The Weapon Labs and the Test Ban Treaty
The final version of this article appeared under the title "Sowing Nuclear Misconception" in Nature, vol. 404, pp. 131-133, January 13, 2000.
by Kurt Gottfried
The US Senate's vote against ratification of the Comprehensive Test Ban Treaty (CTBT) last October casts a dark shadow over the treaty's prospects, and poses a serious threat to the non-proliferation regime. Although bitter partisan politics and the Clinton administration's inept handling of the case for the treaty were dominant factors in the Senate's decision, the views of the directors of the three US nuclear weapon laboratories contributed substantially to this outcome. Indeed, in their testimony of 7 October 1999, the directors offered a critique of the treaty that could well have led an uninformed listener to wonder why presidents going back to Dwight Eisenhower and John Kennedy had sought such a treaty, and why the directors had agreed to President Bill Clinton's signing of it in 1996. As prominent voices on both sides of the controversy appear to have accepted the directors' assessment, an examination of their testimony is in order.
The directors all implied that to ratify now would amount to betting on unproven technologies. In the words of John Browne of the Los Alamos National Laboratory, "We are using a fundamentally different set of tools to ensure the safety, reliability and performance of nuclear weapons, whereas the essential toolkit for stockpile stewardship [without nuclear testing] will not be completed until the middle of the next decade." However, the labs' own joint report [1] on stockpile maintenance during the nuclear testing era documents that the technology and methodology of that time were very similar to today's and that underground tests played only a minor role in establishing confidence in deployed weapons.
Paul Robinson of Sandia National Laboratory went much further by recommending rejection in a widely quoted statement: "If the United States scrupulously restricts itself to zero yield while other nations may conduct experiments up to the threshold of international detectability, we will be at an intolerable disadvantage. I would advise against accepting limitations that permit such asymmetry." This is a politico-military judgement in a technical guise, and is contradicted by testimony of the preceding day from Gen. John Shalikashvili, chairman of the Joint Chiefs of Staff at the time of President Clinton's decision to sign the treaty. The general reported that after consultation with the lab directors, among many others, the Chiefs had concluded that on balance the CTBT was in the US interest. Furthermore, the Chief US Negotiator testified that without the zero-yield limit no ban was achievable, because the non-nuclear states that the US sought to constrain "insisted that the five [nuclear powers] should be allowed no tolerance -- not even for the smallest possible nuclear yield."[2]
The question, of course, is what price the United States is willing to pay to constrain would-be nuclear states. I will argue that in view of the great strategic advantages that the United States enjoys over all potential proliferators, and indeed all other states, the benefits of the CTBT far outweigh the risk that abstaining from testing will generate a credible loss of confidence in the US nuclear arsenal in any quarter.
Nuclear Weapons and Nuclear Testing
The prospects for fielding a nuclear weapon without testing depend on the sophistication of the design, and the proliferator's knowledge, resources and strategic goals. Of course, the acceptance of risk is a subjective matter. Nevertheless, two such different societies as the US and USSR devoted comparable and enormous efforts to attaining a high level of confidence in their nuclear weapons. Here confidence has two aspects:
- safety -- assurance that a violent impact that detonates the chemical explosives, whether by accident or in warfare, will not cause a nuclear explosion or a dispersal of hazardous material
- reliability -- knowledge that the weapon will perform to design specifications within some specified uncertainty.
It is sometimes conjectured that clandestine proliferators will not have such scruples, but given their limited military capabilities they would have ample reason to insure that their small arsenal is not a graver threat to themselves than to their enemies.
The simplest fission weapon is of the Hiroshima type, in which a supercritical mass of isotopically purified uranium was suddenly assembled by a gun-like device. The Hiroshima bomb had not been tested, and presumably a proliferator could repeat this feat, as did South Africa. But such devices have low yields relative to their weight and cannot use plutonium.
All modern designs are descendents of the Nagasaki bomb, in which a sphere of fissile material is compressed by a surrounding shell of chemical explosives to form a supercritical mass. This device can use plutonium -- either weapon-grade or from the spent fuel of a power reactor. It is a trickier design, however, and to bank on such weapons without nuclear testing would entail considerable risk of very low yield or outright failure.
The next level of sophistication is the "boosted" fission device, in which a mixture of hydrogen isotopes, placed at the very center of the imploding fissile material, undergoes a fusion reaction which produces energetic neutrons that greatly amplify the fission efficiency. The result is a far lighter weapon with yields that can reach hundreds of kilotonnes, but without testing it would be burdened by still greater doubts about reliability and performance.
The five major nuclear weapon states use much more complex two-stage designs. Here the energy released by a boosted fission device, called the "primary," ignites a spatially separated thermonuclear "secondary," giving a yield that can range up to many megatons. This whole "physics package" involves more demanding computations and more sophisticated physics. Those building such a weapon for the first time, even with stolen information, would be taking great risks if they did not test.
Stockpile Testing in the Past
In assessing the confidence the United States can have in its stockpile without nuclear testing, one should recognize that the great majority of nuclear tests were always devoted to developing new weapons, rather than to reconfirming the reliability and safety of existing ones. Indeed, there is a logic to the paucity of tests following deployment:
- the human and technical delivery systems are much more likely to fail than the warhead itself
- everything up to ignition of the fission reaction, including detonation of the conventional explosive and implosion, can be tested without a nuclear explosion
- the straightforward step of raising the yield of the primary can raise confidence in the performance of the secondary [3]
- the many thousands of parts in a warhead can be examined and tested in statistically significant numbers
In contrast, it would be prohibitively expensive to conduct enough nuclear tests to add a statistically meaningful measure of confidence to what was acquired from some half-dozen development tests devoted to each new weapon, followed by a test from an early production run to detect any error that might have crept in.
During the nuclear-testing era, stockpile maintenance already involved an extensive variety of destructive and nondestructive laboratory tests of components and subsystem, and high-fidelity flight tests in which the physics package was replaced by a data recording device that also simulates the weight and moment of inertia of the actual weapon. [1]
From 1972 to the cessation of testing in September 1992, out of the total of some 350 underground nuclear tests, [4] 68 were not devoted to development. [1] Of the latter, 10 were "stockpile confidence tests" conducted on all 9 types of deployed weapons in the existing and enduring stockpile, while 33 others had direct application to these weapons. [5]
No separate overview of what was learned from the post-1972 tests is available, but such data for the entire period 1958-1992 has been published. [1] As shown in the figure, some 830 "findings" of potential defects and anomalies in warheads were discovered in a wide variety of activities, of which 1% came from all types of nuclear tests. Note that this compilation includes such findings in the new designs that were deployed after inadequate development testing, or even none, because of the 1958-61 testing moratorium. [1,6]
Activities where findings are first discovered. Approximately 75 percent have been found by the Stockpile Evaluation Program.
AAU = Accelerated Aging Unit
LLCE = Limited Life Component Exchange
R&D = Research and Development
SIP = Stockpile Improvement Program
Source: Johnson, K. et al. "Stockpile Surveillance: Past and Future" (Sandia National Laboratory, January 1996).
Statistics alone cannot, of course, dispose of the possibility that the handful of findings from nuclear tests other than those related to the moratorium were of great significance. But no knowledgeable person has made such a claim, and the evidence indicates that nuclear testing has played only a minor role in maintaining confidence in the modern stockpile.
Stewardship Without Testing
In signing the CTBT, the president in effect bound the United States to the provisions of the treaty, so only dramatic and unpredictable political or technological developments could lead the United States to resume nuclear testing, whether or not the treaty is ratified and comes into force. Furthermore, current policy requires the weapon labs, the Secretary of Defense, and the military to annually certify all nuclear weapon types in the stockpile. This has been done since testing stopped and makes it most improbable that new designs will be deployed in the future unless the de facto CTBT regime collapses.
As a result, the shelf life of a warhead type can no longer be set by the tempo of replacement by more modern types, but will have to be indefinite. To meet this requirement, resources and facilities are needed to refurbish or remanufacture any and all parts of all weapons of a given type to their original and proven specifications when inevitable signs of deterioration appear, and a scientific staff having the requisite esoteric skills must be retained when no new weapons are being developed and tested.
In facing this problem, the labs have placed their greatest emphasis on developing "a sufficiently detailed understanding of the science and technology that govern all aspects of nuclear weapons," to quote Bruce Tarter, the director of Lawrence Livermore National Laboratory. To this end, Congress is funding the science-based Stockpile Stewardship Program (SSP), a suite of diagnostic tools and computers that go far beyond what was available for developing any of the weapons in the enduring stockpile. A major and reasonable purpose of these facilities is to provide the sophisticated tools that scientists of high caliber would expect to have at their disposal.
Much of the work under SSP is devoted to phenomena that could impact aging and to providing early warning thereof, which needed much less attention when new designs were being introduced. But it is hyperbolic to claim, as did John Browne of Los Alamos, that "we are using a new method: a sequence of surveillance -- evaluation -- response. In this new paradigm we are using a fundamentally different set of tools." Aside from the controversial and costly National Ignition Facility, which is of questionable relevance to stockpile maintenance, much of SSP consists of upgrades or sequels to earlier installations, and today's methodology is not so different from that used for monitoring deployed weapons when testing was being done. [1]
The Directors' Testimony
For decades the position of the lab directors has been a major factor both in successful and in failed efforts to constrain nuclear testing. As Robinson testified, the White House met the labs' conditions after asking "what would it take to get you on-board?" These negotiations produced the Stockpile Stewardship Program, but apparently no enduring understanding.
It is natural and proper for the directors to keep the needs of their laboratories in mind when testifying. Weighing the technical issues in the light of the nation's geopolitical interests is ultimately the responsibility of the president and the Senate. Nevertheless, widespread ignorance and anxiety about nuclear weapons imposes a responsibility on the directors to not stimulate misperceptions that are likely to flow from what they say.
Yet two widely stated misconceptions arose directly from the directors' testimony: that today's stewardship programme replaces a programme that relied primarily on underground nuclear tests and that this "new paradigm" could fail and put the US deterrent in doubt unless future computer simulations live up to promise on schedule. As described already, however, the essential features of current stewardship programme were already in place when tests were still being conducted.
The directors largely ignored the enormous strategic transformation produced by the implosion of the Soviet Union, and that US nuclear weapons are designed to meet requirements set by Soviet targets. To varying degrees, they spoke as if the nation's survival would be at stake unless there were an ever-ready ability to promptly begin testing of new weapon designs should some party "break out" of the test-ban regime or should some dire threat suddenly arise which could only be countered by a novel nuclear warhead that cannot be found either in the strategic arsenal or the wide array of proven design for tactical nuclear weapon. In fact, the CTBT imposes no constraints on means of delivery, so that a virtually unlimited range of new requirements for nuclear weapon systems can be met without reliance on untested nuclear designs.
The most extreme warning was Robinson's statement that "we will be at an intolerable disadvantage" should others conduct tests with very low yields while the United States adheres to the zero-yield limit. This issue has been examined in a high-level technical report whose unclassified conclusions [3] state that such low-yield tests would actually require modifications of warheads that would turn them into "at best a partial and possibly misleading performance indicator."
Of course, Robinson's nightmare raise more than a technical question. The broader issue was addressed by Gen. Shalikashvili when he testified that the Joint Chiefs had concluded that "with a very mature and sophisticated nuclear program ... we would be better off to go to a zero-yield testing [limit], because it would have less of an impact on our stockpile" than on other countries "we're concerned about."
The Price of Rejection
The US formally committed itself to the CTBT in leading the campaign to gain indefinite extension of the Non-Proliferation Treaty.[2] The Senate's rejection has, therefore, done severe damage to the nation's ability to conduct diplomacy, for in essence the Senate stated that the United States will not abide any shadow of doubt about its nuclear weapons, whereas the vital interests of non-nuclear states are best served by not acquiring such weapons!
Rejection makes it improbable that India and Pakistan will now sign, and quite likely that they will resume testing to develop weapons more reliable and potent than those they have tested. A nuclear arms race on the Indian subcontinent, given the proximity of China, would have global consequences. Nor can it be excluded that China, and perhaps even Russia, will resume testing as both have more plausible reasons than does the US to be concerned about abstention.
India and Pakistan advertised their tests, but they could not have evaded detection even though the surveillance system that the CTBT would produce is not in place. While doubt has been cast on India's claim that it successfully tested a two-stage thermonuclear device, there is little question that it is working towards such a weapon or that it will reach this objective if it continues to test.[7] According to US experience, rooted in an enormous data base, test with a yield of 10 kilotonnes are required to establish confidence in such a weapon. [3] To conceal the test of such a device would be a very difficult undertaking for a proliferator, and could well misfire and produce a detectable yield with untoward political consequence.
All agree that nuclear proliferation poses a grave threat to the international order and to the vital interests of the United States. While no one believes that the CTBT is a magic bullet that will eliminate this danger, its benefits are very substantial and must be kept in focus. The existing verification system suffices to detect tests needed to develop sophisticated thermonuclear weapons, whether or not a state already possesses a fission weapon, or whether it seeks to turn stolen design information into a deployable weapon. Should the treaty go into force, it will create a much stronger verification system than the one now at work [8], which already sets a low ceiling on the yield of clandestine tests. Only far-fetched scenarios can conjure up risks to the US that are comparable to these benefits.
One must hope that the US will be able to reverse course before rejection has such consequences. If that opportunity arises, the directors will have to conduct an embarrassing retreat from the branch on which they are now perched.
Sources
- Johnson, K. et al., "Stockpile Surveillance: Past and Future," Sandia National Laboratory (SAND 95-751/UCUC-700), January 1996.
- Ledogar, S.J., testimony to Senate Foreign Relations Committee, October 7, 1999.
- "Nuclear Testing," Summary and Conclusions (unclassified), the JASON Nuclear Testing Study, S. Drell (chair), MITRE Corp., August 1995.
- Mikhailov, V.N. (ed.), "Catalog of Worldwide Nuclear Testing," New York: Begell House (1999).
- Johnson, K., Lawrence Livermore National Laboratory, private communication.
- Kidder, R.E., "Maintaining the US Stockpile of Nuclear Weapons During a Low-Threshold or Comprehensive Test Ban," Lawrence Livermore National Laboratory (UCRL-53820), October 1997.
- Wallace, T.C., "The May 1998 India and Pakistan Nuclear Tests," Seismological Research Letters 69, No. 5, September/October (1998); Perkovich, G., "India's Nuclear Bomb -- The Impact on Global Proliferation," University of California Press (1999).
- Richards, P.G., and Kim, W-E. Nature 389, 781-782 (1997).
