![]() |
||||
|
Home
Process Simulation
Device Simulation
Interactive Tools
Virtual Wafer Fab
Licensing
Platforms
Services
Design Flows
Technical Library
Downloads and Support
Corporate
Contact Us
Learn more
|
Simulation of ESD Pulse in a MOSFET DeviceATLAS is capable of simulating MOS transistors under transient ESD pulses. Pulses to model standard ESD conditions such as the Human Body Model (HBM) or Charge Device Model (CDM) can be defined in ATLAS. The example presented here is for the HBM. For the HBM, the standard is a 10ns linear current ramp followed by an exponential current drop with a time constant of 150ns. For the CDM a set of passive elements is added in series with the MOSFET. CDM simulations are done using MIXEDMODE. For realistic simulation it is necessary to model the self-heating at high current levels. The local heating has to be combined with temperature dependent models for mobility, recombination and impact ionization. The combination of the complex transient pulse and advanced models means that ESD pulse simulations are computationally intensive. The MOS structure was created using ATHENA. A thermal boundary condition was applied to the substrate including a lumped thermal resistor to represent the heat sinking of the total wafer thickness and package. A high current pulse was applied to the device according to the HBM standard. ATLAS records the peak temperature in the device at each time step of the simulation. This is shown in Figure 1 along with the input current. Due to heat capacity the maximum temperature occurs significantly later than the peak current. Figure 2 shows the location of the hot spot within the device. Figure 1. Peak temperature in MOSFET during an HBM ESD current pulse. Figure 2. Hot spot location
within the MOS device. Typical applications of ESD simulations are to examine the ESD performance of different drain engineering designs, to examine the position of the drain contact with respect to the hot spot and to observe the peak electric field across the gate oxide. This simulation was executed using parallel ATLAS with increasing number of processors. The reduction in execution time vs the number of processors is shown in Figure 3. A high efficiency is seen even though the mesh used is not large. Figure 3. Execution time
improvement with
|
|||
|
© 1984 -
Silvaco Data Systems Inc. -
Trademarks - Privacy Policy
|
||||