How does a heat pipe carry heat?

A sealed copper tube holds a little water. Water evaporates at the hot end (evaporator), vapour flows to the cold end (condenser), condenses and releases heat. Condensate returns to the evaporator by capillary action in a wick (sintered copper powder or fine grooves) on the inner wall. Using latent heat of evaporation and condensation, it moves large heat loads with small temperature differences.

How does it compare with a copper rod?

Carrying 30 W over 200 mm at 6 mm diameter, a copper rod (400 W/mK) has a thermal resistance of 17.7 K/W and needs a 531 K difference — impractical. A heat pipe of 0.15 K/W needs 4.5 K, an effective conductivity of about 47,000 W/mK, roughly 120 times copper.

Why does tilting reduce performance?

Only capillary force returns condensate to the evaporator, so with the evaporator above, lifting liquid against gravity consumes part of it. In a sintered-wick example, maximum transport was about 34 W horizontal, 21 W (64%) at 30° and 9 W (27%) vertical. With the evaporator below, gravity helps instead.

Heat Pipes for Electronic Devices

Category: Thermal Analysis – Electronics Cooling | Revised 2026-10-01

CAE visualization for heat pipe electronics theory - technical simulation diagram

Heat pipes for electronic devices

Theory: carry by latent heat, return by capillarity

Overview

🙋

Why do laptops move CPU heat with heat pipes rather than copper plates?

🎓

In thin devices the cross-section for heat is small and the distance long. Copper conducts well for a metal, but a thin rod carrying tens of watts needs hundreds of kelvin end to end. A heat pipe evaporates water inside, carries it as vapour and condenses it, exploiting large latent heat and fast vapour flow; only a few kelvin are needed. But liquid returns solely by capillary action in the wall wick, so transport has a ceiling that also depends on orientation. Estimating that ceiling is central to design.

Capillary limit

$$ \Delta P_{cap} = \frac{2\sigma}{r_{eff}} \;\ge\; \Delta P_l + \Delta P_g,\qquad \Delta P_l = \frac{\mu_l L_{eff}\,Q}{\rho_l K A_w h_{fg}},\qquad \Delta P_g = \rho_l g L \sin\theta $$

$$ Q_{max} = \left(\frac{2\sigma}{r_{eff}} - \rho_l g L\sin\theta\right)\frac{\rho_l K A_w h_{fg}}{\mu_l L_{eff}},\qquad k_{eff} = \frac{L}{R_{hp}A} $$

σ is surface tension, r_eff effective wick pore radius, K permeability, A_w wick cross-section, h_fg latent heat, L_eff effective length and θ tilt with the evaporator raised. Vapour pressure drop is neglected.

Coffee Break Trivia

From space to the desktop

The heat-pipe principle was patented in the 1940s but developed in earnest in the 1960s by Grover and colleagues at Los Alamos. Because capillarity returns liquid even without gravity, heat pipes were used for satellite thermal control. When laptops spread in the 1990s they became mass-produced parts carrying CPU heat to fans in thin cases, and today smartphones contain thin vapour chambers.

Worked examples

Example 1: copper rod vs heat pipe (6 mm, 200 mm, 30 W)

Item	Copper rod	Heat pipe
Thermal resistance	17.7 K/W	0.15 K/W
ΔT at 30 W	531 K	4.5 K
Effective conductivity	400 W/mK	about 47,000 W/mK

Example 2: tilt and maximum heat transport (sintered wick)

Tilt (evaporator up positive)	Gravity loss	Maximum transport	vs horizontal
−90°	−1,929 Pa	58.3 W	173%
0°	0 Pa	33.7 W	100%
+15°	499 Pa	27.4 W	81%
+30°	964 Pa	21.4 W	64%
+60°	1,670 Pa	12.5 W	37%
+90°	1,929 Pa	9.2 W	27%

Conditions: water at 60 °C (surface tension 0.0662 N/m, viscosity 4.67×10⁻⁴ Pa·s, density 983 kg/m³, latent heat 2,358 kJ/kg), wick 0.5 mm thick (7.7×10⁻⁶ m²), permeability 5×10⁻¹¹ m², pore radius 50 μm (capillary pressure 2,648 Pa), effective length 0.15 m.

🙋

With the evaporator below, it carries 1.7 times the horizontal value.

🎓

In terms of the capillary limit alone, yes: gravity pushes liquid back to the evaporator. In practice other limits may come first, such as the boiling limit (too vigorous boiling in the evaporator) or the entrainment limit (fast vapour stripping liquid), so you can't always use that much. The design problem is the opposite orientation: laptops used upright, or layouts with the CPU above the fan, lose a lot. Finer pores raise capillary pressure but lower permeability, slowing liquid flow, so wick design balances the two.

Design and analysis workflow

Set heat load and the worst plausible orientation (evaporator up).
Estimate capillary, boiling, entrainment and sonic limits; the lowest governs.
Model the heat pipe in thermal analysis as a high-conductivity rod or thermal resistance.
Account for performance loss from bending and flattening.
Measure temperatures at various orientations to confirm margin.

Coffee Break Trivia

“A tablet that overheated when upright”

A tablet passed tests lying on a desk, but throttled during long video playback when stood upright on a stand. The processor sat at the upper end of the heat pipe, so standing it up put the evaporator above the condenser. Calculation showed maximum transport fell to about 30% of horizontal in that orientation, too little for the load. A finer-pored wick and a rerouted heat pipe with the evaporator lower solved it.

Common mistakes

Mistakes and fixes

Mistake	Effect	Fix
Horizontal only	Shortfall in other orientations	Check worst orientation
Capillary limit only	Other limits exceeded	Minimum of all limits
Ignoring bends and flattening	Capacity overestimated	Use post-forming values
Treating as infinite conductivity	ΔT missed	Include resistance
Ignoring cold start	Freezing, poor start-up	Check operating range

🙋

I'd like to learn related topics.

🎓

Rate this article

Thank you for your feedback!

Helpful

More details

Report error