Finding the coefficient of friction between two surfaces comes down to measuring how much force it takes to slide one across the other, then dividing that friction force by the force pressing them together. That ratio is the coefficient of friction. The concept dates back centuries, but the specific number you get depends heavily on how you measure it, what conditions the surfaces are in, and whether you care about the force needed to start motion or to keep it going. Several well-established lab methods exist, from tilting a ramp until an object slides to dragging a weighted sled with a force sensor, and newer techniques can measure friction at scales as small as a single atom or as inaccessible as a buried interface inside a machine.
The Inclined Plane Method
The simplest and oldest way to find a coefficient of friction requires nothing more than a flat surface you can tilt. You place your object on the surface, raise one end slowly, and note the angle at which the object just begins to slide. That critical angle, sometimes called the angle of repose, gives you the static coefficient of friction directly: it equals the tangent of that angle. If the object starts moving at 27 degrees, for instance, the tangent of 27 degrees is about 0.51, so the static coefficient of friction between those two materials is roughly 0.51.
This approach works because at the tipping point, the component of gravity pulling the object down the slope exactly matches the maximum friction force holding it in place. No force sensor is needed, just a protractor and a steady hand. A study testing several material pairs on an inclined plane apparatus confirmed that the method reliably reproduces known friction values from the literature while also allowing you to extract additional information like the mechanical advantage and efficiency of the inclined surface.1International Journal of Engineering, Science and Mathematics. Study of the co-efficient of static friction between different pairs of surfaces using Inclined Plane Apparatus
The main limitation is that the inclined plane method gives you only the static coefficient. Once the object starts sliding, it accelerates down the ramp, and measuring the kinetic coefficient this way requires more sophisticated tracking of the object’s acceleration, which is possible but less straightforward. It also works best with rigid, flat-bottomed objects. Anything that rolls, deforms, or has an irregular contact patch introduces errors.
The Horizontal Pull Method
If you want both the static and kinetic coefficients in one test, the horizontal pull method is the standard approach. You place an object on a flat surface, attach a force gauge or load cell, and pull horizontally. The peak force just before the object starts moving, divided by the object’s weight, gives you the static coefficient. The steady force required to keep it sliding at a constant speed, divided by the weight, gives you the kinetic coefficient.
A study measuring friction during weighted sled pulls on artificial turf illustrates how this works in practice. Researchers tethered a sled to a force transducer and pulled it at constant velocity with a winch, testing three different loads. The static friction coefficient ranged from about 0.39 to 0.47, while the kinetic coefficient ranged from about 0.31 to 0.35, with the heavier loads consistently producing slightly lower values.2PubMed. Determination of friction and pulling forces during a weighted sled pull That load dependence is a useful detail: even in a controlled setup, the number you get can shift depending on how hard the surfaces are pressed together.
The horizontal pull method is the workhorse of introductory physics labs and industrial quality control alike. Spring scales work for rough measurements, while electronic load cells and data-acquisition systems give you continuous force-versus-time curves that clearly show the spike at breakaway (static friction) and the plateau during steady sliding (kinetic friction).
Pin-on-Disc and Rotary Tribometry
When you need friction data for engineering materials under controlled, repeatable conditions, the pin-on-disc tribometer is the industry standard. A small pin or ball of one material is pressed against a rotating disc of another material under a known load. Sensors measure the tangential (friction) force continuously, and dividing by the applied normal force gives you a running coefficient of friction over time, distance, or number of revolutions.
One advantage of this setup is that it captures how friction evolves. A pair of surfaces might start with a high friction coefficient that drops as they wear in, or start low and climb as surface films break down. Research on pin-on-disc experiments found that the friction force trace often oscillates rather than holding steady, because the wear track develops non-uniform surface features and uneven lubrication as the test runs. The researchers concluded that reporting friction as a range of values, rather than a single number, gives a more honest picture of real sliding conditions.3Tribology International. Friction oscillations with a pin-on-disc tribometer
Ball-on-disc tribometers, a close variant, are commonly used to generate data across different lubrication regimes, covering everything from dry contact to fully lubricated sliding. Researchers have used these instruments to build large datasets spanning multiple materials, lubricants, temperatures, loads, and sliding speeds, capturing the full transition from boundary lubrication through mixed lubrication to hydrodynamic lubrication.4Lubricants. Deep Learning Data-Driven Model for Stribeck Curve Prediction of Lubricated Tribo-Pairs
Measuring Friction at the Nanoscale
At very small scales, traditional methods break down. You cannot put a micromechanical component on a ramp and tilt it. Atomic force microscopy (AFM) and friction force microscopy (FFM) fill that gap. An extremely sharp tip mounted on a flexible cantilever is dragged across a surface while sensors track how much the cantilever twists. That twist is proportional to the friction force. By scanning the tip at different applied loads, you can plot friction force versus normal force, and the slope of that line gives the coefficient of friction.
Researchers have used this technique to study friction and wear on thin films at scan sizes as small as one micrometer across, obtaining friction coefficients from that force-versus-load slope and generating spatial maps of friction across the surface.5Acta Materialia. Micro/nanoscale friction and wear mechanisms of thin films using atomic force and friction force microscopy These maps reveal that friction is not uniform even across a seemingly smooth surface. Tiny variations in composition, crystal orientation, or contamination create patches of higher and lower friction that would be invisible to any macroscopic test.
At the atomic scale, computational methods supplement physical measurements. First-principles molecular dynamics simulations have been used to study how adhesive bonds forming and breaking across a sliding interface affect static friction, showing that the number and strength of chemical interactions at the contact fundamentally control the force needed to initiate sliding.6PubMed. Effect of adhesive interactions on static friction at the atomic scale
Why the Number Changes With Conditions
A common frustration when measuring friction is that the number refuses to stay put. You test the same two materials on Monday and Friday and get different results. That is not your equipment failing. Friction is genuinely sensitive to conditions that might seem minor.
Surface roughness is one of the biggest variables. Research on floor-safety friction found that among the roughness parameters studied, the average peak height above the surface mean line and the average slope of the surface asperities showed the strongest correlation with friction.7Safety Science. The effect of surface roughness on dynamic friction between neolite and quarry tile In plain terms, it is not just how rough a surface is in a generic sense but the specific shape and spacing of its peaks and valleys that determine how much grip it provides. Two surfaces with the same “average roughness” can have very different friction coefficients if their microscopic geometry differs.
Temperature matters too, and not in a simple “hotter means more friction” way. Experiments on thermoplastic materials sliding against aluminum found that the friction coefficient changed in complex, non-linear patterns as temperature and contact pressure varied. For one polymer tested, friction was lowest (around 0.05) at moderate pressures and temperatures near freezing, while the highest values (around 0.19) appeared at low pressure and low temperature. Another polymer showed the opposite trend, with friction climbing steadily as temperature dropped at low pressure, ranging from about 0.04 up to 0.18.8PubMed Central. Effect of Temperature on the Tribological Properties of Selected Thermoplastic Materials Cooperating with Aluminium Alloy The takeaway is that you should measure friction at or near the actual operating temperature rather than assuming room-temperature values will transfer.
Humidity can be just as tricky. Studies on magnetic tape found that at temperatures above 10°C, friction decreased with rising relative humidity, reaching a minimum around 50–80% humidity before climbing again. But below 5°C, the relationship flipped: friction increased with humidity, peaking in that same 50–80% range. The coefficient of friction values ranged from lows around 0.06–0.08 to highs of 0.15–0.18 depending on the combination of temperature and humidity.9Elsevier / ScienceDirect (Wear). The effect of temperature and humidity on the friction and wear of magnetic tape If your application involves surfaces exposed to the atmosphere, controlling or at least recording humidity during testing is essential for reproducible results.
Sliding Speed and Lubrication Regimes
How fast one surface moves across another can dramatically change the friction coefficient, especially when a lubricant is present. The classical way to visualize this is the Stribeck curve, which plots friction against a parameter that combines sliding speed, lubricant viscosity, and applied load. At very low speeds or high loads, the lubricant film is too thin to separate the surfaces, and direct contact between surface peaks dominates. This is boundary lubrication, and friction is relatively high. As speed increases or load drops, a mixed regime develops where the lubricant partially separates the surfaces. At high enough speed, a full fluid film forms between the surfaces and friction drops to its lowest levels in the hydrodynamic regime.
Molecular dynamics simulations of this transition have shown that different lubricant molecules behave differently across these regimes. One study found that a longer-chain hydrocarbon performed better under hydrodynamic conditions but actually produced more friction under boundary lubrication than a shorter-chain molecule, because the two formed structurally different films when squeezed between nearly touching surfaces.10Friction. Molecular dynamics simulation of the Stribeck curve: Boundary lubrication, mixed lubrication, and hydrodynamic lubrication on the atomistic level
Even without a lubricant, sliding velocity matters. Research in geophysics on granular fault materials has shown that friction parameters change with velocity, with some parameters remaining essentially constant while others show a clear negative dependence on speed.11Journal of Geophysical Research: Solid Earth. Velocity Dependence of Rate‐And‐State Friction in Granular Fault Gouge and Implications for Slow‐Slip Events Depending on the specific parameters, friction can either increase with speed (rate-strengthening, which leads to smooth steady sliding) or decrease with speed (rate-weakening, which leads to jerky stick-slip behavior).12arXiv. Learning a potential formulation for rate-and-state friction This is why specifying the sliding velocity in your test protocol is not optional: two measurements on the same material pair can give you completely different coefficients if one was done at a crawl and the other at speed.
Measuring Friction Without Taking Things Apart
All the methods described so far require direct access to the surfaces. But what if the interface is buried inside an assembled machine or a sealed bearing? Ultrasonic techniques offer a way to probe friction at interfaces you cannot see or reach.
Researchers have developed a method that sends high-power ultrasonic waves through a contact interface and analyzes the nonlinear response. Because friction changes how the surfaces interact with the wave, it becomes possible to back-calculate the friction coefficient from the signal. In tests on different surface configurations, this approach yielded friction values ranging from 0.22 to 0.61, consistent with what you would expect from those material pairs using conventional methods.13PubMed Central. Measuring friction at an interface using ultrasonic response The technique is still relatively new but points toward a future where friction monitoring happens in real time inside operating equipment, without stopping anything.
Friction in Biological Tissues
Friction measurement gets considerably more complicated when the surfaces are alive or were recently part of a living organism. Articular cartilage, the smooth tissue covering the ends of bones inside joints, has been extensively studied with tribological methods because understanding its friction behavior matters for designing joint replacements.
The challenge is that cartilage is not a simple solid. It is a porous, fluid-filled material whose friction behavior changes over time as fluid is squeezed out under load. Researchers measuring friction of bovine cartilage under both static and dynamic loading have found that the friction coefficient evolves throughout a test as the tissue creeps and deforms.14PubMed Central. Effect of dynamic loading on the frictional response of bovine articular cartilage Studies measuring cartilage friction against different orthopedic implant materials have tracked this evolution over multiple hours, extracting friction values at the start of the test, at near-equilibrium, and at the final time point. The initial friction coefficient did not differ among counterface materials, but as the tissue settled into equilibrium, significant differences emerged depending on which metal or ceramic the cartilage was sliding against.15PubMed Central. Wear and Damage of Articular Cartilage with Friction Against Orthopaedic Implant Materials
For evaluating replacement biomaterials meant to fill cartilage defects, researchers have developed specialized pin-on-disc setups where the biomaterial is placed inside an artificial defect in real cartilage, then tested in sliding against another cartilage surface in a physiological solution. One study using this approach found median friction coefficients of 0.36 for cartilage sliding on cartilage, 0.38 for a cartilage-hydrogel composite against cartilage, and 0.32 for cartilage against aluminum.16PubMed. A method for the assessment of the coefficient of friction of articular cartilage and a replacement biomaterial The fact that a living tissue’s friction coefficient depends on loading history, fluid content, and test duration makes biological friction measurement an area where the test protocol is inseparable from the result.
Testing in Unusual Atmospheres
Most friction measurements happen in ambient air, but some applications demand data in environments you cannot just walk into. A preliminary study tested the friction and wear of aluminum, titanium, and polymer samples in a simulated Martian atmosphere of carbon dioxide, comparing results with those in nitrogen and air. Nitrogen and air produced statistically identical friction for both metals, but carbon dioxide increased friction for aluminum while decreasing it for titanium. The polymer samples showed no sensitivity to atmosphere at all.17Elsevier (Wear). Friction and wear behavior of engineering materials in a simulated Martian (CO2) environment, a preliminary study For anyone designing hardware for space missions or pressurized industrial systems, the lesson is straightforward: you cannot assume that a friction coefficient measured in air will hold in a different gas.
Road and Pavement Friction Testing
One of the most practical applications of friction measurement happens on roads. Transportation engineers routinely assess pavement skid resistance to decide when surfaces need treatment. Two of the most common instruments for this are the British Pendulum Tester, a portable device that swings a rubber slider across the pavement and measures energy lost to friction, and the Dynamic Friction Tester, which spins a set of rubber pads on the surface at decreasing speed to map friction across a range of velocities.
These two instruments do not always agree. A study comparing them on polished pavements found that the British Pendulum Tester could detect directional differences in skid resistance caused by traffic polishing, meaning friction measured in the direction of traffic flow differed from friction measured across it. The Dynamic Friction Tester could not pick up those directional differences and tended to overestimate friction in the traffic direction.18Transportation Research Record: Journal of the Transportation Research Board. Directional Skid Resistance Characteristics of Road Pavement: Implications for Friction Measurements by British Pendulum Tester and Dynamic Friction Tester If you are relying on friction measurements for safety decisions, knowing which instrument was used and what its limitations are is just as important as the number itself.
The Persistent Myth of Fixed Friction Coefficients
Physics textbooks often present tables of friction coefficients as though they are material constants, the way density or melting point are constants. Steel on steel: 0.74 static, 0.57 kinetic. Rubber on concrete: 1.0. These tables are useful starting points, but treating them as fixed truths will get you into trouble. As every section above illustrates, the coefficient of friction is not a property of a material pair alone. It is a property of a system: the two materials, their surface finish, the temperature, the humidity, the sliding speed, the load, the lubricant or lack thereof, and even the atmosphere.
Leonardo da Vinci recognized the basic proportionality between friction force and normal load over five hundred years ago, along with the observation that friction does not depend on apparent contact area. Researchers who repeated his experiments found that they could reproduce da Vinci’s reported friction coefficient of 0.25 for wood, but only when using roughly cut and hand-handled samples, not clean laboratory specimens.19SpringerLink. Leonardo da Vinci’s Friction Experiments: An Old Story Acknowledged and Repeated Surface preparation changed the number. Modern modeling work confirms that the Amontons-Coulomb laws, stating that friction force is proportional to normal load and independent of apparent contact area, emerge naturally from the mechanics of interlocking surface bumps under load.20PubMed. Model of Friction with Plastic Contact Nudging: Amontons-Coulomb Laws, Aging of Static Friction, and Nonmonotonic Stribeck Curves with Finite Quasistatic Limit The laws are real, but the specific coefficient you plug into them is not a universal constant. It is an empirical measurement that belongs to a specific set of conditions.
For anyone setting up a friction test, the practical takeaway is to document everything: the surface preparation, the normal load, the sliding speed, the temperature, the humidity, and how many times the surfaces have already slid against each other. Without that context, a friction coefficient is just a number floating in space. With it, the number tells you something you can actually use.