The road from Alamogordo

The road from Alamogordo

INDEX TERMS Nuclear weapons|testing, safe alternatives;Defence|nuclear, safe testing;Computers|nuclear-testing simulation;

DATE 24-Jun-95

WORDS 1675

IN JULY 1945 the first nuclear test lit up the sky over Alamogordo, New Mexico, more or less as powerfully as theory had predicted; the Manhattan project's designers had got it right first time. Over the past 50 years this success has been built on with at least 2,000 further tests of nuclear weapons around the world. You might think that, with all this experience, the weaponeers now have a pretty good idea of what is going on in the heart of their bombs. But some still want more. France has announced plans for eight new tests, enraging people in the South Pacific. America is wondering whether to prolong its moratorium on tests. Fifty years on, is there really no alternative?
The obvious place to look for one is inside a computer. Computer models, too, trace their origins back to the Manhattan project. They are a great help to weapons designers - who have, in return, been a generous source of money for supercomputer builders. But even supercomputers need data before they can get down to work. The traditional source for that data is nuclear testing, but various alternatives and supplements have been devised over the years. Those now on offer are pretty good - not good enough, perhaps, for someone who wants to design a sophisticated new weapon from scratch, but good enough to verify the effectiveness and reliability of an already tested bomb design.
Liquid assets

There are three sources of data other than bombs. Hydrodynamic (HD) experiments reveal the behaviour of a warhead from the point at which it is triggered to the point at which a nuclear chain reaction begins. Inertial-confinement fusion (ICF) uses lasers to heat and compress little beads of deuterium and tritium - isotopes of hydrogen distinguished from the run-of-the-mill stuff by the extra neutrons in their nuclei. These heavy hydrogen nuclei then fuse as they would in a hydrogen bomb. The last source, hydronuclear (HN) experiments, is the trickiest; HN experiments are aborted bombs, their yields reduced to that of kilograms of high explosive, rather than kilotonnes.
Nuclear bangs start with a squeeze. In fission weapons balls of metal are squashed together, making the free-flying neutrons within them more likely to bang into, and thus split, the metal's heavy atomic nuclei. Fusion weapons squeeze together light nuclei rich in neutrons.
Fission comes first. Nuclear bombs use chemical explosives to squeeze fissile isotopes of uranium and plutonium - isotopes with nuclei that are easily split. In HD testing sturdier, non-fissile isotopes, such as uranium-238 and plutonium-242, are treated in much the same way, subjected to enough pressure and shock that they start to behave like liquids (hence the 'hydro' in hydrodynamic). Gas-powered guns fire them into targets at enormous speed, pneumatic devices squash them, diamond anvils crush them. In the most faithful duplication of the conditions in a weapon, a shell of chemical explosives is stuck to a non-fissile blob and ignited, the shock waves compressing the metal just as they do in a real explosion.
The scientists observe this compression with tiny sensors and large X-ray machines. These machines have to be phenomenally good; both uranium and plutonium are better X-ray shields than lead, so the challenge is harder than taking an X-ray of the centre of a lead block. At the moment, scientists can get only a snapshot view of the core - and that only in one direction. Soon, they hope to do better. 'One goal is to create holographic images of the core imploding,' says Paul Brown of Lawrence Livermore National Laboratory in California.
Putting these data into their computers, Dr Brown and his colleagues can work out how well the high explosive burns, whether it ignites symmetrically, and how the metal behaves as it is compressed. But because they are using sturdy nuclei, there are no nuclear chain reactions within the compressed metal - and, in the end, nuclear reactions are what the bomb is about.
This is where HN comes in. In these experiments, fission is allowed to get under way. Neutrons split nuclei, thus producing more neutrons, which split more nuclei, and so on: the N, after all, is for nuclear. There are various ways of stopping this reaction before it goes too far. Impurities in the fissile material can leave the chain reaction unable to sustain itself; so can a neutron-absorbing gas in the core. The compression can be weakened, or applied unevenly, so that the fissile material blows itself apart before giving out much energy. But despite all this cleverness there is always what the scientists call the 'whoops' factor; a bomb intended to fizzle occasionally misbehaves, and winds up producing a full-fledged nuclear explosion. HN experiments, unlike HD ones, are seldom done in laboratories.
Because they provide insights into processes not seen in HD experiments, HN tests have a wider variety of applications. They can be used to test the reliability of weapons, to improve the safety of warheads, and to devise methods of disarming terrorist nuclear bombs. For instance, a safety experiment might involve igniting one point of the high-explosive shell; such a situation might occur because of a fire. Because of the asymmetry of this explosion, it should not trigger a nuclear detonation - but there is only one way to find out for sure. Disarming a nuclear bomb is a similar sort of problem.
The line between an HN experiment and a nuclear test is a fuzzy one. Livermore and Los Alamos, America's other weapons-design laboratory, developed the technique so that they could keep up research during a previous moratorium at the end of the 1950s. America now says that any nuclear explosion that gives off less energy than two kilograms (four pounds) of high explosive is a hydronuclear experiment and should be exempted from a future test-ban treaty. France, on the other hand, argues that bannable tests begin at yields of 100 or 200 tonnes, and anything smaller is fair game. This is hardly hydronuclear: it is the size of a small tactical nuclear weapon.
Light elements

In most modern bombs, fission is just the beginning. The force of a primary fission reaction is used to squeeze deuterium and tritium together in a more powerful secondary fusion reaction. Neither HD nor HN experiments help much here; by the time the second-stage fusion reaction gets started, a bomb is well past the 'whoops' phase and into the 'uh-oh' one.
To simulate the second stage without a first stage, physicists use a different sort of squeeze, hammering samples of heavy hydrogen from all sides with beams of light or sub-atomic particles. It is not necessary to go all the way to fusion for this approach to be useful; heating and compressing the hydrogen into an electrically charged plasma provides useful data in itself.
Livermore has plans for a new ICF machine called the National Ignition Facility (NIF), which will routinely set the fires of fusion burning. The NIF is to be the most powerful system of lasers in the world; for three billionths of a second at a time, it will focus 500,000 gigawatts, split among 192 laser beams, on to deuterium and tritium targets a centimetre or so across. It is a big enough project, at $ 1.1 billion, to make the difference between keeping Livermore open and closing it down. The French have plans for a similar facility outside Bordeaux as part of PALEN - a weapons-experimentation programme set to gobble up FFr10 billion ($ 2 billion) over the next five years.
Hopes and fears

To justify such spending, enthusiasts point to a possibility much more alluring than better replacements for nuclear tests: fusion power. A pellet ignited with the aid of a laser can give out approximately 11 times the amount of energy put in. But there is a lot of engineering still to be done before this technique finds commercial expression. Pellets have to be very smoothly and precisely made if they are to implode correctly. The energy, given out in a burst, has to be collected and stored. And the lasers need to operate ceaselessly; at the moment it takes hours for their optics to cool down and their capacitors to charge up. The world already has one set of expensive prototype fusion reactors, which control plasmas with magnetic fields; it is not clear that it needs two.
Even with America's NIF and France's PALEN, weapons designers would still not really be able to model the explosion of a nuclear warhead. 'You simply cannot get complete information without (nuclear) testing,' says Patricia Lewis, head of the Verification Technology Information Centre in London. None of the experiments models the interaction between the first and second stages, nor do they give valuable insight into fancier nuclear effects such as 'boosting' fission yields with a small injection of deuterium and tritium. They can let you test ways in which bombs might fail; but they cannot show you the full extent of a bomb's possible success.
This suggests that the real motive for further testing is to find designs for new weapons. The French have complained that their current TN70, 71 and 75 warheads are 'sensitive' because they are highly optimised; that is, they are prone to failure because they are so finely tuned. But this fine tuning is the result of their having been developed with the help of extensive testing. And that makes it hard to believe that still more tests are needed - unless there are new warheads in the works, such as small warheads for new M-5 submarine-launched missiles and ASLP cruise missiles.
The French deny that they want to design a new warhead, and argue that the eight tests they plan are anyway too few for such a purpose. It has previously taken France 13 or more test explosions to design a weapon. But Dr Brown says that, thanks in part to better simulation, America can design a weapon with as few as four tests. Eight might well be enough for France. It only took one for the Manhattan project.
© 1995 The Economist Newspaper Limited. All rights reserved

			The road from Alamogordo

	INDEX TERMS		Nuclear weapons\|testing, safe alternatives;Defence\|nuclear, safe testing;Computers\|nuclear-testing simulation;
	DATE		24-Jun-95
	WORDS		1675

			IN JULY 1945 the first nuclear test lit up the sky over Alamogordo, New Mexico, more or less as powerfully as theory had predicted; the Manhattan project's designers had got it right first time. Over the past 50 years this success has been built on with at least 2,000 further tests of nuclear weapons around the world. You might think that, with all this experience, the weaponeers now have a pretty good idea of what is going on in the heart of their bombs. But some still want more. France has announced plans for eight new tests, enraging people in the South Pacific. America is wondering whether to prolong its moratorium on tests. Fifty years on, is there really no alternative? The obvious place to look for one is inside a computer. Computer models, too, trace their origins back to the Manhattan project. They are a great help to weapons designers - who have, in return, been a generous source of money for supercomputer builders. But even supercomputers need data before they can get down to work. The traditional source for that data is nuclear testing, but various alternatives and supplements have been devised over the years. Those now on offer are pretty good - not good enough, perhaps, for someone who wants to design a sophisticated new weapon from scratch, but good enough to verify the effectiveness and reliability of an already tested bomb design. Liquid assets There are three sources of data other than bombs. Hydrodynamic (HD) experiments reveal the behaviour of a warhead from the point at which it is triggered to the point at which a nuclear chain reaction begins. Inertial-confinement fusion (ICF) uses lasers to heat and compress little beads of deuterium and tritium - isotopes of hydrogen distinguished from the run-of-the-mill stuff by the extra neutrons in their nuclei. These heavy hydrogen nuclei then fuse as they would in a hydrogen bomb. The last source, hydronuclear (HN) experiments, is the trickiest; HN experiments are aborted bombs, their yields reduced to that of kilograms of high explosive, rather than kilotonnes. Nuclear bangs start with a squeeze. In fission weapons balls of metal are squashed together, making the free-flying neutrons within them more likely to bang into, and thus split, the metal's heavy atomic nuclei. Fusion weapons squeeze together light nuclei rich in neutrons. Fission comes first. Nuclear bombs use chemical explosives to squeeze fissile isotopes of uranium and plutonium - isotopes with nuclei that are easily split. In HD testing sturdier, non-fissile isotopes, such as uranium-238 and plutonium-242, are treated in much the same way, subjected to enough pressure and shock that they start to behave like liquids (hence the 'hydro' in hydrodynamic). Gas-powered guns fire them into targets at enormous speed, pneumatic devices squash them, diamond anvils crush them. In the most faithful duplication of the conditions in a weapon, a shell of chemical explosives is stuck to a non-fissile blob and ignited, the shock waves compressing the metal just as they do in a real explosion. The scientists observe this compression with tiny sensors and large X-ray machines. These machines have to be phenomenally good; both uranium and plutonium are better X-ray shields than lead, so the challenge is harder than taking an X-ray of the centre of a lead block. At the moment, scientists can get only a snapshot view of the core - and that only in one direction. Soon, they hope to do better. 'One goal is to create holographic images of the core imploding,' says Paul Brown of Lawrence Livermore National Laboratory in California. Putting these data into their computers, Dr Brown and his colleagues can work out how well the high explosive burns, whether it ignites symmetrically, and how the metal behaves as it is compressed. But because they are using sturdy nuclei, there are no nuclear chain reactions within the compressed metal - and, in the end, nuclear reactions are what the bomb is about. This is where HN comes in. In these experiments, fission is allowed to get under way. Neutrons split nuclei, thus producing more neutrons, which split more nuclei, and so on: the N, after all, is for nuclear. There are various ways of stopping this reaction before it goes too far. Impurities in the fissile material can leave the chain reaction unable to sustain itself; so can a neutron-absorbing gas in the core. The compression can be weakened, or applied unevenly, so that the fissile material blows itself apart before giving out much energy. But despite all this cleverness there is always what the scientists call the 'whoops' factor; a bomb intended to fizzle occasionally misbehaves, and winds up producing a full-fledged nuclear explosion. HN experiments, unlike HD ones, are seldom done in laboratories. Because they provide insights into processes not seen in HD experiments, HN tests have a wider variety of applications. They can be used to test the reliability of weapons, to improve the safety of warheads, and to devise methods of disarming terrorist nuclear bombs. For instance, a safety experiment might involve igniting one point of the high-explosive shell; such a situation might occur because of a fire. Because of the asymmetry of this explosion, it should not trigger a nuclear detonation - but there is only one way to find out for sure. Disarming a nuclear bomb is a similar sort of problem. The line between an HN experiment and a nuclear test is a fuzzy one. Livermore and Los Alamos, America's other weapons-design laboratory, developed the technique so that they could keep up research during a previous moratorium at the end of the 1950s. America now says that any nuclear explosion that gives off less energy than two kilograms (four pounds) of high explosive is a hydronuclear experiment and should be exempted from a future test-ban treaty. France, on the other hand, argues that bannable tests begin at yields of 100 or 200 tonnes, and anything smaller is fair game. This is hardly hydronuclear: it is the size of a small tactical nuclear weapon. Light elements In most modern bombs, fission is just the beginning. The force of a primary fission reaction is used to squeeze deuterium and tritium together in a more powerful secondary fusion reaction. Neither HD nor HN experiments help much here; by the time the second-stage fusion reaction gets started, a bomb is well past the 'whoops' phase and into the 'uh-oh' one. To simulate the second stage without a first stage, physicists use a different sort of squeeze, hammering samples of heavy hydrogen from all sides with beams of light or sub-atomic particles. It is not necessary to go all the way to fusion for this approach to be useful; heating and compressing the hydrogen into an electrically charged plasma provides useful data in itself. Livermore has plans for a new ICF machine called the National Ignition Facility (NIF), which will routinely set the fires of fusion burning. The NIF is to be the most powerful system of lasers in the world; for three billionths of a second at a time, it will focus 500,000 gigawatts, split among 192 laser beams, on to deuterium and tritium targets a centimetre or so across. It is a big enough project, at $ 1.1 billion, to make the difference between keeping Livermore open and closing it down. The French have plans for a similar facility outside Bordeaux as part of PALEN - a weapons-experimentation programme set to gobble up FFr10 billion ($ 2 billion) over the next five years. Hopes and fears To justify such spending, enthusiasts point to a possibility much more alluring than better replacements for nuclear tests: fusion power. A pellet ignited with the aid of a laser can give out approximately 11 times the amount of energy put in. But there is a lot of engineering still to be done before this technique finds commercial expression. Pellets have to be very smoothly and precisely made if they are to implode correctly. The energy, given out in a burst, has to be collected and stored. And the lasers need to operate ceaselessly; at the moment it takes hours for their optics to cool down and their capacitors to charge up. The world already has one set of expensive prototype fusion reactors, which control plasmas with magnetic fields; it is not clear that it needs two. Even with America's NIF and France's PALEN, weapons designers would still not really be able to model the explosion of a nuclear warhead. 'You simply cannot get complete information without (nuclear) testing,' says Patricia Lewis, head of the Verification Technology Information Centre in London. None of the experiments models the interaction between the first and second stages, nor do they give valuable insight into fancier nuclear effects such as 'boosting' fission yields with a small injection of deuterium and tritium. They can let you test ways in which bombs might fail; but they cannot show you the full extent of a bomb's possible success. This suggests that the real motive for further testing is to find designs for new weapons. The French have complained that their current TN70, 71 and 75 warheads are 'sensitive' because they are highly optimised; that is, they are prone to failure because they are so finely tuned. But this fine tuning is the result of their having been developed with the help of extensive testing. And that makes it hard to believe that still more tests are needed - unless there are new warheads in the works, such as small warheads for new M-5 submarine-launched missiles and ASLP cruise missiles. The French deny that they want to design a new warhead, and argue that the eight tests they plan are anyway too few for such a purpose. It has previously taken France 13 or more test explosions to design a weapon. But Dr Brown says that, thanks in part to better simulation, America can design a weapon with as few as four tests. Eight might well be enough for France. It only took one for the Manhattan project. © 1995 The Economist Newspaper Limited. All rights reserved