Spectral Imaging–Based Optical Loss Mapping for Efficient Failure Analysis and Wafer-Level Screening of Silicon Photonic Integrated Circuits

下午 4:40 - 下午 5:05

With the rapid advancement of Silicon Photonics (SiPh) technology in Co-Packaged Optics (CPO) and high-speed optical communications, testing and failure analysis (FA) of Photonic Integrated Circuits (PICs) have increasingly become critical bottlenecks for yield improvement and mass production deployment. Conventional optical measurement methods typically rely on Fiber Array Unit (FAU) coupling and wavelength scanning techniques to obtain parameters such as insertion loss (IL) and polarization-dependent loss (PDL). However, these approaches suffer from significant limitations in testing efficiency, spatial information acquisition, and defect diagnosis capability, making them inadequate for advanced processes that require fast, large-area, and high-resolution diagnostics.

In this study, we propose a spectral imaging–based optical measurement and analysis method. By utilizing high-sensitivity imaging across the visible to short-wave infrared (VIS–SWIR) spectrum, the method directly visualizes and reconstructs the propagation distribution of light within waveguide structures. Compared with traditional approaches that only provide aggregate loss values, this technique enables spatial mapping of optical loss (optical loss imaging) and effectively reveals the impact of localized scattering, structural non-uniformity, and interface roughness on light propagation.

With submicron spatial resolution and picowatt-level optical sensitivity, the proposed method supports rapid inspection and defect localization at both wafer-level and die-level. It also enables panoramic imaging and multi-wavelength analysis. By correlating optical imaging results with physical failure analysis techniques such as scanning electron microscopy (SEM) and focused ion beam (FIB), we establish a clear relationship between optical anomalies and actual structural defects, confirming that sidewall roughness and non-uniform profiles are the primary sources of scattering and loss.

Furthermore, this work demonstrates the potential of the proposed method for Optical Wafer Acceptance Testing (OWAT). By simultaneously performing optical mapping and defect diagnosis in a single measurement, the testing flow is significantly simplified. Compared with traditional point-by-point scanning, the analysis time per die can be reduced to a few minutes or less, greatly improving throughput while reducing equipment load. In addition, within an Optical Critical Parameter (OCP) pre-screening strategy, this imaging-based approach effectively distinguishes good dies from potentially defective ones, enabling an optimized “test early, skip selectively” methodology.

Moreover, the integration of this technique with Electrical Failure Analysis (EFA) and Optical Failure Analysis (OFA) is explored to enhance the understanding of complex failure mechanisms through multimodal characterization. By incorporating 3D tomography, the method further reconstructs light field distributions at different depths, offering deeper insights into internal scattering, coupling, and radiation behavior within waveguides.

The results demonstrate that spectral imaging–based measurement not only significantly improves PIC testing efficiency but also provides spatial distribution information that conventional methods cannot achieve, thereby strengthening defect diagnosis and process optimization capabilities. This technology offers a high-efficiency, high-resolution, and scalable solution for next-generation silicon photonics measurement and reliability evaluation, and holds significant promise for the mass production of CPO and large-scale photonic integration systems.

Key Technologies Covered

Hyperspectral imaging for multi-location optical signal mapping
High-resolution defect localization and mapping
Wafer-level leakage light and optical loss visualization
Optical loss localization imaging
Absolute optical power distribution mapping
Diagnostic tool for identifying loss sources in OEFA systems
Quantitative analysis module for coupling efficiency at the CPO packaging interface

Featured Speakers

Dr. Joseph Liao

CTO, ENLI TECHNOLOGY CO., LTD.

Dr. Joseph Liao is the founder of EnliTech, a company established in 2009 in Kaohsiung, Taiwan. Initially, EnliTech focused on developing photovoltaic efficiency measurement systems for solar cells. In the optoelectronics and semiconductor industries, EnliTech has successfully transitioned to focus on advanced testing solutions for semiconductor devices, CMOS image sensors (CIS), LiDAR, and silicon photonics (SiPh). EnliTech’s technological strengths are built upon three core capabilities: precise artificial light source simulation that accurately replicates a wide range of illumination conditions—including sunlight, infrared, and LED; high-sensitivity photovoltaic efficiency testing for reliable evaluation of devices such as solar cells and image sensors; and wafer-level, non-destructive optical inspection, enabling high-resolution, contactless measurements during semiconductor fabrication.

As of 2026, EnliTech has actively expanded its silicon photonics portfolio through multiple patent filings and the continued advancement of its silicon photonics inspection platform. This product line highlights the company’s innovative capabilities and reinforces its position in the global market as a leader in next-generation optoelectronic testing solutions.