The impact of gate–image charge on RTS amplitudes in ultra-thin gate oxide n-MOSFETs

Abstract

An experimental study of RTS (random telegraph signal) noise amplitudes in ultra-thin gate oxide n-MOSFETs biased in sub-threshold and linear regions was conducted. At the same time, a modified model integrating effects of trapped charge and its gate image charge on the carrier number and mobility fluctuations in the conducting channel was presented. Compared with the widely used model, the newly built model fits the experimental RTS noise data of ultra-thin gate oxide n-MOSFETs better than the former one. By employing this model in the extraction of dynamic border trap characteristics by RTS noise means in ultra-thin gate oxide n-MOSFETs, a more accurate result can be expected.

A percolation study of RTS noise amplitudes in nano-MOSFETs by Monte Carlo simulation

“…Fig. 5 also gives the fitted line of Zhang's data [15], which were acquired on larger area MOSFETs (L eff  W eff = 90  150 nm 2 ), so demonstrates relative smaller values of DI D /I D , whereas pertains similar trends as our curve.…”

“…The maximum value of L eff  W eff = 30  30 nm 2 n-MOSFETs reaches approximately 42% in subthreshold region according to our simulations. In all the results of our model show 1-2% larger values than those of Asenov's when mobility fluctuations were considered, though the omission of image charge effect and use of random dopant distribution in Asenov's work will cancel out part of the RTS amplitudes induced by mobility fluctuation [7,12,15].…”

“…According to Fermi distribution function, once the energy elevation induced by the trapped charge is larger than 1kT, almost no carries exist in the energy elevation region of the channel. For a nanoscale n-MOSFET with acceptor-like oxide trap, L t can be derived from the following equation [7,8,15] q 4pe ox ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi…”

Lessons Learned from Low-Frequency Noise Studies on Fully Depleted UTBOX Silicon-On-Insulator nMOSFETs

“…The origin of RTN is related to charge capture and emission by a single trap, residing generally in the gate dielectric of a MOS device (10). Ample evidence has been presented that the corresponding normalized amplitude of the drain current fluctuation (∆I D /I D ) can range over several orders of magnitude (16)(17)(18)(19)(20)(21). While this wide variation was puzzling at first, it has become clear that it can be understood in the frame of a non-uniform filamentary channel, defined by the random location of dopant atoms and fixed oxide charges, with a trap in its neighborhood (1)(2)(3)18,19).…”

Low-Frequency Noise Studies on Fully Depleted UTBOX Silicon-on-Insulator nMOSFETs: Challenges and Opportunities

“…11 Ample evidence has been presented that the corresponding normalized amplitude of the drain current fluctuation ( I D /I D ) can range over several orders of magnitude in a large set of similar devices. [17][18][19][20][21][22] While this wide variation was puzzling at first, it has become clear that it can be understood in the frame of a non-uniform filamentary channel, defined by the random location of dopant atoms and fixed oxide charges, with a trap in its neighborhood. [1][2][3]19,20 In this picture, the spread in RTN amplitude is mainly defined by the trap position with respect to the non-uniform potential landscape of the channel.…”