The Harvey Rosten Award – 2024 Winner - Simulating the Real World

By Robin Bornoff

The Harvey Rosten Award, sponsored by Siemens Digital Industries Software, recognises outstanding research in electronics cooling and thermal management. Established in memory of Harvey Rosten, a founder of Flomerics and instrumental in the development of the Simcenter Flotherm software, the award celebrates not only technical excellence, but also the curiosity, creativity, and engineering insight that continue to drive innovation in the field. In this series, we speak with past winners about their research, the ideas behind it, and the personal experiences that shaped their journey into electronics thermal engineering. The award has been running since 1997.

2024 Winner

The recipient of the 2024 award was for this publication:

Static and Dynamic Thermal Modelling of Si Photonic Thermo-Optic Phase Shifter

David Coenen¹, M. Kim¹, H. Oprins¹, K. Croes¹, P. De Heyn¹, J. Van Campenhout¹, I. De Wolf¹˒²
¹imec, Leuven, Belgium
²KU Leuven, Leuven, Belgium

2024 23rd IEEE Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronic Systems (ITherm)

DOI: 10.1109/ITherm55375.2024.10709411

An Interview with…

The corresponding author for the winning publication was David Coenen, a senior researcher at imec in Belgium.

Here I ask him some questions about his paper, his research journey, and the ideas and experiences that have shaped his interest in electronics thermal management…

What first sparked your interest in electronics thermal management, or brought you into the field?

During my Master’s thesis I researched thermal management for automotive electric motors. Coincidently I came across a PhD opening at imec that talked about ‘a showerhead for chips’. As a mechanical engineer, this seemed very strange at the time. Intuitively, I would assume that it’s best to avoid splashing water on a computer chip, to avoid short circuits and what not. This was my initial introduction into the world of electronics thermal management. Since then, I have learned a great deal about chip cooling. And apparently the showerhead for chips was a reference to an impingement cooler, attached to the Si back side. This anecdote shows that there still is a gap in education between electrical (electronic) and mechanical engineering that should be bridged if we want to train the next generation of electronics thermal engineers.

Your Harvey Rosten Award recognised your work on Dynamic Thermal Modelling of Si Photonic Thermo-Optic Phase Shifters. Could you briefly describe what you were exploring, and what you found most interesting about it?

This paper is really a collection of results obtained during my Ph.D. concerning thermo-optic phase shifters. Initially, we were trying to validate thermal models with experimental data and exploring the design space with the models to optimize the efficiency of the devices. I would consider this to be a standard methodology. What really sets this work apart, is the section on dynamic modelling.

At one point, one of our partners came to us with the question on how to interpret a certain experimental data set. They presented a large discrepancy between heating and cooling time constants for these devices. From pure thermal physics, we could not explain this. This prompted us to develop non-linear, coupled thermo-optic models. This was challenging, but in the end allowed us to explain the experimentally observed phenomena. As cherry on top, we discovered that many publications erroneously report widely varying heating- and cooling time constants, without proper interpretation.

*(a) 3D view and (b) 2D view of Si photonic Mach-Zehnder modulator equipped with a metal thermo-optic phase shifter.*

Was there a particular moment during that work where things suddenly clicked, or surprised you?

Writing Python code to couple optical and thermal physics is always exciting. Visualizing the simulation results and interpreting the data is a satisfying process if all the puzzle pieces come together.

What do you personally find most engaging or rewarding about working in this area?

Working in electronics (or photonics) thermal management opens many opportunities to collaborate with experts in different fields. Over my (brief) career of about 6 years I have worked with experts on: machine learning, fluid mechanics, photonic device and circuits, packaging, 3D integration… Finally, I must also mention to opportunity to work with students. Teaching and mentoring bachelor, master and Ph.D. level students is part of my job and is a very rewarding activity.

Is there anything about electronics cooling that you think is often misunderstood, or perhaps oversimplified?

If we are talking about the average person, outside of the field, I think the most common misconception is that ‘only a small part’ of the power in a chip is converted into heat. A 100 W chip will dissipate 100 W of heat, from thermodynamics point of view it’s quite simple. Rejecting all this heat to the environment can be a real concern ecologically, for example if the heat is dumped using river water.

A very specific misconception in the industry is that reducing a layer thickness in a chip package will reduce the overall thermal resistance. Intuitively, this would be true, if all heat conduction was purely one-dimensional. However, chip power maps are far from uniform. Consequently, heat spreading occurs mostly in the highly conductive layers. For example, reducing the silicon substrate thickness below a certain threshold can incur a large penalty on heat spreading.

Are there any developments or trends in electronics cooling that you’re particularly interested in at the moment?

At imec we are researching how electronic systems will look like in the next 10-15 years. A key question is how the thermal wall of these systems will be addressed. Two exciting developments are: the integration of heat spreading materials and embedded liquid cooling. Heat spreaders such as AlN, SiC, diamond… will become necessary if we want to keep up with Moore’s law. Embedded liquid cooling, i.e. bringing µchannels with coolant directly inside silicon, shows potential for minimizing the convective thermal resistance, right at the limit with what’s physically possible.

For someone starting out in this field, what would you suggest they focus on, or stay curious about?

For a thermal engineer, it is important to keep up to date with all aspects of thermal management: heat spreaders, thermal interface materials, heat sink design, single and two-phase cooling, immersion cooling, data centre cooling infrastructure… It’s quite a list, but this will enable you to make educated decisions in thermal designs of electronic systems. Don’t hyperfocus on one single aspect.

Has anything outside of engineering, perhaps a hobby or interest, shaped how you think about your work?

Ever since I was little, I have been fascinated by airplanes, rockets… This prompted me to study aerospace engineering. During my studies, I discovered many high-quality scientific Youtube channels, just to learn in a fun and light way. Since then, I have broadened my scope to topics outside of the aerospace industry. Some recommendations: Veratisium, Fluid Mechanics 101, Kurzgesagt, ColdFusion, Steve Brunton, Asianometry.

Is there anything I didn’t ask that you wish I had, and what would have been your answer?

Proposed question: Is it difficult for a mechanical engineer to start working in photonics?

Answer: No! During my studies, I never encountered the topic of photonics. This was not a problem. During engineering studies, you mainly learn problem solving skills, analytical thinking, working together to achieve targets. In the end, you can apply those skills to any kind of problem. For me, this problem happened to be photonics. I know it’s a niche field, to be a mechanical engineer in photonics, but still, I encourage people to get outside of their comfort zone and try something new. The future of advanced chip interconnect is photonic.

Thank you David for your answers!

The Harvey Rosten Award Committee

The committee convenes annually and considers over 40 publications from the previous year that conform to the specific requirements of: context, pragmatism, innovation, broad applicability and accessibility. A long list is identified, whereupon a more rigorous selection criteria is applied to determine a shortlist, from which the committee subsequently identifies a winner via a democratic ballot.

The committee comprises: