Glass-network-dominated instability of laser modification and sidewall undulation formation in through-glass vias fabricated by laser-induced deep etching
Year: 2026
Type: article
Abstract: Through-glass vias (TGVs) fabricated by femtosecond laser-induced deep etching (LIDE) are critical vertical interconnect structures for glass-based heterogeneous integration. However, severe sidewall undulation and excessive roughness remain major challenges that degrade metallization reliability, electrical performance, and thermomechanical stability. In this work, eight commercially available glass substrates with different silica contents and non-bridging oxygen (NBO) concentrations, were systematically investigated to reveal the material-dependent origin of TGV sidewall undulation. The morphology of TGVs and embedded laser-modified channels was quantitatively characterized using multiscale microscopy. SEM, EDS, and Raman spectroscopy were further employed to correlate laser-induced structural evolution with glass-network characteristics and compositional heterogeneity, and the effects of laser pulse energy and focus offset on the development of distorted modification regions were systematically examined. Beyond confirming that high-NBO glasses are prone to filamentation instability, this work provides three specific advances: (i) a quantitative correlation between the Raman-derived fraction of depolymerized structural units and TGV sidewall roughness, established under identical LIDE conditions; (ii) the identification of filament-associated crescent-shaped microvoids and their tentatively inferred laser-induced compositional heterogeneity as the structural features that distort modification channels and drive nonuniform etching; and (iii) a pulse-energy regulation strategy that suppresses the axial propagation of unstable filamentary modification. These results connect glass chemistry, nonlinear laser–matter interaction, and selective etching kinetics within a single framework and provide a material-selection and process-optimization route toward low-roughness TGVs for advanced semiconductor packaging.
Institutions: Chinese Academy of Sciences, Institute of Microelectronics
Language: English
Subfield: Computational Mechanics
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $3,540
Funder: Chinese Academy of Sciences
Award: G2685239806
OpenAlex ID: W7214447491
