Donald MacKenzie on nuclear weapons
It always helps to see the ordinariness of things. Despite the end of the Cold War, nuclear weapons remain very un-ordinary in the popular mind. The world’s nuclear arsenals still contain over twenty thousand warheads. Yet nuclear weapons are an ordinary technology and can, like other technologies, become obsolete. They can, perhaps, be abolished. There is even a meaningful sense in which nuclear weapons can be disinvented.
Surprisingly, the abolition of nuclear weapons is discussed less often now than it was during the Cold War, when the odds were stacked hopelessly against it. Public alarm has diminished, and as a result, nuclear disarmament movements have lost impetus. The familiar arguments in support of abolition remain: among them, the risk of nuclear war breaking out by accident, miscalculation and irrationality. An important new factor is that half the world’s nuclear weapons are in the hands of a power which is likely to be unstable for many years. To date, those in charge of the former Soviet nuclear forces have shown restraint and responsibility: the serious cases of smuggling of nuclear materials have largely resulted from ‘stings’ by Western, especially German, intelligence agencies. However, restraint and responsibility cannot be guaranteed when soldiers, unable to feed their families, sell firearms from the backs of trucks. Against abolition is the fear that, without nuclear weapons, conventional war will be more likely, and that ‘pariah’ states might turn to chemical or biological weapons. Above all, there is the possibility that some states might cheat on a nuclear disarmament agreement by hiding weapons away or by re-arming in the future.
The most common way technologies disappear is by obsolescence: the task they performed is no longer needed, or a new, better, way of doing it emerges. To a striking extent, nuclear weapons are already obsolete. Our dominant images of them are framed by Hiroshima and Nagasaki; we see them as terror weapons used to destroy entire cities. From the Fifties to the Eighties, however, this was not the purpose of most nuclear weapons, especially American ones, which were designed to be used against a massive invasion of Western Europe by Soviet tanks; to destroy Soviet missiles in their silos before they could be launched; and in a variety of other roles that now seem bizarre – as landmines, anti-aircraft missiles and depth-charges. The variety of purposes for which they were intended explains the enormous number accumulated in the American and Soviet arsenals, even though the most hawkish of nuclear strategists accepted that a few hundred were sufficient to devastate an enemy’s cities.
Two things have changed. First, the United States and its allies now enjoy overwhelming superiority in conventional armaments over any likely foe. Economic and organisational collapse has eroded Russia’s armed services and the military-industrial complex that used to stand behind them. Ten years ago, the West feared that the Red Army could reach the Atlantic unless stopped by nuclear weapons; now, it can’t subdue a rebellious province. Nor does any other country approach parity with the Western Allies in conventional weaponry. Saddam Hussein’s Iraq was one of the best armed countries outside Nato. The Gulf War, in which there were around a thousand Iraqi casualties for each Allied one, showed how great the imbalance is. Second, the progress of military technology, especially in precision weaponry, means that many of the targets against which nuclear weapons used to be directed can now be destroyed by non-nuclear warheads. True, the cruise missiles used against Iraq still have many shortcomings. For example, the scene-matching systems currently employed to correct their trajectories look downwards rather than forwards, so a missile has to rake its last positional fix several miles from its target; also feeding in data on new targets is still a protracted process. Even so, the missile that can be guided to enter a particular window of a particular building will shortly be a reality. With that accuracy, few military targets require the explosive force of a nuclear warhead.
Another way technologies lose their potency is by disconnection from the infrastructure that supports them. A washing machine is no use if the electricity fails, or if a key part of the machine breaks and cannot be mended or replaced. The Third World is littered with real, rusting examples of disconnection: with the remains of technologies that work well enough under First World conditions, but fail in the absence of supporting networks. Even a simple nuclear weapon can contain up to several thousand different parts. Many of these parts need periodic inspection, maintenance and replacement: corrosion, radioactive decay or chemical decomposition can render a weapon useless, even make it dangerous to those who possess it. After the collapse of the Soviet Union, hydrogen gas began to build up in some of the nuclear weapons left behind, unmaintained, in the Ukraine. Fortunately, senior figures in the Ukrainian Armed Forces recognised the risk of an explosion which, though not nuclear in itself, could have scattered radioactive materials over the surrounding area – and resolved the problem with the help of experts from the Russian nuclear weapons laboratories.
A realistic worldwide agreement to abolish nuclear weapons would involve their phased destruction over a twenty-year period: that much time would be needed to allow trust to build up, verification systems to be put in place, and existing arsenals to be safely disposed of. A country cheating on such an agreement would have to hide its weapons away during the whole of that time, but after twenty years a nuclear weapon left unattended would not work properly. What makes a hidden arsenal most vulnerable is the dependence of all sophisticated nuclear weapons on tritium, a rare radioactive gas. In most designs of atomic bomb, whether simple or sophisticated, a ‘core’ made of plutonium or enriched uranium is imploded by chemical explosives (see diagram). In sophisticated weapons, tritium and deuterium (which is, like tritium, an isotope of hydrogen) are injected from a pressure vessel into the heart of the core as nuclear fission begins. The tritium and deuterium fuse, producing large numbers of fast neutrons, and the extra neutrons increase the proportion of the core that undergoes fission. ‘Boosting’, as this is called, enables nuclear missile warheads to be far smaller than the four-ton monsters that had to be carried to Hiroshima and Nagasaki by lumbering, heavy bombers.
Unlike uranium and plutonium, the tritium needed for boosting wastes away relatively quickly. Within just over twelve years, half of any given quantity of tritium will no longer exist. The vanishing of tritium progressively weakens boosting, and the gas it turns into, helium-3, absorbs neutrons. Eventually, boosting may fail so badly that the most destructive part of a hydrogen bomb – its ‘secondary’ – fails to ignite. Published figures suggest that if this were to happen in a modern, miniaturised warhead, its yield, or explosive force, could be reduced to less than 1 per cent of what it should be. Instead of the intended explosion, equivalent to hundreds of thousands of tons of chemical explosive, the yield could be as little as 500 tons – a fortieth of the explosive power of the Nagasaki bomb.
Tritium exists in nature only in unusably minuscule quantities. It is created in nuclear reactors, and then separated out by a complex process. Tritium dependence is of particular concern to the United States, which can no longer make the substance because safety problems have closed all the suitable military reactors. An existing civil nuclear reactor could be modified to produce tritium, but this would violate the spirit of the Nuclear Non-Proliferation Treaty, which attempts to maintain a strict separation between the civil and military uses of nuclear energy. Another possibility would be to construct a 4000-feet-long underground particle accelerator, powered by an amount of electricity equivalent to the consumption of a medium-sized city, and use it to make tritium by bombarding lithium or helium targets. The other nuclear weapons states still make tritium in military reactors (although those in which Britain’s tritium is produced are also ageing). With reactors placed under international supervision, it would be possible to detect whether tritium was being produced. ‘Tritium disconnection’ – a cut-off in tritium supplies – would mean that even if a small number of sophisticated nuclear weapons could be hidden away (a form of cheating which vigorous verification could, in any case, make very difficult), such weapons would be, literally, wasting assets. True, the timescales involved are long. The US, for example, will be able to get by for another fifteen to twenty years by recycling and purifying the tritium contained in warheads decommissioned by arms-control measures: but recycling of this kind can also be controlled, as can leakage of tritium from the small civil market that deals in the substance.
Crude nuclear weapons do not require tritium, but they too are vulnerable to disconnection from their supporting infrastructure. At their heart is a small device known as an initiator, which produces a sudden flux of neutrons to kick-start the fission chain reaction. Early weapons programmes typically used polonium as their initiator, which decays far faster than tritium: the half-life of polonium is a little over three months. This means that polonium cannot usefully be stored, and a nuclear weapons programme reliant on it needs regular access to reactors that produce it. The more common nuclear substances – plutonium and enriched uranium – present more of a problem. Nuclear bookkeeping has been lax: 2.8 tons of plutonium which ought to be in the US nuclear arsenal cannot be found. No one seriously thinks it has been stolen, but the inability to account for quantities of that magnitude (and, in the case of Russia, almost certainly much larger amounts) means that it will be hard to be sure that no fissile materials have been hidden by an existing nuclear state. Controls over plutonium and enriched uranium could, however, prevent the emergence of new nuclear states.
The defenders of nuclear weapons will point out that ‘you can’t disinvent the Bomb.’ Admittedly, the physics of nuclear fission and fusion is written down in the textbooks forever, but there is much more to creating a functioning technology than explicit knowledge of this kind. Also important is ‘tacit knowledge’, the analogue of the knowledge we deploy in being able to ride a bicycle. Those in charge of nuclear weapons programmes, typically physicists, have underestimated the need for tacit knowledge and craft skill. The original plans for the first nuclear weapons laboratory, at Los Alamos, envisaged it as no bigger than the physics department of a large university, but as the engineering problems of building real, working bombs became apparent, the staff grew to be several thousand strong.
Over the years since 1945, much of this originally tacit knowledge has been systemised, codified, and embodied in machines performing tasks that previously needed skilled workers. However, Iraq’s nuclear weapons programme is evidence of the continuing importance of tacit skill. This was no amateur effort: it had existed for around a decade; no expense was spared; all available literature was searched; up to seven thousand scientists and twenty thousand workers were employed; and Western companies were willing to supply the necessary hardware. Yet the Iraqis made only slow progress: the necessary skills – in the use of imported equipment, for example – could not be acquired quickly enough. If, in 1990, Saddam had thought his scientists were about to give him a nuclear arsenal, he would surely have postponed the invasion of Kuwait until they had done so. What this shows is that, even with slack international controls, there are barriers of skill to be overcome in acquiring the materials for, and in designing and building, a simple, unboosted atomic bomb.
These problems intensify as one moves to the more sophisticated, miniaturised weapons needed to arm intercontinental ballistic missiles or long-range cruise missiles. Much of the bulk of an atomic weapon is the chemical high explosive. This has to be formed into a structure, loosely analogous to the lens of a magnifying glass, that focuses the blast of the high explosive into an inward-moving spherical wave. More realistic designs than my diagram can be found in published literature, but actually building them is skilled, critical work, involving the shaping of two different types of high explosive into a precise structure (an imperfect explosive lens leads to a blast wave that is not spherical, and may simply blow the core apart rather than compress it). Early British nuclear weapons scientists, for instance, had a good understanding of the theory of explosive lenses: it was a member of the British delegation to Los Alamos, the experimental physicist James Tuck, who first suggested how a three-dimensional lens could be used to create an adequately spherical blast wave. The early British bomb team couldn’t, however, find a way of stopping the components of the lens from shrinking unevenly in their casts as they were being made.
The official history of the British nuclear programme reveals that the first British atomic bomb was, literally, held together with sticky tape, which was used to fill in the gaps in the lens structure and to minimise settlement. Such homely remedies can work with a big, simple bomb, but not with a small, sophisticated one. Designers have to have a good grasp of what they are doing – for example, a detailed understanding of the process of boosting cannot be derived simply from theoretical physics – and fabrication must be exact. It is possible to have reasonable confidence in a simple atomic bomb without actually testing it in a nuclear explosion, but testing becomes more necessary as sophistication grows, and nuclear testing is incompatible with keeping a nuclear programme secret.
In the more advanced nuclear states, much of the designing and analysis of nuclear warheads has for some time been done by computer, but even the most detailed computer simulation of a warhead is an approximation to reality, not a mirror of it. A large part of the skill of a nuclear weapons designer derives from an intuitive ‘feel’ for these computer simulations: for when they can be relied on, and when they cannot. As the design of a weapon becomes more sophisticated, so the ‘feel’ becomes more important. The margin for error, quite large in the case of a simple weapon, tends to shrink as the weapon becomes more complex.
Attempts are being made to capture and to store the tacit knowledge of the current generation of nuclear weapons designers (who have experienced the whole design process through to nuclear testing) before they retire and die: the US Department of Energy, which is responsible for nuclear weapons laboratories, has a large Knowledge Preservation Project. But such efforts are likely to be only partly successful: after sufficient time has passed, even re-creating weapons on the basis of surviving blueprints becomes increasingly problematic. The characteristics of some components – high explosives, most notably – are remarkably sensitive to the particularities of how they are made, and production processes often change with time in ways that are hard to document.
Explicit, written-down, knowledge is effectively immortal. Tacit knowledge, in contrast, dies with its possessor, unless it is passed on face to face and hand to hand. If the design and production of nuclear weapons were to cease, let us say for a couple of generations, then re-creating them – especially re-creating the sophisticated, miniaturised warheads needed for relatively easy delivery – would be significantly harder than conventional wisdom allows.