A vehicle returning from orbit must shed an enormous kinetic energy, and it does so by heating the gas around it. Only a small fraction of that energy reaches the surface, yet the fraction is enough to raise unprotected aluminium or composite structure far above its working temperature within seconds. The thermal protection system, or TPS, is the set of materials and attachments that keeps the heat out. Capsules and probes have generally relied on ablators, which consume themselves in a controlled way, and the modern lightweight ablators descended from phenolic impregnated carbon are very effective (Tran et al., 1997; Natali et al., 2016). An ablator that recedes on every flight, however, sits uneasily with the idea of a vehicle that is turned around like an aircraft.

Reusable systems therefore depend chiefly on materials that reject heat by re-radiation and do not change shape: insulators that run hot at the surface and cool at the bond line, and hot structures that carry load at temperature. The Space Shuttle Orbiter showed that such a system could fly repeatedly, and also how much labour it could demand between flights (Curry, 1993). Interest has revived with partially and fully reusable launchers and with spaceplanes and lifting re-entry demonstrators, and reviews of the wider field are available (Uyanna and Najafi, 2020). This article concentrates on the reusable case, and on the link between materials and the tests used to qualify them.

The article is a narrative review of the open literature: journal papers, agency technical reports and standard texts on high-temperature materials and hypersonic heating. It makes no claim to exhaustive coverage. Much of the detailed design data for operational vehicles is proprietary or export controlled, and the published record is correspondingly weighted towards government programmes and laboratory studies. Property values are discussed qualitatively unless they are firmly established, because reported figures depend strongly on processing route, fibre architecture and test condition.

Convective heating at the stagnation point of a blunt body scales inversely with the square root of the nose radius, a result that follows from the classical analysis of Fay and Riddell (1958) and that explains why the Orbiter was blunt and why sharp leading edges are so much harder to protect. Behind the bow shock the air is dissociated, and atoms that recombine on the surface release their heat of formation there. A surface that is a poor catalyst for recombination therefore receives noticeably less heat than a fully catalytic one (Goulard, 1958), which makes catalytic efficiency a material property of the first importance alongside emissivity. The gas chemistry and its coupling to the wall are treated at length by Anderson (2006).

Heating is only part of the load. The TPS must tolerate launch acoustics and vibration, the strain of a cryogenic tank shrinking beneath it, rain and ground handling, impact by debris and ice, and the oxidising, low-pressure atmosphere of entry, all of it repeatedly. For a reusable vehicle the governing figure of merit is less the peak temperature capability than the cost of returning each square metre to flight condition. Design issues of this kind for hypersonic vehicles generally are surveyed by Sziroczak and Smith (2016).

Reusable TPS concepts can be grouped by how they handle the heat and by the severity of the environment they suit. Figure 1 arranges the main families in that way, and Table 1 summarises their principal strengths and life-limiting weaknesses.

No ground facility reproduces entry in full. Each matches some parameters and distorts others, and a qualification programme is an exercise in combining partial views (Laub and Venkatapathy, 2004).

Arc-jet facilities heat air with an electric arc and expand it over a sample, giving a long-duration, high-enthalpy, chemically reacting flow. They are the standard for demonstrating a TPS at flight-like heat flux and surface pressure. Their limitations are well known: the stream may be contaminated by electrode material, its composition and enthalpy are hard to measure, and model sizes are small. Inductively coupled plasma torches produce a cleaner subsonic stream and are favoured for catalycity and oxidation studies. In both, interpreting the result requires a model of the facility flow, and in situ diagnostics have become part of good practice (Marschall and Fletcher, 2010). Because arc-jet time is costly, a material rarely accumulates many cycles there.

Simpler tests fill the gap. The oxyacetylene torch test standardised as ASTM E285 is cheap and severe and is widely used for screening, although its combustion products are not air and its results rank materials more reliably than they predict flight performance. Radiant lamp arrays and furnaces apply realistic temperature histories to full-size panels and can run hundreds of cycles, but without flow there is no shear, no atomic oxygen and no catalytic heating. Laser and solar furnace heating reach very high flux on small coupons. Mechanical, acoustic, vibration, impact and rain-erosion tests are run separately, usually at room temperature. The weakness of the whole approach is that mechanisms interact in flight: a coating crazed by acoustic loading oxidises differently from a pristine one, and few rigs combine the loads.

The first open question is life prediction. Oxidation kinetics, coating crack growth and creep can each be characterised, but a validated method that turns a handful of short plasma exposures into a certified number of flights does not yet exist in the public domain. Accelerated testing is hard to justify when the damage mechanism changes with temperature and oxygen pressure, as the passive-to-active transition does.

The second is inspection. Rapid reuse needs a way to clear a large surface quickly, which points towards automated optical and thermographic scanning and embedded sensing, and towards acceptance criteria that say which defects may fly. The Orbiter experience was that inspection and repair, not the materials themselves, set the turnaround time (Curry, 1993).

Joints, seals and attachments form the third. A vehicle combines families, and the interfaces between a hot composite and a cool tank wall, or between adjacent tiles, are where hot gas ingress and thermal mismatch cause trouble. They receive less published attention than the materials on either side. The fourth concerns the balance of modelling and test. Coupled simulation of flow, surface chemistry and material response has advanced, yet the surface reaction data that such models need are sparse and facility dependent (Marschall and Fletcher, 2010). Finally there is the choice between passive and actively cooled concepts. Transpiration or film cooling could relax the demands on materials at the hottest points, at the price of plumbing, coolant mass and new failure modes.

Reusable thermal protection has moved from a single flown architecture, the tile and carbon-carbon system of the Orbiter, to a broader palette in which ceramic matrix composites carry load at temperature, metallic panels and tougher tiles compete for the acreage, and ultra-high temperature ceramics are candidates for the sharpest features. The materials science of each family is reasonably mature. What lags is the evidence on durability across many flights, because the facilities that best represent entry are the least suited to repeated cycling. Progress on reuse will depend as much on test strategy, inspection and joining as on new compositions.

To be declared by the author(s) before publication.

To be declared by the author(s) before publication.