Aubry Transition in Incommensurate Systems

• Aubry transition is the sharp shift in incommensurate systems from an unpinned (superlubric) state with zero static friction to a pinned state with finite friction, marked by breaking of analyticity and phonon gap opening.
• It is studied in models like the 1D Frenkel–Kontorova and extended to 2D colloidal and 3D crystalline interfaces, illustrating both continuous and first-order transitions based on dimensionality and temperature.
• Experimental realizations with trapped ions and colloidal monolayers validate the theory, offering insights into friction control and localization phenomena in nanoscale systems.

Searching arXiv for recent and foundational papers on the Aubry transition and closely related Aubry–André work. The Aubry transition denotes a sharp change between an unpinned, sliding state and a pinned state in systems with competing incommensurate length scales. In its original formulation, it arises in the one-dimensional Frenkel–Kontorova model, where increasing substrate corrugation drives an incommensurate chain from a superlubric phase with zero static friction to a pinned phase with finite static friction, accompanied by breaking of analyticity of the hull function and the opening of a phason gap (Bylinskii et al., 2015). Subsequent work has shown that the concept extends well beyond the classical zero-temperature one-dimensional setting: in two-dimensional incommensurate monolayers it becomes a first-order structural and frictional transition that survives at finite temperature and terminates at a critical point (Mandelli et al., 2017), while in quasiperiodic tight-binding models the same name is also used for the Aubry–André localization transition between extended and localized eigenstates (Lin et al., 11 Aug 2025).

1. Classical formulation in the Frenkel–Kontorova picture

The canonical setting is the Frenkel–Kontorova model, a chain of particles connected by springs and placed in a sinusoidal substrate potential. A standard form is

$$H = \sum_i \left[\frac{k}{2}(x_{i+1}-x_i-a_0)^2 + U_0 \cos\!\left(\frac{2\pi x_i}{a_s}\right) \right],$$

with $a_0$ the natural lattice spacing of the chain, $a_s$ the substrate period, and $U_0$ the corrugation amplitude (Brazda et al., 2018). The essential control parameter is the mismatch ratio

$\rho = \frac{a_0}{a_s},$

which distinguishes commensurate and incommensurate configurations (Brazda et al., 2018).

For an incommensurate chain at $T=0$, Aubry showed that increasing the substrate corrugation drives a transition between an unpinned and a pinned phase. Below a critical corrugation, the chain is in a translationally invariant sliding phase with no Peierls–Nabarro barriers and zero static friction; above it, the lattice potential overcomes the chain stiffness, the atoms reorganize toward lattice minima and avoid lattice maxima, and finite static friction appears (Bylinskii et al., 2015). In the infinite one-dimensional Frenkel–Kontorova chain, this Aubry transition is continuous and is conventionally described as a transition by breaking of analyticity of the hull function (Brazda et al., 2018).

A standard diagnostic is the disorder parameter introduced in later formulations of the Aubry problem. In one dimension it measures whether particles can access the vicinity of substrate maxima. In the sliding phase, maxima remain accessible; in the pinned phase, a neighborhood of maxima becomes strictly unoccupied (Mandelli et al., 2017). Closely related signatures are the onset of a phonon gap at $q=0$, the appearance of finite static friction, and the loss of analyticity of the hull function (Brazda et al., 2018). In the infinite incommensurate limit, the hull function becomes a nowhere-analytic fractal staircase, or Cantorus, above the transition (Bylinskii et al., 2015).

The same framework underlies the language of superlubricity. Below the transition, force cancellation across the incommensurate interface yields vanishing static friction. Above it, the interface becomes structurally pinned even though it remains incommensurate (Wang et al., 19 Jan 2025). This association between analyticity breaking and frictional locking is central to later experimental and theoretical generalizations.

2. Finite chains, quantum effects, and experimental observation in trapped ions

A direct experimental realization of Aubry physics in a finite system was achieved with chains of laser-cooled $^{174}\mathrm{Yb}^+$ ions in a linear Paul trap subject to a periodic optical lattice (Bylinskii et al., 2015). In this setting, the system is described by a generalized Frenkel–Kontorova–Tomlinson model with interatomic stiffness $g$, external stiffness $K=m\omega_0^2$, and optical lattice period $a_0$0 (Bylinskii et al., 2015). The relevant mismatch condition for a finite chain is not irrationality in the strict infinite-chain sense, but cancellation of the net lattice force on the unperturbed chain,

$a_0$1

which defines maximal mismatch and the superlubric case (Bylinskii et al., 2015).

In that experiment, the Aubry transition was identified simultaneously through static and dynamical observables. Hysteresis loops in the ion position as a function of support position reveal both the opening of position gaps and the onset of stick–slip friction. The heights of the hysteresis loops reconstruct the static gaps in the allowed ion positions, while the separation between the slipping events equals twice the static friction force required to pull the chain over the Peierls–Nabarro barrier (Bylinskii et al., 2015). Below the critical lattice depth, the chain follows the global minimum smoothly and no hysteresis is present; above it, bistability, finite static friction, and stick–slip appear (Bylinskii et al., 2015).

The same paper showed that the measured critical depths in finite mismatched chains coincide with the calculated values for the Aubry transition in an infinite golden-ratio chain with the corresponding values of $a_0$2 (Bylinskii et al., 2015). Finite size therefore rounds the transition into a symmetry-breaking crossover, but the onset of friction and analyticity breaking remains sharp enough to define an operational critical depth.

Quantum effects in the Aubry transition were later analyzed for trapped ions in an optical lattice using path-integral Monte Carlo (Bonetti et al., 2020). There the system is a finite chain of ions with long-range Coulomb repulsion in a harmonic trap plus a periodic optical potential,

$a_0$3

Quantum tunneling modifies the finite-chain Aubry crossover and provides new signatures that are robust against thermal and finite-size effects (Bonetti et al., 2020). In particular, the Binder cumulant of the central ion’s position distribution and the total energy difference $a_0$4 display sharp changes near the crossover, and the hull function develops gaps as pinning sets in (Bonetti et al., 2020). This work emphasizes that the universality class of the quantum Aubry transition remains an open question (Bonetti et al., 2020).

3. Two-dimensional generalization: finite-temperature structural and frictional transition

A major extension of Aubry physics concerns two-dimensional incommensurate interfaces. Using a model inspired by colloid monolayers in an optical lattice, simulations showed that a sharp Aubry transition persists in two dimensions at finite temperature, but with properties fundamentally different from the one-dimensional case (Mandelli et al., 2017). The model consists of classical particles interacting via a screened Coulomb potential,

$a_0$5

on a triangular substrate potential

$a_0$6

with overdamped Langevin dynamics at temperature $a_0$7 (Mandelli et al., 2017).

In this system, incommensurability arises from both lattice spacing mismatch and rotational misalignment. The weak-coupling equilibrium configuration is characterized by a nonzero Novaco–McTague angle, and the mismatch produces a moiré superlattice made of nearly commensurate domains separated by a honeycomb network of soliton lines (Mandelli et al., 2017). The two-dimensional Aubry transition occurs when increasing corrugation causes the local moiré domains to rotate from the Novaco–McTague angle toward near alignment with the substrate and to become locally commensurate (Mandelli et al., 2017).

Several observables establish the transition. A two-dimensional disorder parameter,

$a_0$8

counts the fraction of particles in a geometrically defined region around substrate maxima where the potential exceeds the saddle-point value. In the unpinned phase, particles occupy these high-energy regions with finite probability; in the pinned phase, those regions become effectively forbidden and $a_0$9 drops sharply (Mandelli et al., 2017). A local orientation angle,

$a_s$0

shows a simultaneous drop from the Novaco–McTague angle to nearly zero (Mandelli et al., 2017).

Unlike the $a_s$1 Aubry transition, the two-dimensional transition is first order. Simulations show discontinuous jumps in $a_s$2, $a_s$3, substrate energy, and interparticle energy, together with hysteresis and a finite coexistence region (Mandelli et al., 2017). The free-energy balance is written as

$a_s$4

and the strongly oblique first-order line in the $a_s$5–$a_s$6 plane implies that the unpinned phase has larger entropy (Mandelli et al., 2017). This first-order Aubry line terminates at a finite-temperature critical point,

$a_s$7

marked by a susceptibility peak in the substrate-energy fluctuations (Mandelli et al., 2017).

This scenario was then realized experimentally in a two-dimensional colloidal monolayer on an optical lattice at room temperature (Brazda et al., 2018). The colloidal monolayer, with mismatch ratio $a_s$8, exhibits a sharp drop in mobility at a critical corrugation $a_s$9, together with the onset of finite static friction (Brazda et al., 2018). Structural observables such as the fraction of tilted particles and a disorder parameter defined from occupation of the repulsive regions of the substrate unit cell show the same transition from tilted, superlubric domains to aligned, pinned domains (Brazda et al., 2018). The coexistence of pinned and unpinned regions, together with hysteresis in simulations, identifies the transition as first order in $U_0$0 (Brazda et al., 2018).

4. Friction, superlubricity, and load-induced locking in realistic interfaces

From its origin, the Aubry transition has been inseparable from friction. In the Frenkel–Kontorova picture, the sliding phase has vanishing static friction because no Peierls–Nabarro barriers exist, whereas the pinned phase exhibits a finite depinning threshold (Bylinskii et al., 2015). In the ion-chain experiment, the onset of friction and the opening of position gaps were shown to be the dynamic and static aspects, respectively, of the same Aubry transition (Bylinskii et al., 2015). In the two-dimensional colloidal monolayer, the Aubry line likewise separates a superlubric regime with $U_0$1 from a pinned regime where static friction rises rapidly with corrugation, and the frictional upswing disappears as temperature approaches the finite-temperature critical point (Mandelli et al., 2017).

Recent work has pushed this connection into fully three-dimensional crystal interfaces. Simulations of Au(111) twist grain boundaries showed that the load-free moiré is smooth and superlubric at incommensurate twists, but increasing load induces a first-order structural transformation in which the highest-energy AA moiré nodes are removed and replaced by commensurate mini-domains (Wang et al., 19 Jan 2025). This is described as an Aubry-type transition because the highest-energy local configurations disappear, in direct analogy with the elimination of particles at substrate maxima in the classical disorder-parameter picture (Wang et al., 19 Jan 2025).

The corresponding Landau description uses the mini-domain size $U_0$2 as an Aubry order parameter and writes a free energy per moiré cell,

$U_0$3

Under load, this produces a first-order twist–load transition line terminating at a critical point (Wang et al., 19 Jan 2025). Frictionally, the transformation causes a superlubric-locked transition with a huge friction jump, stick–slip, and irreversible plastic flow (Wang et al., 19 Jan 2025). That study makes explicit that Aubry pinning in realistic crystal interfaces can become so strong that sliding relocates from the grain boundary to a nearby pristine crystal plane (Wang et al., 19 Jan 2025).

A distinct but conceptually related realization appears in an active mechanical setting. A snake-like robot undulating through a channel with a bichromatic array of hemispherical obstacles experiences a drag landscape approximating a one-dimensional Aubry–André potential. When the landscape is strictly periodic, the robot traverses the channel ballistically; when the landscape is sufficiently aperiodic, it becomes trapped and fails to exit (Pierce et al., 4 Jul 2025). The average travel distance, the mean-squared displacement exponent, and the distribution of travel distances reproduce the change from extended to localized transport familiar from the Aubry–André transition (Pierce et al., 4 Jul 2025). Although this is not the classical pinned–sliding Aubry transition of the Frenkel–Kontorova type, it illustrates the broader principle that incommensurate structure plus sufficient amplitude can suppress transport (Pierce et al., 4 Jul 2025).

5. Aubry–André localization transition and its generalizations

In the spectral-theory and condensed-matter literature, “Aubry transition” often refers to the localization transition of the Aubry–André or Aubry–André–Harper model. The canonical Hermitian model,

$U_0$4

exhibits a sharp metal–insulator transition controlled by Aubry–André self-duality (Lin et al., 11 Aug 2025). For irrational $U_0$5, all eigenstates are extended for $U_0$6, all are localized for $U_0$7, and all are critical at $U_0$8 (Lin et al., 11 Aug 2025). Equivalent conventions in other papers place the critical point at $U_0$9 when the Hamiltonian is written with hopping $\rho = \frac{a_0}{a_s},$0 and quasiperiodic potential amplitude $\rho = \frac{a_0}{a_s},$1 (Romito et al., 2018).

This localization transition has several generalizations in the supplied literature. Periodic modulation of the Aubry–André potential produces a Floquet localization–delocalization transition controlled by drive frequency and amplitude. In the noninteracting driven model,

$\rho = \frac{a_0}{a_s},$2

delocalization occurs at sufficiently low frequency and sufficiently large amplitude, with a low-frequency threshold

$\rho = \frac{a_0}{a_s},$3

and a characteristic frequency scale $\rho = \frac{a_0}{a_s},$4 (Romito et al., 2018).

Non-Hermitian generalizations alter both spectral and dynamical critical behavior. For the non-Hermitian Aubry–André–Harper model with asymmetric hopping,

$\rho = \frac{a_0}{a_s},$5

the critical potential is $\rho = \frac{a_0}{a_s},$6, and the localization–delocalization transition becomes discontinuous not only in the diffusion exponent but also in the ballistic velocity (Longhi, 2021). In the $\rho = \frac{a_0}{a_s},$7-symmetric non-Hermitian AAH model with complex onsite potential,

$\rho = \frac{a_0}{a_s},$8

the metal–insulator transition at $\rho = \frac{a_0}{a_s},$9 coincides with $T=0$0-symmetry breaking and a topological change in the complex spectrum; the localization length in the insulating phase remains energy-independent,

$T=0$1

exactly as in the Hermitian Aubry–André model (Longhi, 2019).

Further extensions include disorder and hybrid universality. In the non-Hermitian disordered Aubry–André model with nonreciprocal hopping and random onsite disorder, both the quasiperiodic control parameter $T=0$2 and the disorder strength $T=0$3 are relevant directions at the non-Hermitian AA critical point, leading to a new universality class and a hybrid scaling law in the overlap between Aubry–André and Anderson critical regions (Sun et al., 2024). The non-Hermitian interpolating Aubry–André–Fibonacci chain exhibits a cascade of delocalization transitions as the potential is deformed from the Aubry–André limit toward the Fibonacci limit, with self-similar critical modes and only a few plateaux surviving under nonreciprocal hopping (Zhai et al., 2021).

The many-particle ground-state version of the Aubry–André transition also acquires a richer arithmetic structure. For the interacting Aubry–André model, the critical behavior depends on a Diophantine equation relating the incommensurate frequency and the filling fraction,

$T=0$4

which generalizes the dependence of the single-particle critical properties on the continued-fraction expansion of the incommensurate frequency (Cookmeyer et al., 2020). Numerical evidence suggests that nearest-neighbor interactions may be irrelevant at at least some of these critical points, so the Diophantine classification can survive in the interacting case (Cookmeyer et al., 2020).

6. Open directions, reinterpretations, and broader significance

Several papers recast the Aubry transition by changing the microscopic setting while preserving the central competition between incommensurate structure and coupling. One such direction concerns the shape of the substrate potential itself. In a modified Frenkel–Kontorova model with periodic potential

$T=0$5

the Aubry transition can occur with very small structural distortions if the local substrate force becomes very strong in narrow regions while the potential remains weak elsewhere (Cépas et al., 2024). In that model, the critical coupling $T=0$6 tends to zero as $T=0$7, the phason gap opens in the pinned phase, yet the high-energy part of the phonon spectrum remains close to that of an undistorted chain (Cépas et al., 2024). This suggests that pinned incommensurate phases with small observable distortions need not contradict an Aubry mechanism (Cépas et al., 2024).

Another active direction concerns dissipation and nonequilibrium open-system transport. In the Aubry–André–Harper chain coupled to a non-Markovian bath, the role of bath memory depends qualitatively on the side of the localization transition (Suárez et al., 21 May 2026). In the extended phase, memory reshapes the dynamical generator and produces transport patterns that cannot be reduced to a simple rescaling of time, whereas in the localized phase the bath activates motion between localized states and memory mainly renormalizes the dynamical timescales (Suárez et al., 21 May 2026). This suggests that localization acts as a filter of non-Markovian effects (Suárez et al., 21 May 2026).

Across these formulations, several points recur. First, the Aubry transition is not restricted to a single model, but identifies a class of incommensurate pinning or localization phenomena controlled by the competition of length scales and couplings. Second, dimensionality matters: the one-dimensional pinned–sliding transition is continuous at $T=0$8, whereas realistic two-dimensional interfaces can exhibit first-order, finite-temperature transitions with coexistence and critical endpoints (Mandelli et al., 2017). Third, the transition is inseparable from transport: it manifests as the onset of static friction in tribology and as localization in quasiperiodic spectral problems. Fourth, structural degrees of freedom absent in the original Frenkel–Kontorova chain—rotation of moiré domains, twist elasticity, superconducting pairing, non-Hermiticity, or bath memory—do not abolish Aubry physics but can profoundly alter its order, universality, and observables.

In that sense, the Aubry transition remains both a specific phenomenon and a unifying concept. In one branch of the literature it describes the passage from superlubric sliding to structural pinning at incommensurate interfaces (Bylinskii et al., 2015); in another it denotes the self-duality-driven localization transition of quasiperiodic quantum lattices (Lin et al., 11 Aug 2025). The supplied work shows that both usages continue to expand, into two-dimensional colloids at ambient temperature (Brazda et al., 2018), three-dimensional crystal grain boundaries under load (Wang et al., 19 Jan 2025), active mechanical transport (Pierce et al., 4 Jul 2025), and non-Hermitian, interacting, driven, and dissipative quasiperiodic systems (Sun et al., 2024).