Papers
Topics
Authors
Recent
Search
2000 character limit reached

Single-File Fluids: Anomalous Transport

Updated 14 July 2026
  • Single-file fluids are one-dimensional or quasi-1D systems where particles cannot overtake each other, ensuring preserved order and unique transport properties.
  • They exhibit anomalous tracer subdiffusion with displacement scaling as T^(1/4) and variance proportional to sqrt(T), deviating from classical diffusion laws.
  • Multiple theoretical frameworks—including hydrodynamic, Langevin, and path-integral approaches—reveal the role of confinement, compressibility, and initial state in these systems.

Single-file fluids are one-dimensional or quasi-one-dimensional interacting fluids in which particles cannot overtake one another, so their longitudinal order is preserved for all times. This no-passing constraint is the defining kinematic property and is the source of the characteristic anomalous transport of a tagged particle: the typical displacement scales as XT∼T1/4X_T\sim T^{1/4}, while the variance scales as ⟨XT2⟩∝T\langle X_T^2\rangle \propto \sqrt{T}, rather than the T1/2T^{1/2} and TT laws of ordinary diffusion (Krapivsky et al., 2014). Modern work treats single-file fluids not only as strictly one-dimensional files, but also as strongly confined quasi-one-dimensional fluids with residual transverse structure; in that broader setting, the anomalous exponent survives, while amplitudes, equations of state, and structural correlations can depend sensitively on confinement geometry, compressibility, and initial preparation (Bénichou et al., 9 Jun 2026).

1. Kinematic definition and basic scaling laws

Single-file transport means no overtaking: particles move in a narrow channel while preserving their longitudinal order. In a strict one-dimensional file, particles literally move on a line and have no transverse freedom. In a quasi-one-dimensional file, particles move in a finite-width channel: they still cannot pass one another, but they can explore transverse coordinates inside the cross-section. The strict single-file condition is therefore topological rather than purely dimensional; what matters is that exchanges are geometrically forbidden (Montero et al., 2023).

For hard disks of unit diameter in a slit, the strict single-file regime is realized for channel width

w=1+ϵ,0≤ϵ≤32,w=1+\epsilon,\qquad 0\le \epsilon \le \frac{\sqrt{3}}{2},

within which disks cannot pass each other and only nearest-neighbor interactions along the channel matter (Montero et al., 2023). For hard spheres or disks of diameter ss in a channel of width s+Ws+W, a corresponding geometric condition is

W<32s≈0.866 s,W < \frac{\sqrt{3}}{2}s \approx 0.866\,s,

which ensures that neighboring particles cannot exchange order (Bénichou et al., 9 Jun 2026).

The principal dynamical hallmark is tracer subdiffusion. In macroscopic fluctuation theory and related hydrodynamic descriptions, the tracer displacement satisfies

XT∼T1/4,⟨XT2⟩∝T,X_T\sim T^{1/4},\qquad \langle X_T^2\rangle \propto \sqrt{T},

and all odd cumulants vanish by symmetry in the homogeneous unbiased setting (Krapivsky et al., 2014). Effective tagged-particle theories further resolve three temporal regimes—ballistic, diffusive, and subdiffusive—separated by characteristic crossover times. In the Langevin formulation with inertia, the mean-squared displacement behaves as

δx2(t)∼kBT t2(t≪τb),δx2(t)∼2Dt(τb≪t≪τd),δx2(t)∼21ρDtπ(t≫τd),\delta x^2(t)\sim k_B T\,t^2 \quad (t\ll \tau_b),\qquad \delta x^2(t)\sim 2Dt \quad (\tau_b\ll t\ll \tau_d),\qquad \delta x^2(t)\sim 2\frac{1}{\rho}\sqrt{\frac{Dt}{\pi}} \quad (t\gg \tau_d),

with ⟨XT2⟩∝T\langle X_T^2\rangle \propto \sqrt{T}0 and ⟨XT2⟩∝T\langle X_T^2\rangle \propto \sqrt{T}1 (0810.5210). A conceptually related coarse-grained description maps the file onto an effective discrete Edwards–Wilkinson equation whose roughness shows normal diffusion, subdiffusion, and finite-size saturation, with crossover times

⟨XT2⟩∝T\langle X_T^2\rangle \propto \sqrt{T}2

for particles of unit size and mean spacing ⟨XT2⟩∝T\langle X_T^2\rangle \propto \sqrt{T}3 (Centres et al., 2010).

2. Hydrodynamic, fluctuating, and path-integral descriptions

At large scales, single-file fluids are described by a conserved density field ⟨XT2⟩∝T\langle X_T^2\rangle \propto \sqrt{T}4 obeying fluctuating hydrodynamics,

⟨XT2⟩∝T\langle X_T^2\rangle \propto \sqrt{T}5

with space-time white noise ⟨XT2⟩∝T\langle X_T^2\rangle \propto \sqrt{T}6, diffusion coefficient ⟨XT2⟩∝T\langle X_T^2\rangle \propto \sqrt{T}7, and mobility ⟨XT2⟩∝T\langle X_T^2\rangle \propto \sqrt{T}8 (Krapivsky et al., 2014). These transport coefficients encode the collective relaxation of density gradients and the magnitude of current fluctuations. In the simplest exactly solvable case of Brownian point particles with hard-core repulsion,

⟨XT2⟩∝T\langle X_T^2\rangle \propto \sqrt{T}9

(Krapivsky et al., 2014).

The tracer coordinate is not an independent degree of freedom in this formulation. Because the number of particles to one side of the tagged particle is conserved, the tracer position is an exact functional of the density trajectory. With the tracer initially at the origin,

T1/2T^{1/2}0

so tagged-particle motion is slaved to integrated density/current fluctuations of the entire file (Krapivsky et al., 2014). This identity is the basis of the macroscopic fluctuation theory of tracer statistics, in which one studies the cumulant generating function

T1/2T^{1/2}1

through a saddle-point problem for an action functional in the density and conjugate response fields (Krapivsky et al., 2015).

The same tracer problem admits complementary stochastic formulations. The effective generalized Langevin equation for a tagged particle,

T1/2T^{1/2}2

encodes the ordinary bath through the T1/2T^{1/2}3-kernel and the single-file constraint through the long-memory T1/2T^{1/2}4 term (0810.5210). This produces the characteristic negative velocity autocorrelation tail

T1/2T^{1/2}5

together with a generalized fluctuation-dissipation relation for the colored effective noise (0810.5210). A different effective picture rewrites the file as a fluctuating elastic chain with displacement field T1/2T^{1/2}6,

T1/2T^{1/2}7

and effective elasticity

T1/2T^{1/2}8

which compactly reproduces the crossover from independent diffusion to correlated roughening and saturation (Centres et al., 2010).

At the microscopic level, overdamped interacting Brownian motion can also be rewritten as a path integral over full many-body trajectories. In the harmonic-chain approximation, this route yields an exact all-time tracer mean-squared displacement in the thermodynamic limit,

T1/2T^{1/2}9

which crosses from TT0 at short times to a TT1 law at long times (Castro-Villarreal et al., 2021). This suggests that path-integral, fluctuating-hydrodynamic, and effective-memory descriptions are best viewed as complementary representations of the same collective no-passing mechanism.

3. Tracer fluctuations, large deviations, and bath–tracer correlations

For arbitrary single-file systems with transport coefficients TT2 and TT3, perturbative macroscopic fluctuation theory yields explicit universal variance formulas. In the quenched setting,

TT4

while in the annealed setting,

TT5

The TT6 ratio expresses the persistent memory of the initial state: in the annealed case, initial density fluctuations assist tracer motion; in the quenched case they are absent (Krapivsky et al., 2015).

For Brownian hard-core point particles, the full long-time statistics can be obtained exactly. One explicit annealed large-deviation function is

TT7

and the far tails differ sharply between initial ensembles: TT8 Thus annealed and quenched preparations are not only quantitatively distinct; they have qualitatively different rare-event structure (Krapivsky et al., 2015).

Trajectory-level statistics are also accessible. For arbitrary single-file systems, the asymptotic two-time covariance of the tagged-particle displacement takes the universal form

TT9

w=1+ϵ,0≤ϵ≤32,w=1+\epsilon,\qquad 0\le \epsilon \le \frac{\sqrt{3}}{2},0

and the full cumulant generating functional of the tracer trajectory can be written within macroscopic fluctuation theory (Krapivsky et al., 2015). A plausible implication is that single-file subdiffusion is more naturally understood as a non-Markovian Gaussian process with universal covariance than as an isolated anomalous one-time observable.

A second major advance is the explicit determination of bath–tracer spatial correlations in the tracer frame. The generalized density profile

w=1+ϵ,0≤ϵ≤32,w=1+\epsilon,\qquad 0\le \epsilon \le \frac{\sqrt{3}}{2},1

generates mixed cumulants between tracer position and local bath density (Poncet et al., 2021). In the SEP, the first nontrivial scaling function is

w=1+ϵ,0≤ϵ≤32,w=1+\epsilon,\qquad 0\le \epsilon \le \frac{\sqrt{3}}{2},2

which quantifies the correlated density excess in front of a positively displaced tracer (Poncet et al., 2021). Exact hydrodynamic-limit equations for these bath–tracer correlations close as Wiener–Hopf equations for suitably normalized profile derivatives, turning an infinite many-body hierarchy into a solvable problem (Grabsch et al., 2023). This makes precise the common physical statement that tracer subdiffusion is generated by collective density rearrangements.

Single-file fluctuating hydrodynamics also possesses a general particle–gap duality. If the original transport coefficients are w=1+ϵ,0≤ϵ≤32,w=1+\epsilon,\qquad 0\le \epsilon \le \frac{\sqrt{3}}{2},3 and w=1+ϵ,0≤ϵ≤32,w=1+\epsilon,\qquad 0\le \epsilon \le \frac{\sqrt{3}}{2},4, the dual gap description has

w=1+ϵ,0≤ϵ≤32,w=1+\epsilon,\qquad 0\le \epsilon \le \frac{\sqrt{3}}{2},5

Under this map, tracer and integrated-current observables are exchanged at the hydrodynamic level (Rizkallah et al., 2022). This suggests that particles and gaps are not merely complementary interpretations, but universally dual hydrodynamic variables in single-file transport.

4. Equilibrium structure, transfer matrices, and thermodynamics under confinement

Single-file fluids are not exhausted by tracer anomalies. Strong confinement also creates a distinct equilibrium statistical mechanics. For hard disks in a slit of width w=1+ϵ,0≤ϵ≤32,w=1+\epsilon,\qquad 0\le \epsilon \le \frac{\sqrt{3}}{2},6, the longitudinal contact distance between neighboring disks with transverse separation w=1+ϵ,0≤ϵ≤32,w=1+\epsilon,\qquad 0\le \epsilon \le \frac{\sqrt{3}}{2},7 is

w=1+ϵ,0≤ϵ≤32,w=1+\epsilon,\qquad 0\le \epsilon \le \frac{\sqrt{3}}{2},8

so transverse configuration directly modulates longitudinal exclusion (Montero et al., 2023). This geometry can be mapped onto a one-dimensional non-additive hard-rod mixture, with the transverse coordinate reinterpreted as a species label. In the continuum limit the mapping is exact, and all species must share the same chemical potential because they represent different transverse states of the same physical particle (Montero et al., 2023).

The resulting transfer-matrix/eigenvalue problem yields the equation of state, the transverse density profile, w=1+ϵ,0≤ϵ≤32,w=1+\epsilon,\qquad 0\le \epsilon \le \frac{\sqrt{3}}{2},9th-neighbor distributions, the longitudinal radial distribution function ss0, the structure factor ss1, defect statistics in the zigzag regime, and asymptotic correlation lengths (Montero et al., 2023). At high pressure the preferred arrangement is a zigzag,

ss2

with particles alternating near opposite walls. Defects correspond to same-side contacts such as ss3 or ss4, and their probability is exponentially suppressed: ss5 This is a sharp structural signature of confinement-induced packing, not of any thermodynamic phase transition (Montero et al., 2023).

The confined hard-disk equation of state differs in an instructive way from the Tonks gas. With reduced longitudinal pressure ss6, linear density ss7, and close-packing density

ss8

the high-pressure asymptotics are

ss9

so the close-packing pole has residue s+Ws+W0, not the Tonks value s+Ws+W1 (Montero et al., 2022). The same transfer-matrix treatment yields exact low-pressure virial coefficients and shows that standard irreducible Mayer diagrams do not provide a complete account of virial coefficients in confined geometries, because reducible diagrams do not cancel in the usual bulk manner (Montero et al., 2022). This directly refutes the common simplification that a quasi-one-dimensional single-file fluid is merely a Tonks gas with a renormalized diameter.

The same mapping philosophy extends to a broader equilibrium theory of strongly confined fluids. A general Q1D single-file fluid with nearest-neighbor interactions can be mapped exactly onto a one-dimensional mixture—discrete or polydisperse—whose species labels encode transverse position, orientation, or other confined variables, subject to equal chemical potential for all effective species (Montero, 6 Oct 2025). In this framework one obtains closed expressions for the Gibbs free energy, longitudinal and transverse compressibility factors, exact longitudinal RDFs, and anisotropy-adapted RDF-like quantities (Montero, 6 Oct 2025). The thesis-level synthesis is that strong confinement produces a statistically anisotropic fluid that is neither a bulk fluid nor a strict 1D gas: transverse freedom survives in the equation of state, pressure tensor, and structural correlations (Montero, 6 Oct 2025).

5. Geometry, inhomogeneity, activity, and nonclassical transport regimes

Recent work has sharpened the distinction between strict one-dimensional files and finite-width confined files. Starting from Brownian dynamics in the full confined geometry, one can derive an exact large-scale one-dimensional fluctuating hydrodynamics for the cross-section-averaged density,

s+Ws+W2

Here s+Ws+W3 is fixed by the confined equilibrium equation of state of the full channel, not by a strict-1D rod model (Bénichou et al., 9 Jun 2026). In the minimal hard-core setting, s+Ws+W4 can become non-monotonic in density, so even without overtaking the collective diffusivity may differ qualitatively from the Tonks prediction. This geometry-dressed hydrodynamics preserves the usual s+Ws+W5 current and tracer fluctuation laws, but with geometry-dependent amplitudes (Bénichou et al., 9 Jun 2026).

A related homogenization result applies to periodic inhomogeneous environments. For noninteracting particles diffusing in a one-dimensional periodic medium with potential s+Ws+W6 and local diffusivity s+Ws+W7, the effective single-particle diffusivity is

s+Ws+W8

In a hard-core single file built from such particles, the long-time tracer variances are exactly those of homogeneous single-file diffusion with s+Ws+W9 replaced by W<32s≈0.866 s,W < \frac{\sqrt{3}}{2}s \approx 0.866\,s,0: W<32s≈0.866 s,W < \frac{\sqrt{3}}{2}s \approx 0.866\,s,1 Thus periodic spatial inhomogeneity renormalizes the transport coefficient but not the asymptotic single-file scaling form (Sorkin et al., 2023).

Activity introduces a different modification. For active Brownian particles in a periodic one-dimensional channel, the long-time law

W<32s≈0.866 s,W < \frac{\sqrt{3}}{2}s \approx 0.866\,s,2

survives, but the mobility is controlled by active-state mechanical quantities rather than equilibrium compressibility. The paper defines the reduced kinetic temperature

W<32s≈0.866 s,W < \frac{\sqrt{3}}{2}s \approx 0.866\,s,3

the free active diffusivity

W<32s≈0.866 s,W < \frac{\sqrt{3}}{2}s \approx 0.866\,s,4

and the constant-Pe compressibility

W<32s≈0.866 s,W < \frac{\sqrt{3}}{2}s \approx 0.866\,s,5

The generalized one-dimensional mobility is then

W<32s≈0.866 s,W < \frac{\sqrt{3}}{2}s \approx 0.866\,s,6

so the tagged-particle MSD interpolates from a ballistic short-time regime,

W<32s≈0.866 s,W < \frac{\sqrt{3}}{2}s \approx 0.866\,s,7

to the asymptotic single-file law W<32s≈0.866 s,W < \frac{\sqrt{3}}{2}s \approx 0.866\,s,8 (Akintunde et al., 2024). This suggests that activity renormalizes amplitudes and crossover physics while leaving the no-passing exponent intact.

Single-file constraints can also generate non-Fickian collective transport in open pores. In a symmetric exclusion process with reversible wall binding and unbinding, short channels with slow binding kinetics exhibit

W<32s≈0.866 s,W < \frac{\sqrt{3}}{2}s \approx 0.866\,s,9

rather than the usual XT∼T1/4,⟨XT2⟩∝T,X_T\sim T^{1/4},\qquad \langle X_T^2\rangle \propto \sqrt{T},0 scaling. The mechanism is two-state intermittency: rare high-flux bursts through an entirely unbound pore,

XT∼T1/4,⟨XT2⟩∝T,X_T\sim T^{1/4},\qquad \langle X_T^2\rangle \propto \sqrt{T},1

are weighted by a fast-state probability

XT∼T1/4,⟨XT2⟩∝T,X_T\sim T^{1/4},\qquad \langle X_T^2\rangle \propto \sqrt{T},2

because a single bound particle acts as a plug that mobile particles cannot bypass (Farrell et al., 2021). This is not a change of the constitutive law within a steady flowing state, but a transport regime generated by stochastic switching between conducting and jammed configurations.

6. Experimental realizations and the boundary of the single-file regime

A particularly direct experimental realization uses electrically interacting metallic beads in a narrow annular channel. The beads are confined strongly enough that they remain ordered and cannot pass one another, while random mechanical shaking supplies an effective thermal bath (Coste et al., 2010). The measured orthoradial mean-squared displacement shows a ballistic short-time regime, an approximately Fickian intermediate regime, and a long-time single-file regime,

XT∼T1/4,⟨XT2⟩∝T,X_T\sim T^{1/4},\qquad \langle X_T^2\rangle \propto \sqrt{T},3

with XT∼T1/4,⟨XT2⟩∝T,X_T\sim T^{1/4},\qquad \langle X_T^2\rangle \propto \sqrt{T},4 the single-file mobility (Coste et al., 2010).

The experiment is notable because the mobility depends on interaction strength, density, and effective temperature. The classical hard-sphere prediction,

XT∼T1/4,⟨XT2⟩∝T,X_T\sim T^{1/4},\qquad \langle X_T^2\rangle \propto \sqrt{T},5

fails for this system because it is insensitive to the tunable electrostatic repulsion. A compressibility-based expression,

XT∼T1/4,⟨XT2⟩∝T,X_T\sim T^{1/4},\qquad \langle X_T^2\rangle \propto \sqrt{T},6

or equivalently

XT∼T1/4,⟨XT2⟩∝T,X_T\sim T^{1/4},\qquad \langle X_T^2\rangle \propto \sqrt{T},7

captures the observed trends much better (Coste et al., 2010). The comparison is described as qualitatively excellent and quantitatively correct but not perfect; the remaining discrepancy is attributed to violation of the overdamped assumption, since the metallic-bead dynamics are not overdamped (Coste et al., 2010). This experiment therefore separates two ingredients of single-file transport that theory often conflates: the static part controlled by compressibility and interactions, and the dynamic part controlled by overdamped versus underdamped motion.

The single-file condition itself also has a clear boundary. For parallel hard squares in a slit pore, the strict single-file regime is

XT∼T1/4,⟨XT2⟩∝T,X_T\sim T^{1/4},\qquad \langle X_T^2\rangle \propto \sqrt{T},8

where particles cannot pass and the system reduces exactly to the Tonks gas. In the wider but still quasi-one-dimensional regime

XT∼T1/4,⟨XT2⟩∝T,X_T\sim T^{1/4},\qquad \langle X_T^2\rangle \propto \sqrt{T},9

particles can pass each other and at most two layers can form; interactions remain limited to first and second neighbors, and an exact transfer-matrix treatment based on dimer states is still possible (Gurin et al., 2015). The equation of state and nearest-neighbor distributions then show three continuous structural regimes: a one-layer fluid, a two-layer fluid, and a solid-like structure with strongly correlated layers (Gurin et al., 2015). This is the clearest exact illustration that the physics of single-file fluids is not robust to arbitrarily small relaxations of the no-passing constraint: once overtaking is allowed, Tonks-like one-dimensional order gives way to layering, second-neighbor couplings, and different high-density packing.

Taken together, theoretical, numerical, and experimental work shows that single-file fluids are best understood as a family of strongly constrained many-body systems organized by a single kinematic rule—preserved particle order—but not exhausted by it. The anomalous δx2(t)∼kBT t2(t≪τb),δx2(t)∼2Dt(τb≪t≪τd),δx2(t)∼21ρDtπ(t≫τd),\delta x^2(t)\sim k_B T\,t^2 \quad (t\ll \tau_b),\qquad \delta x^2(t)\sim 2Dt \quad (\tau_b\ll t\ll \tau_d),\qquad \delta x^2(t)\sim 2\frac{1}{\rho}\sqrt{\frac{Dt}{\pi}} \quad (t\gg \tau_d),0 transport law of the tracer is the most visible consequence of that rule, yet the full theory also requires transport coefficients δx2(t)∼kBT t2(t≪τb),δx2(t)∼2Dt(τb≪t≪τd),δx2(t)∼21ρDtπ(t≫τd),\delta x^2(t)\sim k_B T\,t^2 \quad (t\ll \tau_b),\qquad \delta x^2(t)\sim 2Dt \quad (\tau_b\ll t\ll \tau_d),\qquad \delta x^2(t)\sim 2\frac{1}{\rho}\sqrt{\frac{Dt}{\pi}} \quad (t\gg \tau_d),1 and δx2(t)∼kBT t2(t≪τb),δx2(t)∼2Dt(τb≪t≪τd),δx2(t)∼21ρDtπ(t≫τd),\delta x^2(t)\sim k_B T\,t^2 \quad (t\ll \tau_b),\qquad \delta x^2(t)\sim 2Dt \quad (\tau_b\ll t\ll \tau_d),\qquad \delta x^2(t)\sim 2\frac{1}{\rho}\sqrt{\frac{Dt}{\pi}} \quad (t\gg \tau_d),2, tracer–bath correlation fields, initial-condition dependence, and, in quasi-one-dimensional confinement, the equilibrium thermodynamics of the full confined geometry.

Topic to Video (Beta)

No one has generated a video about this topic yet.

Whiteboard

No one has generated a whiteboard explanation for this topic yet.

Follow Topic

Get notified by email when new papers are published related to Single-File Fluids.