1. The short version, what to know first, and where this fits

• The star of the heat is compressed hot gas. Low-orbit speed is about 17,500 mph (about 7.8 km/s), and the kinetic energy is about 30 MJ per kilogram, about 46 times the energy needed to warm aluminum from room temperature to its melting point.Our calculation[Refs. 23, 25] The vehicle takes this heat from the hot gas ahead of it.
• There are four ways to get rid of it: wear away and carry it off (ablators), radiate it as light (heat-resistant tiles and the like), store it in thick metal (heat sink), or cool with a coolant or seeping gas. Real vehicles rely mainly on either single-use ablators or reusable tiles and heat-resistant composites.Sourced[Ref. 2]
• The material is decided by properties, process and attachment. The Shuttle tile is mostly air (94% by volume) and needed a compliant mount. On Artemis I, NASA reported that how easily gas escapes from the Orion ablator was behind the char loss.Sourced[Refs. 1, 11]

2. Why it gets hot: compressed hot gas

First, the speed. SpaceX says staying in low orbit needs a sideways speed of about 17,500 mph (about 7.8 km/s).Sourced[Ref. 23] NASA says the Shuttle orbiter re-entered at over 17,000 mph,Sourced[Ref. 1] and that Apollo returned from the Moon at more than 25,000 mph.Sourced[Ref. 1] The Stardust capsule re-entered at 12.6 km/s.Sourced[Ref. 5]

At this speed, the air ahead cannot get out of the way and is compressed, forming a layer of hot gas behind a shock wave. The vehicle takes heat from this hot gas. The ECSS handbook says the thermal loads on a protection system are characterized by very large heat fluxes during relatively short times, and that the surrounding flow is hypersonic, three-dimensional and reactive (involving chemical reactions).Sourced[Ref. 2]

View original SVG, values and sources

Original SVG reference: Conceptual diagram (vector drawing). Schematic of why re-entry is hot, centered on compression heating of the gas. The point that it is not friction heat is our commentary.

(1) Seeing the energy of speed in numbers

The kinetic energy per kilogram is ½ × speed squared. At about 7.82 km/s it is about 30.6 MJ (about 8.5 kWh).Our calculation Meanwhile, warming solid aluminum from 298 K to its melting point of 933 K takes about 0.67 MJ/kg, calculated from NIST-JANAF heat-capacity data.Our calculation[Ref. 25] So the kinetic energy at orbital speed is about 46 times the energy needed to warm aluminum to its melting point; for a lunar return (about 11.2 km/s) about 94 times, and for Stardust (12.6 km/s) about 119 times.Our calculation

More detail: the calculation

Speed conversion uses 1 mph = 0.44704 m/s. 17,500 mph = 7.823 km/s and 25,000 mph = 11.176 km/s. Kinetic energies: 7.823² / 2 = 30.6, 11.176² / 2 = 62.5 and 12.6² / 2 = 79.4 (all MJ/kg).Our calculation For aluminum, the NIST-JANAF solid heat-capacity equation (298 to 933 K) was integrated to give an enthalpy rise of about 17.97 kJ/mol, divided by the molar mass of 26.98 g/mol (about 0.666 MJ/kg).Our calculation[Ref. 25] The heat of fusion is not included. The heat that actually enters the vehicle is a fraction of the kinetic energy, and that fraction was not checked in this article.

3. How the heat is shed: four concepts

The ECSS handbook divides re-entry thermal protection into ablation, radiation, heat sink and transpiration cooling.Sourced[Ref. 2] The figure summarizes each.

View original SVG, values and sources

Original SVG reference: Conceptual diagram (vector drawing). The four concepts of thermal protection. Classification and descriptions follow ECSS-E-HB-31-01 Part 16A; examples are those confirmed in the references.

A closer reading of the ECSS handbook. Ablation handles heat through loss of surface material (mass loss); phase changes absorb a lot of energy, giving good thermal characteristics, but it cannot be reused. Radiation has two layers: the outer layer withstands the "radiative equilibrium temperature" (the temperature at which the incoming heat is shed as light) and the inner layer insulates the structure. A heat sink is the simplest approach, storing heat in a thick layer of high heat capacity, used on early crewed vehicles (the first two crewed Mercury flights); the melting temperature is its limit, and it is hard to apply to the large heat loads of short re-entries.Sourced[Ref. 2]

4. Radiation can shed only so much

A radiative approach, such as the Shuttle tiles, can be reused. How much can radiation alone shed? When a surface sheds heat by radiation alone, the incoming heat flux q and the surface temperature T balance as q = ε × σ × T⁴ (ε is emissivity and σ is the Stefan-Boltzmann constant, 5.670×10⁻⁸ W m⁻² K⁻⁴).Sourced[Ref. 26] This is the "radiative equilibrium temperature".

Assuming an emissivity of 0.85, at RCC's maximum-use temperature of above 3,000 °F (about 1,922 K), the heat flux radiation can shed is about 66 W/cm².Our calculation[Ref. 1] By contrast, the PICA test condition for Stardust had a peak convective heating of about 1,200 W/cm².Sourced[Ref. 6] Shedding 1,200 W/cm² by radiation alone would put the surface at about 3,970 K,Our calculation far beyond the temperature of any reusable heat-resistant material. This is a rough guide to the dividing line: ablators are used where the heat flux is large, while tiles work where it is lower, as in a gliding return from orbit (our commentary).

5. Ablators: protection by wearing away (PICA, Avcoat)

An ablator protects a vehicle by deliberately wearing away. NASA describes Avcoat, the main material of the Orion heat shield, as a material designed to wear away as it heats up.Sourced[Ref. 13] Materials fall roughly into these groups. The ECSS handbook lists carbon phenolic (a high-strength charring ablator, used on the Galileo probe), silica phenolic (selected for the Huygens probe), phenolic nylon, graphite, silicone polymers, Teflon and AVCOAT 5026 used on the Apollo capsules, and says almost all heat shields are made of carbon-based materials.Sourced[Ref. 2]

View original SVG, values and sources

Original SVG reference: Conceptual diagram (vector drawing). Schematic cross-section of an ablator. The heat-shedding functions organize the ECSS handbook; the cause of spalling is NASA's report. Layer thicknesses are illustrative.

(1) PICA: a light carbon skeleton soaked with resin

PICA (Phenolic Impregnated Carbon Ablator) is a carbon-fiber skeleton (FiberForm) soaked with phenolic resin. Its density is a light 0.27 g/cm³ and it ablates efficiently under large heat fluxes.Sourced[Ref. 9] NASA Ames developed it in the Lightweight Ceramic Ablators program, and it served as a single-piece forebody heat shield on Stardust. A 1996 paper said the Stardust re-entry at 12.6 km/s would be the fastest to date.Sourced[Ref. 5]

PICA was tested in arc jets that simulated the Stardust condition (peak convective heating of 1,200 W/cm² at 0.5 atm), and a code's predictions of recession and internal temperature were matched by adjusting material properties.Sourced[Ref. 6] In another exploratory test at a cold-wall heat flux of about 1,800 W/cm² and a stagnation pressure of 130 kPa, all samples survived.Sourced[Ref. 7] On the other hand, PICA has good thermal properties but is structurally weak; fracture tests showed cracks starting and growing in the phenolic phase.Sourced[Ref. 8] PICA has been used on Stardust, the Mars Science Laboratory (MSL) and Dragon.Sourced[Ref. 8]

On MSL, temperatures inside the heat shield were measured with thermocouples. The data show that the boundary layer transition (when the flow becomes turbulent) occurred at 5 of 7 thermocouple locations before peak heating.Sourced[Ref. 17] It is an example of comparing design assumptions (a fully turbulent heat pulse) with real flight data.

(2) Avcoat: resin filling a honeycomb

Avcoat descends from AVCOAT 5026 used on the Apollo capsules.Sourced[Ref. 2] The heat shield of Orion's first flight test, EFT-1 (2014), followed the Apollo approach and was a single piece made by filling each cell of a honeycomb (a cellular skeleton) with ablator using dedicated injection guns.Sourced[Ref. 10] EFT-1 experienced about 80% of the lunar-return velocity and about 50% of the lunar-return heating rate.Sourced[Ref. 10] For Artemis I (EM-1) in 2022, the Avcoat was changed to blocks (tiles) bonded to a composite skin, on a carrier structure of a titanium skeleton and composite skin.Sourced[Ref. 10]

6. Reusable silica tiles and heat-resistant composites

The Space Shuttle needed light, reusable protection with a design life of 100 missions. NASA explains that the Mercury, Gemini and Apollo capsules were protected by heat shields of phenolic epoxy resin in a nickel-alloy honeycomb, but ablators were bonded directly to the vehicle, heavy and not reusable.Sourced[Ref. 1] The airframe was mainly aluminum, which cannot exceed 350 °F without annealing (softening); the purpose of the thermal protection system is to hold that limit.Sourced[Ref. 1]

View original SVG, values and sources

Original SVG reference: Conceptual diagram (vector drawing). Schematic cross-section of a Shuttle tile. Numbers are those in NASA Facts (coating thickness, gap, temperature swing, LI-900 composition); thicknesses of the other layers are illustrative.

According to NASA Facts, most lower-surface tiles are LI-900 (a high-temperature reusable surface insulation) with a bulk density of 9 pounds per cubic foot (about 144 kg/m³), made of pure silica glass fibers 1 to 3 µm in diameter, 6% solid and 94% air by volume.Sourced [Ref. 1] It is designed to minimize thermal conductivity and mass and maximize thermal-shock resistance: it can be heated to 2,200 °F and plunged into cold water without damage. In exchange it is weak, so the higher-strength LI-2200 (22 pounds per cubic foot, about 352 kg/m³) was used around landing-gear doors and windows.Sourced[Ref. 1] The surface is covered by a coating 0.01 to 0.1 inch thick, sprayed from a glass-powder paint and fired at 2,200 °F for 90 minutes.Sourced[Ref. 1]

Converted with 1 lb/ft³ = 16.0185 kg/m³.Our calculation The FRCI 12-pound figure is from the tile-count table in NASA Facts. Densities are not a ranking, since uses and structures differ.Sourced[Refs. 1, 9]

(1) Attachment: loose and soft

Tiles are bonded to the airframe with a silicone adhesive through a strain isolator pad (a needled Nomex felt). Unlike many adhesives, silicone stays flexible in the cold part of the orbit and keeps good bond strength at re-entry temperatures. The pad limits vibration damage to the tiles during ascent and absorbs the difference in thermal expansion between tile and airframe.Sourced[Ref. 1] Gaps of 0.028 to 0.2 inch between tiles are needed to absorb the difference in expansion and contraction as the exterior temperature swings by up to about 400 °F in orbit. But if a gap is too wide, hot gas flows in during re-entry and damages the seal behind it, so gaps were controlled with "gap fillers" in many places.Sourced[Ref. 1]

(2) Reinforced carbon-carbon (RCC): the hottest places

The hottest places, like the nose cap and wing leading edge, used RCC (reinforced carbon-carbon). According to NASA Facts, RCC has a carbon-fiber composite substrate with layers added against oxidation: a silicon-carbide conversion coating, tetraethylorthosilane (TEOS) impregnation and a sodium silicate sealant. The maximum-use temperature is above 3,000 °F (multi-mission). Because its thermal conductivity is higher than other materials, heat moves internally by radiation from the hottest apex to cooler areas, softening temperature differences. But the internal insulation that protects the structure slows RCC cooling, so it stays hot longer, oxidizes more, and shortens the component's life.Sourced[Ref. 1]

NASA Facts covers other materials too. AFRSI (fibrous insulation blankets) sandwiches pure silica felt between a silica fabric on the hot side and an S-glass fabric, through-stitched with silica thread in a 1-inch grid; one blanket can replace up to 25 tiles. In 1996 AETB-8, with added alumina fiber, was introduced, and tougher "TUFI" tiles were used in high-impact areas such as around the engines. The tile water-repellent treatment (injection of dimethylethoxysilane) burns out at 1,050 °F, so it had to be redone after every flight.Sourced[Ref. 1]

NASA Ames later developed TUFROC (Toughened Uni-piece Fibrous Reinforced Oxidation-resistant Composite). It is described as suitable for reusable entry heating above 2,900 °F, with single-use potential to at least 3,600 °F; it was developed for the X-37 project and used as the wing leading edge of the Air Force X-37B. It is a two-layer system, from a carbon-composite surface into a Shuttle-tile-type material in depth. An improved surface treatment, Advanced TUFROC, is reported to run cooler by reducing catalytic atom recombination at the surface.Sourced[Refs. 18, 19]

7. Learning from damage: Columbia and Artemis I

Damage to thermal protection can lead to loss of the vehicle. Here we organize two cases through a materials lens, within what official investigations have published. Causes are not asserted by inference.

(1) Columbia (2003): the reinforced carbon-carbon leading edge

The Columbia Accident Investigation Board (CAIB) reported that the physical cause of the loss was a breach in the thermal protection on the leading edge of the left wing. Insulating foam separated from the left bipod ramp of the external tank 81.7 seconds after launch and struck the wing near the lower half of RCC panel 8. During re-entry, superheated air entered through this breach, penetrated the leading-edge insulation and progressively melted the aluminum structure of the left wing, weakening it until aerodynamic forces caused loss of control.Sourced[Ref. 15] The Board recommended that NASA develop, validate and maintain a model able to predict the damage threshold of debris impacts on the RCC leading edge and similar parts.Sourced[Ref. 16] Foam impact tests on flat RCC panels were later compared with finite-element analysis.Sourced[Ref. 16]

(2) Orion on Artemis I (2022): char loss

Artemis I returned on 11 December 2022 using a skip re-entry (dipping into the upper atmosphere, slowing, and using lift to extend range). After splashdown, charred material on the heat shield was found to have broken off unexpectedly. NASA announced the cause on 5 December 2024: gases generated inside the Avcoat could not vent and dissipate as expected, so pressure built up, horizontal cracks formed near the surface of the char layer, and some charred material broke off.Sourced[Ref. 11]

A 2026 abstract by NASA authors concludes that the low-permeability Avcoat built up extreme internal gas pressure during the skip entry that could not adequately outgas, leading to cracking and char loss. It also reports that permeable and impermeable regions performed very differently and that ground testing reproduced the anomaly. Artemis II flew a modified trajectory without the skip, splashed down on 10 April 2026 and is reported to have had significantly reduced char loss. For Artemis III and beyond, a corrective action to make a more permeable Avcoat has been reported.Sourced[Ref. 12] This is an abstract, not a final report.

8. Heat protection for a stainless-steel vehicle

The vehicles so far had an aluminum skeleton protected by tiles or ablators. In a stainless-steel vehicle, the metal itself withstands higher temperatures than aluminum, so the thickness and role of the tiles could change. However, the primary information this article could confirm is very limited. In an official update dated 15 September 2026, SpaceX says: Starship is designed to be fully reusable and built to survive re-entry; multiple Starship vehicles survived re-entry despite flight experiments that removed tiles to mimic heat-shield failures and maneuvers designed to stress the vehicle's structure; because of its size and robustness the vehicle can get through re-entry, so the company deliberately took a suborbital approach to flight testing before moving to orbital flights.Sourced[Ref. 23] (company statement)

Separate what is known from what is not. What is known: the company states that it deliberately ran experiments with tiles removed and that vehicles still survived re-entry. What is not known: the stainless-steel alloy, the tile material and structure, what temperatures reach which parts of the vehicle, and the number of reuses. These could not be confirmed in official materials. Also, re-entry from a suborbital flight is a lower-speed condition than re-entry from orbit (about 17,500 mph),Our calculation so such demonstrations do not carry the same meaning as an orbital return (our commentary). In its 15 September 2026 update SpaceX says orbital flights begin with Flight 14.Not yet confirmed[Ref. 23] The completed-launch list on its site shows Flight 14 (28 September 2026, return column "EXPENDED"),Sourced[Ref. 24] but the outcome of that re-entry is not covered in this article.

9. Testing on the ground: arc jets and flight data

Thermal protection materials are tested and qualified on the ground in arc jets, which blow a hot gas stream on a specimen. NASA Ames facilities include the 60-MW Interaction Heating Facility (IHF) and the 10-MW TP3.Sourced[Refs. 21, 22] A 2025 NASA Ames short course explains that test targets are set on a cold wall (a value measured with a water-cooled calorimeter), matching heat flux and pressure at the specimen's stagnation point. It also notes that the hot-wall environment computed by CFD may not be close to the cold-wall environment, depending on the flow enthalpy and the specimen's catalycity.Sourced[Ref. 20]

In real flights, vehicles carry instruments that record temperatures inside the heat shield. On MSL, the MEDLI instrument suite measured temperatures within the PICA.Sourced[Ref. 17] Comparing ground-test predictions with flight data validates the design assumptions. But ground tests have limits in flow uniformity and in reproducing the chemistry and radiation of real flight. In TP3, the flow was reported to be non-uniform at most conditions.Sourced[Ref. 22]

10. How to read the yardsticks and standards

The performance of thermal protection is measured by what thermal environment (heat flux, pressure, time) it withstands and how well. Heat-shield materials are tested in ground arc jets for recession and in-depth temperature, and checked against flight data. A single number such as "withstands X degrees" changes meaning without its conditions (reusable or single use, duration, cold wall or hot wall), so it must be read with care.

The arrangement of the table is Our calculation (organizing). Numbers, titles and status were checked on ecss.nl and the NASA Technical Standards System (17 October 2026). Standard texts were not read; only titles and scope are given.

Reading notes (for materials engineers)

• Ground tests and flight differ. Arc-jet heat fluxes are cold-wall values and drop on a real hot wall. Surviving on the ground does not fully reproduce the chemistry and radiation of flight.
• "Up to how many degrees" is a conditional number. 3,000 °F appears as both the Shuttle re-entry exterior temperature and RCC's maximum-use temperature. Meaning changes with reusable versus single use, and with duration.
• Company or agency statements differ from independent verification. For Starship heat-shield demonstrations only company statements were confirmed here. The Artemis investigation is likewise a NASA authors' abstract; independent verification or a final report could not be confirmed.
• Standards include handbooks and requirements. ECSS-E-HB-31-01 is a design reference, not a requirement. If something says "compliant with a standard", check which document and clause.

11. Glossary

12. References

1. NASA (Kennedy Space Center) "NASA Facts: Thermal Protection System" (Space Shuttle orbiter) — www3.nasa.gov/centers/kennedy/pdf/167473main_TPS-08.pdf
2. ECSS ECSS-E-HB-31-01 Part 16A "Thermal design handbook – Part 16: Thermal Protection System" (5 December 2011) (a handbook, not a standard that sets requirements; checked 17 October 2026) — ecss.nl/hbstms/ecss-e-hb-31-01-part-16a-thermal-design-handbook-part-16-thermal-protection-system-5-december-2011/
3. ECSS ECSS-E-ST-31C "Thermal control" (15 November 2008) (checked 17 October 2026) — ecss.nl/standard/ecss-e-st-31c-thermal-control/
4. NASA NASA-STD-6016 "Standard Materials and Processes Requirements for Spacecraft" (Version C with Change 1, change date 15 November 2023) (NASA Technical Standards System; Active; checked 17 October 2026) — standards.nasa.gov/standard/NASA/NASA-STD-6016
5. Tran H. K. et al. (NASA Ames) "PICA Forebody Heatshield Qualification for the Stardust Discovery Class Mission" (1996) — ntrs.nasa.gov/citations/20020040895
6. Covington M. A. (NASA Ames) "Performance of a Light-Weight Ablative Thermal Protection Material for the Stardust Mission Sample Return Capsule" (2005) — ntrs.nasa.gov/citations/20070014634
7. Agrawal P. et al. (NASA Ames) "Investigation of Performance Envelope for Phenolic Impregnated Carbon Ablator (PICA)" (2016) — ntrs.nasa.gov/citations/20160000305
8. Agrawal P., Chavez-Garcia J. F. (NASA Ames) "Fracture in Phenolic Impregnated Carbon Ablator" (2011) — ntrs.nasa.gov/citations/20110015019
9. Stackpoole M. et al. (NASA Ames) "Development of Low Density Flexible Carbon Phenolic Ablators" (2011) — ntrs.nasa.gov/citations/20110015026
10. Koenig W. J. et al. "Orion Heat Shield Manufacturing Producibility Improvements for the EM-1 Flight Test Program" (2019) — ntrs.nasa.gov/citations/20190001310
11. NASA "NASA Identifies Cause of Artemis I Orion Heat Shield Char Loss" (5 December 2024) (checked 17 October 2026) — www.nasa.gov/missions/artemis/nasa-identifies-cause-of-artemis-i-orion-heat-shield-char-loss/
12. Ferguson J. C. et al. (NASA) "Under Pressure: The Artemis I Heatshield Char Loss Investigation and Artemis II's Successful Entry" (conference abstract, 2026) (an abstract by NASA authors, not a final report) — ntrs.nasa.gov/citations/20260006849
13. NASA "FAQ: NASA's Artemis Campaign and Recent Updates" (Orion heat shield: about 5,000 °F, 16-foot diameter, Avcoat) (checked 17 October 2026) — www.nasa.gov/missions/artemis/faq-nasas-artemis-campaign-and-recent-updates/
14. NASA Johnson Space Center "5,000°F" (Houston We Have a Podcast, episode 66, 12 October 2018) — www.nasa.gov/podcasts/houston-we-have-a-podcast/5000f/
15. Columbia Accident Investigation Board "Report Volume I" (August 2003) — govinfo.library.unt.edu/caib/news/report/pdf/vol1/full/caib_report_volume1.pdf
16. Lyle K. H. (NASA Langley) "Test-Analysis Correlation for Space Shuttle External Tank Foam Impacting RCC Wing Leading Edge Component Panels" (2008) — ntrs.nasa.gov/citations/20080014232
17. Edquist K. T. et al. (NASA) "Mars Science Laboratory Heatshield Aerothermodynamics: Design and Reconstruction" (2013) — ntrs.nasa.gov/citations/20140000343
18. Feldman J. et al. (NASA Ames) "TUFROC Thermal Protection System" (2019) — ntrs.nasa.gov/citations/20190030273
19. Feldman J. D. (NASA Ames) "Advanced TUFROC Thermal Protection System" (2021) — ntrs.nasa.gov/citations/20210022886
20. Morgan J., Skokova K. (NASA Ames) "Thermal Protection System Testing and Evaluation at NASA Ames Arc-Jet Complex" (short course, 2025) — ntrs.nasa.gov/citations/20250007899
21. Gokcen T., Skokova K. A. (NASA Ames) "CFD Simulations of the IHF Arc-Jet Flow: Compression-Pad/Separation Bolt Wedge Tests" (2017) — ntrs.nasa.gov/citations/20160014906
22. Gokcen T., Alunni A. I. (NASA Ames) "Flow Characterization Studies of the 10-MW TP3 Arc-Jet Facility: Probe Sweeps" (2016) — ntrs.nasa.gov/citations/20190000376
23. SpaceX "Updates" (15 September 2026 entry "Starship to Orbit") (company statement; checked 17 October 2026; the page is updated over time) — www.spacex.com/updates
24. SpaceX "Launches" (completed-mission list, Starship Flight 14) (company listing; checked 17 October 2026) — www.spacex.com/launches
25. NIST "NIST Chemistry WebBook: Aluminum" (NIST-JANAF solid heat-capacity data, Chase 1998) (checked 17 October 2026) — webbook.nist.gov/cgi/cbook.cgi?ID=C7429905&Units=SI&Mask=2
26. NIST "Stefan-Boltzmann constant" (CODATA value) (checked 17 October 2026) — physics.nist.gov/cgi-bin/cuu/Value?sigma

13. Claim-to-source audit

Last updated: 17 October 2026. Sources are limited to primary materials (NASA, ECSS, CAIB, NIST, official SpaceX pages). Company announcements are treated as company statements. Numerical examples (energy, radiative equilibrium temperature, density conversions) are explanatory calculations made for this article. Details of Starship's stainless steel and tiles, independent verification, inspection methods for ablator permeability, and an international standard for thermal-protection test methods could not be confirmed and are not stated. The Artemis investigation is a NASA authors' abstract, not a final report. Figs. 1 and 9 are AI-generated concept images and do not depict real facilities or products.