Ablative Heat Shield

The problem: a spacecraft is a meteor it does not want to be

A capsule returning from low Earth orbit hits the upper atmosphere at about 7.8 km/s. A capsule returning from the Moon arrives at 11 km/s. Stardust, carrying comet dust, came back at 12.9 km/s — the fastest any human-made object has ever reentered Earth's atmosphere. At those speeds the vehicle is not fast enough to be slowed by friction alone; what slows it is a shock wave. The air piles up in front of the blunt nose faster than it can get out of the way, compresses violently, and heats to thousands of degrees. The shock layer ahead of an Apollo capsule reached roughly 11,000 K — hotter than the surface of the Sun — and the capsule's own surface saw around 1,650 °C.

That heat reaches the vehicle two ways: convection, where the hot shock-layer gas physically scrubs against the surface, and radiation, where the glowing plasma radiates straight at it. At lunar-return speeds convection dominates; at the extreme entry speeds of probes like Galileo into Jupiter, radiation does. Either way, no structural material survives direct exposure. Aluminum melts at 660 °C, titanium at 1,668 °C, steel around 1,500 °C. The shield's entire job is to stand between that plasma and the structure and make sure almost none of the heat gets through.

There are two families of solutions. A reusable insulator — the Space Shuttle's silica tiles, or reinforced carbon-carbon — re-radiates heat and conducts almost none inward, surviving intact to fly again, but it is limited to relatively gentle low-orbit reentry. An ablative heat shield takes the opposite approach: it is designed to be partly destroyed. By sacrificing its own mass it can absorb and reject far more heat than any inert insulator, which is why every capsule that has ever returned from the Moon, and every sample-return probe from deep space, has used one.

The mechanism: the jobs the shield does as it dies

"Ablation" simply means the controlled removal of material from a surface by erosion, melting, or vaporization. In a modern charring ablator the process is more subtle than just melting away. Walk through what happens to a column of shield material as the heat pulse builds:

1. Heating and pyrolysis. The outer resin (almost always a phenolic — a thermosetting polymer that chars rather than melts) heats past ~500 °C and begins to pyrolyze: its long polymer chains break apart chemically into volatile gas (water vapor, CO, CO₂, hydrocarbons, hydrogen) and a solid carbon residue. This reaction is strongly endothermic — it consumes energy to break the chemical bonds, soaking up incoming heat directly.
2. Char formation. What's left behind is a porous, near-pure-carbon skeleton: the char layer. Carbon does not melt at any realistic pressure; it sublimes near 3,500 °C. So the char can run glowing-hot at the surface while remaining solid, and because it is porous and made of carbon fiber it conducts heat very poorly (≈0.5–1.5 W/m·K). It is a self-forming insulating blanket.
3. Radiation. A char surface at 2,000 K radiates energy as the fourth power of temperature (Stefan–Boltzmann, q = εσT⁴). With emissivity ≈0.9 that is about 900 kW/m² thrown back out as infrared — a large fraction of the incoming flux simply re-radiated to the sky and never conducted inward.
4. Blowing. The pyrolysis gas generated deep in the material has to escape, and it does so by percolating outward through the porous char and injecting into the boundary layer. This "blowing" or "transpiration" thickens the boundary layer and physically pushes the hottest shock-layer gas away from the wall. It is the single largest convective-heating reducer in the system, cutting surface heat transfer by anywhere from 30 to 70 percent depending on the blowing rate.
5. Recession. Finally the char itself erodes — it oxidizes, sublimes, and mechanically spalls away — and the surface recedes. Every kilogram of char that leaves carries its absorbed thermal energy off the vehicle as ejected mass. The recession rate is slow, a few millimeters per second at most, which is why a shield only a few centimeters thick survives a heat pulse lasting a minute or more.

The genius of the design is that these five effects stack. The structure behind the shield only ever feels the small fraction of heat that conducts through the low-conductivity char and virgin material after all of that rejection. Engineers track the bondline temperature — the temperature of the glue holding the shield to the aluminum or composite structure — and design the thickness so it never exceeds its limit (about 250 °C for Apollo's adhesive) before the heat pulse ends.

The governing numbers: heat flux, heat of ablation, recession

Design begins with the stagnation-point heat flux — the heating at the nose, where it is worst. The classic engineering estimate is the Sutton–Graves correlation for convective heating:

q_conv = k · sqrt(rho / R_n) · V³

  q_conv = convective heat flux at stagnation point   [W/m²]
  k      = 1.7415 × 10⁻⁴   (SI, Earth air)
  rho    = freestream density                          [kg/m³]
  R_n    = nose radius                                 [m]
  V      = velocity                                    [m/s]

The cube on velocity is the headline. Double the entry speed and the heat flux rises eightfold; that is why a lunar return at 11 km/s sees roughly three times the convective stagnation heating of a low-orbit return at 7.8 km/s — and, once the longer, deeper trajectory and added radiative heating are folded in, a far larger total heat load — which is why interplanetary sample-return missions need exotic ablators. The √(1/R_n) term is why reentry capsules are blunt: a big nose radius spreads the shock and lowers peak heating, the counter-intuitive insight (Harvey Allen, 1951) that made crewed reentry possible.

You don't size a shield from peak flux alone; you integrate flux over the whole trajectory to get the total heat load (J/m²). Then the material's effective heat of ablation, Q*, tells you how much energy each kilogram of sacrificed shield can absorb and reject:

mass loss per area  ≈  (integrated heat load)  /  Q*

  Q* (effective heat of ablation):
    Carbon-phenolic     ~  20 – 40 MJ/kg
    PICA                 ~  high Q* per unit mass (low density wins)
    Sublimation of carbon  ~  ~60 MJ/kg (the ceiling)

A worked sketch for an Apollo-class lunar return makes the scale concrete:

Entry velocity        V    = 11,000 m/s
Nose radius           R_n  = 4.7 m   (Apollo CM, very blunt)
Peak density          rho  ≈ 1×10⁻³ kg/m³ at ~60 km

q_conv = 1.74e-4 · sqrt(1e-3 / 4.7) · (11000)³
       = 1.74e-4 · 0.0146 · 1.33e12
       ≈ 3.4 MW/m²   (order-of-magnitude stagnation heating)

Stagnation shield thickness (AVCOAT, as flown)  ≈ 4 cm
Bondline temperature limit                       ≈ 250 °C
Recession at stagnation point                    ≈ 1–2 cm of the 4 cm

So about half the stagnation-point shield is consumed; the rest is insulation reserve and margin. The exact numbers come out of coupled aerothermodynamic and material-response codes (NASA's FIAT and CHAR solvers), not the hand calculation above — but the hand calculation tells you the thickness will be centimeters, not millimeters or meters, and that velocity is the variable that dominates everything.

Real materials and real specs

Charring ablators are a small, well-characterized family. The two most important are AVCOAT and PICA.

• AVCOAT 5026-39. The Apollo shield. An epoxy-novolac resin loaded with silica fibers and phenolic microballoons, gunned by hand into a fiberglass-phenolic honeycomb matrix bonded to the capsule. Roughly 370,000 honeycomb cells were filled one at a time and X-rayed. Density ≈0.5 g/cm³. Robust, mature, and reflown on NASA's Orion crew vehicle (with the honeycomb now filled by molded blocks rather than by hand-gunning, after Artemis I revealed unexpected char loss).
• PICA (Phenolic Impregnated Carbon Ablator). Developed at NASA Ames in the 1990s. A rigid carbon-fiber felt impregnated with phenolic resin; density ≈0.27 g/cm³, about a quarter of AVCOAT. The low density makes it both a superb insulator and very light, ideal for high-flux but mass-critical missions. It flew on Stardust (12.9 km/s, 2006) and on Mars Science Laboratory / Curiosity. SpaceX manufactures a proprietary variant, PICA-X, for the Dragon capsule heat shield.
• Carbon-phenolic (tape-wrapped). Dense, high-strength, used where heating is brutal: the Galileo Jupiter entry probe (entry at 47.4 km/s, the most extreme atmospheric entry ever) lost nearly half its 152 kg carbon-phenolic shield mass to ablation. Also used on ICBM reentry vehicles.
• SLA-561V. A silicone-based, cork-filled low-density ablator used on the Viking and earlier Mars landers; effective at the lower heat flux of Mars entry but found inadequate for the higher loads of MSL, which switched to PICA.

Ablative shield vs reusable insulator

The choice between an ablator and a reusable thermal protection system (TPS) is the central architectural decision in reentry vehicle design. They are not interchangeable; they live at different points on the heat-flux/reusability curve.

Property	Ablative shield (PICA / AVCOAT)	Reusable insulator (Shuttle tile / RCC)
Survival mechanism	Sacrifices mass: chars, blows, recedes	Re-radiates heat, conducts little; stays intact
Max usable heat flux	Multiple MW/m² (lunar / interplanetary)	~100s of kW/m² (low-Earth-orbit return)
Peak surface temp	Up to ~3,000 °C (carbon char sublimation)	~1,260 °C tile / ~1,650 °C RCC limit
Reusable?	Largely single-use; inspect / replace	Designed for many flights
Mass per area	Low (PICA) to moderate (AVCOAT)	Low for tiles, but fragile
Failure mode	Burn-through, spallation, bondline overheat	Tile loss / RCC breach (Columbia, 2003)
Best for	Capsules, sample return, lunar/Mars/deep-space	Winged, frequently-flown LEO vehicles

The Space Shuttle chose reusability and paid for it in fragility and a ceiling on entry energy: it could only ever come back from low Earth orbit. Apollo, Soyuz, Orion, Dragon, and Starliner all chose ablation because they had to survive lunar-return energy or wanted a simple, robust, single-mission shield. Newer vehicles blur the line — SpaceX's Starship aims for reusable ceramic tiles at near-orbital energy, betting that frequent flights pay back the maintenance.

Why the shield is a blunt dish, not a sharp point

Intuition says a sharp nose should cut through the air with less heating. The opposite is true. A sharp body sits close behind an attached shock, the gas barely deflects, and almost all the kinetic energy converts to heat right at the tiny tip — which promptly melts. A blunt body throws a strong detached bow shock well ahead of itself; most of the shock-heated gas flows around the vehicle and carries its energy downstream rather than dumping it into the surface. H. Julian Allen's blunt-body insight in 1951 is the reason every crewed reentry vehicle is a shallow dish presenting a huge nose radius — and, via the √(1/R_n) term in Sutton–Graves, a large nose radius directly lowers stagnation-point heat flux. The shape that looks least aerodynamic is the one that survives.

Failure modes — where ablators actually break

• Burn-through. The terminal failure. If recession plus margin runs out before the heat pulse ends, the char is gone and the structure cooks. There is no fault tolerance once carbon is consumed, which is why recession allowance is generous.
• Spallation. Chunks of char break off mechanically — from thermal stress, aerodynamic shear, or pressure from trapped pyrolysis gas — before they have finished doing their thermal job. Premature material loss eats into margin and can roughen the surface, raising turbulent heating.
• Bondline failure. The adhesive holding the shield to the structure has a hard temperature limit (~250 °C class). If conduction through the shield raises the glue past that limit, the shield can debond — even if plenty of char remains on the surface. Sizing is driven by bondline temperature, not surface temperature.
• Asymmetric recession / shape change. Uneven ablation shifts the vehicle's center of pressure relative to its center of mass. Because reentry stability depends on that offset, a shape that ablates asymmetrically can pitch, lose its trim angle, and tumble.
• Gap and seam flow. Tiled ablators (MSL's PICA, Orion's blocks) have seams. Hot gas getting into a gap can heat the structure laterally; gap fillers and step/gap tolerances are a major qualification concern.
• Unexpected char loss. Orion's Artemis I shield (2022) came back with more char erosion and cracking than predicted, traced to pyrolysis gases building pressure inside the AVCOAT and cracking the char. The fix for Artemis II changed the trajectory and the material processing — a reminder that ablator response is hard to model precisely even today.

Where ablative heat shields fly

• Crewed return capsules. Apollo (AVCOAT), Soyuz, Orion (AVCOAT), Crew Dragon (PICA-X), Starliner. Every spacecraft that has returned humans from orbit or the Moon used an ablator.
• Sample-return probes. Stardust (12.9 km/s, PICA), Hayabusa and Hayabusa2 (carbon-phenolic), OSIRIS-REx (AVCOAT-derived). The fastest, hottest Earth reentries on record.
• Planetary entry probes. Galileo's Jupiter probe (carbon-phenolic, 47 km/s — the most extreme entry ever flown), the Mars landers Viking through Perseverance, Huygens at Titan, Pioneer Venus.
• Ballistic-missile reentry vehicles. Carbon-phenolic and graphite nose tips, where the entry environment is comparable to a lunar return but the vehicle is deliberately sharp and slender for accuracy.
• Recoverable instruments and demonstrators. NASA's ADEPT and HEART deployable/inflatable ablators, and a long line of arc-jet test articles that never fly but characterize material Q* and recession behavior on the ground.

Common pitfalls when designing an ablative TPS

• Sizing to peak flux instead of integrated load. A short, intense pulse and a long, mild one can need very different shields for the same peak number. Integrate the whole trajectory.
• Forgetting that blowing changes the boundary condition. The heat flux into a blowing surface is much lower than the cold-wall flux you'd compute ignoring ablation. Use it, but don't over-credit it — blowing reduction depends on the gas injection rate, which couples back to the surface temperature.
• Designing to surface temperature, not bondline. The shield can glow at 2,000 °C all day; what kills the vehicle is the glue reaching 250 °C. Bondline temperature at end-of-heating is the real constraint.
• Ignoring trapped-gas pressure in the char. Pyrolysis gas that can't escape fast enough builds internal pressure and cracks or spalls the char (the Orion Artemis I surprise). Permeability matters as much as conductivity.
• Trusting the material model too far. Ablator response is genuinely hard to predict; arc-jet ground tests and flight data still surprise engineers. Carry margin, instrument the shield, and expect to recalibrate.