All of us are an integral part of a digital society and make directly or indirectly use of online services such as streaming, storage, computation and more. The advent of the internet-of-things not to mention. This causes an exponential growth in data rates for optical communication reaching Terabit links in a few years. The situation is most severe for datacentres, where the current infrastructure is almost at its limits and new paradigms are searched for. Our work presents such a new paradigm with the introduction of a monolithic high-speed platform that might impact our future way of life.
Before diving into the details, I want to share my path towards this results completing my PhD studies. When I started my PhD at the Institute of Electromagnetic Fields (IEF) of Prof. Dr. Juerg Leuthold, my research interest were mostly on numerical methods and computational electromagnetics and I saw optical communications just as an interesting field to apply what I have learned. Yet, my PhD journey was not meant to go only this way and threw me into the experimental world of plasmonics. Plasmonics, the photonic technology with high‑performance devices (Nature Photonics 9, 525-528, (2015)) and flexibility for integration on many substrates (Science 358, 630-632, (2017)), offers unique advantages such as compactness and bandwidth. So I challenged myself to design and demonstrate the world’s smallest photonic integrated circuit for data modulation using plasmonics. Indeed, we achieved data modulation of 0.8 Tbit/s on a record-small footprint of only 90 × 5.5 µm2 by using a plasmonic modulator array with four channels (Journal of Lightwave Technology 37, 1484-1491, (2019)). The small size suddenly enabled new possibilities and raised my ambition and interest to merge them with electronics.
In 2017, I had the chance to join a European consortium with exactly this goal as part of the Horizon 2020 project PLASMOfab. In close collaboration with the electronics specialists from the Saarland University and Micram, and many other partners, we have created a novel monolithic BiCMOS-plasmonic platform for compact, high-speed optical interconnects that is expected to keep up with the datacentre demands of the next decades. We have achieved a breakthrough in electronic-photonic co‑integration and demonstrated for the first time more than 100 Gbit/s data modulation in a monolithic transmitter (see Figure).
Key to the success of monolithic integration was the co‑design of electronics and photonics, including assembly and packaging, thermal optimisation, and a new temperature-stable electro‑optic material. The electronic layers perform a 4:1 multiplexing to generate high-speed electrical signal by mixing of four lower‑speed inputs. The photonic layer uses a plasmonic intensity modulator to convert the electrical signal into the optical domain for transmission via an optical fibre. These layers are connected by on‑chip wires to guarantee shortest distances and best signal quality.
We have verified the potential of the bipolar CMOS electronic-plasmonic platform by testing two modulator concepts: a silicon-plasmonic modulator and an ultra-compact plasmonic modulator. We have processed both of them directly onto the bipolar CMOS electronics. Then, we have tested the transmitter under uncooled ambient air conditions and without encapsulation. We were able to demonstrate symbol rates of up to 120 GBd (corresponding to 120 Gb/s for two-level modulation) in a data modulation experiment.
For more information, please read our article “A monolithic bipolar CMOS electronic-plasmonic high-speed transmitter” published in Nature Electronics.