduction-band edge. Trap sites are located at 4.24 nm from the gate interface for 6.8-9.6-nm-thick films, while the 5.1 nm film exhibits a slightly different trap site location of 4.08 nm. The trapped sheet charge density Qtrap increases linearly with oxide thickness from -0.34×10-6 to -1.29×10-6 C/cm2. As a result, the thickness dependence of Qtrap suppresses the local tunneling current between the gate injection interface and trap sites by a reduction of the local oxide electric field. This fact explains the decrease of SILC with an increase in oxide thickness.

Original languageEnglish
Pages (from-to)5238-5245
Number of pages8
JournalJournal of Applied Physics
Volume88
Issue number9
DOIs
Publication statusPublished - 2000 Nov 1

ASJC Scopus subject areas

  • Physics and Astronomy(all)

Fingerprint

Dive into the research topics of 'Analytical modeling of stress-induced leakage currents in 5.1-9.6-nm-thick silicon-dioxide films based on two-step inelastic trap-assisted tunneling'. Together they form a unique fingerprint.

Cite this