Structure de mise en forme 2 colonnes

WP1 results for M18

BU: BU investigated the influence of additional uniaxial stress on a biaxially strained 100GHz SiGe HBT by physics-based simulation. Results show a significant dependence of maximum fT on uniaxial strain, especially in 010 direction (lateral direction of the intrinsic base). Additional tensile stress reduces peak fT while additional compressive stress increases it.
The same structure has been used to calculate impact ionization rates. The results are as expected with the majority of the impact ionization taking place in collector. The data can be used to benchmark classical device simulation tools. For such tools BU provided band edge data, which is consistent with MC. This solves a previous issue with differences in base potential.
BU also improved our bulk SHE solver to handle band structures fitted to full-band. Resulting distribution functions show good agreement with Monte Carlo simulations and suggest that implementation in a device simulator could be especially advantageous when the electron distribution function at higher energies is of interest. One application is the calculation of impact ionization and breakdown voltage.

UN: MC calibrated transport models (developed in table form by BU) were transferred in analytical form that improves simulation convergence and computation time. A critical overview of HD models was performed. These results are included in the following publication: “Accurate Mobility and Energy Relaxation Time Models for SiGe HBTs Numerical Simulation,” G. Sasso, G. Matz, C. Jungemann, and N. Rinaldi, accepted at SISPAD 2009, San Diego, CA.
Thermal simulations of IFX, ST, IHP devices were completed. Detailed simulations were carried out to investigate the relevance of different technology details: DTI, STI, SiGe layer, base polysilicon, emitter polysilicon, base/emitter contacts. Scaling analysis was also performed to study the scaling behaviour of the thermal resistance (IFX HBT). These results are a useful input to WP4 (RTH model and RTH scaling equations) Transient analysis by numerical thermal simulations also started. The results will also provide an input to WP4 (ZTH(t) model and CTH scaling equations). The results are summarized in D1.3.3.

TUD:  Concerning the scaling effects at the ultimate limit of a SiGe HBT, sensitivity analyses are being performed in order to estimate future HBT profiles and their FoM limits. These investigations are based on the results of TUD’s in house HD simulator and a 1D SHE solver provided by BU in order to take into account non-local effects. For the above mentioned task calibrated physical models are crucial, which are also under development at TUD in based on SHE results.

ST: The main goal for the period was to complete the work related to “FSA SEG architecture exploration and optimization” D112 (M18) report. In addition, continuous support to the advanced modules developed within WP2 was provided. Finally, the exploration of alternative architecture towards B5T architecture started.

IMEC: IMEC has fully updated the process/device simulation files for accurate simulation of the latest fully self-aligned version of their process of reference. The updated 2D TCAD platform accurately reproduces the DC and AC characteristics of state-of-the-art SiGe:C HBTs, with maximum oscillation frequency of 400GHz. The updated TCAD platform is used as the starting point for device architecture exploration of two promising novel FSA architectures conceived for half terahertz RF performance.

IMS: The impact of strain engineering in SiGe HBTs has been investigated by TCAD tools. The proposed structure that was considered includes an extrinsic SiGe layer to introduce external strain. The doping profiles in the emitter, base, and collector regions of the device have been taken from an IMEC bipolar device profile. For the device simulations, specific models for SiGe band-gap, bandgap-narrowing, effective mass, energy relaxation, mobility for hydrodynamic simulation and drift-diffusion simulation have been calculated by BU. Preliminary results show that introducing extrinsic stress layer on our HBT device will enhance maximum oscillation frequency fmax (5%), and cut-off frequency ft (3%) compared to a conventional HBT device without extrinsic stress layer.