What is the chemistry behind this “petrol from air” technology?

A recent news report in the UK claimed a breakthrough in making a petrol equivalent from carbon dioxide and water:

A small British company has produced the first "petrol from air" using a revolutionary technology that promises to solve the energy crisis as well as helping to curb global warming by removing carbon dioxide from the atmosphere.

It sounds like the news reports are over-hyping it a little (and I'm not sure it is that new), so what is the real chemistry behind it? What are the key reactions and how much input energy is required?

• Higher rep users might want to think of useful tags for this one. I would have liked something about industrial processes and maybe catalysis but can't yet create tags. – matt_black Oct 19 '12 at 14:04
• The energy required is certainly more than the amount you'll get out by burning the result, you can't cheat the laws of thermodynamics. – Mad Scientist Oct 20 '12 at 9:45
• The gist is: water + electrolysis -> hydrogen; hydrogen + CO2 from air -> methanol; methanol + gasoline fuel reactor -> petrol – 410 gone Oct 20 '12 at 9:55
• @EnergyNumbers that would make a great answer if you just summarised the chemistry of the key steps (eg what mechanism, conditions and catalysts covert hydrogen and CO2 into methanol). – matt_black Oct 20 '12 at 10:00
• @matt_black unfortunately that's pretty much all I know. AIUI they were planning to go from CO2 to CO with a reverse water-gas shift, and then add Hydrogen in a Fischer Tropsch to make the hydrocarbons. But I've seen no tech lit from them, and I guess they're playing their tech cards very close to their chests. – 410 gone Oct 20 '12 at 10:30

Step one: trap $\ce{CO2}$. This can be done using specially selected amines, like triethanolamine. Aminosilicones (a common chemical used in hair conditioner), can also do this.

Step two: release trapped $\ce{CO2}$. In case of triethanolamine you'll have to boil mixture.

Step three: mix $\ce{CO2}$ with hydrogen (hydrogen is produced by electrolysis of water). Volume proportion should be somewhere 1:3.

Step four: threat the mixture with proper catalyst. This step may be divided into two stages: reverse water-shift reaction and some variant of Fischer–Tropsch process with some type of alcohol synthesis as very interesting candidate. There are some multi-functional catalysts that can perform both stages at once.

Another variation on step 4 is the Sabatier process, which involves generating methane from $\ce{CO2}$ and hydrogen. Typically the catalyst is Rainy nickel (powder).

The required energy in form of electrical current is probably from 1.5 to 3 times of heat of combustion of resulting fuel, which is in turn at least twice more that can be converted into actual work again.

• While your estimate of energy efficiency sounds reasonable to me, I'd highly appreciate pointers to literature where such estimates are given - that would help tremendously in discussions... – cbeleites unhappy with SX Nov 4 '12 at 21:08
• @cbeleites well... internal combustion engines have energy conversion efficiency 30-50% and the process is essentially fuel combustion reversed, so at the very least all energy released in combustion must be provided at the time of process. However, some energy is lost during $H_2$+CO_2 -> CH_x + H_2O\$ part to shift chemical equilibrium. 30% loss is actually a point-blank shot, but process with 10% losses is actually a very good one and we have at least three steps here. Upper estimate have no actual foundation under it except unbelief that something with ridiculous losses may be used. – permeakra Nov 5 '12 at 17:34
• I agree with the internal combustion engine efficiency; but I was wondering at the synthesis part. Anyways, thanks. I'm not so sure about not having ridiculous losses, though (given what I found here chemistry.stackexchange.com/a/1361/160). I had a quick look at coal liquefaction (which is not the same, but possibly related in difficulty, and has a rather long industrial history): aspo2012.at/wp-content/uploads/2012/06/…. So: 60-70% theoretically possible, existing plant reaches 20-28%. Ouch... – cbeleites unhappy with SX Nov 5 '12 at 18:19