Sadi Carnot: The Theoretical Limits of Heat Engines

In June 1824 a 28-year-old French engineer paid for 600 copies of a short book and asked a question nobody had asked properly: how much work can you get out of heat, at most? Sadi Carnot's answer was that the limit depends on temperatures and on nothing else, not on the fuel, the working fluid, or the cleverness of the machine. Every steam engine, car engine, jet, and power plant built since runs under that ceiling. The limit is not a failure of engineering. It is a property of how heat behaves, and it led by a straight line to the second law of thermodynamics and to entropy.

The problem Carnot was trying to solve

By the 1820s, steam power was reshaping Britain, and French engineers were anxious about falling behind. Engines pumped water out of mines, drove mills, and were starting to pull carriages on rails. They also wasted most of the heat from their fuel. Improvements came from trial and error: tighter fittings, higher pressures, better boilers. Nobody could say whether those improvements had a destination or would simply keep going.

Carnot approached the matter the way an engineer-physicist would. Instead of studying one engine in detail, he asked what any engine must do. Whatever its parts, a heat engine takes heat from something hot, turns a portion of it into work, and gives the rest to something cold. His book, Réflexions sur la puissance motrice du feu ("Reflections on the Motive Power of Fire"), was published in Paris by Bachelier. It runs 118 pages, and Carnot paid for the printing of 600 copies (Wikipedia: Reflections on the Motive Power of Fire; Wikipedia: Nicolas Léonard Sadi Carnot).

The surprising part: Carnot reached his central result while believing something that turned out to be wrong. He treated heat as a conserved fluid called caloric. The argument survived the mistake.

What a heat engine does

A heat engine is any device that converts heat into mechanical work by cycling a working substance (steam, air, combustion gas) through a repeating sequence of states. Three things happen in every cycle:

1. The engine absorbs heat from a hot source, such as burning coal, a nuclear core, or burning gasoline.
2. It converts part of that heat into work, such as a turning shaft or a moving piston.
3. It releases the remaining heat to a cold sink, such as the atmosphere, a river, or a radiator.

Efficiency is the fraction of the absorbed heat that becomes work: work out divided by heat in. The question Carnot raised is why this fraction can never reach one, and what sets its maximum.

An analogy helps. Carnot himself compared a heat engine to a water wheel. A wheel gets work from water falling from a high level to a low one, and the work depends on the height of the drop as well as on the amount of water. In Carnot's picture, a heat engine gets work from heat falling from a high temperature to a low one. He used the French phrases chute d'eau and chute de calorique, the fall of water and the fall of caloric (Wikipedia: Reflections on the Motive Power of Fire).

The caloric part of the picture was wrong, because heat is not a conserved fluid. The temperature-drop part was right, and it is still the best intuition for why a bigger temperature gap lets an engine do more with each unit of heat.

The Carnot cycle, step by step

Carnot described an idealized engine whose cycle he argued could not be beaten. In the modern description, it has four steps, with a gas in a cylinder as the working substance. The same cycle was later drawn on a pressure-volume diagram by Émile Clapeyron in 1834, which is the form most textbooks use (Wikipedia: Carnot cycle).

Step	What happens	Heat flow	Gas temperature
1. Isothermal expansion	Gas expands slowly while touching the hot reservoir	Absorbs heat from the hot source	Constant at the hot temperature
2. Adiabatic expansion	Gas keeps expanding, insulated from everything	None	Falls to just above the cold temperature
3. Isothermal compression	Gas is compressed slowly while touching the cold reservoir	Releases heat to the cold sink	Constant at the cold temperature
4. Adiabatic compression	Gas is compressed further, insulated	None	Rises back to just below the hot temperature

The word "isothermal" means constant temperature, and "adiabatic" means no heat is exchanged with the surroundings. On a pressure-volume diagram, the area enclosed by the four curves is the work done per cycle. On a temperature-entropy diagram the cycle becomes a rectangle, because isotherms are horizontal lines and adiabats are vertical ones, and the enclosed area is the heat converted to work.

Why the cycle must be reversible

Every step in the Carnot cycle is carried out so slowly that the system is always almost in equilibrium. Heat only flows between objects whose temperatures differ by a vanishingly small amount, and there is no friction or turbulence. Such a process is called reversible: run it backward and everything returns exactly to where it started, with no leftover change in the surroundings.

That requirement is what makes the cycle the best possible and also what makes it impossible to build as a practical machine. A reversible engine has to run infinitely slowly, so it produces no power in any finite time (Wikipedia: Carnot cycle). Think of it as a speed limit that no car can reach: not a car anyone expects to drive, but a number that tells you how far a real car sits from perfection.

Running the cycle backward

Because every step is reversible, the Carnot cycle can be run in reverse. Work goes in, heat is pulled from the cold reservoir, and a larger amount of heat is delivered to the hot one. That is a refrigerator or a heat pump. The same logic that caps engines also caps refrigerators and heat pumps, which is why the same temperature ratio shows up in both.

Carnot's theorem: the claim that matters

Carnot's central result, in modern wording, has two parts:

• No heat engine operating between two heat reservoirs can be more efficient than a reversible engine operating between the same two reservoirs.
• All reversible engines operating between the same two reservoirs are equally efficient, regardless of what they are made of or what substance they use.

The second part is the startling one. It means efficiency cannot depend on whether the working fluid is steam, air, alcohol vapor, or something exotic. Only the two temperatures can matter. If it were otherwise, you could couple a more efficient reversible engine to a less efficient one run backward and extract net work from one heat reservoir alone, with nothing else changing. That is the sort of machine that a correct physical theory should forbid, and it is the core of the argument that Carnot's followers turned into the second law.

The formula, and who wrote it down

The formula usually credited to Carnot is:

Maximum efficiency = 1 − (T_cold / T_hot)

with both temperatures measured on an absolute scale such as kelvins (K). The history is more tangled than the textbook credit suggests. Carnot's 1824 text did not contain this formula, because the absolute temperature scale did not yet exist. William Thomson, later Lord Kelvin, proposed an absolute scale in 1848, building it on Carnot's reasoning, and the clean formula came out of that work and the later work of Clausius. Carnot's own estimate of absolute zero was about −267 degrees on the Celsius scale, compared with the accepted −273.15 (Wikipedia: Lord Kelvin).

Here is what the formula says in numbers, for an engine whose cold side sits at 300 K (about 27°C, a warm room or a cooling pond):

Hot side (K)	Hot side (°C, approx.)	Maximum efficiency
400	127	25%
600	327	50%
900	627	67%
1,500	1,227	80%

Three consequences follow directly.

• No engine reaches 100%. That would require a cold side at absolute zero, which cannot be reached in practice.
• Only the temperatures matter. The working fluid drops out of the formula entirely.
• The gap matters, and the cold side counts more than people expect. Take a 600 K hot side and a 300 K cold side, for a 50% ceiling. Raising the hot side by 50 K lifts the ceiling to about 53.8%. Lowering the cold side by 50 K lifts it to about 58.3%. A given temperature change at the cold end buys more than the same change at the hot end (Wikipedia: Carnot cycle). The cold side is usually the environment, which is why power plants are built beside rivers and the sea.

How real engines measure up

The honest way to use the Carnot limit is as a ceiling: real engines fall below it, and the gap shows how much is lost to friction, heat leaks, finite-speed processes, and the limits of materials. The table below uses the temperatures from a standard comparison of three real plants. The "endoreversible" column is a second, more realistic estimate explained in the next section.

Plant	Cold side	Hot side	Carnot limit	Endoreversible estimate	Observed
West Thurrock (UK, coal)	25°C	565°C	0.64	0.40	0.36
CANDU (Canada, nuclear)	25°C	300°C	0.48	0.28	0.30
Larderello (Italy, geothermal)	80°C	250°C	0.33	0.178	0.16

Source: the comparison table in Wikipedia: Endoreversible thermodynamics, which cites Curzon and Ahlborn (1975).

The Carnot values are a long way above what the plants actually deliver. A coal plant with a 565°C boiler and a 25°C cooling water supply has a ceiling around 64%, yet the plant achieved about 36%.

A real fleet, measured in the United States

Power plant performance in the US is reported as a heat rate, the amount of fuel energy in British thermal units (Btu) burned per kilowatt-hour of electricity delivered. One kilowatt-hour is equal to 3,412 Btu, so dividing 3,412 by the heat rate gives efficiency. The US Energy Information Administration reported these average operating heat rates for 2024 (EIA, Electric Power Annual, Table 8.1):

Plant type (US average, 2024)	Heat rate (Btu/kWh)	Implied efficiency
Coal	10,777	about 31.7%
Natural gas (all plants)	7,754	about 44.0%
Nuclear	10,443	about 32.7%

The efficiency column is the author's calculation from the EIA heat rates and the standard conversion of 3,412 Btu per kilowatt-hour. The natural gas row averages every gas plant in the survey, from simple peaking turbines to modern combined-cycle plants, so it understates what the best new plants achieve.

Why gas plants do better than coal

Part of the answer is temperature. A modern gas turbine burns fuel at a much higher temperature than a steam boiler can tolerate, so its hot side is hotter. The gas leaving the turbine is still hot, however, and discarding it wastes the gain. A combined-cycle plant uses that exhaust to make steam for a second turbine. Think of two water wheels on the same stream: the second catches the water that falls past the first.

Reported efficiencies for such plants reach the low 60s. The certified record, awarded by Guinness, is 62.22%, at an EDF plant at Bouchain, France, in 2016, built around a GE 9HA turbine. The same turbine alone, without the steam cycle, was reported at 41.5% (Wikipedia: Combined cycle power plant). Efficiency figures come in two flavors, depending on whether the fuel's lower or higher heating value is used, so two numbers for the same plant can differ by several percentage points. A single steam (Rankine) cycle sits lower, with supercritical plants (steam at about 24.1 MPa and 538–566°C) reaching roughly 40% and ultra-supercritical plants (steam near 600°C) about 42% (Wikipedia: Rankine cycle).

Cars

The gap between ideal and actual is widest in the engine under a car's hood. The US Department of Energy reports that only about 12%–30% of the energy in a gasoline car's fuel reaches the wheels, depending on whether the driving is stop-and-go city traffic or steady highway cruising. In the city, engine losses (mostly heat) account for 71%–75% of the fuel energy. On the highway, they account for 64%–69% (FuelEconomy.gov: Where the energy goes).

A reference calculation shows what Carnot predicts for such an engine. A heat source at 816°C and an environment at 21°C give a ceiling of about 73% (Wikipedia: Thermal efficiency). Real gasoline engines achieve far less, because of friction, throttling, incomplete combustion, cooling needs, and the fact that the gas in the cylinder is not at that peak temperature for the whole cycle. The same source reports about 25% as typical for a gasoline automobile engine, and a peak of 51.7% for the largest diesel engine, the Wärtsilä-Sulzer RTA96-C. That is a peak figure for a ship engine, not a typical one. Diesels are generally more efficient than gasoline engines.

Why real engines miss the Carnot limit

The Carnot limit is not simply a high bar that engineering will someday clear. The gap has distinct causes.

• Speed. A reversible cycle takes infinite time. Real engines need power, so they run fast, and fast processes create temperature differences and turbulence that waste useful energy.
• Heat transfer across a gap. To push heat from a flame into a gas, the flame has to be hotter than the gas. To dump heat, the gas has to be hotter than the coolant. Those temperature differences are real losses relative to the idealized cycle.
• Friction and leaks. Moving parts lose energy to friction, and seals leak.
• Materials. Turbine blades and boiler tubes weaken at high temperature, so the hot side is capped by metallurgy, not by physics.
• The sink. The cold side is set by the environment, which an engineer cannot choose freely.

The more realistic ceiling: maximum power

In 1975, Frank Curzon and Boye Ahlborn published a short paper, "Efficiency of a Carnot engine at maximum power output," in the American Journal of Physics (43(1), 22–24). It asks what the efficiency would be if the engine were reversible inside, but had to exchange heat with the two reservoirs at a finite rate, as real engines do. The result is that the efficiency at maximum power is 1 − √(T_cold / T_hot), a smaller figure than the Carnot limit. The formula had been found earlier by Reitlinger (1929), Novikov (1957), and Chambadal (1957), but is most often attributed to Curzon and Ahlborn (Wikipedia: Endoreversible thermodynamics).

As the table above shows, this estimate tracks real plants more closely than the Carnot limit does, within about four percentage points for all three. The point of the comparison is not that Carnot was wrong. It is that Carnot's limit answers "what is physically forbidden?" while the maximum-power estimate answers "what is reasonable for a machine that has to deliver power?"

From Carnot to the second law and entropy

Carnot's work had almost no immediate effect. The next step came from Émile Clapeyron, a French mining engineer, who in 1834 published a memoir presenting Carnot's analysis graphically. It was translated into English in 1837 and into German in 1843. William Thomson read Clapeyron's paper in 1845 and Carnot's original work at the end of 1848 (Wikipedia: Nicolas Léonard Sadi Carnot). In his 1849 paper "An Account of Carnot's Theory of the Motive Power of Heat," Thomson wrote that nothing in natural philosophy is more remarkable than the establishment of general laws by such a process of reasoning.

The caloric problem

There was a difficulty. Carnot had assumed that heat is conserved: the same amount flows in and flows out, and work comes from the fall. But James Joule's experiments in the 1840s, in which a paddle wheel stirred water and warmed it, indicated that work can be turned into heat and that the two are interchangeable. (The experiment is described in the Joule paddle wheel experiment.) Earlier, Count Rumford's cannon-boring experiment had suggested heat is a form of motion. If heat is converted into work, then the heat released at the cold end must be less than the heat taken in.

Clausius resolved the conflict. In 1850 he published "On the Moving Force of Heat," which first stated the basic ideas of the second law. He kept Carnot's central insight, that the fall from hot to cold is what yields work, and replaced conservation of caloric with the equivalence of heat and work. In 1854 he gave a famous statement of the second law: heat cannot spontaneously flow from a colder body to a warmer one without some other change occurring (Wikipedia: Rudolf Clausius). Kelvin published a parallel statement in 1851, about the impossibility of deriving mechanical effect from matter by cooling it below the temperature of the coldest surrounding objects.

The two statements look different but are logically equivalent, and the Carnot cycle is the bridge that proves it (Wikipedia: Reflections on the Motive Power of Fire).

Entropy

In 1865, Clausius gave the first mathematical version of entropy and named it, choosing a word that resembles "energy" and comes from Greek roots meaning roughly "transformation content" (Wikipedia: Rudolf Clausius). The concept grows straight out of Carnot's cycle. In a reversible cycle, the ratio of heat to temperature absorbed from the hot reservoir equals the same ratio released to the cold one. Entropy is that ratio, heat divided by temperature, tracked as a quantity that is conserved in reversible processes and increases in real ones.

A worked example makes this concrete. Suppose an engine absorbs 1,000 J of heat from a source at 500 K and rejects heat to a sink at 300 K.

Quantity	Reversible (Carnot) engine	Real, irreversible engine
Heat absorbed at 500 K	1,000 J	1,000 J
Entropy taken in (heat ÷ temperature)	2.00 J/K	2.00 J/K
Heat rejected at 300 K	600 J	700 J
Entropy dumped to the sink	2.00 J/K	2.33 J/K
Work produced	400 J (40%)	300 J (30%)
Entropy created inside the engine	0	about 0.33 J/K

The reversible engine dumps exactly as much entropy as it took in, so it produces the maximum 40% (which equals 1 − 300/500). The real engine dumps more entropy than it took in, with the difference created by irreversible processes inside, and every unit of that extra entropy corresponds to extra heat rejected and less work delivered. This is the second law in bookkeeping form: the entropy of an isolated system left to itself cannot decrease. For the wider picture, see thermodynamics and entropy, how thermodynamics explains everything, and entropy and the arrow of time.

What the second law rules out

• Perpetual motion machines of the second kind. These would draw on the large internal energy of the environment to do work, with a single heat reservoir and no cold sink. The second law rules them out (Wikipedia: Second law of thermodynamics).
• Spontaneous cooling of hot things by warmer surroundings. Heat flows from hot to cold on its own. Reversing that flow requires work, as in a refrigerator.
• A free arrow of time at the microscopic scale. The statistical meaning of the law, developed by Ludwig Boltzmann, connects entropy to the number of microscopic arrangements consistent with what we observe.

The man behind the book

Nicolas Léonard Sadi Carnot was born on June 1, 1796, in the Petit Luxembourg palace in Paris. His father, Lazare Carnot, was a mathematician and military organizer who served as Napoleon's minister of the interior during the Hundred Days and then lived in exile in Magdeburg until his death in 1823 (Wikipedia: Nicolas Léonard Sadi Carnot). Sadi trained as a military engineer. The 1824 book was his only publication, and it drew little attention during his lifetime.

Carnot died on August 24, 1832, at Ivry-sur-Seine, aged 36. The hospital record says he was cured of "mania" before dying of cholera. Most of his personal papers were destroyed afterward, by his younger brother Hippolyte. Popular accounts attribute the destruction to cholera quarantine rules, but the available biography states only that Hippolyte destroyed most of them and obscured details of the death, so the reasons are not fully documented.

The notes that changed his mind

Some notes survived, and Hippolyte published them in 1878. Written in spring 1832, they show Carnot had abandoned the caloric theory by then. They estimate that 1 kilocalorie of heat equals 370 kg·m of work, against the currently accepted value of about 427 kg·m, a notable early estimate of the mechanical equivalent of heat. One passage reads: "Heat is simply motive power, or rather motion that has changed form." He never published this work (Wikipedia: Nicolas Léonard Sadi Carnot).

That puts him close to the first law of thermodynamics, the conservation of energy, at about the time Joule and others were making the same step. Whether he would have gone further is a question the surviving papers cannot answer, and any claim that he had already reached the full second law or entropy goes beyond what the sources show.

Heat pumps, solar cells, and electric cars

The Carnot logic reaches well beyond steam.

Heat pumps run the cycle in reverse

A heat pump uses work to move heat from a cold place to a warm one. Its performance is described by the coefficient of performance (COP): heat delivered divided by electricity used. A resistive electric heater has a COP of 1.0. A well-designed heat pump typically achieves about 3 to 5 with an outdoor temperature of 10°C and an indoor temperature of 20°C (Wikipedia: Heat pump).

The Carnot version of the formula gives an upper bound: indoor temperature in kelvins divided by the temperature difference. For 20°C indoors and 10°C outdoors that is 293 ÷ 10, about 29, far above what real units reach, and the bound falls as the outdoor air gets colder. The same source notes that real air-source units fall to roughly 1 to 4 below about −8°C. The thermodynamic reason heat pumps beat resistive heating is that they move heat instead of making it, and the smaller the temperature lift, the less work is needed. For the engineering side, see how heat pumps move heat efficiently. The same reasoning applies in a refrigerator, as described in how a refrigerator works.

Solar cells are not heat engines, but they have a similar limit

A photovoltaic cell turns light directly into electricity, so it does not run a heat cycle, and the Carnot formula does not apply to it directly. Still, the physics of thermal radiation sets a ceiling of a related kind. William Shockley and Hans-Joachim Queisser published the detailed-balance limit for single-junction cells in 1961, in the Journal of Applied Physics (32(3), 510–519). Their original calculation approximated sunlight as a 6,000 K blackbody and gave a maximum of about 30% at a band gap of 1.1 electron-volts. Later calculations with measured sunlight spectra give about 33.7% under the standard AM1.5 spectrum (Wikipedia: Shockley–Queisser limit).

Electric vehicles are a different case

Electric motors convert electrical energy into motion directly, so the Carnot limit does not cap them. The electricity that charges the battery often comes from a heat engine, however, and that part of the chain does. A fair comparison between electric and gasoline cars therefore has to account for how the electricity is generated. A grid supplied by wind, solar, or hydropower has no heat-engine step at all, while a grid supplied by coal runs into the coal-plant figures above.

What Carnot did not claim

The limit is often stretched beyond what it says. Four common misreadings are worth correcting.

"Carnot proved engines can't beat a fixed figure like 40%"

The Carnot limit is not a fixed number. It depends on the two temperatures, and for a hot enough source and a cold enough sink it is very high. The modest efficiencies of real engines come from the practical limits above, not from the theorem.

"Carnot's formula applies to everything that makes energy"

It applies to devices that turn heat into work. A fuel cell, which converts chemical energy directly into electrical work, is not bound by it (Wikipedia: Thermal efficiency). A battery, a hydroelectric turbine, and a wind turbine are not heat engines either.

"Waste heat is a design flaw"

Some heat must be rejected because the cycle needs a cold sink to run at all. The rejected heat is the price of the temperature drop that drives the engine. The practical response is to use it where possible: combined-cycle plants run a second turbine on the exhaust, and combined heat and power systems pipe it to buildings. Neither beats Carnot. They simply put the unavoidable heat to work.

"Better efficiency will fix everything"

Higher efficiency means less fuel per unit of work, which cuts emissions from fossil plants. It cannot change the fact that burning fossil fuel releases carbon dioxide. The thermodynamic limit tells you how much work a given fuel can yield at most. It says nothing about what happens to the combustion products. Discussion of which power sources dominate the world's electricity is covered in countries with the most nuclear power.

Where the idea shows up in other machines

The history of the steam engine, from Newcomen's pumps to the machines Carnot studied, is covered in the invention of the steam engine, and the machine that powers most road vehicles is described in the internal combustion engine. Each can be placed on the same map: a hot source, a cold sink, a working substance, and a ceiling determined by the two temperatures.

The Carnot framework also gives engineers a way to look for improvements. If the gap between the ceiling and the observed efficiency is large, the loss is probably in irreversibility, and better design can recover some of it. If the gap is small, the remaining gains must come from raising the hot side, lowering the cold side, or adding a second cycle, as combined-cycle plants do.

FAQ

Q: Why can't we build a 100% efficient heat engine? A: The Carnot efficiency, 1 − (T_cold / T_hot), equals 1 only if the cold temperature is absolute zero, which cannot be reached. More fundamentally, an engine that rejected no heat would convert heat from a single reservoir entirely into work, which the Kelvin statement of the second law forbids. Some heat must always be rejected to a cold sink.

Q: Did Carnot invent the formula that carries his name? A: Not in its modern form. Carnot's 1824 book established that no engine can beat a reversible one working between the same temperatures, and that all reversible engines are equally efficient. The formula with absolute temperatures followed from Kelvin's absolute scale of 1848, which was built on Carnot's reasoning, and from the later work of Clausius.

Q: Does Carnot's theorem apply to solar panels? A: Not directly. Solar cells convert light to electricity through the photovoltaic effect and are not heat engines. A related ceiling exists, though: the Shockley–Queisser detailed-balance limit of 1961, which is about 33.7% for a single-junction cell under standard AM1.5 sunlight. It is derived from thermal radiation physics, not from the Carnot cycle.

Q: How efficient are real power plants compared with their Carnot limit? A: Far below it. A UK coal plant studied by Curzon and Ahlborn operated at a Carnot limit of 0.64 and an observed efficiency of 0.36. In the United States in 2024, average operating heat rates implied roughly 32% for coal, 44% for natural gas, and 33% for nuclear. Combined-cycle plants reach the low 60s at best, with a certified record of 62.22%.

Q: What is entropy and how does it relate to Carnot? A: Clausius defined entropy in 1865 as heat divided by temperature, tracked across a process. In a reversible Carnot cycle, the entropy taken from the hot reservoir equals the entropy delivered to the cold one, so the net change is zero. In real engines, irreversible processes create additional entropy, which shows up as extra rejected heat and less work.

Q: Is the Carnot limit relevant to electric vehicles? A: Only indirectly. Electric motors convert electricity to motion without a heat-engine step, so the Carnot limit does not cap them. The power plant that generates the electricity may be a heat engine, and then its efficiency is bound by Carnot. For a grid supplied by wind, solar, or hydropower, no such step exists.

Q: What did Carnot believe about heat, and was he right? A: In 1824 he treated heat as caloric, a conserved fluid, which is incorrect. His private notes of 1832, published in 1878, show he had rejected that view and recognized heat as motion that has changed form. His central theorem survived the correction.

The takeaway

Sadi Carnot gave engineers something unusual: a limit that does not shrink as machines get better. By reasoning about reversible cycles instead of measuring any particular engine, he showed that efficiency is governed by temperatures, and that every engine pays a toll in rejected heat. His assumption about caloric was wrong, his book drew little attention in his lifetime, and his papers were mostly destroyed. Yet Clapeyron, Kelvin, and Clausius built the second law and entropy on his argument within three decades. The practical reading is modest and durable: when you see an efficiency figure for any engine that burns something, the first question is not how good the machine is but how far its temperatures allow it to go.

Sources

• Carnot, S. (1824). Réflexions sur la puissance motrice du feu. Bachelier, Paris.
• Clausius, R. (1850). "Über die bewegende Kraft der Wärme." Annalen der Physik. (Described in Wikipedia: Rudolf Clausius.)
• Curzon, F.L. and Ahlborn, B. (1975). "Efficiency of a Carnot engine at maximum power output." American Journal of Physics 43(1), 22–24.
• Shockley, W. and Queisser, H.J. (1961). "Detailed Balance Limit of Efficiency of p-n Junction Solar Cells." Journal of Applied Physics 32(3), 510–519.
• US Energy Information Administration. Electric Power Annual, Table 8.1: Average operating heat rate for selected energy sources.
• US Department of Energy. FuelEconomy.gov: Where the Energy Goes, Gasoline Vehicles.
• Overview pages consulted: Carnot cycle, Sadi Carnot, Reflections on the Motive Power of Fire, Second law of thermodynamics, Thermal efficiency.