The band alignment problem at the Si-high-k dielectric interface
- PDF / 372,462 Bytes
- 6 Pages / 612 x 792 pts (letter) Page_size
- 48 Downloads / 240 Views
E5.6.1
The band alignment problem at the Si-high-k dielectric interface A.A. Demkova*, L.R.C. Fonsecab, J. Tomfohrc, and O.F. Sankeyc a Motorola, Inc. Austin, Texas, USA b Motorola, Inc. Tempe, Arizona, USA c Arizona State University, Tempe, Arizona, USA Abstract We investigate the use of the complex band structure of high-k gate dielectrics to estimate their charge neutrality levels, and compute band offsets to Si. Results of these model calculations are then compared to those obtained with direct electronic structure methods and available experiment. It appears that charge neutrality levels thus obtained indeed provide a consistent picture. However, the uncertainty in the conduction band position inherent in the local density approximation may render the theory inadequate for the engineering support. Despite this limitation, linear re-scaling of the charge neutrality levels based on the experimental band gaps has shown excellent agreement with experimental data.
Introduction To insure continuous downscaling of CMOS technology the semiconductor industry must make a transition from the Si-SiO2-poly-Si triad to a much more complex Si-dielectric-metal system [1]. The higher than silicon dioxide dielectric constant of the new gate dielectric will allow maintaining the gate capacitance and therefore the drain-source saturation current without reducing the oxide thickness. The integration of this new stack into the current CMOS flow is one of the most urgent tasks of today's electronics. The oxide's gate action, among other factors, depends on the barrier height at the oxide-semiconductor and oxide-metal interfaces. The band alignment is often estimated within the so-called metal-induced gap states (MIGS) model [2,3]. The MIGS model describes both the Bardeen and Schottky limits and interpolates between the two in a linear fashion, provided that electron affinities, charge neutralities, and the pinning factor are known. The theory was successfully used to describe the band discontinuity in heterojunctions between covalent semiconductors. It is not obvious whether this approach should work for junctions between Si and high-k dielectrics. A consistent procedure to determine the charge neutrality level is also not clear. The reference potential method of van de Walle and Martin [4] produces reliable valence band offsets, and if the band gaps are known from experiment the conduction band offset can be inferred. However, these calculations are rather time consuming, and extremely sensitive to the exact structure of the interface. It would be very useful to have a simple back of the envelope model to estimate the discontinuity. In this paper, following the recent work of Robertson [5], we apply the simple MIGS model to the Si-high-k dielectric interface. We use the complex band structure to determine the charge neutrality level. We then compare these estimates of the band offset with those obtained via density functional theory (DFT) calculations. We find that the later need to be performed in this case with special care tak
Data Loading...
