There is already a market for hydrogen. In its 2023 Global Hydrogen Review the International Energy Agency (IEA) said global production was almost 95 Mt in 2022, 3% higher than in the previous year. Of that, 41 Mt was produced and used on-site in refineries. Around 32Mt was used for ammonia production, largely for the fertiliser industry. Nearly 16Mt was used for methanol production and the remainder, about 5Mt, in the iron and steel sector.
The hydrogen was almost entirely produced using unabated fossil fuels. IEA listed the main sources as natural gas (62%), coal production largely in China (21%) and as a by-product of processes in the petrochemical sector (16%).
That existing hydrogen industry must be decarbonised, while in addition low-carbon hydrogen must also be available for new roles. A key use for hydrogen is as energy storage, to help manage periods when electricity generation from renewables is low. In northern Europe this is seen as a key use of bulk hydrogen, fuelling gas turbines to meet power needs during extended winter ‘lulls’ with no wind. Other global regions also exhibit medium term renewable generation and seasonal fluctuations that can be smoothed out using hydrogen for power.
But there are other potential markets in production of hydrogen-based fuels (such as synthetic hydrocarbons), biofuels, high-temperature heating in industry, transport and more. For some of these industries, such as steel processing, baseload supplies are needed for commercial reasons, to ensure high load factors and keep breakdown risks low.
The Hydrogen Council, a global grouping of 141 companies with interests in the sector, believes demand for hydrogen by 2050 will reach 660Mt per year.
Currently, there are two main ways to produce ‘low emission hydrogen’ that are being scaled up. (A third, thermochemical water splitting, using high temperatures and chemical reactions to produce hydrogen and oxygen from water, has opportunities to use heat from nuclear but is less of a development focus).
Of the two options being scaled up, one is based on the most common existing process – steam methane reforming – but with the addition of carbon capture and storage; the other uses low carbon electricity to electrolyse water into hydrogen and oxygen. Is this where nuclear can make its mark? Does it have a business case that can compete with the other options?
Chris Harris, a long-standing member of the UK energy utility sector who is now at the University of Bath, says the case for hydrogen is made by its use in industrial process, in transport (direct or as vector) and as storage, replacing methane. He has examined the consequences of some hydrogen production models and the capital and operating regimes, and considered whether and how nuclear can become a major player in the industry.
Steam methane reforming
As noted above, the current method of hydrogen production is steam reforming of methane. This is an energy intensive process – around a third of the gas used is consumed in providing heat to the process – and one that produces carbon dioxide. Additional processes have to be incorporated that capture the carbon dioxide emitted (generally by absorption into a chemical substrate) and transport it (by pipe or other means) to a permanent storage site (such as injecting into closed offshore oil and gas reservoirs). This option is attractive to policymakers because of the effect of clustering: industry has tended to locate along estuaries where it has access to water and transport options.
How does nuclear compare as an option for producing hydrogen? Nuclear and fossil carbon capture and storage (CCS) generation share economic characteristics that drive them to baseload operation, such as the need to recover capital costs by high load factor operation. Both are large-scale centralised facilities, and are likely to have a long lead time before permits are in place and construction – which in itself may take two decades – can begin.
The process of methane reforming with CCS may have an edge on nuclear. The first reason is that steam methane reforming is already in operation and it may be possible – if economic – to retrofit CCS to existing plants. The reforming technology is proven.
The second reason is that the methane/CCS option does not require an electricity network connection. It is possible that a nuclear plant could be entirely focused on providing hydrogen – but it will not have industrial customers for the hydrogen until it is up and running. Industry has to be local to the plant to use the hydrogen, but will not invest, or sign contracts for hydrogen supply ahead of plant operation (although as the plant completes construction they may be ready for ‘in principle’ agreements). That contrasts with nuclear’s history in the electricity sector: investors in nuclear, as in other power generating plant, know it will have permanent and open access to a customer base via the electricity network.
Until a nuclear plant intended to provide hydrogen has secured customers, it will require the option of electricity supply and it will have to invest in the physical and contractual arrangements that allow it to do so – as well as the infrastructure required to supply hydrogen to nearby customers.
Once the plant is in operation, electricity contracts can be unwound or lapse as new contracts to supply hydrogen go into effect.
The ‘back end’ of the various cycles also have to be considered in comparing the methane/CCS and nuclear options. The core question is how much the cost of disposal is taken into account (carbon dioxide for fossil and spent fuel and decommissioning for nuclear).
Current CCS technologies are limited because they do not capture all the carbon produced. That means CCS at 90% capture is a not a long term solution, as there are not enough sequestration resources to deal with the residual 10%. Potential technologies such as oxyfuel are more efficient at capturing carbon dioxide – up to around 95% capture may be achievable long term – and may be a long term solution but it still requires a positive benefit to outweigh the cost of managing residual carbon dioxide emissions.
Harris suggests the best option is to run nuclear flat out and spill non-used hydrogen into storage caverns. The storage supplies flexibility as well as energy security. They are highly valuable to the electricity system and to industrial users, respectively. But this involves the cost of storing hydrogen. Ideally, the nuclear plant and cavern will be very close to each other to minimise pipework. This does add another siting requirement to the hydrogen-fuelled industrial cluster though.
Alternatively there are numerous non-cavern ways to store hydrogen. That may involve low temperatures (below 20K) or high pressure. There are also chemical options, such as hydrogen adsorbed on a substrate possibly combined with low temperatures, which include metal organic, interstitial or chemical hydride, liquid organic or complex hydride and metal hydride options. But they are likely to be used in niche applications.
‘Free’ hydrogen from renewables?
The most frequently cited hydrogen model is using excess green energy from wind and solar sources, which are expected to be built out at massive scale in the next decades. Huge renewables capacity is required to meet increasing demand as industries electrify, but it will supply its energy in surges, as weather conditions are right, that have to be captured, stored and provided to customers as required. It has led to a simple narrative about hydrogen production: use the excess to produce hydrogen by electrolysis of water. That seems to suggest hydrogen production might be cheap or even free, as the excess renewable energy is currently discarded, or pays a negative price to export to the grid and this appears to be far more attractive than investing in a nuclear plant. In fact, there are hidden costs.
Renewables, especially wind, are variable generators, but for off-takers, steady and predictable power is more valuable (and realises a better price for the generators). Typically a wind farm operator will sell a proportion of its generation as baseload on a long term contract (possibly to an electrolyser operator).
At the moment electrolysers tend to be of the alkaline variety, and are not flexible in operation. Running hydrogen plant at full capacity requires it to be under-sized relative to an associated wind farm, as wind farm peak hours will be infrequent. If the baseload feed is guaranteed, the generator will have to source low-carbon power from another supplier to ‘fill in’ at times when the wind plant is not generating – and at times when one wind farm is in low-wind still weather many others are also stationary, so low-carbon grid power is at a price premium.
Meanwhile, the variable component will be sold as ‘merchant’ power, but because the ‘baseload’ portion of generation has been sold on a long term contract the merchant sale is now ‘very variable’ and typically less valuable.
This issue has to be managed by using flexibility that exists in the wider system and this is why electric vehicles are so important, because they offer the scale of flexibility required for the ‘very variable’ wind fraction. Electric home heating may also be important, but either way, the flex has to be accessed from downstream.
What changes?
Nuclear’s long lead time means much can change while it is being deployed – even with the development of small modular reactors – there is little opportunity for rolling out multiple projects to reduce costs via ‘learning by doing’. Electrolysis, in contrast, is likely to develop more like renewables or batteries: by replication of hundreds and thousands of units that can also begin to provide a partial return to investors at an early stage in the project.
It also means electrolysis technologies are likely to undergo significant development and innovation while nuclear is being developed and licensed – especially in regards to flexibility and efficiency, which are seen as major targets. That raises the possibility that the case for nuclear hydrogen will be ‘hollowed out’. Another competitor may be hydrogen imports.
Harris says, “a key question about nuclear is how it competes with wind on a £/MWh basis if we will electrolyse”. If it cannot reduce its costs nuclear must bring something new, either to support wind through low wind periods or find a role in hot electrolysis”. Nevertheless, he sees value in the nuclear option for hydrogen. That is partly because of the opportunity for bulk production, but there are other important advantages: “There is an advantage in having baseload power to reduce the double flex need of green hydrogen. [In Northern Europe] winter heat load is a key driver and hydrogen with caverns can do this”.
He says storing hydrogen in caverns “seems to come out on top”.
What is the strategy?
This exploration suggests a long-term strategy for nuclear to secure a role in the hydrogen industry by aiming for a twin track solution.
Electrolysers benefit most in a flexible system. Wind powers baseload electrolysers, with the power spill going to EVs or other flexible users. In addition, for the system as a whole there is a security benefit from including electrolysers. That is because even if they are not generally variable in operation, there is an option to shut them down on occasions, to free up low wind or other lack of supply for other users.
Meanwhile nuclear’s strength is be to provide consistent power and hydrogen in bulk.
It can have a strong synergy with wind in situations where demand does not achieve the extreme flexibility that is most desirable. In this scenario the provision of baseload power from nuclear capacity to meet baseload hydrogen demand reduces the strain on demand flexibility. The two hydrogen sources can combine to power industrial processes plus some other uses and spill any excess hydrogen to caverns. The caverns absorb the imbalance between hydrogen production and consumption. The nuclear has reduced the double flex requirement of the wind-hydrogen system.
To maximise its attractiveness, nuclear should focus on siting in industrial regions that also have access to bulk hydrogen storage.
In this future, steam methane reforming with carbon capture and storage may be a short-term solution for industrial customers and initially nuclear competes head on with high capture fossil CCS. Which wins is unknown, but both are extremely expensive, and indicate more efforts on the demand side are needed. But nuclear does not require transport and disposal of carbon dioxide – which may in any case be under used in a future where electrolysis from renewables is under way at large scale.
This in turn has implications for the infrastructure investment in these areas. In the event of a switch to nuclear hydrogen production, hydrogen transport options remain important and so does bulk hydrogen storage, but carbon dioxide infrastructure becomes obsolete.
Policymakers should have this in mind when considering business models for carbon dioxide infrastructure. It should focus on whether flexible options such as tankering would be a cheaper option if nuclear has potential to take over bulk supply of hydrogen for industrial customers. In this event, fixed pipelines for carbon dioxide transport quickly become stranded assets and so do connections to carbon dioxide sequestration.
