Ceramics and glasses

In brief

A stone column can stand for two thousand years carrying the full compressive weight of the structure above it. The same stone, cut into a thin slab and loaded in bending, snaps almost without warning at a fraction of that load. Both facts describe the same material. Ceramics and glasses are not simply weak or fragile; they are asymmetric, extremely tolerant of compression and poorly tolerant of tension, and once you understand why, choosing when to use them stops being guesswork about "how brittle" something feels and becomes a question of what kind of stress the part will actually see.

The full treatment

First look: the column that stands and the slab that snaps

Roman aqueducts and Egyptian pyramids are built almost entirely from stone loaded in compression, and many still stand. Try to support the same weight on a thin stone beam spanning a gap, loading it in bending rather than pure compression, and it fails at a small fraction of the load the column carried. The stone has not changed. The way it is being loaded has.

Building the idea: bonds with no slip route

Ceramics are held together by ionic bonds, electrostatic attraction between oppositely charged ions, as in many oxide ceramics, by covalent bonds, shared electron pairs pointing in fixed directions between specific atoms, as in silica-based glasses, or by a mixture of both. Both bond types are strong, which is why ceramics resist stretching and compressing so stiffly. But both are also directional in a way metallic bonding is not. Sliding one plane of ions past another would force like-charged ions into direct contact, which the electrostatic repulsion between them strongly resists; sliding one plane of covalently bonded atoms past another requires literally breaking rigid, angled bonds rather than gliding through a nondirectional electron sea. There is no equivalent, at ordinary temperatures, to the dislocation glide that lets a metal's atomic planes slide past one another and absorb strain. With no low-energy way to rearrange under load, a ceramic has nowhere to send the strain energy building up at any imperfection in its structure except into breaking bonds outright, a crack.

Building the idea: strength is set by the flaw, not the bond

In the 1920s, the aeronautical engineer A. A. Griffith noticed something that did not fit a picture of strength as a simple property of the material: glass fibers were dramatically stronger than glass rods of the same composition, and the strength of a given sample depended on its size and surface condition, not only on what it was made of. His explanation was that stress concentrates sharply at the tip of any pre-existing flaw, a microcrack, a surface scratch, a pore, and the material fails when that locally concentrated stress, not the average stress across the whole part, reaches the bond-breaking threshold. Griffith showed that the failure stress a brittle sample can support is proportional to the square root of a quantity built from the material's stiffness and surface energy, divided by the length of the worst flaw present, in words, doubling the flaw length lowers the strength by roughly a factor of the square root of two, not by half. A small glass fiber is strong because it is statistically unlikely to contain a large flaw; a large ceramic plate made of the identical material is weaker because, across its much larger surface and volume, a large flaw becomes far more likely to be present somewhere in it.

The tension-compression asymmetry

This flaw mechanism explains the asymmetry directly. Under tension, a crack's two faces are pulled apart, the stress concentration at its tip is fully active, and the crack can grow catastrophically at a low overall stress. Under compression, the crack's faces are pushed together instead, friction across the closed crack resists any sliding, and the stress concentration that drives crack growth in tension is largely absent. The same flaw that would be lethal in tension is nearly harmless in compression. This is precisely why ancient masonry structures, arches and domes, deliberately shaped so that the material is everywhere in compression, could be built reliably from a material that would fail immediately if the same stones were loaded in bending, and it is why modern engineering keeps ceramics in compression-dominated roles, engine components under compressive preload, cutting tool inserts, brake friction materials, refractory furnace brick, and reaches for reinforcement, prestressing, composite fibers, wherever a ceramic must also carry tension.

Glasses as ceramics without long-range order

A glass is built from the same kind of ionic or covalent bonding as a crystalline ceramic, but its atoms are arranged without the long-range repeating order of a crystal, an amorphous network rather than a lattice. The same logic applies without modification: no low-energy slip mechanism, strength set by the worst flaw present rather than by bond strength, and strong tolerance for compression paired with poor tolerance for tension. Glass is, if anything, more flaw-sensitive than many crystalline ceramics, because ordinary handling leaves microscopic surface flaws that dominate its strength; tempered glass exploits the compression tolerance directly, by rapidly cooling the surface to leave it in a state of permanent residual compression, so that an applied tensile load must first cancel that residual compression before any crack at the surface can begin to open.

Lineage

Fired clay pottery is among the oldest human technologies, tens of thousands of years old, and glassmaking was practiced in ancient Egypt, Mesopotamia, and Phoenicia long before anyone had a theory of bonds or crystal structure. Roman engineers refined compression-dominated masonry, arches, vaults, and domes, into structures still standing today, entirely through empirical trial rather than any theory of fracture. The first quantitative theory of brittle fracture came from Griffith's 1920s work on why glass fibers were so much stronger than glass rods, motivated by his search for lighter, stronger materials for early aircraft; his flaw-based criterion was extended in the mid-twentieth century, notably by George Irwin, into the modern field of fracture mechanics used across ceramics, glasses, and metals alike.

The strongest case for it

The bonding-to-flaw-sensitivity chain correctly predicts a wide range of observed behavior: it explains why thin fibers of a ceramic are reliably stronger than thick sections of the identical material, which underlies the strength of high-performance glass and ceramic fibers used in composites; it explains why tempering and prestressing, putting the surface into residual compression, reliably raise the effective strength of glass and concrete; and it explains directly why millennia-old masonry compression structures endure while similarly-aged tensile stone members do not survive.

The strongest case against it

Griffith's theory in its cleanest form assumes an ideally brittle material with a sharp crack and linear elastic behavior around it, and several advanced ceramics beat that simple prediction on purpose: transformation toughening in zirconia ceramics and fiber bridging in ceramic-matrix composites both raise fracture toughness well above what the basic flaw model alone would predict, by adding a mechanism that absorbs energy at the crack tip instead of letting it propagate freely. At high enough temperature, ceramics can show real plasticity and creep, slow deformation given time and heat, which the room-temperature "brittle always" picture in this entry does not capture. A common misconception is treating "ceramic" as a synonym for fragile in every sense, which overlooks the very real compressive and thermal strength that makes ceramics indispensable in furnace linings, cutting tools, and structural masonry. A second, more consequential misconception is assuming that a strength value measured on a small laboratory sample scales directly to a full-size part; because failure is governed by the worst flaw present, and larger volumes are statistically more likely to contain a larger flaw, the same ceramic is reliably weaker as a large plate than as a small bearing ball, a fact captured formally by Weibull statistics rather than by a single strength number.

Where it stands now

The chain from bonding, to the absence of low-energy slip, to flaw-governed strength, to the tension-compression asymmetry, is broad consensus and has not been seriously challenged since Griffith's theory was confirmed and extended in the twentieth century. Current engineering effort continues in toughening mechanisms for advanced ceramics and in the statistical treatment of reliability for ceramic components, refinements to the framework rather than disputes with its core mechanism.

Test yourself

You are designing two parts for a furnace at twelve hundred degrees Celsius: a load-bearing interior pillar, and a thin, flexible gasket sealing the furnace door. Decide, using the ideas in this entry, whether a ceramic is an appropriate choice for each part as initially described, and explain your reasoning in terms of the stress each part actually experiences. If you conclude that a ceramic is used in the gasket location, describe a redesign, either a change of geometry or the addition of reinforcement, that would keep the ceramic in compression rather than tension in service, and explain specifically why that redesign works using the crack-opening mechanism described above.