“It will take all of us working together – government officials, and diplomats, academic experts, and scientists, activists, and organizers – to come up with new and innovative approaches to strengthen transparency and predictability, reduce risk, and forge the next generation of arms control agreements.”
– Wendy Sherman
U.S. Deputy Secretary of State
June 2, 2022
Designing Denuclearization Regimes: Agreement, Declarations, Objectives, and Inspections

Latest ACA Resources

November 2018
By Geoffrey Forden

How should a country plan a new denuclearization verification regime? Should it concentrate on accounting for fissile material, should it try to “verify” a declaration of the program’s history, or, as some analysts have suggested,1 should it simply redirect the weapons scientists toward peaceful purposes after dismantling the weapons infrastructure and monitor their activity?

Iraqi officials present the entire 11,807-page “currently accurate, full and complete” Iraqi declaration of weapons of mass destruction programs and dual-use facilities on December 7, 2002 in Baghdad. This is an example of the kind of extensive documentation that can be used to help verify denuclearization. (Photo: Ahmed Al Rubayyh/Getty Images)Enough historical examples have occurred to permit a review of the lessons learned from past denuclearizations and try to draw some conclusions on what has and has not worked. This article highlights the denuclearization of South Africa and Iraq as exemplars at opposite ends of the cooperation spectrum. What is perhaps most surprising is that even with the wide gulf in cooperation provided by each country, there is a consistent theme to both denuclearizations: the necessity to develop a “coherent technical picture” of their nuclear weapons program.

To do so requires using inspections to search for the evidence necessary to construct and to verify a coherent understanding of the country’s nuclear weapons and related infrastructure. Contingency plans for inspections with the goal of developing a coherent technical picture can be worked out in advance, but what is actually implemented can only be decided after an agreement has been negotiated. From such an agreement flows the requirements for any declarations the country being denuclearized must submit, as well as the objectives of the inspection regime. Only after the objectives are clear can the right contingency plan for the inspections be implemented.

Historical Examples

Although all inspection regimes are unique, a case study of them reveals a striking number of similarities in how verification is achieved. To illustrate this, consider South Africa and Iraq, although there are considerably more. The denuclearization of South Africa started out as a fairly standard comprehensive safeguards inspection done when a nation accedes to the nuclear Nonproliferation Treaty (NPT) and signs a comprehensive safeguards agreement with the International Atomic Energy Agency (IAEA).2 The initial declarations the IAEA received October 30, 1991,3 were the standard materials accounting declarations; and although the inspectors had their suspicions4 of why South Africa had such a large quantity of highly enrichment uranium (HEU) in metal form, they kept those suspicions to themselves. Naturally, that changed in March 1993, when President F.W. de Klerk announced that South Africa had produced a limited number of nuclear weapons. All of those had been dismantled, and their HEU had been demilitarized5 and returned to South Africa’s civilian Atomic Energy Commission.

As usual for the IAEA inspectors during that period and corresponding to their mandate, the initial emphasis was given to material accountancy.6 When South Africa admitted that they had had a nuclear weapons program, the inspectors were given a series of “detailed briefings on the various phases of the program”7 and a new list of associated facilities. These newly declared facilities were mostly associated with the weaponization of South Africa’s fissile material rather than its production and included facilities for the manufacture of the first demonstration nuclear device,8 the laboratories for the development of the neutron generators, and the complex where most of the final arsenal was assembled,9 among others.10

From the start, the IAEA had difficulties confirming South Africa’s fissile material balance. This was due to the developmental status of the aerodynamic enrichment process South Africa was employing, which led to an unexpectedly large amount of material “held up” in the enrichment cascades,11 and to the IAEA’s unfamiliarity with this technique.12 Countrywide material balances had to be abandoned, and IAEA efforts in this area shifted to understanding the Y-plant (South Africa’s enrichment facility designed to produce HEU) material balances based on daily operational logs of the facility.13 With these records and the examination of fragments of the non-nuclear components surviving South Africa’s “do-it-yourself” denuclearization,14 the IAEA was finally able to declare that the “HEU produced by the pilot enrichment plant are consistent with the declared scope of the nuclear weapons program.”15 Full verification of South Africa’s uranium-material balance did not occur for 18 years.16

If South Africa was a model of cooperation after the president revealed the existence of its nuclear deterrent, Iraq is at the extreme other end of the cooperation scale. Nevertheless, these inspections shared a recognition that verification cannot be absolute either because of practical issues such as humans making mistakes or because the country being inspected has willfully tried to conceal significant pieces of its program.

U.S. President George W. Bush, during a 2004 tour of Oak Ridge National Laboratory in Tennessee, is shown nuclear-related materials and equipment surrendered by Libya. Under pressure from the United States, Libya dismantled its nuclear weapons infrastructure in 2003–4. More than 55,000 pounds of Libyan nuclear equipment and documents were shipped from Libya and to Oak Ridge. (Photo: Tim Sloan/AFP/Getty Images)Iraq tried to conceal every aspect of its nuclear weapons program, including denying it had ever had any nuclear material. That was a patently false claim since its Russian-supplied, HEU-fueled research reactor had been under IAEA safeguards for years.17 Over the next seven years, the IAEA patiently tracked down evidence, typically to refute the most significant portions of the most recent version of the Full Final and Complete Declaration issued by Iraq. Arguably the most important discoveries were contained in the so-called Haider House Chicken Farm, which concealed a cache of documents including CDs containing information from the Iraqi group responsible for weaponization.18 This trove of documents might not have ever been found had not Hussein Kamel, Saddam’s son-in-law, defected to the West. That defection prompted Iraq to try to appear proactive by turning the documents over to the IAEA.

Of course, it is politically unacceptable that the success or failure of an inspection regime hinge on such unpredictable events. Be that as it may, these weaponization documents and the other information gathered through careful inspections over the years allowed the IAEA to state that, “over the many years[, the IAEA inspections] yielded a technically coherent picture of Iraq’s clandestine nuclear program” and that “there are no indications to suggest that Iraq was successful in its attempt to produce nuclear weapons.”19

In both cases, inspections yielded a “technically coherent picture” that was accepted as the truth because no significant evidence was found to contradict it. This should be a reminder that deciding when to declare a successful denuclearization is a technical and political decision because there will always be inconsistencies of some nature. It helps, of course, that the Iraq Survey Group, which went into Iraq after the 2003 Iraq War with the ability to jail and question Iraqi scientists involved with weapons of mass destruction, concluded that “the aggressive UN inspections after Desert Storm forced Saddam to admit the existence of the [nuclear] program and destroy or surrender components of the program.”20

Central to this methodology is the gathering of as much corroborative or, potentially, contradictory information as possible. Ideally, this information should come from as many different sources as possible, minimizing the chances that the country being denuclearized could create false information meant to divert the inspectors. Declarations of the program’s history are very helpful in this and in other aspects, and negotiations should seek to obtain those if at all possible. Inspection planners should be aware, however, that some information might be too politically sensitive to be released, for example South Africa’s suppliers. Yet, it is possible to verify the dismantlement without such complete declarations.

Planning for an Inspection Regime

Planning for the verification of a country’s denuclearization should take the lessons discussed above into account from the start. The nuclear weapons program may be so complex that verifying denuclearization can be expected to be a long and intricate process, so complex that some analysts have advocated giving up on verification and propose that the “best” approach is to dismantle what is known and then have Western scientists work shoulder to shoulder with scientists from the denuclearized country. This, they maintain, is the best way of assuring the world that they are not working on banned activities.21

Such an approach would be a radical departure from past denuclearizations and was not applied even to South Africa, where there was a considerable amount of trust. This article takes a different position and advocates a method similar to past successful denuclearizations. The planners, however, should be fully cognizant that a strict fissile material balance is unlikely to be achieved on a politically meaningful timescale. Instead, inspections should be planned from the beginning with a goal of establishing a politically and technically acceptable understanding of the country’s nuclear weapons program. This understanding, or “picture,” needs to be tested by inspections that search out contradictory information to convince the international community there are no significant omissions or discrepancies.

Peaceful denuclearization is a complicated process that can only move forward to the extent that the inspecting party and the country being denuclearized can agree on terms for the process. The entire process is controlled by the agreement negotiated between these two parties. This agreement sets objectives for the inspections and their day-to-day function. For instance, the inspecting party clearly might find it advantageous to remove the existing nuclear weapons as soon as possible, while the country being denuclearized might insist on retaining its weapons until adequate security guarantees are put in place. Another important aspect of the agreement is establishing rights and responsibilities of the inspectors during inspections. Contingency plans for a satisfactory inspection regime can be created for both these cases and everything in between, but each inspection contingency plan must start from some assumptions about the final agreement.

Importantly, without a document to which inspectors can point, each facility is free to set new and more limiting constraints on the inspectors. This can start a process of “obstructionism creep” where each new constraint acts as a precedent for all future inspections.

Hopefully, one of the points of the agreement is that a country being denuclearized provides a full declaration of its nuclear weapons program and its history.22 This declaration has multiple purposes. First, it should offer an initial and, hopefully, final account of the weapons program that can be used as a benchmark against which to compare data gathered from the inspections. The most important comparison should be looking for significant inconsistencies and contradictions. There are many physics and engineering relationships between components inside a nuclear weapon and inside a nuclear weapons program, and these would likely be impossible to counterfeit if enough of them were utilized. It was exactly these inconsistencies that allowed the UN Special Commission to detect the existence of Iraq’s biological weapons program despite Iraq’s determined efforts to completely conceal it.23

A second factor that is in some sense just as important is that declarations allow inspectors to discuss topics with experts from the country being denuclearized without too much worry about revealing information that would actually help that country with its bombs, something that might be called “proliferation by inspection.” Sticking strictly to concepts, techniques, and devices mentioned in the declaration can greatly reduce this danger, even if it does not eliminate it completely.

Once the agreement has been negotiated, inspectors can develop objectives for the inspection regime. They can start this even before the declarations are received or select the set of objectives from a contingency library. The actual objectives used are dependent on the agreement. For instance, the inspection objectives will clearly be different for a regime that allows open-ended inspections as opposed to one that limits the number of visits or even sets the inspection regime to a single visit to a single site.

International inspectors in Iraq conduct a gamma-radiation test at the remains of a research reactor in Tuwaitha in the aftermath of the 1991 Gulf War. (Photo: IAEA Action Team)Placing inspection objectives into the business school framework that other major businesses use may be instructive. The inspection agency’s vision is that they will denuclearize the country; their mission statement is that they will provide the international community with the confidence needed to remove sanctions and accept the country back into the international community; the objectives are the major steps the inspectors will take to accomplish their mission and build the necessary confidence. Many people have used the shorthand of “verification” for this step, but that ignores the inspectors’ other objectives, such as reducing the threat the country’s nuclear weapons present by removing them, isolating them in a remotely monitored storage area, or effectively “de-alerting” the weapons, or dismantling their fissile material production capacity.

Objectives must be practical given the agreement. One cannot dismantle the country’s fissile material production capacity under an agreement that only allows a single visit to single site because there is no certainty that there are no other such facilities. Under a single-site/single-visit inspection regime, the priorities must be carefully conceptualized and will certainly be different than an open-ended inspection regime. Of course, the vision and mission of the inspection agency also will likely be very different under such an agreement.

The inspections performed by the agency are a major part of the inspectors’ tool box. Each inspection should work to contribute to one or more of the objectives. Historical experience shows that a complete understanding of a country’s nuclear weapons program is impossible in practice. There will always be practical problems in establishing the facts about how much fissile material was produced, how many nuclear weapons were built, or even how many centrifuge plants the country has. These problems could arise from benign issues such as the fact that all human beings make mistakes or serious ones such as a wide-scale program of deception.

Because of this, inspection agencies should acknowledge these limitations at the start and plan on contributing to the inspection objectives by verifying a technically coherent picture of the weapons program. For instance, if the primary objective is the removal of all nuclear weapons from the country, secondary objectives that contribute to this primary objective include establishing how many weapons were produced. It is too much to hope for that this could be established through accounting for the fissile material produced and where it was shipped on a politically meaningful timescale. Instead, the inspections should look for corroborating information, such as production and shipment records for fissile material and estimates of fissile material used in each weapon from sizes of casting molds.

Fissile Material Accounting

Should an inspection regime concentrate on accounting for fissile material? Fissile material accounting can play an important role in any inspection regime, but there are a number of practical limitations. First, it is likely to take much longer than is politically acceptable. The time necessary for material accounting increases exponentially as the complexity of the weapons program increases. Verifying a complete material balance for South Africa took at least 13 years, until an industrial-scale machine was developed to measure isotopic content of barrels of enrichment tails.24 Further, a country being denuclearized might refuse to reveal its sources of imported materials. If that is consistent with the agreement and it might be for a number of reasons, then establishing a global material balance becomes much more difficult because of the uncertainty created by not being able to verify both sides of the deal.

Given these difficulties, inspectors should utilize fissile material production for what it can contribute to the technically coherent picture but not expect it to be the final answer.
Some examples:

  • Measure isotopics from environmental samples and samplings from tails and other “waste products” to look for consistency with weapons
    designs and other declarations. Inspectors should not expect a quantitative constraint arising from fissile materials on a politically meaningful timescale.
  • Check production records from known or declared production plants with shipment records
    from weapons assembly facilities. Are they consistent? Do the assembly records use all of the material known to be produced, or are assembly plants missing? Is more material needed to produce the number of weapons than has been shipped to the assembly plant (are there more fissile material production centers)?
  • Check employment records at known production plants for unexplained long periods of absence, which might signal experts leaving to work at other plants.25

Should the inspection team work to verify a declaration of the program’s history? Arriving at a technically coherent picture inherently involves an understanding of the history of the weapons program. It is possible to reconstruct this history as the inspections progress, which is essentially what the IAEA team did with Iraq’s first nuclear declaration, but it is much easier to have some basis from which to start. For instance, any program will inevitably make false starts or other digressions that will have wasted resources and time and created lasting infrastructure that can mislead inspectors. Iraq’s attempts to weaponize aflatoxin is a perfect example, although one even Iraq did not recognize as a dead end until after the inspections started.26

Starting from scratch will take much longer and could involve much more intrusive inspections than might otherwise be necessary. Of course, the country being inspected could simply deliver detailed briefings on the program, as did South Africa.27 Briefings have the advantage of being interactive, and the inspectors can ask questions and seek clarifications right then. Declarations have the advantage that they can form part of the inspectors’ tool kit and can be used as starting points for discussions at various facilities while avoiding the “proliferation by inspection” trap.

Should a nuclear weapons program’s infrastructure be eliminated without verification and the outside world only rely on “monitoring” the day-to-day activities of former weapons scientists after they have been redirected into other, peaceful work? If this is the only practical way of eliminating a country’s known nuclear weapons infrastructure, then it might be the preferred process.28

It does not address the issue of an unknown weapons infrastructure, and it places the inspecting agency in a similar situation to what the IAEA faced with Iraq’s nuclear weapons program before Operation Desert Storm.29 In this case, working day to day with the “redirected” weapons scientists, as has been suggested,30 should be avoided at all costs. Redirecting scientists after the country’s nuclear weapons program has been dismantled is a key component of preventing it from being reconstituted, but every day the foreign scientists collaborate with the “redirected” scientists without verifying the absence of any covert infrastructure is a day when the country being denuclearized has its cooperation endorsed without evidence.


Practical difficulties will almost certainly prevent any inspection from gaining complete certainty that a country has denuclearized. In the end, it will be a judgement call—both technical and political—as to how much corroborative evidence is necessary. The chances of success can be maximized by accepting these limitations at the very beginning and planning for them. This article has suggested a workflow for that planning process that emphasizes using as much information about the procedures in conjunction with material balance.

Formulating the agreement with the country being denuclearized, specifying the required information from the declarations, designing the objectives to be accomplished during inspections, and executing those inspections should all be oriented toward constructing a technically coherent picture of the weapons program. That picture can be used to seek evidence supporting it and contradicting it. If the contradictions are not significant while the major outlines of the program are supported by the evidence, then the completeness of the denuclearization can be approached in an objective fashion.

Not only will this workflow result in a systematic approach to verification, it will also involve stakeholders throughout the government in the discussion necessary to making the technical and political decisions about how much verification will be necessary.


1. Siegfried S. Hecker, Robert L. Carlin, and Elliot A. Serbin, “A Technically-Informed Roadmap for North Korea’s Denuclearization,” Center for International Security and Cooperation, May 28, 2018, https://fsi-live.s3.us-west-1.amazonaws.com/s3fs-public/hecker_carlin-serbin_denuc_rlc.pdf.

2. Adolf von Baeckman, Garry Dillon, and Demetrius Perricos, “Nuclear Verification in South Africa,” IAEA Bulletin, Vol. 37, No. 1 (1995): 42–48; International Atomic Energy Agency (IAEA), “Agreement of 16 September 1991 Between the Government of the Republic of South Africa and the International Atomic Energy Agency for the Application of Safeguards in Connection With the Treaty on the Non-Proliferation of Nuclear Weapons,” INFCIRC/394, October 1991.

3. David Albright and Andrea Stricker, Revisiting South Africa’s Nuclear Weapons Program: Its History, Dismantlement, and Lessons for Today (North Charleston, SC: CreateSpace Independent Publishing Platform, 2016).

4. Von Baeckman, Dillon, and Perricos, “Nuclear Verification in South Africa,” p. 48.

5. “Demilitarized” means that various highly enriched uranium components in South Africa’s gun-type weapon had been reshaped into very different configurations, partially concealing their use in nuclear weapons. Albright and Stricker, Revisiting South Africa’s Nuclear Weapons Program, p. 197.

6. Garry Dillon and Demetrius Perricos, “IAEA 1991-1995 Safeguards Criteria: Experience Gained in the Verification of the Completeness of the Inventory of South Africa’s Nuclear Installations and Material,” in International Nuclear Safeguards 1994: Vision for the Future (Vienna: IAEA, 1994), pp. 231–241.

7. Ibid., p. 238.

8. These were the Building 5000 complex. The “demonstration device” was re-engineered with increased environmental production and safety features from a design built by South Africa’s Atomic Energy Commission (AEC). Albright and Stricker, Revisiting South Africa’s Nuclear Weapons Program, p. 101.

9. The so-called Armscor/Circle facility.

10. Albright and Stricker, Revisiting South Africa’s Nuclear Weapons Program, p. 239.

11. Ibid., p. 68.

12. Ibid., p. 231.

13. The South African AEC insisted on keeping secret the names of its suppliers and did not originally provide shipping records to the IAEA.

14. Albright and Stricker, Revisiting South Africa’s Nuclear Weapons Program.

15. IAEA General Conference, “The Denuclearization of Africa: Report by the Director General,” GC(XXXVII)/1075, September 9, 1993; UN Security Council, S/1997/779, October 8, 1997.

16. B. Rollen et al., “Validation of IQ3 Measurements for High-Density Low-Enriched-Uranium Waste Drums at Pelindaba,” in Proceedings of the 7th International Conference on Facility Operations-Safeguards Interface, 2004.

17. Garry Dillon and Jacques Baute, “An Overview of the IAEA Action Team Activities in Iraq,” n.d., www.iaea.org/inis/collection/NCLCollectionStore/_Public/33/034/33034358.pdf.

18. UN Security Council, S/1997/779, October 8, 1997.

19. Ibid., para. 21.

20. Mark Hosenball, “Held: Iraq’s Scientist,” Newsweek, June 19, 2005; Charles Duelfer, “Comprehensive Report of the Special Advisor to the DCI on Iraq’s WMD,” September 30, 2004, https://www.cia.gov/library/reports/general-reports-1/iraq_wmd_2004/.

21. Hecker, Carlin, and Serbin, “Technically-Informed Roadmap for North Korea’s Denuclearization,” p. 9.

22. Writing declarations is a complicated and difficult process that can initially contain many errors and omissions. For this reason, negotiators from the inspecting country might want to consider offering classes to technical experts from the country being denuclearized. In no case should the inspectors offer to assist in writing the actual declarations.

23. Rod Barton, “The Application of the UNSCOM Experience to International Biological Arms,” Critical Reviews in Microbiology, Vol. 24, No. 3 (January 1, 1998): 219–233.

24. B. Rollen et al., “Validation of IQ3 Measurements for High-Density Low-Enriched-Uranium Waste Drums at Pelindaba.”

25. The examples of developing a technically coherent picture are illustrative, not exhaustive, and some of the specifics listed here might be impractical.

26. Dany Shoham, “Iraq’s Biological Warfare Agents: A Comprehensive Analysis,” Critical Reviews in Microbiology, Vol. 26, No. 3 (January 1, 2000): 179–204.

27. Baeckman, Dillion, and Perricos, “Nuclear Verification in South Africa.”

28. Because of the way the negotiations developed, the choice might be between elimination with no verification or no elimination. This then might be the best practical method of denuclearization.

29. Some might point to the IAEA Model Additional Protocol as a sufficient improvement in the inspection regime, and the author agrees. Yet, it is difficult to imagine a country refusing to allow the verification of the dismantlement of its nuclear program and still signing an additional protocol.

30. Hecker, Carlin, and Serbin, “Technically-Informed Roadmap for North Korea’s Denuclearization.”


Geoffrey Forden is a physicist and principal member of the technical staff at the Cooperative Monitoring Center at Sandia National Laboratories. This article describes objective technical results and analysis. Any subjective views or opinions that might be expressed in the paper do not necessarily represent the views of the U.S. Department of Energy or the U.S. government. Sandia National Laboratories is a multimission laboratory managed and operated by National Technology and Engineering Solutions of Sandia, LLC, a wholly owned subsidiary of Honeywell International Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525.