Spent Nuclear Fuel – Waste to Resource, Part 2: Radiolytic Hydrogen Production – A Novel Approach to Sustainably Meeting the World’s Hydrogen Needs from the Nuclear Fuel Cycle

1. Introduction and Context

Co-generation or cogeneration (CoGen) is defined as the exploitation of both heat and electrical power from energy generating sources (Royal Society, 2020; Matthews et al., 2024), including the recovery of waste heat and combined heat and power (CHP) (Khamis et al., 2013). In a nuclear context, this means the use of heat produced from reactors for various processes such as saltwater desalination, district and industrial heating, and hydrogen (H₂) production, in addition to conventional electricity generation (Figure 1) (Royal Society, 2020; Matthews et al., 2024). For the purposes of this work, we shall designate this concept as reactor co-generation (RCG). The primary goal of RCG is the decarbonisation of carbon-intensive operations currently driven by fossil fuels, by instead employing the vast amounts of heat generated within nuclear reactors (IAEA, 1999), in addition to conventional baseload power generation (Matthews et al., 2024). Secondary goals include load following more transient renewables such as wind and solar (Royal Society, 2020), and the flexible use of excess electrical or thermal output for other cogeneration applications (including hydrogen production) where renewables are generating at scale (Dong et al., 2021; Verfondern et al., 2017; Kowalczyk et al., 2019; IAEA, 1999; Novotny et al., 2024). Both conventional, large and small modular reactors (SMRs) are noted as being suitable for this concept (Alabbadi and AlZahrani, 2024, Vanatta, 2025), sometimes where these are coupled with high-temperature energy storage for alternative pathways to load following (Jin et al., 2024). In concert with nuclear energy, hydrogen is viewed as the optimal energy storage medium as a zero-carbon alternative to fossil fuels (Agyekum et al., 2025; Ali et al., 2025). It has been forecast that hydrogen and derivatives will make up 12% of worldwide energy consumption by 2050 (Guerrero-Rodriguez et al., 2024).

Despite RCG’s potential, several drawbacks have limited uptake beyond small-scale test beds. Firstly, most current reactors are water-cooled types with limited output temperature (≤ 335 ^oC), well below the minimum temperatures required to drive many industrial processes (Khamis et al., 2013; Royal Society, 2020). Consequently, such high temperature heat must be provided by resistive heating, with 60-70% losses in convert reactor heat to electricity. Secondly, the advanced, high-temperature reactors capable of driving thermally demanding industrial processes (T > 700 ^oC (Navotny et al., 2024)) are developmental at the time of writing, with lead times of a decade or more before construction starts except for a single prototype and twin-reactor first-of-a-kind station operating/under construction in China (IAEA, 2024). In any case, some of these designs have issues with the long-term sustainability of their fuel cycle as their spent nuclear fuel (SNF) cannot be readily recycled and lack turnkey disposal options for open cycles (Holdsworth and Ireland, 2024). Regardless, these designs have historically proven troublesome in operation (Holdsworth and Ireland, 2024; Ramana, 2016), with further potential concerns surrounding the development and efficacy of thermochemical cycles for H₂ production in time for high-temperature gas-cooled reactor HTGR deployment (Boretti, 2021; Li et al., 2024; Ahmaruzzaman, et al., 2025; Parkinson et al., 2019). This means that even HTGRs will likely rely on electrolysis for hydrogen production, albeit with a significant increase in thermal efficiency over pressurised water reactors (PWRs). Other technologies driven by HTGRs, such as methane pyrolysis (with a carbon by-product), and high-temperature steam electrolysis using solid oxide fuel cells, are under consideration (Onishi et al., 2025). The round-trip efficiencies of nuclear-electrolysis hydrogen production, factoring in compression/liquefaction for storage, are overall low, around 20-30%.

Realising fuel cycle opportunities will be key to the long-term sustainability of nuclear fission; in this context the definition of sustainability being the ability to maintain the NFC in the long-term (centuries or millennia) (Taylor et al., 2024). In addition to RCG proposals, increased interest in fission more generally has arisen given urgent imperatives to decarbonise in the face of rapidly accelerating climate change, with ~50 new reactors under construction at the time of writing and many more planned (IAEA PRIS). This will increase demands on finite natural uranium (and thorium) reserves (Peel, 2021) already threatened by recent geopolitical shifts, especially if fuel continues to be utilised only once in the open fuel cycle (Figure 2a) (Holdsworth and Ireland, 2024). Closing the fuel cycle by recycling spent nuclear fuel (SNF) to recover the fissile and fertile actinides (U and Pu, Figure 2b, or with addition of the minor actinides (MAs – Np, Am, Cm), Figure 2c) for further energy generation can greatly increase sustainability (IAEA, 2008). This sustainability improvement is achieved by reducing environmental demands of mining, as well as by decreasing waste volumes for disposal. Furthermore, careful selection of separation and recycling procedures can reduce radiotoxicity and duration of post disposal hazards markedly (IAEA, 2008, Holdsworth et al., 2025a). Uptake of SNF recycle is limited however due to high costs, proliferation concerns , and historical lack of commercial necessity in the form of low U prices and/or demand for MOX (mixed oxide) fuel (Holdsworth and Ireland, 2024).

More recent proposals for recovery of valuable fission products (FPs) such as the platinum group metals (Ru, Rh, Pd, (and Ag)), rare earth elements (Y, La-Lu), and noble gases (He, Kr, Xe) alongside useful isotopes from SNF during recycle would further increase nuclear fuel cycle (NFC) sustainability. These separations would further reduce waste volumes, whilst offering direct value recovery to offset the high costs of fuel recycle (Holdsworth et al., 2023; Holdsworth and Ireland, 2024; Bourg and Poinssot, 2017; Holdsworth et al., 2025a; Hodgson et al., 2023; Holdsworth et al., 2025b; Rohrmann, 1965; Kolarik and Renard, 2003; Jensen et al., 1984). These are in addition to further refinements to SNF recycle developed in recent decades (Baron et al., 2019). Around 300 000 USD of valuable FPs could be recovered from each ton of SNF, representing an untapped stream of valuable, naturally-scarce resources (Holdsworth et al., 2023; Holdsworth et al., 2025a; OECD, 2024).

Separation of the high-heat radionuclides (HHRs – ¹³⁷Cs, ¹³⁴Cs, and ⁹⁰Sr/⁹⁰Y as high-yielding fission products (Holdsworth et al., 2019)) has been proposed several times in the literature (McKibben, 1984; Venkatesan et al., 2009). This includes in service of advanced waste management (Forsberg, 2000a, 2000b, and 2000c), and more recently in some of the authors’ research for additional mitigation of operation challenges in SNF recycle (Bond et al., 2019; Holdsworth et al., 2019a, 2019b, and 2019c; Holdsworth et al., 2020; Holdsworth et al., 2021a, 2021b, 2021c and 2021d). This could reduce all other outputs of recycle operations from high level to intermediate level wastes (Holdsworth et al., 2025) at the expense of producing a “screaming hot” Cs-Sr wasteform. This waste stream would then either be stored pending decay or disposed of in a dedicated repository. If these materials could be utilised, however, their disposal would become a secondary consideration, and the radiotoxicity of NFC would outputs reduced to a fraction of present day values. The use of neutron activation products (e.g., ⁶⁰Co) could also be considered for waste valorisation; producing these intentionally in a reactor would otherwise negatively affect the neutron economy of the fuel cycle and should ideally be avoided (Rohrmann, 1965).

2. Putting “Waste” Radiation to Use

Although the generation of hydrogen by radiolysis of water has long been recognised as a problematic side-effect of radioactive materials (Bibler, 1998; Agyekum et al., 2025; IAEA, 1999; Vinson et al., 2002; Bottomley et al., 2011), research has proposed harnessing this for energy generation since the 1970s (Sauer et al., 1976; Yoshida et al., 2007; Agyekum et al., 2025; Vandenborre et al., 2024; Ali et al., 2024; Ali et al., 2025; Sawasaki et al., 2003, Blibler, 1998; Jafarov et al., 2024; Cecal et al., 2008), a concept we shall henceforth call radiolytic hydrogen production (RHP). The most interesting and likely practical of these concepts proposes using individual SNF assemblies stored in “silos” filled with water containing a TiO₂ photocatalyst to increase the efficiency of this process (Figure 3).

This would function by harnessing the intense gamma emissions arising from this material (Vandeborre, 2024), most of which are produced Cs isotopes, and is reported to function by harnessing photoelectric and Compton effects (Yoshida et al., 2007; Ali et al., 2024). Vandenborre et al. calculated that the current worldwide stockpile of SNF (390 000 t as of 2017) could produce at least 42.9 MtH₂/y (millions of tons of H₂ per year; 60% of world demand) via this approach, or 110 tH₂/tSNF.y (tons of hydrogen per ton of spent fuel per year) in what they reported to be a technologically simple system. We assume that these values are based on the bulk of the world SNF stockpile being both longer-cooled and relatively low burnup; SNF from modern, Gen III(+) reactors will be both higher burnup and shorter cooled, and thus potentially much more effective for H₂ production. We use the conservative values presented for discussions here until this can be more accurately modelled. The radiolysis of water itself is a complex process, producing a range of reactive species in addition to H₂ and H₂O₂, the minutiae of which, including studies with other photocatalysts, have been reported extensively elsewhere and are beyond the scope of this work (Agyekum et al., 2025; Ali et al., 2024; Yoshida et al., 2007; Ali et al., 2025; Yoshida et al., 2007; Sawasaki et al., 2003; Jafarov et al., 2024; Cecal et al., 2008).

In fuel cycle terms, we view this approach as a replacement for traditional SNF cooling ponds. Hydrogen produced in this manner would be essentially passive and “free” as natural radiolysis is exploited to harness energy that is otherwise wasted as heat. In contrast, to produce the same volume of H₂ from current light water reactors driving electrolysis would require around 164 large current-generation fission plants such as EPRs (European Pressurised water Reactors) devoting their entire electrical output to this operation. This assumes an electrical output of 1650 MWe and a conversion efficiency of 70%, taking 55 kWh to produce 1 kg of H₂, and represents around half of the installed fission generating capacity at the time of writing.

The H₂ produced could be used to generation zero-carbon energy (electricity generation or heating) or other industrial uses. However, this approach would essentially “lock out” the actinides present in such SNF from fuel cycle use and thus be less sustainable than a closed fuel cycle (IAEA, 2008), though this would be beneficial from criticality and security standpoints if individual SNF bundles are used in isolated silos. This is dependent on the availability of fissile and fertile materials from existing stockpiles and the primary economic drivers to reuse U and Pu as MOX, and/or the MAs for waste management purposes. Preliminary results indicate that single assemblies of even fresh fuel in such a silo (representing a severe maloperation) would be sufficiently subcritical (k_eff = 0.832) to be satisfactorily safe; this can be reduced further by the addition of neutron absorbers to the system (See Supplementary Information). Other applications for radiation such as conversion of bio-feedstocks (Plant et al., 2023), polymer grafting and cross-linking (Naikwadi et al., 2022) medical device sterilisation (McEvoy et al., 2023), and water purification (Sanchez-Polo et al., 2009) could also be harnessed.

While it has been demonstrated that gamma radiation, produced by ¹³⁷Cs, ¹³⁴Cs, and the like, can effectively produce H₂ from water at significantly enhanced yields (relative to pure water alone) in the presence of a photocatalyst such as TiO₂, the same effects has yet to be verified for beta radiation, although the effects should be similar to gamma (Leotlela, 2024; Spinks and Woods, 1990). The most prominent source of this in SNF is ⁹⁰Sr, which could further increase the theoretical H₂ yields if effective with the same photocatalyst as demonstrated for gamma radiation. As TiO₂ likely enhances H₂ production yields in water by more conversion of gamma rays into energetic electrons than the photoelectric and Compton effects, this would seem likely (Ni et al., 2007). These, amongst many other optimisations of the system, remain to be explored, as was noted by the authors of that study (Vadenborre et al., 2024) and other previous proposals (Sawasaki, et al., 2003). As alpha emitters deposit the majority of their energy within a short range of the source, these are not considered for hydrogen production here.

3. Recovering and Using the HHRs for H₂ Production

A combination of our own previous work in HHR separations and the H₂ production proposed by Vandenborre et al., could provide a revolutionary approach to the NFC with numerous, unrealised benefits. In contrast to the open fuel cycle (Figure 4a), closed fuel cycle (Figure 4b), and Vandenborre’s proposal (Figure 4c), we would combine the closed fuel cycle with two stages of hydrogen production (Figure 4d). Firstly, SNF would be cooled on-site at the reactor until it can be safely shipped to a cooling pond. At this time, we would instead transfer individual SNF bundles to individual silos, using of their intense radiation field to produce large quantities of H₂.

Once the SNF bundles have cooled sufficiently to permit it (typically 5-10 years depending upon burnup) they would be recycled. This would enable the recovery the U, Pu, and MAs for further energy generation, ideally alongside valuable FPs, the HHRs, and other useful isotopes. The HHRs would then be processed and concentrated into suitable sources for long-term H₂ production and eventual disposal, returned to the silos or similar apparatus, and retained in-place until radiolysis is no longer effective. The HHRs could be further recycled or “refreshed” midlife to increase activity.

The rationale behind this approach is that most of the radiolysis will arise from only a few kg of Cs and perhaps Sr present within each tonne of SNF (Forsberg, 2000a, 2000b, and 2000c). By separating these sources from SNF during recycle operations, the actinides and other valuable FP materials are also released for further energy generation and use respectively. The remaining HHRs would then be converted from a hazardous waste into potentially one of the most valuable components of SNF and no longer require separate disposal. As demonstrated by our previous work, separating the HHRs from SNF during recycle results in all remaining raffinates being reduced to intermediate level wastes rather than high level (Holdsworth et al., 2025a); this introduces new options in waste disposal and reduces, if not entirely eliminates, the necessity for a geological disposal facility. An intermediate approach, whereby U and the HHRs (and possibly valuable FPs) are recovered from SNF in a pseudo-recycle process to reduce waste volumes could provide a proliferation-resistant fuel cycle approach for nations where such prohibitions exist. It should also be noted that separations for Cs, Sr, and other HHRs would not be isotopically-selective, and would also co-extract sTable Snd long-lived species (e.g., ¹³³Cs and ¹³⁵Cs) alongside the desired radioactive species.

4. Market Potential and Environmental Implications

The market potential for RHP is significant. Based on the conservative values utilised by Vandenborre et al., (Vandenborre et al., 2024) and the potential system optimisations that could be realised, each tonne of SNF, or the HHRs present therein, would produce 110 tH₂ annually, decreasing slightly each year due to radioactive decay. This corresponds to an annual return of 220,000 USD per t_SNF at a base price (for H₂ produced by fossil fuels, gas price valuation after 2008 crisis recovery) of 2 USD/kg of H₂ (Curson and Hassan, 2025), though prices are often variable and in the EU are often higher than this (Li et al., 2024). For the world SNF stockpile (390 kt), this corresponds to a potential income of 85.8b USD per year generating 42.9 Mt per year, representing ~60% of world demand in 2024 (Vandenborre, 2024).

While the fuel cycle and other processing costs of RHP are not currently well understood, the inherently low operating costs arising from the ability to generate large volumes of H₂ in a fuel-less, passive system make this concept highly desirable compared to incumbent approaches such as methane steam reforming, and may even compete with conventional electrolysis. As additional SNF is produced every year with current rates and recycle capacity is limited, the production potential of RHP increases cumulatively. This continued fuel production has higher burnup than legacy fuels and thus more radioactive, and therefore potentially capable of greater H₂ production.

There are several critical use-cases for low-carbon hydrogen: first and foremost, energy storage for later use, either in compressed or liquefied form, and secondly for industrial use for production of important feedstocks, such as ammonia. Direct energy generation could be achieved via fuel cells or combustion. Despite the promise of fuel cell technology, their application to grid-scale energy generation for load following is currently insufficient relative to other technologies due to longer start-up times, challenges in load following, high costs, and the efficiency losses incurred with conversion from DC to AC. In contrast, combined-cycle gas turbines (CCGTs), which normally run on natural gas, can be readily converted to partially or completely run on hydrogen and directly output 50 or 60 Hz AC at grid voltages with thermal efficiencies over 60% (Boretti, 2025).

The approximate energy contained within a kg of H₂ is 33.3 kWh which, if continuously turned into electrical energy in a CCGT (at 60% efficiency) at a grid price of 0.12 USD/kWh, would afford 264k USD in energy generation per tonne of SNF annually, not accounting for energy consumption in H₂ processing, storage, transfer, overheads, and so on. This would offset the equivalent of 670 tons of CO₂ per year per tonne of SNF from burning natural gas to generate the same volume of energy, assuming CO₂ production of 0.183 kgCO₂/kWh of natural gas. One ton of light water reactor fuel is estimated to have a total carbon footprint (including all upstream activities such as mining, enrichment, etc.) of 1.58 tCO₂-equivalent (Ecoinvent, 2025). Therefore, this pathway effectively offsets 42% of the fuel’s carbon footprint annually. Individually, these numbers may seem relatively small, but given the potential scalability of this system, a large impact could be reached relatively quickly, especially if integrated alongside other H₂ producing and consuming (i.e., energy or industrial) infrastructure. These values assume continuous operation without significant H₂ storage; use for load following would have a lower capacity factor and higher costs for storage.

Moreover, RHP might offer a route to mitigate some of the increasingly clear environmental challenges associated with other hydrogen production routes. Environmental life cycle assessments of hydrogen production via both electrolysis and steam methane reforming (with and without carbon capture) have suggested that, while hydrogen’s carbon footprint is lower than that of fossil fuels such as natural gas, worsening of other environmental impacts is likely: one study found that hydrogen boilers had higher environmental impacts than gas boilers for 18 out of 19 environmental impact categories, with climate change being the exception (Slorach & Stamford, 2021). Meanwhile, another study analysed scenarios for the production of 500 Mt H₂/y, finding a high likelihood of transgressing 4 out of 9 ‘safe operating spaces’ under the Planetary Boundaries framework, as well as high overall impacts on ecotoxicity and metal depletion (Weidner et al., 2023). While the life cycle implications of RHP are currently unknown, its seemingly low material and energy inputs might enable hydrogen production that outcompetes other options on environmental grounds, mitigating the above concerns.

5. Technical and Regulatory Considerations

The necessary development and regulatory challenges for this concept to be implemented in full are manyfold. As noted in their work, Vandenborre et al. state that RHP with SNF bundles could be implemented with little technological development, though further optimisations are required (Vandenborre et al., 2024) but the second stage we have outlined requires significant further study, advancement, and technological fusion.

Ideally, RHP would be co-sited with SNF recycle operations, augmenting or replacing current cooling ponds as H₂ production from SNF assemblies would essentially fulfil the same role. Once SNF bundles have been sufficiently cooled to allow for recycle (10 years for high-burnup PWR fuel), those operations would be performed and the separated HHRs converted into targets for further RHP. From a security standpoint, this second step would best be performed co-sited with SNF recycle and generation on site. However, given the nuclear industry’s extensive expertise in the transport of active sources (especially SNF), it may be possible for RHP to occur on suitably licensed non-nuclear sites; given that such HHR targets would be neither fissile nor fertile. This would mitigate the challenges of compressing and/or liquefying and transporting or piping H₂, but compressed storage for later energy generation or industrial use would be necessary anyway, and could provide the means to achieve localised H₂ production “off-grid” with minimal supporting infrastructure.

A further approach would be integration into clean energy hubs where fission reactors and hydrogen production are co-sited, for example where excess generation capacity is used to electrolyse water, the facilities for handling H₂ would already be present and could accommodate RHP without duplication of resources, thus lowering costs. This would also negate the need to ship SNF assemblies to a recycle facility prematurely. In this regard, RHP complements, rather than a replaces, reactor H₂ production, and could serve to simplify the safety case and economics of such operations. .

The separation of Cs in hydrometallurgical SNF recycle has seen significant development by some of the authors, and can be assessed, based on a modified TRL scale (Baron et al., 2019), to be at TRL 3-4. The Sr separations in this context, and the processing of both Cs and Sr into inert matrices suitable for long-term RHP (and disposal) are all at TRL 1. Several options for these matrices are discussed in the supplementary information, as are the various isotopes present in SNF for H₂ production from both beta and gamma-emitting nuclides (Leotlela, 2024). The only other isotope in addition to ¹³⁴Cs, ¹³⁷Cs, and ⁹⁰Sr that may appreciably contribute to HHR RHP is ¹⁵⁴Eu (see Supplementary Information), which could be recovered from REE-rich feeds from MA separations (Holdsworth et al., 2025b), if this is economical and practical. The level of “dilution” to which the HHRs would need to be incorporated in an inert matrix for RHP must also be determined. Industry-standard H₂ purification and handling technologies (TRL 9) are employed worldwide and would not need significant further development for system integration.

Vandenborre et al. demonstrated that gamma emissions effectively produce H₂ with a TiO₂ photocatalyst (Vandenborre et al., 2024), but these results remain to be verified experimentally for beta-emitting isotopes (Leotlela, 2024), though the yields should be similar (Spinks and Wood, 1990). If the radiolytic mechanism by which beta-emitters (⁹⁰Sr/⁹⁰Y) produce hydrogen in water is different from gamma, then a different (photo)catalyst may be required, and would logically dictate that Cs and Sr are used separately. Alternatively, ⁹⁰Sr could (as outlined in the supplementary) be used as a heat source for low-intensity but high-temperature applications, if not effective in H₂ production, as an isotope with negligible gamma emission (each tonne of high-burnup PWR SNF contains about 700 Wth of ⁹⁰Sr (Ando and Takano, 1999, Shor et al., 1971; Uherka, 1987; Rohrmann, 1963; Wood, 1971; Hoisington, 1982)), and as a source of ⁹⁰Y for medical applications (Chakravarty, 2012). A portion of separated ¹³⁷Cs could similarly be recovered for other industrial processes requiring radiation sources, such as conversion of bio feedstocks for other co-generation purposes (Plant et al., 2023) and polymer grafting and cross-linking (Naikwadi et al., 2022), medical device sterilisation (McEvoy et al., 2023), and water purification (Sanchez-Polo et al., 2009). Further optimisations in H₂ yields can likely be found in “silo” geometry, photocatalyst loadings and doping, flow rates, water chemistry, and so on. Given the sensitivities around the use of nuclear-derived materials, valuable FPs recovered from SNF, such as Pd and Ag, could be used in remote operations to purify hydrogen produced using this concept (Bourg and Poinssot, 2017; Holdsworth et al., 2023; Holdsworth et al., 2025). At all stages, the safety implications of producing hydrogen, as a highly explosive gas, especially in contact with oxygen or air, must be considered; these have been extensively explored in a RCG context elsewhere, but will require specific modifications for the context outlined here in future works (Verfondern et al., 2017), especially given the presence of high radiation fields (Agyekum et al., 2025; IAEA, 1999).

Current regulations in industrial setting require official certification of hydrogen related technologies and the financial support from the UK Government (under the Energy Act 2023 law) was restricted to a set of projects on hydrogen transport and storage. A wider approach/consideration for hydrogen production has been taken since 2023, encouraging more technologies to be used but only if they meet the Low Carbon Hydrogen Standard (CMS, 2024). A lack of regulations regarding the interface of nuclear (both fission and fusion) and hydrogen is apparent.

Safety considerations around the handling of hydrogen in the presence of radioactive materials must be a foremost concern. Radiolytic production of H₂ occurs naturally in spent fuel pools and within water-cooled reactors and are actively suppressed within existing fuel cycle operations. The explosion risks would be mitigated through proper process design, ensuring that H₂ is never mixed with sufficient oxygen to present a risk, alongside the omission of ignition sources and comprehensive monitoring. This likely means that new regulations will be needed to implement RHP and as such, early engagement with national and international regulators must be sought to remove roadblocks on the path to implementation.

6. Discussion, Conclusions, and Further Work

Building upon both our own research and that of peers, we have outlined a novel approach to both the NFC and sustainable H₂ production using RHP. This has the potential to drive substantial paradigm shift in the sustainability of the NFC and decarbonisation of both energy grids and, more generally, industrial hydrogen production. This could facilitate modern, sustainable, fully-closed fuel cycles for recovery of valuable resources and, most importantly, utilising materials previously considered as wastes (SNF and more importantly the HHRs) as an incredibly valuable resource.

Our approach would offer further NFC benefits, including mitigating a large proportion of the highly active wastes from SNF recycle, and this minimising or potentially even eliminating the need for costly geological disposal (Holdsworth et al., 2025), in addition to offsetting a large proportion, if not the entire costs of SNF recycle assuming the full range of valuable FPs are recovered (Bourg and Poinssot, 2017; Holdsworth et al., 2023; Hodgson et al., 2023; Holdsworth et al., 2025b). This would hopefully eliminate the barriers to SNF recycle and facilitate a wider closing of the fuel cycle, which will be essential to the long-term sustainability of nuclear fission (Holdsworth and Ireland, 2024). The ability to decouple hydrogen production at least partially and if not entirely from reactor operations could significantly simplify the safety and economic cases of nuclear hydrogen energy storage and production, with the potential for decentralised operations if the security and safety challenges can be addressed.

Significant development is needed to realise our concept in its entirety, despite the enormous potential to the NFC and Net Zero transition more generally. The first stage, where H₂ is produced using SNF assemblies, as proposed by Vandenborre, is achievable using current technology with little further development, albeit with substantial regulatory and licensing barriers to be overcome, and more comprehensive modelling required to better understand the production potential and safety aspects. We invite collaboration to achieve these aims.

The second stage requires comprehensive technological development of HHR separations, recovery, and processing to usable forms, optimisation of hydrogen production with respect to Cs and Sr partitioning (i.e., beta/gamma effectiveness), the potential inclusion of other nuclides, disposal route at the end of life, and so on, though these should all be achievable with present-day technology. A comprehensive life-cycle assessment of this potential approach to hydrogen generation ideally should be undertaken soon, accounting for environmental, socio-economic, and regulatory factors, to inform the following research and best approach to implementation. New regulatory and licensing frameworks may be needed to achieve what has been outlined here alongside a step-change in the attitudes of the nuclear industry, whose conservatism has often held back emerging technologies from having meaningful impact (Holdsworth and Ireland, 2024). These will be assessed in future works.

On a final note, as radiolytically-produced H₂ is distinct from the other “colours” of hydrogen (Figure 5) – red, pink, and purple normally being assigned to nuclear-derived hydrogen, from electrolysis, (catalytic) thermolysis, and combinations of these respectively (Arcos and Santos, 2023; Ahmaruzzaman, et al., 2025) – we propose that this method be designated as PLATINUM HYDROGEN. This is to distinguish this method versus the other proposed approaches, all of which require some external or natural (underground) chemical and/or energy feedstock in order to function (Osselin, et al., 2025).