KMS Fusion, an early ICF group mostly remembered for its mashup of good science and bad politics. This is our historical addendum.
An excellent article on the beginning of ICF (inertial confinement fusion) was posted on January 28th by ScienceLine’s Chelsey Coombes. This is a summary of Alex Wellerstein‘s presentation at the October 2014 meeting of the New York Academy of Science.
Wellerstein is a respected science historian specializing in American Classified research, and is located at Stevens Institute of Technology.
Coombes’ (and Wellerstein’s) good discussion of the early days of inertial confinement fusion traces development of the LLNL indirect drive approach and the KMS fusion direct drive alternative (see Fig 1). We also discuss, in passing, ICF efforts at LANL.
We add a missing link to the KMSf story, one that I have never seen mention in any published description of the early days of fusion energy research. This small band of researchers produced some of the best results in ICF physics in the 1970s and even the ’80s, but their leadership made stunningly bad political mistakes and succeeded in being ignored by History.
KMSf lost government funding in 1990 and closed as a fusion research and support organization. Since then, their science results have been marginalized, their people disparaged.
Click any image for full resolution.
Fig 2 is our diagram of the direct and indirect drive interaction that is central to the ICF process. We show a simple shell; LLNL, in particular, designs highly complex configurations.
I have not seen any configuration that is more successful than the one shown here, although NIF estimates that its target provides perhaps 7½% lower growth for RT instabilities than this.
The target is a fuel-containing sphere, with a hard outer shell; the beam strikes the shell and causes ablative burn-off of the outer layer. The explosively burned off shell-plasma generates the reactive force that compresses the fuel layer into a compact core, for fusion density and temperature.
Fusion energy research started in the 1960s by two different physicists independently developing their ideas along two very different paths for ICF.
A “drive beam” could be from a laser, a particle accelerator (different types of ion beams have been tried or discussed), or, it could be radiation from a nearby explosion.
Hohlraum geometry was a severely classified concept until the early 1990s; but in the late 1980s, a German scientist decided to describe it to me, personally, to show know he knew everything, despite U.S. secrecy.
John Nuckolls In the late 1950s, Nuckolls (Fig 3) developed the idea of using a laser to cause a small capsule to implode, reach the condition needed for fusion, and generate power. He did this work in the ultra classified hydrogen bomb effort at LLNL and shared the ideas with other bomb workers in 1960.
His idea – use a variant of the classified hohlraum idea and focus laser beams onto the cavity’s inner surface.
Nuckolls’ work was born classified and stayed that way. But, in 1972, he was allowed to publish a carefully composed article in the journal Nature. By 1988, when the New York Times published a clear description of hohlraums, every interested physicist in the world knew how indirect drive ICF worked.
Nuckolls became head of the LLNL laser fusion program that built the series of laser fusion test labs, including the current NIF facility. In 1988 he became Director of the laboratory but was forced to resign in 1995 – management improprieties.
Keith Brueckner In 1969, Brueckner (Fig 4) worked out the basic physics for what is now called the direct drive technique, to separate it from hohlraum indirect drive method.
Brueckner was a theoretical physicist at UCSD in San Diego and had also worked at Los Alamos (LANL) in New Mexico, among other places. His LANL experience was the justification for the government attempt to label direct drive as classified material.
He was allowed to publish details in the early 1970s, but only after Nuckolls had been allowed to publish is indirect drive work.
By the end of the late 1960s, Brueckner was in discussion with Keeve M (Kip) Siegel, professor at the University of Michigan, and a pugnacious risk-taker.
Siegel (Fig 5) sold off his resources from his patents and many companies, to form KMS fusion, Inc. By 1973, the group had built Chroma, the then-largest (IR) laser in the world.
In 1974, they published neutron yield data demonstrating the first laser-driven fusion results of any type, and demonstrated that their proprietary method worked. This embarrassed the LLNL effort that had worked longer and used H-Bomb technologies.
The KMSf results used direct drive IR (infra red) beams from Chroma and the results were taken skeptically by the outside fusion community. The KMSf team never claimed they had made a fusing core, just that they had neutron evidence that laser beams had caused fusion reactions. The results were valid, but they generated anger and even venom.
KMS fusion’s ICF Technology
The key to getting good implosion structure is to have completely uniform illumination all about the surface of the target shell.
“Hot” spots in a beam will cause dimples to form on the shell and the implosive compression can be highly non-symmetric. The central collapse region will become a core of multiple knots, none reaching ignition conditions – each knot so small that ions can drift out of it before they participate in fusion. Inertially confined ions require that the fusion region be large enough so that thermal speeds will not dissipate a knot before the ions fuse.
DBIS for uniform illumination The KMS technique was to use what they called their double bounce illumination system (DBIS) meaning that 2 mirrors were needed so that one beam can be expanded and reflected to strike 1/2 of the target. The original beam is pre-split into two beams which symmetrically illuminate the entire target surface.
Fig 6 shows the original DBIS mirror cavity with Kent Moncur, the Head of the Chroma laser program. His group developed the first shaped pulse capability in laser fusion, along with its variable pulse width capability. He was supported by the entire talented group of optical engineers and technicians.
Gar Busch, in the laser group, tapped a portion of the Chroma beam to make the uniquely valuable holographic interferometer system that took up to 4 snapshots during a pulse that displayed shape and density contours of the plasma. Fig 7 is one record for a flat plate receiving laser power from opposite directions. Our post NIF-3 shows a different shot.
I am proud to have worked with the entire team of talented and innovative laser engineers.
The DBIS-1 mirror was damaged a bit more every pulse; it was upgraded by DBIS-2.
Fig 8 shows the path that the laser beams follow through the mirror chamber and to the target. The main beam is split into Left and Right beamlets. The paths are matched so that the arrival-on-target time is identical to within a tiny fraction of a nanosecond. (Firm numbers are no longer available to me.) Only the Right beam path is shown for simplicity.
Each beam goes through a lens that focuses it through a small hole in its mirror and immediately expands to cross the target and reflect off the opposite mirror surface. The mirrors are carefully figured a-spheres and the reflected rays spread out further to the opposite mirror which focuses the beam directly onto the target surface. The double bounce is required to properly spread the beam for nearly uniform convergence over the target, with the least power damage to the mirrors.
Fig 9 shows the dramatically beautiful art that was at the entrance to the building lobby. The beams are entering the illumination chamber from both right and left directions, but, again, only the right side is fully detailed.
The DBIS-2 mirror segments are shown pulled apart for clarity, the paths are not to scale.
Fig 10 shows the DBIS-2 mirror chamber cleaned up and mounted for display after DOE informed management that the 1986 sequence was the last target implosion fusion test campaign KMSf would ever be allowed to run.
The display was built by volunteers, a work of love by the employees. I took a DBIS diagram from one of the annual reports and sketched a 3D diagram. Larry Fleeman, a young designer in Engineering made a careful CAD diagram using my drawing and also engineering prints. The commercial artists did sketches of the device and used the CAD results for the accurate painting. Meanwhile, Clark Charnetski painstakingly rebuilt the device that had been in storage for a several years, and assembled it into its best form. All the wiring was correct and the glass and mirror surfaces were cleaned with care. Clark was the master of the optical system, his skill made our implosion tests successful. He was responsible for both DBIS -1 and -2. The Design Engineering Head, Bill Groves, constructed the display stand, made up of carefully hand selected special wood, pieced together and hand rubbed to a museum-like finish. He made the seamless acrylic box cover and hinged it to the stand.
DBIS-2 operated in the final 1986 sequence where results demonstrated the very important need for targets with cryogenic (cold solid) DT fuel on the inner wall, shaped laser pulses (Intensity proportional to t2) to generate near-adiabatic compression, the need for short pulse times, and special care toward symmetry of targets and illumination schemes. These results were too spectacular for the time and were discounted. But these are lessons NIF seems to be re-inventing.
Roy Johnson and his team published the 1986 campaign results (Phys Rev A, 41, 2 (15 Jan 1990), pp 1058-1070) after very careful data analysis, and computer modeling. They were very aware that their results would be criticized and they did a thorough job. Good scientists worked at KMSf from its start. Good, innovative scientists worked there when it was closed.
Technical postscript. When I was with the National Center for Manufacturing Sciences in the early 1990s, I visited the Los Alamos National Laboratory. In one of the hallways I passed through, I saw our display stand. It had been bashed in with splintered sections and the finish as marred and gouged. The Acrylic top was missing, as was any sign of DBIS-2. When I asked about this piece of junk, a LANL physicist just laughed. I asked about the mirror chamber. “It makes a great boat anchor” said he. How did this reach LANL hands?
By the late 1970s, it was well understood that IR (long wavelength infra red) beams made fusion success impossible. The IR beam loses much of its power to heating the electrons in the abating plasma, which in turn, preheats the imploding fuel. The back pressure in the fuel prevents an ignitable core from ever forming. The immediate fix was to frequency double the deep, deep red light to green. By the mid 1980s, virtually every ICF lab was investigating frequency tripled light into the blue. Los Alamos started their innovative Aurora KrF laser to directly emit UV (short wavelength ultra violet) beams for high compression efficiency, and had began start-up before 1990. I never learned why this underfunded program was terminated, nor what happened to its KrF facility.
KMS fusion Conclusion
Management retrospective KMSf management started with respected physicists, who engaged in very silly politics. In the late 1960s and early 1970s, the Department of Defense tried to classify their work, tried to take away even the original notes, and blocked them from access to laser fusion information. In retaliation, KMSf minimized government access to its lab.
I was told an amusing story several times by the older physicists and engineers in our group. After the 1974 neutron results, LLNL staffers said that what was done was impossible and demanded that they be allowed on-site to review the data. This was not unreasonable. But the KMS staff wanted revenge, so they declared DBIS to be Corporate Classified material, that is, deep proprietary. They moved it to a closet during the visit and stationed guards at the door, to emphasize the point. Pretty funny, right?
ICF fusion leaders never forgave. When the DBIS information was released, there was no way the bunch of KMS hyenas could do anything right. Kip Siegel had his wealth and even insurance invested in KMS, he gave his all by stroking out at a 1975 Congressional hearing during a plea for support. So KMSf survived, but he did not. Pretty funny, right?
Ownership change Top management changed about 1980 when (very) non-technical entrepreneurs from Canada bought the company and saved it from closing. I was at General Atomic when a couple real estate sales guys from Colorado, the Blue brothers, bought it and saved GA from closing. I thought this would be a disaster but leadership change worked for GA – very well indeed. Wish I could say the same for KMSf.
IR lasers cannot drive fusion KMS really needed to upgrade its Chroma. Frequency doubling to green caused a crucial loss of power delivered to target and, even though the Target Development Group had invented ways of making low mass shells with highly uniform cryogenic DT fuel layers, target physics demanded more power for best success.
A uniform solid DT target layer is due to a self correcting process named Beta Heating by the discoverers, scientists in our target group. Recently I have seen it referred to by a different name. Maybe to divorce the process from KMS associations?
CEO to the rescue(?) Our CEO (& Chairman-For-Life of the Board of Directors) took up the challenge and called all the laser labs to ask for support. His way to generate support was to inform them of the consequences of dissent. For example, a lab leader like Bob McCrory, head of LLE (Rochester New York) was told that unless he supported our upgrade, KMS would focus its congressional lobby to assure that LLE lost major funding. Funny story, right?
I was at an ICF leadership conference in the late 1980s. I was half way across the room but recall Bob bouncing up and down on his toes screaming at our Tim Henderson (VP at the time, and one of our experienced fusion engineers). He was yelling something like: You think I can forget? Huh? I will NEVER forget! over and over, inches from Tim’s face. I was new to ICF management and the shock blanked out of my memory his exact words. That was when I understood, truly, how much a fool our owner was.
The KMSf final program review by DOE was a joke. Some physicists, such as Marshall Rosenbluth, had no intention of wasting his time – so he filed a devastating review on material he had never read and did not show up to see. The only real question – how fast could KMS be taken down? (This is where Ed Gabl first presented possibly important results showing plasma jets driven from regions near laser spots.)
Roy Johnson’s 1986 publication on the DBIS-2 campaign was indeed discounted by the ICF community.
Appropriate scientific response to spectacular but uncertain results would be to insist on a repeat of the experiment with outside observers and diagnostics that the critics would accept. The actual response was not scientific criticism, it was herd-behavior bloviation. (oops! I meant to add IMHO.) The upstart KMS-ers dared to use data to differ from those by LLNL and its direct-drive sidekicks at LLE, such things were not meant to happen.
I happened to lead a tour for a visiting congressman in 1989. During that walk-through, he told me that he knew that Chroma was obsolete: too small to be at all useful for American research. During our final review, the KMSf target team was characterized as incapable of understanding what they were doing, had actually schemed to make bad targets for its customers at LLNL, LANL, and LLE. In addition: The physics team was made of of disconnected people who never worked together, never did anything innovative and were not worth consideration.
KMSf, the first and only private company to be engaged in ICF studies, closed in 1990.
- In 1991, General Atomics (the new target manufacturer) eagerly hired every single person from the target group they could convince to move to California.
- Two thirds of the obsolete Chroma laser was sent to Los Alamos to be dedicated as the Trident facility, where it has been upgraded over the years and is still doing valid research.
- I am not sure what happened to all in the physics staff. I do know of several who were hired directly and several others who refused to work in government projects again, but I lost contact with nearly everyone.
No one has ever tried to build another visible light implosion cavity similar to DBIS. In fact, we in America seem to have put ICF fusion on hold, waiting for our intrepid LLNL miracle workers. NIF estimates were very clear. Based on their LASNEX wonder code, they knew then would have success by 2012. Didn’t happen
In late 1989, as Head of the physics group, I made contact with the Smithsonian to see if we could donate DBIS-2 and DBIS-1 as a part of our American technical heritage. I made several contacts with a Mr Bernard Finn, who responded as though he were talking to a crackpot. I was informed that physics historian Paul Forman, the nominal contact for this kind of gift would too busy to be involved with our discussion. Several other KMS people also contacted Mr. Finn. In my last contact, Mr. Finn asked me to send written documentation about our achievements, maybe he would get back to me. My memory of this is clear – he seemed to be chuckling during this conversation. … our custodians of American science history.
Root causes I see KMSf as a sad story, an early demonstration of over enthusiastic pronouncements prior to actual data arriving – their code phrase in 1974 was “on line by ’79.” NIF is a current demonstration; the companies in our Paths Not Taken sequence are others. But it also is a brutal consequence of the harsh cutbacks in American funding for fusion. As we have pointed out many times, funding was appropriated in 1980 but the Reaganites cut the legs off our fusion programs – an amputation that is going strong today. Our culture draws ever closer to our day of reckoning for our change of the American social balance.
I have mentioned only few of the people who should be recognized and I apologize; it’s been many years and I have certainly missed some who should be in this list, especially those working to make the wonderful DBIS entry display. Points of pride for us all – KMS work with physics testing; KMS work with laser development; KMS excellence in target development and manufacture; and our volunteer efforts at the end to showcase our company.
The KMSf images here are my own digitized photographs. They are all high resolution TIFF files, many tens of megabytes large, with (much) smaller JPG images. You can request a copy by clicking the [LastTechAge] menu under our banner and select [Send Message].
UPDATE: The pictures are archived online at box.com at
Everyone who links into box.com, will find all of my images there. Some are nicely small and some are huge. If you have trouble with that site, let me know!
You should be able to freely view and download the files, but, to protect the file integrity, no one may upload into the archive without special permission. I would be very pleased if others would care to contribute to the image archive. Use this post and send me a request for download rights. There must be hundreds of images still existing. Wouldn’t it be nice to be able to view them? – Charles Armentrout
Our final topic is the speculative idea sketched in the Postscript after the sign off.
Most comments seem to have come to our obsolete and abandoned website.
To see all those other interesting comments, click here
Charles J. Armentrout, Ann Arbor
2015 Feb 15
Listed under Technology… a post in the thread > Fusion–ICF
Have a comment? Click on the title of this post, go to bottom, let us know.
Related posts: Click the INDEX menu under the Banner picture
Postscript: A speculative idea
In early 1991, Tim Henderson (VP) called a meeting of senior physics staff for an idea exchange. What would we do if we could do what we wanted?
To me, the main problem with our DBIS-2 was (A) mirror damage by beam hot spots and (B) cost of replacing the damaged mirrors.
I sketched a DBIS upgrade, very similar to Fig 11, here. This was inspired by the multi-mirror telescope proposals of the time. I still think it might hold merit.
We would use a single beam (to maintain temporal symmetry) that is split into two beamlets for entry into the evacuated mirror reflection cavity, maybe 2 to 3 m in diameter.
Inside would be a scaffold holding tightly packed flat mirrors, perhaps 1 cm across. The proper figure would be maintained by the scaffold and the mirrors would be aimed by piezo actuators. When the surface of any mirror was damaged, it could be replaced by an inexpensive clone. The advantage of small flat mirrors is light is not focused into a pinpoint, but into a wider target-sized area. This spreads beam hot spots across the target, and will not drive RT dimples. Many such computer controlled mirrors would be needed and I am not sure how tightly the mirrors would have to fit together to send most of the drive beam onto target.