SandBox provides software and services to improve recipe development for next-generation chips. Our software suite SandBox Studio™ uses a combination of proprietary physical models and machine learning to dramatically increase the speed of recipe development.
The challenges incurred during etch and deposition recipe development are significant barriers for the manufacturing of next generation technologies focused on device integration or the extension of Moore’s Law. Today’s leading-edge 3D NAND flash memory devices are expected to hit their ceiling at or near 128 layers limiting achievable memory densities. SandBox aims to remove this ceiling by using advanced physical and statistical models to virtually simulate optimal processing conditions.
Current etch and deposition equipment used in industry rely on careful optimization of dozens of variables such temperature, pressure, and gas chemistry. To achieve the critical dimensions required for next generation devices, each of these process knobs must be precisely tuned for a given application.
With the analytics capabilities in SandBox™ Studio, users can run predictive models to make develop and optimize deposition and etch recipes. A unique feature of SandBox™ Analytics is its ability to rapidly calibrate models based on a physical framework to new incoming data. This has been experimentally validated to enable engineers to identify optimal process parameters to achieve a given target objective at a fraction of the time and cost of conventional methods.