The world as I see it
RSS icon Email icon Home icon
  • No Ignition at the National Ignition Facility: A Post Mortem

    Posted on March 21st, 2015 Helian No comments

    The National Ignition Facility, or NIF, at Lawrence Livermore National Laboratory (LLNL) in California was designed and built, as its name implies, to achieve fusion ignition.  The first experimental campaign intended to achieve that goal, the National Ignition Campaign, or NIC, ended in failure.  Scientists at LLNL recently published a paper in the journal Physics of Plasmas outlining, to the best of their knowledge to date, why the experiments failed.  Entitled “Radiation hydrodynamics modeling of the highest compression inertial confinement fusion ignition experiment from the National Ignition Campaign,” the paper concedes that,

    The recently completed National Ignition Campaign (NIC) on the National Ignition Facility (NIF) showed significant discrepancies between post-shot simulations of implosion performance and experimentally measured performance, particularly in thermonuclear yield.

    To understand what went wrong, it’s necessary to know some facts about the fusion process and the nature of scientific attempts to achieve fusion in the laboratory.  Here’s the short version:  The neutrons and protons in an atomic nucleus are held together by the strong force, which is about 100 times stronger than the electromagnetic force, and operates only over tiny distances measured in femtometers.  The average binding energy per nucleon (proton or neutron) due to the strong force is greatest for the elements in the middle of the periodic table, and gradually decreases in the directions of both the lighter and heavier elements.  That’s why energy is released by fissioning heavy atoms like uranium into lighter atoms, or fusing light atoms like hydrogen into heavier atoms.  Fusion of light elements isn’t easy.  Before the strong force that holds atomic nuclei together can take effect, two light nuclei must be brought very close to each other.  However, atomic nuclei are all positively charged, and like charges repel.  The closer they get, the stronger the repulsion becomes.  The sun solves the problem with its crushing gravitational force.  On earth, the energy of fission can also provide the necessary force in nuclear weapons.  However, concentrating enough energy to accomplish the same thing in the laboratory has proved a great deal more difficult.

    The problem is to confine incredibly hot material at sufficiently high densities for a long enough time for significant fusion to take place.  At the moment there are two mainstream approaches to solving it:  magnetic fusion and inertial confinement fusion, or ICF.  In the former, confinement is achieved with powerful magnetic lines of force.  That’s the approach at the international ITER fusion reactor project currently under construction in France.  In ICF, the idea is to first implode a small target of fuel material to extremely high density, and then heat it to the necessary high temperature so quickly that its own inertia holds it in place long enough for fusion to happen.  That’s the approach being pursued at the NIF.

    The NIF consists of 192 powerful laser beams, which can concentrate about 1.8 megajoules of light on a tiny spot, delivering all that energy in a time of only a few nanoseconds.  It is much larger than the next biggest similar facility, the OMEGA laser system at the Laboratory for Laser Energetics in Rochester, NY, which maxes out at about 40 kilojoules.  The NIC experiments were indirect drive experiments, meaning that the lasers weren’t aimed directly at the BB-sized, spherical target, or “capsule,” containing the fuel material (a mixture of deuterium and tritium, two heavy isotopes of hydrogen).  Instead, the target was mounted inside of a tiny, cylindrical enclosure known as a hohlraum with the aid of a thin, plastic “tent.”  The lasers were fired through holes on each end of the hohlraum, striking the walls of the cylinder, generating a pulse of x-rays.  These x-rays then struck the target, ablating material from its surface at high speed.  In a manner similar to a rocket exhaust, this drove the remaining target material inward, causing it to implode to extremely high densities, about 40 times heavier than the heaviest naturally occurring elements.  As it implodes, the material must be kept as “cold” as possible, because it’s easier to squeeze and compress things that are cold than those that are hot.  However, when it reaches maximum density, a way must be found to heat a small fraction of this “cold” material to the very high temperatures needed for significant fusion to occur.  This is accomplished by setting off a series of shocks during the implosion process that converge at the center of the target at just the right time, generating the necessary “hot spot.”  The resulting fusion reactions release highly energetic alpha particles, which spread out into the surrounding “cold” material, heating it and causing it to fuse as well, in a “burn wave” that propagates outward.  “Ignition” occurs when the amount of fusion energy released in this way is equal to the energy in the laser beams that drove the target.

    As noted above, things didn’t go as planned.  The actual fusion yield achieved in the best experiment was less than that predicted by the best radiation hydrodynamics computer codes available at the time by a factor of about 50, give or take.  The LLNL paper in Physics of Plasmas discusses some of the reasons for this, and describes subsequent improvements to the codes that account for some, but not all, of the experimental discrepancies.  According to the paper,

    Since these simulation studies were completed, experiments have continued on NIF and have identified several important effects – absent in the previous simulations – that have the potential to resolve at least some of the large discrepancies between simulated and experimental yields.  Briefly, these effects include larger than anticipated low-mode distortions of the imploded core – due primarily to asymmetries in the x-ray flux incident on the capsule, – a larger than anticipated perturbation to the implosion caused by the thin plastic membrane or “tent” used to support the capsule in the hohlraum prior to the shot, and the presence, in some cases, of larger than expected amounts of ablator material mixed into the hot spot.

    In a later section, the LLNL scientists also note,

    Since this study was undertaken, some evidence has also arisen suggesting an additional perturbation source other than the three specifically considered here.  That is, larger than anticipated fuel pre-heat due to energetic electrons produced from laser-plasma interactions in the hohlraum.

    In simple terms, the first of these passages means that the implosions weren’t symmetric enough, and the second means that the fuel may not have been “cold” enough during the implosion process.  Any variation from perfectly spherical symmetry during the implosion can rob energy from the central hot spot, allow material to escape before fusion can occur, mix cold fuel material into the hot spot, quenching it, etc., potentially causing the experiment to fail.  The asymmetries in the x-ray flux mentioned in the paper mean that the target surface would have been pushed harder in some places than in others, resulting in asymmetries to the implosion itself.  A larger than anticipated perturbation due to the “tent” would have seeded instabilities, such as the Rayleigh-Taylor instability.  Imagine holding a straw filled with water upside down.  Atmospheric pressure will prevent the water from running out.  Now imagine filling a perfectly cylindrical bucket with water to the same depth.  If you hold it upside down, the atmospheric pressure over the surface of the water is the same.  Based on the straw experiment, the water should stay in the bucket, just as it did in the straw.  Nevertheless, the water comes pouring out.  As they say in the physics business, the straw experiment doesn’t “scale.”  The reason for this anomaly is the Rayleigh-Taylor instability.  Over such a large surface, small variations from perfect smoothness are gradually amplified, growing to the point that the surface becomes “unstable,” and the water comes splashing out.  Another, related instability, the Richtmeyer-Meshkov instability, leads to similar results in material where shocks are present, as in the NIF experiments.

    Now, with the benefit of hindsight, it’s interesting to look back at some of the events leading up to the decision to build the NIF.  At the time, government used a “key decision” process to approve major proposed projects.  The first key decision, known as Key Decision 0, or KD0, was approval to go forward with conceptual design.  The second was KD1, approval of engineering design and acquisition.  There were more “key decisions” in the process, but after passing KD1, it could safely be assumed that most projects were “in the bag.”  In the early 90’s, a federal advisory committee, known as the Inertial Confinement Fusion Advisory Committee, or ICFAC, had been formed to advise the responsible agency, the Department of Energy (DOE), on matters relating to the national ICF program.  Among other things, its mandate including advising the government on whether it should proceed with key decisions on the NIF project.  The Committee’s advice was normally followed by DOE.

    At the time, there were six major “program elements” in the national ICF program.  These included the three weapons laboratories, LLNL, Los Alamos National Laboratory (LANL), and Sandia National Laboratories (SNL).  The remaining three included the Laboratory for Laser Energetics at the University of Rochester (UR/LLE), the Naval Research Laboratory (NRL), and General Atomics (GA).  Spokespersons from all these “program elements” appeared before the ICFAC at a series of meetings in the early 90’s.  The critical meeting as far as approval of the decision to pass through KD1 is concerned took place in May 1994.  Prior to that time, extensive experimental programs at LLNL’s Nova laser, UR/LLE’s OMEGA, and a host of other facilities had been conducted to address potential uncertainties concerning whether the NIF could achieve ignition.  The best computer codes available at the time had modeled proposed ignition targets, and predicted that several different designs would ignite, typically producing “gains,” the ratio of the fusion energy out to the laser energy in, of from 1 to 10.  There was just one major fly in the ointment – a brilliant physicist named Steve Bodner, who directed the ICF program at NRL at the time.

    Bodner told the ICFAC that the chances of achieving ignition on the NIF were minimal, providing his reasons in the form of a detailed physics analysis.  Among other things, he noted that there was no way of controlling the symmetry because of blow-off of material from the hohlraum wall, which could absorb both laser light and x-rays.  Ablated material from the capsule itself could also absorb laser and x-ray radiation, again destroying symmetry.  He pointed out that codes had raised the possibility of pressure perturbations on the capsule surface due to stagnation of the blow-off material on the hohlraum axis.  LLNL’s response was that these problems could be successfully addressed by filling the hohlraum with a gas such as helium, which would hold back the blow-off from the walls and target.  Bodner replied that such “solutions” had never really been tested because of the inability to do experiments on Nova with sufficient pulse length.  In other words, it was impossible to conduct experiments that would “scale” to the NIF on existing facilities.  In building the NIF, we might be passing from the “straw” to the “bucket.”  He noted several other areas of major uncertainty with NIF-scale targets, such as the possibility of unaccounted for reflection of the laser light, and the possibility of major perturbations due to so-called laser-plasma instabilities.

    In light of these uncertainties, Bodner suggested delaying approval of KD1 for a year or two until these issues could be more carefully studied.  At that point, we may have gained the technological confidence to proceed.  However, I suspect he knew that two years would never be enough to resolve the issues he had raised.  What Bodner really wanted to do was build a much larger facility, known as the Laboratory Microfusion Facility, or LMF.  The LMF would have a driver energy of from 5 to 10 megajoules compared to the NIF’s 1.8.  It had been seriously discussed in the late 80’s and early 90’s.  Potentially, such a facility could be built with Bodner’s favored KrF laser drivers, the kind used on the Nike laser system at NRL, instead of the glass lasers that had been chosen for NIF.  It would be powerful enough to erase the physics uncertainties he had raised by “brute force.”  Bodner’s proposed approach was plausible and reasonable.  It was also a forlorn hope.

    Funding for the ICF program had been cut in the early 90’s.  Chances of gaining approval for a beast as expensive as LMF were minimal.  As a result, it was now officially considered a “follow-on” facility to the NIF.  No one took this seriously at the time.  Everyone knew that, if NIF failed, there would be no “follow-on.”  Bodner knew this, the scientists at the other program elements knew it, and so did the members of the ICFAC.  The ICFAC was composed of brilliant scientists.  However, none of them had any real insight into the guts of the computer codes that were predicting ignition on the NIF.  Still, they had to choose between the results of the big codes, and Bodner’s physical insight bolstered by what were, in comparison, “back of the envelope” calculations.  They chose the big codes.  With the exception of Tim Coffey, then Director of NRL, they voted to approve passing through KD1 at the May meeting.

    In retrospect, Bodner’s objections seem prophetic.  The NIC has failed, and he was not far off the mark concerning the reasons for the failure.  It’s easy to construe the whole affair as a morality tale, with Bodner playing the role of neglected Cassandra, and the LLNL scientists villains whose overweening technological hubris finally collided with the grim realities of physics.  Things aren’t that simple.  The LLNL people, not to mention the supporters of NIF from the other program elements, included many responsible and brilliant scientists.  They were not as pessimistic as Bodner, but none of them was 100% positive that the NIF would succeed.  They decided the risk was warranted, and they may well yet prove to be right.

    In the first place, as noted above, chances that an LMF might be substituted for the NIF after another year or two of study were very slim.  The funding just wasn’t there.  Indeed, the number of laser beams on the NIF itself had been reduced from the originally proposed 240 to 192, at least in part, for that very reason.  It was basically a question of the NIF or nothing.  Studying the problem to death, now such a typical feature of the culture at our national research laboratories, would have led nowhere.  The NIF was never conceived as an energy project, although many scientists preferred to see it in that light.  Rather, it was built to serve the national nuclear weapons program.  It’s supporters were aware that it would be of great value to that program even if it didn’t achieve ignition.  In fact, it is, and is now providing us with a technological advantage that rival nuclear powers can’t match in this post-testing era.  Furthermore, LLNL and the other weapons laboratories were up against another problem – what you might call a demographic cliff.  The old, testing-era weapons designers were getting decidedly long in the tooth, and it was necessary to find some way to attract new talent.  A facility like the NIF, capable of exploring issues in inertial fusion energy, astrophysics, and other non-weapons-related areas of high energy density physics, would certainly help address that problem as well.

    Finally, the results of the NIC in no way “proved” that ignition on the NIF is impossible.  There are alternatives to the current indirect drive approach with frequency-tripled “blue” laser beams.  Much more energy, up to around 4 megajoules, might be available if the known problems of using longer wavelength “green” light can be solved.  Thanks to theoretical and experimental work done by the ICF team at UR/LLE under the leadership of Dr. Robert McCrory, the possibility of direct drive experiments on the NIF, hitting the target directly instead of shooting the laser beams into a “hohlraum” can, was also left open, using a so-called “polar” illumination approach.  Another possibility is the “fast ignitor” approach to ICF, which would dispense with the need for complicated converging shocks to produce a central “hot spot.”  Instead, once the target had achieved maximum density, the hot spot would be created on the outer surface using a separate driver beam.

    In other words, while the results of the NIC are disappointing, stay tuned.  Pace Dr. Bodner, the scientists at LLNL may yet pull a rabbit out of their hats.

    ICF

    Leave a reply