Ceramics and thermal protection

Ceramics survive temperatures that would melt any metal, resist oxidation and wear, and weigh little, but they are brittle. In aerospace they appear in three main roles: thin coatings that protect metal parts in engines, structural ceramic matrix composites that now replace metal in some hot engine parts, and thermal protection systems that shield spacecraft and return capsules from the heat of atmospheric entry. This lesson covers all three. It brings back the civil cargo-return capsule from the AERO-320 wind tunnel campaign, for which Prairiewind Aerospace is now studying a heat shield as a partner to a university space program, and the thermal barrier coatings in the Tern Block 2's turboshaft generator.

Why ceramics are brittle

Ceramics such as alumina, zirconia, silicon carbide and silicon nitride are held together by strong ionic and covalent bonds. Those bonds give high melting points, high stiffness and hardness, but they do not allow dislocations to move easily, so ceramics cannot yield to blunt a crack tip. A small flaw, a pore or a surface scratch, concentrates stress until a crack runs through the part. Fracture toughness of monolithic ceramics is typically 2 to 10 MPa√m, against 30 to 100 for aerospace metals. Their strength is also statistical: two nominally identical parts may fail at very different loads because of different flaw populations, which is described by Weibull statistics. Designers therefore keep monolithic ceramics in compression, small, or in roles where cracking does not lead to loss of the part.

Thermal barrier coatings

In a modern turbine, gas temperatures leaving the combustor can exceed 1500 °C, well above the 1100 to 1150 °C at which nickel superalloys begin to lose strength and melt. Two technologies make this possible. Internal cooling passes compressor air through passages in the blade and out through small holes that form a cooling film. Thermal barrier coatings, TBCs, add a thin insulating ceramic layer on the hot surface. A typical TBC system has a metallic bond coat, often a nickel-cobalt-chromium-aluminum-yttrium alloy or an aluminide, that resists oxidation and grows a thin, protective alumina layer, and a top coat of yttria-stabilized zirconia about 100 to 300 µm thick. The zirconia's low thermal conductivity, about 1 to 2 W/(m K), can lower the metal temperature by 100 °C or more.

TBCs are applied by air plasma spray, which builds a layered, porous coating suited to combustors and vanes, or by electron beam physical vapor deposition, EB-PVD, which grows columnar grains that tolerate the strain of thermal cycling and are used on blades. Their life is limited by spallation: oxidation of the bond coat and thermal cycling eventually crack the interface and the ceramic flakes off. Engine life and inspection planning assume that coating loss will happen and that the cooled metal underneath can survive for a time without it. Teresa Nakamura's group checks that the Tern's turboshaft, at its new constant-power duty, stays within the coating's qualified temperature and cycle range.

Ceramic matrix composites

Ceramic matrix composites, CMCs, embed ceramic fibers, usually silicon carbide, in a silicon carbide or oxide matrix, with a weak interface coating such as boron nitride between fiber and matrix. When a crack runs through the matrix, the weak interface lets it deflect along the fibers, and intact fibers bridge the crack. The material fails gradually instead of shattering. SiC/SiC CMCs have about one third the density of nickel superalloys and can operate at about 1200 to 1300 °C with little cooling. They now fly in turbine shrouds, combustor liners, nozzles and some turbine blades and vanes in recent commercial engines. Oxide/oxide CMCs, cheaper but limited to about 1000 °C, appear in exhaust nozzles and acoustic liners. CMCs need environmental barrier coatings, because water vapor in combustion gases attacks silicon carbide at high temperature.

Atmospheric entry and thermal protection

A spacecraft returning from orbit arrives at about 7.5 km/s with enormous kinetic energy, about 28 MJ for every kilogram of mass. Almost all of it must be converted to heat in the air around the vehicle, and only a small fraction can be allowed to reach the vehicle itself. A blunt shape helps: as Allen and Eggers showed in the 1950s, a blunt body pushes a strong bow shock ahead of it, and most of the heat goes into the shock-heated air rather than the surface. Even so, the stagnation point of a small capsule may see heat fluxes of one to several megawatts per square meter for a few minutes.

Two numbers drive heat shield design. The peak heat flux sets which material can survive at the surface, because each material has a maximum temperature and surface recession rate. The total heat load, the integral of heat flux over time, sets how thick the shield must be to keep the structure underneath, typically aluminum or composite, below its temperature limit. For convective heating at the stagnation point, the Sutton-Graves relation gives a quick estimate: \(\dot q = k\sqrt{\rho/R_n}\,V^3\), with \(\rho\) the air density, \(R_n\) the nose radius, \(V\) the speed and \(k \approx 1.74 \times 10^{-4}\) in SI units for air. A larger nose radius lowers the heat flux, which is one reason capsules are blunt.

# Thermal protection for a small return capsule (the civil cargo-return capsule from AERO-320):
# Sutton-Graves stagnation heating, total heat load and a first ablator thickness estimate.
import math
k_sg = 1.7415e-4                     # Sutton-Graves constant for air, SI units (kg^0.5 / m)
Rn = 0.6                             # nose radius, m
def q_dot(rho, V): return k_sg * math.sqrt(rho / Rn) * V ** 3    # W/m2
# Illustrative entry: speed and density through the heating pulse (time s, V m/s, rho kg/m3)
traj = [(0, 7600, 1e-6), (60, 7550, 2e-5), (100, 7300, 1.2e-4), (130, 6600, 3.5e-4), (160, 5300, 8e-4),
        (200, 3600, 1.6e-3), (250, 2000, 4e-3), (320, 900, 1.2e-2)]
Q, peak = 0.0, (0, 0)
for (t1, V1, r1), (t2, V2, r2) in zip(traj, traj[1:]):
    q1, q2 = q_dot(r1, V1), q_dot(r2, V2)
    Q += 0.5 * (q1 + q2) * (t2 - t1)
    for t, q in ((t1, q1), (t2, q2)):
        if q > peak[1]: peak = (t, q)
print(f"peak stagnation heat flux about {peak[1]/1e6:.2f} MW/m2 at t = {peak[0]} s")
print(f"total heat load about {Q/1e6:.0f} MJ/m2")
# Ablator sizing: recession from effective heat of ablation, plus insulation depth so the bondline
# stays below the structure limit. Values illustrative of a low-density charring ablator.
H_eff, rho_abl = 30e6, 270.0          # effective heat of ablation J/kg, ablator density kg/m3
recession = Q / (H_eff * rho_abl) * 1000
insul = 18.0                          # mm needed so the bondline stays below about 250 C (from a thermal model)
margin = 1.25
print(f"recession about {recession:.1f} mm; insulation depth {insul:.0f} mm; "
      f"design thickness with margin {margin}: {(recession + insul) * margin:.0f} mm")

peak stagnation heat flux about 1.21 MW/m2 at t = 130 s
total heat load about 154 MJ/m2
recession about 19.0 mm; insulation depth 18 mm; design thickness with margin 1.25: 46 mm

The program finds a peak heat flux of about 1.2 MW/m² around 130 seconds after entry interface and a total heat load of about 150 MJ/m². With an illustrative low-density charring ablator, recession is about 19 mm and an insulation depth of about 18 mm keeps the bondline below its limit; with a 1.25 margin the first estimate is about 46 mm. Odile cautions that the estimate ignores radiative heating, which grows quickly at higher entry speeds, and turbulent heating, which can raise flux on the shoulder of the capsule. The real design will use flow solvers and arc-jet tests, in which samples are exposed to a plasma jet that reproduces the heat flux and pressure of entry.

Reusable and ablative thermal protection

Thermal protection systems, TPS, are either reusable or ablative. Reusable systems survive the heating without consuming themselves. The Space Shuttle used silica tiles of very low density and conductivity on most of its underside, felt blankets on cooler areas, and reinforced carbon-carbon on the nose and wing leading edges, which saw the highest temperatures. Damage to the carbon-carbon leading edge from a piece of insulating foam during launch led to the loss of Columbia in 2003, a reminder that TPS must survive the whole mission, not only entry. Reusable metallic and ceramic TPS concepts continue to be developed for space planes.

Ablative systems absorb heat by charring and eroding. As the resin in a phenolic-impregnated carbon or a silica-phenolic material heats, it pyrolyzes, absorbing energy and releasing gases that flow out through the char and thicken the boundary layer, blocking some of the incoming heat. The char layer then erodes slowly. Ablators handle very high heat fluxes and are light for one-time use, which is why they are used on capsules returning from orbit and from the Moon. Low-density ablators such as phenolic-impregnated carbon, PICA, and its variants have protected recent sample return and crewed capsules; denser carbon-phenolic materials protect probes entering atmospheres at much higher speeds.

Inspecting and maintaining thermal protection

Thermal protection must be inspected as carefully as any structure, because small damage can grow catastrophically during entry. The loss of Columbia in 2003 began with foam debris that struck the reinforced carbon-carbon leading edge of the left wing during launch; the breach let hot gas into the wing structure on entry. Since then, inspection of leading edges and heat shields before entry has been standard practice for reusable vehicles. Tiles are inspected for cracks, gouges and missing gap fillers; reinforced carbon-carbon is checked by thermography and ultrasound for internal damage and oxidation of its protective coating; ablative heat shields are inspected after manufacture by radiography and ultrasound for voids and bond defects, because a void in the ablator concentrates heating. Each recovered capsule's heat shield is cored and measured, so that predicted recession and char depth can be compared with what actually happened and the margins confirmed or adjusted.

Gas turbine ceramics have their own maintenance. Thermal barrier coatings fail mainly by spallation, when the thermally grown oxide between the bond coat and ceramic thickens over thousands of hours and the coating flakes off under thermal cycling. Borescope inspections at intervals look for spalled areas on blades, vanes and combustor liners; parts are then removed, stripped, recoated and returned to service. Ceramic matrix composite parts are inspected for oxidation and recession of their environmental barrier coatings, which protect the silicon carbide from water vapor in the combustion gases. Teresa Nakamura notes that engine shop visit intervals are often set by the life of these coatings rather than by the metal parts beneath them.