The Physics of Bone: Why the Strongest Material in Your Body Is Lighter Than Concrete

The Material That Builds Itself

Engineers spend careers designing composite materials. They layer carbon fibre into epoxy matrices, orient the fibres along load paths, cure them under heat and pressure, and test them to destruction. The best engineering composites — the ones in aircraft wings and Formula 1 chassis — are expensive, difficult to manufacture, and fundamentally inert. If they crack, they stay cracked.

Your skeleton is made of a composite material that exceeds most engineering composites in fracture toughness, weighs less than concrete, and has a trick that no synthetic material can match: it repairs itself. Crack a bone, and it will regenerate — not scar, not patch, but genuinely rebuild — restoring its original strength within months. Load it repeatedly, and it gets stronger in precisely the places where stress is highest. Unload it, and it removes material where it’s not needed.

Bone is not a dead scaffold. It’s a living, responsive, self-optimising material, and the physics of how it works — from the nanoscale structure of its composite architecture to the piezoelectric signals that guide its remodelling — is a masterclass in biological materials engineering.

The Nanocomposite: Collagen Plus Mineral

Bone is, at its core, a two-phase composite. The same design principle that makes fibreglass, carbon fibre, and reinforced concrete work — combine a stiff, brittle phase with a flexible, tough phase — is at work in every bone in your body.

The flexible phase is collagen — specifically, type I collagen, the most abundant protein in the human body. Collagen molecules are long, rope-like triple helices about 300 nm long and 1.5 nm in diameter. They self-assemble into fibrils (50–500 nm diameter) with a characteristic staggered arrangement: adjacent molecules are offset by about 67 nm, creating a periodic pattern of “gap” and “overlap” zones along the fibril. This 67 nm periodicity is visible in electron microscopy and is one of the most distinctive structural features in biology.

The stiff phase is hydroxyapatite — a calcium phosphate mineral with the chemical formula Ca₁₀(PO₄)₆(OH)₂. In bone, the hydroxyapatite forms as tiny plate-shaped crystals, roughly 50 × 25 nm in area and only 2–3 nm thick. These crystals nucleate preferentially in the gap zones between staggered collagen molecules, eventually growing to fill both the gaps and the spaces between fibrils.

The proportions: about 60–70% mineral and 30–40% collagen by weight, or about 45% mineral and 55% collagen by volume. The remaining ~10% is water, non-collagenous proteins, and living cells.

What matters is how the two phases interact.

Collagen alone is flexible and tough. It has a Young’s modulus of about 1.5 GPa (similar to nylon) and can stretch 10–20% before failing. It’s great under tension but buckles easily under compression.

Hydroxyapatite alone is stiff and brittle. It has a modulus of about 100 GPa (similar to glass) and a fracture toughness of only about 1 MPa·√m — it shatters if you look at it wrong.

Bone — the composite — has a modulus of about 15–25 GPa and a fracture toughness of 2–12 MPa·√m, depending on orientation and loading mode. That fracture toughness is 3–10 times higher than pure hydroxyapatite. The stiffness comes from the mineral. The toughness comes from the collagen. And the combination is far better than either component alone.

This is the fundamental principle of composite design: the whole exceeds the sum of its parts. Carbon fibre in epoxy works the same way. So does steel rebar in concrete. Bone got there about 500 million years ago.

Seven Levels of Architecture

What makes bone exceptional isn’t just that it’s a composite — it’s that the composite architecture is hierarchical, spanning seven orders of magnitude from nanometres to the whole bone.

Level 1 — Molecular (1–10 nm). Collagen triple helices and individual hydroxyapatite crystals. The crystals sit within and between collagen molecules, bonded by electrostatic and van der Waals interactions.

Level 2 — Fibrillar (50–500 nm). Mineralised collagen fibrils. The 67 nm stagger pattern creates a periodic structure where mineral crystals are organised along the fibril axis. Each fibril is a nanoscale composite beam.

Level 3 — Fibre (1–10 µm). Bundles of mineralised fibrils, often arranged in layers with different orientations. The fibrils are bound together by non-collagenous proteins and a thin layer of extrafibrillar mineral.

Level 4 — Lamellar (3–7 µm). In mature bone, fibres are arranged in concentric sheets called lamellae, each about 3–7 µm thick. Within each lamella, the fibre orientation is roughly uniform, but adjacent lamellae have different orientations — typically rotated by 30–90°. This cross-ply layup is exactly the same strategy used in plywood and cross-laminated timber. It provides quasi-isotropic mechanical properties in the plane of the lamellae.

Level 5 — Osteonal (100–300 µm). In cortical (compact) bone, lamellae are arranged concentrically around a central vascular channel to form osteons (Haversian systems) — cylinders about 200–300 µm in diameter. Each osteon is a multi-layered composite tube, with blood vessels running through the centre. The boundary between adjacent osteons — the cement line — is a weak interface that plays a critical role in fracture mechanics (more on this shortly).

Level 6 — Tissue (1–10 mm). Bone tissue comes in two types: cortical bone (dense, making up the outer shell of all bones and the shafts of long bones) and trabecular bone (spongy, porous, found inside the ends of long bones and inside vertebrae and flat bones). Cortical bone has a porosity of about 5–10%. Trabecular bone has a porosity of 50–90% — it’s a foam, with thin struts (trabeculae) arranged along the principal stress trajectories.

Level 7 — Whole bone (10 cm–1 m). The overall geometry of the bone: a long bone is a hollow tube with flared ends; a vertebra is a cylinder of trabecular bone with a cortical shell; the skull is a sandwich panel (two layers of cortical bone with trabecular bone in between).

Each level contributes to the overall mechanical behaviour. Damage a few collagen molecules? The fibrillar structure distributes the load around the defect. Crack at the osteonal level? The cement lines deflect and blunt the crack. Fracture the whole bone? The trabecular structure in the metaphysis distributes impact loads and the cortical shell provides bending stiffness.

This hierarchy of energy-dissipating mechanisms is what gives bone its remarkable fracture toughness — far exceeding what you’d predict from the properties of its constituent materials.

Fracture Mechanics: Why Bones Are Tough

A materials engineer looking at bone’s composition would predict a brittle material. It’s 60% ceramic by weight. Ceramics are stiff but brittle — they shatter. A plate of pure hydroxyapatite would break like a china cup.

But bone is tough. It absorbs enormous energy before fracturing. A femur can withstand compressive loads of 7,000–10,000 newtons (roughly 1,000 kg or the weight of a small car) before failing. How?

The answer lies in toughening mechanisms at multiple length scales:

Intrinsic Toughening (Ahead of the Crack Tip)

At the nanoscale, collagen fibrils bridge microcracks. When a crack tries to propagate through mineralised collagen, it must break or pull out fibrils — and each fibril absorbs energy through stretching, uncoiling, and interfibrillar sliding. The sacrificial bonds in the collagen matrix (weak, reversible cross-links between fibrils mediated by non-collagenous proteins) break sequentially, dissipating energy at each step. These bonds can reform after loading — effectively giving bone a molecular-scale self-repair mechanism that operates continuously, not just after fracture.

Extrinsic Toughening (Behind the Crack Tip)

At the microstructural level, the cement lines between osteons act as crack deflectors. When a crack propagating through bone reaches a cement line — which is a thin, highly mineralised interface — it tends to deflect along the interface rather than continuing straight through the next osteon. This crack deflection increases the total crack path length and therefore the total energy required for fracture.

If the crack tries to open (Mode I fracture), uncracked ligaments — bridges of intact bone spanning the crack behind the crack tip — apply closing forces that reduce the stress intensity at the crack tip. These ligament bridges can carry significant load, effectively shielding the crack tip and requiring higher applied stress to drive further growth.

Microcracking ahead of the main crack tip creates a process zone of distributed damage. Each microcrack absorbs energy, and the collective effect is to blunt the main crack tip and distribute strain over a larger volume. This is similar to what happens in concrete — but bone does it better because the microcracks can be repaired by biological remodelling.

The result of these combined mechanisms: bone’s fracture toughness increases as a crack grows (a property called a rising R-curve). Short cracks are easier to propagate; longer cracks are harder because more toughening mechanisms engage as the crack extends. This is the opposite of what happens in most engineering ceramics, where cracks become easier to propagate as they grow (leading to catastrophic, brittle failure).

Wolff’s Law: The Feedback Loop

In 1892, the German surgeon Julius Wolff published Das Gesetz der Transformation der Knochen — “The Law of Bone Transformation.” His observation, now called Wolff’s law, is simple: bone adapts its internal architecture and external shape to the mechanical loads it habitually bears.

Load a bone more → it gets stronger. Unload it → it gets weaker.

The evidence is everywhere:

The dominant arm of professional tennis players has 30–40% more cortical bone than the non-dominant arm. The difference develops over years of asymmetric loading and is visible on X-rays.

Astronauts in microgravity lose 1–2% of bone mass per month in load-bearing bones (spine, pelvis, legs). After six months on the International Space Station, a crew member may have lost 10% of their hip bone density — equivalent to about a decade of age-related bone loss on Earth.

Bedridden patients lose bone rapidly. In contrast, high-impact athletes (gymnasts, sprinters) have bone densities 10–20% above average.

The trabecular bone inside the femoral head is arranged along the principal stress trajectories — the paths of maximum compressive and tensile stress during walking. The alignment is so precise that it matches predictions from finite element analysis of the load distribution. When the loading pattern changes (due to a hip replacement, for example), the trabecular architecture remodels to match the new stress field. Karl Culmann, a Swiss engineer, noticed the correspondence between trabecular patterns and the stress trajectories in a crane design as early as 1866 — a connection that directly inspired Wolff’s work.

The Mechanosensors: Osteocytes

How does bone sense mechanical load? The sensors are osteocytes — mature bone cells embedded within the mineralised matrix, connected to each other and to the bone surface by a network of long cellular processes running through tiny channels called canaliculi (about 0.1–0.5 µm in diameter).

There are about 25,000 osteocytes per cubic millimetre of bone, connected by roughly 100 canaliculi each — forming a sensing network of extraordinary density. No point in a piece of cortical bone is more than about 100 µm from the nearest osteocyte.

When bone is loaded, the matrix deforms very slightly (strains of 0.1–0.3% during normal activity). This deformation squeezes interstitial fluid through the canaliculi, creating fluid shear stress on the osteocyte cell processes. The shear stress is detected by mechanosensitive structures on the cell surface — including primary cilia, integrins, and ion channels — which trigger intracellular signalling cascades.

The osteocytes then communicate with the bone-forming cells (osteoblasts) and bone-resorbing cells (osteoclasts) on the bone surface. In regions of high mechanical stress, osteocytes signal osteoblasts to deposit new bone. In regions of low stress, they signal osteoclasts to resorb bone. The result: material is added where it’s needed and removed where it’s not — a continuous optimisation algorithm running in your skeleton.

In engineering terms, Wolff’s law describes a structure with a built-in feedback loop: load → strain → fluid flow → cellular signal → remodelling → adjusted structure → changed strain distribution. The loop runs continuously, with a time constant of weeks to months. Your skeleton today is not the same skeleton you had a year ago — about 10% of the adult skeleton is remodelled every year.

Piezoelectricity: Electricity From Stress

Bone generates electrical voltages when stressed. Compress one side of a bone, and that side develops a negative surface charge. The stretched side develops a positive charge. The voltages are small — microvolts to millivolts — but they’re real and measurable.

This is piezoelectricity — the same property that makes quartz crystals vibrate in watches and ignite gas lighters. In bone, the piezoelectric response comes primarily from the collagen component. Collagen has a non-centrosymmetric crystal structure (it lacks inversion symmetry), which is the fundamental crystallographic requirement for piezoelectricity. When the collagen lattice is deformed, the displacement of charged atoms creates a net electrical polarisation.

In addition to true piezoelectricity, bone generates streaming potentials — voltages created by the pressure-driven flow of charged interstitial fluid through the canalicular network. The fluid contains ions, and when it flows past the charged surfaces of the canaliculi, it creates a measurable voltage (essentially the same physics as an electrokinetic flow sensor).

Both mechanisms produce the same pattern: negative charge on compressed surfaces, positive charge on tensile surfaces. And here’s the biological payoff: osteoblasts (bone-forming cells) are preferentially stimulated by negative electrical potentials, while osteoclasts (bone-resorbing cells) are stimulated by positive potentials.

This means the piezoelectric effect provides a direct physical mechanism for Wolff’s law. When a bone is loaded in bending, the compressed (concave) side develops a negative charge → attracts osteoblasts → bone is deposited → the compressed side gets thicker. The tensile (convex) side develops a positive charge → attracts osteoclasts → bone is resorbed → the tensile side gets thinner. The net effect: bone material migrates toward the compression side, exactly where it’s most needed structurally.

Electrical stimulation of bone healing — a clinical technique where small electric currents are applied to fracture sites to accelerate healing — is directly based on this physics. Clinical trials show measurable improvements in healing rates for non-union fractures when electrical stimulation is applied. The body’s own piezoelectric signalling mechanism is augmented artificially.

Why Bones Are Hollow Tubes

Look at a cross-section of a femur or tibia — it’s a tube. Dense cortical bone on the outside, marrow cavity on the inside. Why not a solid rod?

The answer is a standard result from structural mechanics: for resisting bending and torsion, a hollow tube is far more efficient than a solid rod of the same mass.

The second moment of area (I) determines bending stiffness. For a solid cylinder of radius R:

I_solid = πR⁴/4

For a hollow cylinder with outer radius R and inner radius r:

I_hollow = π(R⁴ − r⁴)/4

The key insight: the material near the centre of a solid rod contributes almost nothing to bending resistance (I depends on R⁴, so material close to the centre, where R is small, contributes very little). By removing this low-contribution central material and redistributing it to the outer surface (increasing R), you dramatically increase the second moment of area — and therefore the bending stiffness — for the same total mass.

A typical long bone has an outer-to-inner radius ratio of about 0.5–0.6. This geometry provides roughly 3–5 times the bending stiffness of a solid bone of the same mass. The mass saving is substantial: a solid femur of the same bending stiffness would weigh about twice as much.

The trade-off: thin-walled tubes are susceptible to local buckling — the wall can crush inward under compression. Long bones avoid this by maintaining a minimum wall thickness and by filling the cavity with marrow (which provides some internal pressure support and serves the entirely separate function of producing blood cells). The thickness-to-radius ratio of cortical bone in long bones is optimised to balance bending stiffness against buckling risk — a solution that aerospace engineers would recognise immediately from aircraft fuselage design.

Trabecular bone in the flared ends of long bones (the epiphyses) serves a different structural function: it distributes the concentrated joint loads over a larger area, reducing peak stresses in the cortical shell. Trabecular architecture functions like a foam core in a sandwich panel — low density, moderate stiffness, excellent energy absorption.

Bone as a Material: The Numbers

How does bone compare to engineering materials?

Property	Cortical Bone	Steel	Aluminium	Carbon Fibre Composite	Concrete
Density (kg/m³)	1,800–2,000	7,800	2,700	1,600	2,400
Young’s Modulus (GPa)	15–25	200	70	70–150	30
Tensile Strength (MPa)	100–150	400–800	200–500	500–2,000	2–5
Compressive Strength (MPa)	150–250	400–800	200–500	200–700	20–40
Fracture Toughness (MPa·√m)	2–12	50–200	20–40	20–50	0.5–1.5

Bone is not the stiffest, not the strongest, not the toughest material on the table. But consider what it does that none of the others can: it self-repairs, it adapts to loading, it’s produced at body temperature from dietary calcium and protein, and it lasts for decades under millions of loading cycles.

The fatigue life of bone is particularly impressive. Under the cyclic loading of walking (about 2,000 cycles per day per leg), bone sustains strains that would cause fatigue failure in most engineering materials within years. Bone survives because it continuously repairs fatigue-induced microcracks through remodelling. The osteocyte network detects microcracks (a crack disrupts the canalicular network, cutting off fluid flow signals to downstream osteocytes), which triggers targeted remodelling: osteoclasts tunnel into the damaged region, remove the cracked bone, and osteoblasts refill the tunnel with fresh lamellar bone. This targeted repair process replaces damaged material before microcracks can grow to critical size.

Osteoporosis — the loss of bone mass and deterioration of bone architecture with age — is fundamentally a failure of this remodelling balance. When osteoclast activity exceeds osteoblast activity (due to hormonal changes, disuse, nutritional deficiency, or genetic factors), bone mass declines, trabeculae thin and disconnect, cortical bone becomes more porous, and fracture toughness drops. The physics doesn’t change. The biology’s ability to maintain the structure does.

What Bone Teaches Us About Design

Materials engineers study bone because it demonstrates principles that we struggle to replicate synthetically.

Hierarchical architecture provides toughening at every length scale. Engineering composites typically have 2–3 levels of structure. Bone has seven. Each level contributes distinct mechanical properties, and the interfaces between levels (especially the cement lines between osteons) serve as crack-deflection barriers that dramatically increase fracture toughness.

Self-sensing and self-optimisation through the osteocyte network and Wolff’s law. Imagine an aircraft wing that could sense where fatigue damage was accumulating and automatically add material to those locations while removing material from overstressed regions. That’s what bone does, continuously.

Self-repair through targeted remodelling of microdamage. A bridge that could detect and repair its own cracks before they reached critical size would be a revolution in structural engineering. Bone has been doing it for half a billion years.

Piezoelectric feedback that links mechanical loading directly to biological remodelling. The physics of electromagnetism and the biology of cell signalling are coupled through a mechanism that automatically directs material deposition to where it’s most structurally efficient.

Some of the most promising directions in materials science are directly inspired by bone: self-healing concrete (embedding microcapsules of healing agent that release when a crack ruptures them), adaptive structures (embedding sensors and actuators in composites to adjust properties under load), and hierarchical composites (designing structures with multiple nested length scales of reinforcement).

We’re getting closer. But we haven’t matched bone yet. The femur you’re standing on right now is a self-sensing, self-healing, load-adaptive, hierarchically structured nanocomposite tube that grows itself from calcium and protein, operates for 80 years under millions of loading cycles, and weighs less than half a kilogram.

That’s not bad for a material that assembles itself at 37 degrees.