WTF? Has Wimberley finally taken leave of his diminishing senses? Why laud a commonplace and unpleasantly acrid standard chemical, used to make fertilizer and explosives, and still popular as a cleaning agent with traditionalist housewives in Spain and Brazil? (The nasty smell tells germs you mean business.)
Hear me out, kind readers. Ammonia is about to take its place as a worthy piece of the complex jigsaw puzzle of the energy transition.
This will be down to its additional potential use as a carbon-free fuel. Burn ammonia, in an engine or fuel cell, and you ideally get:
4.NH3 (ammonia) + 3.O2 (oxygen) → 2.N2 (nitrogen) + 6.H2O (water)
In an engine, in practice you also get some nasty nitrous oxides, NOX: controllable by clever engine management, scrubbing the exhaust with more ammonia, or reforming to hydrogen just before burning.
Ammonia is not a greenhouse gas, and nor are its main combustion products. So it’s a candidate for a storable renewable fuel to replace oil-based liquid fuels and natural gas. The rivals are plant-based liquid biofuels (diesel, ethanol, kerosene), biogas from digesters fed with biomass, and catalytic hydrogen.
In this competition it has some attractive technical characteristics. Ammonia’s energy density is half that of the hydrocarbons, but twice that of hydrogen. This is a major issue for aircraft, but not really for the other major potential uses: heavy trucks, ships, and power stations.
Its boiling point is -33° C. Compare butane: – 12° C; propane: – 42° C; methane/natural gas: – 164° C; hydrogen: – 253° C. The last two require very expensive refrigeration plants to reach the liquid state, and expensive refrigerated transport to stay in it. Ammonia, like butane and propane, can be pumped around as a gas at ambient temperature, or liquefied by standard compressors and transported in simple pressurised containers, again at ambient temperatures. It’s clearly a better bet all round than hydrogen. Methane has the benefit of incumbency: a massive legacy piped distribution system, even to many houses. But methane is a greenhouse gas, in the short run twenty times as powerful as CO2, and leaks can’t be avoided.
[Update 2/1/2019]Ammonia does not emit greenhouse gases at all. The others burn sustainably sourced hydrocarbons; the CO2 is recycled back into plants. These are not morally identical. Partly it’s the time lag. Partly it’s because net zero won’t be enough, and we need to enlist the biosphere as a net carbon sink, through reafforestation and burial. [/update]
The existing ammonia industry of about 150 MT a year may be small by oil standards, but it is quite big enough for the handling and distribution technology to be reliable and mature. It doesn’t need anything new to become a fuel.
The main lack is engines and fuel cells to burn it. You can’t go out and buy a marine diesel designed to burn ammonia either by itself or in a dual-fuel configuration. This is not SFIK a major engineering challenge – create a demand, and the engines will come quickly.
The main factor militating against ammonia is cost. Ships in particular burn a vile sludge called heavy bunker fuel, the residue left after all the good fuels have been refined away. It’s cheap, and likely to stay that way. Besides, there is just now no climate reason to go for ammonia. It’s produced from oil and gas by the Haber-Bosch process, which vents the carbon as CO2 anyway.
Things are changing here. A team in Japan have built a working pilot plant using catalysed hydrogen as the feedstock. So a renewable production chain for ammonia now looks feasible: and hence a pathway to sustainable decarbonised shipping.
Cost? I’ve no idea, and ruthenium catalysts are presumably expensive. But cheapness is not a reasonable demand to make of a process at this exploratory stage. The normal pattern is that once a technology works, it can be made cheaper with experience and scale.
How does this fit into the wider energy transition? Very nicely, but it needs some background.
Let’s do a crude thought experiment for the UK electricity supply. Recall that the baseline 100%-renewable electricity scenario is now wind, solar, transmission, and pumped hydro storage (see Blakers for Australia). Simplify this even more, and take an initial solution for the UK without solar, which we will add back if it’s cheaper. UK peak electricity demand is 48 GW in winter. Allowing a 50% capacity factor, we need 96 GW of turbines, a doable increase from the 20.5 GW today. But there are lulls even in winter. The late David Mackay did some back-of-the-envelope math in 2008 for a 5-day maximum lull and 33 GW of wind at 33% CF, and the pumped storage needed was 1,200 Gwh. Scaling up we need 5,760 Gwh. It’s a mere 30 Gwh today, so it’s a truly massive expansion, so large as to look impossible.
There are plenty of ways to cut this. The storage can be in Norway, which has a very large supply of steep, wet and unpopulated mountains. The price is adding undersea cables at £1.4 bn per GW capacity, on top of what you pay the Norwegians for the electricity. You can add solar, whose variation is mainly diurnal and on longer timescales is uncorrelated or inversely correlated to wind. (BTW, if you can handle the variation in a large wind park, adding a more consistent solar one logically poses few problems. The “solar needs storage” meme is largely bunk.) You can hook up underused car batteries, V2G in the trade: range anxiety by purchasers will ensure massive excess capacity. There’s demand response (contracts to cut usage on request), and intelligent management of water heaters, a/c, and freezers. Finally, there is conversion to storable fuel, which is where we came in.
You can also tackle the issue by just building more wind turbines at a modest £1.5 bn per GW onshore (IRENA, pdf page 94). This will lower capacity factors. But so what? On the LCOE cost metric, which assumes full takeoff, wind is already competitive with fossil generation in most countries and much cheaper in some, including the USA. There is a steadily growing cost gap that allows ever higher curtailment. 2.5c/kwh LCOE with a drastic 50% curtailment is still only an acceptable 5c/kwh net.
Whatever mix is settled on for firming a wind-and-solar dominated electricity supply, it will very probably include significant overbuild of nominal capacity against peak demand, and massive overbuild against trough demand. I don’t think 2 x is unrealistic economically.
This means that in any such world, there is certain to be large oversupply for significant periods when the wind is blowing hard, or the sun shining from a clear summer sky. This electricity will be practically free at the point of production, and available for the low cost of transmission.
Similar but more professional thought experiments have been carried out elsewhere. These use hour-by-hour simulations of grid demand, and recently they have been coming up much more cheerful than Mackay. A lot of money and effort has been going into P2X projects, especially in Germany (pdf). The acronym stands for “power to (some) synthetic fuel.” The working assumption is that there will in due course be a lot of very cheap surplus wind and solar power to use.
The cost-benefit calculations will get complicated. My guess is that it’s a race between green ammonia and green methane, with pure hydrogen a distant third because of the transport problem. The crucial variables are whether methane is taxed to pay a proper carbon price for its leaks, and the comparative efficiency of the upgrade reformers. On a level playing field, ammonia looks pretty good.
Wonkish note on methodology
My thought experiment is crude and many of the numbers are guesses, so don’t take the results too seriously. However, I defend the method. What I remember from a brush with linear programming years ago is that you start an optimisation with a simple non-optimal solution, and tweak the variables from there until it looks good. This is not guaranteed to work but it normally does. Under uncertainty, it is sensible to start with a solution of known feasibility, say technologies x and y. This allows you to tweak with hypotheticals: what if technology z also works? In my example x is wind, y is solar, and z is synthetic green ammonia.
Nuclear power offers a case where the technologies are not well-behaved and there may be several local optima. If you start with wind and solar, adding nuclear doesn’t help, as you want despatchable firming not baseload. Start with all nuclear, and adding wind and solar doesn’t help, as they aren’t despatchable backup. They aren’t complements at all, which helps explain the bitterness of the disputes around nuclear (see Jacobson v. Clack). It’s nuclear or renewables, not both. These disputes are now academic in the pejorative sense, as nuclear power is hopelessly expensive and slow to build.