How Heat Pumps Actually Work (and Why They Beat 100%)

How Heat Pumps Actually Work (and Why They Beat 100%)

How Heat Pumps Actually Work (and Why They Beat 100%)

Here is a number that sounds like a scam: a good heat pump delivers three to four units of heat for every one unit of electricity it draws. That is 300–400% “efficiency.” Furnaces, gas boilers, and resistance heaters — devices most of us trust completely — top out at roughly 100%, and usually less. So either every HVAC engineer is lying, or we have been quietly mislabeling what a heat pump does. It is the second one. Understanding how heat pumps work means letting go of the idea that a heater makes heat. A heat pump doesn’t make heat. It moves heat — and moving something is almost always cheaper than building it from scratch.

That single reframing dissolves the paradox and opens a richer story about phase changes, pressure, and a refrigerant that can boil in freezing air. By the end you will understand the refrigeration cycle in detail, why the coefficient of performance shifts with the weather, where the hard physical limits sit, and why your next furnace might not be a furnace at all.

What this post covers: the move-not-make principle, the vapor-compression cycle stage by stage, COP and the Carnot ceiling, cold-climate behavior, air-source versus ground-source, and the refrigerants quietly changing underneath it all.

Why “over 100% efficient” isn’t cheating

A heat pump can exceed 100% efficiency because efficiency, defined as heat delivered divided by energy created, is the wrong yardstick. The pump barely creates heat at all. It uses electricity to relocate existing thermal energy from outside to inside, so the heat you feel is mostly imported, not manufactured — no law of thermodynamics is bent.

Think of a bucket brigade versus a bonfire. A resistance heater is a bonfire: it converts electrical energy directly into heat, and the very best it can do is turn one joule of electricity into one joule of warmth — that is 100%, the theoretical ceiling for making heat. A heat pump is the bucket brigade. It spends a little energy carrying heat that already exists outdoors and dumps it indoors. Because it is hauling free ambient energy, the warmth arriving in your living room far exceeds the electricity spent doing the hauling.

This is why engineers refuse to call the ratio “efficiency” and instead use coefficient of performance, or COP. A COP of 3.5 means 3.5 joules of heat delivered per joule of electricity consumed. The “extra” 2.5 joules were never created — they were scooped out of the cold air outside and ferried indoors. The First Law of Thermodynamics (energy is conserved) is perfectly satisfied: every joule is accounted for, it just came from two places at once. And the Second Law is satisfied too, because we are spending high-grade electrical work to push heat “uphill” from cold to warm, which is exactly the price the universe demands for that direction of travel. The U.S. Department of Energy’s heat pump guidance frames the same idea in plain terms: these systems move heat rather than generate it, which is what makes them so much more economical than combustion or resistance heating.

The core idea: there’s heat in cold air, and a fluid that boils to grab it

The whole machine rests on one counterintuitive fact: air well below 0°C still contains an enormous amount of thermal energy, and a refrigerant engineered to boil at very low temperatures can absorb that energy. By cycling that refrigerant through pressure changes, a heat pump repeatedly grabs heat where it is cold and releases it where it is warm.

Vapor-compression cycle loop showing compressor, condenser, expansion valve, and evaporator moving heat from outdoor air to indoor space
Figure 1: The closed refrigerant loop — heat enters at the evaporator outdoors and exits at the condenser indoors, driven by the compressor.

“Cold” is not “no heat”

Our intuition treats freezing air as empty of warmth, but that is a trick of perspective. Temperature measures the average kinetic energy of molecules relative to absolute zero at −273.15°C. Air at −10°C is still 263 degrees above the floor where molecular motion stops. Those nitrogen and oxygen molecules are zipping around at hundreds of metres per second; a cubic metre of winter air holds a vast reservoir of thermal energy. It only feels empty because your body, at 37°C, is warmer than the air and loses heat to it. Heat flows from hot to cold on its own, so you feel the drain.

A heat pump’s job is to tap that reservoir anyway. The trick is to introduce something even colder than the outdoor air — a refrigerant chilled to, say, −20°C — so that heat flows naturally from the (relatively warm) −10°C air into the (genuinely colder) refrigerant. Nature still moves heat from hot to cold; we just engineered a “colder” to make the cold air count as “hot.”

Boiling is a heat sponge

The second half of the trick is phase change. When a liquid boils into a gas, it absorbs a large amount of energy — the latent heat of vaporization — without its temperature rising. That energy goes entirely into breaking the molecular bonds that hold the liquid together. This is why sweat cools you and why a pot of boiling water stays at 100°C no matter how high you crank the burner. A refrigerant chosen to boil at a usefully low temperature becomes a thermal sponge: as it vaporizes in the outdoor coil, it soaks up heat from the air while barely warming. That latent-heat absorption is doing the heavy lifting, and it dwarfs what you could move by simply warming a liquid.

Pressure is the steering wheel

So how do you make a fluid boil at −20°C in one place and condense at +40°C in another, on demand? Pressure. A fluid’s boiling point depends on the pressure pressing down on it — water boils at a lower temperature on a mountaintop, higher in a pressure cooker. Refrigerants exaggerate this dependence dramatically. By dropping the refrigerant to low pressure, you force it to boil even in freezing air. By squeezing it to high pressure, you raise its condensing temperature above your room temperature so it will gladly dump heat indoors. Pressure is the single lever the machine pulls to decide where the refrigerant grabs heat and where it releases it. The compressor and the expansion valve are nothing more than the two ends of that lever.

The vapor-compression cycle, stage by stage

The vapor-compression cycle is a closed loop of four components — evaporator, compressor, condenser, expansion valve — that the refrigerant travels endlessly. At each stage its pressure, temperature, and phase change in a coordinated way so that it absorbs heat outdoors as a boiling liquid and releases it indoors as a condensing gas. Nothing is consumed; the same refrigerant goes around forever.

Four phase-change stages of the refrigeration cycle showing heat absorbed at the evaporator and released at the condenser
Figure 2: The four stages as a phase-change story — boil, squeeze, condense, expand — repeating around the loop.

Stage 1 — Evaporator: cold liquid drinks heat from the air

The refrigerant enters the outdoor coil as a cold, low-pressure mixture of liquid and mist, colder than the outside air. Heat flows from the air into the refrigerant, and because the refrigerant is at its boiling point for that low pressure, the incoming energy goes into boiling it rather than warming it. It exits as a cool low-pressure vapor, having quietly stolen warmth from air that felt freezing to you. A fan blows outdoor air across the coil to keep fresh, heat-laden air arriving — this is the unit humming away outside your house.

Stage 2 — Compressor: the only part that costs you

The compressor is the heart of the system and the one place electricity actually goes in. It takes that cool low-pressure vapor and squeezes it hard. Compressing a gas concentrates its thermal energy into a smaller volume, so its temperature shoots up — the same reason a bicycle pump gets hot. The refrigerant leaves as a hot, high-pressure gas, now hotter than your home’s interior. Crucially, the compressor did not create most of that heat; it concentrated the heat carried in from outside and added the modest work needed to raise its temperature above room level. That added work is the electricity bill. Everything else was a free ride.

Stage 3 — Condenser: dumping heat into the house

The hot high-pressure gas now flows through the indoor coil. Because it is hotter than the room, heat flows naturally out of the refrigerant and into your living space — the reverse of stage one. As it sheds energy, the gas condenses back into a liquid, releasing its latent heat of vaporization all over again, this time into the room. This latent release is why the condenser delivers so much warmth: it is paying back all the boiling energy from outdoors plus the compressor’s work. The refrigerant leaves as a warm, high-pressure liquid. A blower or hydronic loop carries the warmth to your rooms.

Stage 4 — Expansion valve: resetting to cold

The warm high-pressure liquid hits a narrow restriction — an expansion valve or capillary tube. As it squeezes through to the low-pressure side, it expands suddenly. Expansion does the opposite of compression: the fluid’s energy spreads into a larger volume, its pressure collapses, and its temperature plummets. Some of it flashes to vapor instantly, cooling the rest. It emerges as the cold low-pressure mist we started with, ready to drink heat from outdoor air again. The loop closes and repeats many times a minute. This same vapor-compression logic is exactly what your refrigerator and air conditioner use — a heat pump is simply that machine pointed the way you want.

For the same physics applied to materials and failure rather than comfort, our piece on how lithium-ion batteries work and why thermal runaway happens walks through heat that you very much do not want to move around.

COP, temperature lift, and the Carnot ceiling

COP measures how much heat a pump moves per unit of electricity, and it falls as the temperature gap between indoors and outdoors grows. The bigger that “lift,” the harder the compressor must work to push heat uphill, so a mild-day COP near four can sink toward one on a frigid night. The theoretical maximum for any given lift is set by Carnot’s law — and real machines reach only a fraction of it.

Conceptual chart of coefficient of performance falling as outdoor temperature drops and temperature lift increases
Figure 3: COP versus temperature lift, conceptually — a small gap means a high COP; a large gap drags it toward one.

Why temperature lift dominates everything

The defining variable is temperature lift: the gap between the temperature at which the refrigerant releases heat indoors and the temperature at which it collects heat outdoors. On a 7°C autumn day heating a 20°C room, the lift is modest, the compressor barely strains, and COP can sit around 3.5–4. On a −15°C night the lift roughly doubles, the compressor must work the vapor far harder to reach a useful condensing temperature, and COP can drop toward 2 or approach 1 in extreme cold. This is not a flaw; it is the cost of pushing heat across a steeper hill. Two identical units in Lisbon and Winnipeg will post very different seasonal numbers purely because of the weather they face.

The Carnot limit — the speed of light for heat pumps

There is a hard ceiling no machine can beat, set in the 1820s by Sadi Carnot. For a heat pump, the maximum possible COP equals the warm absolute temperature divided by the difference between the warm and cold absolute temperatures, both in kelvin. The consequences: as the cold side gets colder, the denominator grows and the ceiling falls; and the ceiling is always finite. You can never get infinite heat for free, but for small lifts the theoretical maximum is enormous — sometimes 10 or 20. Real machines, dragged down by compressor inefficiency, friction, imperfect heat exchangers, and the energy for fans and defrost, typically reach 40–60% of the Carnot ceiling. That fraction is why a Carnot ceiling of 8 becomes a real-world COP near 3.5–4. The gap between ideal and real is where HVAC engineering lives.

Reading the rating labels

Manufacturers rarely quote a single COP, because it changes with conditions. Instead you will see seasonal metrics that average performance across a heating season — HSPF in North America, SCOP under European standards — plus SEER for cooling. These roll the COP-versus-temperature curve into one headline number for a typical climate. Treat them as comparative benchmarks rather than promises; your actual seasonal COP depends on your local winter, your home’s heat demand, and how the system was sized and installed. Benchmark against your own climate, not a glossy spec sheet.

Heating, cooling, and the clever reversing valve

A heat pump heats and cools with the same hardware by flipping the direction heat travels, using a reversing valve that swaps the roles of the indoor and outdoor coils. In summer it pulls heat out of your house and dumps it outside (an air conditioner); in winter it does the reverse. One machine, two seasons, one elegant component doing the switching.

The reversing valve is a four-port solenoid that redirects high-pressure refrigerant flow. In cooling mode the indoor coil becomes the evaporator — boiling refrigerant and absorbing heat from your rooms — while the outdoor coil becomes the condenser that rejects that heat to the summer air. In heating mode the valve flips and the coils swap jobs. This is why “a heat pump is just an air conditioner running backwards” is literally true: an air conditioner already moves heat against the natural gradient, so pointing it the other way is a plumbing decision, not a new invention. The cooling COP is sometimes labeled EER or SEER, but the physics — boil here, condense there, pressure as the lever — is identical.

Cold-climate heat pumps and the defrost problem

When the outdoor coil drops below freezing while absorbing heat from humid air, moisture condenses and freezes onto it, choking heat transfer. Cold-climate heat pumps handle this with periodic defrost cycles and aggressive engineering, letting them keep useful COPs at temperatures that would have stalled units a decade ago.

Here is the catch the brochures gloss over. To pull heat from −10°C air, the outdoor coil must be even colder, perhaps −20°C. Any humidity in the air freezes on contact, building a layer of frost that insulates the coil and blocks the very heat transfer you need. So the system periodically runs a defrost cycle: it briefly reverses into cooling mode, sending hot refrigerant out to the coil to melt the ice, sometimes with a backup electric heater to keep your rooms from going cold during the swap. Defrost costs energy and momentarily dents COP, which is part of why real seasonal performance trails the headline lab figure.

Modern cold-climate models — the kind certified under programs like the ENERGY STAR Cold Climate spec — fight back with inverter-driven variable-speed compressors that modulate instead of slamming on and off, enhanced vapor injection that boosts low-temperature capacity, and smarter defrost logic that runs only when frost is present. The result is real heating capacity and COPs comfortably above 1.5–2 even at −15°C to −25°C, dispelling the myth that heat pumps “don’t work in real winters.” They do; they simply work less efficiently as the lift climbs, exactly as the physics predicts.

Air-source vs ground-source: where you steal the heat

Air-source and ground-source heat pumps differ only in where they collect heat: ambient air versus the ground. Air-source units are cheaper and easier to install but lose efficiency as winter air cools; ground-source units cost far more upfront but tap the stable, mild temperature of the earth and hold a high COP all year.

Comparison of air-source and ground-source heat pumps showing outdoor air versus buried ground loop as heat sources
Figure 4: Two ways to source heat — open-air coil versus a buried loop in stable ground.

The overwhelming majority of installed heat pumps are air-source. They sit outside, breathe ambient air across a coil, and need no excavation, which keeps installation affordable. Their weakness follows from the COP-versus-lift story: as outdoor air plunges in midwinter, the lift grows and efficiency sags right when you need heat most. For most temperate and even many cold climates, modern cold-climate air-source units are now more than adequate, and they dominate the residential retrofit market because they are cheap to fit.

Ground-source: expensive, stable, quietly brilliant

Ground-source (or “geothermal”) heat pumps bury their heat-collecting loop in soil or deep boreholes. A few metres down, the ground holds a steady temperature year-round — roughly your local annual average, often around 10–15°C — because soil is a thermal flywheel that ignores daily and seasonal swings. That stability is the prize: the lift never balloons the way it does for an air-source unit in a cold snap, so COP stays high through the depths of winter. The price is brutal upfront cost — trenching or drilling dominates — and a longer payback. But over decades, in a harsh climate, the savings and consistency can be decisive. Water-source variants apply the same idea to lakes, aquifers, or wastewater: find a stable, mild reservoir and tap it.

The engineering instinct here — exploit a quiet, stable physical reference instead of fighting a noisy environment — echoes through other precision systems. It is the same philosophy behind how GPS atomic clocks achieve precision timekeeping, where stability is everything.

Heat pump vs furnace: the honest comparison

In the heat-pump-vs-furnace matchup, the heat pump almost always wins on energy efficiency and carbon, while a furnace can win on raw output during extreme cold and on upfront cost. A furnace burns fuel to make heat and is capped near 100% efficiency; a heat pump moves heat at 200–400%, but its advantage shrinks as temperatures fall.

A gas furnace or boiler is a combustion device. It burns natural gas, oil, or propane and transfers the resulting heat to your home, with the best condensing furnaces reaching the mid-90s percent — meaning a few percent of the fuel’s energy escapes up the flue. It is fundamentally a make-heat machine, and physics caps it: you cannot extract more energy from fuel than the fuel contains. A heat pump, by contrast, sidesteps that cap entirely by importing ambient heat, which is how it posts numbers a furnace can never touch.

The honest caveats matter. First, electricity often costs more per delivered joule than gas, so a heat pump’s cost advantage depends on local prices, not just its COP. Second, in a deep cold snap a furnace’s output does not falter while a heat pump’s capacity and COP both sag — which is why cold regions sometimes pair a heat pump with a backup source in a “hybrid” or “dual-fuel” setup. Third, the carbon story increasingly favors heat pumps as grids decarbonize: a heat pump on renewable electricity is dramatically cleaner than any combustion appliance, and that gap widens every year. For most homeowners in 2026, a well-sized cold-climate heat pump is the efficiency and emissions winner, with a furnace relegated to backup duty or the harshest climates.

Trade-offs, gotchas, and what goes wrong

Heat pumps are not magic boxes, and the most common disappointments trace back to sizing, climate, and refrigerant choices rather than the underlying physics. The cycle is sound; the failures are practical.

The biggest real-world gotcha is oversizing and short-cycling. A unit sized for the coldest hour of the year is wildly oversized for mild days, switching on and off rapidly, wearing out faster and never settling into its efficient steady state. Variable-speed inverter units largely solve this, but a poorly specified single-stage unit can perform far below its rating. Defrost penalties in humid cold climates eat into seasonal COP and can briefly blow cold air if backup heat is undersized. Electricity pricing can quietly erase the cost benefit even when COP is excellent — run the numbers for your tariff. Distribution mismatch bites homes built for high-temperature radiators: heat pumps deliver heat at lower water temperatures, so they pair best with underfloor heating or oversized emitters.

Then there is refrigerant. The widely used R-410A has a high global warming potential, so the industry is shifting toward lower-GWP options like R-32 and the hydrocarbon R-290 (propane), which are far gentler on the climate if they leak — at the cost of mild flammability that demands careful handling and charge limits. Regulations in many regions are accelerating this transition, so the refrigerant in a unit bought today may differ from one bought five years ago. None of these issues is a dealbreaker; they are simply the difference between a heat pump that delights and one that underwhelms.

Practical recommendations

If you are weighing a heat pump, the physics points to a short, practical checklist:

  • Size for the load, not the worst hour. Insist on a proper heat-loss calculation; prefer variable-speed inverter units that modulate rather than cycle.
  • Match the climate. In real winters, choose a certified cold-climate model and check its rated capacity and COP at low temperature, not just at mild conditions.
  • Mind the emitters. Heat pumps love low-temperature distribution — underfloor or generously sized radiators. Retrofitting onto small high-temperature radiators may need upgrades.
  • Run your own numbers. Compare delivered-heat cost using your real electricity and gas tariffs and a realistic seasonal COP — not a headline lab figure.
  • Consider ground-source for harsh, long winters if you can absorb the upfront cost; the steady COP pays back over decades.
  • Ask about the refrigerant. Lower-GWP options like R-32 or R-290 are the future-proof choice; understand the handling implications.
  • Plan backup wisely. In extreme climates, a hybrid setup with a backup heat source covers the coldest nights without oversizing the pump.

Frequently asked questions

Why are heat pumps more than 100% efficient if that breaks physics?

It doesn’t break physics, because “efficiency” is the wrong word. A heat pump moves existing heat from outside to inside rather than creating it, so the warmth delivered indoors far exceeds the electricity used to do the moving. The proper metric is coefficient of performance (COP), typically 2–4, and every joule is still conserved exactly as the First Law of Thermodynamics requires.

How can a heat pump pull heat from freezing air?

Cold air still contains huge thermal energy — even −10°C is over 260 degrees above absolute zero. A heat pump introduces a refrigerant chilled to an even lower temperature, so heat flows naturally from the “warmer” cold air into the colder refrigerant, which boils to absorb it. Pressure control lets the same fluid boil outdoors and condense indoors, ferrying that heat into your home.

What is a good COP for a heat pump?

A COP of 3–4 is excellent for typical heating conditions, meaning three to four units of heat per unit of electricity. The figure falls as outdoor temperatures drop and the temperature lift grows, sometimes toward 2 or lower in severe cold. Look at seasonal ratings like HSPF or SCOP for a climate-averaged picture rather than a single best-case number.

Heat pump vs furnace — which is cheaper to run?

It depends on your local energy prices. A heat pump’s COP of 3–4 usually beats a furnace’s roughly 95% efficiency on energy terms, but if electricity is far pricier than gas per joule, the cost gap narrows or reverses. As electricity grids decarbonize, heat pumps win decisively on carbon emissions, and increasingly on cost too.

Do heat pumps work in very cold climates?

Yes. Modern cold-climate heat pumps with variable-speed compressors and vapor injection deliver useful heat and COPs above 1.5–2 even at −15°C to −25°C. They run less efficiently than on mild days because the temperature lift is larger, but the old claim that they “don’t work in real winters” is outdated. Many cold regions now rely on them, sometimes with a backup source for extreme nights.

What’s the difference between air-source and ground-source heat pumps?

Air-source units collect heat from outdoor air; they are cheaper and easier to install but lose efficiency as winter air cools. Ground-source (geothermal) units draw heat from the stable, mild temperature of the earth via buried loops, holding a high COP all year. Ground-source costs far more upfront due to excavation but pays back over decades in harsh climates through steadier, cheaper operation.

Further reading

By Riju — about.

Comments

No comments yet. Why don’t you start the discussion?

Leave a Reply

Your email address will not be published. Required fields are marked *