Friction does not depend on how much surface is touching. The reason is that the surfaces are barely touching at all

Every physics course teaches that the force of friction depends on how hard two surfaces are pressed together and on what they are made of, but not on how large the contact is. A brick sliding on its broad face and the same brick sliding on its narrow edge experience the same friction.

Almost nobody believes this on first hearing, and the disbelief is reasonable. It contradicts an intuition so strong that people assume the rule must be a simplification hiding the real answer.

It is not a simplification. It is true, and the reason is that the question contains a false assumption. The two surfaces are not touching over the area you can see. They are touching at a scatter of tiny points that add up to roughly one per cent of the area that appears to be in contact — and once you know that, the law stops being strange and becomes almost obvious.

Real contact area is not the area you can see

No surface is flat. Machine it, grind it, polish it to a mirror, and at the scale of micrometres it is still a landscape of peaks and valleys. Press two such landscapes together and they meet only where a peak on one lands on a peak on the other. Engineers call those points asperities.

The load is therefore carried by a small number of very small junctions, at enormous local pressure — high enough that the asperity tips deform plastically, squashing until there is just enough contact area to support the load without deforming further.

That last sentence is the whole explanation. Real contact area is set by the load, not by the apparent area. Put the brick on its narrow edge and the same weight is carried by fewer asperities squashed harder, producing the same total real contact. Put it on its broad face and the load spreads over more asperities, each less deformed — again the same total. Double the weight and you double the real contact, which is why friction is proportional to load.

This is the same fact that appeared in our article on cooling a chip, where the microscopic air gap between two nominally flat surfaces is what actually limits heat flow. One physical reality — surfaces touch far less than they appear to — produces a thermal bottleneck in one context and an apparently paradoxical law in another.

What friction actually is

At those junctions, the two materials are pressed into intimate contact hard enough to bond. Clean metal surfaces in contact form genuine adhesive junctions — a kind of cold welding — and sliding means continuously tearing those junctions apart and forming new ones a moment later. Friction is largely the force needed to keep shearing them.

A second contribution is ploughing: a hard asperity gouging through a softer surface, which costs energy and also removes material. That is the beginning of wear, and it is why friction and wear are related but not the same thing — a system can have low friction and high wear, or the reverse.

The junction picture explains a family of everyday phenomena at once. Static friction exceeds kinetic friction because a stationary contact gives junctions time to grow and settle, so the first movement needs more force than continued movement does. And once you have that, you have stick-slip: the surface sticks, force builds, the junctions snap, it slides, it sticks again — repeating fast enough to make a sound.

A squeaking door hinge, a screeching brake, a chair dragged across a floor and the note of a violin string under a bow are all the same mechanism, running at different frequencies. The violin is stick-slip that somebody learned to control.

The three ways a lubricant works, and why startup is the dangerous part

Lubrication is the business of keeping asperities from touching. It happens in three distinct regimes, and confusing them is behind a lot of bad maintenance advice.

Hydrodynamic lubrication is the ideal. Relative motion drags oil into a converging gap and generates enough pressure to lift the surfaces completely apart. The solids never touch; the only resistance is the oil shearing against itself. In this regime friction is set by viscosity, wear is essentially zero, and a properly designed bearing can run almost indefinitely.

Boundary lubrication is what happens when that fails — at low speed, high load, or the moment of starting, when there is not enough motion to build a film. Asperities do touch, and the only thing between them is a molecular layer of additives that have chemically attached themselves to the metal. This is where the additive package in an oil earns its cost: anti-wear and extreme-pressure compounds that bond to the surface and form a sacrificial film which shears instead of the metal.

Mixed lubrication sits between the two, with part of the load on a fluid film and part on asperity contact.

The practical consequence is worth stating plainly: most engine wear happens in the first seconds after a cold start, when oil has drained from the upper surfaces and the film has not yet formed. Not during hard driving, not at high speed — at startup. This is why frequent short trips are harder on an engine than long runs, and it is the same short-trip penalty that appeared in our article on catalytic converters for entirely different reasons.

It also disposes of a common belief. Thicker oil is not automatically safer. A grade heavier than specified takes longer to reach the parts that need it at startup, absorbs more power to shear, and can worsen precisely the regime where wear actually occurs. The specified grade is an engineering decision about clearances and pump capacity, not a manufacturer being stingy.

Friction is not a property of a material. It is a property of a pair of surfaces, in a particular environment, at a particular speed and load — which is why a coefficient of friction quoted without those conditions is not a number you can use.

Why some solids are slippery

Some materials lubricate without any oil at all, and the reason is structural rather than chemical.

Graphite, molybdenum disulphide (MoS₂), tungsten disulphide (WS₂) and hexagonal boron nitride are all layered materials. Within a layer the atoms are held by strong covalent bonds; between layers there is only weak van der Waals attraction. Shear the material and the layers slide over one another easily while each layer stays intact. It is the same structural argument that makes graphene strong in-plane, viewed from the opposite direction: strong bonds one way, weak bonds the other, and a solid lubricant exploits the weak direction.

Then comes a genuinely surprising fact, and it is one of the better cautionary tales in materials engineering.

Graphite is not intrinsically slippery. Its lubricating behaviour depends on water vapour adsorbed between the layers, which weakens the interlayer interaction. In dry air or in vacuum, that adsorbed layer is absent, graphite's friction rises sharply, and it fails as a lubricant — a discovery made expensively when graphite-lubricated mechanisms were taken to high altitude and later into space, where they seized.

Molybdenum disulphide has no such dependence. It lubricates in dry conditions and in vacuum, which is why spacecraft mechanisms, satellite deployment hinges and vacuum equipment use MoS₂ coatings rather than graphite. Two layered solids that look interchangeable in a textbook behave oppositely in the one environment where getting it wrong is unrecoverable.

Nanoparticles in oil: what is real and what is sold

Nanomaterials enter this field mainly as oil additives, and it is worth separating the mechanisms from the marketing.

The proposed mechanisms are plausible and several are demonstrated. Layered nanoparticles of MoS₂ or WS₂ can exfoliate under pressure and deposit a thin lubricating film directly onto the asperity contacts — a tribofilm formed where it is needed. Roughly spherical fullerene-like WS₂ nanoparticles are argued to act partly as rolling elements before exfoliating. Hard nanoparticles can also polish asperity tips, and some additives fill in surface valleys, both of which reduce contact severity.

Laboratory results on these are real, and commercial products based on them exist in industrial settings where the oil, the load and the maintenance interval are all controlled.

The honest caveats are three. Dispersion stability is the usual nanomaterial problem — particles that agglomerate settle out or, worse, become abrasive third bodies that accelerate wear rather than reducing it. Filtration is a genuine conflict: an engine's oil filter is designed to remove particles, and a nanoparticle additive either passes through it or is removed by it, and which of those happens matters. And consumer additive claims are largely untested independently; a modern engine oil is already a carefully balanced additive package, and introducing an unqualified extra can interfere with the chemistry that is already working — which is why manufacturers generally advise against aftermarket additives.

At the frontier is superlubricity: when two crystalline surfaces slide with their lattices rotationally mismatched, the atomic-scale forces cancel out across the contact and friction falls to nearly nothing. This has been demonstrated convincingly with graphite flakes at micrometre scale. Whether it can be maintained over large, imperfect, contaminated engineering surfaces is an open question, and it is the kind of result that is entirely real and entirely not yet a product.

Why it matters for students and researchers

Tribology is one of the most economically significant subjects that most engineering graduates never formally study. Widely cited estimates put the share of global energy consumption spent overcoming friction at close to a fifth, with a further large cost in wear, replacement parts and downtime. Even a modest fractional improvement in either is an enormous absolute number, which is why the field attracts serious industrial funding and comparatively little classroom time.

It is also a good corrective to the habit of treating properties as belonging to materials. There is no such thing as "the coefficient of friction of steel". There is a coefficient for steel against a particular counterface, at a particular roughness, load, speed, temperature and humidity, with a particular film in between — and changing any of those changes the answer. Friction is a system property, and learning to think that way transfers directly to adhesion, wear, contact resistance and heat transfer.

The open problems are substantial. Predicting friction from first principles for real engineering surfaces remains largely out of reach, and design still leans heavily on empirical data. Lubrication for electric vehicles is a genuinely new problem, because electric drivetrains have different speeds, no fuel dilution, and stray currents through bearings that conventional oils were never formulated for. Extending superlubricity beyond ideal laboratory contacts is a live goal. And texturing surfaces deliberately at the micro and nano scale to trap lubricant and control contact, rather than simply polishing them smoother, is an approach that keeps producing results counter to intuition.

Frequently asked questions

If friction really does not depend on area, why do racing cars use wide tyres?

Because rubber on road is not the simple case the law describes. Tyre grip includes a large adhesion component that does scale with contact area, and rubber deforms around road texture rather than contacting at rigid asperities. Wider tyres also spread heat and wear over more material, which matters enormously over a race distance. The classical law is a very good description of dry sliding between hard, stiff materials, and tyres are deliberately none of those things.

Why does a door hinge squeak, and why does oil fix it?

The squeak is stick-slip: the surfaces alternately grip and release, and the frequency of that cycle falls in the audible range. Oil works by separating the surfaces so junctions never fully form, which removes the grip-release cycle rather than merely quietening it. If oil stops fixing it, the surfaces have usually worn enough that the geometry itself is the problem.

Is thicker engine oil better protection?

Usually not, and often worse. The specified grade is chosen for the engine's clearances, oil pump and operating temperatures. Heavier oil is slower to reach bearings after a cold start, which is when most wear occurs, and it consumes more power in shear. Where a heavier grade is genuinely appropriate — a worn engine with enlarged clearances, or an unusually hot duty cycle — it is a deliberate trade rather than a general upgrade.

Do nanoparticle oil additives actually work?

The mechanisms are real and demonstrated in laboratory and controlled industrial use. For a consumer pouring a bottle into a modern engine, the case is much weaker: the oil already contains a balanced additive package, independent verification of retail products is thin, and poorly dispersed particles can act as abrasives. If a manufacturer specifies an oil, meeting that specification reliably is worth more than any additive.

Does WD-40 lubricate?

Only lightly and briefly. It is primarily a penetrant and water displacer — the WD stands for water displacement — with a low-viscosity carrier that evaporates and leaves very little behind. It is excellent for freeing a seized fastener or displacing moisture, and a poor choice for a hinge, chain or bearing that needs lasting lubrication, where a proper grease or oil belongs. Using it as a general lubricant is one of the most common maintenance mistakes there is.