Stirling Engine

An engine with no explosions and no valves

Almost every engine you have met burns its fuel inside the cylinder. A spark or a compression flash detonates a fuel-air charge, the explosion slams a piston down, the burnt gas is thrown out of an exhaust valve, and a fresh charge is sucked in for the next bang. The working gas is consumed and replaced on every stroke, valves orchestrate the breathing, and the whole thing only runs because something keeps exploding.

A Stirling engine does none of that. It seals one fixed mass of gas inside a closed space, applies heat to one end from outside the engine, removes heat from the other end, and lets that single trapped charge of gas cycle between hot and cold forever. Nothing is burned inside. Nothing is exhausted. There are no intake or exhaust valves at all. Heat goes in through a hot metal wall, work comes out at the crankshaft, and waste heat leaves through a cold metal wall — a continuous, sealed loop. Robert Stirling, a Scottish minister, patented the idea in 1816, two decades before practical thermodynamics existed to explain why it worked.

That sealed, valveless, external-combustion architecture is the whole identity of the machine. Because combustion happens outside, in steady open-air conditions rather than in confined pulses, the heat source can be literally anything that gets one end hotter than the other: a gas burner, concentrated sunlight, a decaying nuclear isotope, factory waste heat, a wood stove, or — for a toy demonstration engine — the warmth of your hand against a glass of ice water. And because there is no explosion and no exhaust blowdown, a Stirling engine runs astonishingly quietly. The trade-off, as we will see, is power density.

The mechanism — displacer, power piston, 90° out of phase

The trick that makes a Stirling engine run is getting the gas to expand while it is hot and contract while it is cold. If you could do that, the gas would push hard on the way out and pull weakly on the way back, leaving a net positive swing of work every cycle. Two pistons, ganged together with a 90° phase offset, arrange exactly that.

• The displacer. A loose-fitting plug — usually larger in diameter than the power piston — that slides inside the hot/cold cylinder but does not seal against the wall. Gas flows freely around its edges. Its only job is to push the working gas from the hot end to the cold end and back, like a spoon stirring soup from one side of the pot to the other. Because gas passes freely past it, the pressure on its two faces is nearly equal, so the displacer takes almost no force to move and absorbs almost no work.
• The power piston. A gas-tight piston that feels the rising and falling pressure of the trapped gas and is the only component that converts that pressure swing into shaft work. It is the engine's actual output.
• The regenerator. A heat-storing matrix — fine wire mesh, stacked metal screens, or packed foil — sitting in the gas path between hot and cold. More on it below; it is the component that lifts the cycle to Carnot efficiency.

The displacer and power piston are coupled to the crankshaft so that the power piston lags the displacer by about 90°. Walk through one revolution of a beta engine:

1. Gas on the hot side, expanding. The displacer has pushed the gas onto the hot end. It absorbs heat through the wall, its pressure climbs, and it shoves the power piston outward — this is the power stroke.
2. Displacer moves gas to the cold side. As the crank turns, the displacer shuttles the gas through the regenerator to the cold end, where it rejects heat and its pressure falls.
3. Gas on the cold side, contracting. Now at low pressure, the gas no longer resists, and the flywheel's momentum returns the power piston with little effort.
4. Displacer moves gas back to the hot side. The gas flows back through the regenerator — pre-heated by the heat it dumped on the way out — and the cycle repeats.

The net effect: high pressure pushes the power piston out, low pressure lets it return cheaply, and the difference is the work the engine delivers. The 90° phasing is what guarantees the gas is always hot when expanding and cold when contracting. Shift the phase to 0° or 180° and the engine stalls — there is no net work.

The ideal Stirling cycle and its governing numbers

The idealised thermodynamic cycle has four steps, two at constant temperature (isothermal) and two at constant volume (isochoric / isometric):

Step	Process	What the gas does	Heat
1 → 2	Isothermal expansion at T_h	Expands on the hot side, doing work	Absorbs Q_h from the heater
2 → 3	Isochoric (constant-volume) cooling	Pushed through the regenerator to the cold side	Dumps heat into the regenerator
3 → 4	Isothermal compression at T_c	Compressed on the cold side	Rejects Q_c to the cooler
4 → 1	Isochoric (constant-volume) heating	Pushed back through the regenerator to the hot side	Reabsorbs heat from the regenerator

The two isothermal legs each move heat in or out across the cylinder wall. The two constant-volume legs are where the regenerator earns its keep: the heat dumped in step 2 → 3 is stored in the matrix and handed straight back in step 4 → 1, so the heater never has to supply it. With perfect regeneration the only external heat in is the isothermal expansion heat and the only external heat out is the isothermal compression heat, and the efficiency collapses to the Carnot value:

η_ideal = 1 − T_c / T_h        (T in kelvin)

Worked example — a solar-dish Stirling:
  T_h = 700 °C = 973 K   (heater head, concentrated sunlight)
  T_c =  50 °C = 323 K   (water-cooled cold side)

  η_ideal = 1 − 323 / 973 = 1 − 0.332 = 0.668  →  66.8 %

That 66.8 % is the same ceiling a Carnot engine would have between those two temperatures — the Stirling cycle is one of only a handful of cycles that reaches it in the ideal limit. The work per cycle, for an isothermal ideal-gas model, is the difference of the two isothermal areas:

W_net = n R (T_h − T_c) ln(V_max / V_min)

  n = moles of working gas,  R = 8.314 J/(mol·K)
  V_max / V_min = compression ratio (volume ratio)

Example: n = 0.05 mol helium, T_h = 973 K, T_c = 323 K, V_max/V_min = 2.0
  W_net = 0.05 × 8.314 × (973 − 323) × ln(2)
        = 0.05 × 8.314 × 650 × 0.693
        ≈ 187 J per cycle
  At 1500 rpm (25 rev/s):  P ≈ 187 × 25 ≈ 4.7 kW (indicated, ideal)

Real engines need the Schmidt analysis — a closed-form isothermal model that accounts for the sinusoidal motion of both pistons, the phase angle, and the dead volume — to predict pressure and power with any accuracy, and detailed nodal or CFD models beyond that. But the two equations above capture the essential levers: raise T_h, lower T_c, pressurise the gas (more moles n), and increase the volume ratio.

Why helium and hydrogen instead of air

The working gas choice is a real engineering decision, not a footnote. Power scales with the mass of gas you cycle and how fast you can move heat into and out of it, and that pushes designers toward light, high-conductivity gases run at high pressure:

• Air is free, safe, and used in hobby and low-temperature-difference engines. But it has modest thermal conductivity and you cannot pressurise an open-loop air engine much, so power density is low.
• Helium is the workhorse for serious engines (NASA convertors, many solar units). High thermal conductivity moves heat fast, it is inert, and it is run at 4 to 15 MPa to pack in mass. The catch: helium is a tiny molecule that leaks past seals, so the engine must be hermetically sealed or continuously topped up.
• Hydrogen gives the highest power density of all — best conductivity, lowest flow friction — and was used in the Ford/Philips and United Stirling automotive prototypes of the 1970s. But it leaks even worse than helium, embrittles steel, and is flammable, which kept it out of consumer products.

This is why a hermetic seal and high charge pressure are the hallmarks of a high-performance Stirling: more moles of a fast-conducting gas means more work per cycle and faster heat exchange, both of which directly raise power.

The three configurations — alpha, beta, gamma

Stirling engines are classified by how the displacer and power piston are arranged. All three obey the same thermodynamics; they differ in packaging and sealing.

• Alpha. Two separate power pistons in two separate cylinders — one hot, one cold — connected by a duct that holds the regenerator. There is no distinct displacer; each piston does double duty. Compact and high-power, but both pistons run at high temperature and both need good seals. Used in the Swedish-Navy submarine engines.
• Beta. A displacer and a power piston share a single cylinder, the displacer above the power piston. The most thermodynamically efficient layout because the hot and cold spaces overlap with minimal dead volume. The classic Philips and Robert Stirling original layout.
• Gamma. A displacer and a power piston in separate cylinders joined by a passage. Mechanically the simplest to build and the easiest to fit a large cold-side cooler to, at the cost of extra dead volume that slightly lowers efficiency. The configuration shown in this page's animation, and the one most desktop demonstration engines use.

Configuration	Pistons	Cylinders	Strength	Weakness
Alpha	Two power pistons	Two (hot + cold)	Compact, high power density	Both pistons run hot; sealing both is hard
Beta	Displacer + power piston	One (shared)	Least dead volume → best efficiency	Mechanism inside one bore is intricate
Gamma	Displacer + power piston	Two (linked)	Simplest to build and cool	Extra dead volume lowers efficiency

There is also the free-piston Stirling, which deserves its own mention. Invented by William Beale in the 1960s, it has no crankshaft, no connecting rods, and no rotating shaft at all. The displacer and power piston are sprung masses that bounce on gas springs; the power piston drives a linear alternator directly to make electricity. With no rubbing rotary seals and nothing to wear out, free-piston Stirlings can run hermetically sealed for decades unattended — which is precisely why NASA chose them for deep-space radioisotope power, where you cannot send a mechanic.

The regenerator — the component that makes it Carnot

If you remember one thing about Stirling engines beyond "run by a temperature difference," make it the regenerator. It is a passive wad of fine metal mesh, stacked screens, or packed foil sitting directly in the gas path between the hot and cold spaces, and it behaves like a thermal sponge with a temperature gradient frozen across it: hot at the heater end, cold at the cooler end.

As gas streams from the hot side to the cold side it pours its heat into the matrix, leaving the matrix hotter and the gas already pre-cooled before it ever reaches the cooler. On the return stroke the same gas flows back through the now-hot matrix and picks that heat right back up, arriving at the hot space already pre-heated. The heat is being recycled internally instead of being thrown to the heater and cooler every cycle.

The numbers are dramatic. A good regenerator handles four to five times more heat per cycle than the heater itself adds, and recovers 80 to 95 percent of the heat that would otherwise have to be re-supplied. Remove the regenerator and the heater must reheat the gas from cold to hot every single stroke — which roughly halves the engine's power and efficiency and collapses the cycle far below Carnot. The regenerator is the difference between a Stirling engine and a feeble hot-air toy. It is also a design tightrope: pack the mesh denser for better heat transfer and you choke gas flow with friction (pumping loss); open it up to flow freely and you lose regeneration. Tuning that balance is most of the art of Stirling design.

Where Stirling engines actually earn their keep

• Air-independent submarine propulsion. The Swedish Navy's Gotland-class submarines carry Kockums V4-275R Stirling engines, ~75 kW each, burning diesel with stored liquid oxygen. Running submerged with no air intake, near-silent and with no explosive combustion to detect, they let the boat patrol for weeks without surfacing. Stirling AIP has since spread to Japanese (Sōryū-class) and other navies.
• Spacecraft power. NASA's free-piston Advanced Stirling Radioisotope Generator (ASRG) and the later Stirling convertors turn the ~250 °C heat of decaying plutonium-238 into electricity at about 30 percent efficiency — roughly four times better than the radioisotope thermoelectric generators (RTGs) flown on Voyager and Curiosity, stretching scarce Pu-238 four times further.
• Solar dish-Stirling. A parabolic mirror focuses sunlight onto a Stirling heater head at its focal point. The Stirling Energy Systems / SunCatcher units demonstrated over 31 percent sunlight-to-electricity — among the highest of any solar technology — though they lost the cost race to flat photovoltaic panels.
• Cryocoolers. Run the cycle backwards (drive the shaft, and one end gets cold) and a Stirling becomes a refrigerator. Stirling cryocoolers cool infrared sensors in missiles and telescopes to ~80 K and liquefy air and natural gas industrially.
• Micro-CHP and off-grid generators. Home combined-heat-and-power units (e.g. the Whispergen) and quiet marine/RV generators use the Stirling's fuel flexibility and low noise. Wood-stove-top fan generators and developing-world cookstove chargers run on the cycle too.
• Education. Low-temperature-difference (LTD) demonstration engines spin from a few degrees of difference — a cup of hot coffee, or just the warmth of your hand against a cold plate.

Failure modes and trade-offs — why it never replaced the car engine

• Low power density. All heat must conduct through a metal wall rather than detonate inside the cylinder, so heat transfer is the bottleneck. A Stirling making the power of a car engine is several times larger and heavier. This single fact is why the 1970s automotive Stirling programs (Ford, GM, Philips, MAN) were abandoned despite excellent efficiency and emissions.
• Seal and gas leakage. Helium and hydrogen leak past any sliding seal over time, dropping charge pressure and power. Crankshaft (kinematic) engines need rod seals that are a perennial maintenance item; free-piston designs sidestep this entirely by being hermetic.
• Slow throttle response. You cannot change a Stirling's power quickly because you cannot change the heater-head temperature quickly — the thermal mass has inertia. Power is usually controlled by varying the mean gas pressure or dead volume, neither of which is instant. Bad for vehicles, irrelevant for steady generators.
• Heater-head creep and oxidation. The hot head sits at 700–800 °C under pressure for thousands of hours; it suffers high-temperature creep and oxidation, and is usually built from costly nickel superalloys (Inconel). This is the limiting durability item in high-temperature units.
• Dead volume penalty. Any gas trapped in the regenerator, ducts, and clearances that does not get fully cycled between hot and cold dilutes the pressure swing and cuts work. Minimising dead volume — the great strength of the beta layout — is a constant design pressure.
• Cost. Precision pistons, hermetic seals, a finely engineered regenerator, and superalloy hot ends make a Stirling expensive per kilowatt compared with a mass-produced gasoline engine — acceptable only where its quiet, sealed, fuel-flexible operation is worth paying for.

Stirling vs. the internal-combustion alternatives

Property	Stirling (external)	Otto / gasoline (internal)	Diesel (internal)	Steam / Rankine (external)
Combustion location	Outside, continuous	Inside, pulsed	Inside, pulsed	Outside, continuous
Working fluid	Sealed gas (He, H₂, air)	Air + fuel, replaced each cycle	Air + fuel, replaced each cycle	Water / steam (phase change)
Real efficiency	38–40 % (best)	25–35 %	35–45 %	30–42 % (large plant)
Ideal-cycle ceiling	Carnot	Otto (below Carnot)	Diesel (below Carnot)	Rankine (below Carnot)
Power density	Low (heavy)	High	High	Low
Noise / emissions	Very low	High	High	Low
Fuel flexibility	Any heat source	Gasoline only	Diesel only	Any heat source
Throttle response	Slow	Fast	Fast	Slow

The pattern is clear: Stirling and Rankine, the two external-combustion cycles, trade power density and responsiveness for quiet, clean, fuel-flexible operation and (for Stirling) a Carnot-limited efficiency ceiling. Internal combustion wins anywhere weight and instant throttle matter — which is most vehicles. Stirling wins in the quiet, sealed, slow-and-steady niches: submarines, spacecraft, solar dishes, and silent generators.

Common pitfalls when designing or building one

• Wrong phase angle. The power piston must lag the displacer by roughly 90°. At 0° or 180° the gas expands and contracts at the wrong temperatures and the engine produces no net work and will not self-start. Always verify the phasing before chasing other faults.
• Sealing the displacer. Beginners often try to make the displacer gas-tight. It must not seal — gas has to flow freely around it. A sealing displacer just compresses gas and stalls the engine.
• Skimping on the regenerator. Leave it out (or use too little mesh) and you lose roughly half the power and efficiency. Pack it too densely and flow friction (pumping loss) eats the gains. The mesh density is a tuned compromise, not a maximise-or-minimise variable.
• Too much dead volume. Long connecting ducts, an oversized cooler, and loose clearances trap gas that never fully cycles, diluting the pressure swing. Keep the hot and cold spaces and the path between them as tight as the layout allows.
• Insufficient cooling. Efficiency is 1 − T_c/T_h, so a poorly cooled cold side raises T_c and quietly kills performance. The cold-side heat exchanger matters as much as the heater.
• Ignoring the heater-head limit. Pushing T_h higher always helps efficiency on paper, but the head creeps and oxidises above ~800 °C. Material temperature limits, not thermodynamics, cap the practical T_h.