The 2001 IEEE Medal of Honor recipient is recognized for his pioneering research and leadership in developing lasers and optoelectronics
PHOTO: JIM ALLEN
Name: Herwig Kogelnik
Date of birth: 2 June, 1932, in Graz, Austria
Family: wife, Christa; sons Christoph, Florian, and Andreas
Present position: adjunct photonics systems research vice president, Bell Laboratories
Education: Dipl. Ing., 1955, and doctor of technology, 1958, in electrical engineering from the Vienna University of Technology; D.Phil. in physics from Oxford University, England, 1960
Patents: 34, four still active
Hero: Karl Jansky, founder of radioastronomy
Favorite composers: "Several. Georg Phillip Telemann and Handel are certainly near the top."
Favorite movie: Four Feathers
Last book read: God's Equation: Einstein, Relativity, and the Expanding Universe, by Amir D. Aczel
Leisure activities: tennis (former president of the Seabright Lawn Tennis and Cricket Club), skiing, swimming, hiking, paddle tennis
Memberships: Fellow of the IEEE, Optical Society of America, National Academy of Engineering, National Academy of Sciences
Awards: Frederic Ives Medal, Optical Society of America, 1984; David Sarnoff Award, IEEE, 1989; Joseph Johann Ritter von Prechtl Medal, Vienna University of Technology, 1990; Quantum Electronics Award, IEEE Lasers and Electro Optics Society, 1991
Herwig Kogelnik stands in a cramped closet in a Bell Laboratories research building, surrounded by file cabinets and cardboard boxes full of paper. Having pushed an office chair weighed down by yet another box out into the hallway, he's got room to stand before an open file drawer and go through its contents. He's been at it for maybe 15 minutes, trying to pull out only the highlights of his 40-year career.
"Boy, was I in so many things?" he mutters to himself. "This is amazing."
Amazing enough that the IEEE awarded him the 2001 Medal of Honor, "for fundamental contributions to the science and technology of lasers and optoelectronics, and for leadership in research and development of photonics and lightwave communications systems."
The "things" he's been involved with during his four-decade career as an experimentalist, theorist, and administrator at Bell Laboratories include the basics of how lasers operate, the underpinnings of the multichannel optical networks that make the Internet possible, and holographic data storage, to name a few. He was also a co-inventor of distributed feedback lasers, so critical to today's optical communications, and he was working on optical amplifiers as long ago as the 1960s. In fact, he wrote the book, or at least the chapter, on the modes of resonator cavities, the part of the laser that produces a beam.
Described by co-workers as a gentleman in the European mold, Kogelnik displays an easy-going manner as he patiently puts into words, tinged with a sonorous Austrian accent, years of research that is more often presented as mathematical formulas. As he shows a visitor around the conference room in the Crawford Hill laboratory in Holmdel, N.J., Kogelnik points to a black-and-white photograph that shows astronomer Karl Jansky in front of a huge metal lattice-work that served as his antenna. In 1932 Jansky was working on ways to improve microwave communications when he picked up a signal coming from the center of the Milky Way galaxy, essentially inventing radioastronomy. "Jansky's my hero," said Kogelnik. The only reason he didn't win a Nobel Prize, Kogelnik contends, was that it took people 15 years to realize the magnitude of what he had accomplished, and by then he was dead.
Born in Graz, Austria, in 1932, Kogelnik grew up in Bleiburg, a country town of about 1500 people. Situated in southern Austria not far from the Slovenian border, the town was out of the way then--and still is today. "You really have to make an effort to get there," he said. Because of its remoteness, Bleiburg was mostly untouched by World War II.
His father was a commercial manager for a newspaper, and his mother taught English and physical education. When it came time, in 1950, to decide what career path to follow in college, he considered medicine. His grandfather was a doctor, cousins were doctors, even his brother became a doctor, so there was a fair amount of family tradition. But there was also no shortage of doctors in his homeland.
"There was a lot of advice coming from friends and family: 'Don't become a doctor, because there are so many doctors in Austria, you'll wind up selling cigarettes in the drug store,' " he said. He took their advice, but the medical tradition in the Kogelnik family has been maintained by his youngest son, now doing a medical internship at Stanford University in California. Another son works for a computer chip company in California, and the third is a lawyer in Washington, D.C.
One other career Kogelnik considered was music, the violin in particular. Playing the violin requires that the wrist be bent softly, but when Kogelnik was 18, his violin teacher noticed that the youth's wrist was getting stiff, the result of all the tennis he loved to play. "He wanted me to give up tennis, and kid that I was, I went to college and gave up the violin," Kogelnik said. He thinks the fact that he let the violin go so easily shows he wasn't that committed to music. To this day, he plays tennis with his wife, Christa. He hasn't picked up a violin in years.
Shortly before he finished high school, a career counselor told him that electronics was the hardest course colleges had to offer. Kogelnik was impressed, though he had no real background in engineering. As a child, he didn't build his own radios or cobble together inventions out of household objects. But the promise of hard work proved to be a persuasive argument, and he went to the Vienna University of Technology.
There he earned a diploma in engineering in 1955 and a doctor of technology degree in electrical engineering in 1958. While working on the doctorate, he taught in the college's electrical engineering department and did research on microwave transmissions. Kogelnik next won a British Council Scholarship to the University of Oxford in England, where he studied plasma physics with an eye toward fusion and the generation of electricity. Asked to submit his work as a thesis, he was awarded a Ph.D. in physics in 1960.
It was at that point that he had a visitor who would change the course of his career. Rudi Kompfner was then a director of research at Bell Labs. The first laser--Theodore Maiman's ruby laser--had recently been demonstrated, and its potential as a means of communication was already apparent to leaders in the field. Kompfner, whom Kogelnik calls Bell Labs' primary champion of optical communications, was on the road, putting together a research team that he hoped would figure out how to achieve that potential. Kompfner, himself a 1973 IEEE Medal of Honor winner for inventing the traveling-wave tube used for broadband microwave communications, talked to Kogelnik at Oxford, but was unable to fully convince him to join Bell Labs.
This was four years after the Soviet Union launched Sputnik and frightened the United States into a renewed emphasis on science. Kogelnik took advantage of a government program that provided free transportation to the United States and a grant for three months while he interviewed for jobs. He met with Kompfner again, this time in New Jersey.
During that meeting, Kogelnik was ushered into a laboratory to see something few had seen before, the first continuously operating laser, a device that utilized a mixture of helium and neon gases. "They said, 'Here it is. Wouldn't you like to work on such things?' And that was very intriguing," Kogelnik recalls.
Kompfner finally convinced Kogelnik to switch from plasma physics to optical communications with the simple come-on: "Think of all that bandwidth."
Theory and practice
The first gas laser was built entirely inside a vacuum envelope, which was then filled with gas. Electric energy would excite the gas atoms to emit photons, which would bounce between mirrors at either end of the tube, stimulating the emission of more photons as they traveled back and forth through the gas. Eventually this lasing action would produce a narrow beam of light in a small range of wavelengths. To adjust the mirrors, engineers had to use levers running from outside the vacuum envelope to the mirrors. The setup was impractical, and Kogelnik was assigned to build a better gas laser.
His team came up with the idea of placing glass windows at each end of the vacuum tube [see photo]. To avoid losing photons to reflections in the glass, the team slanted the windows at Brewster's angle--that is, the angle at which linearly polarized light of a particular wavelength will not reflect off glass. Since the light can pass through glass with no reflection, designers could place the mirrors outside the vacuum tube, where they are easier to manipulate.
To improve the lasing process, they replaced the flat mirrors with curved ones, which were easier to align and which helped create beams with a narrower range of wavelengths. They then put obstacles such as pins into the laser cavity to see what effect they would have on the patterns of the beam, known as modes.
"To understand all these modes was a big challenge," Kogelnik said. But it was a challenge he readily tackled, laying the theoretical groundwork for understanding how laser modes work. At the time, no one really knew how laser beams could be expected to behave or what effect one change in the system would have on the whole lasing setup. Understanding the modes allowed engineers to make the beams more coherent and control the wavelengths better.
Once engineers started using curved mirrors, many questions popped up, Kogelnik told IEEE Spectrum. For instance, what would happen if the mirror on one end of the laser's resonator cavity had a different curvature from that of the other mirror? What if the angle of the mirrors was changed, or the distance between them? Kogelnik set about finding out.
"For some cases, you got very nice, stable behavior. For others, the beam would want to blow up in size" and diverge, he said. "That was a big surprise to everybody, that laser cavities could be stable or unstable."
And what if lenses or prisms or some other element were introduced? He applied the old principle of ray tracing, which tells where light will go when it passes through a lens or a prism, to laser beams. He derived a rule, the ABCD law, which describes how laser beams propagate. The rule is now a basic tool engineers use to design lasers that behave predictably.
Kogelnik's ability to develop the theories that explained laser behavior helped him to improve the understanding of holograms, which was a hot field of research in the late 1960s. A hologram is made by creating an interference pattern between two laser beams. One beam strikes the object to be recorded, and then interferes with a reference beam in a photosensitive material, where it creates a holographic grating. Although researchers knew what happened to a light wave as it propagated, they knew little about what would happen when two waves interacted. Kogelnik derived the mathematical formulas that explained how a holographic grating feeds the electromagnetic energy from one light wave to another. What he came up with was dubbed coupled wave theory.
Coupled wave theory gave optics researchers a vital mathematical tool. "They knew how to make [holograms], but there was not the kind of understanding that people could use to improve how they made them," said Rod Alferness, now senior vice president of research at Bell Labs' Optical Networking Group in Holmdel, N.J., but once a young researcher brought into the fold by Kogelnik.
Because they provide a compact way of storing data, holograms are once again an important area for research. By changing the angles of the holographic gratings slightly, engineers can store multiple images in the same hologram. The knowledge of how to do that, and how many images can be stored, traces back to Kogelnik's analytical work, Alferness said.
The path to distributed feedback lasers
Coupled wave theory played an even bigger role in the laser field. Early lasers were rather messy in their output. Their beams covered a fairly wide wavelength range, not the single frequency that researchers knew would be needed to make optical communications practical.
In 1970, two items that were crucial to optical telecommunications were invented. One was an optical fiber with low enough light loss that data could be transmitted over a practical distance. The other was a semiconductor laser that could work at room temperature and be used to send data down the fiber. Because the beam from the lasers had to fit into the end of a tiny optical fiber, thinner than a human hair, the salt-grainsized semiconductor, seemed the most promising. But as the semiconductor laser, made of the alloy gallium arsenide, operated, its temperature changed, and the wavelength of the emerging light changed along with it. For optical communications to work, stable, single-frequency lasers were required.
Kogelnik and his colleague Charles Shank hit upon the idea of using gratings, similar to holographic gratings, to stabilize the laser's output [see figure]. They made tiny gratings, with periodic spacings measured in nanometers, to match the wavelength of light they wanted. Placed in the laser's resonator cavity, the gratings, which Kogelnik described as "little wiggles," acted as tiny mirrors distributed throughout the resonator cavity. The gratings made the light of the desired wavelength scatter back to the end of the cavity and be reflected forward again, forcing all the light gain to come out at the right wavelength.
The two researchers demonstrated their so-called distributed feedback (DFB) concept in a device that used a nitrogen laser to pump a fluorescent organic dye dissolved in a gelatin film. The dye, like gas and semiconductors, can also produce lasing action. In those days, semiconductor devices were too delicate and difficult to make. "Very early on, we were thrown out of the labs of the semiconductor people," Kogelnik said.
Shank, now director of the Lawrence Berkeley National Laboratory, in California, was a young man just out of graduate school when he collaborated with Kogelnik on the project. The DFB laser was fairly easy to make, he recalls, once the pair figured out how big the wiggles needed to be. "It relied very heavily on Herwig's analysis and what he had done in holography," Shank said.
Eventually, the two built a semiconductor version of their device. But years of work, in their own lab and others, would be needed to figure out how to integrate the gratings into semiconductor material in an efficient way. "It took almost 17 years for it to go into an actual lightwave system," Shank said. The semiconductor DFB laser, built from layers of alloys with slightly different conducting properties, finally became practical in 1988, and now it forms a basic part of optical network systems.
"Someone once said, 'You obviously invented that too early,' " said Kogelnik. By the time the DFB laser became a hot commercial property, the Kogelnik and Shank patent had expired. Another patent he filed too early was for optical networking and switching using tunable filters and tunable lasers, in 1976. Those are just now becoming part of optical networks.
Bell Labs' Alferness points out that the DFB was important to the development of integrated optics. From practically the beginning of the laser era--at least as early as 1962--researchers were anticipating the need to put semiconductor lasers less than a millimeter in size on a semiconductor chip. What's more, they knew they had to be able to connect them with other devices, such as electrical modulators to encode the signal and switches to reroute it.
To make the mirrors on the ends of such tiny lasers, the semiconductor crystals had to be carefully cut at precise angles, a difficult achievement. Also hard was integrating such a laser into a system. But the gratings, built into the semiconductor crystal at strategic points, make the mirror ends unnecessary. They also allow engineers to shape the ends to fit with other components on the chip.
The first aim of the DFB laser was to send signals over long distances. If light pulses cover too many wavelengths, they spread out in time as they travel along an optical fiber at high speeds. The pulses will overlap, blurring into noise. But the single wavelength provided by the DFB laser solved that problem.
Tight control of the wavelength is also necessary in dense wavelength-division multiplexing (WDM), in which many wavelengths of laser light travel down a single optical fiber simultaneously. Without DFB lasers, the different channels would interfere with each other.
The first four-channel WDM systems were introduced in 1995, immediately quadrupling the capacity of the fiber in the ground. Current systems have as many as 160 channels available. Without WDM, the demand for data capacity in communications networks might have so far outrun supply that the Internet might have ground to a halt.
Think of all that bandwidth
But back in the early 1970s, Kogelnik and his colleagues were already looking into WDM, then called simply wavelength multiplexing. Kogelnik said the idea was actually around from the very beginning--it's mentioned in a 1961 memo from Kogelnik's old boss, Kompfner. By 1974, Kogelnik and his team were using slides of glass of varying indexes of refraction, coating them with a thin film and etching a grating into the film. The waveguides they thus created could combine multiple wavelengths within a single fiber.
"It's the same effect you use for distributed feedback, really, except you don't put any gain in, and then the whole thing is a filter," he told Spectrum. "Distributed feedback is very much an integral part of the WDM idea. You can hardly take them apart."
At the head of AT&T's coherent optical research department since 1967, Kogelnik was named director of the Electronics Research Laboratory in 1976, then director of the Photonics Research Laboratory in 1983. During those decades, he helped push forward AT&T's (and later Lucent's) WDM development.
If the first wave of WDM was traceable to Kompfner's speculations in the early 1960s, and the second wave arose from the distributed feedback filters of the mid-1970s, the third wave, and the one that finally led to practical systems, came in the late 1980s, with the invention of the erbium-doped fiber amplifier. That invention was developed simultaneously at Bell Labs and the University of Southampton in Britain. Its development was crucial because without an amplifier, it would be impossible to send the multiple wavelength signals far enough along the fiber to create a useful transmission system--a drawback Kogelnik had been aware of for years.
Another long-time colleague, Andrew Chraplyvy, director of lightwave systems research at Bell Labs, notes that Kogelnik understood what it was going to take to develop WDM systems. As director of the Photonics Research Laboratory, Kogelnik never balked when his people wanted to buy, say, a $300 000 or $400 000bit-error-rate test set for measuring the performance of very-high-speed transmission systems.
It was thanks to that sort of support, Chraplyvy said, that the Crawford Hill lab became the first to reach an experimental data rate of 1 terabit per second in 1996, sending 50 channels at 20 Gb/s each over 55 km of fiber. Being well-funded also let the lab invent non-zero dispersion-shifted fiber, a type of fiber that can handle the high-speed signals of the newest WDM systems. Somebody would have invented it eventually, he said, but Lucent got there first, thanks in part to Kogelnik.
"He always protected the research that was going on and was very much appreciated for that," said Hermann Haus, Institute Professor in the Research Laboratory of Electronics at the Massachusetts Institute of Technology (MIT) in Cambridge, Mass., and a friend who worked briefly with Kogelnik in 1975.
Co-workers also cite Kogelnik's intuitive understanding of the issues they bring to him and his ability to present complex ideas in elegant ways. "He would think of the problem on so many different levels and that would allow him to solve seemingly impossible problems," said Chraplyvy.
Ron Schmidt, now a consultant to start-up companies, saw the value of Kogelnik's leadership,when he worked with him to develop integrated optics, including a switch that used electricity to reroute an optical signal. "[Kogelnik is] a mentor, so sometimes you don't know what he has led you to and what you discovered yourself," he said.
Bell Labs retires its executives at age 65. Kogelnik hit that mark in 1997, and so relinquished his adminstrative duties the day before his birthday. "Now I devote myself to research and research strategy, which is what I love anyway," he said.
Some of the problems he's been tackling lately involve getting WDM systems to work at much higher capacities. As the rate at which data is transmitted down fiber increases, from rates of 10 Gb/s to 40 Gb/s expected later this year, a number of so-called nonlinearities crop up, instances in which the signal becomes distorted in hard-to-predict ways. Perhaps the biggest problem is something called polarization mode dispersion, which blurs the signal into noise.
Just last year Kogelnik and James Gordon, a pioneer of laser research, wrote a "very beautiful paper on the mathematics of polarization," according to MIT's Haus. Kogelnik himself described the paper as "very interesting, very complicated stuff."
He expects such stuff will keep him busy for the foreseeable future. "There's no end to problems that need to be solved," he said. "They still want more bandwidth and it's still hard."
But for a man who chose his field because it was the toughest course offered, what could be better?
Samuel K. Moore, Editor
About the Author
NEIL SAVAGE is a free-lance science and technology writer based in Lowell, Mass. He has written for Discover, Technology Review, and OE Magazine.
To Probe Further
For a historical view of the development of optical communications, see City of Light: The Story of Fiber Optics, by Jeff Hecht (Oxford University Press, New York, 1999).
Herwig Kogelnik described his role in that development in "High-Capacity Optical Communications: Personal Recollections." The paper appeared in the IEEE Journal on Selected Topics in Quantum Electronics, Vol. 6, no. 6, pp. 127986.
The Medalist is known for a chapter on laser fundamentals, "Propagation of Laser Beams," Chapter 6 in Applied Optics and Optical Engineering, Vol. VII, pp. 15590 (Academic Press, New York, 1979).
"Coupled Wave Theory of Distributed Feedback Lasers," by Kogelnik and C.V. Shank is a seminal paper on the subject; it was published in the Journal of Applied Physics (Vol. 43, no. 5, 1972, pp. 232735).