Skip to content
DraconiticResearch
Search Sources API
203,203 studies31 sourcesUpdated 3 min ago
Peer-reviewedCC BY 4.0Materials Science in Semiconductor Processing

Terahertz nondestructive evaluation of heterogeneous packaging stackup: Multi-layer interposer

Abstract

With increasing integration density of semiconductor devices, high-speed interconnects and advanced packaging techniques are essential to meet the growing demands of artificial intelligence and large-scale computing. Three-dimensional packaging, where multiple chips are stacked to improve bandwidth and reduce latency have emerged as a promising solution; however, the complex multilayer integration introduces significant manufacturing challenges, including defects such as misalignment, voids, and warpage. Ensuring high manufacturing yields and product reliability requires advanced inspection techniques. In this study, we explore the potential of terahertz electromagnetic waves as a nondestructive-evaluation method for an 8-layer interposer. The effect of polarization on the ability to detect features at depth is discussed. With deconvolution and unsupervised learning algorithms, several defect areas are revealed.

Related

  1. Packaging and long-term temperature-cycling evaluation of a nine-channel polymer AWG for O-band LAN-WDM
    Shangci Yang, Daming Zhang, Bing Li +4
    Optics ExpressPeer-reviewedPublisher termsOptics & photonicsCrossref
  2. Mitigating Sensor Cracking using the Interposer Strategy in the ATLAS ITk Strip Detector
    Syed Haider Abidi, Kirsten Marty Affolder, Anthony Allen Affolder +443
    JINSTPeer-reviewedarXiv licenceFull textApplied physics & instrumentationHigh-energy & particle physicsHALarXiv
  3. Nondestructive 3D Metrology of Through-Glass Via Geometry and Internal Wall Texture Using Holotomography
    Dohyeon Lee, Herve Hugonnet, YongKeun Park
    arXivPreprintCC BY 4.0Full textOptics & photonicsarXiv
  4. Compact nondestructive generator for microsecond-class ultrafast pulsed magnetic fields
    Y. Ishii, A. Ikeda, H. Sun +2
    arXivPreprintarXiv licenceFull textApplied physics & instrumentationCondensed matterarXiv
  5. Nondestructive 3D Quantification of the Magnitude and Extent of Indentation-Induced Densification in Glass
    Sang-min Lee, Dohyeon Lee, Juheon Lee +2
    arXivPreprintCC BY 4.0Full textOptics & photonicsarXiv
BibTeX
@article{shi2026terahertz,
  title = {{Terahertz nondestructive evaluation of heterogeneous packaging stackup: Multi-layer interposer}},
  author = {Haolian Shi and Christopher Blancher and Meghna Narayanan and Pragna Bhaskar and Erwan Emile and Mohanalingam Kathaperumal and Alexandre Locquet and D.S. Citrin},
  journal = {Materials Science in Semiconductor Processing},
  year = {2026},
  month = oct,
  doi = {10.1016/j.mssp.2026.111189},
  url = {https://hal.science/hal-05785326v1},
  note = {CC BY 4.0},
}
RIS
TY  - JOUR
TI  - Terahertz nondestructive evaluation of heterogeneous packaging stackup: Multi-layer interposer
AU  - Haolian Shi
AU  - Christopher Blancher
AU  - Meghna Narayanan
AU  - Pragna Bhaskar
AU  - Erwan Emile
AU  - Mohanalingam Kathaperumal
AU  - Alexandre Locquet
AU  - D.S. Citrin
PY  - 2026
DA  - 2026/10/05
JO  - Materials Science in Semiconductor Processing
DO  - 10.1016/j.mssp.2026.111189
UR  - https://hal.science/hal-05785326v1
AB  - With increasing integration density of semiconductor devices, high-speed interconnects and advanced packaging techniques are essential to meet the growing demands of artificial intelligence and large-scale computing. Three-dimensional packaging, where multiple chips are stacked to improve bandwidth and reduce latency have emerged as a promising solution; however, the complex multilayer integration introduces significant manufacturing challenges, including defects such as misalignment, voids, and warpage. Ensuring high manufacturing yields and product reliability requires advanced inspection techniques. In this study, we explore the potential of terahertz electromagnetic waves as a nondestructive-evaluation method for an 8-layer interposer. The effect of polarization on the ability to detect features at depth is discussed. With deconvolution and unsupervised learning algorithms, several defect areas are revealed.
SN  - 1369-8001
ER  -