Enabling Terabit capacity optical interconnects requires a paradigm shift in the packaging approach. The electrical interconnect distance between the optical engine (OE) and the digital switching chip must be minimised, removing signal conditioning chips and unwanted components like sockets that would otherwise be required and would inevitably lead to increased power consumption and lower signal integrity. It also requires the right combination of photonic and electronic technology that will be integrated to deliver high performance, low-cost and energy efficient optical engines. This approach has the potential to remove the optical interconnect bandwidth bottlenecks and allow DC networks and the 5G wired infrastructure, which heavily rely on them, to grow. POETICS is a research proposal that targets the development of novel Terabit optical engines and optical switching circuits and to copackage them with digital switching chips to realise Multi-Chip Modules (MCM) with Tb/s capacities, very high energy efficiency that fit into the roadmap of vendors. In order to do so POETICS will utilize SiGe BiCMOS, InP, PolyBoard and TriPleX technologies and rely on hybrid integration, which allows the selection and combination of the best performing components. POETICS in specific targets 1) MCM with 1.6Tb/s OEs based on 8-fold InP-EML arrays (200Gb/s per lane) and PolyBoard with parallel SMFs on par with the PSM/DR spec for 500m-2km intra-DC connectivity. 2) MCM with 1.6Tb/s OEs based on 8-fold InP-EML arrays (200Gb/s per lane) and 3D PolyBoard with duplex MCFs for 5G optical fronthaul applications 3) low power consumption 3D Benes optical switch 4) MCM coherent 64Gbaud OEs with up to 600Gb/s capacity of DC interconnect applications within 80-120 km reach on par with 400G-ZR specification.

Despite the significant advances that photonic integrated circuits (PICs) offer in terms of miniaturization, power consumption and functionalities, they run into scalability and cost issues, related to the fabrication yield, the increased integration and packaging complexity, the lack of wafer scale compatible processes and the lack of integration and packaging standards. Furthermore, so far photonic packaging considered the sub-GHz electrical connections to the PICs as a separate and second priority issue, until the number of electrical IOs of the PICs was too large to ignore. POLYNICES aims to address these challenges with the development of a novel general purpose photonic integration technology, compatible with wafer scale processes that will reduce the production costs of photonic modules by at least 10x. POLYNICES will develop for the first time a polymer based Electro-Optic PCB (EOPCB) motherboard that will host Si3N4 chiplets, InP components and micro-optical elements. POLYNICES invests in Si3N4 platform with PZT actuators to realize complex structures in only 1x1 cm2 chiplets with ultra-low power consumption. The chiplets’ grid array electrical pads and the use of flip-chip integration on vertical alignment stops will allow optical alignment and electrical connection in one step. The standard size and interfaces of the chiplets as well as the electronic IC co-packaging on the same EOPCB, provides excellent scalability and customization, and significantly simplifies packaging. Dielectric rod THz antennas will be integrated on the EOPCB taking advantage of its good HF properties. Using the above novel concepts and building blocks, POLYNICES will develop a fully integrated optoelectronic FMCW THz spectrometer with THz antenna array and beam steering abilities for quality control in plastics, a 16x16 quantum processor with integrated 780 nm light source and non-linear crystals and a 24x24 quantum processor with integrated squeezed light state source.

3PEAT will develop a powerful photonic integration technology with all size, functionality and quality credentials in order to help a broad range of optical applications like optical switching and remote sensing, to achieve a strong commercial impact. In order to do so, the project will introduce a fully functional 3D photonic integration platform based on the use of multiple waveguiding layers and vertical couplers in a polymer technology (PolyBoard), as a means to disrupt the integration scale and functionality. Moreover, 3PEAT will combine this powerful 3D photonic technology with a silicon-nitride platform (TriPleX), via the development of a methodology for the deposition and processing of multilayer polymers inside etched windows on TriPleX chips. In parallel with the development of this hybrid 3D technology, 3PeaT will bring a number of key innovations at the integration and component level relating to: a) the heterogeneous integration of PZT films on TriPleX platform for development of phase shifters and switches for operation up to 50 MHz, b) the development of a disruptive external cavity laser on the same platform with linewidth less than 1 kHz, c) the development for the first time of an integrated circulator on PolyBoard with isolation more than 25 dB, and d) the development of flexible types of PolyBoards for the purpose of physical interconnection of other PICs. This enormous breadth of innovations can remove the current limitations and unleash the full potential of optical switching and remote sensing and ranging applications. The main switching module that will be fabricated will be a 36×36 optical switch with 20 ns switching time and possibility for power and cost savings of almost 95% compared to standard electronic solutions. The main sensing module on the other hand will be a disruptive Laser Doppler Vibrometer (LDV) with all of its optical units, including its optical beam scanning unit, integrated on a very large, hybrid 3D PIC.

The bioreactor industry is currently flourishing with a global market valued estimated at 2.3 B€ in 2020 and predicted to exceed 6.6 B€ euro by 2030, growing at a rate of 10.7% CAGR. Despite this impressive growth, there are challenges which can significantly impede the further advancement of bioreactors: Bioproducts can be sustainable and competitive only if reliable and contamination-free production is ensured. Currently, there is no catholic solution to this issue. To this end, LIBRA project introduces a benchtop smart multi-sensing system for the in-line automatable screening of cultivation processes in bioreactors. The LIBRA sensing technology lies in the use of light based integrated on-chip, real time sensors. A novel integration procedure of the photonic platforms together with disposable microfluidic modules and biofunctionalization units will result in a modular system with interchangeable components enabling the screening of nutrients and pathogens in bioreactor samples, according to the end users need. Furthermore, the LIBRA system will be able to be attached and integrated to various bioreactor systems regardless of their form factors, spanning from stirred tank bioreactors to single use bioreactors (SUB). To achieve this, LIBRA will rely on a highly multi-disciplinary consortium comprising expertise and specialization in several fields spanning photonics, surface functionalization, microfluidics, advanced packaging and assembly, artificial intelligence and bioreactor manufacturers. The exploitable results of LIBRA are expected to disrupt the current PIC-based sensing landscape, as estimated by the two business cases stemming from the project: the market revenues one year after the end of this project are expected to be €7.8 million growing to almost €59 million in 2032, and plethora of new IP and new business opportunities for the partners involved in the joint venture of LIBRA.

Despite the huge progress by photonics, extended spectral bands at wavelengths below 1100 nm remain heavily underserved in terms of integration solutions. At the same time, the silicon nitride is booming and the lithium niobate is making an impressive comeback in the form of lithium niobate on insulator (LNOI), with both materials being transparent both in the visible and the NIR. With all these viewed as a unique opportunity, LOLIPOP steps in to develop a disruptive platform that will offer the highest integration, modulation and second order nonlinear performance in the entire spectrum from 400 up to 1600 nm, based on the combination of the LNOI and the silicon-nitride (TriPleX) technology. To this end, LOLIPOP will develop die-bonding and micro-transfer-printing methods for low-loss (<0.5 dB) integration of LNOI films on TriPleX without compromise in the functionality of the two platforms. It will also develop a process for growth of Ge photodiodes (PDs) inside pockets and a process for flip-chip bonding of active elements inside recesses on TriPleX. Given the possibility of the Ge-PDs to operate in the entire 400-1600 nm spectrum, and the flexibility of the bonding process to adapt to different actives and wavelengths, the picture of this ultra-wideband technology is complete. LOLIPOP will demonstrate its potential via the development of: 1) The first ever integrated laser Doppler vibrometer at 532 nm with ultra-narrow linewidth (<5 kHz) and ultra-high modulation (6 GHz), 2) The first ever integrated FMCW-LIDAR at 905 nm with ultra-high linear chirp (10 GHz) and optical phased array-based 2D beam scanning, 3) Photonic convolutional neural networks with record scale, computation speed (24 TOPS) and power consumption reduction compared to electronic solutions, and 4) The first ever integrated squeezed-state source with 6 dB squeezing level for quantum applications at 1550 nm. A roadmap for the offering of LOLIPOP technology as commercial service will be prepared.