Engine Evolution: How 1 Brilliant Frenchman Defined the Limits of Energy

In the winter of 1698, an Englishman named Thomas Savery patented a device that used steam to suck water out of flooded coal mines. It had no piston, wasted most of the heat it consumed, and could lift water barely thirty feet. It was also the beginning of something enormous: the moment human beings started building engines, machines that could turn the heat of a fire into useful motion, on purpose and at scale. A century and a quarter later, a quiet French army engineer named Sadi Carnot sat down and asked a question nobody had thought to ask properly: is there a limit to how much work any engine can ever squeeze out of a given amount of heat? His answer, published in 1824, didn’t just describe steam engines. It founded a new branch of physics, thermodynamics, and along the way introduced an idea — entropy — that would eventually be used to explain everything from why ice melts to why time seems to move in only one direction.

Also read: How Newton Explained the Universe

An age powered by fire it didn’t understand

By the early 1700s, Britain had a problem. It had cut down most of the forests near its cities, and it needed a new source of energy to keep growing. It found one underground, in coal — but coal mines kept flooding, and pumping them dry took more muscle and more horses than anyone could spare. Savery’s steam pump was a crude answer to that specific problem, and it worked just well enough to matter.

Over the following decades, the machine got better in fits and starts. Around 1712, Thomas Newcomen built an engine with an actual piston, using steam to push it up and a spray of cold water to cool and collapse it back down, which pulled water from mines far more reliably. In the 1760s, a Glasgow instrument-maker named James Watt was asked to repair a Newcomen engine and noticed something wasteful in its design: the same cylinder was being heated by steam and then cooled by water, over and over, wasting huge amounts of fuel just reheating metal that didn’t need to move anything. Watt’s fix, adding a separate condenser so the cylinder itself never had to cool down, dramatically improved efficiency and became one of the key mechanical ideas that made the Industrial Revolution possible. Watt went on to convert the up-and-down piston motion into rotary motion, turning the steam engine from a mine pump into a general-purpose power source that could run textile mills, forges, and eventually locomotives and steamboats.

This is the part of the story most people know: clever inventors, incremental tinkering, a machine that reshaped the world. What’s less well known is that all of this happened almost entirely without theory. Engineers built engines by trial and error, tweaking one part at a time and keeping whatever worked. Nobody had a scientific account of why an engine produced the work it did, or whether there was some hard ceiling on how efficient it could ever become. Even the best engines of the early 1820s converted only around 3 percent of their fuel’s heat into usable work — the other 97 percent simply vanished as waste. That gap, between what engines were doing and what anyone understood about why, is exactly where Sadi Carnot stepped in.

Carnot and the engine that existed only on paper

Sadi Carnot was not primarily an engineer of steam engines himself — he was a French military engineer, the son of a former leader of Revolutionary France, and by most accounts a reserved, private man. What drew him to the subject was partly wounded national pride: French steam technology lagged behind British engines, and Carnot came to believe that this technological gap had real consequences for France’s fortunes. Determined to understand the underlying science that nobody had bothered to work out, he set himself two questions. First, was there an upper limit to how much motive power heat could ever produce? Second, was steam actually the best possible medium for extracting that power, or could some other substance do better?

Carnot’s method was unusual, and it’s the reason his 1824 book, Reflections on the Motive Power of Fire, still gets taught two centuries later. Rather than trying to analyze a real, messy steam engine full of friction and leaks, he imagined an idealized one: a theoretical engine operating in a repeating cycle between just two temperatures, a hot reservoir and a cold one, with every step either perfectly isothermal or perfectly insulated from heat loss. This imagined device is now called the Carnot engine, and the cycle it runs on is the Carnot cycle. By stripping away every real-world imperfection, Carnot could ask a purely theoretical question: even in a perfect world, with no friction and no wasted motion, how much work can you get out of a flow of heat?

His answer was striking. The maximum possible efficiency of any heat engine, he showed, doesn’t depend on what the engine is built from, what working fluid it uses, or any detail of its construction. It depends only on the temperatures of the hot and cold reservoirs it operates between. A bigger temperature gap between hot and cold means a more efficient engine is possible; a smaller gap caps the efficiency lower, no matter how cleverly the engine is designed. This was a genuinely new kind of result — not an engineering trick, but a law of nature that any engine, real or imagined, has to obey.

Carnot worked within the caloric theory of heat, the belief, standard for his era, that heat was a weightless fluid that flowed from hot objects to cold ones without being used up. That framework turned out to be wrong. But it’s a testament to how carefully Carnot reasoned that his central conclusion about maximum efficiency survived the theory’s later collapse almost untouched, because he had grounded it in something deeper: the impossibility of getting useful work for free, with nothing else changing. Carnot also came remarkably close to a second landmark idea, later called the second law of thermodynamics, when he recognized that any real engine will always lose some of that theoretical maximum to friction, vibration, and heat leaking between objects at different temperatures — meaning a perfectly efficient engine can be approached on paper but never actually built.

Reflections on the Motive Power of Fire sold poorly and drew little attention when it appeared. Carnot died young, at 36, during the Paris cholera epidemic of 1832, and most of his papers were destroyed as a precaution against spreading the disease. It would take another two decades before other physicists picked up the thread he’d left behind and realized how much was buried in that short, plainly written book.

From motive power to energy and efficiency

The idea that would eventually connect Carnot’s engine to the rest of physics was energy conservation — the recognition that heat and mechanical work are two forms of the same underlying quantity, convertible into one another but never created or destroyed. This wasn’t obvious at the time. In the 1840s, several people working independently, including the physician Julius Robert Mayer and the physicist James Prescott Joule, built the experimental and theoretical case that heat is not a separate, indestructible fluid but a form of energy, interchangeable with mechanical motion. Joule’s careful experiments, dropping weights to turn a paddle wheel in water and measuring the resulting temperature rise, gave the first solid numbers for how much mechanical work corresponds to a given amount of heat.

This reframing changed what an engine’s efficiency actually meant. Once heat and work were understood as the same currency, efficiency became a simple, honest ratio: how much of the heat energy put into an engine comes back out as useful work, versus how much is unavoidably lost. Carnot’s original result, translated into this new energy-based language by the physicist Rudolf Clausius and the engineer-physicist William Thomson (later Lord Kelvin) in the early 1850s, turned out to still hold: no engine operating between two temperatures can beat the Carnot limit, and that limit depends only on the ratio of those two temperatures, measured on an absolute scale. This is why, in a real power plant or a car engine today, engineers care enormously about how hot the combustion gets and how cold the exhaust or coolant runs — the wider that temperature spread, the closer an engine can get to its theoretical ceiling. It’s also why no engine, however advanced, will ever convert 100 percent of its fuel’s heat into work. Some loss isn’t a flaw to be engineered away; it’s built into the physics of turning heat into motion at all.

Where entropy comes from

Rudolf Clausius did more than translate Carnot’s result into the language of energy. Working through the mathematics of the Carnot cycle in the early 1850s, he noticed that a certain quantity — heat divided by the temperature at which it’s exchanged — behaves in a very particular way. Add it up around one full cycle of a perfectly reversible engine, and it always comes back to exactly zero. That meant this quantity, whatever it was, behaved like a genuine property of a system, similar to pressure, volume, or temperature itself, rather than something that depended on the particular path a process took to get there. In 1865, Clausius gave this new quantity a name, borrowing from the Greek word for transformation: entropy.

The deeper claim Clausius extracted from this was the second law of thermodynamics: heat does not spontaneously flow from a colder body to a hotter one, and in any real, imperfect process — as opposed to the idealized, perfectly reversible ones Carnot imagined — the total entropy of a system and its surroundings can only increase or, in the limiting case of a perfectly reversible process, stay the same. It can never decrease. This single principle explains why so many everyday processes only run in one direction. A hot cup of coffee cools to room temperature; room-temperature coffee never spontaneously heats itself back up. An ice cube melts in a warm room; puddles don’t spontaneously refreeze on their own. None of these one-way processes violate the conservation of energy — the total energy is the same either way — but they do violate the second law if run backward, which is precisely why they don’t happen.

For an engine, entropy is what explains the gap between the theoretical Carnot limit and the messier reality of an actual machine. Every real engine involves friction, turbulence, and heat leaking across finite temperature differences instead of the idealized, infinitely slow steps Carnot’s cycle assumes — and each of those imperfections generates additional entropy, which corresponds directly to potential work that’s been permanently lost, not converted to motion and not recoverable afterward. That’s why engineers still refer to the Carnot efficiency as an unreachable ceiling: it marks the best any engine could possibly do, with every real design falling somewhere below it depending on how well it manages friction, heat leakage, and irreversibility.

Why heat still matters

It would be easy to file all of this away as 19th-century history — clever Frenchman analyzes steam engine, physics textbook chapter follows. But the reach of what Carnot, Clausius, and Thomson worked out goes far beyond steam. Every engine humanity has built since, from the internal combustion engine in a car to the gas turbines in a jet aircraft to the steam turbines that still generate most of the world’s electricity today, whether the heat source is coal, natural gas, or a nuclear reactor, is bound by the same Carnot limit and generates entropy according to the same second law. Later engineers, including Rudolf Diesel, used Carnot’s own theoretical framework to design new kinds of engines that pushed operating temperatures higher specifically to chase a higher theoretical ceiling.

The reach goes further still. The second law of thermodynamics is one of the very few physical laws that has a built-in sense of direction — it’s the reason physicists talk about an “arrow of time,” since entropy increasing toward the future, rather than the past, is one of the only ways physics distinguishes one time direction from the other at all. Applied to the universe as a whole, the same logic behind a cooling cup of coffee suggests a long-term trend toward increasing disorder and decreasing usable energy, an idea sometimes called the heat death of the universe. And in the 20th century, entropy escaped physics altogether: Claude Shannon borrowed both the mathematics and the name for his theory of information, and the connection between thermodynamic entropy and information turned out to be more than a loose metaphor, showing up in serious work on the physical limits of computation.

None of that was remotely on Sadi Carnot’s mind in 1824. He was trying to answer a practical, almost narrow question: how good can a heat engine get, and why. What makes his short, poorly received book one of the genuine turning points in physics is that a rigorously honest answer to a narrow engineering question turned out to expose one of the deepest and most universal patterns in nature.

What Carnot didn’t get to

It’s worth being clear about the limits of what Carnot himself established. He worked entirely within the caloric theory, treating heat as a conserved fluid rather than a form of energy, so his own explanation of why his efficiency limit held was, strictly speaking, built on a false premise. He did not discover the second law of thermodynamics in the form we use it today, nor did he coin the term entropy — both of those came from Clausius and Thomson, working a generation later and building explicitly on Carnot’s result. And Carnot’s own manuscripts show he was privately beginning to doubt the caloric theory and edging toward the idea of heat as a form of mechanical work in the years just before his death, a shift in his thinking that was only discovered when his notes were finally examined more than a century later. He simply didn’t live long enough to make that case publicly, or to see how far his one clean, careful result about an idealized engine would end up traveling.

The takeaway

A quiet, underappreciated book about the theoretical limits of heat engines ended up giving physics one of its most far-reaching ideas. Sadi Carnot set out to answer a practical question about steam and efficiency, and in doing so handed later physicists the tools to define entropy, state the second law of thermodynamics, and eventually explain why time itself seems to run in only one direction. The next time a car engine, a power plant, or even a refrigerator hums along nearby, doing its ordinary work of turning heat into something useful, it’s still operating inside the boundary a French engineer worked out on paper two hundred years ago — and still losing a little more to entropy than Carnot’s perfect, imaginary engine ever would.

Sources

• APS Physics — June 12, 1824: Sadi Carnot Publishes Treatise on Heat Engines
• NASA Technical Reports Server — The Transformation of Heat in an Engine (NACA-TM-509)
• Encyclopaedia Britannica — Steam Engine
• Encyclopaedia Britannica — Thermodynamics
• Encyclopaedia Britannica — Sadi Carnot
• Encyclopaedia Britannica — Entropy
• Entropy (MDPI) — A History of Thermodynamics: The Missing Manual
• PMC — A History of Thermodynamics: The Missing Manual
• PMC — Thermoelectric Cycle and the Second Law of Thermodynamics
• PMC — Entropy: From Thermodynamics to Information Processing
• Taylor & Francis — Historical Development of the Theory of Heat and Thermodynamics
• IEEE Technology Navigator — Heat Engines
• MIT OpenCourseWare — Thermodynamics & Kinetics
• MIT OpenCourseWare — Unified Engineering Thermodynamics
• NASA Glenn Research Center — Beginner’s Guide to Propulsion (Thermodynamics)
• NASA Glenn Research Center — Heat Engines and Energy Resources
• New York Academy of Sciences — Exploring the Science and History of Thermodynamics
• Caltech — Feynman Lectures on Physics, Volume I (Heat & Thermodynamics)
• NIST Digital Library — Thermodynamic Data
• CERN Document Server — Thermodynamics-related educational material
• SpringerLink — Thermodynamics Collection
• Elsevier ScienceDirect — Thermodynamics Collection
• Cambridge University Press — Thermodynamics textbooks
• Oxford University Press — Thermodynamics resources
• arXiv — Formulation of Entropy through Work by Carnot Machine and Direct Derivation of Law of Entropy Non-Decrease from Kelvin Principle
• arXiv — Second Law for Active Heat Engines
• arXiv — Information Processing and the Second Law of Thermodynamics
• The Feynman Lectures on Physics — Heat and Thermodynamics
• David V. Schroeder — An Introduction to Thermal Physics (textbook)
• Herbert B. Callen — Thermodynamics and an Introduction to Thermostatistics (textbook)
• Enrico Fermi — Thermodynamics (textbook)
• Richard P. Feynman — The Feynman Lectures on Physics, Vol. I (Heat)
• Yunus A. Çengel & Michael A. Boles — Thermodynamics: An Engineering Approach (textbook)
• Atkins & de Paula — Physical Chemistry (Thermodynamics chapters)
• Moran, Shapiro, Boettner & Bailey — Fundamentals of Engineering Thermodynamics (textbook)
• Stephen G. Brush — The Kind of Motion We Call Heat (historical reference)
• Thomas S. Kuhn — Black-Body Theory and the Quantum Discontinuity (historical context for thermodynamics)