№ 34 · engineering

Why no engine can turn all its heat into work

Every real power plant throws most of its fuel's heat away — not because the parts are bad, but because a share of it can never become work, and two temperatures fix that share.

What a heat engine is

A heat engine takes in heat from something hot, turns part of it into work, and dumps the rest into something colder: a car engine through the exhaust, a power station into a river or cooling tower. Its efficiency is the fraction of heat taken in that comes out as work.

Why the ceiling matters

A well-built engine cannot approach 100%. In 1824 Sadi Carnot showed that between a hot source and a cold sink, the best possible efficiency depends on nothing but the two temperatures. In Kelvin's absolute scale, no engine of any design can beat

ηmax = 1 − Tcold / Thot

with both temperatures in kelvin. That ceiling is the first reason real machines sit where they do; useful speed and boiler and stack losses take more on top. In 2024 the average US coal plant needed about 10,800 BTU of fuel per kilowatt-hour of electricity, and a kilowatt-hour is only 3,412 BTU: an efficiency near 32%. Gas plants averaged about 44%. Nobody wastes two thirds of the fuel out of carelessness; the ceiling was set before the first bolt.

Interactive Drag the hot and cold temperatures, then press run a cycle and watch 100 J of heat split into work and dumped heat; the heavy dashed curve is the Carnot ceiling nothing can cross.

T hot = 800 K T cold = 300 K
Each cycle takes exactly 100 J from the hot block. The engine drawn is an ideal reversible one, so it sits on the Carnot ceiling 1 − Tcold/Thot; the amount it must dump is 100 J × Tcold/Thot, and no design can dump less. The lower, solid curve is 1 − √(Tcold/Thot), one estimate of the efficiency an engine reaches when tuned for maximum power rather than maximum efficiency; it holds when the hot and cold sides waste heat symmetrically, and the general answer lies in a band around it. The animation of packets is schematic; the numbers are the formula.

Where the limit comes from

Heat, left alone, flows from hot to cold and never the other way — an observation never contradicted. Carnot turned that one-way street into a bound.

Picture an engine that runs so gently it could be run backwards, step for step. Backwards it is a refrigerator: it takes in work and pumps heat from cold to hot. Call it reversible.

Suppose a rival engine were more efficient. Use its work to drive the reversible engine backwards as a refrigerator. The rival draws heat from the hot source and makes work; the refrigerator, being less efficient forwards, pushes more heat back into the hot source than the rival took out, using only that work. Net result: heat has moved from cold to hot with nothing else changed — exactly what never happens. So no engine can beat a reversible one, and every reversible engine between the same temperatures has the same efficiency, whatever it is built from.

That common value depends on the temperatures alone; working it through for a gas gives 1 − Tcold/Thot. Efficiency reaches 100% only if the cold side is at absolute zero or the hot side infinitely hot. Neither exists, so some heat is always dumped. The dumped share is Tcold/Thot: a steam plant at 800 K dumping into a 300 K river keeps at most 1 − 300/800 = 62.5% of its heat as work, however cleverly built.

Real engines sit lower still — a trade-off, not a flaw. A reversible engine delivers no power: it must run infinitely slowly. Ask for power and heat must flow across temperature differences, itself a one-way, wasteful step. Engines tuned for maximum power rather than maximum efficiency land near 1 − √(Tcold/Thot) — about 39% for the plant above. The coal fleet's 32% is in the same neighbourhood once boiler and stack losses, which the formula ignores, are added.

In short

A heat engine takes heat from hot, returns part as work, and dumps the rest cold. Because heat never flows from cold to hot by itself, no engine can beat a reversible one, whose efficiency is fixed by temperature: 1 − Tcold/Thot. The cold side is never at absolute zero, so the ceiling is always below 100%, and useful speed costs more on top.

Where this comes from

  1. Efficiency at maximum power of low dissipation Carnot engines linked only, not reproduced
    Massimiliano Esposito, Ryoichi Kawai, Katja Lindenberg, Christian Van den Broeck · arXiv:1008.2464 · 2010
    arxiv.org/abs/1008.2464
  2. Universal trade-off relation between power and efficiency for heat engines linked only, not reproduced
    Naoto Shiraishi, Keiji Saito, Hal Tasaki · arXiv:1605.00356 · 2016
    arxiv.org/abs/1605.00356
  3. University Physics Volume 2, §4.5 The Carnot Cycle linked only, not reproduced
    OpenStax · 2016
    openstax.org/books/university-physics-volume-2/pages/4-5-the-carnot-cycle
  4. Electric Power Annual, Table 8.1: Average Operating Heat Rate for Selected Energy Sources, 2014 through 2024 linked only, not reproduced
    U.S. Energy Information Administration · 2025
    www.eia.gov/electricity/annual/html/epa_08_01.html