Nuclear Stockpile Stewardship

John C. Hopkins

physicist learns at his mother's knee that if he cannot work a problem to change it ever so slightly to one that he can work. This is never more applicable than in the design of nuclear weapons. No one knows all of the ins and outs of the science of modern weapons. The first two devices of World War II fame, the gun-type weapon and the simple implosion weapon, are explained quite well, but the scientific understanding of our very sophisticated, and efficient, modern weapons has so far defied our best technical efforts. Nuclear weapon design has been largely a cut and try affair, being guided by mathematical models that are adjusted to fit the experimental data observed in what have been called nuclear tests. Because of this crucial dependence upon full-scale nuclear testing it is not at all clear that the stockpile can be maintained indefinitely without testing. This is one place where corporate memory and years of experience are invaluable. Judgment is absolutely necessary, as necessary for stockpile stewardship as for the design of new weapons.

The label "test" is an unfortunate choice, because it conveys the image of detonating a nuclear device to see whether it will go off. In fact the so-called tests are really very complex physics, chemistry, and engineering experiments formulated to measure quantities that can be related to terms in theoretical models. Before a nuclear test of a new design many of these terms are either unknown, or not known accurately enough to permit the designers complete confidence in the calculations. Weapons theoretical physicists are working toward a generalized model that can explain a wide spectrum of nuclear designs. A generalized model is the Holy Grail of nuclear weapons theory. So far that is not at hand. Each family of weapons has its own parametric values that "work" for that design and for modest extrapolations from the base-line design. Experience has historically been the best teacher in exploring how far one could extrapolate any particular design with confidence.

The present stockpile was designed during a period when nuclear testing was allowed and when it could reasonably be expected that the weapons, and weapons systems, would have at most a twenty to thirty year lifetime. Now we have a new set of ground rules and a quite different political environment. Today we have very sophisticated weapons, where sophistication is not an asset. We also have numerous high yield weapons on MIRVed missiles where one or a very few low yield weapons per delivery platform would provide a more credible deterrent.

The present weapons are as light and as small as possible so that many would fit on a MIRVed missile. They use minimal amounts of nuclear material, so that as many as possible could be built. These weapon characteristics do not fit a prescription of long life weapons in a non-testing regime. Being designed with a minimal performance margin they are less forgiving of slight variations, or changes during the life of the weapon. The stockpile stewardship program is aimed at preserving these more or less inappropriate weapons indefinitely as safe, secure, and reliable as they are today. Can this be done? No one knows. Nothing even remotely similar has been attempted.

The biggest uncertainty is with the primary. Modern strategic weapons have two major parts: the primary, which is a highly developed atomic bomb, and the secondary, which is the main thermonuclear part that must be driven by the tremendous energy of a nuclear explosion to set it off. Modern nuclear weapons, i.e. those designed within the past thirty years, are built to extremely close tolerances and everything must be just right for them to work. Even without nuclear testing we can deal with the question of whether our weapons might go off unexpectedly: The more difficult problem is whether we can be assured that they will give the appropriate yield (i.e. energy release) when needed.

Primaries are usually shells, called "pits", of plutonium, or uranium, or both, surrounded by high explosive and very precise detonation hardware. The shells are usually filled with a deuterium/tritium gas mixture that provides a small thermonuclear reaction that "boosts" the primary to greatly increase the yield. The design community does not yet have an adequate understanding the details of the pit configuration, the pit surface, or the boost cavity characteristics as the device goes critical, boosts, and explodes. Much work is being devoted to the understanding of these phenomena.

The primary and secondary are in a case, or container, that confines the energy output of the primary until the secondary is compressed and develops a nuclear and thermonuclear chain reaction. All of our weapons were designed during the Cold War and have as light a case as was deemed adequate at the time. The weapons were optimized for a strategic situation that has passed.

One concern with the primary is that tritium decays with about a 12-year half-life, and therefore the boost reaction varies over the life of the weapon. The question for the designers is how long will the primary last before the yield is too low to drive the secondary? Before 1992 questions like this were addressed by testing bombs of various ages.

An additional concern, when contemplating keeping weapons indefinitely, is the alpha decay of plutonium, resulting, eventually, in microscopic inclusions of helium within the plutonium. How long might the plutonium parts be expected to last before the pits would have to be refabricated? How much confidence should one have in the answers? How does one even quantify the confidence levels? In the future, with weapons of indefinite age and no nuclear testing, the designers must rely even more on calculations and judgment.

Stockpile stewardship includes a complex suite of theoretical and experimental laboratory programs to try to understand how nuclear devices work from a very fundamental basis. When a nuclear weapon is to be used in a military setting the sequence is approximately as follows: After use-authorization is verified, the arming system checks certain variables to be sure that the weapon is enroute to its designated target, the altitude sensors or fuses ascertain that the proper detonation point has been reached, detonators are fired, the high explosive compresses the nuclear material, the compressed material goes supercritical, the deuterium and tritium boost the primary yield, the energy of the primary "flows" to the secondary, the secondary is compressed and ultimately gives a very high yield. Only the first four steps can really be "tested" without resorting to a full-scale nuclear detonation. Nothing after the device goes critical can be done in the lab. It is relatively straightforward to check the fusing and firing systems, but it quickly gets very difficult beyond that.

Elaborate experimental facilities are being developed to shed light on the nature of plutonium and uranium components as they are compressed with high explosives. One such facility incorporates two modern flash x-ray machines of enormous power and size to probe the internal features of a high explosive implosion. Clearly the actual nuclear materials cannot be used or the experiments would result in a nuclear explosion. Other techniques must be developed to cope with the experimental challenges posed by the investigations. Further experiments in the laboratory or at the Nevada Test Site address the plutonium equation of state, or the physical characteristics of plutonium under the extremes of pressure that are encountered in a nuclear device. Additional experiments, and experimental facilities, address a whole raft of design details that are affected by the high explosive.

To carry the research into the realm of criticality, boosting, energy flow, or secondary behavior requires very sophisticated computational work employing the world's largest computers. It is imperative to convince oneself that the calculations make sense and really represent the nuclear phenomena correctly. Calculations of three-dimensional hydrodynamic behavior are horrendously complicated and even with enormous computational capability can take weeks or months to accomplish. It is anticipated that some scaled work relevant to stockpile stewardship will eventually be done on radiation flow and related phenomena with large lasers and pulsed power machines.

There is no doubt that the computational simulations will be extremely valuable, but whether in the long run they really provide the confidence that we are doing things correctly will, for the time being, remain unknown.

This issue brings me to my final point: How do we know whether stockpile stewardship is working? I don't think that anyone knows the whole answer to this today, but we can make some observations. When we are able to calculate correctly the more complicated phenomena that will be produced in the laboratory, and when all the results from the previous full-scale Nevada and Pacific tests can be explained, then the Nation should feel some confidence that the stockpile stewardship role is indeed being properly performed. It should be recognized, however, that it is possible that we will not achieve such a level of understanding, or mathematical modeling capability, and may be forced to declare that we no longer have confidence that our weapons will work when they are supposed to and not work when they aren't.

October 26, 2000

Dr. Hopkins is a Sumner Associate. He is a retired nuclear physicist and a 34-year veteran of the Los Alamos National Laboratory. He was the leader of the nuclear testing program, the leader of a nuclear policy think tank, and eventually the leader of the entire nuclear weapons program at Los Alamos.

  Return to Sumner Associates, Inc. Home Page