Spanish photonics pioneer iPronics Programmable Photonics is making photonic computing and telecommunications applications commercially viable with the world’s first reprogrammable general purpose photonics integrated circuit (PIC). The company’s technology enables hardware developers to programme light signals on chips with a level of flexibility that has never been seen before now.
Image Credit: amgun/Shutterstock.com
Advanced Manufacturing from a Solid Research Base
iPronics was founded in 2019 as a way to commercialize research carried out at the Universitat Politècnica de València.
The company is bringing a photonics innovation to market that would significantly reduce development time and costs for a wide range of photonics applications. This is the concept of a general-purpose integrated programmable photonic chip that uses common optical hardware and which can be configured and reconfigured from a computer.
This would mean that companies could re-use the same hardware platform for different applications.
This innovation would make high-speed photonic information processing much more time and cost effective to develop. iPronics says that this technology could be applied in any information processing task requiring high speeds, such as next-generation mobile data (5G and 6G), large data centers, artificial intelligence (AI), autonomous driving, quantum computing, and the Internet of Things (IoT).
What is Photonics?
Photonics applies the science of optics to generate, detect, and manipulate photons for various applications. It involves emitting, transmitting, modulating, signal processing, switching, amplifying, and sensing light.
In information processing, photonics can be applied with photonic integrated circuits (PICs). These pieces of hardware benefit from a small form factor and footprint, stability, low power consumption, and potentially economic manufacturing processes. Relative to conventional electronic chips, PICs also benefit from the properties of optical waveguides: propagation loss is minimal, there is no diffraction, power can be well confined, there is low crosstalk, and there is no risk of electromagnetic interference.
Due to these features, integrated photonics could have a large impact in a number of emerging technology areas. High-speed fiber communications, for example, could use PICs to support all multiplexer domains (space, polarization, and wavelength) in multiplexers. High-speed mobile data systems (5G and 6G) will rely heavily on PICs. PICs enable high-speed signal processing, and can form the basis of quantum logic gates required for quantum computing.
To date, one photonic platform has dominated in terms of research and focus: the application-specific photonic integrated circuit (ASPIC). This approach configures a particular circuit through which photons will travel and optimizes it for a particular function. The optimization includes managing propagation loss, power consumption, footprint, and the number and type of components in the device.
This optimization is carried out through a number of design iterations, which may need to be physically produced and tested for robust performance results. This means that ASPIC development processes are significantly drawn out, which in turn means that photonic devices are significantly expensive to develop.
Making Photonics Viable with Programmable PICs
iPronics is seeking to overcome this challenge by developing general-purpose processor architecture that can be integrated on a PIC. Such a platform would feature either single or multiple input/output generation, or both, and would be able to carry out different signal processing tasks after reprogramming on a computer.
This approach is inspired by electronic Field Programmable Gate Arrays, a mainstay of electronic hardware. Programmable PICs implement a common hardware with a two-dimensional photonic waveguide mesh.
In the underlying research that the company is commercializing, scientists have already demonstrated a reconfigurable waveguide mesh in silicon. This programmable PIC had a simple structure of seven hexagonal cells and carried out over 20 different functions.
These functions included optical ring resonators (ORRs), coupled resonator waveguides (CROWs), and side-coupled integrated spaced sequences of optical resonators (SCISSORs).
The research also proposes other mesh topologies, but notes that hexagonal and triangular meshes performed best in terms of reconfiguration.
iPronics costs development of each iteration of an ASPIC at between €550,000 and €1.7 million, with approximately 12 months of development time. Typically, two or three iterations are required to perfect the ASPIC design, bringing the total development time to two or three years and costs ranging from €940,000 and €4.6 million.
The company says its products reduce ASPIC development time by 90%, and costs by 95%. This is because new designs can be simply reprogrammed, downloaded onto the PIC, and physically tested in just a few hours.
There are applications in a number of fields, including quantum computing, multiprocessor networks, signal processing, chemical and biomedical sensing, and communications.
Next Steps for iPronics
iPronics is currently working on two major projects that will bring its technology closer to the market.
The “INSPIRE” project is developing programmable photonic processors in the form of TRL5/6 demonstrators, a world first. This project improves the company’s current technological readiness by adding more programmable unit cells to each chip.
This will result in enhanced performance in several areas, including chip coupling losses, space requirements, and power consumption. The new layer design will be produced in a pilot batch to be functionally tested, validated, and demonstrated.
The “PROMETHEUS” project is leveraging programmable PICs for neuromorphic computing architectures. Neuromorphic computing is inspired by the structure and working principles of biological brains. Neuromorphic chips create artificial “neurons” that distribute computer processing in a way that is analogous to brains.
iPronics’s PIC can be applied in neuromorphic computing through large scale photonic spiking neural networks that exploit the gigahertz firing rate of laser-neurons integrated in the chip.
With this project, iPronics will also put quantum random generators into practice, which, because they cannot be cloned, will embed physical layer encryption and authentication to the chip.
Before long, iPronics’s programmable PICs may form the basis of high-speed computing and sensing that will power the next generations of numerous cutting edge technologies.
More from AZoOptics: Recent Research on Quantum Cascade Lasers
References and Further Reading
About. [Online] iPronics. Available at: https://ipronics.com/about/ (Accessed on 26 January 2023).
Ham, D., et al (2021). Neuromorphic electronics based on copying and pasting the brain. Nature Electronics. doi.org/10.1038/s41928-021-00646-1.
Pérez, D., et al (2017). Multipurpose silicon photonics signal processor core. Nature Communications. https://doi.org/10.1038/s41467-017-00714-1.