Charles Kao, the father of fiber-optic communications, was recently awarded a share of the Nobel prize for his work in physics. On October 16th, 2001, Larry Johnson had the opportunity to film an interview with Dr. Kao at the Stanford Park Hotel in Palo Alto, California. This interview is intended to be released in the near future as part of the History of Fiber Optics compilation. The following is excerpted from that interview. For a full transcript, contact Larry at ljohnson@FOHistory.org

Johnson: When did the attention [at STL] turn to the possibility of using optical fibers as a potential medium?

Dr. Kao: I really joined when STL (Standard Telecommunications Lab) was already thinking about optical communications, because the laser was invented and everybody got excited. But at STL, Alec Reeves recognized that optical frequency as a carrier wave should be exploited as early as possible and they were already doing experiments for optical communications. So when I joined, the work was started but I was not directly involved, only peripherally involved because the two labs were working very closely. By 1963, Karbowiak decided that besides the millimetric waveguide project, which was ongoing, STL should start looking at alternatives to free space propagation as well as real guided propagation using dielectric material.

Johnson: Can you tell us how the research on fiber evolved?

Kao: Well, that’s a long story. First of all, we had to make some computations to understand what it takes to guide electromagnetic wave at optical frequencies, and there are many ways of guiding it. So the starting point was, “Let’s just shoot the beam in free space without guiding it since the lens can focus the beam almost parallel and it should shoot over distances of many meters.” To our surprise, we found that when you shoot a beam over nearly a kilometer distance, the beam is very unstable because it travels through the atmosphere and the atmosphere is not uniform in density and that gave us a first warning that free space is not a very friendly place. It’s not free because it’s really filled with air, which happens to be changing in density and shifting the beam.

Johnson: Dr. Kao, can you tell us the challenges first encountered and how you addressed them as far as the use of glass as a conceptual transmission medium?

Kao: Well, we first had to make sure that transmission could not take place in free space or in any form of guidance that involves too much of free space. First, we had to see whether one could guide it by using lenses, and that generated this so-called confocal lens guide. It turned out that the free space still had a great deal of trouble for the beam to stay within the access that it’s supposed to be flowing, so we decided that maybe we should look at a true waveguide. A true waveguide had one major problem: we didn’t know what material was transparent. There were two types of materials that we could choose from. One is a low temperature dielectric material like Perspex, or Plexiglas as some people know it. This is low temperature glass and the transparency is pretty good. The other is the normal glass that we associate with glass being transparent. So we had to tackle and understand it by measuring it to see what sort of losses these transparent materials were, and whether we could do any improvement on them.

So that’s when we started looking, first of all, at a waveguide which is designed such that the light does not couple very strongly to the waveguide but travels with only 10% of the light inside the dielectric and 90% of it outside the dielectric. This way we said, “Well, this will make sure that only 10% of the loss of the guiding material will be imparted into the light, thereby reducing the transparency.”

Johnson: So you must have had to develop specialized test equipment or measurement equipment because you were breaking barriers.

Kao: That’s right. Everything was nonexistent. We didn’t have any test equipment. We didn’t have any clue of any of the fundamentals of these things, so we had to answer those questions first. So we built our test equipment. We were trying to measure things like 1 mm distance, and we had only a ruler, which has calibrations in inches. So we had a very difficult time of trying to make sure that our measurements were right to determine a sample of glass or other material which has losses that are pretty low.

Johnson: Now at this point, when you were looking at the use of optical fibers-pulling this technology together, performing your measurements, and defining the characteristics-was this looking at a single-mode concept or a multi-mode concept?

Kao: Well, at the beginning, we obtained a sample of a fiber, which was made by an American optical company to demonstrate that the optical wave is like an electromagnetic wave and that you can get different modes in a fiber. Sowe obtained samples and did some of the mode work to show that you can excite different modes by changing the excitation conditions using [a] helium-neon laser as a light source. Once that is demonstrated, you know that given the right condition, you can get single-mode to multi-mode fiber, and then we had to look at what we really needed. The theoretical calculation that we did at that time convinced us that we really should go for a single-mode fiber, and the single-mode fiber should be designed such that that’s the only mode that it can propagate on it and that it should have a fairly substantial cladding in order that outside interference by touching it or something would not disturb the propagation. So we defined the geometry and also defined the type of refractive index difference and so on and showed that that’s what the guiding medium should look like.

The next one is we had to look for more material to see whether indeed we could get the loss of the transparent material to come down. That meant that we needed to know what causes losses, so we did some very fundamental work to look at the basic scattering loss due to small particle scattering in the glass as well as absorption due to impurity ions. This information in gross orders were in existence to people that were making colored glass. They said, “Oh, if we put few percent of ion into glass, you will get a colored piece of glass. If you put some copper in it, you will get another color and so on.” So I had to answer the question to what extent we must remove this material before the transparency can be improved. So the paper calculation showed that we must reduce the impurity ions to less than one part per million for some materials and one part per billion in other types of materials, and these materials that we don’t want are the transitional elements of the periodic table such as copper, manganese ion, and so on. So that gave us the theoretical basis to know what we must get in order to get the transparency and that took two to three years-two years or so-to get an essential confirmation. It was a very difficult confirmation because it was very difficult to find the material.

Johnson: It has been said that others looked for the best glass available where you looked at the theoretical limits on what was possible.

Kao: That is true. That is precisely the question that I answered ourselves. We said, “In order to correct this problem, we have to understand what these problems are, and how do we overcome these problems?” So we tackled the most fundamental problem, first of all. If we made a fiber, and if the tolerance limit of say 5% variation might make the guide not work, then it would be futile for us to try to make it. That’s why we did the microwave experiments-to show the work that George Hockham did, which was to try to see what are the radiation losses that can be caused by essentially dimensional intolerance-the dimension of variations. We tackled that both theoretically on the electromagnetic theory as well as experiments to verify that indeed if you put periodic variations, you can get very strong radiation coming from it. George Hockham was so fascinated by that experiment that later on in his life he concentrated on those radiative elements as his career, which is very interesting.

Johnson: Can you tell us about how you calculated the allowable loss for optical fiber for using a transmission system?

Kao: This is not calculated at all. In fact, this is a figure that we say if we can achieve 20 dB/km, then it becomes practically useful. If we cannot achieve that, then there will be too many repeaters from point A to B that it makes the system very awkward and not really practical for real use. So we set this limit that we must reach before we would say this is something that is practical and will do something for us. 20 dB happens to be the last that can span the repeater spacing at 1 km interval, which was the repeater spacing of coaxial line systems at that time.

Johnson: Okay. So in November of 1965, George Hockham and you released a paper to The Proceedings of the Institution of Electrical Engineers titled “Dielectric-fibre surface waveguides for optical frequencies.” This paper, even today, addresses the issues that are viable. The six main topics address the dielectric fiber waveguide material, electromagnetic and physical aspects of the fibers, and review the experiments and results. The conclusions themselves open the door of the visionary that you must have been as well. Can you expand on the paper and the vision beyond the paper?

Kao: Well, at that time when optical communication was regarded as a possibility, all the people in the transmission business said, “This is something that’s going to change the communication world very significantly.” So we were all very excited that we could put so much more information onto the transmission medium and so on and so forth. So the excitement was there. When this paper was written, we saw that it really could meet all the requirements. I was really very excited about that and that’s why in the publicity that occurred afterwards we were saying, “Hey, this thing will allow us to do all sorts of things,” that were at that time very brave to be stated by the PR people-like it might go across the ocean and send very many messages simultaneously across great distances. That was our speculation at that time. However, based on the paper, you can see that should be a reality. Interestingly, the paper’s accuracy was very, very high. In fact, the only thing that we did not cover in the preliminary sense is the problem of hydrogen penetration into the core of the fiber.

Johnson: After the paper was released, you came to travel quite extensively, meeting with different organizations and traveling around the world. Can you tell us about how that evolved?

Kao: Well, I think one of the things is this optical fiber work was originated at our lab, and I felt that we had to tell everybody that this was forthcoming and could be very exciting. I wanted to essentially wet the appetite of people so that there would be more people interested in working. I did not have any concept that this was releasing information to enable other people to become competitors of ours. In fact, my thoughts were the more people that work on it, the earlier we might be able to make this thing a practical reality. Later on, I felt that this was absolutely the right move.

Johnson: So in 1966, you traveled extensively in North America, Japan, and Europe to talk, discuss, and promote your findings. I find it unique that STL would invite potential competitors about your findings. Can you expand a little more on STL’s philosophies and objectives?

Kao: Well, as soon as we started to really look into this project, we needed people to pour more money into it. At that time, the first supporter was the British Post Office, and they asked us at STL to expand the project. So by the time I went over to the United States as well as Asia, particularly Japan, there were several centers that were already working on preliminary work in this direction. I was very happy because as soon as I wanted people to make fiber, I had to approach the glass people, and glass people think very differently. So there were lots of different ideas. It would be very, very useful, as far as I was concerned, for lots of people to try lots of different techniques to really make fiber that would perform with a specification like what we were looking for. So we went to companies like Pilkington in the UK, Bausch and Lomb, the American Optical Company in the United States, and some companies in Europe, as well as Japan. Japan Tohoku University is looking at different ways of making fibers, and they were the ones later on who invented the concept of the graded index fiber. I took the opportunity to tell NTT about it, so that’s why in some ways Japan felt that I was letting them know early enough, and they were very pleased about that. Anyway, my intention was to promote this project since I knew that we were going to have to do a very, very large amount of work before this thing would become a real, practical system.

Johnson: How did the paper and the following press release affect your work and role at STL?

Kao: Well, first of all, there was a very significant amount of work that had to be done, so we were trying to put many experiments into a mode of operation that includes making a fiber. Making a fiber was quite unpredictable, so we were testing very many different ways of making fiber from either molten glass or with rods of glass that are surrounded by other rods and try to pull it into fiber. There were lots of experiments going on. In fact, everything was new. Nobody knew any of the things, and the suppliers didn’t have any of the right low-loss glass, so we were experimenting-essentially looking at the techniques. We were discovering, to our dismay, that many of the techniques produced fibers that are very high loss indeed because the interface was not very good, meaning the cladding and the core interface are rough. Therefore scattering loss was very high and so on. So it was quite a period in which we anticipated a lot of pioneering work that had to be done at very many places with different techniques. So by 1970 when Corning came up with a result which showed that the 20 dB/km was up, I was very pleased. That announcement essentially indicated the paper’s prediction that such a fiber can be made.

Johnson: Since we mentioned the Corning fiber in 1970, when did you first hear about this breakthrough?

Kao: They wrote me a letter to tell me, as well as the fact that at the long-haul waveguide symposium when we presented our papers including this paper and then the variant. Corning took the opportunity of that conference NIEE in the UK to announce that they got this result, which I’m sure later on-- It was revealed that it was actually a very, very experimental fiber that they produced. Nevertheless, it was a demonstration of the possibility of achieving that result.

Johnson: How does it feel now to look back at all the changes and achievements that have occurred due to your perseverance and achievements?

Kao: Thinking about it now instead of sort of reflecting on how I envisaged it then, I think perseverance is necessary. However, I was so enthused about it that time seemed to pass by very quickly, and progress seems to be continuously being made, so I had a very satisfying life with respect to my association with the fiber. I regard the fiber success as one of demonstration of really a concerted work from many, many people in many countries. It was very satisfying to see that fiber indeed could live up to its expectations. At one time, the graded index fiber was seemingly taking over as the main fiber transmission system, but fortunately, the single-mode fiber came back very strongly, which was what I wanted that to be. So in the end, I think my interest has shifted back to the entire system and in fact to the networks and this continues to the present day.