Editorial Welcome to the IEEE Journal of Selected Topics in Quantum Electronics (JSTQE) Special Issue on Hybrid Integration for Silicon Photonics (HISP)

Guo-Qiang Lo, Juejun Hu, Luo Xian-Shu, Joyce K. S. Poon, Dries Van Thourhout, Fengnian XiaIEEE Journal of Selected Topics in Quantum Electronics 2022 · DOI

TL;DR

This special issue of JSTQE highlights the transition of silicon photonics from laboratory curiosity to a commercially mature technology through hybrid and heterogeneous integration. By marrying silicon with III-V materials, 2D materials, and phase-change materials, researchers are now achieving terabit-scale aggregate bandwidths and shipping millions of units annually with high manufacturing yields.

What Problem Does This Paper Address?

While silicon is the gold standard for electronics, it has intrinsic limitations for photonics—most notably, its indirect bandgap makes it a poor light emitter. To build high-performance Photonic Integrated Circuits (PICs), researchers must find ways to integrate "non-silicon" functions like lasers, high-speed modulators, and non-volatile switches onto a silicon substrate.

The challenge isn't just making one device work in a lab; it’s about scalability and packaging. This editorial addresses the need for standardized design rules, reliable 3D packaging schemes, and the integration of diverse materials (like III-V semiconductors and 2D materials) to meet the exploding bandwidth and energy-efficiency demands of modern data centers and emerging applications like LiDAR.

Methods

As an editorial, this document summarizes the diverse methodologies presented across 20 papers (8 invited, 12 contributed). Key technical approaches include:

  • Heterogeneous Integration: Bonding III-V materials (like InAs/GaAs quantum dots) onto silicon wafers to provide optical gain and high-efficiency lasing.
  • 3D Packaging: Utilizing Cu-bumps with a tight 36µm pitch to achieve high-density electrical-optical interconnects.
  • Pilot Line Standardization: The establishment of the PIXAPP pilot line, which provides standardized design rules and Assembly Design Kits (ADKs) to bridge the gap between chip-level design and module-level production.
  • Novel Material Deposition: Integrating Phase Change Materials (PCM) for non-volatile switching and mechanical exfoliation of 2D materials (like $Si_2Te_3$) for high-responsivity photodetectors.

Datasets

Not discussed in the paper. (This is an editorial summary of hardware research rather than a data-driven machine learning study).

Key Insights and Ideas

  • Commercial Viability is Proven: Hybrid integration has moved past the "proof of concept" phase. The editors note that millions of units are now shipped yearly, with laser wavelength control across 300 mm wafers reaching a standard deviation of just 0.2 nm.
  • Packaging as a First-Class Citizen: The success of the PIXAPP project suggests that the "bottleneck" in silicon photonics is shifting from device physics to standardized packaging and assembly.
  • Beyond Datacom: While transceivers remain the primary driver, the integration of Yttria-Stabilized Zirconia (YSZ) and Magneto-Optical (MO) garnets is opening doors for mid-IR sensing and non-reciprocal devices (like isolators) on-chip.
  • The "Chip-let" Paradigm: The editorial anticipates that the "chip-let" trend in CMOS electronics will soon become standard in Optical-Electronic Integrated Circuits (OEIC), allowing for more modular and flexible system designs.

Key Results

  • Terabit Performance: Demonstration of a fully integrated 8 × 25 Gb/s DWDM link on silicon, achieving Terabit/s aggregate bandwidth.
  • Energy Efficiency: Achieving energy consumption as low as ~100 fJ/bit for optical links.
  • High-Density Interconnects: 3D packaging schemes achieving bandwidth densities >1 Tb/s/mm².
  • Laser Efficiency: Short-cavity III-V-on-Si DFB lasers achieving >12% wall-plug efficiency.
  • Extreme Responsivity: 2D-photodetectors ($Si_2Te_3$) showing photo-responsivities of 170.5 A/W at 1310 nm.

Limitations and Future Work

  • Missing Modalities: The editors explicitly mention that this issue lacks coverage of polymer-based devices, which remains an area for future exploration.
  • Market Expansion: While datacom is mature, the technology needs to find its way into "untapped" markets where silicon photonics was previously thought to be impossible or impractical.
  • Non-Reciprocal Components: While MO garnets are mentioned, the full-scale integration of high-performance on-chip isolators remains a significant ongoing challenge for the community.

Summary updated

Abstract

Presents the introductory editorial for this issue of the publication.
BibTeX
@article{lo2022editorial,
  author = {Guo-Qiang Lo and Juejun Hu and Luo Xian-Shu and Joyce K. S. Poon and Dries Van Thourhout and Fengnian Xia},
  title = {{Editorial Welcome to the IEEE Journal of Selected Topics in Quantum Electronics (JSTQE) Special Issue on Hybrid Integration for Silicon Photonics (HISP)}},
  journal = {IEEE Journal of Selected Topics in Quantum Electronics},
  year = {2022},
  doi = {10.1109/jstqe.2022.3164759},
}
Figure