Pyrolysis for transporting solar energy
An expensive option for baseload power if nuclear, batteries, or CCS aren't available for some reason.
Unfortunately, this idea didn’t turn out to be as exciting as I’d hoped. If the value of an idea was proportional to the time spent on it, I’d be rich. Sigh.
In a previous post I proposed and rejected several options for “shipping” solar energy to other parts of the world. The best option was to ship solar fuels, but but a BOTEC shows that the resulting cost of energy is a little high.
Part of the problem with solar fuels is how expensive they are, even with future advances. An optimistic price for green hydrogen is $2/kg, and an optimistic price for CO2 capture is $100/tonne. Making 1 kg of hexane from these two components requires 0.44 kg of hydrogen and 3 kg of CO2 even with perfectly efficient reactions and zero reactor costs. That comes out to about $1.2/kg of hexane before shipping, whereas oil trades at around $0.50/kg. Not so good.
Direct air capture adds almost 50% of the cost and CO2 doubles the required hydrogen. In the last post I considered whether to capture the CO2 from power plants and ship it back to the places producing solar fuels, but CO2 shipping costs were more expensive than air capture.
What if we didn’t burn the carbon at all? A process called pyrolysis1 heats organics to release hydrogen while leaving behind carbon. Solid carbon is easy to ship back, it’s basically coal.
Here’s the plan:
A sunny and/or windy place like Australia, Namibia, or west Texas uses renewable energy to produce green hydrogen.
The hydrogen is combined with solid carbon on-site to produce fuel via the now-defunct Bergius process.
The fuel is shipped across the world in oil tankers.
When the consumer needs energy, the fuel is pyrolyzed into hydrogen and carbon.
The hydrogen is turned into energy with a hydrogen fuel cell.
The carbon is shipped back via bulk carrier.
Besides pyrolysis, there are existing technologies for each step. But careful design is necessary to make the cycle economical. As we saw in the solar shipping post, you can only share cheap solar energy if your transmission method is cheap.
Solar and wind are cheapest if you design your process around that intermittency rather than demand constant output. To do this, your equipment has to run for part of the day; each piece of equipment is going to generate less product. In that case, it’s better to select equipment that’s simpler and cheaper per unit of input in order to keep your equipment costs low.
The result of all this is a method for shipping and storing solar energy anywhere in the world with the ability to produce industrial grade heat and minimal carbon emissions at a price of $200/MWh. This is too expensive for most users, but could provide an emergency source of power or baseload power for regions lacking renewable energy.
Hydrogen electrolysis
The hydrogen electrolyzers have to be cheap. That means dropping high temperature, pressurized designs and making them out of plastic. We can do away with the fancy membrane if we use alkaline water electrolysis.
We can increase our efficiency slightly by running hot without boiling, perhaps 90 C (EDIT see footnote2). This will require a plastic like PEEK or PPS that can withstand the hot alkaline solution. The downside of higher temperatures is that our nickel electrodes will corrode faster, so some recycling from the solution is important. Stainless steel anodes might be a good option. Fortunately, sunny Australia is the 5th largest producer of nickel3.
Because the Bergius process requires hot hydrogen input and produces excess heat, there’s room to optimize the temperature and materials used during electrolysis. There’s also an option to add hydrogen storage so the equipment can achieve higher utilization. Locating near a salt dome would make this particularly attractive.
If onsite hydrogen storage is very cheap, high utilization could be feasible. In that case, you could run a high temperature electrolysis system powered by excess heat from the Bergius process.
For the cost calculation at the end, I’m going to make the optimistic assumption that green hydrogen can be produced for $2/kg.
The Bergius Process
The Bergius process exposes powdered carbon to high pressure hydrogen to produce liquid fuels. A hydrogen donor solvent and a catalyst are used to accelerate the process. Metals and sulfur can act as catalysts, but it would be better to use a catalyst that can be generated from the product such as oleic acid. Hydrocarbon catalysts can enter the pyrolysis step without being removed. Supplementing coal to account for roundtrip carbon losses can provide catalytic sulfur and metal.
After each batch, there will be some unreacted hydrogen and tar which can be recycled into the next step leading to very little material lost.
The process generates 5x-10x more heat than is needed to warm the water for electrolysis. Some of this could be put to use desalinating seawater or cracking the Bergius products into something more valuable, but some heat will have to be stored or sold.
Oil is better than methane
There are two liquid fuels that we ship at scale, oil and liquid natural gas. Oil is much cheaper to transport per MWh of energy delivered because it’s stable at ambient temperature and less corrosive than alternatives. The same is true for shipping via pipelines, trucks, and rail. It’s easier to deal with at every step.
As long as the hydrocarbons we make are longer than butane, they should be liquid during shipping. It turns out the length of the carbon chain doesn’t matter much for delivered energy costs, so I’m going to do the calculations assuming we’re working with hexane. In practice, shorter chains are a little better since they store more hydrogen per unit mass and are less viscous.
Methane has some promise as a transitional fuel since methane pyrolysis is much more established. This requires turning the solid carbon into carbon monoxide and using a modified Sabatier process rather than the Bergius process. This wastes more energy and increases hydrogen requirements by 50% but might still pencil out if hydrogen gets cheap enough.
Shipping
Shipping is easy. We have very large crude carriers for shipping oil and bulk carriers for moving coal. Oil tankers cost around $20/tonne of oil. Bulk carriers vary in cost but I’m going to assume $25/tonne is reasonable.
One trick to reduce costs is backhauling. Oil tankers and bulk carriers can’t return empty. They need a ballast of around 30% of their normal tonnage to maintain stability. The oil tanker might be able to carry a small amount of coal back to the producer, and the bulk carrier might be able to deliver a small amount of oil to the consumer.
Even better would be to modify an oil tanker so that it can carry both oil and coal slurry. If each reaction can handle a little contamination, there’s no need to clean the container between shipments. This could halve shipping costs in the best case. However, processing coal slurry is tricky and comes with environmental risks. It’s safer to stick with bulk carriers for the return trip.
(EDIT: it looks like ore-bulk-oil (OBO) carriers are a perfect fit for this. Unfortunately they have declined in popularity since the 80’s.)
Another opportunity is to find an alternative fuel for the ships. Burning the fuel out of the Bergius process would be tricky since it’s not optimized for the ship (though many ships can burn low-quality bunker fuel). Battery power could work for shorter distances or trips with several hops. A smaller version of the pyrolysis and fuel cell system might also work.
Pyrolysis
First, a short note on terminology (everyone’s favorite). Pyrolysis (or thermolysis) heats organic matter in the absence of oxygen. It comes in different forms, often producing oils or tars. What I’m going for might be better described as carbonization or coking. Hopefully this delightful image from Wikipedia clears it up:
Pyrolysis is the step with the most technical risk. While converting methane into carbon and hydrogen is well established4, there’s little research on pyrolyzing oil to generate hydrogen. That can be a good sign or a bad sign depending on how you look at it.
Pyrolysis requires energy in the form of heat. Some of the heat output from the fuel cell will have to be recycled to run the pyrolysis step. Pyrolysis also isn’t perfectly efficient at squeezing the hydrogen out of the hydrocarbon. This turns out not to have a big effect on costs, since the unused hydrogen gets shipped back to be recycled in the Bergius process and adds little mass during shipping. I’ll assume 70% of the hydrogen present in the fuel gets released.
The power plant
My first thought was to burn the hydrogen in a modified natural gas plant. But this turns out to be a lot of work. Hydrogen is awful stuff, it damages metals, it’s one of the few gases that heats up as it leaks out of small pores, and it burns with an invisible, superhot flame. A combined cycle plant would have to completely redesign their first turbine and piping to accommodate hydrogen’s unique properties. Better to keep these plants on the job of grid stabilization.
The other problem with combined cycle plants is that they’re “only” 60% efficient and require a lot of startup capital intensive. Fuel cells are easier to scale down and can achieve higher efficiency if we’re just using hydrogen5.
We have to capture some of the waste heat from burning hydrogen to achieve these higher efficiencies. These systems can achieve 85% conversion efficiency, 60% in the form of electricity and an additional 25% as heat. Molten carbonate fuel cells (MCFC’s) operate above 600 C and can sell high quality industrial heat.
MCFC’s don’t require precious metals and are more resilient to contaminants in the gas stream. That means we don’t have to worry as much about the outputs from the pyrolysis and Bergius steps.
The drawback is that the molten carbonate is corrosive, requiring regular replacement of the container and nickel electrodes.
Final energy cost and conclusion
Starting with the long list of assumptions in the appendix, the variable cost of electricity and heat is $105/MWh. Adding in the capital costs of the MCFC, that rises to $188/MWh, so let’s call it $200/MWh. This may be somewhat of an overestimate since I assumed we’re not going to recycle the nickel or sell the heat for a higher price. Clever shipping arrangements could knock the price down by $8/MWh.
For comparison, industrial energy costs about $100/MWh today and the cheapest sources reach $20-$40/MWh under good conditions. So pyrolysis energy can only act as a small, stabilizing supplement to the grid rather than the main source of power.
If hydrogen costs reach the vaunted $1/kg the scheme starts looking better, $53/MWh variable cost. But in such a world a lot of things are different and this process might not be competitive for other reasons6.
Instead, I treat this as an upper-bound on the cost of clean baseload power. It would be better to just build more nuclear, batteries, and flue gas capture, but in the absence of a better solution, there’s a way to ship solar energy anywhere and store it for indefinite periods at a high-but-not-impossible cost.
For this to work there needs to be innovation at both ends. A sunny place could start by making e-methane with local solar energy and DAC. A climate-conscious city could start pyrolyzing natural gas rather than burning it. Then form a long-term shipping arrangement between the two. Optimizing roundtrip costs will naturally lead to bespoke reactors, different fuels, and new ship designs.
EDIT: Another implication of all this is that we can get clean power from natural gas today. Turquoise hydrogen from methane pyrolysis would generate hydrogen around $2/kg. Finding a way to sell the carbon byproduct would lower costs further7. That being said, there are better technologies to meet baseload power needs.
Appendix
Assumptions
Assumptions for the cost calculation:
On-site hydrogen at $2/kg8.
No sale of Bergius process excess heat.
The cost of Bergius reactor and pyrolysis reactor add 20% to final variable costs.
Bergius fully utilizes hydrogen and carbon.
Crude oil shipped at $20/tonne.
Ignoring intermediate shipping costs (trucks, pipelines, rail, transfers, etc).
70% release of pyrolysis hydrogen.
Pyrolysis requires 10 kJ/mol H2 released (see table 1 here).
85% efficient MCFC.
MCFC lifetime of 50,000 hours.
MCFC capital costs of $2000/kW.
MCFC supplies 30% of energy needs with capacity for 50%.
High quality heat sold at same price as electricity.
Carbon shipped back at price of $25/tonne.
No nickel recycling.
Ignoring catalyst generation costs.
Ignoring carbon and hydrogen losses.
MCFC financing costs of 10%.
To sketch out how I did this calculation, our journey starts with 1 kg of hydrogen ($2 cost). It’s combined with hexane that has been depleted of 70% of its hydrogen, C10H7. The overall reaction is:
49 H2 + 6 C10H7 → 10 C6H14
1 kg of hydrogen becomes 8.72 kg of hexane. Our crude shipping cost adds $0.17. The pyrolysis reactor generates 0.85*33.3 kWh/kg = 28.3 kWh. The heat of pyrolysis is 1.38 kWh. Net energy production is 26.9 kWh. Returning the 7.72 kg of coal adds $0.19.
Dividing total cost by net energy output gives $88/MWh, the 20% reactor cost brings us to $105/MWh.
For the MCFC cost, we’re paying $2000/kW and dividing that cost over the 50,000 hour lifetime. Then we scale this cost by 50%/30% since we’re buying the capacity to supply 50% of grid energy in emergencies, but only getting 30% of our energy out on average. Then we multiply by 10% for financing costs with a final cost of $73/MWh.
Further reading
Molten carbonate fuel cell cost models
Molten Carbonate Fuel Cells for 90% Post Combustion CO2 Capture From a New Build CCGT
Fuel Cell Power Model Version 2
Fuel Cells and Renewable Hydrogen | WBDG
Performance and Cost Analysis for a 300 kW Tri-generation Molten Carbonate Fuel Cell System
Molten Carbonate - Hydrogen fuel cell electric vehicle
Hexane and hydrocarbon pyrolysis
Hydrogen Production by Thermal Plasma Pyrolysis of Hydrocarbon Gases
Methane pyrolysis
Methane pyrolysis: the case for cleaner hydrogen with existing infrastructure
Hydrogen Shot Summit: Methane Pyrolysis Panel
On-Site Hydrogen Production via Distributed Methane Pyrolysis
A techno-economic analysis of future hydrogen reconversion technologies - ScienceDirect
Companies doing methane pyrolysis
Solid carbon products
One option is to convert the solid carbon waste into a valuable product. In that case, the solar factory can just import coal and the power plant can upsell its waste product. A list of ideas is included below.
This assumes there’s a market for solid-carbon products that’s large enough. Unfortunately everything I can think of is far too small to match the huge amount of carbon we use for energy.
The climate impact is also iffy. On one end, higher demand for coal reduces it’s use as an energy source. On the other hand, displacing solid carbon products frees up the fuels that were previously used to make them.
Asphalt
Biochar
Battery graphite
Carbon black
Graphite composites
Carbon foam insulation
Carbon fiber composites
Diamond powder
Full size diamonds
Diamond structural beams
Diamond palaces
Nuclear graphite
Space tethers
Government strategic energy reserves
Carbon capture
Plastics
Silicon carbide and other carbides
Polluting uses:
Coal
Hall-Heroult electrodes
Coke for carbothermic reductions
Electric arc furnace electrodes
See the note on terminology in the pyrolysis section.
Turns out since the solution is ~40 wt % base, the boiling point is much higher, perhaps closer to 120-140 C. This means the electrolyzer can be a little more efficient. I wonder what the optimal concentration and temperature combination is in the presence of the cheap heat from the bergius process.
Neighboring Indonesia, Philippines, and New Caledonia are 1st, 2nd, and 4th respectively.
Methane pyrolysis benefits from a catalyst. Fortunately, a specialized form of solid carbon can act as a catalyst, which means it can be regenerated using some of the solid carbon output from the pyrolysis step.
Another option is to use lightcell’s approach. Their target efficiency is only 40%, but it may be more capital efficient.
For example, in this case it may make sense to overbuild solar and use hydrogen storage to patch seasonal variability.
Many are working on this, see appendix. Unfortunately no solid carbon market is large enough today.