Opendata, web and dolomites

Report

Teaser, summary, work performed and final results

Periodic Reporting for period 1 - MORPHIC (Mems-based zerO-power Reconfigurable PHotonic ICs)

Teaser

MORPHIC develops MEMS technology for large-scale silicon photonics to create programmable photonic integrated circuits (PIC) that can be configured in software for different applications.PICs are chips that manipulate light in waveguides, and today they are mostly used for...

Summary

MORPHIC develops MEMS technology for large-scale silicon photonics to create programmable photonic integrated circuits (PIC) that can be configured in software for different applications.
PICs are chips that manipulate light in waveguides, and today they are mostly used for fibre-optic communications, but also for sensors, spectroscopy and quantum optics. Silicon photonics is a unique PIC technology that builds on the fabrication technologies developed for CMOS electronics, allowing for very compact building waveguides that can be integrated into very large circuits.
Developing PICs is a costly proposition: it can take a year to design, fabricate and test a new chip, similar to electronic chips. However, in electronics, programmable chips, such as FPGAs, can be purchased off the shelf, and these have made electronics a widely accessible technology.
MORPHIC wants to enable a similar model for photonics, creating an optical equivalent of the FPGA. For this, we develop a complete technology stack based on silicon photonics. A programmable PIC consists of many identical building blocks connected in a dense mesh of waveguides, where the routing of light is controlled by electronics. Such a chip therefore requires electronic drivers for all the optical elements, an assembly scheme for fibers and high-speed connections, and the algorithms and software that will allow a designer to actually implement a useful function.
On programmable PICs, light passes through many more components than on a custom-designed chip. Therefore, the building blocks should be efficient, compact and have low power consumption. That is why MORPHIC is extending silicon photonics with micro-mechanical tuning elements (MEMS). Photonic MEMS consist of suspended waveguides that can be moved by applying a voltage, performing a phase shift or coupling between waveguides. Unlike the commonly used heaters, MEMS consume very little power and, with mechanical latching, can even maintain their state without external power. The exposed MEMS waveguides need to be protected from the environment, so MORPHIC also develops a hermetic sealing procedure for the MEMS.
MORPHIC will demonstrate this technology in three application settings: low-power optical switches for fiber-optic communication, optical processing of microwave signals for 5G communication, and optical beamforming for free-space light communication and LiDAR. MORPHIC will implement dedicated demonstrators for these applications, but also prototype a generic programmable PIC that can be configured in software to be used in all three cases, testing the viability of this photonic equivalent of the FPGA.
MORPHIC is testing the full supply chain for programmable PICs, from the MEMS processing to the packaging schemes and the programming interface. This supply chain can establish a new way of working with PICs, where generic chips are produced in high volume, and can be customized in software. This lowers the threshold for trying new concepts and deployment of PICs into new applications, reducing the time for first prototypes from months to weeks.

Work performed

In the first period of the MORPHIC project, the ground layers were laid out for MEMS-enhanced silicon photonic programmable chips.
Starting from IMEC’s state-of-the-art silicon photonics platform, we developed the fabrication process for the MEMS waveguides. After applying a protective layer we used a vapour etch to locally remove the glass layer supporting the silicon waveguides, creating the suspended structure that can be electrically actuated. We demonstrated working MEMS phase shifters and tunable couplers. In parallel, the hermitical sealing process was tested, showing the compatibility with the photonic MEMS.
At the circuit level, we explored different concepts for programmable circuits, and how they behave when not all the circuit elements are perfect. In large circuits, the effect of small variations can accumulate, which also imposes requirements on the electrical control. A new circuit models was built and applied not only to programmable circuits, but also to switch matrix and beamforming circuits, and even to explore the configuration algorithms needed to define functionality in a programmable PIC.
Translating a programming strategy into a working circuit goes through several interface layers: software, electronics and the physical interfaces on the chip. We elaborated these interfaces for the fabrication RUN2 and the final demonstrators. With >3000 electrical connections, 72 optical fiber ports and 24 high-speed microwave lines, this is one of the most ambitious packaging endeavours in silicon photonics. We selected the packaging technologies, and devised a solution where significant reuse between different chips is possible. All packaging rules, together with the process rules for the MEMS components and the hermetic sealing, are integrated into a design kit for RUN2.
The design of RUN2 contains a multitude experiments, ranging from next-generation MEMS building blocks to fully-functional programmable photonic circuits and small-scale implementation of the application demonstrators for switching, microwave filtering and beamforming.

Final results

At this point MORPHIC is laying the groundwork for an entirely new ecosystem for programmable silicon photonic chips powered with MEMS technology. At this stage, we have not yet realized the full potential of this technology, but with the results of the first fabrication run, we have a identified the implementation route for the entire technology stack for the programmable photonics. The demonstration of photonic MEMS components compatible with a standard silicon photonics platform has garnered significant interest from the community, as the quest for compact, low-power, optical phase shifters in silicon photonics is still ongoing.
Our simulation framework allows us to assess the viability of large-scale circuits, and especially programmable PICs, in different application settings. We combined this with a mapping of the application spaces, showing great potential for communication, microwave processing, beamforming, and sensing, but also in emerging fields like quantum information processing and artificial intelligence.
The process and building blocks, the circuit concepts and the packaging strategies were brought together in the design of RUN2. This second mask contains many designs that are beyond the state of the art: Photonic MEMS with latching for zero power consumption, large-scale programmable circuits, and a packaging strategy for large-scale photonics that could well become an example for prototyping such large-scale systems.
The potential impact of MORPHIC is not just in its technology, but in the realization that programmable photonics could well become a game-changer in how people use PIC technology. By providing a software interface to the manipulation of light on off-the-shelf chips (fabricated in large volumes), prototyping of new photonic concepts will become easier, postponing costly custom PICs to a later stage in product development. Also, this new ecosystem can put integrated photonics in the hands of the ever-growing maker community, spurring an entirely new range of functionalities.

Website & more info

More info: http://www.h2020morphic.eu.