Heat Flows From Cold to Hot Inside Every Refrigerator, All Day

“Heat flows from hot to cold, never the reverse.” Every student hears this before they hear the words “second law.” Then they go to the kitchen, open the fridge, and watch it pull heat out of a 4°C interior and dump it into a 25°C room, continuously, for years. If heat genuinely never moved from cold to hot, that machine could not exist. It does, it works, and nothing about it is broken. The sentence everyone memorised is missing a word, and that missing word is the entire reason refrigerators are possible.

The word everyone drops: “spontaneously”

The Clausius statement of the second law says heat cannot flow from a colder body to a hotter one spontaneously, with no other effect. Both qualifiers matter. A refrigerator extracts heat Qc from its cold interior and rejects a larger heat Qh to the warmer room, but only because a compressor does work W on the refrigerant first, drawing that energy from the electricity grid. The “other effect” is right there on your electricity bill. Over one full cycle the refrigerant returns to its starting state, so by the first law, ΔU = Q − W, applied to the whole loop, the books close as Qh = Qc + W. Nothing is created, nothing vanishes, and nothing moves for free. The second law was never a ban on cold-to-hot heat flow, it is a ban on getting that flow for free.

This is also why refrigerator performance is measured differently from engine efficiency. A heat engine is graded on how much work it squeezes out of a given heat input, always less than one. A refrigerator is graded on its coefficient of performance, COP = Qc/W, how much cold you buy per unit of work spent, and this number is routinely greater than one. That is not a loophole, it is the entire point of the machine.

Running the numbers on an actual fridge

Take a domestic refrigerator holding its interior at Tc = 277 K (4°C) inside a kitchen at Th = 308 K (35°C), and say it needs to move Qc = 200 J of heat out of the interior. Carnot’s theorem caps the best possible performance at COP(carnot) = Tc/(Th − Tc) = 277/31 ≈ 8.9, and no real compressor gets there, irreversibilities in the compression and heat exchange see to that. A realistic domestic unit runs closer to COP ≈ 3. At that COP, shifting 200 J of cold out of the interior costs W = Qc/COP = 200/3 ≈ 66.7 J of compressor work, and the kitchen receives Qh = Qc + W ≈ 266.7 J, more heat than was removed from inside, by exactly the work that was spent. Set W = 0 in that equation and you get a machine that cools your kitchen forever at zero running cost, a genuine free lunch. That is precisely the outcome the second law rules out, which is why the equation never lets W reach zero for a net cooling effect.

Same mechanism, one stroke at a time

The compressor stroke that makes this possible is not exotic physics, it is the same work-to-heat exchange you can watch happen in a single adiabatic compression: squeeze a gas with zero heat allowed in or out, and its temperature still rises, visibly, purely from the mechanical work done on it. The live P-V simulation on the thermodynamics page shows exactly that reddening under compression. A refrigerator just repeats that stroke in a closed cycle, pairing it with an expansion that cools the refrigerant back down before it re-enters the cold interior. Heat climbing from cold to hot was never the forbidden move, getting it to climb without paying in work is, and no kitchen appliance has ever managed that.