Thermal control — multi-layer insulation (MLI), heaters, heat pipes, radiators: Question bank

This page is a workout for your intuition, not your calculator. Every item below targets a place where thermal control "common sense" quietly lies to you — the sign of a temperature difference, what a vacuum lets you do, why gravity ruins a heat pipe on your lab bench. Read the question, commit to an answer out loud, then reveal.

If any word here feels new, the parent topic builds every tool from scratch. Prerequisite pictures live in Stefan-Boltzmann Law, Heat Transfer in Vacuum, Phase Change Heat Transfer, and Capillary Action.

Definition Every symbol used on this page (read once, refer back)

$T$ , $T_{1}$ , $T_{2}$ — absolute temperatures in kelvin (K) of a surface or of two facing surfaces.
$T_{r}$ — temperature (K) of a radiator surface; $T_{s}$ — temperature (K) of the space sink it faces (~3 K).
$σ_{SB} = 5.67 \times 1 0^{- 8} W/m^{2} K^{4}$ — the Stefan–Boltzmann constant (fixed number of nature).
$σ_{γ}$ — surface tension of a liquid, units N/m. (Different quantity from $σ_{SB}$ — we give them different subscripts so they never blur together.)
$ε$ — emissivity, a dimensionless fraction from 0 to 1: how well a surface radiates compared to a perfect blackbody ( $ε = 1$ ).
$ε_{eff}$ — the combined emissivity of two facing surfaces exchanging heat; $ε_{MLI}$ — the whole-blanket effective emissivity of $N$ shields.
$A$ — radiating/exchanging area, m². $N$ — number of MLI shield layers.
$\overset{m}{˙}$ — vapor mass flow rate in a heat pipe, kg/s. $L_{v}$ — latent heat of vaporization, J/kg (energy to boil 1 kg).
$r_{pore}$ — wick pore radius, m; $θ$ — liquid–wall contact angle.
$F_{12}$ — view factor: the fraction of radiation leaving surface 1 that actually lands on surface 2 (geometry only, 0 to 1).

Formula Where the three MLI/radiator formulas come from

Two-surface exchange. A blackbody emits $σ_{SB} T^{4}$ per unit area. Two facing surfaces each emit toward the other and each absorb part of what arrives; tracking the endless back-and-forth of reflected radiation and summing the geometric series gives the net flow $Q = ε_{eff} σ_{SB} A F_{12} (T_{1}^{4} - T_{2}^{4}), \frac{1}{ε _{eff}} = \frac{1}{ε _{1}} + \frac{1}{ε _{2}} - 1.$ The $- 1$ appears because a photon leaving surface 1 is only finally lost after it fails to be reflected back — the resistances $1/ ε_{1}$ and $1/ ε_{2}$ share one common "gap," so you subtract one to avoid double-counting it. For large parallel plates $F_{12} = 1$ .

N-shield stack. Insert $N$ floating foils, all with the same emissivity $ε$ on both faces, each well-spaced so no two touch. Each foil is one more radiation "resistor" in series; $N$ shields create $N - 1$ equal gaps between the enclosing walls, so the total resistance is $(N - 1)$ times one gap's, giving $ε_{MLI} \approx \frac{ε}{N - 1} .$ Domain: this holds only for identical, non-touching foils with negligible edge and conduction effects; real blankets with seams and compression do worse.

True or false — justify

TF1. "MLI keeps a spacecraft warm the same way a wool blanket keeps you warm."

False. A blanket traps air to block convection; in vacuum there is no air, so MLI instead blocks radiation by stacking mirror-like reflective (low-

ε

) layers, each bouncing ~95% of infrared back. See Heat Transfer in Vacuum.

TF2. "Adding more MLI layers always lowers the heat leak."

False. The formula

ε_{MLI} \approx ε / (N - 1)

suggests it, but past ~30 layers the blanket's own weight and launch vibration press layers together, creating conduction shorts that add heat back, plus a mass penalty.

TF3. "A radiator radiates only when the spacecraft is hot; when it cools down it stops."

False. A radiator emits

ε σ_{SB} A T^{4}

for any

T > 0

; it never switches off. That is exactly why cold-side components need heaters — they keep radiating into 3 K space even when you want them warm.

TF4. "Because space is at 3 K, the radiator's net rejection depends strongly on the space temperature."

False. With

T_{r} \sim 300 K

and

T_{s} = 3 K

the ratio

(300/3)^{4} = 1 0^{8}

, so

T_{s}^{4}

is about a hundred-millionth of

T_{r}^{4}

— utterly negligible, and the radiator behaves as if space were at absolute zero.

TF5. "Heat pipes need electrical power to move heat."

False. A heat pipe is fully passive: it runs on evaporation at the hot end and capillary return through a wick, with no pump and no power. See Phase Change Heat Transfer.

TF6. "A heat pipe works equally well in any orientation."

False on the ground. Capillary pressure (~1 kPa) cannot beat gravity's head (~10 kPa per metre) if the condenser sits below the evaporator. In microgravity capillary forces dominate and orientation truly stops mattering.

TF7. "Two surfaces at the same temperature still exchange net radiative heat."

False.

Q \propto (T_{1}^{4} - T_{2}^{4})

; equal temperatures make this zero. Radiation still flies both ways, but the net flow is zero — thermal equilibrium.

TF8. "High emissivity is good for a radiator, so it must be good everywhere on the spacecraft."

False. High

ε

is desired only where you want to dump heat. On insulated surfaces you want the opposite — very low

ε

(the aluminized Mylar of MLI) so heat stays put.

TF9. "A heater and a radiator are opposites, so a well-designed spacecraft uses one or the other."

False. Real spacecraft use both simultaneously — radiators sized for the hot case, heaters filling in during eclipse — because the environment swings between sun-facing

+ 120° C

and shadowed

- 270° C

. See Orbital Thermal Environment.

Spot the error

SE1. "Since

Q = ε_{eff} σ_{SB} A (T_{1}^{4} - T_{2}^{4})

, doubling both surface temperatures doubles the heat flow."

Wrong — the flow depends on

T^{4}

, not

T

. Doubling both temperatures multiplies each

T^{4}

term by 16, so the difference scales by 16, not 2.

SE2. "For a single MLI layer between two surfaces I add the emissivities:

ε_{eff} = ε_{1} + ε_{2}

Wrong. You combine resistances:

1/ ε_{eff} = 1/ ε_{1} + 1/ ε_{2} - 1

. Two poor emitters in series give an even smaller effective emissivity, not a sum.

SE3. "A copper rod and an ammonia heat pipe of equal cross-section move similar heat because copper conducts so well."

Wrong. The heat pipe carries latent heat of a boiling/condensing fluid, moving hundreds of times more heat per unit area than solid copper conduction over the same length. See Phase Change Heat Transfer.

SE4. "The heat a pipe carries is

Q = \overset{m}{˙} L_{v}

, so making

\overset{m}{˙}

arbitrarily large gives unlimited transport."

Wrong. As

\overset{m}{˙}

rises the vapor pressure drop

Δ P

grows, and once it exceeds the wick's capillary pressure

2 σ_{γ} cos θ / r_{pore}

the liquid can no longer return — the pipe "dries out" (the capillary limit).

SE5. "A blackbody radiator emits per unit area

j = σ_{SB} T

Wrong. Stefan-Boltzmann is

j = σ_{SB} T^{4}

; the fourth power is what makes small temperature increases dump vastly more heat. See Stefan-Boltzmann Law.

SE6. "To size a radiator for more waste heat, just cool it down — lower

T

means less leftover heat."

Wrong, backwards. Since

Q = ε σ_{SB} A T^{4}

, a colder radiator rejects less heat, forcing a larger area. Higher allowed temperature means a smaller radiator.

SE7. "MLI's effective emissivity for

N

shields is

ε_{MLI} = εN

Wrong. Each added shield reduces leak:

ε_{MLI} \approx ε / (N - 1)

. More layers make the number smaller, not larger.

SE8. "A heater sized to

5 W

average should be built as a

5 W

element run continuously."

Wrong in practice. Thermostatic heaters cycle on/off, so you install extra capacity (e.g.

10 W

at ~50% duty) to hold setpoint with margin against colder-than-expected cases. See Spacecraft Power Systems.

SE9. "Two facing radiators exchange heat

ε_{eff} σ_{SB} A (T_{1}^{4} - T_{2}^{4})

regardless of how they are aimed."

Wrong. The exchange also carries a view factor

F_{12}

— if the surfaces face away from each other most emitted radiation misses, and the net exchange shrinks toward zero.

Why questions

WHY1. Why can't a spacecraft simply dump heat by conduction into space the way a hot pan cools on a countertop?

Because space is a near-vacuum with essentially no matter to touch; conduction and convection both require a medium, leaving radiation as the only path. See Heat Transfer in Vacuum.

WHY2. Why does the Stefan-Boltzmann law use a fourth power rather than a first power of temperature?

It comes from integrating Planck's spectral curve over all wavelengths; hotter bodies emit both more photons and higher-energy (shorter-wavelength) ones, and these effects compound into a

T^{4}

dependence. See Stefan-Boltzmann Law.

WHY3. Why does a heat pipe move so much more heat than the raw temperature difference suggests?

Because the fluid absorbs a large latent heat when it boils and releases it all when it condenses; the temperature barely changes end-to-end, yet enormous energy per kilogram is shuttled along. See Phase Change Heat Transfer.

WHY4. Why does the returning liquid in a heat pipe climb through the wick without any pump?

Surface tension (

σ_{γ}

) in the fine wick pores creates a pressure difference (capillary action) that literally sucks liquid from the cold condenser back to the hot evaporator. See Capillary Action.

WHY5. Why is aluminized Mylar chosen for MLI rather than, say, thick foam?

MLI fights radiation, so it needs many thin, highly reflective (low-emissivity,

ε \approx 0.03

) mirror layers; foam would add mass and conduction paths without the reflective shielding. See Materials Science — Kapton & Mylar.

WHY6. Why do batteries and propellant tanks specifically get dedicated heaters?

They have hard lower temperature limits — a cold battery can't deliver current and propellant can freeze — so during long eclipse periods heaters replace the heat these masses steadily radiate away. See Spacecraft Power Systems.

WHY7. Why does MLI performance measured in a lab often beat performance on the real vehicle?

Lab blankets are pristine and uncompressed; on a real spacecraft, seams, cable penetrations, and clamped edges create conduction shorts that raise the effective emissivity above the ideal

ε / (N - 1)

WHY8. Why is running a radiator hotter attractive despite tighter component limits?

Because

Q \propto T^{4}

, a modest temperature rise sharply shrinks the required area and mass — but you pay for it with larger heat pipes and stricter component derating to survive the higher temperature.

WHY9. Why does the two-surface exchange formula carry a view factor

F_{12}

at all?

Because radiation travels in straight rays; only the fraction of rays leaving one surface that actually strike the other transfers heat, and that fraction depends entirely on their shape, size, and orientation. See Heat Transfer in Vacuum.

Edge cases

EC1. What is the net radiative exchange the instant two facing surfaces reach identical temperature?

Exactly zero net flow, because

T_{1}^{4} - T_{2}^{4} = 0

; both surfaces still emit, but the incoming and outgoing streams cancel.

EC2. As

N \to \infty

layers, what does the MLI formula predict, and why is that unphysical?

It predicts

ε_{MLI} \to 0

(perfect insulation), which never happens because compression, mass, and conduction shorts impose real limits around 30 layers.

EC3. What happens to a heat pipe if the entire pipe is below the working fluid's freezing point at startup?

The fluid is frozen, so no vapor forms and no heat moves until an external heater thaws the evaporator — a real "startup from frozen" hazard for cryogenic-range pipes. See Cryogenics.

EC4. In the radiator equation

Q_{net} = ε σ_{SB} A (T_{r}^{4} - T_{s}^{4})

, what does letting

T_{s} \to 0

do?

Almost nothing — since

T_{s} \approx 3 K

already makes

T_{s}^{4}

negligible, the deep-space sink behaves as absolute zero and

Q_{net} \approx ε σ_{SB} A T_{r}^{4}

EC5. What happens to heater demand if a satellite's eclipse period suddenly lengthens?

Components radiate for longer with no solar input, so their temperature falls further; heater energy demand (and battery drain to supply it) rises, potentially exceeding the power budget. See Orbital Thermal Environment.

EC6. On the ground, what is the worst orientation to test a heat pipe, and why?

Condenser mounted below the evaporator, forcing the liquid to climb against gravity; the ~1 kPa capillary pressure loses to the ~10 kPa/m gravity head and the pipe fails.

EC7. What limits a radiator's usefulness when the spacecraft surface faces the Sun?

Absorbed solar flux can exceed emitted heat, so the radiator gains net heat instead of rejecting it — designers point radiators at deep space or use optical solar reflectors that emit IR while reflecting sunlight. See Orbital Thermal Environment.

EC8. What happens to MLI's benefit if a single small hole or unblanketed cable penetration exists?

That gap becomes a high-emissivity "leak" whose heat flow can rival many square metres of good MLI, because the surrounding blanket is so good — the weakest patch dominates the total.

EC9. Two radiators sit side by side facing mostly each other rather than space — what is the trap?

Their view factor to deep space drops and their view factor to each other rises, so each partly re-absorbs the other's heat; the net rejection is far less than area alone predicts. See Heat Transfer in Vacuum.

EC10. Is "vacuum stops convection completely" exactly true in a partial vacuum?

No — in low-pressure (rarefied) gas, residual molecules can still conduct heat across small gaps, which is exactly why compressed MLI or a poorly evacuated blanket leaks more than the pure-radiation model predicts.

Recall One-line self-test before you leave

Why does almost every trap on this page trace back to one fact? ::: Because in vacuum only radiation ( $\propto T^{4}$ ) transfers heat, and it never turns off and always flows both ways — forget that and the intuition breaks.