Solar-powered jet fuel (and diesel, and...)

Sounds kind of backwards, I suppose, but there is in fact research happening on creating jet fuel, and other liquid fuels, using solar energy. One of the big advantages of liquid fuels like gasoline, diesel, or jet fuel is the large amount of energy contained in a small mass—much more energy per gram than batteries. Just recently, in fact, one such research group announced that they had produced a jar of jet fuel, starting from sunlight and CO₂.

Taking out the quotes and the hyperbole about revolutionizing anything, what they've done is still pretty neat: turned CO₂, the low energy end state of carbon-based fuel combustion, back into usable fuel.

Because CO₂ is the low energy end state, to get it back into a high energy form such as kerosene (jet fuel) or diesel, a whole lot of energy has to be put into it. In this case, the energy is solar.

Well, the energy is intended to be solar.

Because it's a lab scale test and proof of concept, rather than use the sun and a bunch of mirrors to concentrate the sunlight and heat up the reactor, they used an array of Xenon arc lamps with a bunch of mirrors to simulate sunlight in a controlled and repeatable way. This lets them be certain that a change in productivity is due to a change in reactor design, instead of a change in the sunlight input.

The process in question is the first step of the CO₂ to fuel sequence: turning CO₂ and water vapour into CO and H₂. Note that in both cases, the product contains one less "O" than the feed.

The mixture of CO and H₂ is called syngas, and producing kerosene and diesel from syngas has been around for a while: Germany used it on an industrial scale in WWII. The typical source of syngas, and the one Germany used, is coal. In recent years, waste biomass has also been used to make syngas.

This reactor, however, uses CO₂ as its carbon source.

The basic process involves heating a cerium oxide catalyst to about 1800K to drive off the oxygen, then letting it cool somewhat to 1300K and exposing the de-oxygenated cerium oxide to the CO₂ and H₂O mixture, at which time the cerium steals some oxygen atoms from those molecules, converting them to CO and H₂, respectively.

Chemically it sounds simple, but getting it to work in a useful and consistent way is always the catch. One "catch" described then resolved in papers from 2012 and 2014 is that the catalyst, being heated by concentrated sunlight, is only heated where the sunlight hits it; deeper sections of the catalyst do not get heated up and so don't give up their oxygen, meaning there's no drive to steal an oxygen from the CO₂. By changing the shape and structure of the catalyst, the researchers got it to heat more evenly with the solar power, reducing both cool spots and hot spots.

There are other things that make it not as simple as it sounds. (I'm sure I've missed a few, not being connected to the research; likewise I may have listed some things that they plan to or have already addressed. These are just things my engineering brain thought of.)

One is that this cannot run on the CO₂ and water present in air; if any oxygen is mixed in with the CO₂, that will be used up by the catalyst before it uses the CO₂. The feed has to be fairly pure CO₂ and water; the lab test setup used compressed purified gases with non-reactive argon as a carrier gas to push everything else through. This means that in order for this to work, there has to be some source of pure CO₂. Maybe the gases the carbon capture and storage people are dealing with could be a candidate feed stock, but vehicle (including aeroplane) exhaust is not likely as a source.

Using a carrier gas means that has to be separated from the syngas product at some point and preferably recycled, otherwise buying carrier gas gets expensive. The researchers mentioned potentially using vacuum to pull the gases through the reactor instead of a carrier gas to push them.

The amount of sunlight needed to get enough heat to drive the de-oxygenation phase is impressive: about 3000 suns, or sunlight concentrated by 3000 times. Existing solar concentrators that use solar energy for heat, such as solar power towers, concentrates sunlight by up to 600 times. That's a pretty big gap. I don't know if the towers stopped at 600 because of a technology limitation or because 600 was all they needed.

There is probably still lots of room for efficiency improvements, but the test setup used 3.8kW for about 55 minutes based on the temperature graph; the text indicates 30 minutes held at the maximum temperature but the "heater" is on for some time before that to get the system up to temperature. That amount of power input produces an average of 5.4mL of CO gas and an amount of H₂ based on the ratio of steam added. 5.4mL of gas is 5.4mL * 1mmol/22.4mL = 0.24mmol CO.

Since jet fuel is a mixture of molecules ranging from 8-16 carbons (see chapter 4), that means about: 0.24mmol/12 = 0.02mmol * 170mg/mmol = 3.4g of jet fuel is produced for that energy input, assuming an average of 12 carbons per molecule to represent the range 8-16, straight-chain C₁₂H₂₆ for the molecular weight, and 100% conversion of CO to fuel.

For comparison, burning 3.4g of jet fuel will provide: 3.4g * 43.28 kJ/g = 147kJ of energy. The 3.8kW for 55min works out to: 3.8kJ/s * 60s/min * 55min = 12,540kJ

So, roughly 85 times more energy had to be put into the system than can be had by burning the product, and that's before the energy use of all the equipment involved in the stages before and after.

Now, I also don't know whether scaling this system up to produce more syngas will be linear (energy in/energy out ratio stays the same) or whether the scale changes will change that ratio and if so in what ways. Some things get more energy efficient as they get bigger.

Obviously, to be viable, this system can't be run using hydrocarbon fuels to provide the heat. It has to run on solar or some other source of energy that doesn't produce CO₂, and even then the question remains: is this more or less efficient than solar-to-electric? There do exist electrically powered aeroplanes, and they're less theoretical than this technology; Airbus recently demonstrated one and apparently intends to eventually have 80-passenger hybrid electric regional jets.