SBIR Phase II: Innovative Solid-State Phase Change Cooling to Supercharge Central Processing Unit (CPU) Performance

The broader/commercial impact of this Small Business Innovation Research (SBIR) Phase II project is the development of a novel central processing unit (CPU) cooler that will offer enhanced performance for existing and next-generation CPUs. Increasingly, steady-state cooling solutions are unable to keep up with the thermal loads of temperature-sensitive computing and electronic components. As a result, chip performance is artificially limited. By treating the chips as dynamic systems, this project represents a paradigm shift in CPU thermal management. The immediately addressable segment of the CPU cooler market, United States (US)-based sales of CPU coolers to people who build their own machines, is anticipated to reach >$1.5B by 2030. The impact of this new type of CPU cooler will be to enable higher CPU performance of existing CPU technologies, effectively improving computational capabilities in a much more cost-effective way than increasing the number of transistors on a given chip. The broader impact of widespread adoption of this technology would be a reduction in the amount of active cooling needed by, and increase the lifespan of, computer chips thereby saving energy and resources. Further, this effort is synergistic with the broader goal of the US government to onshore semiconductor technology as a matter of national security. The intellectual merit of this project is the development and utilization of metallic, solid-state phase change materials to greatly enhance the thermal storage capacity of cooling systems for CPUs. The key innovations will be the scalable production of these novel phase change materials, the design of a CPU cooler that utilizes them to effectively enhance CPU performance, and the development of a machine-learning algorithm to drive the CPU and cooler in tandem to maximize performance. The material development effort will focus on the casting and rolling of the novel alloys to produce sheet material while maintaining a phase transition temperature within 5 °Celsius (C) of the target. The CPU cooler design and manufacturing tasks will focus on optimizing performance for a small form-factor cooler while allowing for cost-effective production. The control algorithm will modulate CPU power and cooler fan speed to maximize the effective use of the thermal capacity of the cooler. It is anticipated that the cooler and algorithm will reduce throttling events by more than 50% for a given workload, resulting in >20% reductions in processing time, a leap-ahead improvement in thermal management for computing at a fraction of the cost of a similar improvement in chips themselves. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.