As part of the Corporate Europe Observatory (CEO) and The Transnational Institute (TNI) report on Europe’s efforts to turn Northern African into a hydrogen manufacturing center for Europe, I dove into the economic seesaw problem of green hydrogen. Many people are claiming that it will be cheap, but the evidence isn’t strong for that being true — quite the opposite in fact. Quite reasonably, the CEO folks edited my nerdy explanation down to make it more palatable to their audiences, but CleanTechnica readers are made of sterner stuff, so here’s the original version.
The standard process to make green hydrogen is electrolysis. Put a couple of electrodes in water, add electricity, and water separates into oxygen and hydrogen. That’s something that can be done in a basic high school science lab.
But to capture the hydrogen, protect the electrodes, and have a useful volume of hydrogen produced requires many more components. IRENA has a useful schematic of the basic components that include and surround a proton exchange membrane (PEM) electrolysis unit in a report they did on scaling green hydrogen.
The takeaway here is that there are many additional components surrounding the actual electrolyzer, and they are very standard technology that’s already achieved high optimization due to large scale manufacturing and deployment. As Paul Martin of the Hydrogen Science Coalition Points out, the electrolyzer might decrease in cost, but the surrounding components are much less likely to decrease in cost. Capital cost of the hydrogen manufacturing plant is a significant part of the total cost of hydrogen, and reductions in the cost of the electrolyzer won’t reduce overall capital costs as much as many assume. They are still subject to good plant design and vertical scaling efficiencies, so overall CAPEX will decline, but that’s incremental, not Wright’s Law levels of improvements.
The next point comes from Lazard’s 2021 levelized cost of hydrogen analysis.
This is a subset of their overall analysis, which also includes the pyrolysis alternative for hydrogen production and capex sensitivity for interested parties. But this table provides the key point. Electricity has to be dirt cheap and the system has to run 24/7/365 for hydrogen to be anywhere near the current cost of black hydrogen from natural gas or coal, which cost under $1.00 to manufacture before gas prices spiked, as long as all emissions can be vented to the atmosphere.
When we consider renewable solutions, in other words, it isn’t sufficient to use renewableably generated electricity just when it’s cheap, as the utilization costs will rapidly increase. And it can’t be run off grid retail electricity prices or those will make it expensive.
Areas with good onshore wind energy resources which are inexpensive to build see capacity factors — the percentage of potential generation for a year that is actually generated — in the range of 37% across the industry and wholesale cost of electricity prior to transmission in the $20 per MWh range. Offshore is better, seeing 45% to 50% regularly, but at costs above $50 per MWh.
Solar plants in good solar resource areas see capacity factors of 25% over a year, wholesale costs exclusive of transmission in the $30 per MWh range, and that’s expected to level off with wind energy by around 2030.
The combination of wind and solar does not add up to 62% capacity factors as there will be days when the wind and sunshine are strong at the same time. Building a mix of 100 MW of wind and solar, another significant capital cost might deliver 100 MW of electricity 25-30% of the time, as while wind energy generates electricity roughly 85% of the time, it’s often below maximum. Similarly, solar generates electricity for potentially a dozen hours a day near the equator, but it’s weaker in the morning and evening. Doubling capacity of wind and solar increases capacity of course, but wind and solar farms in a limited geographical region are intermittent by their very nature. Assuming you get to 50% capacity factor with $20-$30 per MWh electricity, you are still far off the right hand side of the table in terms of utilization. Adding storage specifically to firm the energy increases capital costs further. Running it only when there is excess and it’s very cheap has the same problem, as the conditions under which renewables will have excess that isn’t being used to displace electricity generated by fossil fuels will be limited until late in the deployment effort.
If you balance the wind and solar generation with transmission from further away to limit intermittency challenges and maximize electrolyzer utilization, you add transmission costs to the base wholesale cost, and likely pay retail rates well over $50 per MWh, dropping off the bottom of the table into steadily increasing prices. Further, you are buying grid electricity, not renewable electricity, which means that the CO2e footprint of the hydrogen goes up, usually significantly.
It’s only very late in the decarbonization of grids, when large-scale transmission smooths intermittencies and renewables across continent-scale areas to balance one another, and allows sufficient grid storage is built, that electricity costs will stabilize at a new low normal due to the advantage of zero fuel costs. That’s decades away, and I project the average price of electricity in 2100 to be in the range of $20 per MWh in 2020 dollars.
It’s more reasonable to assume that the bottom right corner of Lazard’s table is probably the best case scenario for hydrogen manufacturing costs. And as Lazard points out:
“Other factors would also have a potentially significant effect on the results contained herein, but have not been examined in the scope of this analysis. These additional factors, among others, could include: development costs of the electrolyzer and associated renewable energy generation facility; conversion, storage or transportation costs of the hydrogen once produced; additional costs to produce alternate fuels (eg, ammonia); costs to upgrade existing infrastructure to facilitate the transportation of hydrogen (eg, natural gas distribution pipelines); electrical grid upgrades; costs associated with modifying end-use infrastructure/equipment to use hydrogen as a fuel source;”
Yes, the high-cost hydrogen that gets produced has a lot of other costs which must still be added on. In more extracts from the report, I’ll deal with specific projects, shipping of hydrogen, piping hydrogen and transforming it to ammonia.
Manufacturing hydrogen is relatively energy inefficient, and using it as a source of energy is inefficient as well. Best in class PEM electrolysis is 70% efficient at turning electricity into an equivalent heat energy of hydrogen using the lower heat value (LHV) efficiency that determines usable energy and excludes the vaporization of excess water. The absolute minimum of energy required for storage, just compression, represents another 10% efficiency loss. Turning hydrogen back into electricity is at best 60% LHV efficient in modern fuel cells. The combination of those three things, if green hydrogen were used as a storage medium for electricity, represents about 63% of the energy from electricity being thrown away. The inefficiency of generating electricity from hydrogen doesn’t apply if it is used directly as a source of high-quality heat, but burning gases always loses efficiency to the lower vs high heating value efficiency differences, and other factors such as distribution weigh heavily upon this approach.
This is why projections of cheap green hydrogen, even manufactured at a point of consumption, are incredibly optimistic. We’ll need to pay the premium for green hydrogen to displace gray and black hydrogen in the demand areas today that, unlike desulphurizing oil or manufacturing fertilizer, aren’t going to diminish radically. But assuming that hydrogen is going to be anywhere as cheap per unit of energy as natural gas, coal, or petroleum fuel products is deeply naive, even before distribution and storage costs.
Appreciate CleanTechnica’s originality? Consider becoming a CleanTechnica Member, Supporter, Technician, or Ambassador — or a patron on Patreon.