Hydrogen made by electrolysis is only low carbon if made when there is surplus low carbon generation on the system. This will likely become much more common in future. Deployment of electrolysers in the meantime should focus on developing capacity for future deployment, for example by developing electrolysers capable of running flexibly.
The UK government recently consulted on defining a standard for low carbon hydrogen[i]. Similar discussions are taking place elsewhere, including in the EU. However these discussions have generally failed to deal adequately with the problems associated with diverting scarce renewable electricity to make hydrogen.
Using renewable electricity to displace fossil generation achieves much greater emissions savings than making hydrogen by electrolysis, and is much cheaper. Using renewables to displace gas in power generation leads to around two and a half too three times the emissions savings from making hydrogen to displace natural gas in industry, either in boilers or direct combustion (see my previous post for more comparisons of this type). This because of:
- The energy losses in making hydrogen.
- The lower efficiency with which gas is used in power generation (and thus the greater gains from displacing it).
This is shown in simplified form in the chart below.
Charts 1: Comparison of emissions avoided by displacing electricity from natural gas in CCGT and in heat for industry
Abatement using hydrogen to reduce emissions is also much more expensive per tonne of emissions reduction. This is both because of the cost of making the larger amount of renewable energy required per tonne of savings, and the cost of the electrolysis. These comparisons will not change fundamentally as technology improves. For example, electrolysis will always be an additional cost.
This implies that diverting renewable electricity to make hydrogen increases both emissions and costs.
In practice, renewables will be diverted as long as there is fossil generation on the system. This is almost always the case at present. Low carbon electricity (nuclear and renewables) is low marginal cost, so will almost always run at full capacity ahead of higher marginal cost gas plant. Consequently, when an electrolyser is switched on, the additional demand leads to additional natural gas generation, because there is no unused low carbon generation available to run. Correspondingly, when an electrolyser is switched off natural gas generation is reduced. The emissions from the electricity used in hydrogen production are therefore those from a natural gas plant.
The relevant consideration is thus the marginal plant on the system – the additional plant that runs when electrolyser is switched on. One implication is that reducing grid average carbon intensity at any point in time does not change the emissions from hydrogen production while natural gas plant is at the margin. For example, whether the system is running at 50% or 75% low carbon power, additional emissions will still be caused by switching on electrolysers. This is illustrated in the chart, which shows additional demand from electrolysers must be met by additional high carbon power, when there is no surplus of renewables.
Chart: Running electrolyses increases generation from high carbon power
Calling renewables for hydrogen additional, signing PPAs, looking to guaranties of origin, and similar measures do not change this physical reality. Renewables are still diverted from displacing fossil generation. For example, it is sometimes specified that renewables for hydrogen production must be “additional” for hydrogen to be low carbon. However “additional” renewables for hydrogen manufacture still suffer the same problems, because even if labelled “additional” renewables still diverted from other potential uses. The electrolyser can be switched off and the renewables used to displace fossil generation on the grid.
Even when there is a grid constraint or lack of grid connection it will usually be much cheaper, and yield much greater emissions savings, to rectify this than to make hydrogen. These circumstances will be rare in any case. Most new renewable electricty in the UK will be offshore wind, so will have good connections to the onshore grid. Furthermore, electrolysers are unlikely to be situated far from the, grid as they will need to be close to centres of hydrogen demand, for example powering fuel cell buses or industry. Indeed creating standards based on lack of grid connection risks creating perverse incentives to disconnect from the grid.
Provision of a PPA does not materially change the situation. The PPA is financed from the support contract with funds from the government. Financial support for hydrogen could instead be used to build extra offshore wind capacity directed to decarbonising the grid. (This would not apply in the same way if hydrogen production were unsubsidised, and in that case a genuinely additional supply of renewables may be created.) And funds freed up from avoiding the other costs of producing hydrogen can be used to fund yet more renewable capacity, further increasing the benefits.
There are several ways in which additional renewables could be sold into wholesale markets. For example:
- Funds could be spent on developing new renewables to serve wholesale markets rather than electrolysers – recognising the greater value this has for abatement.
- Selling power from the electrolyser PPA into the grid. It may be possible to prevent this by terms in the contract, but this would not be justified on emissions savings or cost grounds.
When there is surplus low carbon electricity, for example because there is a large amount of generation from wind when demand is low, there will be little no natural gas generation on the system. In this case switching of electrolysers will absorb this surplus without leading to additional natural gas generation. The hydrogen is then genuinely being made using low carbon power. This situation is currently rare but expected to be much more common over the next decade or so.
The reality is that making hydrogen from electricity is only low carbon when there are surplus renewables on the system. At all other times there will be additional natural gas-fuelled generation as a result of the hydrogen being made. This point was made in a recent report by respected consultants Element Energy:
“To be truly using renewable electricity, the electrolysers must not be diverting existing renewable electricity production from other sources of demand. Electrolysis performed using curtailed renewable generation is zero carbon.”[ii]
A low carbon standard should recognise this.
Furthermore, policy choices should be made in the light of this reality. They should focus on enabling deployment at scale in the 2030s (and perhaps late 2020s) when there will likely be significant amounts of surplus renewable electricity. Any investment in electrolysis in the meantime should recognise that electrolysers currently increase emissions, and are justified only as part of a pathway to future use. For example, there will likely be a large scale demand for electrolysers that can run flexibly enough to take advantage of periods of surplus renewable electricity.
The priority for both reducing costs and increasing emissions savings must always be to use renewables to displace fossil fuels in power generation. Renewable electricity should not be diverted to making hydrogen. The only exception to this now should be deployment of electrolysis to develop technologies and infrastructure necessary to enable future deployment at scale. This recognises that the hydrogen produced is high carbon, but regards this as a necessary investment to serve future needs.
Adam Whitmore – 27th October 2021