The Comprehensive Test Ban Treaty and US National Security

By Richard L. Garwin

Richard L. Garwin
Richard L. Garwin
On October 13, 1999, the US Senate voted against approving the Comprehensive Test Ban Treaty by a vote of 51 to 48. Senate advice to ratify would have required 67 votes. Signed in September 1996, the CTBT had lain before the Senate for two years with no committee action or report, before being scheduled abruptly for a vote within 11 days. The Administration and its Senate supporters, whatever the merit of their case, were outfoxed by the opposition.

Yet the CTBT is greatly in the US security interest. Even opponents of the CTBT generally support the Nonproliferation Treaty, extended indefinitely in 1995, that binds parties with nuclear weapons from transferring them or knowledge to build them, to non-nuclear weapon states (NNWS), while binding Parties who are NNWS not to acquire nuclear weapons. A condition of that extension was the negotiation of a CTBT that would forbid nuclear test explosions by Nuclear Weapon States and by non-members of the NPT such as India, Pakistan, and Israel. As signed, the CTBT bans every nuclear explosion, of whatever yield - it is a "zero threshold" ban on nuclear explosions, and a permanent treaty. It creates an International Monitoring System (IMS), mandating cooperation of the parties in the emplacement of seismic sensors, hydroacoustic sensors, detectors of atmospheric infrasound, and sensors for radioactive particulates and gases from explosions. It provides also for on-site inspections (OSI) of events identified by the IMS, and it permits parties to bring non-IMS evidence, such as that obtained from general intelligence sources. The CTBT cannot enter into force until 44 specified countries sign and ratify. Once a state signs, under conventional international law it should not test unless it declares that it has no intention of pursuing ratification.

The US has conducted more than 1000 nuclear explosions and has about 10,000 nuclear weapons. The Soviet Union, with some 6000 ready strategic weapons and perhaps 18,000 total, could destroy not only the US but much of the civilized world - hence the long-standing US interest in reducing the number of deployed strategic nuclear weapons and ultimately the number of nuclear weapons of all types. The NPT has played a major role in holding the number of states possessing nuclear weapons to eight instead of the dozens foreseen within a decade in the 1960s.

Since 1992, the US has had a moratorium on nuclear testing, signed by President Bush and passed by large majorities in both Houses. Both Bush and Clinton have stated that we have no need for nuclear weapons of new types and no need for nuclear tests. Under these circumstances, there is annual certification by the Secretaries of Energy and Defense that each of the ten types of weapons in the enduring stockpile continues to be safe and reliable. Enabling this certification is a long-standing surveillance program, by which nuclear weapons are drawn from the stockpile at random - 11 of each type, annually - and carefully inspected.

Each US nuclear weapon consists of a primary fission explosive - a metal-sheathed plutonium shell surrounded by high explosive and arranged to be filled with deuterium-tritium (D-T) gas when it is imploded by that explosive. Upon implosion and the provision of neutrons for starting the fission chain reaction, sufficient fission yield is generated to ignite the contained D-T, which "boosts" the primary yield to a level such that thermal x rays carry much of the fission energy to implode the secondary capsule. A radiation case of high-Z material contains the thermal radiation long enough for this to happen. Most of the energy release of the nuclear package comes from the fusion fuel and the neutron-induced fission in the surrounding U-235 or U-238 elements of the secondary.

US nuclear explosions were used for development of new-type weapons, for production verification tests after a stockpile of such weapons had been begun, for weapons effects tests on other systems, for investigation of "Peaceful uses of Nuclear Explosions" ( PNEs), and on physics experiments. Among these are the creation of super-heavy nuclei, and the exploration of pellet fusion to guide the laboratory Inertial Confinement Fusion program. There have been rare tests to qualify a nuclear replacement or modification to solve a stockpile problem identified in the surveillance activity. No problem with stockpile aging has even been identified in a nuclear test, although many aging problems are found in surveillance and repaired in weapons of the afflicted type.

Safety of nuclear weapons refers primarily to the prevention of accidental nuclear detonations from lightning or other accidents involving a nuclear weapon. The Sandia National Laboratories provides the Enhanced Nuclear Detonation Safety (ENDS) system, as well as means to prevent unauthorized detonation of a nuclear weapon.

US nuclear weapons have been developed also to be one-point safe, so that detonation of the high-explosive of the primary component of the nuclear weapon at a single point will not produce a significant nuclear yield: either zero or (by design) less than two pounds of high explosive, compared with a typical 100 kilotons of design yield. Nuclear tests were essential in verifying one point safety of the designs in the stockpile, but never need be redone. Nor have nuclear tests ever been used to establish reliability of the nuclear components, which the laboratories have always assumed and stated to be 1.0. Such reliability is established through design and inspection, together with replacement or remanufacturing of parts.

The Pentagon and the directors of the nuclear weapons laboratories support the CTBT as signed, when integrated with a package of six safeguards. One of these is an ongoing $4.5 billion annual Science-Based Stockpile Stewardship Program (SBSSP). Safeguard F, as modified during the negotiations surrounding the Senate vote in October 1999, states that the President will resume testing if a nuclear weapon type in the enduring stockpile, critical to US security, could not be certified safe and reliable without test. Part of the SSP is the creation of new facilities for computation, including the Accelerated Strategic Computing Initiative, the Dual Axis Hydrodynamic Radiographic Test facility, and the National Ignition Facility (NIF).

But the ability to act on the basis of stockpile assessment has not kept pace with the assessment process. For instance, the US has had no ability to manufacture or remanufacture the sealed metal primary "pits" since Rocky Flats, Colorado, was closed in 1989. Los Alamos in its facility at TA-55, has now produced a developmental pit, and will soon be able to manufacture at a rate of 20 or perhaps 50 pits per year. I have long advocated a facility that would produce 300 pits per year, but the need for such a facility has nothing to do with a CTBT.

Opposition to the CTBT in part stemmed from a reluctance to accept any constraints on activities of the US, which is the world's preeminent military power. Specifically, one might want at some time in the future to design nuclear weapons for new purposes or optimize them for a different missile, and that cannot confidently be done and proved without nuclear explosion testing. Beyond that, it was argued that other powers might cheat, so that their explosions could not be detected or identified as nuclear explosions via the IMS, with its goal of detecting one kiloton of underground explosive yield anywhere in the world. Finally, there is the charge that one cannot in fact maintain the stockpile safe and reliable over the decades without nuclear testing either for finding problems or validating fixes.

Even in 1945, the US did not test its gun-assembled U-235 fission weapon before using it on Hiroshima, and a newly nuclear nation in 2000 probably would not need to test either. A conservative military establishment would probably desire a test of a gun-type weapon and certainly of a plutonium implosion weapon. Beyond that, even an advanced power would have difficulty being confident of a new-design two-stage radiation implosion system, such as that demonstrated by the US November 1, 1952.

As for cheating, the existing seismic array in Scandinavia readily detects explosions at the level of a few tons at the Russian test site in Novaya Zemlya, and could detect a fully-decoupled explosion in a large cavity at a level of 0.25 kilotons. A JASON Report to the Department of Energy notes that a yield of about ten kilotons would be required to accommodate the typical full-scale primary output and a telltale secondary yield. On balance, my personal testimony is that explosions that escape the IMS with high probability are not of military significance to the US. Even those who do see military significance in such small tests do not advocate US nuclear testing in return, so that is not properly a CTBT issue.

Weapons with fire-resistant pits and so-called insensitive high explosive (IHE) constitute a portion of the stockpile. Stockpile safety could be further enhanced by redesign. But in fact the armed services in the early 1990s explicitly rejected modifications that would require testing, even though 15 such tests were available under the US moratorium. So the remaining question is that of maintaining the stockpile reliable indefinitely. In addition to the component surveillance, US bombs and missile warheads are subject to Joint Test Flights, in which the plutonium is removed, but almost all of the rest of the warhead or bomb is exercised, right to explosion. These give vital information and will continue to be conducted. Even without a CTBT, they could not be conducted as nuclear explosion in the atmosphere under the Limited Test Ban Treaty of 1963.

The SBSSP will provide increasingly good knowledge of the performance of geriatric weapons, with deeper and deeper wrinkles. I have always recommended a much more conservative approach that would remanufacture weapons when their nuclear components began to exceed the range of parameters in newly manufactured weapons. Of course, all non-nuclear components such as radars, batteries, capacitors, etc., can be upgraded and fully tested as well under a CTBT as in the presence of nuclear testing. The problem comes down to whether the laboratories will have the discipline to reject modifications of processes in the manufacture of the nuclear components likely to impair confidence in stockpile reliability.

Do I have perfect confidence in the reliability of our stockpile? I have more confidence than I did in 1992 in the days of nuclear testing. A recent (November 8, 1999) Panel report on the reliability of the US nuclear stockpile, mandated by Congress, recommends inter alia, "To hedge against future uncertainties, the Stockpile Stewardship Program should include the ability to replicate or design replacement nuclear weapons, for our most critical nuclear systems, that will

  • Have a long shelf life
  • Be at least as safe as current weapons
  • Provide a high degree of confidence that the weapons are acceptably reliable without full-scale nuclear testing
  • Be manufacturable

But if such new-design weapons are mandated under a CTBT, that will surely cause a lack of confidence in the nuclear stockpile. One can clearly far more readily remanufacture weapons of existing type than certify weapons of a new design.

In summary, with a continuation of the long-standing stockpile surveillance program and the weapon manufacturing capability that is needed in any case, the enhancements to the program will provide a more reliable weapon stockpile than the United States has ever had in the past. It would be folly to reject the CTBT and thus fail to erect a barrier to the acquisition by others of effective thermonuclear weapons, in order to permit the US to make further improvements in its already highly advanced stockpile.

Richard L. Garwin is an IBM Fellow Emeritus at the Thomas J. Watson Research Center. He contributed to the development of fission fusion weapons beginning in 1950 at Los Alamos, and has been a consultant to the nuclear weapon laboratories to the present day. His testimony of 10/07/99 can be found online at

APS encourages the redistribution of the materials included in this newspaper provided that attribution to the source is noted and the materials are not truncated or changed.

Editor: Alan Chodos
Associate Editor: Jennifer Ouellette

February 2000 (Volume 9, Number 2)

APS News Home

Issue Table of Contents

APS News Archives

Contact APS News Editor

Articles in this Issue
Chodos is New Associate Executive Officer
Man of the Century
APS Career Efforts Focus on Physics Departments
Mass Media Fellow Learns "Nuts and Bolts" of Journalism
Outreach and Community Service I
In the Journals
APS and BNL Host XXX e-Print Archive Mirror
In Brief
Poll Reveals All-Star Physicists
Cosmic CD Available at Last!
US R&D Spending Trends
Exploring "Who Did It?" with Forensic Science
Cornell "Nanoharp" Studies Vibrating Materials at High Frequencies
Zero Gravity: The Lighter Side of Science
Inside the Beltway: A Washington Analysis
The Back Page