
Summer in a tank: methanol, iron-air or hydrogen for a neighbourhood's winter
For a neighbourhood of eight all-electric homes near Breda that wants to generate 90% of its own electricity, seasonal storage with methanol at 10 kWp per roof costs about 2,100 euros per home per year, with hydrogen 3,500 and with iron-air 3,900. Methanol wins on its cheap tank, while only an eighth of the electricity comes back. At 15 kWp per home that advantage disappears and hydrogen costs 900 euros. None of the three reaches 90% in every weather year.
Winter in a tank
Image above: AI impression, not a design drawing.
This week the American company Terraform Industries announced that it makes methanol from sunlight and air, with its own solar panels on its own land. Five years of work. Methanol is a liquid. You store it in an ordinary tank, without pressure, for as long as you like.
That touches the hardest part of a neighbourhood that wants to generate its own electricity. In summer there is far too much, in January far too little. A battery bridges a night. Half a year is another story.
In this lab I model a neighbourhood of eight all-electric homes near Breda in the Netherlands: heat pumps, solar panels on every roof (10 kWp per home), a shared battery and a grid connection. The goal is to generate 90% of the electricity locally, in every weather year. Until now I modelled the winter with hydrogen. I have put methanol and iron-air next to it, hour by hour, for the same neighbourhood.
Three ways to store summer
Hydrogen. An electrolyser splits water using summer electricity. The hydrogen goes into a vessel under high pressure. In winter a fuel cell turns it back into electricity. Of every summer kWh, a third comes back.
Methanol. The same hydrogen is combined with CO₂ from the air in a small reactor to make methanol, which you store as a liquid. In winter a fuel cell turns it back into electricity. 12% comes back, an eighth. The neighbourhood's entire winter supply fits in 1.7 cubic metres: two IBC containers.
Iron-air. A battery that lets iron rust and un-rust. Ore Energy builds them in Delft, in containers of 4.2 MWh. 40% comes back. The battery does leak: an estimated half to one percent of its charge per day.
What it costs
For each carrier I calculated the cheapest neighbourhood that reaches 90% in each of 2018, 2021 and 2022. Panels, shared battery, seasonal storage and the grid bill together, per home per year. With prices for this scale: small electrolysers, the smallest fuel cell available, and whole containers.
| 90% self-generated, 10 kWp per home | per home per year |
|---|---|
| Methanol | €2,137 |
| Hydrogen | €3,495 |
| Iron-air | €3,889 |
| Batteries only | €12,611 |
Methanol is about 1.6 times cheaper than hydrogen. At 95% it is 1.8 times, at 85% they are equal.
The Verkenner on this site shows a higher amount for 90%: 5,131 euros per home. It prices hydrogen storage at 100 euros per kWh, the cautious upper end of what is known. Here I use the middle of what an installation of this size costs for all three carriers; for hydrogen that is 60 euros.
Methanol wins while only an eighth comes back because of the tank. A kWh of storage costs about 4 euros in a methanol tank, 60 in a hydrogen pressure vessel. And the neighbourhood has summer electricity to spare: even with methanol it exports 28 to 40 MWh a year. Efficiency is cheap here, a kWh of tank is expensive.
Two caveats. The ranges overlap: with favourable or unfavourable prices, any of the three can come out cheapest. And a methanol chain of a few kilowatts does not exist yet. That price is my estimate.
When the roof gets bigger
All results above run into the same limit: 10 kWp per home, what fits on an ordinary roof. I repeated the calculation with 15 kWp per home. That is 40 kWp extra, on a barn or a field next to the neighbourhood.
| 90% self-generated, 15 kWp per home | per home per year |
|---|---|
| Hydrogen | €883 |
| Methanol | €1,035 |
| Batteries only | €2,074 |
| Iron-air | €3,322 |
Hydrogen drops from 3,495 to 883 euros, and methanol's advantage is gone. With more summer electricity the store has less to bridge, and then good efficiency counts again.
Part of that gain is export revenue. Extra panels feed electricity to the grid in summer, and in 2022 that paid well. So I also ran everything with a feed-in payment capped at 5 cents. Then 15 kWp with hydrogen costs 2,110 euros and with methanol 2,032. Still cheaper than 10 kWp with methanol: 2,638 euros at 5 cents. For hydrogen, more sun saves more than any choice of carrier.
What those extra 40 kWp cost in land, permits and grid connection is not in the calculation.
Iron-air
In August this lab's capex dossier called iron-air the most realistic path: available, Dutch, better efficiency than hydrogen. Calculated through, it disappoints. The smallest container is 4.2 MWh, while the neighbourhood needs about 2.5 MWh. The charge leaks away: 2.4 to 4.8 MWh a year, almost as much as comes out. And at this scale I use 65 euros per kWh, where 30 to 60 euros is quoted for large installations.
That puts iron-air at 3,889 euros per home, slightly more than hydrogen. At 85% and at 15 kWp, batteries alone are even cheaper. With favourable assumptions (45 euros per kWh, 50% efficiency, 0.5% self-discharge per day) it becomes 2,448 euros. Ore Energy publishes neither a price nor a measured self-discharge. Those two numbers decide.
Buying methanol
You can also buy methanol instead of making it. Green methanol costs about 1,500 euros per tonne today.
If bought fuel counts as an import, just like grid electricity, the model never buys it. Electricity from bought methanol costs about 78 cents per kWh, from the grid about 30 cents on average.
If it does not count, "90%" suddenly costs 954 euros per home. The neighbourhood then burns about 2 tonnes of bought methanol a year in a fuel cell. Count that fuel as an import and the neighbourhood is almost 80% self-sufficient. It reaches that without seasonal storage too (80%), for 384 euros.
Whether bought fuel counts is a choice in the definition of self-sufficient. I count it as an import. Why, you can read in A neighbourhood that buys methanol is an 80% neighbourhood.
The weather of 2010
All amounts above assume a neighbourhood that reaches 90% in 2018, 2021 and 2022. I also ran the same installations through fifteen weather years between 2005 and 2023. Each carrier reaches 90% in 6 to 8 of those 15 years. In the winter of 2010 they all drop to 82 or 83%.
Whoever wants to make it through 2010 as well builds at more than twice the cost: hydrogen 7,184 euros per home, methanol 4,433, iron-air 8,183. The carrier sets the price level. The weather remains the same problem for all three.
How the result shrank
My first calculation, on the day of Terraform's announcement, had methanol five times cheaper than hydrogen. I had the calculation checked three times by a second AI model that received only the files and the question. Each round found a comparison that was skewed:
- the methanol chain was scaled from a 100 MW plant, which left 700 euros at this size for a complete reactor;
- methanol on paper was set against hydrogen with every extra cost included;
- iron-air got the price of a large installation, without whole containers;
- one carrier got a minimum size, the other did not;
- the gain from extra panels leaned on the electricity prices of 2022.
Every correction made the number smaller. From five times to 1.6 times, and with a bigger roof to equal.
What the numbers say
- At 10 kWp per home and 90 to 95%, methanol is about 1.6 to 1.8 times cheaper than hydrogen. At 85% they are equal.
- The tank makes the difference. A kWh of methanol tank costs 4 euros, a kWh of hydrogen vessel 60.
- At 15 kWp per home methanol's advantage is gone. Hydrogen then costs 883 euros per home per year.
- 15 kWp with hydrogen is cheaper than 10 kWp with methanol, even if feed-in pays only 5 cents.
- Iron-air does not get below hydrogen at this scale as long as you have to buy a whole container.
- Bought methanol only wins if you do not count it as an import.
- None of the three reaches 90% in every weather year. Making it through 2010 too costs more than double.
What this rests on
The weakest spots, for anyone who wants to attack this:
- The methanol chain. A chain of a few kilowatts (electrolyser, CO₂ from the air, reactor) does not exist yet. I use 10,000 euros per kW of charging and the smallest fuel cell of 5 kW. With paper prices methanol becomes 1,057 euros, with prototype prices 3,991.
- Iron-air. Ore Energy publishes neither a price nor a measured self-discharge. I use 65 euros per kWh and 0.75% per day.
- The model knows the weather in advance. It plans the whole year with perfect foresight. A real controller achieves somewhat less.
- The neighbourhood. Eight all-electric homes near Breda, with the weather of Gilze-Rijen and Dutch electricity prices of 2018, 2021 and 2022. Feed-in at the wholesale price, with the 5-cent cap alongside.
- The extra panels. The 15 kWp variant prices them like roof panels, without land, permits or connection. The neighbourhood then exports 95 MWh a year; the grid operator has to allow that.
- The weather years. The hard years 2010 and 2017 partly use estimated electricity prices. For whether the neighbourhood reaches 90%, that makes no difference. For the costs it makes a small one.
- What is missing. Waste heat from the fuel cells, fixed costs per installation, permits and safety for a methanol tank or a container in a residential area.
Check it. If you know the price of a small methanol chain or the measured self-discharge of an iron-air container, I will run it again.