In a recent presentation I was asked a question I found impossible to answer. And after some hours of work, I’m still far from certain my response is correct. But I thought I’d share the analysis even though the answer - get rid of coal in power generation - is probably obvious.

The issue raised is crucially important: of the various transitions in technology we are trying to engineer, which will reduce emissions the most for every unit of extra renewable electricity generation? The world is trying to ‘electrify everything’ but which applications should be given the first priority when we add extra wind and solar?

The alternatives are numerous. Should the world use new renewables to reduce the amount of electricity generated by coal or gas? Or would it better to speed up the growth of EVs to use the extra renewables? How does using the electricity to generate hydrogen for decarbonising steelmaking compare? Or making ammonia as a fuel for ships’ engines? What about the impact of increasing the use of electricity for domestic heating? Or producing hydrogen for fuel cell use in heavy transport? Or manufacturing synthetic fuels using electricity for use in aviation?

Of course we eventually need to stop using fossil fuels across all the energy system and transfer to renewable electricity. But the rate of decline towards eventual net zero also matters. If we decarbonise the most polluting activities first the amount of CO2 eventually in the atmosphere will be lower than if extra electricity replaces fossil fuels in sectors with low emissions per unit of energy. The purpose of this analysis is therefore to suggest which sectors the world should be pushing towards renewable electricity into first, whether in the form of electrons, hydrogen, ammonia or synthetic fuels.

I have tried below to calculate the emissions reduction that arises from applying one incremental megawatt hour of zero-carbon electricity to each of the main alternative uses.

The numbers are approximate; they will also vary somewhat across the world as well from operator to operator. For example, if the extra electricity was used to make hydrogen enabling an inefficient old steel furnace to be closed, it would reduce emissions more than if the manufacturing site was more recently built. Or to use an alternative illustration, a heat pump for a domestic house will save far more emissions if it replaces oil fired central heating than a natural gas boiler.

Here is my league table, suggesting an order of priority for investment in carbon-saving technologies.

The implications of this table seem to be clear. In general, we should be focused first on using the growing amount of renewables in world electricity systems to decarbonise activities which are themselves inefficient users of fossil fuels. So, for example, our extra unit of clean electricity can be used to provide the power for a new EV. A new petrol car will only convert about a quarter of the energy in the fuel into useful motion but an electric car is more efficient. The new renewable electricity therefore offers real leverage in reducing emissions.

The same is true for a heat pump: a central heating boiler offers good transfer of the energy in gas into heat but a heat pump can actually use a unit of electricity to transfer several times as much heat into a house. (I’ve used a typical UK figure of a 2.8 times multiple). A coal-fired power station only works at about 40% efficiency, the energy value of the coal compared to the electricity output. That is why it comes out top of the table.

These top three uses all offer CO2 savings of between 600 and 900 kilogrammes per megawatt hour of electricity produced.

We then move to applications will employ the new electricity to create hydrogen for use in other processes. Hydrogen in steel making replaces the use of coal. Making hydrogen from electricity sees substantial energy loss but the gas is somewhat more efficient than coal in reducing iron ore to liquid iron. So this application is relatively productive. Similarly, there are advantages and disadvantages in using electricity to make hydrogen for use in a fuel cell in a heavy vehicle. These activities offer carbon savings of around 400 kg per megawatt hour of electricity.

At the bottom end of the range are those uses which involve the conversion of electricity into hydrogen and then through a second conversion into ammonia or synthetic fuel. Here, the savings can be as little as just over 100 kg per megawatt hour. The emissions reduction value is therefore about an eighth of the gain if the new electricity is employed to reduce coal-fired electricity output.

The key lesson is that there are real differences in CO2 reductions from different uses for new renewable power. Hydrogen comes lower in the list of priorities than getting coal off the grid. This conclusion must be qualified; when an electricity system has genuine surpluses of supply, an increasingly common phenomenon, making hydrogen is far better than simply disconnecting the generation capacity.

(Thanks to Tim Elliott of Regal Funds Management of Australia for the question and for his patience later discussing these results. Errors are all mine).

Appendix: The key inputs into each calculation

1, Reducing coal use in power generation.

1 MWh of new renewable electricity replaces 1 MWh of coal fired power.

In typical power station, 1 MWh produces 900 kg of CO2

2, Using the power for EVs.

A new EV takes in about 0.85 MWh of electricity (accounting for battery losses) from 1 MWh of new renewables production.

A new EV will typically travel about 6 km for each kWh of battery power used. So 1 MWh of new electricity production will enable a journey of 6*850 km, or about 5,100 km.

A new ICE car would typically emit about 130g per km.

So the saving would be 130g multiplied by 5,100 or 663 kg.

3, Electrifying heating using heat pumps.

1 MWh of electricity delivers 2.8 MWh heat into a building using a heat pump (approximate UK average).

This typically would reduce the consumption of gas by about 3.2 MWh. (The boiler is not 100% efficient).

This would have produced about 650 kg of CO2.

4, Making hydrogen for steel manufacturing.

A tonne of new steel made today typically results in emissions of about 1.85 tonnes

Steel made using hydrogen will probably require about 3 MWh of energy in the form of CO2.

At electrolyser efficiency of 67%, about 4.5 MWh of electricity is needed to make the H2 for a tonne of steel.

So 1 MWh of electricity would save about 1.85 tonnes divided by 4.5, or about 411 kg of CO2.

5, Making hydrogen for a fuel cell in heavy vehicles.

1 MWh of electricity makes hydrogen with an energy value of about 670 kWh, assuming an electrolyser operating at 67% efficiency.

The conversion back to electricity from a hydrogen fuel cell to power the battery in the truck is about 60% efficient. The electricity available for travel is therefore about 402 kWh for each 1 MWh of electricity initially produced.

A truck is assumed to be 25% efficient at converting the energy in diesel into power available for travel. Therefore to be equivalent to the travel power delivered by electricity, the truck would use 1608 kWh of diesel.

At 10.6 kWh per litre of diesel, the truck would need 152.3 litres of fuel.

Each litre of fuel produces about 2.5 kg of CO2. This means that the switch to fuel cell truck from a diesel truck would save 381 kg of emissions for each 1 MWh of electricity used to generate hydrogen.

6, Reducing gas use in power generation.

1 MWh replaces 1 MWh gas fired power.

In typical gas power station, 1 MWh produces about 330 kg of CO2.

This excludes fugitive methane losses at point of production or in transport.

7, Making synthetic fuel for aviation rather than kerosene

About 20 kWh of electricity is needed to make 1 litre of fuel. (Plus about 12 kWh of heat, which is assumed to be free). The source for this estimate is Norsk eFuel.

So 1 MWh of electricity will produce 50 litres of eFuel.

2.5 kg of CO2 arise from each litre of aviation kerosene.

So 125 kg of CO2 is saved for each of I MWh of electricity devoted to making synthetic aviation fuel.

8, Using ammonia instead of heavy fuel oil in shipping.

Each tonne of ammonia requires energy of 9.15 MWh. 1 MWh of electricity will therefore make about 109.3 kg of ammonia.

The energy content of ammonia is about 5.2 MWh per tonne. 1 MWh of electricity will therefore make ammonia with an energy value of 565 kWh.

565 kilowatt hours has the energy equivalent of about 45 kg of Heavy Fuel Oil (HFO). (Key assumption that ammonia engines and HFO engines are equally energy efficient)

A litre of HFO produces about 2.5 litres of CO2. The replacement of HFO by the ammonia produced using 1 MWh of renewable electricity therefore saves about 113 kg of emissions.