Piece of the Sun : The Quest for Fusion Energy (9781468310412) Page 21

  Livermore’s Shiva laser was finished in 1977 and was slowly ramped up to its full energy of 10 kilojoules in its first year. Although the laser was a success, the physics of pumping laser power into a tiny fusion target did not turn out to be as simple as the lab’s computer simulations had suggested. The following year, Rochester started using the first six Omega beamlines in a proof-of-design setup they called Zeta. Researchers there soon encountered the same problems. Apart from the laser-plasma interactions encountered by earlier lower-energy experiments, another problem cropped up too: once the implosion got started the fuel inside the capsule didn’t want to stay in there and had a habit of breaking through the ablator and squirting out in unpredictable directions.

  This phenomenon, called Rayleigh-Taylor (RT) instability, is best understood by imaging a dish containing a layer of oil with a layer of water on top of it. This is an unstable situation because water is the denser liquid and, because of gravity, would naturally sit at the bottom but the oil is in its way. The two layers could remain as they are but if there is any ripple or momentary irregularity in the thickness of the water, a ‘finger’ of oil will push upwards through it and there will be a similar downwards thrust of water and the two layers will rapidly change places. In an imploding capsule, the dense ablator is the water, the D-T fuel is the oil and the applied force is the implosion driver. What researchers wanted was for the unstable denser-on-top arrangement to hold until the fuel is compressed, but that requires no irregularities in the ablator layer or in the applied pressure of the driver which would give the fuel an opportunity to burst up and out.

  Making the ablator layer absolutely smooth and symmetrical is a matter of better manufacturing. Making the driver absolutely even is much trickier. As laser facilities had got bigger they had sprouted many more beams. Shiva boasted twenty beams and Rochester’s Omega would have twenty-four. This meant that they could shine beams at the target from many directions to get an even overall coverage. But that was not enough to prevent RT instabilities. If you imagine looking at a laser beam end on, coming straight towards you, because of various optical effects the beam does not have an even intensity across its face and when the beam is focused down to a tiny spot on the target’s surface that unevenness allows RT instabilities to develop.

  Livermore’s answer to the RT instability problem was to use hohlraums, the radiation cases like tiny tin cans that are modelled on the much larger hohlraums used in H-bombs, because then you don’t have to fire the laser beams onto the target at all. In fusion experiments a hohlraum is about the size of the eraser at the end of a pencil and is made of a heavy metal such as gold. The hohlraum has holes at either end and the spherical fusion capsule is held at its centre. In a laser shot the beams are fired in through the hole at each end at an angle so that they don’t touch the capsule but instead hit the hohlraum’s inside wall. The laser pulse heats the gold of the hohlraum to such a high temperature that it emits a pulse of x-rays and it is these that cause the capsule’s ablator to blow off and implode the fuel. The idea is that the conversion from laser light to x-rays smoothes over any unevenness in the laser beams. But this technique, known as indirect drive, does have drawbacks. Firstly the hohlraum, being cylindrical in shape, is far from symmetrical, so you have to be very clever with your placement of laser beams to get uniform x-ray illumination of the target. It’s also inefficient: much of the laser energy is lost when the light is converted into x-rays. And indirect drive does not get around the problem of hot electrons and plasma interfering with the incoming laser beams; now it was plasma blowing off the hohlraum that was getting in the way. Despite these problems, the Livermore researchers had worked with hohlraums for years, understood them and believed that they could make them work.

  Indirect drive at the National Ignition Facility: how 192 laser beams cause a fusion reaction.

  (Courtesy of M. Twombly/Science. Reprinted with permission from AAAS.)

  The Rochester researchers, by contrast, mostly had backgrounds in optics and lasers rather than weapons design, so their inclination was to try to fix the unevenness of the laser beams. Shining laser beams straight onto the target – known as direct drive – was simpler and more efficient since much more of your beam energy goes into the implosion. As a result the laser fusion community in the United States split into two camps, with direct drive proponents led by Rochester and the Naval Research Lab while Livermore became the champion of indirect drive.

  To combat the problems of laser-plasma interactions and hot electrons, researchers were beginning to realise that what they needed were new lasers with shorter wavelengths. Experiments in France had shown that laser beams with short wavelengths are absorbed better by targets and caused less pre-heating of the fuel by electrons. Rochester began a concerted effort to find other types of laser, but none could produce the right combination of high energy, short pulses and short wavelength.

  However, during 1979 and 1980, Rochester researchers came up with the next best thing: a way to convert the infrared light from Nd:glass lasers to a shorter wavelength. They used a crystal called potassium dihydrogen phosphate, or KDP, which has a very useful property: if a beam of light is shone into it, the crystal structure of KDP interacts with photons of light in such a way that two photons can be merged to create a new photon with twice the energy and hence half the wavelength. This phenomenon only takes place in certain crystals and with light of high intensity, so it wasn’t observed until 1961, just after the discovery of the laser. The laser fusion researchers at Rochester could thus convert the 1.054-micrometre light from their Nd:glass lasers to 0.532 micrometres, green visible light. The problem was that it only converted 43% of the beam’s energy to the shorter wavelength, but the theorists at Rochester worked out a more elaborate scheme. After the first conversion they passed the new green beam and the remaining unconverted infrared beam into a second KDP crystal. Here the different photons combine to create new ones with three times the energy of the original ones and hence one-third of the wavelength, 0.351 micrometres, which is in the ultraviolet. The researchers found that if they carefully set the crystal orientations and the polarisation of the light in a particular way they could convert 80% of the original beam energy into ultraviolet light.

  This opened up exciting new possibilities and LLE experimentalists started investigating how ultraviolet light would work as a fusion driver. But soon their thoughts were elsewhere: Moshe Lubin, LLE’s leader and driving force, resigned to take up a job in industry. Standard Oil of Ohio, which was one of LLE’s main corporate sponsors, offered him the post of vice president for research and it was too tempting to refuse. Jay Eastman, the chief engineer who had spearheaded the building of Omega, became LLE’s new director but he also stepped down in late 1982 to set up a company making barcode scanners. Just when the lab had a new technique that showed such promise, it found itself in administrative turmoil.

  From their experiments with UV light it became clear that laser fusion research at Rochester was not going anywhere unless they added a KDP-crystal conversion system to all twenty-four beams of the new Omega laser. It was a difficult task going back to the Department of Energy – which had taken over responsibility for fusion research in 1977 from the short-lived ERDA – to admit that their brand-new Omega laser wasn’t working as predicted and that they needed to install expensive wavelength converters to it. Following much negotiation, the DoE finally agreed to a phased conversion over three years but it insisted that some economies had to be made at the lab. As conversion of the twenty-four beams began, lab managers had to cut more than 20% of the staff – an unpleasant process in the small, tightly knit LLE team; and some believed that LLE would never recover.

  Laser fusion as a whole had hit a rough patch. Expanding at breakneck speed during the 1970s, spurred on by the oil crisis, the US government spent $1 billion on the field during that decade. At the beginning of the 1980s things looked very different: not only was oil cheap but the problems that the labs
were having getting targets to implode successfully did not make it look like energy break-even was going to arrive any time soon. General Graves’ prediction to Congress in 1975 that break-even would be achieved around 1980 now looked hopelessly optimistic. In 1981, the government’s budget request for laser fusion research went down for the first time.

  At the beginning of 1983, Robert McCrory was appointed as the new director of LLE. A veteran of Los Alamos, he had joined LLE in 1976 just as the construction of Omega was starting. Now he brought a measure of stability to the lab and it gradually got back on its feet. A few months after his appointment, the conversion of the first six beamlines to UV operation was finished and all twenty-four beamlines were converted by the autumn of 1985. The researchers now had to learn how implosion worked all over again with the new shorter wavelength of light. While the UV beams helped with reducing laser-plasma interactions and getting more energy onto the target, the team still had to grapple with RT instabilities – how to ensure that the implosion proceeded smoothly and symmetrically. Shining the beam directly onto the target had to be a better system than the complexity of indirect drive; they just had to work on their lasers to make them more uniform.

  The Rochester team came to realise part of the problem was that laser beams were too perfect. Because the waves are all perfectly in step and all one wavelength, any imperfection in the optics is simply propagated along with the beam to the end target. What they needed to do was rough it up a bit – a bit of fuzziness in the beam’s properties might help smooth over imperfections. One way they did this was to add an optical device at the end of the beamline called a distributed phase plate. This chops the beam into 1,500 beamlets and, while doing so, applies small random time delays to each one. The beamlets are then focused onto the target but, instead of making 1,500 tiny spots, they are all superimposed onto the same spot. Because all of these beamlets are now slightly out of phase with each other, there is an averaged-out illumination over the whole of the spot, which covers up imperfections. Another technique they developed, called smoothing by spectral dispersion, did a similar job with the beam’s wavelength – roughing it up to cover over non-uniformities.

  With the UV light and the new beam smoothing techniques, the LLE researchers were by 1988 able to compress targets to between 100 and 200 as dense as liquid D-T, a goal set for them by DoE in 1986. They were, in their mind, well on the way to demonstrating direct-drive laser fusion as a viable source of energy.

  At Livermore in the late 1970s, Shiva was struggling. Because of the unforeseen complications of laser-plasma interactions, hot electrons and RT instabilities it was not performing as the researchers’ computer codes had predicted. What the Livermore researchers did have on their side was very good diagnostic equipment to figure out what was going on during the implosions. They measured, they tweaked their models, and they made modifications. They adjusted the focusing of the beams into the hohlraum, increased the hohlraum size and changed the shape of the pulse. In the end they were able to achieve the nominal goal of compressing targets to 100 times the density of liquid D-T, but the amount of neutrons produced was much less than predicted and they remained very far from energy break-even.

  In time-honoured fashion, their solution to the problem was to build another, more powerful laser. Again using Nd:glass technology producing infrared light, the proposed Nova laser would generate twenty beams with a total energy of 200 kilojoules (up from Shiva’s ten kilojoules) and cost some $200 million. John Nuckolls was convinced that this one would achieve ignition.

  In 1979 the DoE assembled another panel to make an assessment of its inertial confinement fusion programme. Headed by John Foster, a former Livermore director, it had to consider, among other things, whether to approve Nova. But the researchers at his old lab were already reassessing their design. Like their colleagues at Rochester they had seen the evidence that shorter wavelengths reduced laser-plasma interactions. When they heard about Rochester’s invention of wavelength conversion to ultraviolet using KDP crystals they realised that it would be foolhardy to build Nova without such converters, but adding them seriously inflated Nova’s already substantial price tag. Foster’s committee didn’t play along. It recommended that DoE hold Livermore to its original cost estimate and that the lab should pay for the extra frequency converters by reducing the number of beams from twenty to ten. The goal would no longer be ignition but to refine predictions of what beam energy would be required to get to ignition.

  Around this time occurred the most startling and yet most shrouded part of the inertial confinement fusion programme. The DoE set out to test the viability of inertial confinement fusion using nuclear bombs. Starting in 1978 a series of small underground nuclear tests at the Nevada test site were carried out by teams both from Livermore (which they dubbed the Halite series) and Los Alamos (known as Centurion). The aim was to use the explosions as a source of x-rays and to test DT fusion capsules of various sizes to see how much driver energy is necessary to reach energy break-even. The Halite-Centurion programme went on for ten years and its results remain classified but what information has leaked out suggests they found that 20 megajoules of x-rays are needed to reach break-even. Since hohlraums are only around 20% efficient at converting an incident laser pulse into x-rays on the target, this translates into a laser of 100 megajoules – an energy that was far beyond the state-of-the-art then and still is today. While the amount of energy seemed daunting, the results at least gave researchers confidence that ignition via laser fusion was possible.

  Nova was completed in 1985 and, as the researchers had hoped, using ultraviolet instead of infrared light led to much more of the laser energy reaching the target and much less pre-heating of the fuel by hot electrons. After a few years of operation the Livermore researchers were able to achieve cleanly symmetric implosions to 100 times liquid DT density. But the energy they were getting onto the target was a far cry from Nuckolls’ planned 200 kilojoules for Nova. Cutting the number of beams from twenty to ten halved that figure and the addition of wavelength converters – which were 50% efficient – halved it again to 50 kilojoules. The researchers were not even able to reach that level, however, because they had to moderate the power to prevent damaging the laser optics. So Nova was held back to powers no greater than 30 kilojoules.

  Discerning readers will have noticed a pattern in Livermore’s mode of operation: computer models make bold predictions about future achievements; a new laser is built; it either underperforms or the plasma physics proves more complicated than predicted, or both; models are refined and make new bold predictions; another larger laser is built; and so on. True to form, Livermore began planning for its next big machine but this time there were more than the usual number of complications. First, their computer models were much more optimistic in terms of the beam energy required to get to ignition than the results of the Halite-Centurion experiments suggested. Livermore’s models predicted that energy of a few megajoules would be enough, while the nuclear tests had pointed to 100 megajoules. Even if Livermore was right, to go straight from Nova’s 30 kilojoules to more than a megajoule was a huge leap for Nd:glass technology which many laser experts didn’t think could be done in one jump. Part of the motivation for that big jump was that laser fusion was becoming increasingly important to weapons scientists who were constantly pushing for higher energy implosions. But big also meant expensive and the amount of money required for the next generation of machine – approaching $1 billion – was making the programme very visible to politicians in Washington, DC.

  Livermore’s proposal for its next laser was the Laboratory Microfusion Facility (LMF) which would have a 10 megajoule driver laser and would produce an energy output of between 100 and 1,000 megajoules – a gain of 10 to 100. At this time Rochester was also lobbying for a much more modest upgrade of its Omega laser to further demonstrate the viability of direct-drive fusion. To help decide what to do next, the National Academy of Sciences was asked to set up a panel to re
view the laser fusion programme in 1989 and 1990. Headed by Steven Koonin of the California Institute of Technology, the panel listened to both proponents and technological doubters and concluded that the LMF was too great a technical step and too costly to pursue at that time. Instead, the panel recommended an intermediate step, that Livermore should build a ‘Nova upgrade’ with a laser of a few megajoules of energy at less than half the cost of LMF. Such a facility, the panel said, would probably be able to achieve ignition and even modest gain of between 5 and 10 (down from LMF’s 10 to 100) – despite the fact that there was no evidence ignition could be reached with such a beam apart from Livermore’s own optimistic computer projections. The panel did, however, hedge its bets: recognising that there was no firm evidence of which target technology – direct or indirect drive – would ultimately prove more successful, it also recommended that Rochester should be funded to upgrade Omega from twenty-four beams to sixty, boosting its energy from 10 to 30 kilojoules, for the knock-down price of around $60 million.