Iron as an inexpensive storage medium for hydrogen
ethz.ch
ethz.ch
Unfortunately, I don't see this making any sense for large scale energy storage. Storage tanks for compressed hydrogen enjoy the square-cube law. The larger they are the less expensive they are proportional to the mass of hydrogen they hold.
With this iron oxide method, you need 27 tons of iron oxide for one ton of hydrogen. You can procure right now tanks that can hold 2.7 tons of hydrogen and weigh 77 tons empty [1], the ratio is 28 to 1. But the round-trip efficiency of the tank is virtually 100%. The efficiency of the iron-based storage is only 50%. The tanks are not very expensive.
I can't see the niche that this idea can apply to.
[1] https://www.iberdrola.com/press-room/news/detail/storage-tan...
2.7 tons of hydrogen have a volume of almost exactly 30000 m^3, requiring storing it under high pressure in specialized containers. Hydrogen is famous for being hard to store without losses.
For long-term storage storage and losses are a problem.
> But the round-trip efficiency of the tank is virtually 100%. The efficiency of the iron-based storage is only 50%
Maybe I'm missing something, but why? As you mentioned it takes 29kj to restore 3 moles of H2 out of (3 moles of H20 + 1 mole of Fe2O3). Where does 50% comes from?
the efficiency is super low, but again, according to the paper, "most of the energy input was due to thermal losses at the reactor surface (83.9%)", which also benefits from square/cube law.
[0] https://pubs.rsc.org/en/content/articlelanding/2024/se/d3se0...
Tbh, I am not sure either. I think the main benefit of this is FeO is inert under temps/conditions humans consider normal so maybe it long term storage is not that far fetched. I like the idea. I am just unsure about its practical applications.
Not really. Wall thickness is roughly proportional to diameter, and surface area to the square, so you don't gain anything in terms of storage mass ratio by building bigger tanks.
> But the round-trip efficiency of the tank is virtually 100%
This is oversimplifying quite a bit. Compressing hydrogen, the lightest gas, is very energy intensive per unit of mass, and this energy is not fully recoverable upon decompression (due to general pump efficiency and thermal losses in the intercooler).
Edit: wait I forgot that direct reduced iron powder exothermically reacts with oxygen and water in the air, that's how single-use instant hand warmers work. So yeah you gotta isolate the iron powder a bit more than stick it under a tarp.
I've been playing too much Factorio lately so of course my mind goes towards rail systems (could repurpose coal plants) in combination with pneumatics.
This system doesn't store hydrogen. It stores elemental iron (produced from iron oxide, i.e., iron ore, and hydrogen from solar power splitting water into hydrogen and oxygen), and uses steam to get the hydrogen out (and convert the iron to iron oxide) only when the hydrogen is needed.
Fwiw, cervelat is also very common in France, I grew up eating that stuff. Maybe that's why I liked the article so much. There's something to dig up there
https://www.metallics.org/dri.html
> Being a highly reduced material, DRI has a tendency to re-oxidise, an exothermic reaction. Thus, without appropriate precautions being taken in its handling, transport and storage, there is a risk of self-heating and fires. The International Maritime Organisation's International Maritime Solid Bulk Cargoes Code classifies DRI - Direct Reduced Iron (B) - as Group B (cargo with chemical hazard) and class MHB (material hazardous only in bulk) and requires that DRI be shipped under an inert atmosphere, usually nitrogen.
It would be nice if the iron could be in an alloy that, in addition to being oxidized/reduced, could further absorb hydrogen when in the reduced state. FeTi absorbs hydrogen, but I don't think the titanium would withstand repeated oxidation/reduction cycles. The Ti would go to the +4 oxidation state and stay there.
It's hydrocarbons.
In this case, synfuel hydrocarbons as direct analogues of fossil-fuel based compounds of chain-lengths 1 (methane) to around a dozen or so (kerosene / aviation fuel, at a stretch, diesel fuel).
It stores forever (proved to 300 million years), it is drop-in compatible with extant infrastructure and equipment, it's infinitely miscable with present fuels, it doesn't leak out of storage, it doesn't embrittle metals (and in fact generally lubricates and protects them).
Yes, the round-trip storage efficiencies are low (as low as ~15--20% recovery based on thermal electrical generation, roughly the same as the solution named here), but that's in exchange for something that can readily provide weeks to months of storage capacity in a stable, low-risk form. Where you need storage that's long-term stable, dense, safe, and instantly dispatchable, your options are few.
The technology has been demonstrated in numerous experimental trials, and is similar to processes run at national scale for decades in Germany and South Africa. US-based research has been conducted at Brookhaven National Laboratory, M.I.T., and the US Naval Research Lab, amongst others. The stumbling block to date has been that fossil fuel prices are sufficiently low[1] that synfuels simply are not competitive presuming market-based mechanisms which fail to account for externalities and other market failures.
I've be aware of this for about a decade and have written about the technology, Fischer-Tropsch fuel synthesis, multiple times on HN:
<https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...>
________________________________
Notes:
1. A market failure of staggering proportions, as the under-pricing is on the order of a million-fold. See: Jeffrey S. Dukes, "Burning Buried Sunshine", <https://core.ac.uk/download/pdf/5212176.pdf> (PDF)
As in how many power to gas plants are operating currently? I would guess not many due to the price vs fossil methane, but as the GP notes this is a major market failure due pricing the externalities of fossil methane.
>> And how do you get the energy back out?
Easy, burn it. For heat, or with a turbine, electricity. Zero carbon impact because it started out as atmospheric CO2.
Synthesis efficiency is on the order of 60%. Combustion efficiency, driven by Carnot's Law, ~30--50%, possibly a bit higher for combined and/or power-to-thermal. But the overall round-trip efficiency, based on both synthesis and generating losses will be low, 15--25%, perhaps as high as 30%.
The goal of such a system isn't efficient storage so much as it is persistent and dispatchable. It's possible to store large quantities of liquid fuels for long periods of time and tap into them within seconds or minutes as generating, or other fuel-based energy applications, are required. Other storage/release options either don't have the long-term storage capability, or cannot be ramped up as quickly.
Other storage options range from the exceedingly short-lived but responsive (supercapacitors, kinetic flywheels) which can respond on a sub-second to second basis, but aren't viable for long-term storage at all, to options which work over a few hours, daily supply-demand inconsistencies, etc., such as pumped hydro and high-termperature thermal storage (e.g., molten salt), to seasonal thermal storage (hot water, ground geothermal where geology supports this).
Fuel synthesis makes sense where there's exceedingly high seasonality in energy generating capacity (e.g., summer vs. winter for PV / solar), or where there are longer-term highly-variable generating trends (e.g., wind, with higher and lower output over the course of days or weeks). Or where communities are highly isolated and cannot readily be connected to a grid to balance consumption and generation over time (remote islands, arctic communities), or where a region has a very high generating potential but low domestic demand (say, northwest Australia, with sun-drenched deserts and low electrical demand from sheep and roos).
Banking summer-time / high-production capacity as fuel, and drawing down those fuel reserves or shipping them to areas of high demand could be the shape of a future energy economy. It should dramatically rewrite energy geopolitics.
> Zero carbon impact because it started out as atmospheric CO2.
<https://www.epa.gov/natural-gas-star-program/primary-sources...>
In situ / on-prem methane production, storage, and usage would tend to minimise much of that.
Further, I don't see methane synthesis as a likely option for synfuel storage applications as described here, as heavier fuels (~C-12 -- C-15 chain-length kerosene) is liquid (far easier to handle than a gas), nonvolatile (as contrasted with petroleum or lighter hydrocarbon distillates), and can be stored in simple, unpressurised, unrefrigerated tanks virtually indefinitely. You're creating a problem which needn't exist.
Boiling water, internal combustion (turbines or reciprocating engines), or external combustion (heating a boiler, and using the steam to drive turbines), would cover most of it. There might be some more exotic options of fuel cells, but those tend to be cost-prohibitive.
All combustion processes follow Carnot's laws, which limit fundamental efficiency.
The vast, cheap power that photovoltaic will provide is a giant opportunity. Please review the links below
https://terraformindustries.com/
Terraform industries converting sun and air into natural gas:
https://techcrunch.com/2024/04/01/terraform-industries-conve...
The solar industrial revolution is the biggest investment opportunity in history:
https://caseyhandmer.wordpress.com/2024/05/22/the-solar-indu...
https://www.scientificamerican.com/article/rusty-batteries-c...
This paper uses hydrogen as an intermediary, which has advantages but also adds some questionable margins in efficiency. But I don't know the efficiency of the suggested iron-air batteries either.
This may be nicer if you want hydrogen rather than an electric battery. But if you turn that hydrogen into a fuel cell... the efficiencies of producing and consuming that hydrogen add up.
When storing energy, the idea is to split water into hydrogen and oxygen, and then let the hydrogen recombine with the oxygen from iron oxide... to make water again. Meanwhile the oxygen from the original water is just released (since it's everywhere anyway)? That doesn't really seem to me like "storing hydrogen", since you just get the water back that you already had. Rather, it's using the energy to deoxidize rust.
Then on the recovery side, why use this steam process? Apparently (because the thermodynamics work out so that this whole thing has efficiency > 0) you get energy out of the process of putting the oxygen back into the iron. So why not just, well, burn (i.e. rust) the iron directly? What exactly is the dissociation and re-combination of the steam accomplishing?
Direct use of reduced iron as an energy carrier has been discussed here:
https://news.ycombinator.com/item?id=24996153
Oxodizing iron gives rather low intensity heat, comparable, iirc, to burning lignite. That might be good for some combined cycle applications (but still a logistics nightmare?), but it's not a drop-in replacement for anything. Hydrogen on the other hand is good for an extremely wide range of applications, from fuel cells to e-fuel production (and all kinds of other chemical processes) to flying Neil Armstrong to the moon.
To succeed in decarbonization we will need an entire "cache hierarchy" to solve the intermittency problem, a single storage solution will never be enough. Batteries to make hydrolyzers able to run around the clock during high availability seasons, hydrogen storage to make converters further down the pipeline run continuously during surplus seasons and not only on the best days.
What will definitely not get us to decarbonization is any of the following three approaches:
(a) building enough production capacity that we don't need storage (most production would be idle most of the time for lack of a buyer)
(b) focusing on one kind of storage (same problem again, now with conversion capacity)
Exactly, really what they are storing is electrons.
Rusting the iron directly releases heat, which limits your efficiency to the delta-T of the process. Reducing steam to hydrogen and then converting the hydrogen in a fuel cell allows higher efficiency to produce electricity.
Or would the plan be to slowly heat over fall?
I do wonder if “free” will actually pan out, or whether someone will find a way to demand-shift from winter to summer and use it all up.
At that point, it's always summertime somewhere and it's always daylight somewhere, and if prices were to fall to zero there is always someone who would like more heat for something.
Therefore I suspect zero-priced energy will stop existing.
'Take my energy and allow me to stop accelerating my flywheels which regulate production' seems more plausible than 'someone would always like more heat for something' (what?)
Or possibly 'take my energy and I'll cut off some of the people using spare energy to do low priority, low value computation for free'?
You can imagine home appliances having an 'eco' setting which runs the appliance like the washing or the dishwasher at the cheapest time in the next 12 hours.
Or the water heating systems which heat more water when prices are cheap.
Or heaters which switch between natural gas and heat pump based on price.
Or electric car chargers which charge during the cheapest hours.
(all of these already exist, but none are yet common).
Over the long term there is also plenty of elasticity. If electric heating is expensive, people will install gas/oil heaters when they renovate. If electricity is cheap, more people buy electric cars. With cheap electricity, maybe fewer people decide to add more insulation to their houses. Businesses don't upgrade energy inefficient equipment to be more energy efficient, etc.
Plenty of demand elasticity in both the short and long term. End result: As long as the market is unconstrained, prices won't hit zero more than say ~5% of the time.
If peak to trough is a large gap, say 60% of peak, this tends to make it less likely that peak will be met by overproduction, since that would involve very large capital costs.
The picture you paint above would suggest a very small gap between peak and trough, say 2% of peak. This means that almost certainly there would be enough over capacity to more than meet peak demand. Therefore the total daily demand would be more than met by capacity, leading to some energy being thrown away. So at all times except the peak, the marginal cost would be zero.
You have given an accurate argument for why demand would be elastic at trough. But you haven't given any reason why overall demand would be very elastic.
It's a non-problem, really. Especially at scale.
What I'm not getting is how this process produces more energy than the solar input to power the process.
Unless they're getting solar collectors to try to generate 400 degree temperatures rather than PV solar to electricity, but that seems like a sketchy proposition at best in winter.
Obviously the energy efficiency of this process based on iron is modest. It is likely that the energy efficiency is even lower than for the process of storing energy by making synthetic hydrocarbons (e.g. synthetic gasoline), which are much easier to use once energy is stored in them.
The only advantage is the very low cost even for very large storage capacities.
So far their is even lower, though they claim a theoretical max of 79%. Storing large amounts of energy that's ready to be used is rarely not dangerous in any case. Except maybe potential energy of a tank of water on a mountain.
In the paper, the authors mention 11.4% efficiency for this system and a theoretical maximum efficiency of 79% if scaled up, so it might take a lot of scale.
Both charging and discharging seems to require a lot of heat. Waste heat is essentially lost energy that is released in the form of heat. I assume the discharge reaction is exothermic. That would be the energy stored in the summer months. Heating up a lot of tons of iron during charging is also not going to be free. It doesn't matter whether you do it slowly or quickly.
Creating the hydrogen is also not a loss free process. Nor is doing something useful with it like using it in a fuel cell (0.85), burning it (0.45), etc. These inefficiencies multiply.
All that lost energy comes out of the original budget of energy that came out of the solar panels.
Even if you use some wildly optimistic numbers, they multiply to something well below 0.5 pretty quickly even before you consider charging & discharging.
But lets do something silly and unrealistic and just do the math for an average step efficiency at 0.7, 0.8, and 0.9. We're talking four conversions here so that's 0.7^4 =0.24 vs. 0.41 and 0.66. And forget about getting anywhere near average 0.9 efficiencies with all of those steps. I'm assuming 0.7 would already be on the high side. Add more steps to the process and it only gets worse. Pipes aren't perfect. If you need to pressurize the hydrogen before you use it (like in a car), that isn't free either.
Basically, this takes a system that was already quite inefficient end to end and adds two more steps that sound like they involve some pretty significant energy losses to it (i.e. probably well below 0.5 when combined), thus making the system as a whole a lot more inefficient. Hydrogen as a battery already sucked with normal storage. This doesn't improve things.
There's a good reason that most hydrogen produced is used at or close to its site of production: it minimizes the energy losses and producing hydrogen is really expensive so it's not really desirable to lose 80-90% of the energy unless you really need to.
I see:
1) generation 2) storage efficiency (energy while storing divided by energy upon release)
what are the other 2 you had in mind?
All those steps lose energy. And there's stuff that happens in between involving pipes, leaky valves, tanks, compression, etc.
Somewhere with lots of solar on the grid probably has excess energy, even during winter, during the day, so you'd plan to put in the input energy to start the reaction during the afternoon peak, and if you miss that for some reason, some sort of coordinated startup procedure would likely be used.
I would say it's only worth it if the marginal cost of producing the hydrogen is close to 0.
If your battery's efficiency gets bad enough the added winter power from those extra panels is going to be enough to cover the winter shortage - so you don't even need the battery at all. You'd essentially just be turning a huge amount of power into heat for nothing.
Depending on the latitude/weather/etc, the difference between winter PV production can see a reduction of between 40% and 85% compared to summer output (sampling from [0]).
Panel prices have dropped so much, it's far more likely that, at least with places where summer output is double that of winter or less, that simply doubling the number of panels solves the problem.
Batteries and PV are cheap and getting cheaper all the time, are proven and are absolutely trivial to install, operate and maintain compared to ANY conceivable process that involves hydrogen.
I don't get the apparently insatiable drive/impulse that drives people to put so much effort into shoehorning hydrogen into the energy sector. It's expensive, inefficient, dangerous, leaky and is a potent greenhouse gas[1].
The hydrogen idea has been around for over 100 years, has never worked at scale, despite many efforts and large investments. It's time to walk away and concentrate effort into technologies that have actually delivered and have already reduced carbon emissions at scale.
[0] https://pvwatts.nrel.gov/pvwatts.php
[1] https://www.dnv.com/article/is-hydrogen-a-greenhouse-gas--24...
For example, an aluminium smelter might use free summer energy to turn alumina into aluminium, and turn the plant off in the winter - at which time the labor force can do extended maintenance and do post-processing on the aluminium produced during the summer.
Although finding an industry where such seasonal turnovers are viable isn't trivial, if you find one it's going to be far more profitable than very lossy storage.
In reality, the battery is needed when the battery is needed. It might be a long cold winter, or maybe just few evenings after rainy days, but a cheap long lasting battery is still useful.
Unless you also have a nuclear reactor to perform the high-temperature electrolysis process. https://en.wikipedia.org/wiki/High-temperature_electrolysis
Which only Japan seems to be working on building at the moment. https://www.hydrogeninsight.com/production/japan-plans-hydro...
It's cool tech, and I like that it's researched. But I think it's not a breakthrough that could realistically help the grid at the moment.
It’s annoying that they always seem to contain some hand-wavy efficiency calculations. I think this one didn’t even consider the losses from hydrogen production? Is there a benchmark out there, of these long-term electricity storage solutions? Like: you get 1 MWh at 25 deg C, and 6 months later, it’s measured how much your system restores. Everything taken from the grid during storage for upkeep or kickstarting the process is subtracted as well.
Also, I doubt this will exist in isolation. It will probably be built in places that already have demand for hydrogen, available land, have accessible energy, a good grid connection, and existing iron-ore infrastructure. It will be built in such a way to minimise costs and maximise available customers.
I absolutely agree that we should bet on the battery industry. But there is a difference. With battery technology the cost of storing 1GWh is fully coupled to the cost of the equipment. That is not the case when you can store something. It could run for longer durations with the same cost in equipment. You just store more fuel.
Another factor here is industrial capacity. Battery production has amazing capacity for growth. But that is not a zero-sum game. Other parts of industry could also be scaled up.
Its interesting how similar this thread is to debates I have had with people about solar/wind/battery storage. We take one measure and then focus on that exclusively. Whilst ignoring the wider commercial, industrial and technical factors.
https://www.pnnl.gov/news-media/baking-soda-solution-clean-h...
And for good batteries their density doesn't really matter.
Sodium ion batteries will be fantastic grid storage simply because they're just going to be dirt cheap
Will be interesting to see how their campus power project works out in the next years.
Also, innovative “sausage safety test” for the Fe powder reacting with air (figure 6: https://doi.org/10.1039/D3SE01228J ).
Further is that vehicles can use propane so you don't even have to idle the plants during the winter so long as there's some PV from the southern states. They might be running at 100% capacity in summer and 40% capacity in winter but that's way better than 0%. It's a lot easier to keep people employed to operate the plants if they're needed year round.
A certain faction of the project to decarbonize the electrical grid likes to play a similar childish game: "nuclear power is lava". It causes them to come up with whimsical and absurd epicycles, which make no sense at all unless you're playing that game.
Seasonal storage of 2GWh? Please. A 2GW plant produces 2GWh every hour, with 90% uptime. And it doesn't involve losing more than 90% of the photovoltaic energy, I will eat my whole hat if the ray-to-electricity pipeline for this boondoggle exceeds 10% efficiency.
Can we please stop wasting time and effort, and invest in the buildout of a substantial nuclear fleet to provide baseline power?
Go ahead and look at the lazard lcoe numbers, and this year was an odd increase for solar wind that will basically be the best that nuclear can hope for.
If nuclear could provide a cheap scalable easily approved rapidly deployed safe and low waste reactor that could be price competitive with solar and wind, then they be in the game.
That is not happening without probably 10 to 20 years of research and development with billions of dollars of funding.
Old nuclear was solid rods and gigantic domes and all of that stuff simply is not price competitive, and even if you dropped everything and started attempting to get approvals and construction for hundreds of nuclear plants, they won't come online for 10 to 20 years themselves.
When's going to get cheaper solar? Certainly going to get cheaper with perovskites. Stores will get cheaper with lfp and sodium ion improvements, solid state and hopefully someday sulfur chemistries.
The nuclear industry needed to get its act together decades ago. In my opinion, it made a huge error in abandoning msr in the 70s (which has all the features of a competitive nuclear plant if they could figure out the materials science) and other breeder reactors.
A solid fuel rods reactor is simply not reliably safe in all disaster scenarios (Fukushima).
Is it possible to capture that heat during production and conversion, and use it to run a turbine?
(which is also kinda interesting since in parts of Switzerland district heating through waste management facilities starts to be a problem due to a reduction of the amount of waste available)
<https://en.wikipedia.org/wiki/Combined_cycle_power_plant>
<https://en.wikipedia.org/wiki/Cogeneration>
Seimans has a good video 'splainer: <https://yewtu.be/watch?v=eeiu-wcyEbs>
GE also has numerous offerings in this space, and I've seen trade articles describing this in the past, though I'm not finding any presently.
There isn't some majyckal process by which all thermal energy can be converted to motion (or by extension, electrical generation), but it is possible, with additional complexity and capital equipment, to extract much of it.