/Improving Chip Cooling Using Laser Nano-Patterned Thin Films; Design, Fabrication and Characterization

Improving Chip Cooling Using Laser Nano-Patterned Thin Films; Design, Fabrication and Characterization

Leuven | More than two weeks ago

The Hot AI Chips Need Coolers

Improving Chip Cooling Using Laser Nano-Patterned Thin Films; Design, Fabrication and Characterization

 

Tagline: The Hot AI Chips Need Coolers

 

Abstract:

 

This research aims to develop and optimize advanced cooling techniques for microchips using femtosecond laser nano-patterning of cooling surfaces. The study aims to leverage increased heat transfer by two-phase chip cooling using the properties of nano-textured fluid-solid interfaces and the heat spreading properties of tailored metallic and high thermal conductivity materials for backside cooling and hot-spot mitigation. The project involves (i) thermofluid simulations to optimize the process flow and parameters such as layer thicknesses, bonding approaches, heaters and sensors integration, and topology optimization of nano-patterns (b) fabrication of the nano-patterns using femtosecond lasers considering its process integration constraints and compatibility with CMOS chips (c) testing and characterization using a tailored thermal test chip and thermal rig.

 

Background:

Effective chip cooling is crucial due to the increasing heat generated by powerful microchips, which impacts performance, reliability, and energy efficiency. As devices become smaller and more densely packed, managing heat becomes more challenging. Innovations like direct-to-chip cooling and microfluidics, along with advanced materials and manufacturing techniques, are essential to address these thermal challenges. Efficient cooling solutions not only prevent overheating and damage but also enhance the overall energy efficiency of electronic devices and data centers, supporting the next generation of microchips. For AI chips, effective cooling is even more critical due to their high power consumption and heat generation. Proper cooling ensures AI chips can operate at peak performance, maintain reliability, and support the heavy computational demands of AI tasks.

 

Objectives and Research plan:

 

1.    Design and optimize nano-patterned heat transfer thin film metallic and silicon layers to maximize heat transfer properties e.g. pattern morphology, layer thickness, materials impact. This task will be conducted within the thermal team of imec with support of imec thermal experts. Close cooperation with the manufacturing experts from the mechanical engineering department of KU Leuven is required to take account of manufacturing constraints in the design.

 

2.    Explore the process integration flow for laser nano-patterning using simulations to optimize process compatibility, considering thermal and other constraints that it may impose on manufacturability (e.g. bonding interface roughness and thermal resistance, bonding strength, warpage) and their impact on active devices. This task will be jointly conducted where the thermal-mechanical team of imec will support the student with simulation tasks related to chip process integration and manufacturing such as layer bonding, and the mechanical engineering department of KU Leuven will support the student with simulations aiming at defining laser –nano-patterning strategies having minimal thermal impact on the device so that its functional performance is improved while its structural integrity is maintained.

 

3.    Utilize femtosecond laser nano-patterning to create various surface morphologies and explore the impact of patterning parameters using the simulation findings in 2. This task will be conducted within the mechanical engineering department of KU Leuven, with the support of laser nano-patterning experts from the department.

 

4.    Characterize the chip cooling and hot-spot mitigation using an in-house liquid cooling test rig and custom-fabricated thermal test chip and validate the simulation findings on the cooling performance for pool boiling, flow boiling and impingement boiling conditions. This task will be conducted within the thermal team of imec with support of thermal experts.

 

Expected Outcomes:

Optimized nano-patterned thin films and their CMOS chip compatible integration technology to enhance reliable two-phase direct liquid cooling and heat spreading of thermally critical chips such as AI chips.

 

Required background: Students with science and technology backgrounds e.g. physics and engineering science / technology with an interest in both modelling and experimental research are encouraged to apply. The candidate should have a collaborative mindset and be willing to cooperate with experts with multidisciplinary backgrounds.

 

Type of Work: 40% design and modelling, 20% experimental characterization, 40% prototype manufacturing.

 

Supervisory team:

Promotors: Prof. Dr. Ir. Houman Zahedmanesh (promotor) Prof. Dr. Ir. Sylvie Castagne (co-promotor), Dr. Ir. Heman Oprins (co-promotor)

Daily Advisors: Ir. Georg Elsinger, Dr. Ir. Valdimir Cherman

 

 

References:

[1]https://www.imec-int.com/en/imec-magazine/imec-magazine-february-2019/a-cold-shower-for-chips

[2] How AI is Bringing Liquid Cooling into Chip Manufacturing. https://www.nvent.com/en-us/resources/news/how-ai-is-bringing-liquid-cooling-into-chip-manufacturing.

[3] Cooling High Power Dissipating Artificial Intelligence (AI) Chips Using .... https://www.scirp.org/journal/paperinformation?paperid=135386.

[4] Enhanced nucleate boiling of Novec 649 on thin metal foils via laser-induced periodic surface structures, Applied Thermal Engineering, Volume 236, 2024, 121803, https://doi.org/10.1016/j.applthermaleng.2023.121803.

[5] Webpage of the Laser Micromachining Group @ KU Leuven,

https://www.mech.kuleuven.be/en/research/Femto/research

 

 

 

 

 

 

 



Required background: Science and technology backgrounds e.g. physics and engineering science / technology with an interest in both modelling and experimental research and with a collaborative mindset.

Type of work: 40% design and modelling, 20% experimental characterization, 40% prototype manufacturing

Supervisor: Houman Zahedmanesh

Co-supervisor: Herman Oprins

Daily advisor: Vladimir Cherman, Georg Elsinger

The reference code for this position is 2025-184. Mention this reference code on your application form.

Who we are
Accept marketing-cookies to view this content.
Cookie settings
imec inside out
Accept marketing-cookies to view this content.
Cookie settings

Send this job to your email