Comprehensive Failure Analysis and Mitigation of Glass Core Substrate Defects for Advanced AI Server Packaging
Comprehensive Failure Analysis and Mitigation of Glass Core Substrate Defects for Advanced AI Server Packaging
Thursday, October 8, 2026: 10:20 AM
Summary:
Glass core substrates are attracting significant attention for large AI data server packages due to their superior dimensional stability and reduced warpage in large chiplet integration. However, glass-specific defect mechanisms present critical challenges for achieving high reliability. This study investigates three major defect modes: (1) microcrack formation during TGV (Through-Glass Via) processing, (2) inter-via cracking after metallization, and (3) backside cracking after build-up formation. For TGV-induced microcracks, internal via surfaces after laser modification and etching were analyzed using holotomography, enabling identification of crack initiation mechanisms and process optimization. Inter-via cracks were quantitatively evaluated using automated optical inspection (AOI). Thermo-mechanical modeling revealed stress concentration between vias, and a stress buffer layer with optimized elastic modulus and thickness was introduced to suppress crack formation. The effectiveness was confirmed through reliability tests such as temperature cycling tests (TCT) and highly accelerated stress tests (HAST). For backside cracking, a design window based on the coefficient of thermal expansion and elastic modulus of build-up materials was established, successfully eliminating cracks after TCT.
Glass core substrates are attracting significant attention for large AI data server packages due to their superior dimensional stability and reduced warpage in large chiplet integration. However, glass-specific defect mechanisms present critical challenges for achieving high reliability. This study investigates three major defect modes: (1) microcrack formation during TGV (Through-Glass Via) processing, (2) inter-via cracking after metallization, and (3) backside cracking after build-up formation. For TGV-induced microcracks, internal via surfaces after laser modification and etching were analyzed using holotomography, enabling identification of crack initiation mechanisms and process optimization. Inter-via cracks were quantitatively evaluated using automated optical inspection (AOI). Thermo-mechanical modeling revealed stress concentration between vias, and a stress buffer layer with optimized elastic modulus and thickness was introduced to suppress crack formation. The effectiveness was confirmed through reliability tests such as temperature cycling tests (TCT) and highly accelerated stress tests (HAST). For backside cracking, a design window based on the coefficient of thermal expansion and elastic modulus of build-up materials was established, successfully eliminating cracks after TCT.