# Composite material **Composite material** is a material made from two or more constituents of markedly different properties which remain separate and identifiable inside the finished part, and whose combination has properties that neither constituent has alone. The usual arrangement is a stiff, strong, brittle reinforcement — fibres, whiskers or particles — held in a softer, tougher matrix that transfers load into it, protects it and holds the geometry. The matrix is weak and the fibre cannot be used by itself; together they give the highest [[Specific_modulus|specific stiffness]] and [[Specific_strength|specific strength]] of any structural material in production. In the microsim below the reader raises the fibre volume fraction *V*f from 0 to about 0.7 and watches a lamina fill with parallel fibres while two curves separate. Along the fibres the constituents strain together and the modulus follows the Voigt rule of mixtures, `E_c = V_f·E_f + (1 − V_f)·E_m`, a straight line from matrix to fibre. Across them the constituents carry the same stress instead, and the modulus follows the Reuss form, `1/E_c = V_f/E_f + (1 − V_f)/E_m`, a curve that stays near the matrix value until *V*f is almost 1. Presets cover glass/epoxy, carbon/epoxy, [[Kevlar|Kevlar]]/epoxy and tungsten carbide–[[Cobalt|cobalt]], and the density, which really is linear in *V*f, is plotted on the same modulus-against-density axes the [[Material_selection|material selection]] chart uses. On the [[Materials_science|Materials science]] flagship this article serves the *Composites* section of Part VIII; [[Rule_of_mixtures|rule of mixtures]], [[Composite_laminate|composite laminate]] and [[Fibre-reinforced_plastic|fibre-reinforced plastic]] reuse the same lamina with their own controls. ## History Composites are older than metallurgy. Straw-reinforced mud brick, plywood in Egyptian joinery, laminated horn-and-sinew bows and papier-mâché are all fibre-in-matrix constructions exploited long before anyone could say why they worked. Reinforced concrete, patented in the 1860s and 1870s, is the first industrial composite designed on a calculation: concrete is strong in compression and weak in tension, steel is the reverse, and each is placed where it is needed. The modern synthetic era begins with [[Fiberglass|glass fibre]] in a thermosetting resin around 1935, which gave boat hulls, radomes and body panels a material that was strong, light, mouldable and electrically transparent. Carbon fibre of usable stiffness was developed in Britain in the 1960s; aramid followed at DuPont, reaching the market as [[Kevlar|Kevlar]] in the early 1970s. The aerospace consequence took three decades to mature: from secondary fairings to the empennage to airframes more than half composite by weight. The parallel story in cutting tools began earlier, with cemented tungsten carbide in the 1920s. ## Examples The term covers a range wide enough to be unhelpful without qualification, so it is narrowed by matrix — polymer, metal, ceramic or carbon — and by reinforcement geometry: particles, short fibres, continuous fibres or woven fabric. ### Composite materials The polymer-matrix families dominate by volume. Glass fibre in polyester or epoxy is the cheap workhorse; carbon fibre in epoxy the high-performance structural material; aramid in epoxy the choice where impact and ballistic resistance matter more than compressive strength. Metal-matrix composites put silicon carbide particles or fibres into aluminium for stiffness and wear resistance. Ceramic-matrix composites put carbon or silicon carbide fibres into a ceramic, not to stiffen it but to stop a crack running through it. Cemented carbides put hard tungsten carbide grains into a tough cobalt binder. Natural composites — [[Bone|bone]], wood and [[Nacre|nacre]] — use the same principles with far better interfaces. ### Products The catalogue follows the property being bought. Where stiffness per unit weight is the objective: aircraft structures, [[Wind_turbine|wind turbine]] blades, racing car monocoques, bicycle frames, prosthetic limbs. Where corrosion resistance is: chemical tanks, pipes, boat hulls. Where electrical transparency is: radomes and antenna housings. Where hardness with toughness is: drill bits and milling inserts. And where a tailored expansion coefficient is: satellite optical benches, laid up so the structure barely changes length as it passes in and out of sunlight. ## Overview A composite works because load is shared according to stiffness, and the sharing rule depends on how the phases are connected. If they are in parallel, as fibres are along their own axis, both phases suffer the same strain, the stiffer one takes proportionally more stress, and the modulus is the volume-weighted average — the Voigt bound.[^voigt1889] If they are in series, as they effectively are across the fibres, both carry the same stress, the softer one supplies most of the strain, and the modulus is the harmonic average — the Reuss bound.[^reuss1929] No two-phase material can be stiffer than the first or softer than the second at a given volume fraction, so every real composite lies between them, and design is the business of getting close to the upper one.[^johnson-ch8] The consequence that distinguishes composites from metals is anisotropy. With *E*f = 230 GPa for a standard-modulus carbon fibre, *E*m = 3.5 GPa for epoxy and *V*f = 0.6, the two rules give 139.4 GPa along the fibres and 8.6 GPa across them, a ratio of 16 (derived). A metal has one modulus; a lamina has at least four independent elastic constants, and a part made of laminae has whatever the layup gives it. That is an opportunity and a liability in equal measure: strength can be put exactly where the load is, and is absent everywhere else. ## Cores in composites Because a panel's bending stiffness rises as the cube of its thickness, the cheapest stiffness in any structure is separation. A sandwich panel bonds two thin, stiff composite faces to a thick, light core — polymer [[Foam|foam]], balsa, or aluminium or aramid honeycomb — so the faces carry the bending stresses as a couple while the core carries the shear and holds them apart. The panel weighs a fraction of a solid laminate of the same stiffness, which is why aircraft floors, boat decks, train interiors and wind turbine blade shells are almost all sandwiches. The failure modes are the core's, not the face's: core shear, face wrinkling as a local [[Buckling|buckling]] of the skin on an elastic foundation, indentation under a point load, and skin-to-core disbonding. Water ingress into a honeycomb through an unsealed edge is the classic in-service problem, adding weight and freezing to split the cells. ### Semi-crystalline polymers A semicrystalline thermoplastic is itself a composite at a smaller scale: stiff crystalline lamellae dispersed in an amorphous phase whose modulus depends on whether it is above or below its [[Glass_transition|glass transition]]. The same rule-of-mixtures reasoning applies with crystallinity in place of *V*f, which is why the stiffness of [[Polyethylene|polyethylene]] varies severalfold by crystallinity alone. As matrices, semicrystalline polymers such as PEEK offer toughness and a recyclable, weldable alternative to thermosets, at much higher processing temperatures. ## Methods of fabrication Fabrication has to place fibres in the right direction at a high volume fraction, wet them completely with resin, cure the resin under pressure to suppress voids, and hold the geometry while it happens. Every process is a different compromise between those four. ### Overview of mould Open moulding — hand lay-up and spray-up onto a single gel-coated tool — is cheap, needs no pressure vessel and gives *V*f near 0.3 with substantial void content. Closed moulding raises quality: vacuum bagging draws off volatiles and consolidates at up to one atmosphere; autoclave curing adds several more and gives aerospace laminates with *V*f near 0.6 and porosity below 1 %; resin transfer moulding injects resin into a dry preform in a matched metal tool. Pressure buys fibre fraction and fewer voids, and the microsim's two curves show what that fraction is worth. ### Other fabrication methods Continuous and automated processes take over where the geometry allows. Pultrusion pulls fibre through a resin bath and a heated die to make constant-section profiles. Filament winding lays tow onto a rotating mandrel at a controlled angle, the natural process for pressure vessels and drive shafts. Automated tape laying and fibre placement put prepreg down by robot, which is how large airframes are built. Compression moulding of sheet or bulk moulding compound serves automotive volumes. Additive routes now print short- and continuous-fibre thermoplastic, borrowing the layer-by-layer discipline of [[3D_printing|3D printing]] — and its overhang and feature-size rules — while inheriting its anisotropy between layers.[^barnes-dfam] ### Tooling The tool is a large part of the cost and the risk. It must hold tolerance at cure temperature, survive many cycles, release the part, and — the constraint peculiar to composites — expand compatibly with a laminate whose expansion is near zero along the fibres. A steel tool expands far more than a carbon laminate, so a part cured on it comes off distorted unless the tool is compensated; Invar and carbon-composite tooling avoid that at higher cost. ## Physical properties Density follows the rule of mixtures exactly, because mass is additive: `ρ_c = V_f·ρ_f + (1 − V_f)·ρ_m`. For carbon/epoxy at *V*f = 0.6, with ρf = 1.8 and ρm = 1.2 Mg/m³, that is 1.56 Mg/m³ (derived), and combining it with the 139.4 GPa longitudinal modulus gives a specific modulus of 89 GPa per Mg/m³ against about 27 for [[Steel|steel]] — the single number that explains the aerospace industry's migration. Thermal and electrical properties are anisotropic for the same reason the moduli are. Carbon fibre conducts heat and electricity well along its axis and poorly across it, so a laminate is a fair conductor in-plane and an insulator through the thickness — which is why lightning protection on a composite airframe needs an embedded metal mesh. Expansion along the fibres is near zero or slightly negative for carbon and aramid, while across them it is the matrix's, an order of magnitude larger, and the mismatch puts residual stress into every laminate as it cools from cure. Polymer matrices also absorb moisture, which plasticises them, lowers the glass transition and swells the laminate. ## Mechanical properties of composites Stiffness is predictable from the constituents; strength is not. Modulus is an average and obeys bounds; strength is set by the weakest link in a population of flaws, by the fibre–matrix interface, and by whichever failure mode is reached first. Hence design allowables established by test rather than calculation, and a literature that treats stiffness and failure in separate chapters.[^johnson-ch8][^johnson-ch9] ### Particle reinforcement Particles stiffen and harden without the directionality of fibres, giving a nearly isotropic composite whose modulus lies between the Voigt and Reuss bounds rather than on either; for a statistically isotropic two-phase material the Hashin–Shtrikman variational bounds narrow the gap considerably.[^hashin-shtrikman] Cemented carbide is the important case: with *E* ≈ 700 GPa for tungsten carbide, 210 GPa for cobalt and 90 % carbide by volume, the elementary bounds are 651 and 568 GPa (derived), and commercial grades fall between them. The cobalt binder supplies the toughness that lets a brittle carbide tool survive an interrupted cut. ### Short fiber reinforcement (shear lag theory) A short fibre carries no load at its ends, because stress must be fed into it through shear at the interface. The shear-lag argument gives a critical length `l_c = σ_f·d/(2·τ_y)`, with σf the fibre strength, *d* its diameter and τy the interfacial shear strength: a fibre shorter than *l*c pulls out before it breaks, and the composite never reaches the fibre's strength.[^kelly-tyson] For a carbon fibre of 7 µm diameter with σf = 3,500 MPa in a matrix of τy = 40 MPa, `l_c` is about 0.3 mm, an aspect ratio near 44 (derived, ILLUSTRATIVE interface strength). Injection-moulded short-fibre compounds break fibres during processing and end below that length, which is why they reach perhaps a third of the strength of their continuous-fibre equivalents.[^jensen-plastics] ### Continuous fiber reinforcement Continuous fibre is the case the microsim draws. Along the fibres the parallel assumption is nearly exact, and the Voigt line is confirmed by test to within a few per cent; the fibres carry a fraction `V_f·E_f/(V_f·E_f + (1 − V_f)·E_m)` of the load, which for the carbon/epoxy preset at *V*f = 0.6 is 99 % (derived). Across the fibres the Reuss form is only a lower bound — real transverse moduli sit somewhat above it, because the matrix between fibres is constrained by [[Poisson's_ratio|Poisson]] effects — so the microsim marks that curve as a bound and not a prediction. The upper limit on *V*f is geometric. Parallel circular fibres in a hexagonal array pack at most to π/(2·√3) = 0.907, and in a square array to π/4 = 0.785 (derived); real laminates stop near 0.6 to 0.65, because above that the resin can no longer wet every fibre and voids appear faster than stiffness. That is why the microsim's control stops at 0.7. ### The effect of fiber orientation Rotating a lamina off-axis by an angle θ collapses its stiffness, and the collapse is steeper than intuition allows. The standard transformation gives `1/E_x = cos⁴θ/E₁ + (1/G₁₂ − 2·ν₁₂/E₁)·sin²θ·cos²θ + sin⁴θ/E₂`. With the carbon/epoxy preset — E₁ = 139.4, E₂ = 8.6, G₁₂ = 5 GPa and ν₁₂ = 0.3 — ten degrees of misalignment drops *E*x to 79.5 GPa, a loss of 43 %, and at 45° it is 12.5 GPa, a factor of eleven below the on-axis value (derived; G₁₂ and ν₁₂ are ILLUSTRATIVE typical values). Laminates therefore stack plies at several angles, commonly 0°, ±45° and 90°, trading peak directional stiffness for usable properties in every direction — the subject of [[Composite_laminate|composite laminate]] theory. ### Stiffness and Compliance Elasticity A lamina's in-plane elasticity needs four independent constants: E₁, E₂, G₁₂ and ν₁₂. They assemble into a stiffness matrix relating stress to strain, or its inverse the compliance matrix, each rotated into the laminate's axes ply by ply and summed through the thickness to give extensional, coupling and bending stiffnesses. An unsymmetric layup has non-zero coupling terms and warps when heated or stretched, which is why layups are normally symmetric about the mid-plane. ### Types of fibers and mechanical properties Glass is cheap, dense and moderately stiff; carbon covers a wide range, trading stiffness against strain to failure as precursor and heat treatment change; aramid is tough and light with excellent impact resistance but poor in compression, because the fibre kinks; boron and silicon carbide are stiff, costly and used with metal or ceramic matrices. Matrix choice is a separate axis: thermosets dominate structural work, thermoplastics are advancing on toughness and recyclability, and ceramic or carbon matrices take over above 500 °C. ### Failure Composites do not yield, and they do not fail in one mode. A laminate loaded in tension along the fibres fails by fibre breakage; in compression by fibre microbuckling or kink-band formation, which is why compressive allowables are typically half the tensile ones; transversely and in shear by matrix cracking, at strains low enough that the first cracks appear long before final failure. Between plies, interlaminar shear separates layers — delamination — the damage a dropped tool produces with no visible surface mark, which drives the whole barely-visible-impact-damage design philosophy.[^johnson-fracture] Multi-axial failure is predicted by interactive criteria such as the [[Tsai–Wu_failure_criterion|Tsai–Wu criterion]], which fits a quadratic envelope to measured uniaxial and shear strengths.[^tsai-wu][^johnson-ch9] ### Testing Because strength cannot be calculated from constituents, it is measured — on coupons cut in each principal direction and in shear — and design allowables are statistical lower bounds on a population of them. Non-destructive inspection matters more than for metals: ultrasonic C-scan for delamination and porosity, thermography for disbonds, X-ray tomography for internal geometry. The building-block approach — coupons, elements, subcomponents, full articles — exists because no single test scale tells the whole story. ## See also - [[Rule_of_mixtures]] - [[Fibre-reinforced_plastic]] - [[Carbon_fiber_reinforced_polymer]] - [[Fiberglass]] - [[Kevlar]] - [[Composite_laminate]] - [[Tsai–Wu_failure_criterion]] - [[Nacre]] - [[Material_selection]] - [[Finite_element_method]] ## References [^johnson-ch8]: Johnson, Eric R. (2022). *Aerospace Structures*. Blacksburg: Virginia Tech Publishing (Portal Book 009). Chapter 8, Composite beams and lamina stiffness, pp. 237–284 (page to pin). https://open.umn.edu/opentextbooks/textbooks/aerospace-structures [^johnson-ch9]: Johnson, Eric R. (2022). *Aerospace Structures*. Portal Book 009. Chapter 9, Composite failure, pp. 285–302 (page to pin). [^johnson-fracture]: Johnson, Eric R. (2022). *Aerospace Structures*. Portal Book 009. Chapter 13, Fracture, pp. 389–416 (page to pin) — the mode-I analysis and its restriction to shear stresses small against the normal stress, noted in sub-manual 02 §10.4 as a 4 % check in the worked design. [^voigt1889]: Voigt, W. (1889). "Ueber die Beziehung zwischen den beiden Elasticitätsconstanten isotroper Körper." *Annalen der Physik* 274 (12): 573–587. The uniform-strain average that carries his name. (DOI to pin.) [^reuss1929]: Reuss, A. (1929). "Berechnung der Fließgrenze von Mischkristallen auf Grund der Plastizitätsbedingung für Einkristalle." *Zeitschrift für Angewandte Mathematik und Mechanik* 9 (1): 49–58. The uniform-stress average. (DOI to pin.) [^hashin-shtrikman]: Hashin, Z.; Shtrikman, S. (1963). "A variational approach to the theory of the elastic behaviour of multiphase materials." *Journal of the Mechanics and Physics of Solids* 11 (2): 127–140. Bounds tighter than Voigt and Reuss for a statistically isotropic two-phase material. (DOI to pin.) [^kelly-tyson]: Kelly, A.; Tyson, W. R. (1965). "Tensile properties of fibre-reinforced metals: copper/tungsten and copper/molybdenum." *Journal of the Mechanics and Physics of Solids* 13 (6): 329–350. The critical-length argument used in the short-fibre section. (DOI to pin.) [^tsai-wu]: Tsai, S. W.; Wu, E. M. (1971). "A general theory of strength for anisotropic materials." *Journal of Composite Materials* 5 (1): 58–80. (DOI to pin.) [^jensen-plastics]: Jensen, David (2024). *Introduction to Mechanical Design and Manufacturing*. Portal Book 109. Chapter 16, Manufacturing — Plastic Processes, pp. 335–348 (page to pin) — moulding routes for the polymer matrices described here. [^barnes-dfam]: Barnes, John; Simpson, Timothy (2025). *Additive Manufacturing Essentials*. Portal Book 110. Chapter 4, Design for Additive Manufacturing, pp. 75–100; the 45° overhang rule and per-process feature-size allowables are at pp. 81–86 and 93. ## Further reading - Johnson, Eric R. (2022). *Aerospace Structures*. Virginia Tech Publishing. Portal Book 009 — Chapters 8 and 9 for lamina stiffness, composite beams and composite failure. https://open.umn.edu/opentextbooks/textbooks/aerospace-structures - Jensen, David (2024). *Introduction to Mechanical Design and Manufacturing*. Portal Book 109 — Chapter 16 for polymer processing. https://open.umn.edu/opentextbooks/textbooks/introduction-to-mechanical-design-and-manufacturing - Barnes, John; Simpson, Timothy (2025). *Additive Manufacturing Essentials*. Portal Book 110 — Chapters 3 and 4 for the additive routes now used with short- and continuous-fibre feedstock. ## External links - [Aerospace Structures](https://open.umn.edu/opentextbooks/textbooks/aerospace-structures), Eric R. Johnson, via the Open Textbook Library — the composite-beam and composite-failure chapters behind this page. - Further links, including the industry design handbooks and the test-standard indexes, are listed on the Wikipedia pair. <!-- MATTERSIM:BEGIN g24 — Matter & Energy Cluster microsim (framework build, specs/sims/Composite_material.json); do not hand-edit inside --> **Microsim — three.js (Wikitube framework), pending deploy:** *Composite material* will play here once `https://wikitube-3d-microsims.netlify.app/matter/Composite_material.html` is live. <!-- pending: <div class="wt-sim" data-src="https://wikitube-3d-microsims.netlify.app/matter/Composite_material.html" data-title="Composite material"></div> --> <!-- MATTERSIM:END --> ## Wikipedia : Wikitube **Strict pair:** [Wikipedia](https://en.wikipedia.org/wiki/Composite_material) : [Wikitube](https://en.wikitube.io/wiki/Composite_material) · pinned revision [1362339092](https://en.wikipedia.org/w/index.php?oldid=1362339092) · 2026-09-11 ## Previous hub tags Hubs: `Life_Physics`. Portals: [[PORTAL_Materials_science]]. --- *Matter & Energy Cluster child articles, wave 1 · 2026-09-11 · drafted · Materials_science row M47 · sim pending (matter/Composite_material).*