The Role of Verification in Supporting Irreversible Nuclear Disarmament

ABSTRACT

The links between verification and irreversibility are not well understood. We show that verification and irreversibility are deeply linked and mutually supporting. To support this, we map activities through stages of a disarmament process, and consider how these contribute to irreversibility both individually and in aggregate, and how each is impacted by verification. We then re-examine verification concepts to identify qualities of verification processes that are especially relevant to irreversibility. We conclude with reflections on how irreversibility could be applied in practice and additional research questions that could progress this issue.

KEYWORDS:

• Nuclear disarmament
• irreversibility in nuclear disarmament
• nuclear disarmament verification
• arms control verification

Introduction

Irreversibility is considered to be one of the three pillars of nuclear disarmament, alongside verification and transparency.¹ The term has become widely used and acknowledged in nuclear disarmament discussions; however, a shared understanding of the concept is still lacking (Norway, and United Kingdom 2021). A challenge with defining irreversibility is that in practical terms, the knowledge of splitting the atom cannot be erased or forgotten, nor can associated engineering processes in fissile material production or nuclear weapons design be uninvented.² A strict interpretation of the term is not applicable; rather, it is more useful to conceptualise it as a reversibility-irreversibility spectrum (Cliff, Elbahtimy, and Persbo 2011). We will consider a definition for irreversibility offered by Elbahtimy: “Irreversibility: a feature/quality of a disarmament or an arms control process that involves limiting the capacity for re-armament, including the possible re-constitution of aspects of weapons programmes” (Elbahtimy forthcoming). As already noted in VERTIC’s 2011 report (Cliff, Elbahtimy, and Persbo 2011), verification is not irreversibility, and verification alone does not guarantee irreversibility. However, much in the same way verification supports and complements the pursuit of other goals of arms control and disarmament processes, a well-considered verification regime can support and complement the pursuit of irreversibility in nuclear disarmament.

To explore how verification can contribute to the pursuit of irreversibility of nuclear disarmament we first examine the “life cycle” of a disarmament process, and clarify the concept of greater and lesser degrees of irreversibility. We then consider specific activities that could take place during the different stages of disarmament to increase the level of irreversibility of the process and its end result. Following the overview of a disarmament process, we consider some key characteristics and attributes of verification in arms control and disarmament settings, and re-examine them to discuss their value in supporting the irreversibility of disarmament process. As a result, we identify some qualities and features of verification that are particularly relevant for the pursuit of irreversibility of nuclear disarmament.

Irreversibility and Verification in the Disarmament Process

Achieving complete nuclear disarmament, defined as the end-state of a disarmament process where all states have eliminated their nuclear weapons, from the current global status will require a number of measures and activities that will most likely take place in sequential stages. Progress through such intermediate stages of nuclear disarmament can be supported by verifiable activities such as dismantlement of individual facilities, destruction of weapons, downblending of fissile material and control of sensitive knowledge and knowhow.

The scope of activities considered necessary to achieve complete (and, especially, irreversible) disarmament has gradually expanded over the years: initially, research focused more narrowly on dismantlement of warheads (IPNDV 2020); in time, it grew to encompass sensitive parts of the nuclear fuel cycle (Cliff, Elbahtimy, and Persbo 2011) and activities related to weaponisation and delivery systems (Dalton et al. 2017). Recent literature has considered the role of verification under current proposals such as in the context on the Treaty on Prohibition of Nuclear Weapons (TPNW) (Patton 2020). The approach in this paper does not assume a particular global disarmament treaty or framework.

To consider irreversibility as a spectrum, it is firstly useful to define some metrics for characterising a disarmament activity as more or less irreversible. A disarmament activity is “more irreversible” than another when it requires more time and more resources to reconstitute as shown in the contrast in Figure 1.

Figure 1. Highly irreversible disarmament over time and reversible disarmament over time.

Secondly, it is useful to distinguish between the irreversibility of a single disarmament activity and the irreversibility of a disarmament programme as a whole, arising from the combination of a number of individual activities. A single activity – such as the removal of material or the disablement of a facility – may be more or less irreversible, depending on how it is implemented, according to the simple criterion outlined above. For example, downblending a stockpile of highly-enriched uranium (HEU) to a lower enrichment level, or removing it to a different country, is inherently more irreversible than placing the material in storage under seals, even though all these activities achieve the goal of making the stockpile unavailable for enrichment. Irreversibility of a programme as a whole is harder to directly quantify; in essence, this refers to the systematic removal or degradation of a state’s capability to produce nuclear weapons, measured in the time and resources costs (and arguably, in the risk of discovery by adversaries) that would be incurred to once again produce weapons. For example, even an activity with relatively low irreversibility such as the secure in-country storage of HEU stocks outlined above, combined with other activities such as the reduction or dismantlement of centrifuge facilities and the removal of facilities for uranium metallurgy, could result in a high level of irreversibility. Notably, a situation where several measures need to be reversed offers a greater chance of early detection even when individual measures are relatively reversible.

Verification Requirements for Staged Disarmament Irreversibility

The disarmament process can be divided into:³ an initiation phase, where preparatory work takes place; a “draw-down” phase, which could last many years,⁴ where the main disarmament activities take place; and a complete disarmed end-state, where the main goals of the disarmament plan have been achieved, and the focus rests on preventing re-armament.

Initiation Phase

Before weapons, material or facilities are removed or destroyed, the first step of disarmament is the formal decision by a state to dismantle its nuclear weapons programme, and – most likely – its accession to an international agreement to that effect. After this step, states would submit an initial declaration of their holdings and treaty-accountable items (IPNDV 2022). In addition, while a future disarmament treaty is likely to define activities and timelines for implementation, states will need specific implementation plans as applied to their national fuel cycles and holdings.⁵

The goal of verification activities in the initiation phase is to ensure that states cannot retain an undeclared stockpile of weapons or material before implementation. Even more importantly, verification in the initiation phase creates a baseline of knowledge about the states’ nuclear weapons and infrastructure that underpins the whole verification process. Establishing this initial knowledge, and refining it through observation over time, allows for more precise understanding of a state’s capabilities. Verification of historical activities at older facilities faces uncertainties and inconsistencies (Fetter 1993): baseline knowledge of a state’s capabilities and holdings is crucial to help distinguish between process-related inconsistencies, and possible indications of undeclared activities.

Draw-Down Phase: Implementation of Measures Agreed Under the Disarmament Plan

During this phase, states will take steps to dismantle or disable their nuclear weapons, to reduce their holdings of fissile material, and alter their fuel cycle to reduce their capability to produce weapons-usable material. We divide verified disarmament activities into four types: weapons dismantlement, disposal of fissile material, decommissioning of facilities and long-term knowledge management. A set of activities with the potential to enhance irreversibility are shown in Tables 1–4.

Table 1. IND scenarios related to weapons dismantlement with our estimates of confidence in verification.

Disarmament scenario	Description and examples	Confidence in verification
Unilateral nuclear weapons dismantlement with no external observers	This is the model followed by South Africa. South Africa conducted its disarmament and dismantlement of its weapons before declaring the stockpile (Von Wielligh and Von Wielligh-Steyn 2015).	All states have high confidence that South Africa did disarm (Perkovich and Acton 2009; Von Baeckmann, Dillon, and Perricos 1995).
Dismantlement with NWS verification only	The INF follows this model.	High for the other bilateral NWS party, lower for other parties.
Multi-party nuclear weapons dismantlement with NNWS observers	Since the UK-Norway initiative, NNWS-NWS collaboration has evolved and this is the current model for the IPNDV (Hinderstein 2018; IPNDV 2020).	Highest currently available for all states.

Table 2. Activities that could be conducted under IND for the endpoint of fissile material with our estimates of confidence in verification.

Fissile material activity	Description and examples	Confidence in verification
HEU downblending	Used in Megatons to Megawatts program (Laurence Livermore National Laboratory 2013) and the JCPOA.^6	Very high for parties involved. Third parties have indicated high confidence in all processes so far.
HEU transport to another state	Used in the disarmament of Libya (Davenport 2018), and Project Sapphire which HEU was removed from former Soviet states (Tirpak 1995).	Requires a credible host state.
Placing HEU under safeguards	South Africa has retained HEU under safeguards (Stott 2022; Von Wielligh and Von Wielligh-Steyn 2015).	Material is continuously monitored.
Pu burning in MOX	Mixed Oxide (MOX) reactors use plutonium oxide from spent fuel mixed with reprocessed plutonium. After being burned up, products are immobilised and put in geological disposal. MOX has been irradiated but none has been placed in a geological disposal site.	Requires MOX reactor programme in place. MOX fuel cycles have proved difficult to maintain and operate.
Direct geological disposal of Pu	Immobilisation (Nuclear Decommissioning Authority 2019) and geological disposal. This has occurred in the United States at the Waste Isolation Pilot Plant facility (Schramke, Santillan, and Peake 2020).	Material can be continuously monitored.
Pu storage overseas	The UK and France currently have Japanese civil Pu that was reprocessed there under safeguards.	Requires a credible host state.

Table 3. IND activities related to decommissioning and dismantlement with our estimates of confidence in verification.

Decommissioning and dismantlement Activity	Description and examples	Confidence in verification
IAEA decommissioning/decommissioning with third party oversight	The IAEA has a model for decommissioning civilian facilities (Hogue, Hale, and Cooley 2022).	High
Other decommissioning without third party oversight	NWS have unilaterally decommissioned older weapons-related facilities. Voluntary offer agreements with IAEA and bilateral arms control agreements have not covered some facility/building decommissioning.	This could be very challenging for older, complicated facilities with incomplete operational records.
Removal and transport of key equipment overseas	Used in the disarmament of Libya (Tobey 2017).	Requires a credible host state.
Repurposing of facility for civilian use	Used in the JCPOA as a plan for Fordow to be converted into a stable isotope production facility and science and technology centre (Albright et al. 2019).	Depends on the facility and the degree of dual use potential.
Disablement of key equipment	Used in the DPRK during the Six-party talks at the 5 MWe Magnox-type reactor and the reprocessing facility (Nikitin 2013).	Depends on the specific action taken. During the 6-party talks most disablement activities were designed to be reversible, and some such as the demolition of the cooling tower did not significantly alter operation of the 5MWe Magnox type reactor.

Table 4. IND activities related to knowledge transfer and tacit knowledge with our estimates of confidence in verification.

Disarmament Activity	Description and examples	Confidence in verification
Destruction of documents/files	South Africa destroyed many documents prior to IAEA verification	It can be difficult to verify “completeness” and that no documents are being saved, stored or have been copied.
Handover of documents/files	The “plastic bag” of weapons designs passed to the USA in the disarmament of Libya (Tobey 2017).
Long-term employment of key personnel in other industries	Post-Cold War arrangements for former Soviet Scientists	It can be difficult to verify “completeness” and that no documents are being saved, stored or have been copied.

Verification in the Complete Disarmament End-State

Once the disarmament process has concluded, verification will still be required long-term to monitor compliance and to provide timely detection of re-armament. At this stage, the goals and means of verification will likely resemble those of the existing international non-proliferation regime. However, a number of specific considerations apply.

The first consideration is that long-term verification activities should be prioritised according to, among other things, the degree of irreversibility that was achieved through disarmament activities. Any disarmament process will be the subject of negotiation and, inevitably, compromises. As such, it is likely that some parts of a state’s nuclear infrastructure will be more deeply affected by disarmament activities with a high degree of irreversibility, while others will retain a higher level of latent capacity. For example, a state⁶ may willingly relinquish its plutonium stockpiles and dismantle its plutonium production facilities, while retaining enrichment facilities and enriched uranium stockpiles for use in its civilian nuclear industry. Long-term verification measures should focus on those areas of a state’s nuclear infrastructure where activities or holdings remain that have or contribute to a lower irreversibility level, thus reinforcing any possible “gaps”.⁷

The second consideration is that even once disarmament activities are concluded, it may take time for the verification system to build the same level of confidence states have, for example, in the IAEA safeguards regime, especially when it comes to former nuclear weapons states. These states have vast holdings of material and facilities with a decades-long history, some of which have never been subject to verification activities. Should these fuel cycles be subject to a verification regime that includes provisions like a broader conclusion, these may take significant time to reach, and entail a politically charged process (VERTIC 2023). Moreover, there are issues related to advanced fuel cycles that present challenges even in the current regimes: for example, states with large holdings of highly enriched uranium or separated plutonium.

A further challenge is related to the ways weapons-related facilities and expertise intersect with verification capabilities. Due to their very nature, these facilities are a suitable location to conduct some key disarmament activities, such as dismantlement of warheads; moreover, these facilities and the institutions that host them are often involved in developing tools and approaches for existing verification regimes, and conducting research on future verification issues (such as nuclear disarmament verification).⁸ Some of these capabilities and areas of expertise are likely to remain relevant in the complete disarmament end-state; however, given that they are supported by advanced, in-depth knowledge of not only the nuclear fuel cycle, but cutting-edge nuclear technologies and weaponization activities, disentangling verification expertise from weaponization-usable work may be a significant, yet unavoidable challenge.

The Contribution of Verification to Irreversibility

This section provides an overview of the ways verification processes can contribute to increasing the quality of irreversibility in a disarmament process. It does so by analysing some key functions and goals of verification, and considering the specific ways they support achieving irreversibility.

Building Confidence and Reducing Incentives to Hedging

The first element to consider regarding the role of verification in supporting irreversibility is the ability of verification processes to build confidence and allay fear of cheating. This is key to both achieving a disarmament agreement and sustaining it in the long run.

As states contemplate the political decision to disarm, they must consider the risk that an adversary may retain capabilities that they themselves are giving away; as such, fear of cheating provides an incentive against joining arms control or disarmament agreements, similar to considerations for strategies of proliferation (Narang 2017). A strong verification system can increase confidence that cheating could be detected early enough to be able to respond before a militarily significant benefit is gained, influencing the risk calculus of states and reducing obstacles to participation in verification agreements.⁹ A similar calculus is also relevant during international crises, where heightened perception of risks and uncertainty about an adversary’s actions may prompt states to extreme measures, including potentially pursuing re-armament. Even once the decision to disarm is taken, confidence born of an effective verification supports irreversibility by mitigating the perceived need for hedging. States with low confidence in the compliance of other parties have an incentive to accept a disarmament process with a lower degree of irreversibility, in order to retain a latent nuclear capability that could be weaponised rapidly in response to an adversary. This, in turn, results in an inherently less-irreversible end-state. By offering confidence in timely detection of non-compliance, effective verification protocols can help justify the acceptance of higher degrees of irreversibility.

These issues highlight two qualities of possible verification regimes that support irreversibility: the first is high confidence of all parties in the verification regime itself, and the second is confidence of a timely discovery of non-compliance.

The first quality is especially relevant for multilateral disarmament and verification processes. In a multilateral setting, there are often obstacles to the use of information gathered through national technical means, and emphasis is placed on the official (often treaty-mandated) verification processes. In order to inspire confidence in its findings, the verification regime needs to be comprehensive and robust, and its capabilities must be well understood by all participants. It is worth noting that conclusions reached by a multilateral mechanism are unlikely to be the last word on any case where non-compliance is alleged: states are likely to draw their own conclusions, especially when they see their own security threatened. However, a mechanism whose capabilities are well understood, transparent, and of proven effectiveness can draw on these qualities to speak authoritatively.

The second quality that emerges as a requirement is that of timeliness. This is relevant both in the planning of disarmament activities, and in the goals and means of the verification regime itself. Planning disarmament activities and their verification to maximise timeliness means deliberately constraining a state’s nuclear weapons capabilities, such that reconstituting them would take significant time. Accordingly, a verification regime built with timeliness as a goal would aim to identify early moves to reverse disarmament activities, and would offer assurance that attempts to re-arm would be discovered early in the process, giving the international community an occasion to respond. This is a recognised priority in regimes such as the IAEA Safeguards regime (VERTIC 2023).

Transparency and Norm-Building

Another way in which verification supports irreversible nuclear disarmament is by providing opportunities for voluntary transparency and norm-building. By participating actively in multilateral verification and engaging in deliberate transparency, states can build confidence among other parties about their actions and intentions (UNIDIR and VERTIC 2003). Moreover, states can use their participation in verification regimes as a way to signal their support of disarmament norms, engage in norm-building, and claim roles as leaders or norm entrepreneurs. This can strengthen the political underpinnings of irreversible disarmament, as building a global norm against nuclear weapons – possibly as a peremptory norm eventually – has been identified as a further way to bolster the complete disarmament end-state and dissuade states from re-armament.

This process is not new, and has been observed in a number of settings, such as with Non-Nuclear Weapons States taking on additional responsibilities under IAEA Safeguards (such as the Additional Protocol or the Modified Small Quantities Protocol) even when they have little or no nuclear activities and material, for the explicit purpose of demonstrating their commitment to non-proliferation norms and to the Safeguards regime (Mayhew et al. 2023).

Deterring Non-Compliance

Finally, verification supports irreversibility by deterring non-compliance (UNIDIR and VERTIC 2003). Robust verification processes increase the likelihood that re-armament attempts would be uncovered, exposing the would-be cheater to censure, and likely enforcement action.¹⁰ Moreover, this could happen when the state is still months or years away from being able to once again deploy nuclear weapons, creating a window of vulnerability for the cheating state. While states can take measures to circumvent verification regimes, these can increase the costs, technical complexity, and time required to pursue re-armament. For example, a state may be forced to build clandestine facilities shielded from observation, instead of relying on existing infrastructure; or it may have to procure key material and equipment through illicit procurement networks.

Verification impacts both sides of the traditional deterrence calculus: the risks introduced by a robust and timely verification system can impact the expected benefits of pursuing re-armament; while measures required to carry out and shield clandestine activities, let alone the prospect of repercussions should a clandestine programme be uncovered, can increase the costs. To maximise the deterrent impact of verification measures in service of irreversibility, some key qualities are required. The first, outlined above, is timeliness. The second is that the regime must include measures specifically developed to address attempts to circumvent it. In nuclear non-proliferation, this lesson became evident with the failure of the IAEA Safeguards system to uncover the Iraqi nuclear weapons programme. This failure prompted the IAEA to develop a new set of verification tools specifically aimed at identifying undeclared activities, enshrined in the Additional Protocol. More recently, further measures were developed under the Joint Comprehensive Plan of Action (JCPOA) to provide additional monitoring of the Iranian nuclear fuel cycle. Importantly, measures of this kind are often extremely intrusive, and the negotiation of a future verification protocol for a nuclear disarmament agreement would likely see compromises in this area.

Conclusions

We have argued that while verification and irreversibility are distinct concepts, verification has a significant role to play in supporting disarmament activities with a high degree of irreversibility. Some of the more “traditional” arguments about the utility of verification have specific connotations in the irreversibility debate.

In order for verification measures to truly provide support to irreversibility as outlined, some key qualities must be ensured: high confidence in a robust, transparent, trusted verification process and its attendant institutions; timeliness; and comprehensiveness. The first point, as raised, is that a robust, comprehensive verification system is likely necessary for states to accept disarmament; in particular, a high degree of confidence not just in the detection of rearmament attempts, but specifically in a timely detection, can mitigate the need some states may perceive to maintain high levels of latent capabilities, and as such make more intrusive and irreversible disarmament activities more acceptable. States can also voluntarily engage with the verification process to demonstrate their commitment and engage in norm-building via practical, visible steps. This is particularly relevant because recent discussions on irreversibility noted that even after an agreement on nuclear disarmament is reached, building an international norm rejecting nuclear weapons – potentially to the point of having erga omnes value – is a way to strengthen the disarmament regime and make re-armament less likely. Thirdly, the deterrent effect of verification activities will provide ongoing support for compliance.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by King’s College London under a consortium grant from the UK Government’s Counter Proliferation and Arms Control Centre.

Notes on contributors

Alberto Muti

Alberto Muti is Co-Programme Director for VERTIC’s Verification and Monitoring Programme. He leads the programme’s work on IAEA Safeguards, and carries out research and analysis on nuclear disarmament verification and irreversibility, nuclear non-proliferation and nuclear verification, as well as verification and monitoring of biological weapons. Alberto carries out research on Iran and DPRK nuclear issues and has also contributed to projects on other issues, including chemical security, conventional and cyber security. Alberto’s work and expertise encompass planning and delivering capacity-building programmes, conducting workshops, designing tabletop simulations and “serious games” for research and training purposes, and developing innovative tools to facilitate and enhance research and analytical work.

Grant Christopher

Grant Christopher is Co-Programme Director for VERTIC’s Verification and Monitoring Programme. He researches nuclear disarmament verification using software modelling, non-proliferation in East Asia and the role of emerging technologies in nuclear proliferation.

Noel Stott

Noel Stott joined VERTIC in November 2016 after more than 14 years as a Senior Researcher and then Senior Research Fellow at the South African-based Institute for Security Studies (ISS). He has extensive experience in all aspects of arms control, disarmament and non-proliferation having worked on the challenges facing African states by the proliferation of conventional weapons. In particular, Stott has knowledge of small arms and light weapons. In 2007, he established and led ISS’ programme on ‘Africa’s Development and the Threat of Weapons of Mass Destruction’. This programme aimed to enhance Africa’s role in international efforts to strengthen weapons of mass destruction disarmament and non-proliferation initiatives in the context of Africa’s developmental imperatives through the provision of primary research, policy recommendations and training activities.

Notes

¹ Action 2 of the 64 Point Action plan from the 2010 Nuclear Nonproliferation Treaty (NPT) Review Conference states that “All States parties commit to apply the principles of irreversibility, verifiability and transparency in relation to the implementation of their treaty obligations”. This formulation has been adopted and maintained since the agreement of the 64-point action plan.

² The science of nuclear fission and nuclear physics is standard for many undergraduate degrees in physics and engineering. Design and engineering information for sensitive topics such as advanced centrifuges and weapons design are highly restricted. Nuclear technology is a combination of core knowledge, restricted knowledge and skill and tacit knowledge required for design, construction, maintenance and operation. Yet there is a limit to the notion that no technology can be uninvented or forgotten. For nuclear weapons, this concept is explored in (Bourne 2016).

³ It should be noted that this is a very broad and top-level characterisation, and the practical aspects of implementing a disarmament treaty would likely not be as neatly divided; however, it is useful to consider the changing needs and priorities of a verification regime.

⁴ It has been noted that the faster disarmament goals are achieved, the easier it is reach and maintain a general level of irreversibility in case conditions change mid-process. See for example (Cliff, Elbahtimy, and Persbo 2011)

⁵ For example, both an initial declaration and a state-specific plan for disarmament activities are included in the Chemical Weapons Convention, under Art. III par.1; this process is expanded on in Part IV (A) of the Convention’s Verification Annex.

⁶ In the Joint Comprehensive Plan of Action (JCPOA) section A para. 7, Iran could possess up to 300 kg of uranium enriched to 3.67%. The excess could be sold on international markets or downblended. The IAEA reported that Iran initially downblended to satisfy the JCPOA before later violations.

⁷ This proposal mirrors existing practices in IAEA safeguards, such as the use of Acquisition Pathway Analysis to identify and prioritise safeguards activities.

⁸ For example, the US National Laboratories support international nuclear non-proliferation and nuclear security through the development of equipment and techniques and expert contribution to capacity-building; moreover, both the US National Laboratories and the UK’s Atomic Weapons Establishment have historically worked on nuclear disarmament verification methodologies, in the context of initiatives such as the UK-Norway Initiative (UKNI), the Quad Nuclear Verification Partnership (with Sweden and Norway), and bilateral US-UK cooperation.

⁹ As is argued by Herzog, the risk-calculus by other states is best understood by the nuclear progression of a state – if they are in a “danger zone” for proliferation or not (Herzog 2020). Hymans examines the technical and political calculus that inform a state’s path to a nuclear weapon (Hymans 2012). The decision of a state to abandon nuclear ambitions and pursue a disarmament arrangement follows a comparable logic.

¹⁰ As an additional caveat, discussion of deterrence in verification necessarily raises the issue of enforcement: indeed, for a verification regime to be an effective deterrent, would-be cheaters must be assured that non-compliance would result in enforcement action. A full discussion of this topic is beyond the scope of this paper.