Active Brownian Particle: A Minimal Model

• Active Brownian Particle is a minimal stochastic model that describes self-propelled motion with constant propulsion and rotational diffusion, foundational to active-matter theory.
• The model captures a transition from short-time ballistic behavior to long-time effective diffusion, detailing various transport regimes and scaling laws.
• Extensions include variable propulsion speeds, inertia, and environmental interactions, enabling realistic simulations of synthetic microswimmers and biological systems.

Searching arXiv for recent and foundational ABP papers to ground the article. arXiv search query: "active brownian particle" An active Brownian particle (ABP) is a minimal stochastic model of a self-propelled particle whose position evolves by persistent propulsion along an internal orientation while that orientation itself undergoes rotational diffusion. In its standard form, the propulsion speed is constant and the orientational dynamics are Brownian, so ballistic motion at short times crosses over to effective diffusion at long times; this combination has made the ABP a prototype for synthetic microswimmers, motile microorganisms, and active-matter theory more broadly (Basu et al., 2018, Soto, 2 Jan 2025).

1. Canonical stochastic formulation

The canonical overdamped ABP is defined by a position and an orientation. In two dimensions, with orientation angle $\phi(t)$ or $\varphi(t)$, the standard equations are

$$\dot{\mathbf r}=v_0\,\mathbf u(\varphi)+\sqrt{2D_t}\,\boldsymbol{\xi}(t), \qquad \dot\varphi=\sqrt{2D_r}\,\eta(t),$$

with $\mathbf u(\varphi)=(\cos\varphi,\sin\varphi)$, translational diffusion coefficient $D_t$, rotational diffusion coefficient $D_r$, and Gaussian white noises of zero mean and delta correlation (Jamali, 2020). A commonly used reduced variant omits translational noise entirely and retains only orientational stochasticity, in which case

$$\dot x=v_0\cos\phi(t),\qquad \dot y=v_0\sin\phi(t),\qquad \dot\phi=\sqrt{2D_R}\,\eta_\phi(t),$$

and all positional randomness is inherited from the diffusing heading (Basu et al., 2018).

In three dimensions, the same structure appears with a unit orientation vector $\mathbf e(t)$. One representative formulation is

$$\frac{d}{dt}\mathbf r(t)=v_s\mathbf e(t)-v_g\mathbf z+\boldsymbol{\xi}(t),\qquad \frac{d}{dt}\mathbf e(t)=\boldsymbol{\xi}_{\mathrm e}(t)\times\mathbf e(t),$$

which augments active motion by gravity and translational noise in a sedimentation setting (Vachier et al., 2017). Molecular-dynamics realizations often retain the ABP orientational law while allowing translational motion to emerge from explicit fluid degrees of freedom, as in

$$\dot{\mathbf r}=v_A \mathbf n+\sqrt{2D_T}\,\boldsymbol{\xi},\qquad \dot{\mathbf n}=\sqrt{2D_r}\,\mathbf n\times \boldsymbol{\zeta},$$

with $\varphi(t)$0 as the orientational persistence time (Wang et al., 2019).

The defining structural assumption is therefore persistent self-propulsion combined with rotational decorrelation. A common misunderstanding is that ABP dynamics are synonymous with white-noise-enhanced diffusion. The short-time process is instead non-Markovian in projected coordinates, because the active velocity components are bounded, temporally correlated, and mutually correlated through the common orientation process (Basu et al., 2018).

2. Transport regimes, moments, and effective diffusion

For an isolated ABP, rotational diffusion introduces a persistence timescale $\varphi(t)$1 or $\varphi(t)$2. At times short compared with this scale, motion is persistent; at long times, the particle executes an effective random walk. In two dimensions, the isolated-particle long-time diffusivity is

$\varphi(t)$3

for speed $\varphi(t)$4 and rotational diffusivity $\varphi(t)$5 (Soto, 2 Jan 2025). In three dimensions, the corresponding active contribution is $\varphi(t)$6, as used in molecular-dynamics validation studies (Wang et al., 2019).

The short-time regime is not merely ballistic. For a 2D ABP initialized with orientation along the $\varphi(t)$7-axis and without translational noise, the mean position and variances are exactly anisotropic: $\varphi(t)$8 with

$\varphi(t)$9

at early times (Basu et al., 2018). The full short-time position distribution is non-Gaussian and “sickle-like”; the $$\dot{\mathbf r}=v_0\,\mathbf u(\varphi)+\sqrt{2D_t}\,\boldsymbol{\xi}(t), \qquad \dot\varphi=\sqrt{2D_r}\,\eta(t),$$0-marginal has a ballistic edge at $$\dot{\mathbf r}=v_0\,\mathbf u(\varphi)+\sqrt{2D_t}\,\boldsymbol{\xi}(t), \qquad \dot\varphi=\sqrt{2D_r}\,\eta(t),$$1, while the $$\dot{\mathbf r}=v_0\,\mathbf u(\varphi)+\sqrt{2D_t}\,\boldsymbol{\xi}(t), \qquad \dot\varphi=\sqrt{2D_r}\,\eta(t),$$2-marginal is Gaussian but with non-Brownian variance scaling as $$\dot{\mathbf r}=v_0\,\mathbf u(\varphi)+\sqrt{2D_t}\,\boldsymbol{\xi}(t), \qquad \dot\varphi=\sqrt{2D_r}\,\eta(t),$$3 (Basu et al., 2018). This directly contradicts the frequent heuristic replacement of ABPs by passive Brownian particles with an effective diffusivity at all times.

The same persistence controls first-passage properties. In the transverse direction, short-time ABP motion maps to the random acceleration process, yielding survival decay $$\dot{\mathbf r}=v_0\,\mathbf u(\varphi)+\sqrt{2D_t}\,\boldsymbol{\xi}(t), \qquad \dot\varphi=\sqrt{2D_r}\,\eta(t),$$4 rather than the Brownian $$\dot{\mathbf r}=v_0\,\mathbf u(\varphi)+\sqrt{2D_t}\,\boldsymbol{\xi}(t), \qquad \dot\varphi=\sqrt{2D_r}\,\eta(t),$$5 (Basu et al., 2018). Confinement studies likewise show that the ratio between orientational persistence and trap relaxation time governs a crossover from Boltzmann-like stationary distributions to non-Boltzmann boundary accumulation (Wang et al., 2019).

A second misconception is that the effective-diffusion description is universally exact once one waits long enough. Several extensions show that while a long-time diffusivity exists, the route to it can contain multiple crossover scales and nontrivial transient structure, including ballistic, superdiffusive, or localization-like regimes depending on interactions, switching, confinement, or disorder (Jamali, 2020, Zeitz et al., 2016).

3. Beyond constant-speed overdamped propulsion

The standard constant-speed ABP is analytically convenient but not exhaustive. A minimal extension replaces the constant propulsion speed by a stochastic telegraph process,

$$\dot{\mathbf r}=v_0\,\mathbf u(\varphi)+\sqrt{2D_t}\,\boldsymbol{\xi}(t), \qquad \dot\varphi=\sqrt{2D_r}\,\eta(t),$$6

so that the particle alternates between passive and active states while the orientation continues to diffuse as in the ordinary ABP (Jamali, 2020). This “active-passive Brownian particle” retains orientational persistence but adds a second memory channel through speed-state switching. Its mean-squared displacement contains three relaxation scales,

$$\dot{\mathbf r}=v_0\,\mathbf u(\varphi)+\sqrt{2D_t}\,\boldsymbol{\xi}(t), \qquad \dot\varphi=\sqrt{2D_r}\,\eta(t),$$7

and its long-time diffusivity interpolates continuously between passive Brownian motion and ordinary ABP motion: $$\dot{\mathbf r}=v_0\,\mathbf u(\varphi)+\sqrt{2D_t}\,\boldsymbol{\xi}(t), \qquad \dot\varphi=\sqrt{2D_r}\,\eta(t),$$8 (Jamali, 2020). The same work shows that a run-and-tumble particle can be mapped onto this model at the level of long-time diffusivity, but explicitly does not prove strict pathwise equivalence (Jamali, 2020).

A different extension treats activity mechanically rather than phenomenologically. For an anisotropic particle driven by a fluctuating propulsion force applied at a single off-center point, translational and rotational noises no longer decouple. The overdamped limit contains a diffusion tensor with off-diagonal translation-rotation terms and a nonzero noise-induced drift,

$$\dot{\mathbf r}=v_0\,\mathbf u(\varphi)+\sqrt{2D_t}\,\boldsymbol{\xi}(t), \qquad \dot\varphi=\sqrt{2D_r}\,\eta(t),$$9

which survives the overdamped limit and alters the long-time diffusivity (Thiffeault et al., 2021). This result clarifies that standard ABP equations already encode a modeling choice: they discard mechanistic coupling between propulsion-force fluctuations, torque fluctuations, and anisotropic drag.

Further generalizations change the orientational sector itself. Under stochastic orientational resetting,

$\mathbf u(\varphi)=(\cos\varphi,\sin\varphi)$0

at random times, a 2D ABP develops a finite drift along the resetting direction,

$\mathbf u(\varphi)=(\cos\varphi,\sin\varphi)$1

anisotropic long-time diffusion, and a complex intermediate scattering function with a nonzero imaginary part, which directly signals directed motion at large scales (Baouche et al., 2024). Under velocity-alignment interactions and heading-correlated active fluctuations, active angular noise introduces an effective drift term in the Fokker–Planck description,

$\mathbf u(\varphi)=(\cos\varphi,\sin\varphi)$2

which suppresses collective motion more efficiently than passive noise and can generate a bistable mean-field regime (Grossmann et al., 2012).

Inertia provides another nontrivial extension. An underdamped, shape-generalized ABP with translational and rotational inertia, active force and torque, and a full hydrodynamic resistance matrix exhibits an initial inertial transition before converging to a circular helix in the deterministic zero-noise limit; the paper reports this helical attractor for randomly generated asymmetric particles as well (Martins et al., 2022).

4. Confinement, boundaries, and external potentials

ABPs interacting with boundaries approach nonequilibrium steady states that differ sharply from passive Brownian equilibrium. For non-interacting 2D ABPs outside an inclusion, the steady Smoluchowski equation reduces in the main analysis to

$\mathbf u(\varphi)=(\cos\varphi,\sin\varphi)$3

with persistence length $\mathbf u(\varphi)=(\cos\varphi,\sin\varphi)$4 (Wagner et al., 2021). The asymptotic far-field solution has a diffusion-like structure,

$\mathbf u(\varphi)=(\cos\varphi,\sin\varphi)$5

so the bulk density is harmonic and the steady current obeys

$\mathbf u(\varphi)=(\cos\varphi,\sin\varphi)$6

(Wagner et al., 2021). The crucial point is that the boundary layer sets an effective Neumann condition through the normal flux, so boundary breaking of detailed balance generates long-ranged multipolar density and current fields (Wagner et al., 2021).

In a circular disk with an absorbing boundary, the full time-dependent Fokker–Planck operator can be expanded in the passive eigenbasis. The active problem becomes non-Hermitian, and increasing activity drives real eigenvalues into complex-conjugate pairs at exceptional points (Trapani et al., 2023). This spectral structure produces first-passage behavior that depends strongly on activity and, to a lesser extent, on rotational diffusivity (Trapani et al., 2023).

Harmonic confinement provides an analytically tractable but nonequilibrium benchmark. For a 2D ABP in an isotropic harmonic trap, the Fokker–Planck operator becomes lower diagonal in the passive eigenbasis, so the relaxation spectrum is unchanged by activity even though the steady state is not equilibrium (Caraglio et al., 2022). Exact expressions for the positional autocorrelation function, velocity autocorrelation function, and time-dependent diffusion coefficient show nonmonotonic behavior that serves as a fingerprint of the nonequilibrium dynamics (Caraglio et al., 2022). A stochastic path-integral reformulation later derived exact time-dependent mean position and mean-square displacement, including full dependence on initial conditions, and a systematic expansion of the position distribution in powers of the propulsion speed (Littek et al., 30 Sep 2025).

A wall can also stabilize wetting layers in persistent ABP systems. In a quasi-one-dimensional geometry, simulations show a temporal transition from a homogeneous polarization regime, where the whole dense wetting film is polarized, to a heterogeneous regime in which polarization survives only at the interface (Perez-Bastías et al., 17 Apr 2025). The corresponding two-field continuum theory retains both density and polarization because adiabatic elimination of the latter fails when persistence is large enough that orientational relaxation is not fast compared with accumulation dynamics (Perez-Bastías et al., 17 Apr 2025).

5. ABPs in disordered and complex media

Complex media fundamentally modify ABP transport by coupling persistence to quenched geometry. In a two-dimensional random Lorentz gas of fixed overlapping obstacles, the ABP obeys

$\mathbf u(\varphi)=(\cos\varphi,\sin\varphi)$7

with purely repulsive Weeks–Chandler–Andersen obstacle interactions (Zeitz et al., 2016). Near the void-space percolation threshold, ABPs display the same universal critical subdiffusion as passive and ballistic tracers,

$\mathbf u(\varphi)=(\cos\varphi,\sin\varphi)$8

but persistent motion leads to superdiffusive transients and faster access to the asymptotic regime (Zeitz et al., 2016). Below the threshold, stronger propulsion can suppress, rather than enhance, long-time diffusion because persistent swimmers spend longer pushing into obstacles; this is reflected in reduced effective swimming speed and effective persistence time extracted from the velocity autocorrelation function (Zeitz et al., 2016).

In three-dimensional random and porous environments, quenched disorder creates qualitatively new dynamical subpopulations. Simulations in a random pinned matrix and in a frozen percolating gel show absorbed and localized particles close to obstacles, enhanced dynamical heterogeneity, and geometry-dependent restructuring of motility-induced phase separation (MIPS) (Moore et al., 2023). The mean pore chord length is substantially smaller in the random-pinning case than in the gel,

$\mathbf u(\varphi)=(\cos\varphi,\sin\varphi)$9

and the random environment correspondingly produces stronger slowing and localization (Moore et al., 2023). At high activity, MIPS in random disorder resembles nucleation and growth in a disorder-selected location, whereas in porous gels phase separation becomes pore-by-pore and strongly patterned by the matrix geometry (Moore et al., 2023).

Polymer solutions introduce a different kind of complexity. Langevin dynamics simulations of a single ABP in a three-dimensional semidilute polymer solution find that the long-time diffusion coefficient can become a non-monotonic function of particle radius at sufficiently large active force, so that a larger particle diffuses faster than a smaller one (Du et al., 2018). The mechanism is a competition between persistence, enhanced by the size dependence of $D_t$0, and polymer-induced caging and effective viscosity (Du et al., 2018). The same study introduces a phenomenological active effective viscosity,

$D_t$1

which exceeds the passive effective viscosity and depends strongly on both $D_t$2 and $D_t$3 (Du et al., 2018). This suggests that passive microrheological intuition does not transfer directly to active probes.

6. Collective descriptions, phase coexistence, and substrate patterning

At finite density, ABPs are often coarse-grained into density and orientational fields. One route starts from the single-particle Fokker–Planck equation

$D_t$4

for ABPs on an inhomogeneous motility landscape, and truncates after the first orientational moment: $D_t$5 This yields

$D_t$6

with

$D_t$7

(Mishra et al., 2022). Patterned activity profiles then drive nonequilibrium steady states in which density is lower in high-activity regions, maximal at interfaces, and nearly constant in passive regions, so the ABP assembly can mimic the substrate pattern (Mishra et al., 2022).

For phase coexistence and wetting, a scalar density theory can be insufficient. A recent two-field theory evolves both density and polarization because, in highly persistent systems, polarization is not always slaved to the density gradient (Perez-Bastías et al., 17 Apr 2025). In the proposed one-dimensional closure, the density equation is

$D_t$8

while polarization obeys

$D_t$9

(Perez-Bastías et al., 17 Apr 2025). The theory is constructed to capture both bulk liquid-gas coexistence and wall-induced wetting and attributes the stabilization of the wetting film to a permanent interfacial polarization and its associated active stress (Perez-Bastías et al., 17 Apr 2025).

At the kinetic-theory level, moderate-density self-diffusion can be organized around a density-renormalized propulsion speed. A tracer ABP among hard disks experiences persistent collisions that reduce the effective streaming velocity to

$D_r$0

and generate anisotropic collision-induced diffusivities,

$D_r$1

before rotational diffusion restores isotropic long-time transport (Soto, 2 Jan 2025). To leading consistent order at high Péclet number, the long-time diffusivity again takes the isolated-particle form with $D_r$2 replaced by $D_r$3,

$D_r$4

but the theory becomes unphysical beyond area fraction about $D_r$5, where the binary-collision closure breaks down (Soto, 2 Jan 2025). This clarifies both the utility and the limits of ABP kinetic descriptions in dense active matter.

Taken together, these results define the ABP not as a single equation set but as a modeling class centered on persistent propulsion plus orientational diffusion. The standard constant-speed overdamped form remains the reference point, but modern ABP theory now includes stochastic propulsion amplitudes, translation-rotation coupling, inertia, resetting, confinement, quenched disorder, and multiscale field theories. A plausible implication is that the enduring value of the ABP framework lies less in any one canonical equation than in its role as the minimal platform from which increasingly realistic active-particle physics can be introduced in a controlled way.