Observations of negative bias temperature instability defect generation via on the fly electron spin resonance

Ryan, Jason T.; Lenahan, Patrick M.; Grasser, T.; Enichlmair, H.

doi:10.1063/1.3428783

Cited by 47 publications

(23 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Clearly, the greater the stress temperature the larger the negative V T shift, with a sharp increase observed above 100 °C. These NBTS results are consistent with a previous report on NBTS in Si (11), which attributed an increase in V T instability to an increase in the number of active near-interfacial oxide traps. The time dependence of ∆V T and ∆V GS (taken at I D = 10 -6 A) is illustrated in Figure 6 following either a PBTS or NBTS cycle at 150 °C with V GS = ±15 V for two Group B MOSFETs.…”

Section: Previous Resultssupporting

confidence: 95%

(Invited) Implications for Robust Reliability Testing of Power SiC MOSFETs

2011

View full text Add to dashboard Cite

Threshold voltage (V T ) instability remains an important issue for the performance, reliability, and qualification of large-area SiC power MOSFET devices. The direct application of existing reliability test standards to SiC power MOSFETs can in some cases result in an inconsistent pass/fail response for a given device. Significant variations in the I D -V GS characteristics, with accompanying shift in V T and change in leakage current, can result when these prescribed standards are applied as written. These variations are likely due to the complex time, temperature, and bias dependent nature of the charging and discharging of significant numbers of near-interfacial oxide traps that exists in as-processed SiC MOSFET devices. The significant increase in this V T instability at elevated temperature may suggest the activation of additional performance limiting oxide defects. If the negative shift in V T is large enough, it may increase the OFF-state drain leakage current and limit the long-term device reliability.

show abstract

Section: Previous Resultssupporting

confidence: 95%

(Invited) Implications for Robust Reliability Testing of Power SiC MOSFETs

2011

View full text Add to dashboard Cite

show abstract

“…Their existence and hole trapping properties have been experimentally validated using electron spin resonance (ESR) spectroscopy [6,7]. Oxygen vacancy in SiO 2 has also been theoretically investigated extensively using density functional theory (DFT) [8][9][10].…”

Section: Introductionmentioning

confidence: 99%

Effects of carbon-related oxide defects on the reliability of 4H-SiC MOSFETs

Ettisserry

Goldsman

Akturk

et al. 2014

2014 International Conference on Simulation of Semiconductor Processes and Devices (SISPAD)

View full text Add to dashboard Cite

In this work, we use density functional theorybased calculations to study the hole trapping properties of single carbon-related defects in silicon dioxide. We show that such interstitials are stable in the carboxyl configuration, where the interstitial carbon atom remains three-fold coordinated with chemical bonds to two Si atoms and an oxygen atom (Si-[C=O]-Si). Using formation energy calculations, we observed a +2 to neutral charge transition level for carboxyl defect within the 4H-SiC bandgap. This leads us to propose that carboxyl defects are likely to act as switching oxide border hole traps in the oxide and contribute to threshold voltage instabilities in a 4H-SiC MOSFET. Thus, we provide an additional candidate to the traditional oxygen vacancy hole traps in 4H-SiC MOS systems. The atomic structures of the defect in various charge states are presented. The stability-providing mechanism for the carboxyl defect in the doubly positive state is found to be the puckering of the Si atom, as in the case of positively charged oxygen vacancy hole traps.

show abstract

“…In contrast to other PID publications we do not focus on the PID itself but on the impact on full transistor-lifetime estimations. According to this, we evaluate a large amount of long-term NBTI, PBTI and HCS experiments [3][4][5][6][7][8][9][10][11] instead of fast Fowler-Nordheim stresses as usually optimized for fast-WLR monitoring [12]. We compare parameter degradation at single stress conditions and full lifetime estimations.…”

Section: Introductionmentioning

confidence: 99%