Keynote Speakers

Nowadays Wide Bandgap semiconductor devices are getting an increased attention due to some attractive features and superior electrical properties. As the use of these devices is reaching the market, many concerns about their robustness and long term reliability are arising.

In this talk we will focus on the behavior of Silicon carbide power Mosfets during out-of-SOA events such as short circuit or unclamped inductive switching conditions. In the first part of the talk, a detailed approach to self consistent electrothermal simulations of power devices will be given, while in the final part of the talk some results of infrared characterization of short circuit events in power devices will be reported.

The talk will provide an excursus of a career spent within the semiconductor industry, dealing with package design challenges from thermal metrics definition to advanced and sophisticated compact modeling techniques. Moving from the few thermal data available in device datasheet to ambitious multi-million-dollar projects: this is the pendulum where the semiconductor industry has been swinging its thermal design efforts. Still, a kind of thermal version of Moore’s law seems to be emerging, limiting the pace of the next developments in microelectronics. The chiplet and system in package approaches are simply crashing against the basics of heat transmission physics, raising new puzzling questions for system designers. But design feasibility is not the only issue we are facing. The environmental impact of many new power-hungry technologies relies more than ever on efficient thermal solutions. Reuse of thermal energy is a green win-win possibility on the table. Is the microelectronics industry ready and willing to get there?

Direct-to-chip liquid cooling is rapidly emerging as a mainstream solution for managing the escalating thermal loads of high-performance processing units, particularly GPUs, in modern data centers. While single-phase cold plate technologies are relatively well understood and are progressing toward standardization, the adoption of two-phase cold plates presents both opportunities and challenges. By leveraging evaporation and condensation, two-phase systems can achieve significantly higher heat transfer coefficients, enabling more efficient removal of high heat fluxes. However, the intrinsic phase-change processes also introduce complexities, such as elevated pressure drops, flow instabilities, and challenges in system control and reliability. The talk will provide an in-depth overview of the state of the art in two-phase cold plate design, discuss the key technical hurdles to widespread deployment, and explore potential strategies—ranging from optimized flow management to advanced materials and surface engineering—that could unlock the full potential of two-phase cooling in future data centers.