Oral History Transcript — Dr. Horace Furumoto
This transcript may not be quoted, reproduced or redistributed in whole or in part by any means except with the written permission of the American Institute of Physics.
This transcript is based on a tape-recorded interview deposited at the Center for History of Physics of the American Institute of Physics. The AIP's interviews have generally been transcribed from tape, edited by the interviewer for clarity, and then further edited by the interviewee. If this interview is important to you, you should consult earlier versions of the transcript or listen to the original tape. For many interviews, the AIP retains substantial files with further information about the interviewee and the interview itself. Please contact us for information about accessing these materials.
Please bear in mind that: 1) This material is a transcript of the spoken word rather than a literary product; 2) An interview must be read with the awareness that different people's memories about an event will often differ, and that memories can change with time for many reasons including subsequent experiences, interactions with others, and one's feelings about an event. Disclaimer: This transcript was scanned from a typescript, introducing occasional spelling errors. The original typescript is available.
See the catalog record for this interview and search for other interviews in our collection
Interview with Dr. Horace Furumoto
Horace Furumoto; August 19, 1987
ABSTRACT: Furumoto headed the laser development program for the Jersey Nuclear-AVCO Isotopes (JNAI) laser isotope separation project from 1972 on. In this memoir he discusses, among other issues, the original decision to use a linear-lamp-excited transverse flow dye laser; Charles Pike's injection-locked laser; how they increased the lifetimes and efficiencies of the flash lamps; how they determined some of the design parameters; how the vortex flashlamp dye laser program was terminated; the JNAI High-Density Experiment; the movement of the project from AVCO-Everett Research Laboratory in Everett, Mass, to Richland, Washington. (See also the memoirs of R. Levy and S. Janes, and the interview with A. Katrowitz.)
URANIUM LASER ISOTOPE SEPARATION THE EARLY YEARS
Experience in Dye Lasers: I started in lasers in 1963 at AVCO-Wilmington after getting my doctoral degree at Ohio State. From there I moved to the Space Optics Laboratory of NASA Electronics Research Center. I began working on flash lamp excited dye lasers when we heard Peter Sorokin of IBM had been talking to EG&G about flashlamps. We guessed what he was up to, but Sorokin beat us to the first flashlamp excited dye laser because he chose rhodamine 6G while we worked with scintillator dyes. When we learned that Sorokin got lasing with 6G, we put 6G in our laser system and it lased on the first shot.
Introduction to the Jersey Nuclear AVCO Isotopes Joint Venture: I joined the AVCO-Everett — Jersey Nuclear laser isotope joint venture in 1972. It was a proprietary program, and I only learned of the technical content after I came on board. Jersey Nuclear could see the spectroscopy and physics of laser isotope separation, LIS, coming along well, but the laser development program did not look good, because of the profusion of kinds of laser systems being suggested to them. Raymond Dickeman, the President of Jersey Nuclear, insisted on a more focused approach and I believe I was brought in to do this.
Something must be said about the original group of people who had the foresight to believe in the technology. Although LIS is a well-accepted technology today, the original advocates were considered far out or even spaced out. Drs. Richard Levy and Sargent Janes of AVCO Everett Research Laboratory, AERL, can be credited for the creation of the program. Yes, there may have been others who might have thought of using lasers for isotope separation but only Levy and Janes believed in it strongly to make it a reality.
AERL under Drs. Arthur Kantrowitz and Harry Petschek was an ideal place where such ideas could be made to grow to a large scale. Levy and Janes originally approached the AEC but the agency scoffed at the LIS concept and rejected it. Then AERL presented it to a number of oil/energy companies as well as to other large industrial corporations. For some reason, I think still not understood; Ray Dickeman quickly picked up on the idea and was willing to put down one million dollars a year for two years to see whether Levy and Janes' ideas had any validity. Jersey Nuclear pushed the other parties out and formed the joint venture on an 80-20 basis.
The AEC, later the DOE, only began to get interested when they realized something big was going on at JNAI but by that time JNAI had a fairly large program with a budget of $3.5 million per year. About that time the name of Jersey Nuclear and Esso was changed to Exxon and from here on I will refer to Jersey Nuclear as Exxon Nuclear. The joint venture, however, kept the name JNAI for Jersey Nuclear AVCO Isotopes.
AERL scientists had successfully completed a low density proof-of-principle laser isotope separation experiment on Bastille Day, 1971. They used nitrogen laser pumped dye lasers to accomplish this feat. The original JNAI plan for LIS on a commercial scale combined these small lasers to get the required average power. That would have required 70,000 lasers. Later, that plan was changed to use coaxial flashlamp-pumped dye lasers and I believe I was hired because I had experience in these lasers. But by the time I started at AERL, two new approaches were in vogue; linear lamp dye lasers with the dye flowing transverse to the laser beam, and a second transverse flow laser based on theta-pinch flashlamps. Both were fairly well thought through.
My first job was to decide between them. It was obvious to me that the theta-pinch flashlamp efficiency would be poor, because the energy that could be coupled into the discharge would be small. I just knew it wouldn't work and the theta-pinch flashlamp project died a natural death when it was determined that laser threshold could be reached but efficiency was low.
The linear lamp excited transverse flow dye laser developed by AERL was then putting out 30 pulses per second at a fairly good output — 0.3 joules per pulse. This gave about 10 watts. Drs. Jack Marling and Herb Friedman had been working on this under Levy, who was head of the overall Isotope Separation program. Janes was responsible for the physics, spectroscopy and engineering of the evaporation of uranium.
I was brought in to head laser development and direct the use of lasers in the high density proof of principle experiments. I divided the work into a set of projects. Herb and Jack worked on the flashlamp excited dye lasers, Charles Pike on master oscillators, Henry Aldag on laser hardware engineering, and another group later headed by Bert Plourde to support the lasers used in the high density experiments, HDE. The HDE experiment was the successor to the low density experiments, LDE. The LDE was used to define the spectroscopy of the atomic vapor isotope separation process. HDE were proof of principle experiments not only to show enrichment was possible, but more importantly to show that depletion to acceptable tails level was possible in a single pass at process density condition. Process density was 1013 uranium atoms per cm3.
Another thing that intrigued me on first learning about the JNAI process was the master oscillator-amplifier, MOPA, concept to be used in getting high average powers from many small lasers. Master oscillators would be used to give finely tuned output at low power and the amplifiers would increase output while preserving the fine tuning. Dr. Irv Itzkan, head of the Optics Committee, and others were working on this concept. But I had been thinking along those lines independently so I liked the approach. The oscillator would be nitrogen laser pumped dye laser. But there was a problem. Because the nitrogen laser pumped dye laser would only generate a 10 nanosecond pulse, and the flashlamp excited amplifiers had 1 microsecond long pulse, there was a mismatch in laser duration.
Charles Pike assembled a flashlamp excited dye laser ring amplifier and fed the signal from the oscillator into the ring so that it would amplify the signal in successive passes. Pike's idea was to develop a pulse-stretcher but what he had was an injection-locked oscillator. Vrehen and Breimer are usually credited for the explanation of the injection locked oscillator but Pike also had a working model but was unable to publish the result because of the proprietary nature of the program. Since Pike's system was considered a pulse stretcher, it was not clear that the narrow line width could be preserved until it was demonstrated. AERL had decided on a 3-step laser isotope separation process. Much of the early spectroscopic work on the low density experiment was done by Dr. Lawrence Levin. AERL had given up on blue excitation followed by ionization at 337 nm as not being selective enough. I concurred for other reasons.
Dye lasers were, and still are, difficult in the blue, because blue dyes are inefficient and have a degradation problem. Yellow and red dyes are better, because rhodamine dyes are efficient and long lived. That was an easy, but nevertheless critical decision. Flashlamp Excited Dye Laser, FEDL, and Development Program: In many ways the choice of a high average power tunable laser system was easy to make in the period between 1970 and 1974.
From early on it was known that depletion and not enrichment was the goal for commercial success and that a large energy per pulse passing through the uranium vapor was essential. Although there was a concentrated search for auto ionization states, nothing suitable was found. The small photoionization cross section made for saturation intensities in the order of few hundred kilowatts per square centimeter, and the high vapor densities and long paths demanded hundreds of millijoules per pulse. The only laser capable of doing that at that time was the flashlamp excited dye laser.
When Lawrence Livermore started its program a few years later, under Dr. Snavely, they also started with flashlamp lasers but dropped it when they couldn't achieve the success AERL had with these lasers and turned to copper vapor laser excited dye lasers. AERL continued to look at other types of tunable lasers, but continued to favor FEDL, primarily because of the great advances made in the JNAI program. In 1972, AVCO had about few hundred millijoules output from the pulse stretcher. We also by then proved that the injection locked oscillator preserved the narrow line width. Unfortunately, these FEDL would last for only a few minutes. The lamps would last for 5,000 shots.
The thrust at that time was to get the energy out at any cost, and the flashlamps were being overloaded. I decided to put less power per shot into the flashlamps, since that would make the lamps last orders of magnitude longer. The rule of thumb was that lifetime increases by the 8.5th power of the derating. That was not strictly true, but the exponent was in that range, somewhere between 6 and 9. I decided to make the 6" lamps used in the early Mark I and Mark II designs longer to derate the lamps and chose lamps 12", 18" and 24" long. Marling started looking into this and within a month he had data comparing the 4 different lengths. The long lamps proved to be more efficient as well as longer lived. We got better than 1% efficiency, which was unexpected. We started with a flashlamp capable of 5,000 shots and were aiming for 10 shots, 5 orders of magnitude larger. We hadn't yet decided on 500 pulses per second for the basic f1ashlamp laser, but we knew the overall system had to have a rep rate greater than 10 kilohertz. We contracted out to EG&G for flashlamp development. In this relation, most of the new ideas came from AERL, except electrode development.
After the first year of flashlamp development, about 1973, I decided to shoot for 100 watts on the strength of Marling's work on derated flashlamps. Kantrowitz loved it. We were also scaling up from 30 to 500 Hz. We got the 500 number from the rapidity with which we could flow the dye. Liquid flow velocity scales as the square root of pump pressure. We chose a pressure of 100-200 psi as convenient, calculated the flashlamp parameters, and came up with 1,000 Hz system. Then I cut that in half for prudence. That fixed our design and we also fixed upon 200 mJ per pulse, 0.03 A line width lasers for the rest of the program. Mark III was a quick and dirty attempt using Marling's laboratory results, and that didn't work well.
A significant development at this time was the use of simmer current in flashlamp discharges. Most solid state laser developers were using 50 milliamps. We heard Livermore was using currents up to 200 ma. Dr. Bruce Newell of EG&G and I discussed this and decided to increase simmer current an order of magnitude. With 2 amps simmer we increased lamp life from 5000 to 2 million shots. One day I looked at the simmer current and noticed that the arc was sitting right in the middle of the tube and not hitting the walls. That was why the lifetime increased so dramatically, because the arc never touched the wall.
The main discharge superposed on the simmer was short enough that the peak of the current pulse was over before the arc could expand to the walls. Mark IV was the first heavy-duty flashlamp excited dye laser. It was designed to operate at 500 Hz and 100 watts average power, but because of hot spots in the flow channel, we only got 80 - 90 watts. Input power into the dye lasers was 50 kW. What happened was that after lasing started, a tiny amount of dye fluorescence light would hit the flow channel wall and heat it up.
The Mark IV, which we completed in about 1974, was a failure despite a lot of analysis that went into the design. We had the Mark V out 8 - 9 months later, in 1975. It was a very nice laser which eventually Exxon Nuclear in Richland used to produce 1 kW average power in the green. Those involved in that demonstration were Morton, Dragoo, McAllister and Drake of Exxon Nuclear. Mark VI, developed in 1976, a planar waveguide FEDL was in my judgment a better laser, but Exxon didn't want it because it was less efficient because it operated on lower order modes. I think that was a mistake, but by then AERL had little influence on the laser part of the JNAI LIS program. There was also a vortex flashlamp dye laser being built by Dr. Michael Mack, who came to AERL from United Technologies. He had a small group which included Donna Northam and Les Crawford. Mack had worked on vortex flashlamps at United. This displeased Exxon because it was a return to what was, in their view, too many laser concepts. Mack thought his system was better, and he forced a challenge at a research committee meeting. I took a nominal position of neutrality, since I was in charge of both.
A jury was set up of AERL and Exxon people. But I gave Aldag, in charge of the linear dye laser development, advice on the arguments to use. I told him to emphasize that the efficiency is better in the linear lamp and the dye degradation worse in the vortex lamp. Dye degradation rate became the main issue in the choice of laser systems and the linear lamp approach was chosen. The vortex lamp was actually a very nice system, and could reach higher repetition rates because the arc diameter was smaller. Either system could have worked for laser isotope separation. The High Density Experiments: I was also responsible for the laser part of the High Density Experiment, HDE, as well as the laser development program. Janes was responsible for the generation of uranium vapor, the diagnostics, the magneto-hydrodynamic separation of isotopes after excitation and ionization and such. In the Isotopes Research committee meetings, decisions would be made and trade-offs made between different approaches. A laser engineer would say, for example, "We've got only 30 pulses per second lasers." Janes would say "Use 3 lasers with pulse separation at 50 microseconds and then I can put a chopper into the system at such and such a place to cut off the rest of the vapor."
Kantrowitz would always say why 3, why not 5 and we would compromise at 4. He had an enthusiasm for huge things, while I like small, table-top things. I enjoyed those meetings. You expressed an idea and everyone tried to shoot it down. People were informal and sharp, with no restraints. Kantrowitz and Petschek and other key AERL people would attend. LIS was something of a sore subject at AERL because the meetings were closed to other scientists not on the project. We in LIS, however, could attend theirs, and that caused some friction. For the HDE, with 1013 uranium atoms in the interaction zone, the first step of our 3-step process (4 including the 620 cm level) was a red beam, a very astute choice.
The second excitation step was yellow and the third ionization step was the same yellow color. By this, we avoided a third master oscillator. The energy would be 10 millijoules for the first and second and 300 millijoules for the third. The second color, though broad (.3 A) was centered at the second excitation color. That part in resonance with the second excitation step was used for excitation and whatever spilled over from resonance could be used for ionization. That was clever. We could meet the oscillator specifications using AERL's Dial-a-Line laser. These were the first commercial dye lasers and were nitrogen laser excited. But I felt that since we had to develop a laser for the commercial process anyway, we might as well develop a special oscillator for the HDE.
Professor Ali Javan, our consultant, argued with me about it, recommending we just use the HDE to prove feasibility at high density rather than use it as a platform for laser development. And in fact, it was really hard to get all the lasers working together for the HDE. We needed 32 lasers and getting 32 lasers to work successfully in synchronism is difficult. Yet several thousand were needed for the commercial process and I thought it best to get experience early on how to multiplex. The laser system I designed for the HDE was never put into full operation. We completed the experiment itself, but it was by dint of working around the laser deficiencies. I think we got the good data in the third year of the HDE using 9 lasers rather than 32 lasers. Charles Pike looked at the oscillators again. We needed 20 master oscillators for the commercial factory if pulsed master oscillators were used at 500 Hz. It would have been nicer to have just one.
CW dye lasers were just invented. This opened the possibility that we could use a single oscillator per color. Spectra-Physics and Coherent announced commercial versions and I went to the West Coast and tried to buy one. They would only sell with no guarantee. Then Spectra-Physics and Coherent took their CW dye lasers off the market, so we were really in a bind. Pike built one of our own but of course the windows gave problems, but at least we had finely tuned light. Still we had no enrichment. We were working 24 hours a day and we joked about whether if you put in more than 24 hours it should be counted as overtime or direct time. The oscillator would go in and out of resonance. If the lasers worked, the uranium vapor part would not. For example, we'd start at 8:00 a.m. It would take until 2 or 3 p.m. for the laser to be tuned, and then the vapor people, who had been waiting, would get to work and the uranium evaporator would crash because a droplet of uranium would get on the e-beam filament. Finally, one day I saw absorption dip in U-238, in about mid-1973. I used a silicon diode and a Simpson meter. We did our measurements on U-238 because the absorption signal was larger, and then offset to the U-235 line when we needed depletion.
We started the HDE with Mark II flashlamp excited amplifiers; very handsome, stainless steel devices. We got no enrichment. Tuners for the second color at that time were prisms since we only needed 0.3 angstrom. The tuners were overheating at 5 watts to cause the beam to break up and the frequency was moving back and forth with the net effect that it smudged out. We reduced the rep rate from 30 Hz to 3 Hz and finally saw enrichment. It was by then mid-1973. The operation of the HDE got smoother and a remarkable amount of data was being collected. By the end of 1973, depletion on a single pass was measured by macroscopic assays and the U235 content was calculated to be less than 0.3% after correction for laser duty cycle. Question as to lasing of the uranium vapor from higher excited states was quickly resolved by an experiment. Levin looked at emission cross sections from upper excited states and predicted a 60 cm column of U238 would lase near 2 microns. We slapped in IR mirrors and got lasing on the first shot. It was in my experience that the easiest laser ever invented, though of course undesirable.
In 1974, on Halloween day, we got ionization and depletion without turning the ionization laser on. Even Professor Hans Bethe was initially baffled but we soon realized that we had Rydberg atoms and residual electron impacts had given us the separation. This was many years before the topic became popular in laser spectroscopy. Other Aspects of the JNAI Program at AERL: The LIS program was highly invigorating.
There was interesting ground being covered in atomic spectroscopy, in fluid flow, collective phenomena, propagation through dense gases, lasers, engineering aspects of hot corrosive metals, and the economics of fuel processing. JNAI was always concerned about the molecular approach. There was a bothersome outside patent covering that approach.
Late in 1973 there was quite a bit of controversy about the desirability of the atomic vapor process. But by that time AERL had pretty much solved the atomic vapor problems and held steadfast in their approach. Then when Sarge Janes came up with a way to increase evaporation efficiency by nearly a factor of two, we were confident and let the controversy go on outside the JNAI program. We did, however, have a small molecular vapor program just so as not to short change ourselves and did 16 micron laser analysis and low temperature expansion analysis. We always came up with atomic vapor being the better approach especially if depletion was considered. I don't think JNAI was ever seriously concerned about the molecular vapor approach despite all the scare stories of outsiders, much of it amusing in retrospect. Sarge Janes' cylindrical geometry e-beam and uranium evaporator was a tour de force. I believe the chief beneficiary of this technology is the Department of Energy AVLIS program.
The optical system for LIS in the JNAI scheme was built around 2 rotating mirrors per color. The master oscillator had a rotating element distributing its light to all the amplifier lasers. Then we'd recombine the outputs of the amplifiers with another rotating mirror. I never liked it, however, and I transferred this part of the work to AERL's optical engineering group led by Dr. Michael Smotrich. This was a group outside Levy's jurisdiction and had a contract directly from JNAI. I had in mind to push the program in another direction once I had the lasers, and eliminate the rotating mirrors, especially the combiner mirror. Anyway, the original rotating mirror set-up was called the "Stonehenge" because several thousand amplifiers were arranged in a circle about a kilometer in diameter firing at the beam combiner in the center. From the beginning, I wanted a simpler and cheaper system.
The AERL team had a scheme to divide the 100 meters long uranium absorption path into 30 evaporation modules and use one 3 color beam to handle them all by multiple passing the beam through the uranium vapor. All the modules would have to function simultaneously. But if any module crashed, the whole system crashed so there was an elaborate optical system that could be put into place to bypass a module. I wanted one cheap laser system per module, but this idea was always shot down because it was considered too simple.
There was the proliferation issue. The powers to be didn't want to develop a process that was simple because they were afraid they would not get a license to get into enrichment because questions of proliferation would be raised. Dickeman was the key decision maker on this issue. There was always a question whether the government would allow private industry into uranium enrichment. Everyone was confident that Exxon's wealth and power would get them through. The AT&T divestment judgment which occurred in the mid-70's, however, seemed to shake the faith of the Exxon executives. Technology Transfer from AERL to Exxon Nuclear: Exxon Nuclear became more and more involved in the technical decisions as time went on. With the start of the High Density Experiment in 1972, Exxon had made a full commitment. It had spent $2 million to that point. From then it spent $5-6 million a year at AERL. AERL was the source of ideas. In 1974 Exxon Nuclear decided to build up the JNAI staff to manage the program at AERL. JNAI staff members were really Exxon Nuclear staff.
By 1974, there was talk of a test facility at Exxon's Richland Washington plant. Levy, meanwhile, had been made a Corporate Vice-President at AVCO Corporation and moved out of LIS. He was not there long and moved to Exxon Nuclear when the Exxon Nuclear contribution to the LIS program began to grow. I saw this as improving our chance to keep the project at Everett. But in the discussions to keep it at Everett, Levy turned up on the other side. Eventually Dr. Harold Forsen, who was a consultant to Exxon Nuclear from the University of Wisconsin, was put in charge of JNAI.
The decision was made to move the project to Richland, Washington. Forsen hired Dr. Malcolm Stitch for the laser program and gave him the charter to build a laser development group for JNAI. This was not to the liking of AERL because AERL had had some difficulty with Stitch on a previous program; the change occurred in 1975. Stitch pushed for and obtained a big budget. He did a good recruiting job, hiring a lot of good people. My goal was to have a pilot plant in 1978-80 based on flashlamp lasers, but under the new management they moved it back to 1982.
Meanwhile, some technical breakthroughs were going on outside of the JNAI project. The copper vapor laser was getting much better and Lawrence Livermore AVLIS program chose this approach. AERL pushed for a 1 KW, 2 KHz excimer laser system in 1976 but the development program was vetoed by Exxon Nuclear. Exxon Nuclear wouldn't allow us to publish on any of the lasers or vapor technology. The Mark II and Mark V results would have been publishable and of general interest, but it was kept as proprietary information.
At the end of the project, the results of the 1 KW demonstrations were published. The dozen or so Mark V units at Richland were torched. Approximately 4 Mark V's at AERL were retained and are still in use. (In answer to a question on funding) I had about 1.5 million dollars a year for laser development over the course of about 5 years and in addition about $1million/year to run the laser side of the High Density Experiment.
The total team at AERL on lasers for LIS was about 25-30 people with another 20 or so on the vapor technology spectroscopy enrichment and extraction. We also had a dye chemistry group which eventually moved to Richland.