5.5 Brownian motion

Brownian motion is the random movement of particles suspended in a fluid, caused by collisions with the surrounding molecules. It bridges microscopic particle behavior with macroscopic phenomena like diffusion, providing a direct window into thermal equilibrium and the statistical behavior of many-particle systems.

Named after botanist Robert Brown and explained theoretically by Einstein in 1905, Brownian motion is modeled as a continuous-time stochastic process. Particle displacements follow Gaussian probability distributions, and the framework connects deeply to core ideas in statistical mechanics: equipartition, ergodicity, and the fluctuation-dissipation theorem.

Definition of Brownian motion

Brownian motion refers to the irregular, ceaseless motion of small particles suspended in a fluid. The motion arises not from anything intrinsic to the particle, but from the random thermal motion of the fluid molecules that constantly bombard it. This makes it a direct, observable consequence of molecular kinetic theory.

What makes Brownian motion so important in statistical mechanics is that it connects the invisible world of molecular collisions to something you can actually see under a microscope. It gives concrete, measurable evidence for thermal equilibrium and the statistical behavior of large ensembles of molecules.

Historical background

• Robert Brown (1827) observed pollen grains jittering erratically in water under a microscope. He initially suspected a biological "vital force," but then found that inorganic particles behaved the same way.
• Einstein (1905) provided the first rigorous theoretical explanation, linking the observable motion to the existence and behavior of atoms and molecules. This was one of the key pieces of evidence that atoms are real physical entities, not just a convenient model.
• Jean Perrin (early 1900s) experimentally verified Einstein's predictions using colloidal suspensions, earning the Nobel Prize and providing some of the strongest evidence for atomic theory.

Mathematical description

Brownian motion is modeled as a continuous-time stochastic process with independent increments, meaning the displacement over any time interval is statistically independent of what happened before. The displacement over a time interval $\Delta t$ follows a Gaussian (normal) distribution centered at zero.

The formal mathematical model is the Wiener process $W(t)$, which has three defining properties:

• $W(0) = 0$
• Increments $W(t) - W(s)$ are normally distributed with mean 0 and variance $t - s$
• Increments over non-overlapping time intervals are independent

Physical interpretation

Each fluid molecule that strikes the suspended particle imparts a tiny, random impulse. Because the particle is being hit from all sides simultaneously by enormous numbers of molecules, the net force fluctuates randomly in both magnitude and direction.

• The equipartition theorem applies: the average kinetic energy of the Brownian particle in each degree of freedom is $\frac{1}{2}k_BT$, just like the surrounding fluid molecules.
• Ergodicity holds for Brownian systems: if you track a single particle for a long time, its time-averaged properties converge to the ensemble average over many particles. This is what lets you extract thermodynamic quantities from single-particle tracking experiments.

Microscopic origins

Brownian motion emerges from the collective effect of countless molecular collisions. No single collision produces a visible displacement. Instead, the tiny, random pushes accumulate statistically, producing the characteristic jittery trajectory you observe.

This is a powerful example of how random microscopic events give rise to predictable statistical behavior at macroscopic scales.

Collision with particles

The suspended particle experiences an enormous number of collisions per second from surrounding fluid molecules. Each collision imparts a small, random change in the particle's velocity and direction. The frequency and strength of these collisions depend on:

• Temperature: higher $T$ means faster-moving fluid molecules and stronger kicks
• Fluid viscosity: more viscous fluids slow the particle's response
• Particle size: smaller particles are more strongly affected by individual collisions

Random walk model

A useful discrete approximation of Brownian motion is the random walk: the particle takes steps of fixed size in random directions at regular time intervals.

• The step size corresponds roughly to the mean free path between effective collisions.
• After $N$ steps of size $\ell$, the mean square displacement is $\langle r^2 \rangle = N\ell^2$.
• The Central Limit Theorem explains why the sum of many small random displacements converges to a Gaussian distribution, regardless of the distribution of individual steps. This is why Brownian motion has Gaussian statistics.

Langevin equation

The Langevin equation describes the motion of a Brownian particle by applying Newton's second law with two additional forces: viscous drag and a random (stochastic) force.

$m\frac{d^2x}{dt^2} = -\gamma\frac{dx}{dt} + F(t)$

• $\gamma$ is the drag coefficient, representing the viscous resistance of the fluid.
• $F(t)$ is the random force (also called the Langevin force), representing the net effect of molecular collisions. It has zero mean ($\langle F(t) \rangle = 0$) and is delta-correlated in time: $\langle F(t)F(t') \rangle = 2\gamma k_B T \, \delta(t - t')$.

The delta-correlation condition ensures consistency with the fluctuation-dissipation theorem: the strength of the random force is directly tied to the drag coefficient and temperature.

In the overdamped limit (where inertia is negligible compared to drag, typical for colloidal particles), the $m\ddot{x}$ term drops out, and you get the simpler first-order equation: $\gamma\frac{dx}{dt} = F(t)$.

Statistical properties

The statistical properties of Brownian motion allow you to make quantitative predictions about particle behavior and compare them with experiment. These properties form the foundation for understanding diffusion and thermal fluctuations.

Mean square displacement

The mean square displacement (MSD) measures how far a particle has wandered, on average, after time $t$. For normal (Fickian) diffusion in one dimension:

$\langle x^2(t) \rangle = 2Dt$

In three dimensions, this becomes $\langle r^2(t) \rangle = 6Dt$. The linear scaling with time is the hallmark of normal diffusion. If you plot MSD vs. time and get a straight line, you're dealing with standard Brownian motion.

The MSD is the primary experimental observable for characterizing diffusion. By measuring particle trajectories and computing the MSD, you can extract the diffusion coefficient $D$.

Diffusion coefficient

The diffusion coefficient $D$ quantifies how rapidly particles spread out. It depends on:

• Temperature (higher $T$ means faster diffusion)
• Fluid viscosity (higher viscosity slows diffusion)
• Particle size (larger particles diffuse more slowly)

These dependencies are captured quantitatively by the Stokes-Einstein equation (discussed below). Experimentally, $D$ can be measured from particle tracking data or from the broadening of concentration profiles over time.

Probability distribution

The probability of finding a Brownian particle at position $x$ at time $t$, given it started at the origin, follows a Gaussian:

$P(x, t) = \frac{1}{\sqrt{4\pi Dt}} \exp\left(-\frac{x^2}{4Dt}\right)$

This distribution broadens over time as $\sqrt{t}$, consistent with the MSD growing linearly in $t$. In the continuum limit, this probability density satisfies the diffusion equation.

Einstein's theory

Einstein's 1905 theory was groundbreaking because it connected the microscopically random motion of suspended particles to measurable macroscopic quantities like the diffusion coefficient and Avogadro's number. It provided a concrete, testable prediction that could confirm the atomic hypothesis.

Diffusion equation

The diffusion equation governs how the concentration $c(\mathbf{r}, t)$ of Brownian particles evolves in space and time:

$\frac{\partial c}{\partial t} = D\nabla^2 c$

Solutions to this equation give the probability distribution for particle positions. For a point source at the origin, the solution is the Gaussian distribution described above. The diffusion equation can also be derived from the random walk model in the continuum limit.

Einstein-Smoluchowski relation

This relation connects the diffusion coefficient to the mobility $\mu$ of a particle (how easily it moves in response to an applied force):

$D = \mu k_B T$

Here, $\mu = 1/\gamma$ where $\gamma$ is the drag coefficient. This equation is a specific instance of the fluctuation-dissipation theorem: the same thermal energy that drives random fluctuations (diffusion) also determines the particle's response to systematic forces (mobility). It directly links the random and deterministic aspects of particle motion.

Stokes-Einstein equation

Combining the Einstein-Smoluchowski relation with Stokes' law for the drag on a sphere ($\gamma = 6\pi\eta r$) gives:

$D = \frac{k_B T}{6\pi\eta r}$

• $\eta$ is the dynamic viscosity of the fluid
• $r$ is the hydrodynamic radius of the particle

This equation is widely used in practice. For example, in dynamic light scattering, you measure $D$ from the fluctuations of scattered light and then use the Stokes-Einstein equation to calculate the particle radius. It assumes spherical particles in a continuum fluid, so it breaks down for particles comparable in size to the solvent molecules.

Experimental observations

Experiments on Brownian motion have played a central role in establishing the atomic theory of matter and continue to be important in modern physics and biophysics.

Robert Brown's discovery

In 1827, Brown observed pollen grains suspended in water jittering irregularly under a microscope. He initially thought the motion might be a sign of life, but when he repeated the experiment with ground-up minerals and other inorganic particles, the same erratic motion appeared. This ruled out any biological explanation and pointed toward a physical cause rooted in the fluid itself.

Jean Perrin's experiments

Perrin conducted careful, systematic measurements of Brownian motion in the early 1900s using colloidal suspensions. He tracked individual particle positions over time, measured the MSD, and confirmed that it scaled linearly with time as Einstein predicted. From his data, he extracted a value for Avogadro's number that agreed with other independent measurements, providing powerful evidence for the reality of atoms.

Modern microscopy techniques

• Single-particle tracking with advanced optical microscopy can follow individual nanoparticles with nanometer precision, enabling detailed studies of diffusion in complex environments.
• Fluorescence microscopy allows observation of Brownian motion of labeled molecules in living cells.
• Atomic force microscopy detects the Brownian motion of microcantilevers, which is used in ultrasensitive force and mass measurements.

Applications in physics

The concepts underlying Brownian motion extend far beyond suspended particles. The same mathematical framework applies wherever thermal fluctuations and dissipation coexist.

Diffusion processes

Diffusion describes the spread of particles, heat, or other quantities driven by random thermal motion. Examples include:

• Mixing of gases or solutes in solution
• Heat conduction in solids
• Dopant diffusion in semiconductor fabrication

All of these are governed by the same diffusion equation, with different physical interpretations of $c$ and $D$.

Thermal fluctuations

At the nanoscale, thermal fluctuations become significant relative to the energies involved in mechanical and biological processes. These fluctuations set fundamental limits on the precision of nanoscale measurements and devices. They also drive the operation of biological molecular machines like kinesin and ATP synthase, which harness thermal energy to perform directed work.

Noise in electrical circuits

Johnson-Nyquist noise is the electrical analog of Brownian motion: thermal motion of charge carriers in a resistor produces voltage fluctuations. The mean square noise voltage across a resistor $R$ in bandwidth $\Delta f$ is:

$\langle V^2 \rangle = 4k_B T R \, \Delta f$

This sets a fundamental lower limit on the sensitivity of electronic instruments and can itself be used as a primary thermometer.

Brownian motion in biology

Brownian motion is not just a physics curiosity; it plays a functional role in biological systems at the cellular and molecular scale.

Cellular transport

Inside cells, many small molecules and even some organelles move primarily by Brownian diffusion rather than active transport. Diffusion distributes nutrients, waste products, and signaling molecules throughout the cytoplasm. For small molecules over short distances (micrometers), diffusion is fast enough to be effective, but it becomes too slow over longer distances, which is why cells also use motor proteins and active transport.

Protein dynamics

Thermal fluctuations drive the conformational changes that proteins undergo during folding, enzyme catalysis, and molecular recognition. Enzyme-substrate binding, for instance, relies on Brownian encounters between the enzyme and substrate molecules. The rates of these processes are directly influenced by the diffusion coefficients of the molecules involved.

Membrane diffusion

Lipids and proteins in cell membranes undergo lateral Brownian diffusion within the two-dimensional plane of the membrane. This diffusion governs the formation of lipid rafts, the clustering of receptors, and the kinetics of membrane-bound signaling reactions. Membrane diffusion coefficients are typically much smaller than in bulk solution due to the crowded, viscous membrane environment.

Mathematical models

Several mathematical frameworks formalize Brownian motion at different levels of description. These tools are essential for both analytical calculations and numerical simulations.

Wiener process

The Wiener process $W(t)$ is the standard mathematical model for Brownian motion. Its key properties are:

• $W(0) = 0$
• Continuous sample paths (the trajectory is continuous but nowhere differentiable)
• Independent, Gaussian increments with variance equal to the time interval

It serves as the fundamental building block for more complex stochastic models in physics, finance, and engineering.

Fokker-Planck equation

The Fokker-Planck equation describes the time evolution of the probability density $P(x, t)$ for a stochastic variable:

$\frac{\partial P}{\partial t} = -\frac{\partial}{\partial x}[A(x)P] + \frac{\partial^2}{\partial x^2}[B(x)P]$

Here, $A(x)$ is the drift term (systematic forces) and $B(x)$ is the diffusion term (random forces). For pure Brownian motion with no drift, $A = 0$ and $B = D$, recovering the standard diffusion equation. The Fokker-Planck equation is the natural complement to the Langevin equation: the Langevin equation describes individual trajectories, while the Fokker-Planck equation describes the evolution of the probability distribution.

Itô calculus

Standard calculus doesn't work for functions of Brownian motion because the Wiener process is nowhere differentiable. Itô calculus provides the rules for manipulating stochastic differential equations (SDEs). The key difference from ordinary calculus is Itô's lemma, which includes an extra second-order term due to the nonzero variance of $dW$:

$df = \frac{\partial f}{\partial t}dt + \frac{\partial f}{\partial x}dX + \frac{1}{2}\frac{\partial^2 f}{\partial x^2}(dX)^2$

where $(dW)^2 = dt$. Itô calculus is used extensively in statistical physics and in financial mathematics for modeling asset prices.

Brownian motion vs deterministic motion

Comparing Brownian motion with deterministic dynamics highlights several deep ideas in statistical mechanics.

Time reversibility

The equations governing Brownian motion are statistically time-reversible at equilibrium: the statistical properties of the trajectory look the same whether you play the movie forward or backward. However, any individual trajectory is not reversible in practice because the detailed information about molecular collisions is lost. This contrasts with deterministic classical mechanics, where the equations of motion are exactly time-reversible for each trajectory.

Ergodicity

Brownian systems are typically ergodic: the time average of any observable, computed from a single long trajectory, equals the ensemble average over many independent realizations. This is what justifies using statistical mechanics to describe Brownian motion. You can measure the diffusion coefficient from one particle tracked for a long time, and it will agree with the value obtained from tracking many particles simultaneously.

Ergodicity can break down in systems with anomalous diffusion or in glassy, disordered environments.

Fluctuation-dissipation theorem

The fluctuation-dissipation theorem (FDT) states that the spontaneous fluctuations of a system in thermal equilibrium are quantitatively related to its response to small external perturbations. For Brownian motion, this manifests as the connection between the random force and the drag coefficient in the Langevin equation:

$\langle F(t)F(t') \rangle = 2\gamma k_B T \, \delta(t - t')$

The same friction that dissipates energy also drives the fluctuations. The FDT applies broadly to systems near equilibrium and is one of the most important results in statistical mechanics.

Advanced concepts

These extensions of standard Brownian motion address situations where the simple model breaks down or where additional physics is at play.

Anomalous diffusion

In anomalous diffusion, the MSD does not scale linearly with time. Instead:

$\langle r^2(t) \rangle \propto t^\alpha$

• $\alpha < 1$: subdiffusion (slower than normal). Occurs in crowded cellular environments, porous media, and viscoelastic fluids.
• $\alpha > 1$: superdiffusion (faster than normal). Can arise from active transport, Lévy flights, or turbulent flows.

Identifying the exponent $\alpha$ from experimental MSD data is a standard way to characterize transport in complex systems.

Fractional Brownian motion

Fractional Brownian motion (fBm) generalizes the standard Wiener process by introducing long-range correlations between increments. It is characterized by the Hurst exponent $H$:

• $H = 0.5$: standard Brownian motion (no correlations)
• $H > 0.5$: persistent motion (positive correlations; the particle tends to continue in the same direction)
• $H < 0.5$: anti-persistent motion (negative correlations; the particle tends to reverse direction)

fBm is used to model phenomena with long-term memory or self-similar statistical structure.

Active Brownian motion

Active Brownian particles consume energy from their environment to generate self-propulsion, in addition to experiencing thermal fluctuations. Examples include swimming bacteria, synthetic catalytic nanomotors, and motile cells.

The dynamics combine a directed velocity with rotational diffusion, so the particle moves ballistically at short times but diffuses at long times with an enhanced effective diffusion coefficient. Active Brownian motion is a central topic in the growing field of active matter physics, where systems are driven far from equilibrium.