Inverse Design of Substrate from Warpage Surrogate Model Using Global Optimisation Algorithms in Ultra-Thin Packages

The inverse design approach would be advantageous when applied to ultra-thin packages because we can design the packages for an acceptable warpage profile at the start of the design process, instead of only being able to measure the warpage after the parts are built, as is the norm with a conventional design process. A framework for the inverse design of ultra-thin electronic packages is proposed in this work. We start with the design (desired / acceptable) warpage profile and move through the framework to determine the optimum metal densities at different substrate subsections and layers that would ultimately result in the design warpage profile. The framework consists of three main phases - learning the material properties of the substrate, establishing a link between substrate design parameters and warpage and finally carrying out inverse design using a global optimization algorithm. This study utilizes a unique cocktail of machine learning techniques and algorithms to achieve this inverse design goal. Results indicate that the framework can recommend changes to the metal density distribution across the substrate in order to bring about a 20% reduction in warpage.

[1]  Carl E. Rasmussen,et al.  Gaussian processes for machine learning , 2005, Adaptive computation and machine learning.

[2]  Raúl Rojas,et al.  Neural Networks - A Systematic Introduction , 1996 .

[3]  Desmond J. Higham,et al.  Edinburgh Research Explorer Deep learning: an introduction for applied mathematicians , 2022 .

[4]  S. Sitaraman,et al.  Methodology for modeling substrate warpage using copper trace pattern implementation , 2009, 2008 58th Electronic Components and Technology Conference.

[5]  Mingji Wang,et al.  Substrate Trace Modeling for Package Warpage Simulation , 2016, 2016 IEEE 66th Electronic Components and Technology Conference (ECTC).

[6]  Riccardo Poli,et al.  Particle swarm optimization , 1995, Swarm Intelligence.

[7]  Nagarajan Raghavan,et al.  Learning the Stress-Strain Relationships of Ultra-Thin Package Materials using a Bayesian Approach , 2020, 2020 IEEE 70th Electronic Components and Technology Conference (ECTC).