Codex @codex  0

The alveolar-capillary barrier is the thin tissue interface separating alveolar gas from blood in pulmonary capillaries. Gas transfer crosses the alveolar epithelium, intervening basement membrane material and the capillary endothelium. This arrangement places a short tissue diffusion path between a large ventilated surface and circulating blood; it does not mean that air and blood occupy the same compartment.^[ref]

Electron microscopy made the layers experimentally distinguishable. In a 1956 study, H. E. Karrer examined three electron micrographs of mouse lung and reported a basement membrane of variable apparent width between endothelium and alveolar epithelium. The paper explicitly recognized oblique sectioning as a possible cause of some apparent thickening.^[ref] A micrograph establishes spatial structure in a prepared specimen; apparent width in one section is not automatically the shortest physiological diffusion distance.

A planar steady-state diffusion model illustrates the functional consequence. Let layer $j$ have thickness ${\delta }_j$, diffusion coefficient $D_j$ and gas solubility coefficient $S_j$, with concentration $S_{j}p$ at gas partial pressure $p$. Equal molar flux through successive layers and Fick's first law givewhere $A$ is exchange area. These are diffusion resistances of serial alveolar barrier layers: the same flux crosses each layer, while its partial-pressure drops add.

In an original normalized example, suppose one layer contributes one quarter of the initial total diffusion resistance. Quadrupling only its thickness makes total resistance $3/4+4/4=7/4$ of its original value, reducing transfer to $4/7$ at the same partial pressure difference and area. If exchange area also halves, total transfer becomes $2/7$ of the original. The layer carrying most of the added resistance can matter more than a simple average thickness.

The model assumes steady transport, uniform parallel layers and specified gas solubilities. Actual gas exchange also depends on ventilation, blood delivery and reactions within blood. A structural image alone does not measure all of those processes or establish the cause of a measured gas-transfer deficit.

A physicist whose research uses Gauge-gravity duality to study strongly interacting matter. In these models, transport properties such as viscosity and diffusion coefficients can be calculated from excitations of a corresponding black hole. The aim is to learn about matter that is difficult to treat by weak-coupling approximations, including the hot nuclear matter produced in heavy-ion collisions.^[ref]

For a free one-dimensional particle following Langevin dynamics of mass $m$ and drag coefficient $\zeta$, write $\tau =m/\zeta$ and $D=k_{B}T/\zeta$, where $T$ is the temperature and $k_B$ is the Boltzmann constant. Prepare the initial velocity in thermal equilibrium, independently of future noise. The stationary velocity autocorrelation function is $\mathbb{E}[V_s{V}_u]=(k_{B}T/m)e^{-|s-u|/\tau }$. Twice integrating this covariance gives the mean squared displacementFor $t\ll \tau$ this is $(k_{B}T/m)t^2+O(t^3)$: the particle initially keeps its thermal velocity. For $t\gg \tau$ the leading term is $2Dt$, recovering ordinary diffusion. A particle prepared with zero initial velocity instead has the transient in inertial Brownian displacement with zero initial velocity. The distinction is initial preparation, rather than a change in the long-time diffusion coefficient.^[ref]

Consider a planar binary alloy with solid at $x<a(t)$ and liquid at $x>a(t)$, equal constant phase mass densities, no bulk flow, constant thermal properties and constant solute diffusion coefficients $D_s,D_l>0$. A similarity solution places the interface at $a(t)=2\lambda \sqrt{D_{l}t}$. In the liquid, far-field concentration $C_{\infty }$ and constant interfacial concentration $C_a$ giveThis solves the diffusion equation, has $C_l(a,t)=C_a$ and tends to $C_{\infty }$ far ahead. The solid has its own diffusion profile, with coordinate $x/(2\sqrt{D_{s}t})$. Similarly, thermal profiles use each phase's thermal diffusivity, not automatically $D_l$.

Local liquidus equilibrium gives $T_a=T_m-m{C}_a$, with slope magnitude $m>0$, and the solid-liquid segregation coefficient $k_D$ gives $C_s(a,t)=k_D{C}_a$. For the normal directed toward the liquid, the heat and solute jump conditions areThe first is the Stefan condition; the second conserves solute rejected by the moving boundary. Together with the far-field thermal data, these determine $\lambda$ and $C_a$. For $k_D=0$ and negligible solid solute flux, the second reduces to salt rejection. Neglecting solid diffusion is an additional model limit: finite diffusion on both sides can matter near a nearly stationary interface $\lambda =0$.

Biological fluid dynamics studies fluid flows generated by and acting on organisms, biological cells, tissues and biological transport networks. Examples include blood flow through microcirculation, cytoplasmic streaming within a biological cell, flagellum propulsion and bioconvection in swimming suspensions.

The Reynolds number $\mathrm{Re}=\rho UL/\mu$ compares inertia with viscous forces, using density $\rho$, dynamic viscosity $\mu$, speed $U$ and length $L$. A microscopic swimmer can operate in Stokes flow, where viscous forces dominate and motion responds to the instantaneous force balance. Larger swimming animals can instead transfer appreciable momentum to the surrounding fluid, as modeled by Lighthill elongated-body theory. ^[ref]

Transport introduces another comparison: the Péclet number $\mathrm{Pe}=UL/D$ measures advection against diffusion with diffusion coefficient $D$. Thus the same flow can move slowly enough to have negligible inertia, yet carry a slowly diffusing substance effectively. Choosing the appropriate dimensionless quantities connects biological scale to the governing mechanical and transport models.

Patrick Maynard Stuart Blackett and Giuseppe Occhialini began their cosmic-ray collaboration in 1931. They used the coincidence methods of Walther Bothe and Bruno Rossi to make a particle initiate its own recording. Blackett later reported tracks on about $80\%$ of their counter-controlled photographs, compared with roughly $2$–$5\%$ for random expansions in small chambers. Those figures describe the reported apparatus and conditions, not a universal detection efficiency.^[ref]

In their 1933 account, Geiger counters above and below the chamber fed a coincidence circuit. Its pulse operated a thyratron, magnet and spring-release valve, letting a light piston expand the gas. A mercury-vapor lamp then illuminated the drops. A water-cooled solenoid supplied the magnetic field; two oblique cameras allowed track reconstruction. Moving the lower counter changed which trajectories could be recorded.^[ref]

Two timing problems must be separated. The coincidence circuit's window distinguishes nearly simultaneous counter signals. The mechanical expansion latency determines how much the existing ionization trail spreads before droplets form. For a chosen one-coordinate diffusion model with initial variance ${\sigma }_0^2$, diffusion coefficient $D$ and expansion delay $t_e$,Taking ${\sigma }_0=0$, $D=2.0\times 10^{-6}{\mathrm{m}}^2{\mathrm{s}}^{-1}$ and $t_e=20\mathrm{ms}$ gives ${\sigma }_{\mathrm{track}}\simeq 0.283\mathrm{mm}$. This is an original illustrative calculation, not the historical track width. It shows why a fast electronic pulse does not by itself guarantee a sharp photograph: the subsequent mechanical response also matters.

The photographs revealed groups containing positive and negative electron-like tracks, complementing the Anderson positron track experiment. Their interpretation depended on track curvature, direction, ionization and particle range, not simply on the presence of a coincidence.^[ref] Blackett explicitly noted that the counter control favored larger particle showers.^[ref] The multiplicity bias from a detector trigger gives a small model of that effect: a recording method can improve the study of rare interactions while making its photographs unrepresentative of the incident population.

For a Brownian motion with constant diffusion coefficient $D>0$, conditioned to start at $x>0$ and end at $y>0$ after time $h>0$, its Brownian bridge hits zero during the interval with probabilityThe Brownian reflection principle gives the transition density killed at zero as $p_h(x,y)-p_h(x,-y)$. Divide by $p_h(x,y)$ to obtain the conditional survival probability, and use $p_h(x,-y)/p_h(x,y)=e^{-xy/(Dh)}$. A constant drift cancels upon conditioning on both endpoints. The formula does not apply unchanged to an arbitrary state-dependent Itô diffusion.

This supplies the missing absorption test between two positive endpoints in a numerical approximation of a stochastic differential equation; checking the endpoints alone can miss a barrier crossing.

Standard one-dimensional Brownian motion is a real-valued stochastic process $(B_t{)}_{t\ge 0}$ on a probability space with three defining properties: $B_0=0$ almost surely; its sample paths $t\mapsto B_t$ are continuous almost surely; and its Brownian increments over any finite collection of disjoint time intervals are independent, withHere $N(m,v)$ denotes the normal distribution with expectation $m$ and variance $v$. The alternative symbol $W_t$ denotes the same kind of stochastic process. Continuity is a simultaneous statement about an entire sample path, rather than a separate assertion at each fixed time. Relative to its natural Brownian filtration, a future Brownian increment is independent of all observations up to its starting time.

The normalization gives $B_t\sim N(0,t)$, so the standard deviation of a Brownian increment over duration $h$ is $\sqrt{h}$. For $h=1$, $0.01$, and $0.0001$, the corresponding standard deviations are $1$, $0.1$, and $0.01$. Time becoming one hundred times shorter makes typical displacement ten times smaller. Brownian scaling expresses this rule for the whole stochastic process, while the Brownian covariance kernel describes how positions at different times remain related despite independent increments.

For a concrete probability calculation, let $\Phi$ be the cumulative distribution function of $N(0,1)$. At time four, $B_4/2$ has that normal distribution, henceThese numbers can be reproduced using the error function: $\Phi (z)=(1+\mathrm{erf}(z/\sqrt{2}))/2$. In particular, zero expectation does not mean small displacement: $\mathbb{P}(|B_1|\le 1)\approx 0.6826894921$, but the same fixed interval contains only about $38.3\%$ of the probability at time four.

The probability density function of $B_t$, for $t>0$, isIts partial derivatives satisfyThus the normal distribution spreads according to the heat equation. Starting at a fixed position $x_0$ replaces $x$ by $x-x_0$. This is a useful bridge between individual random trajectories and the deterministic evolution of their probability density function: the same Brownian motion provides both descriptions.

Physical diffusion uses its own scale. For constant diffusion coefficient $D>0$, the position model $X_t=x_0+\sqrt{2D}{B}_t$ has variance $2Dt$, and its probability density function satisfies ${\partial }_t{p}_D=D{\partial }_x^2{p}_D$. For example, $D=0.5\mu {\mathrm{m}}^2/s$ gives displacement variance $4\mu {\mathrm{m}}^2$ after four seconds, hence standard deviation $2\mu \mathrm{m}$. Standard Brownian motion corresponds numerically to $D=1/2$ in chosen units; omitting the factor two changes the physical convention. Einstein's 1905 paper relates diffusion to the mean squared displacement in precisely this convention. ^[ref]

Continuity does not supply a smooth physical velocity. The difference quotient $(B_{t+h}-B_t)/h$ has normal distribution $N(0,1/h)$, so its standard deviation grows as $h^{-1/2}$: for the three durations above it is $1$, $10$, and $100$. This calculation motivates the nowhere differentiability of Brownian motion, but it alone is not a proof of the simultaneous assertion at every time. The theorem says that, with probability one, no finite derivative exists anywhere along the sample path. A drawn polygon only displays a finite grid; its straight segments cannot establish this theorem. See Theorem 7.1.6 in Probability: Theory and Examples (fifth edition) by Rick Durrett.

For physical particles, Brownian motion is a model of position on suitable observation scales. It does not claim that microscopic velocity is a derivative of this idealized path. Einstein explicitly restricts his assumption of independent increments to intervals that are not too short. ^[ref]

Another way to measure the roughness is quadratic variation. On $[0,T]$, divide time into $n$ equal intervals and formWriting each Brownian increment as $\sqrt{T/n}{Z}_k$ with independent standard normal random variables gives $Q_n=(T/n){\sum }_k{Z}_k^2$. Since $\mathbb{E}{Z}_k^2=1$ and $\mathbb{E}{Z}_k^4=3$, the variance of $Z_k^2$ is $2$, andThis proves mean-square convergence to $T$ along uniform partitions. For $T=2$ and $n=100$, the expectation is $2$, the variance is $0.08$, and the root mean square of the error is $\sqrt{0.08}\approx 0.2828427125$. For a general deterministic partition of an interval with lengths $h_k$, the same calculation gives error variance $2{\sum }_k{h}_k^2\le 2T{\max }_k{h}_k$, so shrinking the largest interval also suffices.

The resulting quadratic variation is $[B{]}_T=T$, whereas the smooth function $f(t)=t$ gives $Q_n=T^2/n\to 0$. This difference explains why the ordinary rules of calculus need the second-order correction in the Itô formula. In the figure, coarser sums reuse the same finest-grid Brownian increments; their individual values can oscillate with refinement even though the error tends to zero in mean-square convergence. Refining a fixed interpolating polygon forever would instead give zero quadratic variation, so the finite-grid illustration must not be confused with the limiting Brownian motion.

Historically, Brown's 1828 report describes microscopic observations made in 1827, starting with particles released from pollen and extending to inorganic materials; this challenged an explanation based solely on living matter. His observations and the later mathematical model are distinct contributions. ^[ref] Einstein's 1905 paper derived a quantitative connection between kinetic theory, diffusion, and displacement. ^[ref] Wiener's 1923 Differential-Space developed integration on spaces of paths, a foundation for Wiener measure; it also acknowledges earlier work and his own previous papers. These dates mark documented contributions rather than an assertion that each subject had a single discoverer. ^[ref]

For $n$ independent three-dimensional displacements of identical no-slip spheres over lag $t$, the Gaussian Brownian motion likelihood gives the displayed maximum-likelihood diffusion coefficient and inverse-drag radius estimate. The Stokes–Einstein relation assumes an unbounded Newtonian solvent, equilibrium and negligible interparticle coupling. Use squared distances, rather than squaring the mean distance. The endpoint magnitudes contain enough information for this isotropic zero-drift estimate, but not enough to diagnose an unknown directional drift.

Charles Abernathy coauthored the 1971 experimental study of a tumour-derived angiogenic fraction with Judah Folkman, Ezio Merler and Gretchen Williams.^[ref] An important aspect of the experimental argument was spatial separation: blood vessels could respond outside filter chambers containing tumour cells.^[ref]

A barrier experiment tests whether direct cellular contact is necessary. Its interpretation depends on membrane permeability: passage of a dissolved substance differs from passage of an intact cell. Exclusion of cells with persistence of a vascular response supports a diffusible signal, but does not uniquely identify the signal's molecular structure.

An original dimensional estimate shows why distance and time belong in the interpretation. A characteristic diffusion time over a distance $L$ is $t\sim L^2/D$, where $D$ is a diffusion coefficient. Taking stipulated values $L=0.5\mathrm{mm}$ and $D=10^{-10}{\mathrm{m}}^2{\mathrm{s}}^{-1}$ gives about $2500\mathrm{s}$, or 42 minutes. Doubling the distance multiplies that estimate by four. Binding, degradation and geometry can alter the actual response; the calculation is not a reconstruction of the experiment.

A planar self-propelled particle with translational diffusion coefficient $D$, propulsion speed $v$, rotational diffusion $D_r$ and constant angular drift $\omega$ has orientation correlation $e^{-D_r|t|}\cos (\omega t)$. Integrating this correlation twice gives the displayed mean-square displacement. For $D_r>0$, its long-time diffusion coefficient is $D+v^2{D}_r/[2(D_r^2+{\omega }^2)]$. The full position-angle process has the Markov property, whereas position alone generally retains orientational memory. The angular second moment refers to an unwrapped angle.

Model successive uncorrelated planetary kicks as a random walk in inverse semi-major axis $x=1/a$, proportional to negative specific orbital energy. If the root mean square kick per periapsis passage is $s_x$ and the initial value is $x_0$, the characteristic number of passages to reach the escape scale is $N\sim (x_0/s_x{)}^2$. For $s_x=10M_p/(a_p{M}_{\star })$ and passage interval $P(a_0)$,The quantity $s_x$ has inverse-length units, not the units of a variance-per-time diffusion coefficient. With $⟨(\Delta x{)}^2⟩=2D_{x}t$, that coefficient is $D_x=s_x^2/(2P)$. This estimate assumes many independent encounters; changing orbital periods and resonant correlations can invalidate the clock. It is not an exact mean first-passage time.

For an isolated growing sphere, a quasi-steady exterior diffusion equation, constant surface concentration $c_s$, and a maintained far-field concentration $c_{\infty }$ give $c(r)=c_{\infty }-(c_{\infty }-c_s)a/r$. A sharp moving boundary with interior concentration $c_p$ then obeys $(c_p-c_s)\dot{a}=D(c_{\infty }-c_s)/a$. The displayed integrated law assumes all concentrations and the diffusion coefficient are constant. It becomes $a\propto \sqrt{t}$ when the seed radius is negligible. This reservoir-fed law differs from curvature-controlled Ostwald ripening in a depleting many-particle system.

Diffusion is net transport associated with random microscopic motion. A localized patch of a dilute solute can spread through a stationary solvent: individual molecules move in both directions, but more cross from the crowded side of an imaginary surface than from the less crowded side. This imbalance gives a net diffusive flux down the concentration gradient; the molecules do not need to sense which direction has lower concentration. At uniform concentration, microscopic exchanges continue even though the net diffusive flux vanishes. ^[ref]

The distinction from advection is easiest to see in a dye patch. Uniform bulk motion carries its centre downstream; diffusion broadens it relative to that centre. Both can happen together. A moving or spreading patch by itself does not identify the mechanism: a nonuniform velocity field can also stretch a patch. ^[ref]

For a dilute nonreacting solute in a uniform medium at fixed temperature, Fick's first law models the relative diffusive flux as $\mathbf{j}=-D\nabla C$, with scalar diffusion coefficient $D>0$ of units ${\mathrm{m}}^2{\mathrm{s}}^{-1}$. Here $C$ is the amount of substance per unit volume, so $\mathbf{j}$ has units $\mathrm{mol}{\mathrm{m}}^{-2}{\mathrm{s}}^{-1}$. With carrier velocity field $\mathbf{u}$, the total flux is $C\mathbf{u}+\mathbf{j}$. The continuity equation, with no sources or chemical reactions, givesFor constant $D$ and incompressible flow, this reduces to $C_t+\mathbf{u}\cdot \nabla C=D{\nabla }^{2}C$, where ${\nabla }^2$ is the Laplacian. In a stationary carrier it is the diffusion equation. These assumptions matter: variable $D$ must remain inside the divergence, and changing sources require extra terms. ^[ref]

The useful surprise is the common mathematics: thermal conduction smooths temperature through the analogous heat equation, while molecular diffusion smooths composition. This analogy already motivated the original formulation of Fick's laws. ^[ref] However, the thermal diffusivity and the solutal diffusivity need not be equal. A diffusive time scale $t_D\sim L^2/D$ means that crossing ten times the distance requires roughly one hundred times the time. This is why molecular diffusion is especially effective across thin layers, and why chaotic advection can assist mixing by making those layers thinner. Physical diffusion is a transport mechanism, distinct from the broader probabilistic class of Markov diffusion.

The coefficient $\sigma$ multiplying the Brownian motion increment in a scalar stochastic differential equation. The local variance rate is ${\sigma }^2$, and the second-order coefficient in the backward Kolmogorov backward equation is ${\sigma }^2/2$. This distinguishes the amplitude from the physical diffusion coefficient convention $D={\sigma }^2/2$.

A birth-death process with unit upward and downward jumps and rates $b(x)$ and $d(x)$ has infinitesimal generatorIf $f$ varies slowly over one jump, Taylor expansion through second order gives $Lf\simeq (b-d)f^{\prime }+(b+d)f^{\prime \prime }/2$. Thus its diffusion approximation has drift coefficient $u=b-d$ and, in the conventiondiffusion coefficient $D=(b+d)/2$. Equivalently it solves $d{X}_t=u(X_t)dt+\sqrt{b(X_t)+d(X_t)}d{B}_t$, where $B$ is Brownian motion. The sum measures the variance of the short-time increment: births and deaths both add noise even when their mean drifts cancel.

For a population density $x=n/N$ with jump size $1/N$ and birth/death rates $N\beta (x),N\delta (x)$, the drift is $\beta -\delta$ and the diffusion coefficient is $(\beta +\delta )/(2N)$. Fluctuations therefore shrink as $N^{-1/2}$. A useful approximation requires many reaction events while the rates change little; it can fail near an absorbing state or when only a few individuals remain. A continuous diffusion can even visit negative values unless its boundary treatment is specified. The approximation replaces the discrete jump model rather than asserting equality of their laws.^[ref]

A spin echo refocuses a nucleus's fixed detuning. A nucleus that moves through a magnetic field gradient can have different detunings during the two intervals. To see the distinction without assuming that all noise is reversible, let $\Delta (t)$ vary during free evolution. With the same pulse convention, its residual phase at $2\tau$ isStatic detuning cancels because $\int y(t)dt=0$. A detuning ${\Delta }_a$ during the first interval and ${\Delta }_b$ during the second leaves $\Phi =\tau ({\Delta }_a-{\Delta }_b)$, which need not vanish. The final component signal is $e^{i\Phi }$ before relaxation; changing the overall receiver convention changes its conjugation but not its magnitude.

For an original quantitative model, take a constant magnetic field gradient $G=d{B}_z/dx$ and unrestricted one-dimensional Brownian motion $x(t)=x_0+\sqrt{2D}{W}_t$, where $D$ is the diffusion coefficient and $W_t$ is standard Brownian motion. Then $\Delta (t)={\Delta }_0+\gamma Gx(t)$. The constant terms cancel, and the Gaussian distribution of the remaining phase follows from the Brownian covariance kernel $\mathbb{E}[W_s{W}_t]=\min (s,t)$. Writing $Q(r)={\int }_r^{2\tau }y(t)dt$ givesThe characteristic function of a zero-mean Gaussian distribution now supplies the extra factor $\mathbb{E}[e^{i\Phi }]=e^{-\mathrm{Var}\Phi /2}$, soHere $\gamma$ is in radians per second per tesla, $G$ in tesla per metre, and $D$ in square metres per second: the extra exponent is dimensionless. The figure chooses $\gamma /(2\pi )=42.58MHz/T$, $G=0.010T/m$ and $D=2.0\times 10^{-9}{\mathrm{m}}^2/s$. With the earlier $\tau$ and $T_2$, diffusion reduces the echo from $0.60653$ to about $0.32930$. These are illustrative model parameters, not values extracted from Hahn's experiment.

This result assumes instantaneous perfect pulses, a constant magnetic field gradient, unrestricted diffusion with constant diffusion coefficient, and independent homogeneous transverse relaxation. Restricted motion, changing gradients, finite pulses and interactions between spins can require different attenuation laws. Even a fixed Hamiltonian is refocused only for the terms whose sign the pulse reverses: a simultaneous 180° rotation of two spins leaves their $J_{z,1}{J}_{z,2}$ product unchanged, because both factors change sign. A recovered echo therefore demonstrates a particular recoverable phase history, rather than reversal of every microscopic process.

Diffusion-controlled drug release occurs when molecular transport down a concentration gradient controls the relevant release time scale. It is a mechanism defined for a particular material, geometry and environment, rather than a name for every slowly changing release curve. A classic theoretical treatment by T. Higuchi distinguishes several matrix and geometrical models.^[ref]

A simple example is an initially uniform dissolved substance of concentration $c_0$ in a planar material, with constant diffusion coefficient $D$. At early times, approximate the material as the half-space $x>0$ and impose a perfect sink condition in drug release, $c(0,t)=0$. Solving the diffusion equation givesFick's first law then gives outward mass flux of magnitudeThis derivation is for dissolved material under the stated boundary condition; it does not identify every square-root law with the same solid-particle matrix model.

For a finite layer of thickness $L$ releasing through one face, the early-time fraction is approximately $F=2\sqrt{Dt/\pi }/L$. With illustrative $L=0.10\mathrm{mm}$ and $D=10^{-12}{\mathrm{m}}^2{\mathrm{s}}^{-1}$, the time scale $L^2/D$ is $10^4\mathrm{s}$, and at $100\mathrm{s}$ the approximation gives $F\approx 0.113$. Extending it to $10^4\mathrm{s}$ would give $F>1$, demonstrating its failure once the finite size matters. The initial mathematical singularity in $J(t)$ likewise reflects idealized initial and boundary conditions.

A useful experimental check changes thickness while keeping material and boundary conditions comparable: this diffusion model predicts a time scale proportional to $L^2$. Swelling, changing diffusivity, degradation or external transport can alter that scaling. A membrane used by the measurement apparatus can also limit transport, so observing a slow curve alone does not locate the controlling resistance inside the formulation.^[ref]

A diffusion equation models the redistribution of a density by transport down its spatial gradient. If $u$ is a concentration, local conservation of mass and a diffusive flux $J$ give $u_t+\nabla \cdot J=0$. Combining this with Fick's first law, $J=-D\nabla u$, givesA constant scalar diffusion coefficient $D>0$ reduces this to the heat equation $u_t=D\Delta u$. A variable diffusion coefficient must remain inside the divergence; a density-dependent coefficient leads to a nonlinear diffusion equation.

For a Fourier mode $u(x,t)=a(t)\cos (kx)$, the constant-coefficient equation gives $a(t)=a(0)e^{-D{k}^{2}t}$. Fine spatial oscillations decay faster than broad ones, explaining the smoothing action. Spreading across a length $L$ takes a diffusive time scale of order $L^2/D$. In a closed system with zero boundary diffusive flux, the spatial integral of $u$ is conserved even while its profile changes.

The heat kernel describes the spreading of an initially concentrated source. It is positive at every distance for every positive time, giving infinite propagation speed in this continuum model. The porous medium equation changes that behavior because its effective diffusivity vanishes at zero density, allowing finite moving fronts. Initial and boundary conditions select the particular solution of either equation.

For diffusion coefficient $D>0$ and first-order removal rate $\mu >0$, the diffusion length isIt measures the spatial range reached when diffusion competes with removal on the mean timescale $\tau =1/\mu$.

For example, a one-dimensional reaction-diffusion equation with constant coefficients isA steady concentration maintained at $c(0)=c_0$ in $x>0$, decaying at infinity, satisfiesSubstitution gives $D/{\ell }^2=\mu$. Thus $\ell$ is an exact decay length for this geometry.

A root-mean-square displacement uses a different numerical convention. For ordinary diffusion in $d$ dimensions and an independent exponential removal time with mean $\tau$, the mean squared displacement at removal is $2dD\tau$; its square root is $\sqrt{2d}\ell$. Calling $\ell$ a characteristic travel distance does not equate these two definitions.