First-Hand:The First Continuous Visible Laser
It is difficult to recreate these days the sense of excitement and expectency, and the enormous interest that arose during the early days of the laser in 1959-62. Gordon, Zeiger and Towne's demonstration of gain and oscillation at microwave frequencies had shown inverted populations were a potent new source of coherent electromagnetic energy, and Townes and Schalow had pointed the way to extend this result to optical frequencies. But the sticking point was picking the right medium to use. Numerous suggestions for suitable media appeared in the literature. Often, these were contingent on one or more difficult-to-measure parameters being known with some precision. Ruby, caesium, xxxx were all proposed and some rejected. Picking the right medium was important; no one wanted to waste time studying a medium unless there was some indication that inversion was possible. Inversion seemed to be a rare phenomena presumably because nature seemingly favored thermal equilibrium everywhere. In retrospect, this seems a quaint idea. Now we know if one dumps enough energy very quickly into a system with distinct energy levels, one or more pairs of them will usually invert. One of the seemingly poorer candidates, ruby, later turned out to be the first visible laser to operate successfully; Theodore Maiman was the first to discover the advantages of using a giant pulse of energy to invert a pair of energy levels, and thereby was awarded the honor of being the first over the finish line. One of the better candidates (proposed by Ali Javan), was an electrical discharge in a mixture of neon and helium gases, producing inversion in a pair of neon levels by excitation transfer from excited helium atoms. Gas discharges are quite complicated mixtures of relatively simple electronic excitation processes whose light producing properties have been studied for a long time. One of the earliest publications was by Duffenback who reported the enhancement of certain neon spectral lines in a HeNe dischage through collisions of the second kind. It was this enhancement that caught Javan's attention. Indeed, Javan's intuition was correct; measurements by Bennett confirmed the possibility of inversion between the neon 2s2 and 2p4 (Paschen notation) levels of neon resulting from enhanced transfer of excitation energy to the neon 2s2 level from the23S metastable helium level at almost the exact same energy level. The transition wavelength was in the near infra-red at 1.15 microns. The experimental confirmation of inversion and oscillation was a tour de force. Don Herriott's optical expertise was crucial in the experiment. Curved mirror optical cavities with their large angle tolerance where not well understood those days, so Herriott had to struggle with the problem of aligning plane mirrors (enclosed within a vacuum tube) to within a fraction of an arc-second to find a high Q cavity resonance. Patience paid off in the end and oscillation at 1.15 microns was achieved in Dec 1960. It wasn't visible light, but from the Bell Labs point of view it was coherent and continuous, both telephone company imperatives.
Like Maiman's ruby laser, the infra-red HeNe laser created a stir thruout Bell Labs and the scientific community. The era of coherent light energy had begun. To many, it was the Holy Grail. Herwig Kogelnik, responsible for putting a firm foundation under much that we presently know about laser cavities and laser beams put it well. "Think of all that bandwidth!" Strangely enough, given it's wide usage today in so many fields, the laser was first publicized in the popular media as a 'solution looking for a problem'.
Like most other laboratories, the US Army Signal Corps was fascinated by the potential of lasers for communications. Soon after the infrared HeNe made it's public appearance, the Signal Corps decided it needed one of it's own to play with, so they placed a request with Bell Labs to have one built and shipped down to the Signal Corps facility near Red Bank, NJ. They requested all the latest bells and whistles for their tube which meant Brewster windows, a concave mirror cavity and a D.C. discharge. For one reason or another, the group that originally received the request passed it along to our group. We had just completed development of a gas discharge "talking path" diode which was designed to replace the mechanical relay in switching networks and most of the equipment for making and processing gas discharge tubes was still intact. Ray Sears, our department head at the time, approached us with the Signal Corps request and asked if we were interested. Needless to say, we (that is, Dane Rigden and myself) jumped at the chance to become involved with gas lasers. Dane had experience in optics and I had been working with gas discharges for several years, so the request was right down our alley. The group originally approached by the Signal Corps helped us get started by providing us with Brewster windows and infrared mirrors. Our glass shop quickly learned how to make the tubes and within a few weeks we had the first model set up on the lab bench and running. Since the beam was invisible, we set up an infrared detector and developed a scanning procedure to detect the onset of oscillation. Since the optical cavity was a spherical mirror cavity, mirror alignment was a relatively simple task and it wasn't too long before our tube was generating a few milliwatts of coherent 1.15 micron light. One of the first things we noticed was that the power output was very sensitive to dust on the Brewster windows. If these weren't cleaned regularly, the output power would gradually weaken. There were minor problems with the thoriated tungsten filament we were using as a hot cathode, so after making some design changes and requesting a new tube from the shop, we decided to come back to the lab in the evening and see what we could learn about the properties of the discharge. We also had some half-baked ideas about combining optical pumping with d.c. discharge pumping to increase the power output which we wanted to explore. A priority in our efforts to learn more about the discharge was to take a spectral scan of the spontaneously emitted side light from a neon discharge tube with and without added helium. We were limited to the visible spectrum by the spectrometer we were using, but the scans were a revelation. There were major intensity changes in regions of the neon spectrum when helium was added to the discharge and less dramatic changes in the helium spectum. At the same discharge current, the intensity of a number of neon lines increased dramatically. Our source of information on spectral lines and energy levels for helium and neon was the 1957 edition of the American Institute of Physics Handbook. One neon line immediately caught our attention. It was the red 632.8 nm neon line, corresponding to the the 3s2 to 2p4 (Paschen notation) transition. A relatively weak line in the pure neon spectrum, its intensity increased more than 50 fold when helium was added to the neon discharge. Clearly something interesting was going on. Superimposing the helium energy level diagram onto neon level diagram showed us very quickly the likely origin of the enhancement. The upper 3s2 neon energy level was nearly coincident with the 21S helium metastable energy level. Except for differences in absolute values, this coincidence in energy levels was remarkably similar to the coincidence that existed between the 2s2 neon level and the 2s3 helium level, the source of inversion in the 1.15 micron laser. At this point in our explorations things were clearly beginning to look interesting. Our first thought was that the remarkable enhancement of the red 632.8nm neon line must have been noticed before and probably published, so we made a search of the recent (English language) literature. The search came up empty, but it seems unlikely this effect had never been noticed before. Perhaps a search of early 20th C European literature would be more fruitful. In any case the time we could spend investigating the line and its level lifetimes was limited to evenings and weekends. Any idea of repeating Bennett's lifetime measurements on the new energy levels seemed out of the question. We had neither the instrumentation nor the expertise to do so. The question then became what to profitably do next? We did have a phase sensitive amplifier available so in principle we could make gain/loss measurements of the red line using one HeNe tube as a source and a similar tube as a gain/loss tube. In principle it is a simple measurement to make, but the major problem we faced was getting the source beam of light straight along the axis of the gain/loss tube with enough signal to make a measurement. We were able to make measurements, but the signal was very weak. At best, we concluded, the discharge showed transparency but the results were not good enough to claim gain. Still, this was encouraging. One could always think about ways to increase gain by changing the discharge parameters or the tube geometry. In any case we decided to take the next step; get a pair of high-reflectivity concave mirrors coated to reflect the red line and a pair of mirror mounts, and look for oscillation. Acquiring the mirror blanks and mounts was not a problem; we purchased them from Edmund Scientific quite inexpensively. Getting the mirrors coated with special low-loss dielectric coatings was going to be expensive, perhaps as much as $200, but we were able to persuade our supervisor, Gordon Cooper, to sign the purchase order. Dane, not wanting to give away our plans, ordered the mirrors coated at 636.0 nm, close enough to 632.8 nm for high reflectivity but not so close others could guess what we were up to.