Coal Power

Coal is still the world’s largest single source of electricity. It generated about 10,500 TWh in 2025, a third of global supply, despite the rapid growth of solar and wind. Coal is also the largest single source of energy-related CO₂, roughly 15 billion tonnes a year, and power plants account for the biggest share. Where coal is mined locally and the plants are paid off, coal power is cheap to run. It is dispatchable, and whole mining economies depend on it. Per unit generated, it is also by far the deadliest source of electricity, because of air pollution.

How it works

A coal plant is a steam (Rankine-cycle) power station. Coal is crushed to a fine powder in mills, blown into a boiler with preheated air and burned at about 1,300–1,700 °C. Heat passes to water in thousands of metres of boiler tubes. The high-pressure steam spins a multistage turbine coupled to a generator, then is condensed back to water and pumped round again. The hotter and higher-pressure the steam, the higher the efficiency, and that is how the main classes of plant are defined.

Emissions per kWh depend directly on efficiency and on the coal’s carbon content:

The same arithmetic gives coal consumption. At 25 MJ/kg, the 38%-efficient unit burns about 0.38 kg of coal per kWh, or roughly 3.3 million tonnes a year for a 1 GW unit at 100% output. The fuel energy content calculator compares heating values for hard coal, lignite and other fuels.

Main variants

Above about 22.1 MPa and 374 °C, water passes its critical point and becomes steam without boiling. Supercritical plants use once-through boilers built around this. Ultra-supercritical units go past 600 °C, and the best of them, in Japan, Denmark and China, report net efficiencies of around 44–47% (LHV). Every step up needs more expensive high-alloy steel.

IGCC (integrated gasification combined cycle) turns coal into syngas, cleans it and burns it in a gas turbine with a steam bottoming cycle, which makes pre-combustion CO₂ capture easier. Demonstration plants at Buggenum (Netherlands), Puertollano (Spain), Polk and Wabash River (US) and Edwardsport (Indiana) had mixed availability. In Mississippi, the Kemper County project gave up on gasification in 2017 after costs climbed to about $7.5 billion; it now burns natural gas.

Lignite versus hard coal

Lignite (brown coal) has a heating value of about 5–15 MJ/kg and contains 30–60% moisture. Shipping it doesn’t pay, so it is burned in mine-mouth plants in Germany, Poland, the Czech Republic, Greece, the Balkans, Türkiye, Australia (Victoria) and Indonesia. Boiling off the water costs efficiency, and lignite carries more carbon per unit of energy, so lignite units typically emit 1.0–1.2 kg CO₂/kWh. The open-pit mines that feed them disturb more land than almost any other energy source. Hard coal (bituminous, sub-bituminous and anthracite, roughly 20–32 MJ/kg) is traded by sea. Indonesia and Australia are the largest exporters, and China, India, Japan and Korea the largest importers.

Efficiency and capacity factor

Coal units were built for baseload operation at a 70–85% capacity factor, but in many markets they now run far less. Squeezed by cheap gas and renewables, US coal plants averaged roughly 40–50% in recent years (EIA). China’s fleet ran at around 50% because capacity grew faster than demand. Ramping a coal plant up and down increases wear, lowers efficiency and pushes up emissions per kWh. Minimum stable output is typically 30–50% of rated power, although retrofits in Germany, Denmark and China have pushed it lower. The capacity factor calculator translates these utilisation rates into annual energy.

Costs

• Lazard LCOE+ 2024 (US, unsubsidised, new build): $69–168/MWh. Few new coal plants are proposed in the US because new wind, solar and gas are cheaper.
• IEA/NEA Projected Costs 2020: new coal is typically in the $75–120/MWh range at a 7% discount rate in OECD countries. The figures include an assumed $30/t carbon price, which adds about $25–30/MWh. Plants in China and India are built for far less per kW.
• NREL ATB / EIA AEO: overnight costs for new US pulverised coal plants of roughly $3,500–4,500/kW, and much higher with 90% capture.

For existing plants that are already paid off, the marginal cost is what counts: fuel, variable O&M and carbon. At $100/t delivered hard coal (about $4/GJ), fuel costs about $38/MWh. A carbon price of €80/t adds about €70/MWh. The EU ETS is the main reason European coal generation has roughly halved since 2015. Analysts including Carbon Tracker have found that operating many existing coal plants now costs more than building new renewables. The LCOE calculator shows how fuel and carbon prices shift the result.

Land, water and materials

Coal moves a lot of material and takes a lot of land. A 1 GW plant burns roughly 3 million tonnes of hard coal or 8–10 million tonnes of lignite a year, leaving hundreds of thousands of tonnes of ash and, where scrubbers are fitted, gypsum. Coal ash ponds have failed catastrophically, for example at Kingston, Tennessee in 2008. Mountaintop removal in Appalachia and lignite pits in Germany’s Rhineland and Lusatia have removed whole landscapes and villages. Water use is high. NREL’s survey (Macknick et al., 2012) gives median consumption of about 690 US gallons per MWh for tower-cooled coal plants, and once-through cooling withdraws far larger volumes. In water-stressed India and northern China this has become a siting constraint.

Lifecycle CO2 emissions

• IPCC AR5 WGIII Annex III (2014): pulverised coal median 820 gCO2e/kWh, range 740–910.
• UNECE 2021: roughly 750–1,100 gCO2e/kWh without capture, depending on coal type and region. With CCS the range is about 150–470 gCO2e/kWh, because capture is never 100% and upstream mine methane remains.

Coal-mine methane is a large and often underestimated contributor. The IEA’s Global Methane Tracker estimates coal mining’s methane emissions are comparable to those of the oil or gas sectors. For comparison, gas combined cycle emits about 490 g, and nuclear and wind emit about 10 g.

Safety and health impacts

Most of the harm from coal travels through the air. Burning it releases fine particulate matter (PM2.5), sulfur dioxide and nitrogen oxides (which go on to form secondary particulates and ozone), plus mercury, arsenic and other trace metals. The result is heart disease, stroke, lung cancer and respiratory illness hundreds of kilometres downwind.

• Markandya & Wilkinson (Lancet, 2007): about 24.5 deaths per TWh for European hard coal and 32.6 for lignite, plus many more serious illnesses.
• Our World in Data synthesis: about 24.6 deaths per TWh for coal, against 18.4 for oil, 2.8 for gas and about 0.03 for nuclear, wind and solar.

Where pollution controls are fitted, the toll drops sharply. Flue-gas desulfurisation removes 90–98% of SO₂, selective catalytic reduction 80–90% of NOx, and electrostatic precipitators or fabric filters more than 99% of particulates. Activated-carbon injection captures mercury. China’s “ultra-low emission” standard (about 10/35/50 mg/Nm³ for dust/SO₂/NOx), applied to most of its fleet since 2015, contributed to a large fall in urban PM2.5. Many plants in India and Southeast Asia still lack FGD, and Indian compliance deadlines have slipped repeatedly. Mining adds its own toll from accidents and black lung disease.

Carbon capture on coal

Two flagship projects show what capture can do and how hard it is:

• Boundary Dam 3 (SaskPower, Canada, 2014): a 110–115 MW net retrofit designed to capture about 1 Mt CO₂/yr at 90%. It has run reliably in later years but has typically captured well below design. Much of the CO₂ is sold for enhanced oil recovery.
• Petra Nova (W.A. Parish, Texas, 2017): captured CO₂ from a 240 MW-equivalent slipstream, designed for about 1.4 Mt/yr, for oil recovery. It was mothballed in 2020 when oil prices collapsed and restarted in 2023 under new ownership.

Capture consumes 20–30% of a plant’s output for solvent regeneration and compression. Combined with capital costs, it raises generation cost by roughly 40–80%, which leaves CCS coal more expensive than most low-carbon alternatives.

Retirement trends and outlook

Coal power is now heading in two directions depending on where you look. Britain’s coal share fell from about 40% in 2012 to zero when Ratcliffe-on-Soar closed on 30 September 2024. US coal generation has fallen by more than half from its 2007 peak of about 2,000 TWh, and coal’s share is now below 20%. The EU’s coal output has roughly halved since 2015. Germany plans to exit by 2038 at the latest, and Spain, Italy and others are earlier.

In Asia, coal’s share is falling while absolute output has kept rising. China generates over half of the world’s coal power and approved roughly 100 GW of new coal capacity a year in 2022–2023, according to CREA and Global Energy Monitor. Officially those plants are increasingly positioned as backup for renewables rather than baseload. India gets roughly three-quarters of its electricity from coal and plans further additions to meet demand growth. The IEA’s Coal 2024 report expects global coal demand to plateau through the late 2020s, with Chinese decline offsetting growth in India and Southeast Asia.

How fast coal goes depends on cheap replacement capacity (solar, wind, battery storage and natural gas), carbon pricing, financing for early retirement such as the JETP deals in South Africa, Indonesia and Vietnam, and support for coal-mining regions. The power plant map shows the world’s remaining coal fleet, and the technology comparison sets its costs and emissions against the alternatives.