Engineering Ceramics — Engineering Reference

See also (Tier 3 family index): Ceramics Taxonomy

1. At a glance

A ceramic is an inorganic, non-metallic solid — almost always crystalline, but the glasses that share the chemistry and many of the properties are amorphous. The engineering family splits cleanly into two halves: oxides (Al₂O₃, ZrO₂, MgO, SiO₂-glasses, silicates) and non-oxides (SiC, Si₃N₄, B₄C, AlN, TiB₂, BN, the diamond and CBN superhard pair). Both halves share the same fundamental property package — very high stiffness (E = 200–500 GPa), very high hardness (HV 12–40 GPa), high melting point (T_m 1700–3000 °C), high chemical inertness, and brittleness with low fracture toughness (K_IC = 2–10 MPa√m) and statistical, flaw-dominated strength.

Engineers reach for ceramics where their property advantages outweigh brittleness. The list is narrow but high-value:

• Cutting tools — Al₂O₃, Si₃N₄, c-BN, diamond, WC-Co inserts (Sandvik CBN7050, Kennametal KY4400) — cut steels and superalloys faster than any metallic tool can.
• Rolling-element bearings — Si₃N₄ balls in steel races (hybrid bearings); SKF, NSK, GMN, FAG products run at 1.5–2× the dN limit of all-steel bearings.
• Pump seals, wear plates, slurry components — Al₂O₃, SiC, SSiC, Si₃N₄ where chemical attack or abrasion kill metals.
• Semiconductor processing — alumina, AlN, SiC wafer carriers, susceptors, chamber liners; near-zero outgassing, plasma-resistant.
• Biomedical implants — Al₂O₃ and Y-TZP zirconia femoral heads (Biolox-forte, Biolox-delta ZTA), dental crowns and bridges, zirconia dental implants.
• High-temperature aerospace — SiC/SiC CMC for HPT shrouds in GE/CFM LEAP and PW1100G engines, thermal-protection on Orion and X-37B.
• Electronics substrates — Al₂O₃ (HTCC, LTCC), AlN (power LEDs, IGBT modules), SiC and GaN-on-SiC wide-bandgap power devices.
• Refractories and kiln furniture — recrystallised SiC, fused silica, alumina, MgO, ZrO₂ in steel-making and glass-melting furnaces above 1500 °C.
• Armour — B₄C, SiC, Al₂O₃ tiles on aramid/UHMWPE backing — SAPI/ESAPI plates, vehicle armour, helicopter belly panels.
• Optics and precision substrates — fused silica, Zerodur glass-ceramic, ULE titanium-silicate, SiC mirrors (JWST secondary, LSST, ESO ELT segments).

Where ceramics lose. Tensile-loaded primary structure (brittle, statistical), impact-loaded parts (chips and shatters), high-cycle fatigue under tension (no plastic blunting at flaws), large-section castings (sintering shrinkage limits size), and any geometry that demands secondary machining after firing (diamond grinding alone). They are the opposite of an “iterate during prototyping” material: design intent locked at sintering.

---

2. First principles

2.1 What a ceramic is

A ceramic is an inorganic non-metallic compound, most often a stoichiometric oxide, nitride, carbide, or boride, with mixed ionic + covalent bonding. The bonding is what generates the entire property package:

• Strong directional covalent bonds + non-directional ionic Coulomb attraction → high E, hardness, T_m.
• No free conduction electrons (mostly) → dielectric, optically transparent in pure form (sapphire, fused silica, Y-TZP at thin section).
• No slip systems active at room temperature in ionic and covalent crystals → no dislocation motion → no plastic deformation → brittle fracture from the largest flaw.

The few “ductile” ceramics (rocksalt-structure MgO above 1200 °C, some perovskites near T_m, nanograined alumina under hydrostatic pressure) are laboratory curiosities — engineering practice treats every ceramic as brittle at service temperature.

2.2 Strength is flaw-controlled — the Griffith criterion

Because there is no plastic blunting at stress concentrators, strength is set by the largest defect:

σ_f = K_IC / (Y · √(π · a))

with Y a geometry factor (~1 for surface flaws, 1.12 for half-penny), K_IC the plane-strain fracture toughness, and a the critical flaw size. Typical K_IC: Al₂O₃ 4 MPa√m, Y-TZP 8 MPa√m, Si₃N₄ 7 MPa√m, SiC 3 MPa√m, B₄C 3 MPa√m, glass 0.7 MPa√m. Compare metals (steel 50–200 MPa√m) and engineering polymers (1–3 MPa√m): ceramics live in the K_IC < 10 regime where a 50 µm pore reduces strength by 2–3×.

The corollary: strength varies between parts in the same batch because flaw populations vary. A monolithic ceramic does not have a deterministic σ_y the way steel does. Design uses statistical strength.

2.3 Weibull statistics — the central design tool

Strength follows the two-parameter Weibull distribution:

P_f(σ) = 1 − exp[−(σ/σ_0)^m]

where σ_0 is the characteristic strength (the value at P_f = 0.632) and m is the Weibull modulus — the slope of ln(ln(1/(1−P_f))) vs ln(σ). High m means narrow distribution; low m means broad.

• Metals: m ≈ 50–100 (essentially deterministic).
• Modern hot-pressed Si₃N₄, HIP alumina: m ≈ 15–30.
• Conventional sintered alumina, SiC: m ≈ 10–15.
• Wet-process refractories, glass: m ≈ 5–10.

The standard test for σ_0 and m is ASTM C1239 — Weibull analysis of ≥ 30 flexure bars tested per ASTM C1161 (four-point bend).

Size effect. Larger volumes contain larger flaws on average, so larger parts are weaker:

σ_2 / σ_1 = (V_1 / V_2)^(1/m)

For m = 10, doubling the stressed volume drops mean strength by 6.7 %. For m = 5 it drops by 13 %. Test coupons are tiny (~50 mm³ effective volume); production parts can be 10⁴× larger — strength corrections are non-trivial.

2.4 Toughening mechanisms

Several routes raise K_IC above the intrinsic ~2 MPa√m of pure single-crystal ceramics:

• Transformation toughening — Y-TZP zirconia. Metastable tetragonal grains transform to monoclinic at the crack tip, absorbing energy and putting compressive stress on the crack flanks. Drives K_IC to 8–12 MPa√m. Sensitive to moisture and temperature (LTD, §10m).
• Microcracking toughening — distributed grain-boundary microcracks shield the main crack tip. Common in two-phase oxide systems (Al₂O₃-ZrO₂, ZTA).
• Whisker / fibre reinforcement (CMC) — SiC whiskers in Al₂O₃ (SiC_w-Al₂O₃ cutting inserts, K_IC ≈ 9 MPa√m); continuous SiC fibre in SiC matrix → CMC with quasi-ductile, graceful failure. See [[Engineering/materials-composites]].
• Crack deflection and bridging — at strong second-phase grains (elongated β-Si₃N₄, plate-like α-SiAlON).

2.5 Thermal shock — a separate failure axis

Sudden temperature changes generate transient thermal stresses. The maximum survivable ΔT for crack initiation is captured by Hasselman’s R parameter:

R = σ_f · (1 − ν) / (E · α) [units: K]

Higher R = better resistance. For typical aerospace and process ceramics:

• Fused silica: R ≈ 500 K (α = 0.5 × 10⁻⁶/K, the lowest of any structural material)
• SiC: R ≈ 350 K
• Si₃N₄: R ≈ 450 K
• Al₂O₃: R ≈ 110 K
• Y-TZP: R ≈ 280 K

Two extensions matter: R’ = R · k (high thermal conductivity helps spread the gradient — SiC dominates here at k ≈ 120 W/m·K) and R''' = E / (σ_f² · (1 − ν)) which characterises crack propagation after initiation. Materials selection for thermal shock is usually R’-based.

2.6 Compressive vs tensile strength

A defining ceramic asymmetry: σ_c ≈ 8–15 × σ_t. Alumina σ_t_flex ≈ 350 MPa but σ_c ≈ 2500–3500 MPa. Y-TZP σ_t_flex 1000 MPa, σ_c 2000–2500 MPa. Cracks open under tension but close under compression — only shear and Hertzian-cone flaws propagate. Bearings, dies, anvils, and ball mills exploit this; tensile-loaded designs avoid it.

---

3. Practical math / design equations

3.1 Worked example 1 — Si₃N₄ hybrid bearing ball Hertzian contact

Problem. A 10 mm diameter HIP-Si₃N₄ bearing ball is loaded in pure compression between a flat 52100-steel race and an identical Si₃N₄ ball, total load P = 1000 N. Compute the peak Hertzian contact stress and compare to compressive strength.

Step 1. Contact: ball-on-ball, equal radii R₁ = R₂ = 5 mm. Effective radius R* = R₁·R₂/(R₁+R₂) = 2.5 mm.

Step 2. Effective modulus. Si₃N₄ E = 310 GPa, ν = 0.27. 1/E* = (1 − ν₁²)/E₁ + (1 − ν₂²)/E₂ = 2 · (1 − 0.27²) / 310 GPa. E* = 167 GPa.

Step 3. Contact radius a = (3 · P · R* / (4 · E*))^(1/3) = (3 · 1000 · 2.5×10⁻³ / (4 · 167×10⁹))^(1/3) = (1.123×10⁻¹¹)^(1/3) = 0.224 mm.

Step 4. Peak stress p₀ = 3·P / (2·π·a²) = 3 · 1000 / (2 · π · (2.24×10⁻⁴)²) = 9.5 GPa.

Comparison. Si₃N₄ compressive strength ≈ 3000–3500 MPa one-shot — but Hertzian compressive contact below shear yield is non-damaging. The local sub-surface shear stress τ_max = 0.31 · p₀ = 2.95 GPa, which sits below Si₃N₄ shear yield ≈ 0.5 · σ_c = 1.5–1.75 GPa quasi-statically but is the relevant rolling-contact-fatigue (RCF) driver. Si₃N₄ hybrid bearings achieve RCF L₁₀ life 5–10× a steel ball at the same load because no surface plasticity, lower elastic deflection (smaller contact patch), and better lubricant film generation; the design-allowable p₀ for Si₃N₄ on hardened 52100 is typically 4.0–4.5 GPa for long-life applications. Compare to 52100-on-52100 at 1000 N: E* ≈ 110 GPa, a = 0.256 mm, p₀ = 7.3 GPa — same load gives higher peak stress in ceramic-on-ceramic because the harder ball deflects less and patches are smaller, but ceramic shrugs off the cycle that work-hardens and eventually spalls the steel.

3.2 Worked example 2 — Weibull strength scaling for an alumina substrate

Problem. A 99.5 % alumina electronic substrate is qualified on ASTM C1161 four-point flexure coupons (effective volume V₁ ≈ 0.5 cm³) and gives σ₀ = 350 MPa, m = 12. Predict the design allowable for a 99 % reliability (P_f = 0.01) over a production part with stressed volume V₂ = 100 cm³.

Step 1. Adjust σ₀ for size: σ₀,part = σ₀,coupon · (V₁/V₂)^(1/m) = 350 · (0.5/100)^(1/12) = 350 · (0.005)^(0.0833) = 350 · 0.652 = 228 MPa.

Step 2. Convert to the design allowable at the desired reliability:

σ_design = σ₀,part · [−ln(1 − P_f)]^(1/m) = 228 · [−ln(0.99)]^(1/12) = 228 · (0.01005)^(0.0833) = 228 · 0.681 = 155 MPa

Step 3. Apply a safety factor for time-dependent slow crack growth (SCG): for moist environments with SCG exponent n ≈ 30 (typical alumina), reduce by ~30 % for indefinite life — σ_allow ≈ 110 MPa (16 ksi).

Comment. From a published coupon mean of 350 MPa to a design allowable of 110 MPa is a 3.2× knockdown — the Weibull-times-SCG combination is what makes ceramic design conservative. A higher-m material (HIP Si₃N₄ at m = 25) under identical conditions would give σ_allow ≈ 170 MPa from the same nominal σ₀ — and that is the reason aerospace ceramics are HIPed.

3.3 Worked example 3 — Thermal-shock comparison, SiC vs Al₂O₃ seal face

Problem. A mechanical seal face is splash-cooled by liquid from T = 200 °C process down to ambient T = 25 °C → ΔT = 175 K. Which face survives?

SiC (sintered α-SiC): σ_f = 450 MPa, E = 410 GPa, ν = 0.16, α = 4.3 × 10⁻⁶/K.

R_SiC = 450 × 10⁶ · (1 − 0.16) / (410 × 10⁹ · 4.3 × 10⁻⁶) = 378 / 1.763 = 214 K

Al₂O₃ (96 %): σ_f = 300 MPa, E = 330 GPa, ν = 0.22, α = 8.0 × 10⁻⁶/K.

R_Al₂O₃ = 300 × 10⁶ · (1 − 0.22) / (330 × 10⁹ · 8.0 × 10⁻⁶) = 234 / 2.64 = 89 K

Result. SiC at R = 214 K survives ΔT = 175 K with margin. Alumina at R = 89 K fails by mid-quench. Real pump-seal practice: SiC is the default rotating face in high-thermal-shock duty (water-cooled hot service, slurry pumps), with carbon-graphite or carbon-SiC as the stationary mate; alumina survives only modest ΔT and is restricted to cooler general-service seals.

A second criterion — R’ = R · k — captures the role of thermal conductivity in spreading the gradient. SiC k ≈ 120 W/m·K, alumina k ≈ 25 W/m·K. R’_SiC = 25.7 kW/m, R’_Al₂O₃ = 2.2 kW/m. SiC dominates by 12×. This is the metric that drives selection for plasma chamber liners, kiln furniture, and reactor parts.

---

4. Reference data

4.1 Property summary — engineering ceramics (room temperature, typical)

Material	σ_f flex (MPa)	E (GPa)	K_IC (MPa√m)	HV (GPa)	ρ (g/cm³)	T_max air (°C)
Al₂O₃ 99.5 %	350	380	4.0	17	3.90	1700
Al₂O₃ 96 %	300	330	3.5	14	3.72	1500
Al₂O₃ 92 % (machined wear plate)	250	300	3.0	12	3.65	1400
Y-TZP (3Y)	1000	210	8.0	13	6.05	1000 (LTD limit)
Mg-PSZ	600	200	9.0	12	5.75	1100
Ce-TZP (12 mol%)	500	200	12.0	10	6.18	1100
Zirconia-toughened alumina (ZTA, 20 % ZrO₂)	700	350	6.5	16	4.20	1500
Si₃N₄ (HIPSN)	900	310	7.0	17	3.25	1200
Si₃N₄ (SSN, sintered)	600	290	6.0	15	3.20	1200
Si₃N₄ (RBSN)	250	220	3.0	11	2.50	1400
α-SiC sintered (SSiC, Hexoloy)	450	410	3.0	28	3.15	1600
SiSiC (RBSC)	350	380	4.0	25	3.05	1380
CVD-SiC (optics, semiconductor)	580	460	3.4	30	3.21	1600
B₄C (hot-pressed)	350	460	3.0	30	2.50	1000
AlN	320	320	3.5	12	3.26	1000 (air, oxidises)
TiB₂	400	540	7.0	25	4.50	800 (air)
h-BN	100	30	2.5	1 (soft)	2.10	850
c-BN	800	720	5.0	45	3.48	1200
Diamond (PCD)	1100	1050	6.5	80	3.52	700 (oxidation onset)
Fused silica	50	73	0.8	8	2.20	1100
Zerodur (glass-ceramic)	50	91	1.0	8	2.53	600
Macor (machinable glass-ceramic)	95	67	1.5	2.5	2.52	800
WC-Co (6 % Co cutting grade — cermet, borderline)	2200	630	11	19	14.9	700

σ_f per ASTM C1161 four-point flexure; K_IC per ASTM C1421 (SC or SEPB); HV per ASTM C1326/C1327. Compressive strength is typically 8–15 × tensile flexure. All values are batch-typical; design uses Weibull-adjusted allowables per ASTM C1239.

4.2 Thermal and electrical reference

Material	α (10⁻⁶/K)	k (W/m·K, 25 °C)	ρ_v (Ω·cm)	ε_r (1 MHz)	Dielectric strength (kV/mm)
Al₂O₃ 99.5 %	8.0	30	10¹⁴	9.8	17
Al₂O₃ 96 %	7.5	24	10¹⁴	9.5	14
Y-TZP	10.5	2.0	10¹³	30	12
Si₃N₄ HIPSN	3.2	30	10¹³	8	17
α-SiC	4.3	120	10²–10⁶ (doped)	40	n/a (semiconductor)
B₄C	5.0	30	0.3 (semiconductor)	—	—
AlN	4.5	170	10¹⁴	8.9	14
h-BN (basal)	1.0	600 in-plane / 30 cross	10¹³	4.1	35
Fused silica	0.55	1.4	10¹⁸	3.8	25
Zerodur	0.02 (near-zero)	1.5	10¹⁴	7	20

Read the table. AlN combines metal-grade thermal conductivity (170 W/m·K, ~Cu/3) with full dielectric isolation — the substrate of choice for power IGBT modules and high-brightness LEDs. CVD-SiC mirrors and SiC wafer chucks exploit the same k advantage. Fused silica and Zerodur own precision optics through near-zero α: Zerodur α ≈ 0.02 × 10⁻⁶/K is two orders of magnitude below any structural metal.

---

5m. Composition & microstructure

5m.1 Alumina (Al₂O₃)

The most-used technical ceramic by tonnage. Crystal structure: rhombohedral corundum (α-Al₂O₃, the same as sapphire and ruby). Engineering grades are sintered polycrystalline aggregates, classified by purity:

Grade	Purity	Microstructure	Typical use
AD-90, AD-92	90–92 %	Coarse-grain (5–20 µm), silicate glass phase	Wear plates, grinding media, low-end seals
AD-94, AD-96	94–96 %	Mixed grain, less glass	Pump seals, substrates, valve bodies
AD-98, AD-99	98–99 %	Fine-grain (2–5 µm), minimal grain-boundary phase	High-T electronics, RF windows
99.5 % (AD-995, Coors 998)	99.5 %	Fine, dense, ~< 1 % porosity	Medical implants (with Mg doping), optical
99.9 %	99.9 %	Sub-µm; HIPed	Semiconductor processing, laser tubes
Sapphire (single-crystal)	99.99 %+	Czochralski-pulled single crystal	Watch crystals, IR windows, LED substrate

Purity sets strength and dielectric performance — and cost. 96 % alumina is the workhorse at $20–40/kgfired;99.9$150–400/kg. CoorsTek AD-995, CeramTec Rubalit, Kyocera A-479B, Saint-Gobain Pural are standard reference grades.

5m.2 Zirconia (ZrO₂)

Pure ZrO₂ undergoes destructive monoclinic ↔ tetragonal ↔ cubic phase transitions on cooling — useless monolithically. Stabilised zirconias are the engineering grades:

• Y-TZP (3 mol % yttria, “3Y-TZP”) — metastable tetragonal at room temperature; the transformation-toughened workhorse. σ_f = 800–1200 MPa, K_IC = 6–10 MPa√m. Dental crowns (Ivoclar Vivadent IPS e.max ZirCAD, 3M Lava, Glidewell BruxZir), pump components, knee/hip implants — though hip-head use is now restricted after the 2001 Prozyr LTD field-failure recall (St. Gobain Desmarquest batch).
• Mg-PSZ (Mg-partially-stabilised zirconia) — coarser; tougher (K_IC ≈ 9 MPa√m) but lower strength (~600 MPa). Wear parts, dies, extrusion nozzles.
• Ce-TZP (12 mol % ceria) — highest K_IC of any monolithic ceramic at 12–15 MPa√m; lower strength (~500 MPa). Better LTD resistance than Y-TZP. Used in surgical instruments and ball valves.
• Fully cubic ZrO₂ (8 mol % yttria, YSZ) — no transformation toughening; used as TBC (thermal-barrier coating) on turbine blades and as oxygen sensor electrolyte (Bosch λ-sensor, NTK).
• ZTA (zirconia-toughened alumina) — Al₂O₃ matrix with 15–25 % Y-TZP particles. Combines alumina’s stiffness with zirconia’s toughness. Biolox-delta (CeramTec): the current dominant hip-head material globally.

5m.3 Silicon nitride (Si₃N₄)

Hexagonal α and β phases. β grains develop as elongated rods during sintering — the in-situ toughening mechanism. Grades distinguished by densification route:

• RBSN (reaction-bonded) — Si powder compact nitrided in N₂ at 1250–1400 °C. Net-shape, no shrinkage. Lower strength (250 MPa) because of residual porosity (~20 %). Refractory, thermocouple sheaths.
• SSN (sintered) — pressureless sintering with Y₂O₃ / Al₂O₃ liquid-phase aids. σ_f ≈ 600 MPa. Industrial wear parts, cutting tools.
• HPSN (hot-pressed) — uniaxial pressure during sintering. σ_f ≈ 800 MPa, but only simple shapes (right cylinders, plates).
• HIPSN (hot-isostatically pressed) — isostatic pressure densifies sintered preform. σ_f = 900–1100 MPa, m ≈ 20. Bearing balls, aerospace.
• SRBSN (sintered reaction-bonded) — RBSN preform post-sintered. Combines low shrinkage with higher strength (~700 MPa).

SiAlON (silicon-aluminium oxynitride, α/β-SiAlON) is a Si₃N₄ derivative with Al + O substituted into the lattice. α-SiAlON is harder (HV ~21 GPa); β-SiAlON is tougher. Sandvik CC650, Kennametal KY3500 cast-iron cutting inserts are α/β-SiAlON.

5m.4 Silicon carbide (SiC)

α-SiC (hexagonal 6H/4H polytypes, the equilibrium high-T form) and β-SiC (cubic 3C, the low-T form). Engineering grades:

• SSiC (sintered α-SiC) — pressureless sintered with B + C aids; fully dense, fine grain. Saint-Gobain Hexoloy SA, Wacker SiCWolf, CoorsTek PureSiC. The default engineering SiC: pump seals, slurry valves, kiln furniture.
• RBSC / SiSiC (reaction-bonded, silicon-infiltrated) — green SiC + C preform infiltrated with molten Si. ~10 % residual free Si limits T_max to 1380 °C but enables complex near-net shapes and large parts (telescope mirrors). Saint-Gobain Cerastar, Schunk Sicalit.
• NSiC (nitride-bonded) — refractory grade, Si₃N₄ binder phase. Kiln furniture, aluminum-melt handling.
• R-SiC (recrystallised) — pure SiC, vapour-phase mass transfer at 2200 °C. Porous but very high T capability. Heating elements (Globar), beam supports in firing tunnels.
• CVD-SiC — chemically vapour-deposited from methylsilane. β-phase, ultra-pure, defect-free. Semiconductor wafer chucks, IR mirrors, X-ray mirrors. Trex Enterprises, CoorsTek CVD-SiC, Rohm & Haas (Dow).
• HPSC — hot-pressed; for ballistic armour tiles.

5m.5 Boron carbide (B₄C)

Rhombohedral, second-hardest commercial ceramic after diamond/cBN. Density 2.50 g/cm³ — the lightest hard ceramic. Hot-pressed for armour and nuclear applications: SAPI/ESAPI plates (Saint-Gobain, Ceradyne, CoorsTek), control rods and neutron-shielding for PWRs (boron-10 absorbs thermal neutrons aggressively).

5m.6 AlN (aluminum nitride)

Wurtzite structure. Combines 170 W/m·K thermal conductivity (third behind diamond and cBN, ahead of BeO, ~Cu/2.5) with full dielectric insulation. Substrate of choice for high-power LEDs, IGBT modules, RF amplifiers, and laser-diode pump-bar mounts. Sintered with Y₂O₃ + CaO + Yb₂O₃ aids to scavenge oxygen (oxygen is the killer of thermal conductivity in AlN). Suppliers: Maruwa, Tokuyama Shapal, CoorsTek.

5m.7 TiB₂

Hexagonal AlB₂ structure. HV 25 GPa, electrically conductive (10⁻⁵ Ω·m). Armour tiles, electrodes (Hall-Héroult aluminum-cell cathodes), evaporation boats for vacuum coating. Hard to densify; hot pressing standard.

5m.8 BN (boron nitride)

Two engineering polymorphs with opposite property packages:

• h-BN (hexagonal “white graphite”) — soft (HV ≈ 1 GPa), lubricant, electrically insulating, k = 600 W/m·K in-plane. Crucibles, releasing agent, dry lubricant, microwave-transparent radomes. Saint-Gobain Combat BN, Momentive Combat.
• c-BN (cubic boron nitride) — diamond-structure, second-hardest material at HV 45 GPa, chemically inert to ferrous metals (where diamond fails by graphitisation + Fe-C). Standard grinding wheel and tool insert for hardened steel. Sumitomo Sumiboron, Element Six CBN200.

5m.9 CMC (continuous fibre ceramic composite)

The non-monolithic branch — SiC/SiC, C/C, C/SiC, oxide/oxide. See [[Engineering/materials-composites]] §5m.8 for full treatment. Engineering reality:

• SiC/SiC — GE/CFM LEAP HPT shrouds (in service since 2016), GE9X turbine, Pratt PW1100G in development. Continuous service at 1300 °C, 30 % lighter than Ni-base superalloy, eliminates cooling-air bleed.
• C/C — F1 brake discs, military aircraft brakes (Boeing 787 wheel brakes are C/C), rocket throats, Shuttle wing leading edges. Oxidises catastrophically above 500 °C in air — SiC coating mandatory for re-usable hot service.
• C/SiC (Cf/SiC) — Porsche PCCB, Ferrari, McLaren carbon-ceramic brake discs. Better oxidation than C/C, higher toughness than SiC/SiC.

5m.10 Glasses

Amorphous SiO₂-based networks. Engineering glasses by composition family:

Family	Composition	T_g (°C)	α (10⁻⁶/K)	Use
Soda-lime-silica	73 SiO₂ + 14 Na₂O + 9 CaO	560	9.0	Windows, bottles
Borosilicate (Pyrex, Schott 8330)	81 SiO₂ + 13 B₂O₃ + 4 Na₂O	525	3.3	Lab glass, cookware
Aluminosilicate (Gorilla Glass)	60 SiO₂ + 17 Al₂O₃ + …	850	8	Cellphone covers (ion-exchanged)
Fused silica (quartz)	> 99.99 % SiO₂	1175	0.55	UV optics, semi masks, laser cavities
ULE titanium-silicate (Corning)	92.5 SiO₂ + 7.5 TiO₂	1100	0.02	EUV mask blanks, telescope mirrors
Chalcogenide (As₂S₃, Ge-Sb-Se)	non-oxide	200–400	25	IR optics (3–14 µm)

Ion-exchange strengthening (chemical tempering — Corning Gorilla, AGC Dragontrail): immerse aluminosilicate in molten KNO₃ at ~400 °C; K⁺ exchanges for Na⁺, putting the surface in compression to 850 MPa over ~50 µm depth. Σ_f rises from ~50 MPa annealed to ~700 MPa post-strengthening.

5m.11 Glass-ceramics

Controlled partial crystallisation of a parent glass — typically 50–95 % crystalline volume in a residual glass matrix. The crystal phase is engineered for a target property:

• LAS (Li₂O-Al₂O₃-SiO₂) — β-spodumene + β-eucryptite phases; near-zero α (Zerodur from Schott at α ≈ 0.02 × 10⁻⁶/K). Telescope mirrors (ESO ELT, Keck, Subaru), photolithography stages, gauge blocks. The CorningWare ovenware family uses the same chemistry without optical-grade processing.
• MAS (MgO-Al₂O₃-SiO₂) — cordierite-based; Macor (Corning) is fluorphlogopite-mica + borosilicate residual glass — machinable with HSS tooling, electrical-insulating, vacuum-compatible. Aerospace, fusion-tokamak insulator parts, medical instruments.
• Bioactive (Bioglass 45S5) — Na₂O-CaO-SiO₂-P₂O₅; bonds to bone in vivo. Maxillofacial repair.

---

6m. Mechanical properties

6m.1 Strength is statistical — restate the rule

A ceramic σ_f quoted without (a) test method, (b) Weibull modulus m, (c) effective volume V is engineering-meaningless. ASTM C1161 four-point bend on standard 3×4×45 mm coupons is the reference; ASTM C1499 biaxial flexure (ring-on-ring) for substrates and discs. ASTM C1239 specifies the 30-specimen Weibull regression. Vendor datasheets that quote a single mean σ should be challenged for m and V.

6m.2 Compressive strength dominates

Ceramic σ_c is 8–15× σ_t because cracks close in compression. Typical:

Material	σ_f flex (MPa)	σ_c (MPa)	Ratio
Al₂O₃ 99.5 %	350	3000	8.6
Y-TZP	1000	2500	2.5
Si₃N₄ HIPSN	900	3500	3.9
α-SiC	450	3900	8.7
B₄C	350	4500	12.9
WC-Co 6 %	2200	5400	2.5
Granite (for comparison)	15	200	13.3

Y-TZP and WC-Co have the lowest compressive ratio because transformation-toughening and Co-binder ductility both reduce the tensile-vs-compressive asymmetry. This is one reason WC-Co is the world’s most-used “ceramic” — it behaves more like a hard metal than a glass.

6m.3 Fracture toughness — the second key property

Per ASTM C1421 (SEPB or SC chevron-notch beam) or ISO 23146. Typical values were tabulated in §4.1. The ranking: Y-TZP > Si₃N₄ HIPSN > ZTA > c-BN ≈ Diamond PCD > Al₂O₃ > SiC ≈ B₄C > glass. Toughened oxides beat non-oxides on K_IC; non-oxides beat oxides on hardness and conductivity. This is the central trade-off.

6m.4 Hardness

Vickers HV per ASTM C1326/C1327 at 9.8 N load (HV₁) — the standard reporting condition. Diamond 80 GPa, c-BN 45, B₄C 30, SiC 28, AlN 12, Al₂O₃ 17, Y-TZP 13, h-BN 1. Cutting and grinding work scales with specific hardness HV/ρ; B₄C wins by lightness (ρ 2.50).

6m.5 Slow crack growth (SCG)

Subcritical crack growth in water, humid air, biological fluid: cracks propagate below K_IC at a rate ν = A · (K_I / K_IC)^n. The SCG exponent n is the design lifetime parameter:

Material	n (water)
Y-TZP	25–50
Si₃N₄ HIPSN	70
Al₂O₃ 99.5 %	30
SiC	100
Glass	16

A material with n = 50 sees lifetime drop ~10× per 5 % stress increase. Glass at n = 16 fails over decades from sub-critical residual stress alone — the reason patio doors break “spontaneously” years after installation. Per ASTM C1576 for ceramic SCG measurement.

6m.6 Fatigue

True cyclic-fatigue degradation (separate from SCG) is real in ceramics, particularly transformation-toughened Y-TZP where reversed loading degrades the toughening shield. ASTM C1361 is the cyclic-fatigue method. n_fatigue is typically 20–50 % lower than n_SCG. In practice ceramic-bearing design lumps fatigue + SCG into a single time-temperature-stress design curve.

---

7m. Thermal / electrical / chemical properties

7m.1 Density

Ceramic density spans 2.1–7 g/cm³ — narrower than metals but wider than polymers:

• B₄C 2.50 (lightest hard ceramic — armour driver)
• AlN 3.26, Si₃N₄ 3.25, Al₂O₃ 3.90, SiC 3.15 (the structural mid-band)
• Y-TZP 6.05 (heaviest engineering ceramic — bone-implant weight matters)
• WC-Co 14.5 (cermet, near tungsten — cutting-tool inertia)

7m.2 Coefficient of thermal expansion

Material	α (10⁻⁶/K, 20–1000 °C)
Fused silica	0.55
Zerodur, ULE	0.02 (near-zero)
Si₃N₄	3.2
SiC	4.3
AlN	4.5
Borosilicate glass	3.3
Al₂O₃	8.0
Y-TZP	10.5
Soda-lime glass	9.0
Steel (reference)	12
Aluminum (reference)	24

Si₃N₄ and SiC match Si wafers (α_Si ≈ 2.6) closely — the reason both are preferred for semiconductor wafer carriers and chucks. Fused silica and Zerodur own precision optics by α ≈ 0. CTE mismatch design is the dominant constraint when bonding ceramic to metal: alumina substrate on copper-leadframe joint sees ~16 µm/m·K differential strain; brazed joints accommodate this via active-metal brazing with thin compliant fillers (TiCuSil, Cusil-ABA) or via a Kovar (α ≈ 5.5) intermediate.

7m.3 Thermal conductivity

Wide range: from h-BN basal at 600 W/m·K (copper territory) down to Y-TZP at 2 W/m·K (Pyrex territory). The metallic-grade conductors — AlN (170), SiC (120), BeO (300, toxic and largely retired) — combine k with full dielectric isolation: this combination does not exist in metals. Power-electronics substrates, plasma-chamber liners, and HPT-shroud CMC all depend on it.

7m.4 Electrical properties

Most ceramics are dielectric (Al₂O₃ ρ_v ~ 10¹⁴ Ω·cm). Exceptions:

• SiC — wide-bandgap (E_g = 3.3 eV) semiconductor; the substrate for Wolfspeed/Cree, Infineon, ST SiC MOSFETs. n-doped (N) and p-doped (Al). See [[Engineering/semiconductor-devices]].
• Doped Si₃N₄, TiB₂, ZrB₂ — conductive; allow EDM machining.
• B₄C — modest semiconductor; armour and neutron absorber.
• Cermets (WC-Co, TiC-Ni) — metallically conductive.

Dielectric strength: Al₂O₃ 96 % gives 14 kV/mm at 0.5 mm substrate thickness — comparable to high-grade ceramic-loaded epoxy laminates, far better than FR-4 (3 kV/mm). AlN, h-BN, fused silica all match or exceed alumina.

7m.5 Chemistry and corrosion

• Oxides are stable in air and oxidising environment to T_max. Attacked by HF (all silicates), strong alkalis (silica + alumina above pH 12), molten alkali metals (Na, K destroy alumina and zirconia).
• Non-oxides (SiC, Si₃N₄, B₄C) form passivating SiO₂/B₂O₃ surface scales in oxidising service — protective up to ~1500 °C for SiC and 1200 °C for Si₃N₄ in air. Attacked above scale-failure T or in reducing atmosphere.
• Diamond graphitises in air above 700 °C and reacts with Fe-group metals during turning of steel (its standard limitation — c-BN exists exactly to fill this gap).
• Glasses are attacked by HF (etches all silicates) and concentrated NaOH at temperature. Borosilicate beats soda-lime substantially in aqueous chemistry.
• Biocompatibility — Al₂O₃, Y-TZP, ZTA, bioactive glass (Bioglass 45S5), hydroxyapatite are FDA / ISO 10993 cleared for long-term implant.

7m.6 Outgassing and UHV

Ceramics are the standard structural material for ultra-high-vacuum systems. Alumina, AlN, MACOR all outgas < 10⁻¹⁰ torr·L/(s·cm²) after a 200 °C bakeout — orders of magnitude below polymer levels. Critical in semiconductor process chambers, particle accelerators, satellite vacuum payloads, and EUV lithography.

---

8m. Processing & joining

8m.1 Powder synthesis

The starting point for nearly every dense ceramic is a sub-micron powder. Routes:

• Bayer process + calcination (alumina from bauxite; the volume route).
• Acheson process (SiC from sand + petroleum coke at 2500 °C; > 100 yr old, still the dominant industrial SiC).
• Direct nitridation (Si + N₂ → Si₃N₄ at 1300 °C; basis of RBSN).
• Carbothermal reduction-nitridation (silica + carbon under N₂; sub-µm Si₃N₄, AlN).
• Sol-gel (hydrolysis + condensation of metal-organic precursors; high-purity, high-cost; for substrates, fibres, coatings).
• Plasma synthesis, laser ablation (nano-powders for specialty applications).

8m.2 Forming

Powder + binder formed to a “green” shape, then debinded and fired:

• Uniaxial pressing — die-pressing of granulated powder. Limited to simple shapes < 50 mm thick and aspect ratio < 3 (die-wall friction creates density gradients).
• Cold isostatic pressing (CIP) — powder in a flexible rubber mould under 200–400 MPa fluid pressure. Uniform density, complex shapes, very large parts. Standard for SiC and Si₃N₄ blanks.
• Slip casting — aqueous powder slurry poured into a plaster mould; capillarity removes water leaving a green-body shell. Complex hollow shapes (crucibles, radomes, sanitaryware on its industrial side).
• Injection moulding (CIM, ceramic injection moulding) — powder + thermoplastic binder injected like plastic; debound then sintered. Complex small parts (cell-phone parts, surgical tools, dental copings). 15–25 % linear shrinkage.
• Extrusion — rods, tubes, honeycombs (cordierite catalytic-converter substrates are extruded in mass volumes).
• Tape casting — slurry doctor-bladed to a thin (0.1–1 mm) sheet on Mylar carrier. Electronic substrates (Al₂O₃, AlN, LTCC tapes). Multilayer co-firing builds buried-trace ceramic packages.
• Gel casting — slurry + monomer + initiator; in-situ polymerisation sets a strong green body that survives demoulding and green machining.
• 3D printing — vat photopolymerisation with ceramic-loaded resin (Lithoz CeraFab, 3D Ceram), binder jetting (ExOne), DIW for dense parts. Quality approaches CIP for small-batch and complex geometries.

8m.3 Sintering — the heart of ceramic processing

Densification by atomic diffusion at T_sinter ≈ 0.7–0.85 × T_m, driven by reduction of surface free energy. Three regimes:

• Solid-state sintering — pure-phase diffusion. Slow. T_sinter > 1700 °C for alumina.
• Liquid-phase sintering (LPS) — small fraction of a liquid grain-boundary phase wets particles, accelerates densification. Si₃N₄ + Y₂O₃-Al₂O₃, SiC + Al-B-C, all engineering oxide ceramics with ≥ 1 % glass phase. Lower T_sinter, faster, fully dense.
• Reactive sintering — densification accompanied by chemical reaction. RBSC (Si + C → SiC), RBSN (Si + N₂ → Si₃N₄). Net-shape — no shrinkage — but residual unreacted phase limits properties.

Pressure-assisted densification:

• Hot pressing (HP) — uniaxial pressure during sintering. Higher density, finer grain — limited to simple geometries.
• Hot isostatic pressing (HIP) — isostatic gas pressure (~200 MPa Ar) on a pre-sintered or canned part. Closes residual pores. The standard final step for aerospace Si₃N₄, ceramic-armour B₄C, biomedical Y-TZP.
• Spark plasma sintering (SPS) / field-assisted sintering (FAST) — pulsed DC current through die + sample. Very fast (minutes), fine grain. Standard for laboratory and small-series nano-grain ceramics; emerging in production.

Sintering shrinkage is 15–20 % linear, 40–50 % volumetric. Tooling and green-stage geometry must be enlarged accordingly. Anisotropic shrinkage (CIP vs uniaxial-pressed) shifts shape — corrected via finite-element shrinkage modelling and CMM iteration.

8m.4 Machining — green vs hardened

• Green machining — pre-firing, the ceramic is chalk-like; standard tungsten-carbide tooling at low speeds cuts threads, slots, contours efficiently. Holds tolerance after firing within ±0.5 % linear (good practice). The dominant final-shape route for medium-complexity parts.
• Hard machining (post-firing) — only diamond grinding (resin-bonded or metal-bonded wheels on rigid surface grinders) or, for conductive ceramics, EDM (wire EDM or sinker EDM). Material removal rates are 10–100× slower than green machining; cost typically $1–10 per cm³ removed depending on geometry and tolerance.
• Laser machining — short-pulse (ns or ps) laser ablation for micro-features and small holes in thin substrates. Thermal damage is the limit.
• Ultrasonic-assisted machining — abrasive slurry between vibrating tool and ceramic; used for non-round holes in alumina and Y-TZP.
• Polishing to mirror finish: diamond paste 1 µm → 0.25 µm → CMP for optics-grade. Subsurface damage from grinding must be polished out before strength testing (a roughly 5× σ_f swing is common between as-ground and as-polished surfaces).

8m.5 Joining

Joining is the historic weakness of ceramic engineering. Routes:

• Active-metal brazing (AMB) — uses Ti or Zr in the braze alloy to react with the ceramic surface and wet it. TiCuSil (Ti-Cu-Ag), Cusil-ABA (Ag-Cu-Ti) for alumina-to-Kovar feedthroughs; ticusil-on-AlN for IGBT direct-bonded substrates. Joint shear strength 100–200 MPa. Per AWS C3.2 brazing handbook.
• Diffusion bonding — surface-finished parts pressed at temperature in vacuum. Slow, geometry-limited; used in MEMS and laser-cavity manufacturing.
• Reaction bonding (Si infiltration of C-fibre + SiC preforms) — for CMC C/SiC.
• Glass-frit joining — low-melting glass bonds two oxide parts; common in electronic packaging.
• Adhesive bonding — epoxy or cyanoacrylate for non-load-bearing joints (encoder disks, sensor mounts).
• Mechanical joining — bolted, clamped, or shrink-fit. Always with compliant interlayer (graphite, PTFE) to avoid Hertzian-edge cracking.

Metallisation for solder + braze:

• Moly-Mn (W-process) — refractory-metal paint fired at 1500 °C; standard for high-reliability hermetic feedthroughs (Coors AD-995 + Moly-Mn).
• Thick-film silver / silver-platinum — fired conductors for LTCC and AlN substrates.
• DBC (direct-bonded copper) — eutectic bond of Cu foil to Al₂O₃ or AlN at 1065 °C; substrate of choice for power IGBT.
• AMB substrate — Cu brazed to ceramic by AgCuTi at 800 °C; tougher, higher-cycle than DBC.
• Sputtered Ti/Cu/Au or Cr/Cu/Au — thin-film metallisation for microelectronic packages.

8m.6 Inspection

• Visual + dye penetrant — surface flaws.
• Ultrasonic immersion — sub-surface porosity, delamination, large flaws > 100 µm. Per ASTM E2375 for ceramics.
• X-ray and microfocus CT — porosity, density variation, fibre orientation in CMC.
• Acoustic emission (AE) — real-time crack-initiation detection during proof-load.
• Proof testing — load every part to a fraction (typically 1.5×) of its in-service stress; the survivors form a screened lower-strength population. Standard for bearing balls, pressure-vessel windows, biomedical implants.

---

9m. Applications & selection trade-offs

9m.1 Quick-pick guide

Need	First pick	Reason
Cutting tool for steel (HRC < 55)	WC-Co cermet	Toughness, cost, supply
Cutting tool for hardened steel (HRC > 60)	c-BN insert (Sumiboron, Sandvik CBN7050)	No graphitisation in Fe
Cutting tool for cast iron	Si₃N₄/SiAlON (Kennametal KY3500)	Thermal-shock resistance
Cutting tool for nickel superalloys	SiAlON or whisker-toughened Al₂O₃	Hot hardness
Cutting non-ferrous (Al, brass, composites)	PCD or PCD-tipped carbide	Wear life
Hybrid bearing balls (high-speed spindle)	HIP-Si₃N₄ (SKF Class C, NSK Robust)	RCF life, low density
Pump seal face, water service	Al₂O₃ 96 % rotating + carbon stationary	Cost
Pump seal face, slurry / hot oil	SiC rotating + SiC stationary	Thermal shock, abrasion
Slurry pump throat bushing	α-SiC (Hexoloy)	Abrasion, chemistry
Hip femoral head	ZTA Biolox-delta	Toughness + size availability
Knee tibial component	Y-TZP or oxidised Zr alloy	Wear couple to UHMWPE
Dental crown / bridge	Y-TZP (Lava, BruxZir, e.max ZirCAD)	Aesthetic + strength
Power IGBT substrate	AlN AMB or Si₃N₄ AMB	k + dielectric
Power LED substrate	AlN DBC	k + dielectric
Aerospace HPT shroud	SiC/SiC CMC	T capability beyond superalloy
Spacecraft optical bench	Zerodur or CFRP/cyanate	Near-zero α
Telescope primary mirror	Zerodur, ULE, or SiC	Stiffness/density + α
Body armour SAPI plate	B₄C (lightest) or SiC (multi-hit)	Density vs cost
Nuclear neutron absorber	B₄C	¹⁰B absorption cross-section
Refractory crucible for melting steel	MgO or fused silica	Reactivity to Fe
Refractory for aluminum melting	NSiC, fused silica	Al-attack resistance
Plasma-chamber liner	α-SiC or Al₂O₃ 99.9 %	Plasma erosion, purity

9m.2 Trade-offs

• Alumina vs zirconia. Alumina is harder, stiffer, cheaper, and available in larger parts; zirconia is tougher and stronger in bending but heavier (6.05 vs 3.9 g/cm³), more expensive, and vulnerable to LTD in hot moist environment. ZTA composites are the modern compromise — alumina stiffness with zirconia toughness — and dominate hip-implant ball-heads in 2026.

• SiC vs Si₃N₄. SiC is harder, has higher k, better chemical / oxidation resistance to 1500 °C, and lower α. Si₃N₄ is tougher (K_IC 7 vs 3 MPa√m), has better thermal-shock resistance, and dominates as bearing material because of contact-fatigue performance. Pump seals split the difference (SiC for chemistry and high-T, Si₃N₄ for shock-loaded shaft components).

• Monolithic vs CMC. Monolithic ceramics are cheaper, stiffer, and harder, but brittle and statistical. CMC offers quasi-ductile failure (matrix-crack arrest by fibre pull-out), service to 1300 °C, and engineering-acceptable damage tolerance — at 5–20× the cost and manufacturing complexity. The trade-off is purely economic: where Ni-base superalloy fails on temperature (LEAP HPT shroud), CMC wins; everywhere else monolithic ceramic or metal still wins.

• Ceramic vs WC-Co cermet. WC-Co is tougher (K_IC 8–25 MPa√m), shock-resistant, and electrically conductive (EDM-able). Pure ceramic is harder, lighter, and chemically more inert. Cutting-tool selection breaks along this line: WC-Co for impact-loaded interrupted cuts, ceramic insert for finish + high-speed work.

• Glass vs glass-ceramic. Glass is cheap, transparent, formable. Glass-ceramic adds dimensional stability (near-zero α — Zerodur), strength (50–150 MPa vs 30 MPa annealed), and higher T capability (600–800 °C vs 400 °C). Optical-flat applications routinely justify the 5–20× cost premium of Zerodur over ground borosilicate.

9m.3 Notable applications

• CFM LEAP HPT shrouds (SiC/SiC CMC) — in service on A320neo (LEAP-1A, 2016 EIS), 737 MAX (LEAP-1B), C919. ~30 % mass reduction, eliminates ~50 % of HPT cooling-air bleed → SFC improvement.
• SKF Class C hybrid bearings (Si₃N₄ balls + 52100 races) — standard for high-speed machine-tool spindles, electric-traction motors (Tesla rear-drive unit uses Si₃N₄ hybrid for HF-VFD electrical-isolation), industrial gas turbines.
• CeramTec Biolox-delta ZTA femoral head — > 5 million implanted globally as of 2024.
• Schott Zerodur — primary mirror substrates for ESO ELT (798 segments of 1.4 m hexagonal Zerodur), Keck, GTC, LSST, Subaru.
• SiC mirrors — JWST secondary, ESA Herschel 3.5 m primary (one piece — the largest sintered-SiC structure ever made), ESO ELT M5 tip-tilt.
• Wolfspeed SiC MOSFET wafers — 200 mm 4H-SiC substrates for EV traction inverters (Tesla Model 3 + Y, Lucid Air, GM Ultium).
• Saint-Gobain Hexoloy SA — the global benchmark sintered α-SiC: pump seals, slurry-handling, semiconductor wafer carriers.
• Coors AD-995 alumina — the reference high-purity engineering alumina; substrate base in spark-plug insulators (Champion, NGK), microwave windows, IR-window blanks.

---

10m. Failure modes

1. Brittle fracture from intrinsic flaw. The dominant failure mode. Cracks initiate at the largest defect — surface micro-crack from grinding, sub-surface pore, inclusion, grain-boundary phase pocket — and propagate at sonic speed once K_I = K_IC. No warning, no plastic deformation, no leak-before-break. The design response is statistical (Weibull) allowables + proof testing.

2. Statistical strength variability. Same powder lot, same firing run can produce parts varying 2–3× in strength. Weibull design (ASTM C1239) at A-basis (99 % survival, 95 % confidence) or B-basis (90 % / 95 %) is mandatory for primary structure. Material vendors that publish only mean σ_f without m are incomplete; demand the full Weibull pair.

3. Slow (subcritical) crack growth — SCG. Water + stress drives crack tips forward below K_IC. Lifetime ∝ σ⁻ⁿ; for n = 30 (alumina), a 10 % stress reduction multiplies lifetime by ~26. Static-fatigue limit σ_0 ≈ 0.4–0.5 × σ_f short-term in moist environment. The reason ceramic-bearing design uses generous safety factors and the reason proof testing is mandatory — proofed parts have a bounded initial flaw and predictable SCG lifetime.

4. Cyclic fatigue. Real in transformation-toughened Y-TZP (the toughening shield degrades under reversed loading) and in CMC (matrix microcracking accumulates per cycle). Per ASTM C1361. Generally less aggressive than metal-fatigue per cycle, but with no plastic-deformation warning.

5. Thermal shock. Sudden ΔT exceeds R-parameter capability. The bandwidth between safe and catastrophic is narrow — ΔT_safe is typically the published R minus a 20 % margin. Pump seals are the canonical victim (loss of coolant → process-fluid contact → crack within seconds). Mitigation: select high-R (Si₃N₄, SiC, fused silica), thermally isolate, limit T-gradients with stand-off design.

6. Low-temperature degradation (LTD) of Y-TZP. Tetragonal-to-monoclinic transformation accelerated by moisture at 130–300 °C. Surface roughens, micro-cracks form, strength drops 30–50 %. The 2001 Prozyr (Saint-Gobain Desmarquest) hip-head recall removed Y-TZP from femoral-head use; current implants use ZTA Biolox-delta which is LTD-immune. ISO 13356 specifies the hydrothermal-aging test (steam autoclave at 134 °C / 5 hours equivalent to 15–20 years in vivo).

7. Oxidation of non-oxides. Above T_scale-fail (SiC ~1550 °C, Si₃N₄ ~1450 °C, B₄C ~600 °C in air, AlN ~1000 °C), passivating oxide scales fail and bulk oxidation proceeds. Pesting is the catastrophic variant — accelerated oxidation in narrow T-windows for some carbides and silicides. Mitigation: SiC coating on C/C and B₄C for hot service; protective atmospheres.

8. Wear. Abrasive and tribochemical. Typically very low for hard ceramics — that is the whole point — but micro-fracture wear (loss of material as small chips, not plastic flow) dominates at sliding speeds above ~10 m/s or under particulate contamination. Grinding wheels exploit this; bearings avoid it via clean lubrication.

9. Chemical attack. HF dissolves silicates (fused silica, soda-lime, borosilicate, MACOR). Concentrated NaOH at T > 80 °C attacks alumina and silicates. Molten alkali metals (Na, K) destroy oxides. Hot reducing atmospheres (H₂ + H₂O at > 1500 °C) volatilise SiC scale. Choose chemistry: oxide for oxidising service, non-oxide for reducing; replace ceramic family rather than try to coat through.

10. Surface chipping at edges and corners. Sharp ceramic edges are intrinsically weak. Design rule: break every edge to a chamfer ≥ 0.3 mm or a radius ≥ 0.5 mm. Bolt-hole counterbores require radii. Mechanical-handling damage is the second-largest source of field failures after manufacturing-flaw fracture.

11. Joining failure. Active-metal brazed joints fail at the ceramic-near-interface from CTE-mismatch residual stress, not in the braze itself. Mitigation: thin compliant Cu or Ni interlayers, geometry that loads the joint in shear or compression (never peel), and pre-bake to relieve assembly stress.

12. CMC damage. Matrix microcracking + fibre pull-out is the intended damage mechanism (quasi-ductile failure) but it opens diffusion paths for oxygen to attack interior fibres; long-term oxidation of fibre-matrix interfaces under load is the real CMC service-life limit. SiC/SiC operates indefinitely at 1200 °C in air; service above 1400 °C is presently limited to short-duration mission profiles.

13. Glass-ceramic devitrification. Uncontrolled crystallisation in the residual glass phase reduces optical transparency, raises α, and induces internal stress. Causes: prolonged thermal-cycling above T_g, contamination, off-spec crystallisation hold. Critical for telescope-mirror service in temperature-cycled spacecraft.

14. Diamond and c-BN grit micro-fracture. Both fail by abrasive micro-fracture in grinding service — this is in fact the wear-out mode of grinding wheels and is exploited by design (self-sharpening as fresh edges expose). c-BN tools fracture catastrophically in interrupted cutting; reserve for continuous cuts.

15. Manufacturing defect spectrum. Inclusions (refractory contamination from mill media or furnace lining), large pores (incomplete densification, trapped gas), abnormal grain growth (overshoot of sintering schedule), residual binder pockets in CIM parts, machining-induced subsurface damage (chip, scratch, micro-crack). Inspection (immersion-UT for sub-surface flaws, dye penetrant + microscopy for surface flaws, proof testing for screening) is mandatory for high-reliability applications.

---

11. Cross-references

• [[Engineering/materials-steel]] — sibling structural material; cermets (WC-Co) and ceramic-vs-metal cutting-tool selection
• [[Engineering/materials-aluminum]] — alumina + AlN ceramic substrates bonded to Al power electronics packages; aluminum-matrix composites with SiC
• [[Engineering/materials-polymers]] — ceramic-filled polymer composites (FR-4 silica, PCB substrates, glass-bead-filled engineering plastics)
• [[Engineering/materials-composites]] — CMC (SiC/SiC, C/C, C/SiC, oxide-oxide) extended treatment
• [[Engineering/materials-selection]] — Ashby method placing ceramics on E/ρ and σ/ρ charts; the high-stiffness corner
• [[Engineering/mechanics-of-materials]] — Hertzian contact derivation, Griffith fracture criterion, K_IC plane-strain definition
• [[Engineering/bearings]] — Si₃N₄ hybrid + full-ceramic bearings, ceramic-on-ceramic mechanical seals
• [[Engineering/machining]] — diamond grinding of fired ceramics, green machining, EDM of conductive ceramics
• [[Engineering/heat-transfer]] — ceramic refractory insulation, thermal-shock-resistant design, R-parameter
• [[Engineering/semiconductor-devices]] — SiC + GaN wide-bandgap power devices, AlN substrates
• [[Engineering/joining-welding]] — active-metal brazing of ceramic-to-metal joints
• [[Robotics/end-effectors]] — ceramic cutting tools, pump pistons, wear-resistant grippers
• [[Robotics/compliant-mechanisms]] — Zerodur and SiC structural elements for optical positioning stages
• [[Languages/Tier3/construction-bim]] — STEP AP242 material assignment for ceramic + CMC parts

---

12. Citations

1. Callister, W. D. & Rethwisch, D. G. Materials Science and Engineering: An Introduction, 10th ed. (Wiley, 2018). General-purpose ceramics chapter.
2. Barsoum, M. W. Fundamentals of Ceramics, 2nd ed. (CRC Press / IOP, 2019). The graduate-level standard.
3. Richerson, D. W. Modern Ceramic Engineering: Properties, Processing, and Use in Design, 4th ed. (CRC Press, 2018). The practitioner’s reference.
4. Carter, C. B. & Norton, M. G. Ceramic Materials: Science and Engineering, 2nd ed. (Springer, 2013).
5. Wachtman, J. B., Cannon, W. R., & Matthewson, M. J. Mechanical Properties of Ceramics, 2nd ed. (Wiley, 2009). Definitive treatment of Weibull, SCG, fracture.
6. Ashby, M. F. Materials Selection in Mechanical Design, 5th ed. (Butterworth-Heinemann, 2017). Ceramics on the Ashby charts.
7. ASM Handbook, Vol. 21 — Composites and Engineered Materials Handbook, Vol. 4 — Ceramics and Glasses (ASM International).
8. ASTM C1161-18 — Flexural Strength of Advanced Ceramics at Ambient Temperature.
9. ASTM C1421-18 — Determination of Fracture Toughness of Advanced Ceramics at Ambient Temperature.
10. ASTM C1239-13(2018) — Reporting Uniaxial Strength Data and Estimating Weibull Distribution Parameters for Advanced Ceramics.
11. ASTM C1326-13(2018) — Knoop Indentation Hardness of Advanced Ceramics; ASTM C1327-15 — Vickers.
12. ASTM C1499-19 — Monotonic Equibiaxial Flexural Strength of Advanced Ceramics at Ambient Temperature (ring-on-ring).
13. ASTM C1576-19 — Determination of Slow Crack Growth Parameters of Advanced Ceramics by Constant Stress-Rate Flexural Testing.
14. ASTM C1361-10(2020) — Cyclic Fatigue of Advanced Ceramics.
15. ISO 6872:2024 — Dentistry — Ceramic materials. Y-TZP and lithium-disilicate flexural strength + Weibull requirements.
16. ISO 13356:2015 — Implants for surgery — Ceramic materials based on yttria-stabilised tetragonal zirconia (Y-TZP). Includes hydrothermal-aging test.
17. ASTM F1873-21 — Standard Specification for High-Purity Dense Yttria-Tetragonal Zirconium Oxide Polycrystalline Ceramic for Surgical Implant Applications.
18. CoorsTek datasheets — AD-995 alumina, Pure-SiC, hot-pressed Si₃N₄ (current revisions).
19. CeramTec datasheets — Biolox-forte (Al₂O₃), Biolox-delta (ZTA), Rocar S (SiC), SL200 (Si₃N₄).
20. Saint-Gobain Performance Ceramics — Hexoloy SA / SE / SP (SiC), Norbide (B₄C), Combat (h-BN) product literature.
21. Kyocera Fine Ceramics — A-479B alumina, SN-220 Si₃N₄ datasheets.
22. SKF — Hybrid Bearing Technology (general technical brochure; Si₃N₄ ball specifications).
23. GE Aviation / CFM — LEAP Engine CMC Component briefings (open-publication press materials 2015–2020).
24. Schott AG — Zerodur Product Information; Corning — ULE Titania-Silicate Glass datasheet.
25. NIST Structural Ceramics Database (former Ceramics WebBook); NIST Property Data Summaries for Advanced Materials.