Viewpoint
February 13, 2024

VIEWPOINT 2024: Thomas Rodgers, Senior Director of Market Strategy, Head of Business Sector Electronics, Carl Zeiss Microscopy



VIEWPOINT 2024: Thomas Rodgers, Senior Director of Market Strategy, Head of Business Sector Electronics, Carl Zeiss Microscopy
Thomas Rodgers, Senior Director of Market Strategy, Head of Business Sector Electronics, Carl Zeiss Microscopy
Advancing 3D Packaging with AI-enabled X-ray Microscopy

IC packaging continuously challenges analytical equipment in a virtuous cycle of advancement. Today, the market adoption of advanced 3D packages is rapidly growing to meet performance requirements for artificial intelligence (AI).

The complexity of the latest silicon (Si) and advanced package technology has made the cost of a package failure grow dramatically, while at the same time making some instances of fault isolation and package analysis exceedingly difficult.

These trends will continue in 2024. When packaging the latest Si nodes, new materials and designs require robust physical analysis of chip-package interactions and reliability, to reduce both manufacturing scrap and field returns.

The newly released "3D Heterogeneous Integration" chapter of ASM’s Electronic Device Failure Analysis Society (EDFAS) Roadmap highlights that 3D X-ray microscopy (XRM) is crucial for continuously downscaling high-density through silicon vias (TSVs), bumps and interconnects – structures all common in AI advanced packages.

The failure analysis roadmap cites a "critical need to bring down the 3D X-ray imaging time without sacrificing the resolution of the X-ray image to ensure a highly reliable inspection method." AI itself offers a path to address this challenge in building better AI chips.

ZEISS has responded to the market need by implementing AI in the tomography computation for its 3D X-ray microscopes. Deep learning advanced reconstruction solutions can improve 3D XRM package scan times by 4x with improved contrast-to-noise ratios and better image quality. As reported at 2023 iMAPs, a single 20-hour ground truth 3D XRM scan was used to train a network, enabling a 15-minute scan time to assess global and local alignments in a thermocompression-bonded 15-die stack.

In addition, ZEISS has implemented deep learning to achieve 3D XRM resolution recovery, enabling high resolution across a large field-of-view. These innovations offer a path to enable timely and cost-effective package development and improved manufacturing yields.

Thomas Rodgers, Senior Director of Market Strategy, Head of Business Sector Electronics
Carl Zeiss Microscopy
https://www.zeiss.com/microscopy/en/applications/semiconductors-electronics/advanced-semiconductor-packaging.html
Free Newsletter Subscription
Semiconductor Packaging News is built for professionals who bear the responsibility of looking ahead, imagining the future, and preparing for it.

Insert Your Email Address