By Sophie Kretzschmar. There is a strange concurrence of fundamentally different developments in the world today regarding nuclear disarmament: On the one hand, nuclear disarmament seems to have taken several steps backwards – nuclear weapon states are increasing and modernizing their nuclear arsenals, and debates about new nuclear arrangements, however unreasonable these discussions are, have entered the discourse. On the other hand, nuclear disarmament seems to have taken a step forward – a treaty banning nuclear weapons, supported by 139 of the world’s 197 states, is in force since 2021, and active work is on the way to implement the treaty. Across these divides, Germany’s stated goal is to build bridges: Nuclear disarmament is postponed, “taking into account the current security environment”, but the “commitment to strive towards a safe world free of nuclear weapons” remains. For the disarmament research community, this confirms the mission: to use this time to develop and prepare tools to support nuclear disarmament.

Such tools include measures to build confidence: As long as nuclear weapons have any strategic value, building confidence that no one is cheating is crucial – regardless of the specific path that leads to nuclear disarmament. Ideally, such measures would be agreed upon for the disarmament process (i.e., to ensure that the state is actually dismantling nuclear weapons) and for the disarmament outcome (i.e., to ensure that the state remains nuclear-weapon-free). For both, the measures could be based on two strategies: making a nuclear rollback (a) difficult (i.e. complicated, expensive, time-consuming, …) or (b) detectable.   

The first strategy is often referred to as irreversibility, a concept in which there is growing interest in the disarmament research community; the second is often referred to as verifiability. In practice, however, these terms are not always well defined. What I often observe is that irreversibility is used for measures that make a nuclear rollback “more difficult to hide” – implicitly mixing the concepts. Nevertheless, it might be useful to still distinguish the two concepts for analysis to highlight the different roles in the process: measures for irreversibility would be applied mainly by the disarming state; measures for verifiability by those who verify.

The concept of irreversibility is not meant to be binary as the word itself suggests. Rather, irreversibility should be seen as a spectrum, and work has been done to associate certain irreversibility measures with different levels of irreversibility: For example, the requirement to dismantle all nuclear warheads could be associated with a low level of irreversibility (level 1), the destruction of all highly-enriched uranium (HEU) and plutonium stocks with a higher level (level 3), and the further requirement to eliminate all nuclear facilities could be associated with an even higher level of irreversibility (level 5). The degree of irreversibility required in an application case would then have to be negotiated on a case-by-case basis, as it would depend heavily on the technical details, i.e., the specifics of the former nuclear weapons program, but also on the political circumstances.

The strong focus on nuclear materials and facilities in the context of nuclear weapons activities is based on experience: it is precisely the focus point of the International Atomic Energy Agency’s (IAEA) safeguards, which ensure that the NPT non-nuclear-weapon states comply with their obligations not to acquire a nuclear weapon. For these states, the IAEA monitors the nuclear material inventories and ongoing nuclear operations. The situation is different in NPT nuclear-weapon states: IAEA safeguards are applied, if at all, on a voluntary basis to a very limited number of materials and facilities, and the inventory of existing nuclear material is imprecisely known, even by the states themselves. This means, however, that the verification of the above-mentioned irreversibility measures will be complicated in practice: if the amount of nuclear material in the disarming state is not known in the first place, how can the verifier check that all nuclear material has been destroyed? In other words, how can one be sure that no weapons or materials have been retained to undermine irreversibility? Searching everywhere for secret stockpiles is clearly not an option.

Nuclear archaeologists propose to turn the question around. Instead of focusing on what is there, they assess what should be there – by reconstructing the nuclear history of the state and how much material has ever been produced and removed. These figures could then be compared with the number of weapons dismantled or the amount of fissile material destroyed to determine whether material has been set aside.

To reconstruct the nuclear history of the state, various methods have been developed. In general, historic operational records can be analyzed to understand how programs were run in the past. The reconstruction can also be based on physical evidence from former facilities: For plutonium production reactors, the operation history can be reconstructed by analyzing physical samples taken from the facilities, for example with the Graphite Isotope Ratio Method (GIRM). For GIRM, one uses the fact that graphite remains in a reactor throughout its life and gets irradiated during operations. The irradiation changes the isotopic composition of the graphite, and those changes can inform nuclear archaeologists to draw conclusions on how the reactor was operated and how much plutonium was produced. The method has been tested at the UK’s Trawsfynydd reactor, and similar methods have been proposed for other types of plutonium production reactors. Corresponding methods have been proposed for enrichment facilities, where samples of waste or material residues in the facilities are analyzed. Expanding the toolbox of nuclear archaeology by adding more methods or including historical documents in the analyses is an active area of research being pursued at the Nuclear Disarmament and Verification (NVD) group, the physics partner institute involved in VeSPoTec.

One of the most famous application cases of nuclear archaeology, although not called that at the time, is the verification of the disarmament of South Africa. In the 1970s, South Africa began a clandestine nuclear weapons program that resulted in six uranium-based nuclear weapons by the end of the 1980s. Different domestic, economic, and political factors led the South Africans to give up the program, dismantle those weapons (still kept as a secret), and join the NPT as a non-nuclear weapons state in 1991. Two years later, while IAEA inspectors were conducting inspections in South Africa, President de Klerk revealed to the public that there had been a nuclear weapons program, and he invited the IAEA to extend its inspections to also confirm that the past nuclear weapons program had been completely dismantled – a voluntary measure that South Africa was not legally obliged to take.

In addition to investigating the remnants of the nuclear weapons program, a key part of the IAEA mission was to verify that all nuclear material had been placed under safeguards and that no material had been withheld from the IAEA’s monitoring – a typical application case for nuclear archaeology. For reconstructing South Africa’s nuclear history, the inspectors focused on the enrichment facilities and took two general approaches: (a) establishing U-235 balances, and (b) analyzing the performance of the enrichment facilities.

In the first approach, the inspectors scrutinized the accounting and operational records of the facilities to check the U-235 balances: the amount of U-235 entering the facility (in the natural uranium feed) needs to match the amount leaving the facility (in the enriched uranium product and the depleted uranium waste), since the material is only separated during the process; diverted uranium would then appear as an imbalance between the numbers. In the second approach, the inspectors looked directly at the operation of the plants to calculate how much HEU should have been produced. This required a much deeper dive into the historical documents – the inspectors analyzed the daily operating records to precisely understand all enrichment operations; they conducted interviews with former South African personnel to understand the relevant details of the unique enrichment facility; and they even conducted forensic analyses of the documents to ensure that the documents were authentic.

Although these verification approaches seem simple in theory, their implementation showed the complexity of putting theory into practice. Establishing the uranium balances was complicated by several factors: the inspectors were confronted with a complex operational history – instead of operating the plant in a regular and steady manner, it was used for several campaigns, enriching uranium to different levels. Further complications were the need to account for material stuck in the plant’s filters and pipes, and uncertainty about the amount of U-235 left in the depleted uranium stream – because it was considered waste, the South Africans simply disposed of it without characterizing it. These complications could not be resolved at first – the IAEA’s first report pointed out that there was an “apparent discrepancy” in the balances, which they felt might be explained by these complications in the application.  

The second approach, to reconstruct the amount of HEU produced based on the performance of the plant, was no easy task either: initially, the inspectors overestimated the performance of the plant, and it took a lot of cooperative effort to explain and understand why the plant performed less well than expected by the inspectors. They had to consider that at the beginning of the operation, uranium unexpectedly got stuck in the filters of the plant and even caused some technical difficulties which led the plant being shut down, without any enriched uranium being produced at that time. Eventually, however, after two years of inspections, careful document reviews, and interviews with former staff, the IAEA was able to conclude that the apparent discrepancy had been resolved and that “it is reasonable to conclude that the amounts of HEU which could have been produced by the plant are consistent with the amounts declared”.

Ultimately, the case of South Africa’s nuclear weapons program is remarkable for several reasons: first and foremost, it confronts those who associate nuclear weapons with the beacon of ultimate power that everyone inevitably craves, with an example of a state voluntarily deciding to give them up and join the non-nuclear weapon community. Second, it shows that the narrative of a nuclear Pandora’s box, once opened and never to be closed, falls short: South Africa is now a non-nuclear weapon state, like many others, and barely anyone even remembers that this was different in the past. Finally, the case provides an example of the successful use of nuclear archaeology to verify disarmament – fantasies about remaining nuclear warheads from South Africa are at most the stuff of fictional novels.

As a disarmament verification case, it also showed that verification is indeed not just a technical numbers game; both juggling the numbers and working together, inspectors and South Africans, were essential. And for this, the level of cooperation shown by the South Africans was not only essential to get the numbers right, but it had a clear significance of its own – it helped to overcome the basic fact that no matter how much effort is invested “there can be no absolute assurance that all nuclear material [..] has been declared”.

This hurdle is even more relevant when considering today’s nuclear weapon possessing states: they now have far more complicated nuclear material production histories than South Africa, and many of the facilities and documents have been or are in the process of being destroyed – disarming these programs will require far more tools to deal with uncertainties. One big difference, however, is that we now have a precedent from which we can learn to continue filling the nuclear disarmament verification toolbox, with technical and social knowledge, and thus prepare to eventually repeat the South African success story.