The Heat Pipe: Moving Kilowatts of Heat With No Moving Parts

The core idea: carry heat as latent heat, not sensible heat

Move heat through a solid and you pay for it in temperature. Fourier's law, Q = −kA·(dT/dx), says that to push a heat flow Q through cross-section A of a material with conductivity k, you must accept a temperature gradient dT/dx. For 100 W through a 1 cm² copper bar (k ≈ 400 W/(m·K)) over 0.2 m, the required gradient is enormous — order 500 °C — because sensible heat only rides along on the modest heat capacity of the atoms it warms.

A heat pipe cheats the gradient by transporting heat as a phase change. At the hot (evaporator) end the working fluid boils, absorbing its latent heat of vaporization L. For water, L ≈ 2260 kJ/kg at 100 °C (2450 kJ/kg near room temperature) — colossal compared with the sensible heat you'd store by warming the same water a few degrees (c ≈ 4.18 kJ/(kg·K)). The heat carried is simply

• Q = ṁ · L, where ṁ is the mass evaporation rate.
• To move 100 W with water: ṁ = Q/L = 100 W / (2.26 × 10⁶ J/kg) ≈ 4.4 × 10⁻⁵ kg/s, about 44 milligrams per second.

That trickle of vapor carries a hundred watts. Because evaporation and condensation both happen at (nearly) the saturation temperature set by the vapor pressure, the whole vapor core sits at almost one temperature — the hallmark near-isothermal behavior that makes a heat pipe look like a super-conductor of heat.

The closed loop: evaporate, fly, condense, wick back

A heat pipe is a sealed tube holding a working fluid at its own saturation pressure — the air is pumped out so the pressure inside is set by the fluid, not the atmosphere. Three zones repeat a cycle forever, driven only by the heat you feed in:

• Evaporator: heat in → liquid boils → local vapor pressure rises.
• Adiabatic section: the small pressure difference drives vapor from the hot end toward the cold end. Vapor speeds are typically 1–10 m/s, occasionally near sonic in start-up or high-flux designs.
• Condenser: heat out → vapor condenses back to liquid, releasing L → pressure drops locally.

The returning liquid is the crux. In a wicked heat pipe, a porous structure — sintered copper powder, fine mesh screen, or axial grooves — lines the wall. The liquid soaks into it and is pulled back to the evaporator by capillary action, the same effect that draws water up a paper towel. This is why a properly wicked heat pipe works against gravity, in any orientation, and even in the weightlessness of orbit. A bare tube with no wick (a thermosyphon) relies on gravity to drain condensate downhill, so it only works with the hot end below the cold end.

The engine that pumps it: capillary pressure and the Young–Laplace law

The pump has no moving parts because it is surface tension. Inside the fine pores of the wick, the liquid meets vapor across a curved meniscus, and a curved liquid–vapor interface sustains a pressure jump — the Young–Laplace equation:

• ΔP_cap = 2σ·cos θ / r for a pore of radius r, wetting angle θ, and surface tension σ.
• For water, σ ≈ 0.072 N/m at 20 °C; for a perfectly wetting fluid cos θ ≈ 1, so engineers write ΔP_cap = 2σ/r_eff with an effective pore radius r_eff.

Take r_eff = 20 μm (a typical sintered wick): ΔP_cap = 2(0.072)/(2 × 10⁻⁵) ≈ 7,200 Pa. That is the maximum suction the wick can generate. The pipe runs only if this capillary head beats the sum of every pressure loss around the loop — the capillary limit:

• ΔP_cap,max ≥ ΔP_liquid + ΔP_vapor + ΔP_gravity
• ΔP_liquid: viscous drag pulling liquid back through the wick (Darcy flow, ∝ μ_l·Q/(permeability)).
• ΔP_vapor: friction on the vapor stream flying to the condenser.
• ΔP_gravity = ρ_l·g·L·sin ψ, the hydrostatic penalty when the evaporator sits above the condenser at tilt ψ.

Note the crucial trade-off in r_eff: smaller pores pump harder (bigger 2σ/r) but also fight liquid flow harder (permeability falls with pore size). Wick design is the art of balancing these.

Why the core is isothermal: Clausius–Clapeyron ties pressure to temperature

The magic of near-isothermal operation follows from the Clausius–Clapeyron relation, which links how saturation pressure changes with temperature along a phase boundary:

• dP/dT = L / (T · Δv), and for a vapor treated as ideal gas this becomes dP/dT ≈ L·P / (R·T²).

For water near 100 °C, dP/dT is roughly 3.6 kPa/K. Turn that around: the small vapor pressure difference that drives the flow (a few hundred pascals, easily supplied by the capillary head) corresponds to only a fraction of a kelvin of temperature difference between evaporator and condenser vapor. So the vapor core is nearly isothermal by physics, not by clever engineering.

The temperature drops you actually measure across a working heat pipe are dominated by the wall and wick — conduction through the metal wall and the thin liquid film at each end — not by the vapor transport. That is why an entire 20 cm heat pipe carrying 50 W might show an end-to-end ΔT of only 2–5 °C, giving an effective conductivity of tens of thousands of W/(m·K) when you back it out of Q = k_eff·A·ΔT/L.

The limits: dryout, sonic, entrainment, and boiling

A heat pipe is not magic — push it too hard and it fails, and there are five classic ceilings that each cap the transported power Q as a function of temperature. The lowest curve at a given operating temperature wins:

• Capillary limit: the wick can't return liquid fast enough; the evaporator dries out, its temperature spikes, and heat transport collapses. This dominates most terrestrial designs.
• Sonic limit: at low temperature (thin vapor), the vapor stream can reach Mach 1 and choke, like flow through a nozzle — a start-up problem for liquid-metal pipes.
• Entrainment limit: fast vapor shears droplets off the liquid surface and drags them back to the condenser, starving the wick. Governed by the Weber number We = ρ_v·v²·L_c/σ ~ 1.
• Boiling limit: at very high heat flux, bubbles nucleate inside the wick, block liquid return, and create a vapor blanket at the wall.
• Viscous limit: at ultra-low temperature the vapor is so viscous it barely flows at all.

The working range is bounded by the fluid too: a copper–water pipe is useful roughly 25–150 °C. Below the freezing point water is useless (use ammonia, ethanol, or methanol); for hundreds of °C you switch to sodium or potassium heat pipes with L of millions of J/kg; and near 4 K, helium heat pipes cool superconducting magnets.

Where they hide: laptops, spacecraft, and power reactors

Heat pipes are everywhere precisely because they are silent, passive, and reliable — no bearings to seize, no pump to fail. Real deployments and their numbers:

• Consumer electronics: a flat copper–water heat pipe in a laptop moves 15–60 W from the CPU to a fin stack near the fan. Vapor chambers — flat, two-dimensional heat pipes — spread hundreds of W under a GPU or flagship phone SoC, defeating hot spots.
• Spacecraft: ammonia heat pipes are the backbone of satellite thermal control; the ISS radiators use them to reject tens of kW. With no gravity, the wick's capillary pumping is not a nicety — it is the only return mechanism.
• Loop heat pipes (LHPs): separate vapor and liquid lines with the wick concentrated in a compact evaporator let heat travel many meters around bends — used on spacecraft and in avionics.
• Energy & industry: heat-pipe heat exchangers recover waste heat in HVAC; sodium heat pipes cool leading edges of hypersonic vehicles and are being designed into microreactors (e.g., NASA's Kilopower/KRUSTY used sodium heat pipes to carry reactor heat to Stirling converters).
• Everyday: Apollo used them; ground-source "thermosyphons" keep the Trans-Alaska Pipeline's permafrost frozen with passive ammonia loops.

Common misconceptions

"It's just a good conductor." No — a solid conductor moves sensible heat and obeys Fourier's law with a fixed k. A heat pipe's effective k isn't a material property; it depends on power, temperature, orientation, and how close you are to a limit. Drive it past dryout and its "conductivity" plunges by orders of magnitude in an instant.

"It needs to be vertical / gravity does the work." Only a thermosyphon does. A properly wicked heat pipe pumps liquid uphill by capillarity and works in any orientation, including in orbit. Gravity merely adds or subtracts the ΔP_gravity = ρ_l·g·L·sin ψ term in the capillary balance.

"More fluid is better." Overfilling floods the condenser and shrinks the active area; underfilling starves the wick. The charge is tuned to grams — often just enough to saturate the wick plus a small reserve.

"It creates or amplifies energy." It obeys the first and second laws exactly: heat still flows from hot to cold, and the pipe only transports it, taking a small ΔT as its "toll." What's remarkable is how little ΔT that toll costs — the phase change, not any free lunch, is doing the heavy lifting.

Solid conduction vs. heat-pipe latent transport for ~100 W over 0.2 m in a slender bar

Property	Solid copper rod	Copper–water heat pipe	Why it differs
Transport mechanism	Sensible heat via phonons/electrons	Latent heat of vaporization	Phase change carries ~2260 kJ/kg
Effective k	~400 W/(m·K)	~10⁴–10⁵ W/(m·K)	Vapor moves heat, not the wall
ΔT end-to-end (100 W)	Tens of °C	A few °C	Near-isothermal vapor core
Mass to carry heat	Heavy solid cross-section	Grams of water + thin wall	ṁ ≈ 44 mg/s at 100 W
Directionality	Symmetric	Can be gravity-aided or wick-driven	Orientation sets ΔP_grav term
Failure mode	Just gets hot	Dryout at capillary limit	Wick can't return liquid fast enough