Laser focus on tomorrow’s optical transceivers

Benjamin Wohlfeil
Data center

Integrating a laser onto a silicon electro-photonic integrated chip is the current holy grail in the optical transceiver space. When realized, these advanced semiconductors will help the DCI networking industry tackle critical issues around cost-efficiency, power consumption, and thermal management for optical transceivers. But the addition of a laser to silicon chips that already feature monolithic integration of electronics and photonics presents several challenges. That’s why a multi-disciplinary team of engineers from across the supply chain are combining skills and resources under the banner of the Photonic Embedding of Active Region Laser Chips in Silicon (PEARLS) project. 

Gathering SPEED

To understand how far things have already come and to get an idea of the scale of the challenges ahead, we can rewind to 2016 and the start of the Silicon Photonics Enabling Exascale Data Networks (SPEED) project. Led by ADVA, SPEED’s objective was to combine electronic and optical functions onto a single semiconductor chip. These electro-photonic integrated circuits (ePICs) would be far more economical, improve bit-rate, and reduce the footprint of optical transceivers. It took us and our partners three years of hard work and expert input, but SPEED’s goal was realized and the outcome was a giant leap forward for optical transceiver technology. 

PEARLS is a three-year collaboration bringing together ADVA, FormFactor, Fraunhofer IZM, IHP, IHP Solutions, Sicoya, Technion in Israel, and the University of Kassel, to build on the game-changing technology the SPEED project made possible. It’s the logical next step and aims to once again dramatically improve the semiconductor technology used in current transceiver modules. 

Joining the dots

At the top of the list of challenges the PEARLS engineers will face is the problem of cooling a laser operating on such a small scale. The plan is to use quantum dots, tiny nanocrystals that can withstand a range of temperature extremes. In contrast with lasers that employ quantum wells, quantum dots do not need a thermal-electric cooler and so this will allow the extra space needed to squeeze in an on-chip laser. As an added bonus, quantum dot lasers are also very reflection-tolerant. That means they don’t require an optical isolator, something that is normally needed in an optical package, again saving space on the chip and resulting in more favorable production and running costs. 

The addition of a laser will allow for more compact coherent modules to be fitted per rack unit and this will provide more bandwidth density overall. It’s our belief that operators will want to migrate to these new types of transceivers as soon as they become available. It will enable them to take a crucial step towards taking care of tomorrow’s bandwidth needs, while at the same time hugely improving the efficiency of their current networks. In addition, quantum dot lasers combined with electro-photonic chips will also achieve footprint compatibility between intra-DCI and inter-DCI modules, another huge benefit. 

This innovative new optical transceiver technology represents a world first and will be a major boon for cloud and content providers as well as for sustainability and efficiency. 

The PEARLS project is funded by the German Ministry of Education and Research and runs until April 2022. For more news about PEARLS, watch this space.

Benjamin Wohlfeil

Related articles