Generalized Fick-Jacobs Approach

• Generalized Fick-Jacobs approach is a family of reduced transport models that project diffusion in higher-dimensional confined geometries onto effective one-dimensional dynamics.
• It incorporates variable diffusivity, entropic and free-energy potentials, and nonlinear interactions to accurately capture complex transport phenomena.
• The method enables rigorous analytical treatments and computational validations in systems such as active matter, reaction-diffusion setups, and polymer translocation.

Searching arXiv for recent and foundational papers on generalized Fick–Jacobs approaches. The generalized Fick–Jacobs approach denotes a family of reduced transport descriptions in which diffusion or drift–diffusion in a confined higher-dimensional geometry is projected onto an effective one-dimensional dynamics along the channel axis. Its common premise is that transverse degrees of freedom relax rapidly compared with longitudinal motion, so the geometry can be encoded through a cross-sectional area, width, or an effective free-energy landscape. In the literature, this reduction has been extended well beyond the classical setting of passive non-interacting diffusion to include hard-core exclusion, adsorption–desorption and bulk reactions, conservative and solenoidal forcing, active matter, stochastic resetting, polymer translocation, curved three-dimensional channels, and higher-order flux corrections [(Romero et al., 2012); (Suárez et al., 2016); (Zhao, 2021)].

1. Classical reduction and asymptotic basis

The classical Fick–Jacobs equation replaces diffusion in a narrow channel of cross-sectional area $A(x)$ by the effective one-dimensional evolution

$$\frac{\partial C(x,t)}{\partial t} = D_0 \,\frac{\partial}{\partial x} \left[ A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$

where $C(x,t)$ is the longitudinal concentration and $D_0$ is the diffusion coefficient (Romero et al., 2012). In a more general formulation with inhomogeneous diffusivity, the same structure becomes

$$\frac{\partial C(x,t)}{\partial t} = \frac{\partial}{\partial x} \left[ D(x)\,A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$

so that both confinement and transport inhomogeneity are absorbed into the projected operator (Romero et al., 2011).

The reduction is based on the standard Fick–Jacobs approximation: diffusion in a confined geometry is reduced to an effective one-dimensional description along the channel axis, with the transverse degrees of freedom assumed to equilibrate rapidly compared with longitudinal motion (Romero et al., 2012). This assumption reappears, with model-specific variants, across essentially all generalizations. In passive channels it is often formulated as a smoothly varying geometry; in active systems it is combined with slow width variation and, in some treatments, a channel width larger than the boundary-layer thickness; in reaction and adsorption problems it is written as near-uniform concentration across each transverse section (Zhao, 2021, Ledesma-Durán et al., 2016).

A rigorous asymptotic underpinning was given for anisotropic diffusion in slender impermeable tubes of revolution, where the small parameter

$$\epsilon=\epsilon_s^2\frac{D_\parallel}{D_\perp} = \frac{t_{tr}}{t_L}\ll 1$$

measures the ratio of transverse relaxation time to longitudinal diffusion time. In that framework, the classical Fick–Jacobs equation appears as the leading outer solvability condition of a matched-asymptotic-expansions construction, rather than as a heuristic closure. The reduced concentration $c(z,t)$ satisfies

$$\frac{\partial c(z,t)}{\partial t} = D_\parallel \frac{\partial}{\partial z} \left\{ A(z)\frac{\partial}{\partial z}\left[\frac{c(z,t)}{A(z)}\right] \right\}, \qquad A(z)=\pi r(z)^2,$$

while the full leading approximation also contains spatial boundary layers, an initial temporal layer, and corner layers (Traytak, 2013). This makes explicit that the purely one-dimensional equation is the bulk description, not the entire asymptotic solution.

2. Geometry as entropic potential, mean force, and operator structure

A central feature of generalized Fick–Jacobs theory is that geometry is recast as an effective longitudinal potential. In the simplest passive setting, the local free energy or entropic potential is obtained from the accessible cross-section, so that narrow regions act like barriers and wide regions like wells (Malgaretti et al., 2023, Malgaretti et al., 2023). This viewpoint survives most extensions, but the effective potential changes character once the physics is enlarged.

For Brownian particles subject to a general force field containing both curl-free and divergence-free parts,

$$\mathbf F(\mathbf r)= -\nabla \Phi(\mathbf r)+\nabla\times\boldsymbol\Psi(\mathbf r),$$

the reduced dynamics is governed by a generalized potential of mean force $\mathcal F(x)$. Its first term is the standard entropic contribution from the logarithm of the accessible cross-section; its second term is the cumulative effect of the divergence-free force projected along the channel axis and weighted by the conditional equilibrium distribution in the cross section (Martens et al., 2012). In three-dimensional periodic channels driven by a constant bias and a pressure-driven flow, the same generalized free-energy structure controls the stationary density, mean current, and effective diffusion coefficient (Martens et al., 2014).

For active particles, the passive entropic potential is replaced by a generalized entropy potential. In long tubes with slowly varying width, the effective one-dimensional density is described by

$$\frac{\partial C(x,t)}{\partial t} = D_0 \,\frac{\partial}{\partial x} \left[ A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$0

where $$\frac{\partial C(x,t)}{\partial t} = D_0 \,\frac{\partial}{\partial x} \left[ A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$1, $$\frac{\partial C(x,t)}{\partial t} = D_0 \,\frac{\partial}{\partial x} \left[ A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$2, and $$\frac{\partial C(x,t)}{\partial t} = D_0 \,\frac{\partial}{\partial x} \left[ A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$3 is the wall boundary-layer thickness. The active correction is therefore twofold: $$\frac{\partial C(x,t)}{\partial t} = D_0 \,\frac{\partial}{\partial x} \left[ A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$4 is replaced by the activity-renormalized diffusivity $$\frac{\partial C(x,t)}{\partial t} = D_0 \,\frac{\partial}{\partial x} \left[ A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$5, and the geometric width is replaced by an effective width $$\frac{\partial C(x,t)}{\partial t} = D_0 \,\frac{\partial}{\partial x} \left[ A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$6 due to wall accumulation (Zhao, 2021).

The geometry-to-operator correspondence was pushed much further in work relating the Fick–Jacobs equation to Schrödinger theory and conformal quantum mechanics. With the transformation $$\frac{\partial C(x,t)}{\partial t} = D_0 \,\frac{\partial}{\partial x} \left[ A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$7, the passive equation can be rewritten as a stationary eigenvalue problem

$$\frac{\partial C(x,t)}{\partial t} = D_0 \,\frac{\partial}{\partial x} \left[ A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$8

so the diffusion problem becomes an operator problem entirely determined by the channel geometry (Romero et al., 2012). For the family $$\frac{\partial C(x,t)}{\partial t} = D_0 \,\frac{\partial}{\partial x} \left[ A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$9, the Hamiltonian takes the conformal form

$C(x,t)$0

and a given conformal Hamiltonian is not uniquely associated with one channel, because $C(x,t)$1 and $C(x,t)$2 give the same $C(x,t)$3 (Romero et al., 2012). Related work expressed the constant-$C(x,t)$4 Fick–Jacobs equation as an exact similarity transform with effective potential

$C(x,t)$5

and then extended the construction to variable $C(x,t)$6 through the change of variable $C(x,t)$7 (Romero et al., 2011).

3. Interactions, reactions, and driven transport

One major line of generalization incorporates interactions directly into the longitudinal current. For hard-core particles in a varying-width channel, the modified stationary Fick–Jacobs equation becomes

$C(x,t)$8

where the factor $C(x,t)$9 expresses the reduction of local flux by the hole fraction $D_0$0: a particle can move only if the target site is empty (Suárez et al., 2016). In a derivation based on a lattice exclusion process and a continuum limit, the same exclusion nonlinearity appears in the reduced equation for the transversally averaged density $D_0$1,

$D_0$2

or equivalently

$D_0$3

for the local occupation $D_0$4 (Suárez et al., 2014).

A particularly transparent benchmark is a periodic chain of asymmetric exponential cavities,

$D_0$5

with $D_0$6, $D_0$7, and $D_0$8. For non-interacting particles, this geometry allows an exact stationary solution of the standard Fick–Jacobs equation. For hard-core interacting particles, the modified equation solved approximately by a Fourier expansion up to the 9th harmonic reproduces the interacting-particle concentration profiles very well, including the shift and flattening of the density maximum caused by exclusion; the agreement for current and mobility is also good, especially at higher concentrations where hard-core effects matter (Suárez et al., 2016). The same work highlights the particle-hole symmetry relation

$D_0$9

which implies $$\frac{\partial C(x,t)}{\partial t} = \frac{\partial}{\partial x} \left[ D(x)\,A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$0 at $$\frac{\partial C(x,t)}{\partial t} = \frac{\partial}{\partial x} \left[ D(x)\,A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$1 and explains why current asymmetry becomes small near half filling (Suárez et al., 2016).

Reaction and surface exchange produce a different kind of generalization. For diffusion with adsorption–desorption and bulk reaction in irregular pores delimited by $$\frac{\partial C(x,t)}{\partial t} = \frac{\partial}{\partial x} \left[ D(x)\,A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$2 and $$\frac{\partial C(x,t)}{\partial t} = \frac{\partial}{\partial x} \left[ D(x)\,A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$3, the reduced concentration

$$\frac{\partial C(x,t)}{\partial t} = \frac{\partial}{\partial x} \left[ D(x)\,A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$4

obeys

$$\frac{\partial C(x,t)}{\partial t} = \frac{\partial}{\partial x} \left[ D(x)\,A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$5

where $$\frac{\partial C(x,t)}{\partial t} = \frac{\partial}{\partial x} \left[ D(x)\,A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$6 and

$$\frac{\partial C(x,t)}{\partial t} = \frac{\partial}{\partial x} \left[ D(x)\,A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$7

is the wall arc-length density (Ledesma-Durán et al., 2016). The geometric content is therefore not exhausted by width: for adsorption–desorption, the length and the local curvature of the pore are the relevant geometric quantities for its description, because the reduced surface term is weighted by $$\frac{\partial C(x,t)}{\partial t} = \frac{\partial}{\partial x} \left[ D(x)\,A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$8 (Ledesma-Durán et al., 2016). In the sinusoidal pores studied there, the projected description is reliable with less than about 5% error when $$\frac{\partial C(x,t)}{\partial t} = \frac{\partial}{\partial x} \left[ D(x)\,A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$9 (Ledesma-Durán et al., 2016).

A further generalization arises when Brownian particles move under the competition between a constant bias and a pressure-driven incompressible flow. In that case the reduced one-dimensional equation is still of Smoluchowski type,

$$\epsilon=\epsilon_s^2\frac{D_\parallel}{D_\perp} = \frac{t_{tr}}{t_L}\ll 1$$0

but $$\epsilon=\epsilon_s^2\frac{D_\parallel}{D_\perp} = \frac{t_{tr}}{t_L}\ll 1$$1 now contains both entropic and solenoidal-force contributions (Martens et al., 2012). In thin three-dimensional channels this framework predicts a giant suppression of the effective diffusivity in the hydrodynamically enforced entropic trapping regime, by up to four orders of magnitude compared to the bulk value for the parameters explored (Martens et al., 2014).

4. Active matter, polymers, and first-passage extensions

Generalized Fick–Jacobs reductions have also become a standard device for nonequilibrium stochastic processes whose primary observables are not stationary currents but first-passage or escape statistics. For noninteracting active Brownian particles and run-and-tumble particles in long tubes of varying width, a moment expansion yields closed one-dimensional equations for the longitudinal density and polarization. At leading order, the steady-state density is Boltzmann-like in the generalized entropy potential $$\epsilon=\epsilon_s^2\frac{D_\parallel}{D_\perp} = \frac{t_{tr}}{t_L}\ll 1$$2, and the density is proportional to the effective width rather than to the bare geometric width (Zhao, 2021). This leading-order description predicts steady-state profiles and escape times from spindle chambers, while the next correction distinguishes the density potential $$\epsilon=\epsilon_s^2\frac{D_\parallel}{D_\perp} = \frac{t_{tr}}{t_L}\ll 1$$3 from the polarization potential $$\epsilon=\epsilon_s^2\frac{D_\parallel}{D_\perp} = \frac{t_{tr}}{t_L}\ll 1$$4 and thereby permits spontaneous ratchet flows in asymmetric channels (Zhao, 2021).

A time-dependent extension was developed for active Brownian particles in a pulsating three-dimensional tube with radius

$$\epsilon=\epsilon_s^2\frac{D_\parallel}{D_\perp} = \frac{t_{tr}}{t_L}\ll 1$$5

After integrating over the transverse coordinates, the effective one-dimensional equation contains an explicitly time-dependent entropic free energy

$$\epsilon=\epsilon_s^2\frac{D_\parallel}{D_\perp} = \frac{t_{tr}}{t_L}\ll 1$$6

an active drift $$\epsilon=\epsilon_s^2\frac{D_\parallel}{D_\perp} = \frac{t_{tr}}{t_L}\ll 1$$7, and a geometry-dependent diffusivity

$$\epsilon=\epsilon_s^2\frac{D_\parallel}{D_\perp} = \frac{t_{tr}}{t_L}\ll 1$$8

for the three-dimensional case (Sinha et al., 14 Sep 2025). In that system, both the average velocity and the effective diffusion coefficient show a maximum at an intermediate oscillation frequency, resembling the phenomena of stochastic resonance, while for very large $$\epsilon=\epsilon_s^2\frac{D_\parallel}{D_\perp} = \frac{t_{tr}}{t_L}\ll 1$$9 the Fick–Jacobs approximation eventually breaks down (Sinha et al., 14 Sep 2025).

For polymer translocation across a periodically corrugated channel, the generalized reduction acts on the polymer center of mass rather than on a point particle. The effective free energy

$c(z,t)$0

encodes how confinement and polymer size reduce the number of accessible transverse configurations (1901.10190). The resulting backward equations yield the first two moments of the first-passage-time distribution, and the coefficient of variation

$c(z,t)$1

never significantly drops below 1 and, in fact, can attain very large values. The paper therefore concludes that the mean first passage time alone cannot characterize the first-passage statistics exhaustively well (1901.10190).

Stochastic resetting produces another exact extension. In a narrow conical channel with radius $c(z,t)$2, resetting to a fixed position $c(z,t)$3 at rate $c(z,t)$4 modifies the reduced Fick–Jacobs equation by a sink term $c(z,t)$5 and an injection term $c(z,t)$6. Exact expressions were obtained for conditional mean first passage times, escape probabilities, and the total average lifetime in the channel, together with the criterion that resetting decreases the mean first passage time when the coefficient of variation of the underlying no-reset process satisfies $c(z,t)$7 (Jain et al., 2022). In the two-absorbing-boundary geometry, the optimal resetting rate can undergo both continuous and discontinuous transitions as relevant system parameters are varied (Jain et al., 2022).

Closely related first-passage quantities were analyzed in the form of splitting probabilities. For passive particles in a three-dimensional corrugated channel, the Fick–Jacobs reduction yields

$c(z,t)$8

and hence

$c(z,t)$9

For active Brownian motion, the corresponding reduced theory requires coupled equations for right- and left-moving states and reveals nonmonotonic dependence on corrugation amplitude when entropic and propulsive forces compete (Malgaretti et al., 2023).

5. Exact solutions, higher-order corrections, and computational formulations

One persistent theme in the generalized Fick–Jacobs literature is the search for exact or systematically improvable reductions. The exponential-cavity model already mentioned is important because it is simple enough to allow an exact stationary solution of the standard Fick–Jacobs equation for non-interacting particles,

$$\frac{\partial c(z,t)}{\partial t} = D_\parallel \frac{\partial}{\partial z} \left\{ A(z)\frac{\partial}{\partial z}\left[\frac{c(z,t)}{A(z)}\right] \right\}, \qquad A(z)=\pi r(z)^2,$$0

with constants fixed by continuity, periodicity, constant current, and fixed total particle number $$\frac{\partial c(z,t)}{\partial t} = D_\parallel \frac{\partial}{\partial z} \left\{ A(z)\frac{\partial}{\partial z}\left[\frac{c(z,t)}{A(z)}\right] \right\}, \qquad A(z)=\pi r(z)^2,$$1 (Suárez et al., 2016). This exact benchmark allows a clean comparison between the standard reduction, Monte Carlo simulations, and the modified interacting-particle theory.

The operator mappings discussed earlier also yield exact solutions for special geometries. In the conformal family $$\frac{\partial c(z,t)}{\partial t} = D_\parallel \frac{\partial}{\partial z} \left\{ A(z)\frac{\partial}{\partial z}\left[\frac{c(z,t)}{A(z)}\right] \right\}, \qquad A(z)=\pi r(z)^2,$$2, the stationary equation has Bessel-function solutions, and the concentration can be written in terms of the order $$\frac{\partial c(z,t)}{\partial t} = D_\parallel \frac{\partial}{\partial z} \left\{ A(z)\frac{\partial}{\partial z}\left[\frac{c(z,t)}{A(z)}\right] \right\}, \qquad A(z)=\pi r(z)^2,$$3 (Romero et al., 2012). In the related Schrödinger formulation with constant diffusivity, conical channels $$\frac{\partial c(z,t)}{\partial t} = D_\parallel \frac{\partial}{\partial z} \left\{ A(z)\frac{\partial}{\partial z}\left[\frac{c(z,t)}{A(z)}\right] \right\}, \qquad A(z)=\pi r(z)^2,$$4 give $$\frac{\partial c(z,t)}{\partial t} = D_\parallel \frac{\partial}{\partial z} \left\{ A(z)\frac{\partial}{\partial z}\left[\frac{c(z,t)}{A(z)}\right] \right\}, \qquad A(z)=\pi r(z)^2,$$5, exponential channels $$\frac{\partial c(z,t)}{\partial t} = D_\parallel \frac{\partial}{\partial z} \left\{ A(z)\frac{\partial}{\partial z}\left[\frac{c(z,t)}{A(z)}\right] \right\}, \qquad A(z)=\pi r(z)^2,$$6 give a constant potential $$\frac{\partial c(z,t)}{\partial t} = D_\parallel \frac{\partial}{\partial z} \left\{ A(z)\frac{\partial}{\partial z}\left[\frac{c(z,t)}{A(z)}\right] \right\}, \qquad A(z)=\pi r(z)^2,$$7, sinusoidal channels $$\frac{\partial c(z,t)}{\partial t} = D_\parallel \frac{\partial}{\partial z} \left\{ A(z)\frac{\partial}{\partial z}\left[\frac{c(z,t)}{A(z)}\right] \right\}, \qquad A(z)=\pi r(z)^2,$$8 give $$\frac{\partial c(z,t)}{\partial t} = D_\parallel \frac{\partial}{\partial z} \left\{ A(z)\frac{\partial}{\partial z}\left[\frac{c(z,t)}{A(z)}\right] \right\}, \qquad A(z)=\pi r(z)^2,$$9, and Gaussian channels $$\mathbf F(\mathbf r)= -\nabla \Phi(\mathbf r)+\nabla\times\boldsymbol\Psi(\mathbf r),$$0 reduce to a harmonic-oscillator Hamiltonian (Romero et al., 2011). These constructions do not replace the reduction assumptions; rather, they show that once the reduction is accepted, some geometries admit exact solution theory.

Higher-order asymptotics address the finite-slope corrections neglected by the classical equation. For point-size Brownian particles in a three-dimensional periodic channel with smooth corrugation parameter

$$\mathbf F(\mathbf r)= -\nabla \Phi(\mathbf r)+\nabla\times\boldsymbol\Psi(\mathbf r),$$1

a systematic expansion in powers of $$\mathbf F(\mathbf r)= -\nabla \Phi(\mathbf r)+\nabla\times\boldsymbol\Psi(\mathbf r),$$2 yields a spatially dependent diffusion coefficient

$$\mathbf F(\mathbf r)= -\nabla \Phi(\mathbf r)+\nabla\times\boldsymbol\Psi(\mathbf r),$$3

in the diffusion-dominated regime (Martens et al., 2011). The associated result

$$\mathbf F(\mathbf r)= -\nabla \Phi(\mathbf r)+\nabla\times\boldsymbol\Psi(\mathbf r),$$4

states that the average transport velocity equals the zeroth-order Fick–Jacobs current multiplied by the expectation value of the spatially dependent diffusion coefficient (Martens et al., 2011).

A conceptually different correction was obtained for symmetric two-dimensional periodic channels by abandoning the fiction of an exact one-dimensional Markov description. Instead of deriving a closed reduced diffusion equation, the analysis gives an exact Kubo-type formula for the effective diffusivity,

$$\mathbf F(\mathbf r)= -\nabla \Phi(\mathbf r)+\nabla\times\boldsymbol\Psi(\mathbf r),$$5

where $$\mathbf F(\mathbf r)= -\nabla \Phi(\mathbf r)+\nabla\times\boldsymbol\Psi(\mathbf r),$$6 solves a Laplace problem in the full channel (Mangeat et al., 2017). This formulation confirms that the Kalinay–Percus perturbative result is exact at least to $$\mathbf F(\mathbf r)= -\nabla \Phi(\mathbf r)+\nabla\times\boldsymbol\Psi(\mathbf r),$$7, and it also yields the large-$$\mathbf F(\mathbf r)= -\nabla \Phi(\mathbf r)+\nabla\times\boldsymbol\Psi(\mathbf r),$$8 asymptotic

$$\mathbf F(\mathbf r)= -\nabla \Phi(\mathbf r)+\nabla\times\boldsymbol\Psi(\mathbf r),$$9

with a universal $\mathcal F(x)$0 correction coefficient $\mathcal F(x)$1 (Mangeat et al., 2017).

Recent work has also generalized the reduction computationally. A new derivation for diffusion in variable-radius tubes expands not only the axial face fluxes but also the lateral flux around the cylindrical side surface, leading to an “expanded flux Fick–Jacobs” equation with a temporal correction, higher-order diffusion terms, and a third-derivative term $\mathcal F(x)$2 (Miksis et al., 14 Jan 2025). That formulation was extended to branched diffusion networks, together with second-order finite-difference rules at branch nodes and explicit stability conditions. In numerical tests on steep tapering, the expanded-flux model was roughly twice as accurate as classical Fick–Jacobs, and to reach a relative error tolerance of $\mathcal F(x)$3 it needed a little over 100 nodes, whereas Fick–Jacobs needed about four times as many (Miksis et al., 14 Jan 2025).

6. Validity, limitations, and methodological disputes

Despite its breadth, the generalized Fick–Jacobs approach is not a single universal equation. It is a family of reductions whose precise form depends on the dominant physics. Standard passive transport uses the accessible area or width; active systems add polarization moments, effective temperatures, and boundary-layer renormalization; interacting systems add occupancy nonlinearities; reaction problems add wall and bulk source terms; hydrodynamic transport adds a solenoidal contribution to the potential of mean force [(Zhao, 2021); (Suárez et al., 2016); (Ledesma-Durán et al., 2016); (Martens et al., 2012)]. A common misconception is that “generalized Fick–Jacobs” names one canonical correction. The literature instead shows multiple non-equivalent generalizations.

The conditions of validity are correspondingly model-dependent. Recurrent requirements are effective one-dimensional transport, fast transverse equilibration, smooth geometry, and sufficiently slow variation of the cross section. For active particles in tubes, the assumptions include $\mathcal F(x)$4, $\mathcal F(x)$5, and often $\mathcal F(x)$6 (Zhao, 2021). For pulsating three-dimensional active channels, the paper explicitly assumes $\mathcal F(x)$7 (Sinha et al., 14 Sep 2025). For weak-corrugation hydrodynamic trapping, the asymptotic expansion is controlled by

$\mathcal F(x)$8

(Martens et al., 2012). For adsorption in irregular pores, the reduced scheme is accurate in the cases studied when $\mathcal F(x)$9 (Ledesma-Durán et al., 2016).

Breakdown mechanisms are equally explicit. In asymmetric exponential cavities, the standard non-interacting Fick–Jacobs solution begins to deviate from Monte Carlo as the force increases, since strong driving worsens the assumptions behind the reduction (Suárez et al., 2016). In active channels, the leading-order generalized entropy potential cannot produce ratchet currents; those require higher-order corrections that retain boundary-tilt effects (Zhao, 2021). In the symmetric-channel dispersion problem, the projected motion along $$\frac{\partial C(x,t)}{\partial t} = D_0 \,\frac{\partial}{\partial x} \left[ A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$00 is stated not to be truly Markovian, which is precisely why an exact formula for $$\frac{\partial C(x,t)}{\partial t} = D_0 \,\frac{\partial}{\partial x} \left[ A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$01 was derived without resorting to a reduced one-dimensional diffusion equation (Mangeat et al., 2017).

The strongest methodological critique appears in the study of steady diffusion–reaction in a thin cylindrical tube with an absorbing end. There, cross-section averaging produces a reduced equation that is not closed, because it still contains an unknown wall-flux term and therefore requires an extra assumption such as a local-equilibrium factorization of the radial profile (Traytak et al., 10 Jun 2026). The exact solution shows that near the absorbing end the concentration is not spatially uniform on the cross section, but instead develops a sharp $$\frac{\partial C(x,t)}{\partial t} = D_0 \,\frac{\partial}{\partial x} \left[ A(x)\,\frac{\partial}{\partial x}\left(\frac{C(x,t)}{A(x)}\right) \right],$$02 boundary layer. On that basis, the paper argues that the classical Fick–Jacobs reduction is methodologically incomplete for this reaction-diffusion setup, and that the boundary functions method from singular perturbation theory is superior because it respects the correct boundary-layer structure and reproduces the trapping rate asymptotically (Traytak et al., 10 Jun 2026).

A broader implication is that generalized Fick–Jacobs theory works best when it is treated as a controlled reduction rather than as an automatic replacement for the full transport problem. The literature includes exact solvable cases, rigorous asymptotics, and careful numerical benchmarks, but it also shows that end layers, strong forcing, wall accumulation, nonconservative forcing, and reaction-induced inhomogeneities can all force corrections or even invalidate the closure. The approach remains useful precisely because it can be adapted to these distinct regimes, not because one fixed formula applies to them all.