(Invited) Understanding Diffusion, Activation, and Related Phenomena in SiGe Alloys: Models and Challenges

Abstract

An effective diffusivity modeling approach is described that proposes scaling relations for point defect-mediated dopant diffusion to reflect how SiGe with a given Ge concentration differs from Si. Literature results from density functional theory calculations, statistical arguments, and kinetic lattice Monte Carlo simulations can be used to determine the expected scaling relations, which can be verified and adjusted later by comparisons to experimental data. Considerations for targeting of anneals for SiGe and Ge are given. Approaches to optimize annealing using millisecond anneals and nonequilibrium effects are discussed and expectations of behavior in SiGe are outlined. Additional uncertainties related to these SiGe-related behaviors in nanoscale materials are discussed. Improving Understanding via Process Modeling of SiGe MaterialsDevice structures incorporating SiGe within a Si platform have become more commonplace in the last decade (1). Including SiGe is often motivated by its beneficial carrier transport and electronic properties relative to pure Si. Careful control of stress may or may not be a crucial aspect of the engineering, depending on the application. Certainly the type of stress, whether biaxial or approximately uniaxial in nature, is a key determinant of band gap, mobility, and related electronic properties. Process modeling of diffusion, activation, and related phenomena in SiGe must comprehend the compositional and possibly stress states in the SiGe and Si layers. For mainstream use, continuum diffusion models will be the mainstay, but insight may be gained by studies with other approaches such as kinetic lattice Monte Carlo or density functional theory. An overview of relevant physics of SiGe will be discussed as a motivation for our approach at how we are building continuum diffusion models for dopants in SiGe. The traditional approach of acquiring a large database of diffusion experimental data and creating empirical models requires many resources. Here we discuss an approach that attempts to impose more structure on the compositional dependence to initially see the essential features. These forms might then be fitted to experimental data, and details modified, when a sufficiently wide database is available. Subsequently a discussion of annealing and activation of SiGe, including ideas projected from Si millisecond annealing and nonequilibrium approaches to achieving superactivation (activation exceeding solubility), will be discussed. Finally, a discussion of the relevance of nanosizes on diffusion and thermal transport will motivate the need for more attention to this growing area. 10.1149/05009.0233ecst ©The Electrochemical Society ECS Transactions, 50 (9) 233-243 (2012) 233 ) unless CC License in place (see abstract). ecsdl.org/site/terms_use address. Redistribution subject to ECS terms of use (see 138.251.14.35 Downloaded on 2015-03-20 to IP Trends in Physical Chemistry from Si through SiGe to GeWe wish to exploit the similarities between SiGe and Si as much as possible in order to sca...

Continuum modeling of implantation and thermal processes for advanced devices formation

“…Calibrated continuum process models for chemical and stress effects in SiGe have been developed [50]- [53]. For low Ge mole fractions, implantation, diffusion, and activation of SiGe can be treated like in Si with modified parameters depending on Ge content.…”

“…Unlike in Si, the dominating defects are (doubly negative charged) Vac rather than Int. P and As diffusion are only Vac-mediated, and only B is believed to interact much with Int, probably because of its much smaller atomic size [53]. However, for full process modeling of Ge devices, the existing models still lack some effects, e.g.…”