5: Einstein and Markov Equations for Brownian Motion- Historical Foundations and Modern Perspective

Fluctuations Near Equilibrium · LibreTexts

0%

Contents

    Text

    Size
    100%
    Theme
    Typeface
    Line length
    Spacing

    Read aloud

    How

    Speed: 1×
    Light up
    Highlight

    Symbols: Mulberry Symbols, © 2018–2026 Steve Lee, CC BY-SA 4.0 (licence, the set), the site's own copies.

    From LibreTexts, under CC BY-NC-SA 4.0, as the source states it; its equations are written in TeX, which the PDF shows typeset. This page at LibreTexts.

    5: Einstein and Markov Equations for Brownian Motion- Historical Foundations and Modern Perspective

    Einstein's 1905 theory of Brownian motion represents the first systematic application of a continuous Markov process in physics. Without invoking microscopic equations of motion, Einstein introduced a probabilistic evolution law for the position of a suspended particle, derived the diffusion equation, and connected the diffusion constant to thermodynamic quantities. This work provided decisive evidence for the molecular nature of matter and established the conceptual foundations of stochastic processes in physics. In this chapter we reconstruct Einstein's reasoning in modern notation, clarify its Markov structure, connect it to subsequent developments by Smoluchowski and Langevin, and discuss its relation to the Fokker-Planck equation and the fluctuation-dissipation relation.

    5.1 Introduction

    In 1905 Albert Einstein published his seminal paper On the movement of small particles suspended in a stationary liquid demanded by the molecular kinetic theory of heat ( Über die von der molekularkinetischen Theorie der Wärme geforderte Bewegung von in ruhenden Flüssigkeiten suspendierten Teilchen" [1]. In it, he provided a quantitative theory of the irregular motion observed by Robert Brown in 1827. The work played a decisive role in establishing the atomic theory of matter. Independently, Smoluchowski [2] developed a discrete random-walk picture

    Einstein used both thermodynamic and probabilistic arguments. Beyond its historical importance, Einstein's analysis introduced a new type of physical reasoning. Instead of writing Newton's equations for a particle bombarded by molecules, Einstein described the system probabilistically. He postulated an evolution equation for the probability density of the particle's position and derived macroscopic irreversibility from statistical independence of successive displacements. Independently, Smoluchowski [2] developed a discrete random-walk picture.

    Although Einstein did not use the term "Markov process," his construction satisfies precisely the defining properties of a continuous-time Markov process. In modern language, Brownian motion is a diffusion process whose generator is the Laplacian. Einstein's paper thus represents the first clear appearance of Markovian dynamics in physics[3].

    5.2 Part I: The Thermodynamic Derivation

    Einstein first used thermodynamic reasoning.

    Diffusion current:

    \[ J_{\mathrm{diff}}=-D \frac{\partial n}{\partial x} . \tag{5.1} \label{5.1} \]

    Drift current under force \(F\) :

    \[ J_{\mathrm{drift}}=\mu n F . \tag{5.2} \label{5.2} \]

    Here \(n\) is the particle density, and \(\mu\) the mobility. At equilibrium,

    \[ J_{\mathrm{diff}}+J_{\mathrm{drift}}=0 . \tag{5.3} \label{5.3} \]

    Using \(n \propto e^{-U / k T}\), Einstein obtained

    \[ D=\mu k T . \tag{5.4} \label{5.4} \]

    With Stokes' law,

    \[ \mu=\frac{1}{6 \pi \eta a}, \tag{5.5} \label{5.5} \]

    leading to

    \[ D=\frac{k T}{6 \pi \eta a} . \tag{5.6} \label{5.6} \]

    Equation \ref{5.4} is the first example of a fluctuation-dissipation relation.

    5.3 Part II: The Markovian Derivation

    Let \(P(x, t)\) denote the probability density of finding the Brownian particle at position \(x\) at time \(t\).

    Einstein considered a short time interval \(\tau\) and introduced \(\varphi(\Delta)\), the probability density for a displacement \(\Delta\) during that interval.

    His central assumption was:

    Displacements occurring in non-overlapping time intervals are statistically independent.

    Additionally:

    • The medium is homogeneous.
    • There is no systematic drift.
    • The displacement distribution is symmetric.

    Under these assumptions, the probability density evolves according to

    \[ P(x, t+\tau)=\int_{-\infty}^{\infty} P(x-\Delta, t) \varphi(\Delta) d \Delta \tag{5.7} \label{5.7} \]

    Equation \ref{5.7} is precisely the Chapman-Kolmogorov equation for a time-homogeneous Markov process.

    5.4 Derivation of the Diffusion Equation

    Assume \(\tau\) is small and that \(P(x, t)\) is smooth. Expand the left-hand side:

    \[ P(x, t+\tau)=P(x, t)+\tau \frac{\partial P}{\partial t}+O\left(\tau^{2}\right) . \tag{5.8} \label{5.8} \]

    Expand the integrand:

    \[ P(x-\Delta, t)=P(x, t)-\Delta \frac{\partial P}{\partial x}+\frac{\Delta^{2}}{2} \frac{\partial^{2} P}{\partial x^{2}}+\cdots . \tag{5.9} \label{5.9} \]

    Using normalization and symmetry,

    \[ \begin{align*} \int \varphi(\Delta) d \Delta & =1 \tag{5.10} \label{5.10}\\[4pt] \int \Delta \varphi(\Delta) d \Delta & =0 \tag{5.11}\\[4pt] \int \Delta^{2} \varphi(\Delta) d \Delta & =\left\langle\Delta^{2}\right\rangle \tag{5.12} \label{5.12} \end{align*} \]

    we obtain

    \[ \frac{\partial P}{\partial t}=\frac{\left\langle\Delta^{2}\right\rangle}{2 \tau} \frac{\partial^{2} P}{\partial x^{2}} . \tag{5.13} \label{5.13} \]

    Define

    \[ D=\frac{\left\langle\Delta^{2}\right\rangle}{2 \tau} . \tag{5.14} \label{5.14} \]

    The diffusion equation follows:

    \[ \frac{\partial P}{\partial t}=D \frac{\partial^{2} P}{\partial x^{2}} . \tag{5.15} \label{5.15} \]

    5.5 Solution and Experimental Prediction

    For initial condition \(P(x, 0)=\delta(x)\),

    \[ P(x, t)=\frac{1}{\sqrt{4 \pi D t}} \exp \left(-\frac{x^{2}}{4 D t}\right) . \tag{5.16} \label{5.16} \]

    The mean-square displacement is

    \[ \left\langle x^{2}(t)\right\rangle=2 D t . \tag{5.17} \label{5.17} \]

    Equation \ref{5.17} provided a direct experimental prediction. Jean Perrin later verified this relation and measured Avogadro's number [4].

    5.6 First Fluctuation-Dissipation Relation [4]

    The first fluctuation-dissipation theorem, famously known as the Einstein relation, connects the diffusion coefficient \(D\) (representing fluctuations) to the mobility \(\mu\) (representing dissipation) of a particle in thermal equilibrium, Equation \ref{5.4}

    \[ D=\mu k_{B} T \nonumber \]

    In terms of the velocity autocorrelation function, the relation is expressed via the Green-Kubo formula:

    \[ D=\int_{0}^{\infty}\left\langle u\left(t_{0}\right) u\left(t_{0}+t\right)\right\rangle d t \tag{5.18} \label{5.18} \]

    This demonstrates that the macroscopic transport properties of a system are determined by the correlation of its microscopic thermal fluctuations. We now derive the above equation.

    First recast equation Equation \ref{5.17} in the following form

    \[ D=\lim _{t \rightarrow \infty} \frac{1}{2 t}\left\langle[x(t)-x(0)]^{2}\right\rangle \tag{5.19} \label{5.19} \]

    where the average \(\rangle\) is taken over an ensemble in thermal equilibrium. Since we have

    \[ x(t)-x(0)=\int_{0}^{t} u\left(t^{\prime}\right) d t^{\prime} \nonumber \]

    equation Equation \ref{5.19} is transformed as

    \[ \begin{align*} D & =\lim _{t \rightarrow \infty} \frac{1}{2 t} \int_{0}^{t} d t_{1} \int_{0}^{t} d t_{2}\left\langle u\left(t_{1}\right) u\left(t_{2}\right)\right\rangle \\[4pt] & =\lim _{t \rightarrow \infty} \frac{1}{t} \int_{0}^{t} d t_{1} \int_{0}^{t-t_{1}} d t^{\prime}\left\langle u\left(t_{1}\right) u\left(t_{1}+t^{\prime}\right)\right\rangle \\[4pt] & =\int_{0}^{\infty}\left\langle u\left(t_{0}\right) u\left(t_{0}+t\right)\right\rangle d t \tag{5.20} \label{5.20} \end{align*} \]

    where we assumed that

    \[ \lim _{t \rightarrow \infty}\left\langle u\left(t_{0}\right) u\left(t_{0}+t\right)\right\rangle=0 . \tag{5.21} \label{5.21} \]

    Therefore the Einstein relation, Equation \ref{5.4} can be written as the first fluctuation-dissipation theorem:

    \[ \mu=\frac{1}{m \gamma}=\frac{D}{k T}=\frac{1}{k T} \int_{0}^{\infty}\left\langle u\left(t_{0}\right) u\left(t_{0}+t\right)\right\rangle d t \tag{5.22} \label{5.22} \]

    If a periodic external force

    \[ F=F_{0} \cos \omega t, \quad-\infty<t \nonumber \]

    is applied, the mobility is defined by

    \[ \langle u(t)\rangle=\operatorname{Re}\left[\mu(\omega) F_{0} e^{i \omega t}\right] \tag{5.23} \label{5.23} \]

    where \(\mu(\omega)\) is the complex mobility or the admittance. Now we have a frequency dependent FDT

    \[ \mu(\omega)=\frac{1}{k T} \int_{0}^{\infty}\left\langle u\left(t_{0}\right) u\left(t_{0}+t\right)\right\rangle e^{-i \omega t} d t \tag{5.24} \label{5.24} \]

    We will see that they are really Kubo formula [4] for linear response where the macroscopic response function is given by time correlation function of microscopic currents. Here these relations Eqs.( 5.22, 5.24), follow directly from the definitions. Remarkably, no perturbation theory was required.

    Marian Smoluchowski's Contributions to Brownian Motion

    While Albert Einstein is often the first name associated with the theory of Brownian motion (1905), Marian Smoluchowski [2] independently developed a nearly identical theory published in 1906. His approach was rooted in the kinetic theory of gases and collisions, providing a more "physically direct" microscopic mechanism compared to Einstein's thermodynamic derivation. The following discussion follows cosely Fulinski's article[6].

    CHAPTER 5. EINSTEIN AND MARKOV EQUATIONS FOR BROWNIAN MOTION:HISTORICAL FOUNDATION

    1. Kinetic-Probabilistic Derivation

    Unlike Einstein, who viewed Brownian motion through the lens of diffusion and osmotic pressure, Smoluchowski analyzed the individual collisions between the suspended particle and the molecules of the surrounding fluid.

    • He modeled the motion as a random walk.
    • He calculated the cumulative effect of thousands of unobservable molecular impacts, proving that a statistical imbalance in these impacts results in the observable jittery movement.
    • He arrived at a formula for the mean square displacement (MSD) that was virtually identical to Einstein's, differing only by a small numerical factor due to specific assumptions about collision geometry.

    2. The Smoluchowski Equation

    In 1916, he generalized the description of Brownian motion to include the effects of external force fields (such as gravity or electric fields). This resulted in the Smoluchowski equation, a specific form of the Fokker-Planck equation:

    \[ \frac{\partial P}{\partial t}=\nabla \cdot\left[D \nabla P-\frac{\mathbf{F}}{\gamma} P\right] \tag{5.25} \label{5.25} \]

    where \(P\) is the probability density, \(D\) is the diffusion coefficient, and \(\mathbf{F}\) is the external force. This remains a fundamental tool in soft matter physics and chemical kinetics.

    3. Verification of the Atomic Hypothesis

    Smoluchowski was a fierce defender of the atomistic view of matter. Along with Einstein and Jean Perrin, his work provided the theoretical "smoking gun" that atoms actually existed. By showing that macroscopic fluctuations (Brownian motion) were the direct result of microscopic "graininess" (atoms), he helped defeat the "Energisist" school of thought.

    4. Theory of Coagulation

    Smoluchowski applied his Brownian motion theories to the study of colloids. He developed the first mathematical description of how particles in a suspension collide and stick together-a process known as Smoluchowski coagulation. This is still the standard model for understanding the stability of aerosols and colloidal suspensions.

    Summary Comparison

    Feature Einstein (1905) Smoluchowski (1906)
    Primary Tool Thermodynamics & Diffusion Kinetic Theory & Random Walks
    Perspective Macroscopic/Statistical Microscopic/Statistical
    Key Result \(D=\frac{k T}{6 \pi \eta a}\) Nearly identical result via collisions
    Main Impact Founded the theory of fluctuations Linked diffusion to molecular scale

    5.7 The Langevin Equation

    In 1908 Langevin [7] introduced the "Equation of Motion" method, which is, Langevin's approach is notably different from Einstein's (1905) or Smoluchowski's (1906). While Einstein used a diffusion equation approach, Langevin introduced a stochastic differential equation,

    \[ m \frac{d^{2} x}{d t^{2}}=-6 \pi \eta a \frac{d x}{d t}+X \tag{5.26} \label{5.26} \]

    where \(X\) is the the stochastic force. This paper is the formal birth of the Langevin Equation. EquationEquation \ref{5.26} is generally recast in a more familiar form as

    \[ m \dot{v}=-\gamma v+\xi(t), \tag{5.27} \label{5.27} \]

    where \(\xi(t)\) is a rapidly fluctuating force. The second Fluctuation-Dissipation Theorem relates \(<\xi(t) \xi(0)>\) to the damping \(\gamma\) or the generalized memory function \(\gamma(t)\). We will discuss this in more detail in the following chapters, especially those on Noise, Magalinski's model and Mori-Zwanzig formulation of the generalized Langevin equation.

    5.8 Derivation of the Diffusion Equation result from the Langevin Equation

    The Langevin equation provides a microscopic description of Brownian motion. We derive the macroscopic diffusion equation by considering the stochastic evolution of a particle's position using the specific notation \(m \dot{v}=-\gamma v+\xi(t), \gamma\) is the friction coefficient and \(\xi(t)\) is the stochastic force (random noise).

    The random force \(\xi(t)\) is assumed to be Gaussian, white, noise with the properties:

    \[ \langle\xi(t)\rangle=0, \quad\left\langle\xi(t) \xi\left(t^{\prime}\right)\right\rangle=2 \gamma k_{B} T \delta\left(t-t^{\prime}\right) \tag{5.28} \label{5.28} \]

    For the Generalized Langevin Equation, GLE, the damping is time dependent and the equation above, Eq. 5.28 becomes

    \[ \langle\xi(t)\rangle=0, \quad\left\langle\xi(t) \xi\left(t^{\prime}\right)\right\rangle=2 \gamma k_{B} T \gamma\left(t-t^{\prime}\right) \tag{5.29} \label{5.29} \]

    Equation \ref{5.29} is called the second fluctuation-dissipation theorem[4]. It will be proven classically from classical mechanics in the chapter on Noise.

    5.9 The Over-damped Limit

    In the limit of high friction or long times \((t \gg m / \gamma)\), the inertial term \(m \dot{v}\) is neglected. Equation 5.27 simplifies to:

    \[ \gamma v(t)=\xi(t) \Longrightarrow \frac{d x}{d t}=\frac{1}{\gamma} \xi(t) \tag{5.30} \label{5.30} \]

    5.10 Statistical Moments

    For a small time interval \(\Delta t\), the change in position is:

    \[ \Delta x=\int_{t}^{t+\Delta t} \frac{1}{\gamma} \xi\left(t^{\prime}\right) d t^{\prime} \tag{5.31} \label{5.31} \]

    The first and second moments are:

    \[ \langle\Delta x\rangle=0 \tag{5.32} \label{5.32} \]

    \[ \left\langle(\Delta x)^{2}\right\rangle=\frac{1}{\gamma^{2}} \int_{t}^{t+\Delta t} \int_{t}^{t+\Delta t}\left\langle\xi\left(t^{\prime}\right) \xi\left(t^{\prime \prime}\right)\right\rangle d t^{\prime} d t^{\prime \prime}=\frac{2 \gamma k_{B} T}{\gamma^{2}} \Delta t=\frac{2 k_{B} T}{\gamma} \Delta t \tag{5.33} \label{5.33} \]

    Here we have used Eq. (5.28

    Defining the diffusion coefficient \(D=\frac{k_{B} T}{\gamma}\), we arrive at the diffusion equation result:

    \[ \left\langle(\Delta x)^{2}\right\rangle=D \Delta t \tag{5.34} \label{5.34} \]

    5.11 Modern Interpretation

    Equation \ref{5.7} defines a continuous-time Markov process. Taking the limit \(\tau \rightarrow 0\) yields the generator

    \[ \mathcal{L}=D \frac{\partial^{2}}{\partial x^{2}} . \tag{5.35} \label{5.35} \]

    The diffusion equation is the corresponding Fokker-Planck equation.

    More generally, expansion of Equation \ref{5.7} yields the Kramers-Moyal series. Einstein's derivation truncates at second order because the increment distribution has finite variance.

    Thus Brownian motion is the scaling limit of a random walk with independent increments, anticipating later work by Wiener and Kolmogorov.

    5.12 The Master Equation (Discrete States)

    For systems with discrete states (e.g., energy levels), we define \(P_{n}(t)\) as the probability of being in state \(n\). By defining transition rates \(w_{n m}\) (probability per unit time of \(m \rightarrow n\) ), we derive the "probability balance":

    \[ \frac{d P_{n}(t)}{d t}=\sum_{m \neq n}[\underbrace{w_{n m} P_{m}(t)}_{\text {Gain }}-\underbrace{w_{m n} P_{n}(t)}_{\text {Loss }} \tag{5.36} \label{5.36} \]

    5.13 Connecting the Two: The Kramers-Moyal Expansion

    To bridge the discrete Master Equation and the continuous Fokker-Planck Equation, we consider transitions \(w(x, z)\) where \(z\) is the jump size. The continuous Master Equation is:

    \[ \frac{\partial P(x, t)}{\partial t}=\int[w(x-z, z) P(x-z, t)-w(x,-z) P(x, t)] d z \nonumber \]

    Expanding the integrand \(w(x-z, z) P(x-z, t)\) as a Taylor series in \(z\) :

    \[ \frac{\partial P(x, t)}{\partial t}=\sum_{n=1}^{\infty} \frac{(-1)^{n}}{n!} \frac{\partial^{n}}{\partial x^{n}}\left[a_{n}(x) P(x, t)\right] \nonumber \]

    where the jump moments are \(a_{n}(x)=\int z^{n} w(x, z) d z\).

    5.14 The Fokker-Planck Equation

    Truncating the expansion at \(n=2\) (the Diffusion limit) yields the general framework for particles under external forces:

    \[ \frac{\partial P(x, t)}{\partial t}=-\frac{\partial}{\partial x}[A(x, t) P(x, t)]+\frac{\partial^{2}}{\partial x^{2}}[B(x, t) P(x, t)] \tag{5.37} \label{5.37} \]

    • Drift Term \(A(x, t)\) : Corresponds to \(a_{1}(x)\), the average velocity due to external forces.
    • Diffusion Term \(B(x, t)\) : Corresponds to \(a_{2}(x) / 2\), the stochastic spreading.

    5.15 Conceptual Significance

    Einstein's theory achieved several conceptual breakthroughs:

    • Introduction of probabilistic evolution laws in physics.
    • Derivation of irreversible macroscopic behavior from statistical independence.
    • First quantitative link between fluctuation strength and dissipative transport coefficient.
    • Establishment of diffusion as a universal scaling limit.

    The modern theory of stochastic processes, linear response, and nonequilibrium statistical mechanics can be traced to this 1905 paper.

    5.16 Conclusion

    Einstein's treatment of Brownian motion marks the birth of Markovian stochastic dynamics in physics. From a simple assumption of independent increments, he derived the diffusion equation and connected fluctuations to dissipation through thermodynamic reasoning.

    This synthesis of probability theory and physics laid the foundation for stochastic differential equations, Fokker-Planck theory, and the fluctuation-dissipation framework that underlies modern statistical mechanics.

    Bibliography

    [1] A. Einstein, Ann. Phys. 17, 549 (1905).

    [2] M. von Smoluchowski, Ann. Phys. 21, 756 (1906).

    [3] Dattagupta, Sushanta. "The myth about Einstein." Resonance 11.1 (2006): 63-78.

    [4] J. Perrin, Les Atomes (Alcan, Paris, 1909)

    [5] R. Kubo, Rep. Prog. Phys. 29, 255 (1966).

    [6] A. Fulinski, On Marian Smoluchowski's Life and Contribution to Physics Acta Physica Polonica B, vol. 29, Issue 6, p. 1523 (1998)

    [7] P. Langevin, C. R. Acad. Sci. (Paris) 146, 530 (1908).

    [8] C. W. Gardiner, Handbook of Stochastic Methods (Springer, 1985).


    This page titled 5: Einstein and Markov Equations for Brownian Motion- Historical Foundations and Modern Perspective was last modified on Mon, 02 Nov 2026 06:09:25 GMT and is shared under a CC BY-NC-SA 4.0 license and was authored, remixed, and/or curated by .