The challenge

Photonic integration technology is evolving from simple photonic integrated circuits (PICs) that involved only few functionalities, with light generation and modulation being the most prominent, towards larger, more functional, and thus more complex photonic integrated modules, opening new application fields next to the traditional telecom application field. This evolution has always been fueled from the advantages PICs can offer in terms of size, energy efficiency, speed and has resulted in a constant growth of the photonic integration market.

Despite the increasing popularity, photonic integration technology has not yet delivered all required features such as cost-efficiency and compatibility with high volume production, mainly due to the lack of a simple, low cost, wafer scale compatible production process. The reasons can be traced not only in the fabrication processes of PICs but also in the photonic packaging. Monolithic fabrication becomes expensive and runs into scalability issues, due to: the rapid yield reduction with increasing size and complexity, the heat dissipation and thermal crosstalk from the extensive use of thermal actuators, and the large number of electrical pads that need to be accommodated at the edges of the PIC. Heterogeneous integration such as die-to-wafer bonding, although wafer scale compatible, requires optimization of many processes to achieve good quality devices. Hybrid integration is gaining momentum as the most practical solution for high-performance, multi-functional PICs, due to the availability of commercial assembly machines and other methods that can automate the integration process of PICs and other micro-optical elements such as lenses, at the chip level. However due to the lack of packaging standards, each machine needs to be adjusted to the kind of PIC and assembly task, resulting in fragmentation of the production process.

An additional barrier in terms of cost and time is that photonic packaging still deals with the photonic assembly of the PICs and their electrical connection to other PICs or electronic ICs as separate processes. In fact, it is estimated that it can account for more than 75% of the photonic module cost. Traditional wirebonding of PICs to PCBs becomes challenging when the optical subassembly must combine several PICs with a very large number of electrical pads or needs to bridge height differences and gaps. Optimization of the optical assembly process usually leads to compromises in the electrical assembly and vice versa.

The vision

POLYNICES is a research and innovation action that aims to overcome these barriers and provide a general-purpose photonic integration platform with all the cost, performance, scalability and manufacturability credentials for the next generation photonic modules. More specifically POLYNICES for the first time will spin-coat Fraunhofer’s PolyBoard material on PCBs, to realize a low-cost Electro-Optic PCB (EOPCB) motherboard with low-loss single mode waveguides and good HF properties, that will host in properly formed pockets, silicon nitride chiplets, InP components and micro-optical elements for advanced functionalities. POLYNICES takes advantage of LioniX’s Si3N4 platform with PZT actuators to realize matrices, as well as novel narrow linewidth external cavity lasers in 1×1 cm2 chiplets with ultra-low power consumption. The grid array arrangement of the chiplets’ electr. pads and the use of flip-chip integration on vertical alignment stops will allow passive optical alignment and electrical connection to the EOPCB’s electrical pads in one assembly step. Although the chiplets will accommodate different structures for different functionalities, they will all share the same size, optical and electrical interfaces, thus defining standard building blocks, leading to a tremendous customization and scalability potential with minimal effort and cost. Different functionalities can be installed in the motherboard by selecting the chiplets, or the same chiplet can be installed multiple times to scale the circuit. Most importantly, POLYNICES provides a unified approach to photonic integration and packaging, as the electronic ICs are co-packaged on the same EOPCB, greatly reducing packaging costs. On the other hand, the good HF properties of the EOPCB allow THz antennas to be integrated directly on the substrate.

Using the above novel concepts and building blocks, POLYNICES will develop a fully integrated optoelectronic FMCW THz spectrometer with > 2 THz bandwidth, 8 THz antenna array and beam steering abilities for plastic quality inspection, a 16×16 quantum processor with integrated 780 nm light source and non-linear crystals and a 24×24 quantum processor with integrated squeezed light state source. POLYNICES technology provides a holistic approach in photonic integration and packaging and can certainly make advanced photonic modules affordable to SMEs. The way the project will work towards its targets is presented in detail via the following ten (10) objectives. 

Using the ability of polymer to be spin-coated in almost any surface, the integration methods and functionalities unique to PolyBoard and TriPleX, and the excellent complementarity of the two platforms, POLYNICES will develop a disruptive set of integration and packaging technologies that will allow low-cost photonic system in packages for a range of applications including wide band THz spectroscopy and quantum information processing. The ultimate goal is to provide a technology that will be easily accessible to SMEs in the photonic industry. Overall, POLYNICES is an ambitious, but achievable project within the 3-years of duration as it has been carefully thought of during the past months, well discussed with the consortium regarding the ambition of this project; unanimously all considered it as an ambitious but not over-ambitious project with respect to the plans, budget and timescale, carrying some risks, which appear to be though manageable.