The global semiconductor industry is currently undergoing a momentous transformation, fueled by substantial regional investments and unprecedented technological advancements. According to McKinsey, the global semiconductor market will surge to $1 trillion by 2030. Amidst this whirlwind of progress, the need for robust models for semiconductor devices, from transistors to diodes, remains paramount to first-pass design success. As devices become increasingly intricate and small, device modeling engineers face the Herculean task of keeping model libraries up with new process targets, as process specifications mature continually throughout their lifecycle.
To address this critical need, we are excited to introduce the IC-CAP Model Generator Recentering Tool, poised to revolutionize the recentering process by slashing up to 70% model extraction times compared to traditional methods.
Try IC-CAP Model Generator to speed up your modeling projects
IC-CAP Model Generator Re-Centering and Corners Modeling
What is Model Recentering
Model recentering is an intricate process for fine-tuning semiconductor device models to align them with new process targets and specifications. Recentering is crucial to ensure that these models accurately reflect the behavior of devices when changes occur in process technologies, device design, or performance targets.
However, the traditional model recentering approach involves a resource-intensive process that demands time and expertise. Device modeling engineers find themselves in a constant loop of tweaking models, validating them through simulations and data analysis, and repeating the cycle until they get the desired accuracy. The strain placed on the modeling team often impacts the quality of released libraries and results in a rise in customer complaints.
New Challenges in Model Recentering
The necessity of embracing a more effective approach to model recentering arises from a confluence of economic and technical drivers.
With the introduction of the U.S. CHIPS Act and the European Chips Act, leading semiconductor companies like Intel, TSMC (Taiwan Semiconductor Manufacturing Company), and Samsung have all announced record-setting investments in new semiconductor fabs.
This global trend of multi-billion-dollar investments creates a pressing need for accelerated access to more SPICE (Simulation Program with Integrated Circuit Emphasis) models. As a result, device modeling teams are facing mounting pressure to quickly adapt model libraries in response to the evolving semiconductor process specifications that mature over time.
Figure 2. To improve supply chain resilience, multiple initiatives have driven regional investment in semiconductor fabs
On the technological front, the semiconductor industry has embraced new device architectures, materials, and process technologies in the quest for higher power efficiency, faster processing speed, and smaller device sizes. However, these advancements have taken the complexity of device modeling to new heights.
One notable challenge arises from the exponential growth in the number of parameters that must be accurately represented in models. With the traditional step-by-step approaches, extracting a transistor model card from extensive measurement data takes a minimum of several days, often extending into weeks.
Modernizing Model Recentering Flows
For over 20 years, Keysight’s device modeling solutions have empowered customers including Samsung Foundry to improve the efficiency and accuracy of device modeling and characterization processes. To meet customers’ ever-increasing need for speed and handling larger volumes of data, we’re excited to announce that we have created the IC-CAP Model Generator Recentering Tool.
We have reimagined the process of model recentering to enable every modeling engineer to minimize manual, repetitive tasks and get more time to focus on improving the quality of the model. Here are the six key features tailored to achieve this goal.
Effortless target import
No more tedious data wrangling – the recentering tool streamlines the import of target definitions and data from various devices and temperature ranges, ensuring accuracy in aligning with process targets. temperatures, this tool ensures precision in aligning with process targets.
Scaling plots for target FOMs
The new tool makes creating scaling plots of target Figures of Merit (FOMs) against geometry and temperature effortless. It also enables precise adjustments and optimizations, putting the power of injecting your engineering knowledge to improve accuracy firmly in your hands.
Seamless integration
Converting model data from tool to tool could be challenging. This tool seamlessly integrates with the IC-CAP Model Generator, offering a custom extraction flow that leads to thorough verification and comprehensive reporting. It is all about making the process fluid and integrated.
Real-time visualization
After tuning and optimization, the Target Error table comes to life in real-time, offering invaluable insights into how well the model aligns with new process targets.
Statistical corner visualization
In the world of device modeling, consistency is key. To ensure model performance across multiple scenarios, this tool provides robust support for statistical corner visualization.
Enhanced simulation capabilities
For users of Advanced Design System (ADS), we have added more arrows to the modeling quiver – additional 10-pk simulations can enhance the tool’s versatility.
Figure 3. Keysight RF device model generator adopted by Samsung Foundry in 2023
Ready, set — recenter!
The modernized recentering process begins with using transistor models from the current library as a reference point. Then IC-CAP Model Recentering Tool will provide the updated values for Figure-of-Merits (FOMs) to reflect the new process targets necessitated by technological advancements.
By leveraging advanced tuning and optimization, the tool makes precise adjustments to critical model parameters including threshold voltage, cutoff frequency, and other FOMs, ensuring they align flawlessly with the new process targets.
The tools provide the ability to adjust trend plots via real-time visualization. Users can efficiently simulate data points across different geometries and temperatures. This allows the modeling engineers to save 70% compared to traditional step-by-step model extraction.
Conclusion
As the semiconductor industry continues its relentless pursuit of advancement, the modernized recentering process is key to accelerating model generation.
By significantly reducing extraction times and providing a comprehensive suite of features, the IC-CAP Model Generator Recentering Tool empowers modeling teams to adapt rapidly to evolving process specifications, accelerate their design cycles, and unlock the trillion-dollar opportunities that lie ahead.