On the origin of the mobility reduction in n- and p-metal–oxide–semiconductor field effect transistors with hafnium-based/metal gate stacks

Abstract

We examine the mobility reduction measured in hafnium-based dielectrics in n-and p-MOSFETs by means of extensive comparison between accurate multi-subband Monte Carlo simulations and experimental data for reasonably mature process technologies. We have considered scattering with remote (soft-optical) phonons and remote Coulomb interaction with single layers and dipole charges. A careful examination of model assumptions and limitations leads us to the conclusion that soft optical phonon scattering cannot quantitatively explain by itself the experimental mobility reduction reported by several groups for neither the electron nor the hole inversion layers. Experimental data can be reproduced only assuming consistently large concentrations of Coulomb scattering centers in the gate stack. However, the corresponding charge or dipole density would result in a large threshold voltage shift not observed in the experiments. We thus conclude that the main mechanisms responsible for the mobility reduction in MOSFETs featuring Hafnium-based high-j dielectric have not been completely identified yet. Additional physical mechanisms that could reconcile simulations with experimental results are suggested and critically discussed.

TCAD modeling challenges for 14nm FullyDepleted SOI technology performance assessment

“…The mobility in high-k/MetalGate FDSOI devices is limited by a complex combination of scattering mechanisms: the presence of high-k dielectrics induces mobility degradation through Remote Coulomb (RC) [17], Remote Phonon (RP) [18] and Remote Surface Roughness scatterings (RSR) [19]. Moreover, phonon and Surface Roughness (SR)-limited mobility are strongly dependent on the channel thickness (Tsi) [20].…”

Quantum Modeling of the Carrier Mobility in FDSOI Devices

“…The mobility in UTBB-FDSOI devices is actually limited by a complex interplay between carrier-phonons interactions [6], [7], front interface roughness (FSR) and back interface roughness (BSR) [8], [9], remote Coulomb scattering (RCS) [10]- [12], and possibly remote phonon scattering (RPS) [12], [13]. The strength of most of these mechanisms depends a lot on the applied front and back gate bias voltages.…”

Remote surface roughness scattering in fully depleted silicon-on-insulator devices with high-κ/SiO2 gate stacks

“…A more detailed comparison with experiment 20 would call for a comprehensive de-embedding of the different scattering mechanisms. 14 Yet we can conclude that RSR, together with RPH 5,6 (not accounted for at present in NEGF), are serious candidates in explaining the trends measured in very thin IL devices at strong inversion. Finally, we would like to discuss the effects of nonparabolic (NP) corrections and exchange-correlation (XC) effects.…”

“…There is no doubt that charges trapped at the SiO 2 /HfO 2 interface or in the HfO 2 layer (Remote Coulomb scattering (RCS)) make a major contribution at weak inversion. [3][4][5] Remote scattering by polar optical phonons (RPH) in HfO 2 is also a serious candidate, especially in thin IL devices. 5,6 Much less attention has been given up to now to scattering by roughness at the SiO 2 /HfO 2 interface or equivalently by IL thickness fluctuations.…”

“…[3][4][5] Remote scattering by polar optical phonons (RPH) in HfO 2 is also a serious candidate, especially in thin IL devices. 5,6 Much less attention has been given up to now to scattering by roughness at the SiO 2 /HfO 2 interface or equivalently by IL thickness fluctuations. At variance with RCS, this mechanism, known as remote surface roughness (RSR) scattering, is expected to be dominant in the strong inversion regime.…”