Elsevier

Solid-State Electronics

Volume 115, Part B, January 2016, Pages 232-236
Solid-State Electronics

Impact of non uniform strain configuration on transport properties for FD14+ devices

https://doi.org/10.1016/j.sse.2015.08.013Get rights and content

Abstract

As device dimensions are scaled down, the use of non-geometrical performance boosters becomes of special relevance. In this sense, strained channels are proposed for the 14 nm FDSOI node. However this option may introduce a new source of variability since strain distribution inside the channel is not uniform at such scales. In this work, a MS-EMC study of different strain configurations including non-uniformities is presented showing drain current degradation because of the increase of intervalley phonon scattering and the subsequent variations of transport effective mass and drift velocity. This effect, which has an intrinsic statistical origin, will make necessary further optimizations to keep the expected boosting capabilities of strained channels.

Introduction

In the seek of improved performance and lower power consumption in MOSFETs, the 28/20 nm CMOS technology drives at a milestone as the successful bulk technology became deprecated because of the difficulty of fulfilling the requirements given by the ITRS for the forthcoming technological nodes [1]. In particular, standard bulk-MOSFET technology cannot provide satisfactory solutions for sub-22 nm nodes due to the limited control of SCEs and variability problems arising from a highly doped channel [2]. Different technological choices like multigate architectures [3], [4], [5], planar Fully-Depleted Silicon-On-Insulator (FDSOI) [6], [7], [8], [9], [10] or the co-integration of III–V and Ge devices on SOI substrates [11] are competing trying to positioning themselves for the midterm technological scenario. FDSOI devices constitute the natural continuation in planar technology for future nodes thanks to their technological compatibility, outstanding electrostatic control, simpler fabrication process, and competitive overall cost. For the 14 nm node, two generations have been proposed for commercial applications using FDSOI technology. The first one, FD14, will be implemented considering a gate of 20 nm with an overlapped doping profile which yields a shorter effective channel length. The use of a silicon slab as thin as 6 nm and ultrathin buried oxide (UTBOX) with different back bias configurations will ensure a powerful and flexible platform for both high performance (HP) and low power (LP) applications. The evolution of this node, FD14+, will be focused on improved performance rather than in geometrical scaling. In this way, strained channels will be included in order to boost the carrier mobility [12], [13], [14]. However, at such dimensions, the high lateral field and confinement conditions may affect in an important manner the expected transport properties of FDSOI transistors. On the other hand, it is very difficult to induce uniform stress profiles as already shown by structural characterization techniques such as dark-field electron holography [15]. This fact may add a new source of variability with impact on the performance of future devices and circuits. Within this framework, advanced device simulation represents an invaluable tool for the assessment of upcoming technological options in two ways: predicting the performance of different architectures and technological choices; and evaluating optimization options in order to reduce the development stage in terms of cost and time. Another important advantage is the possibility of performing thorough studies of the impact of each physical effect and technological booster separately to explain experimental results and to determine which is the dominant one on the performance of the considered devices. Therefore, the aim of this work is the study, through Multi-Subband Ensemble Monte Carlo simulations (MS-EMC), of the impact of non-uniform strain distributions in FD14+ technology considering different strain configurations.

Section snippets

Simulation methodology

The fundamentals of Multi-Subband (MS) algorithms are based on the mode-space approach of quantum transport [16], which separates the physical description of the simulation domain into a confined and a non-confined problem. In this way, the device is divided in several slices along the confinement direction where the 1D Schrödinger equation is solved. The transport properties are obtained thanks to the solution of the 2D Boltzmann Transport Equation (BTE) in the perpendicular plane (XY in Fig.

Results and discussion

A set of simulations including linear and saturation bias conditions has been performed to determine the importance of non-uniform strain distributions on device performance. ID vs. VGS curves with VDS=100 mV for the different configurations of biaxial strain (sSOI with Si0.8Ge0.2 virtual substrate) are shown in Fig. 3. Two main effects can be observed in this plot. On the one hand, the increase observed in ION level when uniform strain conditions are considered is reduced as non-uniformities

Conclusion

This work presents a MS-EMC study of the impact of non-uniform strain distributions in FD14+ devices. Our simulations show an important degradation of device performance as the strain configuration varies along the channel. This effect becomes of special interest at high drain voltage conditions where the enhancement of intervalley scattering causes a loss of the boosting capabilities of strained channels not only because of the scattering itself but also due to the increase in the transport

Acknowledgments

The authors would like to thanks the support given by Spanish Government (FIS2011-26005, TEC2011-28660), CEI-BIOTIC mP-TIC-12 and Junta de Andalucia (P10-TIC-6902).

References (29)

  • C. Millar et al.

    Accurate statistical description of random dopant-induced threshold voltage variability

    IEEE Electron Dev Lett

    (2008)
  • F.J. Garcia Ruiz et al.

    A comprehensive study of the corner effects in Pi-gate MOSFETs including quantum effects

    IEEE Trans Electron Dev

    (2007)
  • J.-P. Colinge

    FinFETs and other multi-gate transistors

    (2007)
  • Kuhn K, Liu M, Kennel H. Technology options for 22nm and beyond. In: 2010 International workshop on junction technology...
  • Cited by (0)

    View full text