20nm SRAM Example
This is my favorite part of the year at Coventor: We’re about to do another MAJOR release of SEMulator3D. Developers are sprinting to the finish line, customers are clamoring for the newest features. I’d like to start talking about the new features of SEMulator3D 5.0, but one blog certainly won’t cover it all. Let’s get started, and we’ll do this as many times as we need to get it all written down.
This SEMulator3D release is jam packed with new capabilities. The first thing the user will notice is an all new Process Editor User Interface. But, sitting beneath the new UI are a bunch of new process models. First, we’ve totally reworked the Ion Implantation and Diffusion models to provide process-predictive dopant response. With this change, we also added doping capabilities onto our Deposition and Epitaxy models. And, in order to visualize all these new doping capabilities, we’ve reworked the 3D Viewer with dopant concentration visualization capabilities. In another critical change, we’ve added new capabilities into Etch and Deposition models to account for source visibility, shadowing and off-axis processing. On top of all this, the development wizards have made performance improvements to the modeling engines, improved the data-collection capabilities in Virtual Metrology and Structure Search, added to the DOE capabilities of the Expeditor batch execution engine, and added a bunch of new bells and whistles to the 3D Viewer. Just trying to write this all down is pretty difficult!
It’s simply too much feature content for one blog… so let’s just start with the all-new doping capabilities.
To be clear, this release is not meant to replace the long cycle-time simulations of implant physics that traditional TCAD can offer. However, many challenges in advanced semiconductor technology development do not require traditional TCAD modeling. In SEMulator3D 5.0, we offer new implant models that capture tilt, twist, range, straggle and lateral straggle of any implant species in 3D. We also add a new Diffusion model that captures thermal dopant diffusion in an easily calibrated scheme. With these two new models, technology developers will quickly and easily be able to evaluate the sensitivity of dopant concentrations to process variations and integration changes. This type of fast turn-around 3D process modeling can deliver the type of big data needed to simplify the day-to-day technology development decisions in advanced semiconductor nodes. Just as a sneak peek, I ran our Coventor internal 20nm Planar CMOS SRAM example through SEMulator3D 5.0 to see the doping in the FETs through the flow, including well implants and anneal, halo implants, in-situ doped SiGe PFET source/drain, implanted NFET source/drain and activation RTA. The results are shown below. While the visualization is pretty impressive, it’s the underlying data and responses to process variation that will be of significant value to technology developers across the industry.
OK. I think this is enough for now. But I’ll be writing about some wild new Etch capabilities in SEMulator3D 5.0 next. Stay tuned!