Nanoscale Thermal Transport and Artificial Intelligence for Small and Medium Manufacturing

Ali Shakouri, Ph.D.
Purdue University

Abstract: Thermal dissipation and heating limit the performance of many state-of-the-art electronic and optoelectronic devices and integrated circuits. Full-field thermoreflectance imaging is used to measure the dynamic temperature of active devices with sub-nanosecond time resolution. Image deconvolution techniques are used to achieve far-field sub-diffraction limit resolution. Evidence for quasi-ballistic superdiffusive and hydrodynamic transport of heat is presented. We also describe solid-state microrefrigerators that can selectively cool hot spots. Non-equilibrium electron gas expansion can significantly increase efficiency by decoupling electrons and the lattice.  

The second part of the talk describes the challenges in the digital transformation of small and medium manufacturers. While they constitute more than 98% of US manufacturing, the use of the Internet of Things (IoT) and artificial intelligence (AI) in operations is severely limited. Through an NSF Future Manufacturing project, we are working to establish AI-Commons that bridge multiple sites and companies incentivized information sharing and continuous quality improvement and training.  This addresses a critical shortcoming of the current approach to AI in manufacturing: the limitation of training data to in-company data. Through the ManuFuture Today network, cohorts of manufacturers are working with university researchers to advance their digital transformation.

Bio: Ali Shakouri is a professor of electrical and computer engineering and associate dean of research and innovation at Purdue College of Engineering. He received his Diplome d'Ingenieur in 1990 from Telecom Paris, France, and his Ph.D. in 1995 from the California Institute of Technology in Pasadena, USA. He was a faculty at the University of California in Santa Cruz before moving to Purdue in 2011 to lead the Birck Nanotechnology Center for ten years. He has worked extensively in nanoscale heat transport and electrothermal energy conversion. His group has developed and commercialized a novel lock-in imaging technique to obtain thermal maps of integrated circuits and active devices with submicron spatial and nanosecond time resolution.  He currently leads an NSF Future Manufacturing project focused on deploying artificial intelligence techniques in small and medium manufacturers.