Professor Graham Reed from the University of Southampton is a recognized pioneer and academic leader in the field of silicon photonics. He was one of the first researchers in the UK to work on silicon photonics since the early 1990s. His student Andrew Rickman was the founder of Bookham, Kotura and Rockley Photonics, and one of the pioneers in the field of silicon photonics as well. On the conference of IEEE GFP, a special report entitled "Silicon Photonics: the viewpoint of university" made by Reed, in which he explained his view on the development of silicon photonics from the perspectives of history, collaboration between enterprises， universities，research institutes and the role of universities. Professor Gan, who comes from Shanghai Institute of Microsystem and Information Technology (SIMIT) is a scientist in the field of silicon photonics in China. He got his PHD degree from MIT, and returned to Shanghai with the dream of science and technology innovation in China. He has been engaged in the research and industrialization of silicon photonics for a long time.
After Professor Graham’s speech, CFOL(China Optical Fiber Online) invited Professor Graham and Professor Gan to have a conversation of silicon photonics technology, and how to do silicon photonics’ research and development. We were looking forward to obtaining enlightenment from the exchanges between two generations of scientists. These two experts finally chose a relatively general topic considering limited time and audience acceptance.
Dr. Gan: Could you briefly review the history of the silicon photonics development?
Reed: “Professor Soref was the person who first engaged in this field. He published three papers in the late 1980s and early 1990s, so that he laid the foundation for later research on silicon photonics. These papers were all mentioned in my report. I remember my first published paper about silicon photonics was written with Professor Sore in 1991. Andrew Rickman was one of my earliest doctoral students, and he joined my team in 1990 probably. The first turning point in the development of silicon photonics was probably in 2004. Prior to that, available R&D investment came from small projects, with little industrialization. We went to enterprise to delivery training and educate people on silicon photonics. The turning point came in 2004, when Intel entered the field, silicon photonics aroused big companies’ interest, and I remember numerous media interviewed me at that time. From 2004 to 2010, it became easier to get research grants for silicon photonics, industry investment increased, along with more competition and the difficulty in publishing academic articles. Since 2010 till now, the UK is in the third stage, universities no longer be competitive in industry, instead, universities need to engage some new areas that enterprises won’t do, develop new technologies and new devices, even crazy ideas, as well as put more efforts on patents. At this stage, the progress of silicon photonics is fast. I think you knew many stories about it.”
Dr. Gan: one more general question, what drove the silicon photonics development?
Reed: “The initial impetus was related to the termination of Moore's Law. It was found that semiconductor technology could not follow Moore's Law as previously thought, and new technologies development was required. That's why Intel and IBM made big investment in silicon photonics. The development of supercomputer has also encountered a bottleneck. IO density, power consumption, etc. have seriously affected the progress of supercomputer. To sum up, it is the explosion of data volume that has led to overwhelmingly adjustments in semiconductor technology. Silicon photonics offered new imagination for human. Nowadays, many people think that the first drove the development of silicon photonics was data communications, and data center. Well, maybe not exactly, you know optical communications are good touchstone for silicon photonics to be practical.”
Dr. Gan: so do you think that the chip interconnection in a field like HPC is definitely going to be silicon photonics?
Reed: “I had to say that there are some things you can do today and some things you can do tomorrow. As the technology of silicon photonics advanced, so did the technology of competing electrical connections.”
Dr.Gan: let’s going back to the silicon photonics research, could you talk about the challenges in this field you met?
Reed: “The first challenge is the realization of passive low cost fiber chip coupling technology, and you saw many cases at this conference. The second one is wafer level testing, which to maximize the production yield that I mentioned in the report. They are both headache for me.”
Dr. Gan: how about laser optical source?
Reed: “Silicon optical chips have variety of methods for realizing optical sources right now. I think it finally depends on the application. A chip with only four channels and a chip with perhaps hundreds of channels should be taking different technical solutions for optical sources.”
Dr. Gan: Since we are engaged in silicon photonics field, the shortage of fab is a big limitation all the time. Just like when you move from Surrey to Southampton, because many fab companies over there. What do you think of this problem?
Reed: “Fab is a problem. Big companies like IBM and Intel, which have their own Fab in-house, so they don't have these problems. But it doesn’t mean you can't do anything without Fab. I have been encouraging utilizing ‘open Fab’. Small-scale customization processes provide flexibility for research and development, and standardized processes benefit company product development.”
Dr. Gan: As for the application of silicon photonics, do you think may be in other fields besides communication markets?
Reed: “There are variety developing markets for silicon photonics, and silicon photonics are very prospective for medical applications right now. We are still developing middle/far infrared silicon photonics technology, which has a great future in biological, medical, sensing and other fields. So whether silicon photonics succeed or not in the communication markets will determine how quickly it can be used in other fields.”
Dr. Gan: one last question. Could you talk about ecosystem of silicon photonics research and development, and point out the right direction for young scientist? You said big enterprises are very capable, so people in school, people in research institutes, and in small enterprises. What they gonna do?
Reed: “There's still a lot of things you can do. Take us for example, we have projects with big companies, they give us a lot of requirements, and we help them develop what they want to do. We also have projects from the government and partnerships with small companies. If it is suggested that schools and research institutes should not compete with companies, they should try their best to find partners and cooperate with companies. Choose specific points to break, not everything. Do the valuable project, apply for the appropriate patent, and do not suggest turning the "crazy" direction. I recommend that looking for a Consortium and find out your location and figure out how much you can contribute. Our laboratory welcomes the cooperation of any institutions and enterprises.”