The BSIM3 Model
The BSIM3 model (BSIM = Berkeley Short channel Insulated gate field effect transistor Model) was published by the University of California at Berkeley in July 1993. BSIM3 is a public model and is intended to simulate analog and digital circuits that consist of deep submicron MOS devices down to channel lengths of 0.18 micron. Since this channel length is no longer state-of-the-art for modern MOS devices, the model has been adopted several times to model effects not present in devices with greater channel lengths.
BSIM3 is a physical model with built-in dependencies of important device dimensions and process parameters like the channel length and width, the gate oxide thickness, substrate doping concentration and LDD structures. Due to its physical nature and its built-in geometry dependence, the prediction of device behavior of advanced devices based on the parameters of the existing process is possible. As a further improvement, one set of model parameters covers the whole range of channel lengths and channel widths of a certain process that can be used in circuit designs. Due to the physical meaning of many model parameters, the BSIM3 model is the ideal basis for the statistical analysis of process fluctuations.
BSIM3 can model the following physical effects of modern submicron MOS transistors:
| |
• |
Vertical and lateral non-uniform doping |
| |
• |
Mobility reduction due to vertical fields |
| • |
Carrier Velocity Saturation |
| |
• |
Subthreshold conduction |
| |
• |
Source/drain parasitic resistance |
| |
• |
Drain induced barrier lowering (DIBL) |
| |
• |
Channel length modulation (CLM) |
| |
• |
Substrate current induced body effect (SCBE) |
| • |
Short channel capacitance model |
| • |
Temperature dependence of the device behavior |
For a detailed description of these features, refer to the BSIM3 manual from Berkeley University. You can order this manual from Berkeley or you can get it over the Internet. See References for details.
The BSIM3v3 Modeling Package provides a complete extraction strategy for the model parameters of the BSIM3v3.3.0 model. The extraction routines are based on the BSIM3v3.3.0 device equations to ensure that the extracted model parameters represent as good as possible the original physical meaning. Therefore, no or only a minimum of optimization is needed to get a good fit between measured and simulated device behavior.
The routines of this release refer to version 3.3.0 of the BSIM3 model that was released by University of California at Berkeley in July 2005.
Versions of the BSIM3 Model
University of California at Berkeley released four versions of its BSIM3 model. The first three versions have differences in some model parameters, and the model parameter sets are not compatible.
The following example of the parameter UC, which is a part of the mobility reduction, demonstrates the problem:
In BSIM3v2, the effective mobility  eff was calculated according to the following formula:
In BSIM3v3.2.2, the formula changed to:
It can easily be recognized, that UC has quite different values in both equations.
That means, if BSIM3v2 is implemented in the simulator and the parameter is extracted for BSIM3v3.2.2, the simulation will give catastrophic results (in the case of UC).
Therefore, you must be sure that you use the same version of BSIM3 in both your simulator and your extraction tool.
The latest release, BSIM3v3.3.0 is a minor change to BSIM3v3.2.4 with only a few bug fixes and some enhancements in noise modeling. The model equations used are mainly the same in those versions.
Additionally, a few effects are modeled by introducing the ACNQSMOD as well as the LINTNOI model parameters from BSIM4.
|
