Transient to temporarily permanent and permanent hole trapping transformation in the small area SiON P-MOSFET subjected to negative-bias temperature stress

Abstract

Examining the drain current recovery traces of a small area SiON p-MOSFET subjected to repeated NBTI stress and relaxation cycling reveals direct evidence of transient to permanent hole trapping transformation inferred from previous studies on big area devices. The results show that the emission times of hole traps are not time-invariant (as normally presumed) but can increase due to evolution of the defect sites into more structurally stable forms. In addition, a new type of switching hole traps, exhibiting intermittent charging during stress and occasional increase in emission time by ~5 orders of magnitude, is observed.

On the volatility of oxide defects: Activation, deactivation, and transformation

“…So far, we had only observed defects disappearing into a neutral state. Quite recently, evidence of defects transforming into a quasipermanent positive state was presented [13]. However, no sys- The TDDS Setup: Initially, a reference I D /V G curve for the conversion of ΔI D to ΔV th and a reference trace to check for initial instabilities are measured (IV/Ref).…”

Bias temperature instability study on switching defects in SiON and high-K gate dielectrics

“…After the hole-emission event at monitoring cycle 330, the switching or capture activity of the trap reverts to the previous level of activeness in switching or capturing and emitting a hole consistently in every monitoring cycle. This is the notable evidence substantiating that instead of trap J2 losing the likelihood of capturing a hole under the operating condition during the monitoring cycles, the defect fails to discharge the trapped hole owing to an increase in emission timeengendered by the application of a high VAS at -3 V. This causes a brief absence of emission events from monitoring cycles 301 till 329 after the VAS step[169],[170].…”

“…It has been proposed that the repetitive application of an oxide stress field, coupled with phonon-induced effects, may cause permanent structural relaxation of the oxygen vacancy defects (major source of hole traps) and convert them into stable hole-trapping sites [51], [52], [54]. In recent studies of small-area p-MOSFETs [169], [170], the direct observation of a loss in the ability of a trap to emit the trapped hole after a certain number of stress/relaxation cycles lends strong support to the proposed explanation [51], [52], [54]. hydrogen released from the Si-H bond [59], [101], [131], [148] may be the possible causes in aiding the structural relaxation of trap to become a more permanent charged trap.…”

Oxide-defects driven study on reliability and synaptic response characterization of logic devices

“…Further plausible trap candidates involving hydrogen have also been introduced to clarify the BTI phenomenon and the observed experimental data (i.e., the hydroxyl-E' centers and defects in the hydrogen bridge configuration) [93], [94]. Recently, an examination of the effect of dynamic NBTI on small-area devices by Tung et al [95] has revealed that a defect configuration may transform from a recoverable component (R) of total degradation to a more permanent positively charged state (P). SOTs may also exhibit intermittent charging under the applied stress, in addition to an increased emission time at random time periods.…”

“…The general understanding is that c and e are functions of applied biases and temperature only and are invariant under a given operating condition. Interestingly, it has been found in some studies [88], [95], [126], [127] that accelerated gate stressing, typical during BTI testing, could change the capture/emission characteristics of some switching traps, showing a volatile nature not only in terms of charge occupancy, but also in terms of the atomic structure of the underlying defects. These effects may impact models or applications that presume trap properties are stress invariant.…”