End of Nuclear Testing
Jeremiah D. Sullivan
The technical basis for the U.S. adoption of a "Comprehensive Test Ban" national policy was a Department of Energy sponsored study conducted by the JASON group [1] during the summer of 1995. My remarks here draw from my participation in that study together with the knowledge and experience I have gained from two decades of work as an academic and as a consultant to the US Department of Defense, Department of Energy, and Arms Control and Disarmament Agency on arms control and defense technologies. The study itself is classified, but the Summary and Conclusions are publicly available [2], and I speak primarily about them.
The primary task of the JASON study was to determine the technical utility and importance of information the United States could gain from continued underground nuclear testing...
The study panel was unique in having four senior weapons scientist-engineers as members together with ten physicists, primarily from the academic community, all of whom had considerable experience in the issues addressed by the panel. The panel had full and unrestricted access to information held by the US nuclear weapons design laboratories: Los Alamos National Laboratory (LANL), Lawrence Livermore National Laboratory (LLNL) and Sandia National Laboratories (SNL), and it received full cooperation from these laboratories.
US Policy and Plans Concerning Nuclear Weapons
The key assumptions of the study were given by US policies and plans regarding its nuclear forces operative during the summer of 1995,which remain unchanged today.
- The US intends to maintain a credible nuclear deterrent.
- The US is committed to world-wide nuclear non-proliferation efforts.
- The US will not develop any new nuclear weapons designs. (A policy first announced by President Bush in 1992 and reaffirmed by President Clinton.)
In practice, Policy 1 means that after START II reductions (scheduled for the year 2003), the US will have approximately 3,500 nuclear warheads and associated launchers in its active strategic arsenal, a smaller number of strategic warheads in the inactive reserve, and a sizable number of tactical warheads in reserve as well [3]. The warheads in the active reserve are referred to as the "enduring stockpile." The US is currently dismantling about 1,500 nuclear warheads per year and is manufacturing no new ones.
The enduring stockpile consists of nine distinct designs of various ages; six are LANL designs and three are from LLNL. All are variations of a general design type referred to as "hollow-boosted primaries with fission enhanced secondaries. Policy 1 also means that the US will require guarantees of the surety, safety, and performance (reliability) of weapons in the enduring stockpile.
Surety means that rigorous measures are in place to ensure that no nuclear weapon is used without authorization or falls into the hands of an unauthorized individual. To this end, physical mechanisms are built into warheads to prevent any nuclear yield should an unauthorized party acquire a weapon and permissive action links and related measures are included in the overall weapon systems to prevent use by authorized holders without approval from the National Command Authority - the President under normal circumstances. None of the technical or operational procedures associated with surety require nuclear testing, so it has never been an issue in debates over nuclear test ban policy.
Safety refers to choosing weapon designs and handling procedures are chosen to ensure to the highest possible levels that in the normal storage, transport, and basing of nuclear weapons, no accident, fire, collision, or other mishap will result in any nuclear yield or dispersal of fissile material (weapons grade plutonium or highly enriched uranium).
Reliability means the yield of the weapon will fall within specified limits even in the worst-case environment: just prior to tritium boost gas supply replenishment, operation in the high neutron environment of a nuclear war, or implosion of the primary at sub-freezing initial temperature.
Questions Facing the US Nuclear Weapons Community in a CTBT Era
- What technical capabilities will be required to maintain confidence in the enduring stockpile for an indefinite period?
- What should the response be to aging effects uncovered during inspections of existing weapons?
- How will human expertise in the science and technology of nuclear weapons be maintained?
- What types of zero-nuclear-yield experiments are important in maintaining confidence in the enduring stockpile? (These are often referred to as above-the-ground experiments, or AGEX.)
and most importantly,
- What contributions would testing at low levels of nuclear yield make to maintaining confidence in the enduring stockpile? Such testing might be "permitted" for a finite period of time as a transition into a true CTBT era, or perhaps be allowed indefinitely, in effect defined as "not counting as a nuclear test. "
These technical questions do not exist in a vacuum, especially Question 5. Arms control policy, non-proliferation policy, international relations, and a host of other factors are impacted by the answers. My remarks below focus entirely on the fifth question. DOE initiatives to assure answers to the other four questions are contained in its Science Based Stockpile Stewardship Program (see the following paper in this issue).
Physics of Modern Nuclear Weapons
Modern nuclear weapons consist of a primary and a secondary stage [4, 5]. The primary consists of a hollow shell of fissile material (the "pit"), surrounded by an array of high explosive (HE) charges, which when detonated by an appropriate signal causes a spherically symmetric implosion that compresses the pit to a supercritical configuration. At the optimal moment, a burst of neutrons is released into the imploded pit, triggering a flood of fission chain reactions.
In the primaries of all modern nuclear weapons, a boost gas mixture consisting of deuterium (D) and tritium (T) is introduced into the pit just prior to HE initiation. As the fission energy released from the supercritical assembly builds, the DT gas in the pit is heated above the threshold for thermonuclear processes, the most important of which is D + T -> alpha + n. The sudden, intense flood of neutrons created in these fusion reactions induces vastly more fission chains, thereby greatly enhancing the fraction of the fissile material in the pit that undergoes fission. The direct contribution of fusion to the net primary yield is minor; the indirect contribution is very large.
The energy released from the primary couples to the secondary by radiative transport, creating the conditions for thermonuclear burn. The basic ingredient in a secondary is solid lithium deuteride (6Li-D), which contains the deuterium needed for the D-T fusion process and generates in a timely manner the needed tritium through the "catalytic" process n+ 6Li -> alpha + T. Depleted or enriched uranium can be added to the secondary to enhance substantially the secondary yield given the intense flux of energetic 14 MeV neutrons that result from D-T processes there. Modern secondary designs exploit this synergy between fusion and fission, and for this reason a substantial fraction of the overall secondary yield comes from fission processes. This is the basis for the popular rule of thumb for estimating fallout: 50% of the yield from fusion, 50% from fission [6].
Primaries and secondaries present quite different challenges to designers. Primaries are notoriously sensitive to small changes since the basic implosion processes, being hydrodynamic in nature, is accompanied by non-linear effects and a penchant for turbulent behavior. In addition, one has low density material driving high density material, which raises the specter of Rayleigh-Taylor and other instabilities. Finally, the primary depends critically on the uniform performance of high explosive, a notoriously idiosyncratic chemical that is sensitive to environment conditions, grain size, and other details of manufacture.
In contrast, the physics of secondaries while complex is generally forgiving because of the speed-of-light transport of radiation vs. the relatively slow supersonic speed of material transport. Computer modeling of secondaries, however, presents enormous challenges because of the need for fine time-steps to capture the rates of change of radiation processes and while follow hydrodynamic motion operating on much longer time scales.
Why the Study Was Commissioned
Following the successful indefinite extension of the Nuclear Non-Proliferation Proliferation Treaty in May 1995, a major debate developed within the US security and arms control communities. This debate had three main voices: (i) some argued that the United States needed to maintain the right to do nuclear testing at approximately the half-kiloton level for a fixed period, say ten years, to retain confidence in the safety and reliability of its nuclear weapons; (ii) others argued that retaining a right to do hydronuclear testing (defined roughly as less than four pounds of nuclear yield) indefinitely would be required to retain confidence in the enduring stockpile as it aged; (iii) yet others argued that neither sub-kiloton nor hydronuclear testing was necessary and neither contributed usefully to maintaining confidence in the enduring stockpile. (Hydro tests which study implosive assemblies that never achieve a supercritical configuration were not at issue.)
Conclusions of the Study
The primary task of the JASON study was to determine the technical utility and importance of information the United States could gain from continued underground nuclear testing at various levels of nuclear yield for weapons in its enduring stockpile as they age[2].
Conclusion 1-The first conclusion of the study panel was that the United States could have high confidence in the safety, reliability and performance margins of weapons in the enduring stockpile. This confidence is based on 50 years of experience and analysis of more that 1,000 nuclear tests, including approximately 150 nuclear tests of modern weapon types in the past 20 years.
Conclusion 2-The report recommends a number of activities that are important for retaining confidence in the safety and reliability of the weapons in the enduring stockpile whether or not testing at the half-kiloton level or less is permitted. The recommended non-nuclear tests, most being extensions of above-ground experiments that have long been part of laboratory operations, would be designed to detect, anticipate, and evaluate potential aging problems and to plan for refurbishment and remanufacture as part of the DOE Science Based Stockpile Stewardship Program.
Conclusion 3-Weapons in the enduring stockpile have a range of performance margins, all of which were judged by the panel to be adequate. The performance margin of a nuclear weapon is defined as the difference between the minimum expected primary yield [worst case] and the minimum yield required to ignite the secondary. Simple measures requiring little effort, such as increasing the tritium load in boosting or more frequent change in boost gas reservoirs to compensate for tritium decay, could substantially increase performance margins at little effort to hedge against unforeseen effects.
Conclusion 4-The primary argument for retaining the right to do testing at or around the half-kiloton level for a finite period was that it would give the US valuable information about the effects of aging of weapons designs in the enduring stockpile. Extrapolation using computer codes from yields obtained in tests of primaries that demonstrate the initiation of boosting (half-kiloton, roughly) to an expected full-boosted primary yield is quite robust, provided the codes are well calibrated.
After careful examination the study panel concluded that a finite period of half-kiloton testing would give little or no useful information about aging effects beyond what can be obtained from inspections and other AGEX in a well conceived stewardship program. Namely, nuclear tests at the half-kiloton level for a finite term would help develop codes further and improve theoretical understanding of the boosting process but would not contribute to understanding aging effects. Similarly, such testing would not provide useful checks of refurbished or remanufactured primaries made after the time the testing ceased. In view of its limited utility, the study panel found that half-kiloton testing for a limited time had a very low priority in comparison to the activities recommended in Conclusions 2 and 3.
To be useful, half-kiloton testing would need to be continued indefinitely. This, the study panel observed, would be tantamount into converting the CTBT into a threshold test ban treaty.
Conclusion 5-The study panel concluded that testing of nuclear weapons at any yield below that required to initiate boosting is of limited value to the United States whether done for a finite period or indefinitely. This yield range includes 100 ton testing and hydronuclear testing. The hydronuclear case merited special discussion.
The arguments for continued hydronuclear testing were that such tests would provide a valuable tool for monitoring the expected performance of weapons in the enduring stockpile as they age. The basic idea would be to do a set of baseline hydronuclear tests of (modified) primaries of the types now in the stockpile and to follow up in the future with regular hydronuclear testing of primaries drawn from the stockpile and similarly modified. The test results would be compared to the baseline data.
The panel concluded that a persuasive case could not be made for the utility of hydronuclear testing to detect small changes in the performance of primaries. The fundamental problem with hydronuclear testing is that primaries need to be modified drastically to reduce the nuclear yield to a few pounds of TNT equivalent. This can be done either by removing an appropriate amount of the fissile material from the pit and replacing it by non-fissile material, e.g., depleted uranium, or by leaving the pit intact and replacing the boost gas in the pit by material that will halt the implosion just after the point of supercriticality is achieved. Neither method exercises a pit through its normal sequence, and so one does not obtain data relevant to the real implosion. Extrapolation of hydronuclear test results to actual weapons performance would be of doubtful reliability and low utility.
Hydronuclear testing can be useful for checking one-point safety of a primary design as was done in the past by the U. S. [7] Today, however, all US weapons designs in the enduring stockpile are known and certified to be one-point-safe to a high degree of confidence, so hydronuclear testing is not needed for safety tests. Furthermore, the development of 3-D codes further reduces any need to do actual tests to evaluate issues of one-point safety. Modeling capabilities will further improve as computer systems and codes advance.
Conclusion 6-A repeated concern raised by the nuclear weapons community about entering into a CTBT of unlimited duration is the ultimate unpredictability of the future. No one can predict with certainty the behavior of any complex system as it ages. The study panel found that should the United States encounter problems in an existing stockpile design that lead to unacceptable loss of confidence in the safety or reliability of a weapon type, it is possible that testing of the primary at full yield, and ignition of the secondary, would be required to certify a specified fix. Useful tests to address such problems generate nuclear yields in excess of 10 kt. A "supreme national interest" withdrawal clause-standard in all arms control treaties-would permit the United States to respond appropriately should such a need arise. It is highly unlikely that even major problems with one or two designs would cause the United States to withdraw from the CTBT, given the political implications involved. Major problems with all or almost all of the designs in the enduring stockpile would most likely be required.
Conclusion 7- The study panel observed that its Conclusions 1-6 were consistent with US agreement to enter into a true zero-yield Comprehensive Test Ban Treaty of unending duration, which includes a supreme national interest clause.
Jeremiah D. Sullivan is with the Department of Physics and Program in Arms Control, Disarmament and International Security, University of Illinois at Urbana-Champaign jdsacdis@uxl.cso.uiuc.edu
References
Study panel members were: Sidney Drell (Chair), John Cornwall, Freeman Dyson, Douglas Eardley, Richard Garwin, David Hammer, Jack Kammerdiener, Robert LeLevier, Robert Peurifoy, John Richter, Marshall Rosenbluth, Seymore Sack, Jeremiah Sullivan, and Frederik Zachariasen.
- Congressional Record-Senate, Vol. 141, No. 129, pp. S11368-S11369, August 4, 1995. -Arms Control Today, "JASON Nuclear Testing Study," pp. 36-37, September, 1995. (At this time no other portion of the report is unclassified.)
- R. S. Norris and W. M. Arkin, Bulletin of the Atomic Scientists, "Natural Defense Research Council, Nuclear Notebook," July/August, pp. 76-79, 1995.
- P. P. Craig and J. A. Jungerman, Nuclear Arms Race: Technology and Society, pp. 185-190, McGraw Hill, 1986.
- D. Schroeer, Science, Technology, and The Arms Race, pp. 62-65, John Wiley and Sons, 1984
- S. Glasstone and P. J. Dolan, The Effects of Nuclear Weapons: 3rd Edition, United States Government Printing Office, 1977.
- R. N. Thorne and R. R. Westervelt, "Hydronuclear Experiments," LA-10902-MS UC-2, Los Alamos National Laboratory, 1987.