Essential tech for optical data comms, sensing and AI hits headwinds

Over the decades, development into silicon photonic modulators – technologically advanced devices that enable the manipulation of light properties – has revealed potential in various network and communications applications, but a recent research programme has found such devices currently face bandwidth limitations and operational robustness issues stemming from the fundamental properties of silicon and other practical constraints.

Optical and photonic modulators are technologically advanced devices that enable the manipulation of light properties – such as power and phase – based on input signals. Over the past few decades, research and development programmes have highlighted how silicon photonic modulators could see use in optical data communication, sensing, biomedical technologies, automotive systems, astronomy, aerospace and artificial intelligence (AI).

Yet the study – The future of optical modulation, published in the IEEE Journal of Selected Topics in Quantum Electronics – noted that while promising, advancements in photonic modulators face challenges like high costs, non-uniformity and lack of standardised processes, necessitating collaboration between academia, industry and global stakeholders to drive innovation, as well as the need to explore alternatives beyond the traditional platforms.

Experts quoted in the report emphasised the necessity of moving beyond traditional platforms such as bulk silicon, indium phosphide and conventional LiNbO₃. They noted that diversity in materials used, fabrication processes and photonic integrated circuit design is a crucial driver for innovation in this field. This shift is leading to the development of new modulator materials, configurations and integration technologies.

These new developments are expected to significantly impact numerous emerging next-generation applications, such as those in datacentres, AI, quantum information processing, augmented and virtual reality, neuromorphic computing, frequency-modulated continuous wave light detection and ranging (Lidar), microwave photonics, as well as metrology and spectroscopy. Notably, thin-film LiNbO₃ modulators show promise for quantum-classical interfaces in superconducting circuits.

Yet the experts also noted several challenges that must be addressed, including technological bottlenecks, high production costs, non-uniformity in devices, substantial time investment, lack of standardised procedures, and more. Overcoming these issues requires the development of a comprehensive co-design capability and platform, which necessitates collaboration among photonic and electrical chip designers, suppliers, foundries, and packaging and testing service providers.

“The unprecedented surge in AI and the current global geopolitical situation have led to increased investment in semiconductor technology by governments, industries and private sectors across all major economic regions in the world,” said Di Liang, a professor at the University of Michigan. “This has resulted in more funding opportunities for collaboration between academia and industry. By overcoming the technical and interest barriers between these sectors and preparing young minds for the challenges ahead, we can ensure a continuous flow of innovations to propel the technology forward.”