top of page

UVA Engineering researchers join Quintessent in quest to demonstrate photonic systems-on-chip

Future Applications Include Faster, More Efficient Data Centers and Next-Generation Millimeter-wave Wireless Communication

Two researchers looking at data on computer

The World Economic Forum has dubbed the digital era the “fourth industrial revolution.” The microchip is the new steam engine. Electric current replaces mechanical motion, and photolithography is the means of production.


Photolithography makes it possible to fabricate a large number of integrated electronic microelements—transistors, diodes, resistors and capacitors— in and on a thin layer of semiconductor material. A single semiconductor chip in your smart phone, for example, contains billions of these microdevices.


The next frontier of integration and miniaturization in semiconductors will do for photonics what the integrated transistor did for electronics: vastly increase the number of devices per chip. This will substantially increase the photonic integrated circuit’s functionality, reliability and efficiency.


A research team at the University of Virginia’s School of Engineering and Applied Science has set out to prove that a complex photonic integrated circuit is achievable and commercially viable. They have joined a team led by Quintessent, which has earned a research grant from the Defense Advanced Research Projects Agency’s LUMOS program. Quintessent is a start-up company spun out of the University of California Santa Barbara to commercialize quantum dot-based lasers and photonic integrated circuits for optical connectivity.


Four faculty members in UVA’s Charles L. Brown Department of Electrical and Computer Engineering have a role in Quintessent’s team: Professor Andreas Beling, Associate Professor Steven M. Bowers, Assistant Professor Xu Yi, and Joe Campbell, Lucien Carr III Professor of electrical and computer engineering. The quartet combines expertise in photonics, and devices and circuits, two of the department’s research strengths. Commercial partners Morton Photonics, X-celeprint, and Tower Semiconductor mirror these strengths; each firm plays an integral role in this research project, officially named PATRONUS.


To meet DARPA’s expectations for high performance, the team will need to design an integrated circuit that produces more power than it takes in, a physical property called gain. Whereas the transistor does this well for electronic integrated circuits, achieving gain in photonic integrated circuits has proven elusive.


“Gain is tricky in the optical domain,” Bowers said. The standard practice is to produce light through a laser that is coupled onto rather than embedded in the photonic integrated circuit. “If we can get the light source and the high-speed components on the same chip, we can create systems that are much more complex and have added functionality.”


Amplification is a related challenge. On the electronics side, a transistor can output more signal power than it takes in, drawing additional energy from the battery. Due to the specialized materials required for lasers and optical amplifiers, most photonic integrated circuit platforms lack the ability to increase optical power, limiting the total number of devices per chip to tens or hundreds of devices.


“A breakthrough is needed that will enable this optical power gain to happen on chip, replenishing the optical power as it goes through each device, and enabling future chips to increase the number of devices by several orders of magnitude,” Bowers said.


Commercial viability relies on the existing manufacturing infrastructure for consumer electronics semiconductor chips; silicon substrates are the required base platform for manufacturing low cost, scalable photonic integrated circuits. However, electrically driven lasers and amplification (i.e., gain) cannot be achieved using conventional silicon wafer fabrication methods and materials. Quintessent is leading the integration of III-V gain materials containing quantum dots with silicon photonic integrated circuits to overcome this problem.


“We are using state-of-the-art quantum dot gain material and integrating them onto commercially available silicon photonic wafers to enable a technology platform that combines high-performance electrically driven lasers and amplifiers with leading-edge silicon photonic functionality,” Brian Koch, Quintessent vice president of technology and PATRONUS principal investigator, said.


The initial PATRONUS platform demonstration will include integrated gain with lasers that have over 100 milliwatt output power and benefit from quantum dot gain for insensitivity to high temperatures and optical feedback.


“Through material, design and platform optimization, we expect the final platform to have truly state-of-the-art laser and optical amplifier performance with watt class output powers, low linewidth and noise and high wall-plug efficiency,” Koch said.


Two additional building blocks are critical to the effort. One is the optical modulator, a device that can switch the laser light on and off 100 billion times a second, a rate as fast as 100 GHz. The second is a photodetector, a device that can receive the light at such a fast pace and turn it back into a current.


The UVA Engineering team has significant experience working with the material systems and alloys that are preferred for modulators and photodetectors, specifically III-V compound semiconductors. Team members hold several records for high-speed detectors and have teamed up with Morton Photonics to develop modulators with equally high performance.

“Morton Photonics has been working on advanced silicon photonics devices and photonic integrated circuits for over a decade, with a focus on RF photonics and sensing systems, to provide new capabilities for both DoD and commercial applications,” Paul Morton, chief executive officer and chief technology officer, said. “We have worked with members of the UVA team on multiple projects, taking advantage of their world-leading photonics expertise, to develop new photonic devices for use in our systems, and we see PATRONUS as an extension of this collaboration.”


To achieve enhanced functionality, the individual “chiplets”—the laser, photodiode and modulator—must work seamlessly and conduct a signal to other electronic components of the sensing, computing or communications system.


X-Celeprint’s micro-transfer-printing technology allows for the seamless integration of these components with high-yield and accuracy. Micro-transfer-printing transfers an array of laser devices in parallel to the silicon substrate, resulting in a monolithic photonic integrated circuit that can be further processed by a wide range of advanced packaging techniques.


Tower Semiconductor will manufacture the host silicon photonic circuit wafer using its leading-edge foundry PH18 silicon photonics process.


“Tower believes the technology concepts pursued in PATRONUS may lead to the next leap in PIC integration, in particular for highly integrated datacom applications, as well as for myriad novel photonics markets,” David Howard, Tower Semiconductor executive director and fellow, said.


Passive photonic circuits connect the lasers, amplifiers, modulators and detectors on the wafer, serving as optical routers to multiplex and demultiplex light of different wavelengths. “The passive circuits allow us to leverage the huge bandwidth in the optical domain to achieve high-frequency applications, like next-generation millimeter-wave wireless communication, that are difficult if not impossible for pure electronic systems,” Yi said.


Bowers is exploring systems applications to showcase future capabilities—what could be achieved with the high-performance, second-generation photonic integrated circuit that the PATRONUS team is asked to build. These include photonic systems-on-chip for use in ultra-fast data center switching and routing, detection and creation of photonic signals with unprecedented data rates, and even next-generation, 6G wireless signals capable of transmitting unprecedented data rates using higher millimeter-wave frequency bands.


bottom of page