Weak Localization of Bosons

• Weak localization of bosons is an interference phenomenon in disordered systems where phase coherence and reversed-path interference lead to observable effects like logarithmic temperature corrections in resistance.
• Experimental studies on ultrathin superconducting films reveal a distinct R ∼ ln T behavior, distinguishing bosonic localization from traditional fermionic conductance corrections.
• Microscopic transport theories and numerical models confirm that disorder, interaction-induced dephasing, and platform-specific diagnostics (such as coherent backscattering and fragmentation) are key to understanding bosonic weak localization.

Searching arXiv for the cited works to ground the article in current records. Weak localization of bosons denotes interference-driven localization phenomena in disordered bosonic systems, with manifestations that depend strongly on the platform and observable. In quasi-two-dimensional superconducting films, the localized carriers can be Cooper pairs of charge $2e$, and the characteristic low-temperature correction appears as a logarithm in the resistance, $R(T)=R_0+A_b\ln(T/T_0)$, rather than the familiar fermionic correction in the conductance (Lin et al., 2010). In disordered cold-atom settings, the same interference physics is formulated through crossed diagrams, coherent backscattering, and disorder-averaged transport equations, while in quasiperiodic lattice systems it appears through localization diagnostics such as the inverse participation ratio, superfluid fraction, participation ratio, and decay of one-body correlations (Geiger et al., 2013, Ray et al., 2015, Gonzalez-Garcia et al., 2016, Chakrabarti et al., 20 Feb 2025). Across these realizations, weak localization is distinguished from purely classical trapping by its dependence on phase coherence, reversed-path interference, and disorder-controlled crossover or transition scales.

1. Conceptual basis and distinction from fermionic weak localization

In the conventional two-dimensional fermionic case, weak localization gives a correction to the sheet conductance of the form

$G(T)=G_0-\frac{e^2}{2\pi^2\hbar}\ln(T/T_0),$

or more generally

$G(T)=G_0+A\ln T,$

with $A<0$ set by electron-electron interaction and spin-orbit coupling in the Abrahams-Aronov framework (Lin et al., 2010). Das and Doniach predicted that a dilute “gas” of Cooper pairs subject to disorder and a perpendicular field should instead show a weak-localization correction appearing directly in the resistance,

$R(T)=R_0+A_b\ln(T/T_0),$

with $A_b>0$ arising from interference of bosonic paths; in the strong-field bosonic regime one also expects a field dependence of $A_b(B)$ and a localization length $\xi_b$ set by pair size and disorder (Lin et al., 2010).

In microscopic transport theory for interacting bosons in a weak random potential, the disorder-averaged dynamics can be reduced to a nonlinear integral equation of Boltzmann type for the single-particle diffusive flux. The relevant building blocks are the propagation kernel $P_E$, an elastic collision kernel $R(T)=R_0+A_b\ln(T/T_0)$0, and an inelastic collision kernel $R(T)=R_0+A_b\ln(T/T_0)$1, while weak localization enters through crossed, or Cooperon, corrections rather than the ladder sector that describes incoherent diffusion (Geiger et al., 2013). In a complementary semiclassical treatment of nonlinear bosonic matter waves in billiard geometries, ladder pairings produce no coherent backscattering, whereas crossed pairings of a trajectory and its time reverse generate the weak-localization peak (Hartmann et al., 2011).

These formulations establish a central distinction: for bosons, weak localization is not tied to a single universal observable. In superconducting films it appears as $R(T)=R_0+A_b\ln(T/T_0)$2; in disordered matter-wave transport it appears in the coherent backscattering cone; and in lattice realizations it is tracked by fragmentation and localization diagnostics. This suggests that “weak localization of bosons” is best understood as a family of interference phenomena rather than a single transport law.

2. Quasi-two-dimensional superconducting films and the $R(T)=R_0+A_b\ln(T/T_0)$3 signature

A direct experimental realization was reported for ultrathin amorphous Bi films grown by in-situ quench-condensation at $R(T)=R_0+A_b\ln(T/T_0)$4 under UHV on a polished $R(T)=R_0+A_b\ln(T/T_0)$5 SrTiO$R(T)=R_0+A_b\ln(T/T_0)$6 crystal with ex-situ deposited sputtered Pt four-probe contacts of $R(T)=R_0+A_b\ln(T/T_0)$7. A $R(T)=R_0+A_b\ln(T/T_0)$8 amorphous Sb wetting underlayer was deposited first, followed by amorphous Bi to a total thickness of $R(T)=R_0+A_b\ln(T/T_0)$9. From the Hall-effect value $G(T)=G_0-\frac{e^2}{2\pi^2\hbar}\ln(T/T_0),$0, one obtains $G(T)=G_0-\frac{e^2}{2\pi^2\hbar}\ln(T/T_0),$1 and $G(T)=G_0-\frac{e^2}{2\pi^2\hbar}\ln(T/T_0),$2, with normal-state sheet resistance just above the superconducting transition $G(T)=G_0-\frac{e^2}{2\pi^2\hbar}\ln(T/T_0),$3. Measurements were performed in DC four-probe geometry in a dilution refrigerator down to $G(T)=G_0-\frac{e^2}{2\pi^2\hbar}\ln(T/T_0),$4, with perpendicular field up to $G(T)=G_0-\frac{e^2}{2\pi^2\hbar}\ln(T/T_0),$5 and extensive lead filtering to minimize electromagnetic noise (Lin et al., 2010).

The phase identification was operational. Superconducting behavior was defined by $G(T)=G_0-\frac{e^2}{2\pi^2\hbar}\ln(T/T_0),$6 at the lowest $G(T)=G_0-\frac{e^2}{2\pi^2\hbar}\ln(T/T_0),$7 with $G(T)=G_0-\frac{e^2}{2\pi^2\hbar}\ln(T/T_0),$8; insulating behavior by $G(T)=G_0-\frac{e^2}{2\pi^2\hbar}\ln(T/T_0),$9 at the lowest $G(T)=G_0+A\ln T,$0 with $G(T)=G_0+A\ln T,$1; and a mixed or nonmonotonic regime by a resistance maximum at $G(T)=G_0+A\ln T,$2 followed by a low-temperature downturn, interpreted as superconducting islands in an insulating matrix (Lin et al., 2010).

At zero field and for $G(T)=G_0+A\ln T,$3, $G(T)=G_0+A\ln T,$4, the sheet conductance obeys

$G(T)=G_0+A\ln T,$5

with $G(T)=G_0+A\ln T,$6, in quantitative agreement with weak-localization plus electron-electron interaction theory for strong spin-orbit scattering (Lin et al., 2010). In the intermediate temperature window $G(T)=G_0+A\ln T,$7 and fields $G(T)=G_0+A\ln T,$8, the sheet resistance collapses under Fisher’s quantum-critical form,

$G(T)=G_0+A\ln T,$9

with $A<0$0, $A<0$1, and $A<0$2. This scaling breaks down for $A<0$3, where $A<0$4 becomes strongly nonmonotonic (Lin et al., 2010).

Deep on the insulating side, for $A<0$5 and $A<0$6, the key result is that the resistance itself, not the conductance, is a clear logarithm of temperature over $A<0$7–$A<0$8 and up to $A<0$9: $R(T)=R_0+A_b\ln(T/T_0),$0 At $R(T)=R_0+A_b\ln(T/T_0),$1, the best fit gives $R(T)=R_0+A_b\ln(T/T_0),$2 per $R(T)=R_0+A_b\ln(T/T_0),$3 decade with $R(T)=R_0+A_b\ln(T/T_0),$4, whereas fitting $R(T)=R_0+A_b\ln(T/T_0),$5 gives $R(T)=R_0+A_b\ln(T/T_0),$6. Converting to bulk resistivity using film thickness $R(T)=R_0+A_b\ln(T/T_0),$7 yields $R(T)=R_0+A_b\ln(T/T_0),$8 per $R(T)=R_0+A_b\ln(T/T_0),$9, close to the value $A_b>0$0 reported for underdoped cuprates. By contrast, forcing $A_b>0$1 in this regime yields $A_b>0$2, about three times the zero-field value and therefore inconsistent with electronic quantum corrections (Lin et al., 2010).

The interpretation is specific: $A_b>0$3, rather than $A_b>0$4, with a coefficient of order $A_b>0$5 over a broad field and temperature range, is taken as the hallmark of bosonic weak localization. In this picture, the localized carriers are Cooper pairs, and their interference produces a positive back-reflection correction to conductance, which appears experimentally as increasing $A_b>0$6 on cooling. The absence of a large magnetoresistance peak was noted to argue against simple boson-pair hopping or Josephson-junction-array physics and to suggest weakly localized bosons rather than a true Bose glass (Lin et al., 2010).

3. Microscopic transport theory and coherent backscattering of interacting bosons

A microscopic scattering theory for interacting bosons in a three-dimensional weakly disordered potential begins from

$A_b>0$7

with $A_b>0$8, $A_b>0$9 in units $A_b(B)$0, a Gaussian white-noise disorder potential satisfying $A_b(B)$1, and a two-body interaction $A_b(B)$2 (Geiger et al., 2013). The disorder-averaged single-particle Green function is diagonal in momentum,

$A_b(B)$3

where $A_b(B)$4 is the disorder mean free path (Geiger et al., 2013). At low energy, the two-body $A_b(B)$5-matrix satisfies

$A_b(B)$6

with $A_b(B)$7 the $A_b(B)$8-wave scattering length (Geiger et al., 2013).

After disorder averaging and tracing over $A_b(B)$9 atoms, the transport problem is expressed in terms of a propagation kernel

$\xi_b$0

together with elastic and inelastic collision kernels $\xi_b$1 and $\xi_b$2. The resulting nonlinear transport equation for the disorder-averaged spectral density $\xi_b$3 contains an incoming source term, the propagation term $\xi_b$4, and quadratic collision terms in $\xi_b$5 built from $\xi_b$6 and $\xi_b$7 (Geiger et al., 2013). The formalism reduces to the Gross-Pitaevskii mean-field description when only elastic collisions are retained, but even weak interactions generate inelastic energy redistribution and drive thermalization of the single-particle current (Geiger et al., 2013).

Weak localization is encoded in the crossed sector. The crossed propagator is

$\xi_b$8

with $\xi_b$9, and the crossed density $P_E$0 obeys coupled integral equations whose solution determines the interference part of the backscattered flux (Geiger et al., 2013). In slab far field, the coherent backscattering contribution takes the form

$P_E$1

with static Cooperon

$P_E$2

This is the explicit bosonic counterpart of weak-localization interference in the language of coherent backscattering (Geiger et al., 2013).

The inelastic channel modifies the coherent backscattering cone in a highly structured way. Inelastic scattering broadens and changes the amplitude of the cone in a narrow spectral window around $P_E$3, with width $P_E$4 and enhancement greater than $P_E$5 at the peak energy. At $P_E$6, purely elastic nonlinearity yields a crossover from constructive to destructive coherent backscattering as the dephasing parameter $P_E$7 grows, whereas inelastic collisions slow the suppression of the cone and leave a residual enhancement due to many-wave interference in the inelastic channel (Geiger et al., 2013). This makes clear that interactions do not merely wash out bosonic interference; they can also reshape it spectrally.

4. Nonlinear bosonic matter waves and dephasing of weak localization

For dilute Bose-Einstein condensates propagating through irregularly shaped billiards connected to waveguides, the relevant mean-field equation is the inhomogeneous Gross-Pitaevskii equation

$P_E$8

where $P_E$9 is an artificial vector potential breaking time-reversal symmetry, $R(T)=R_0+A_b\ln(T/T_0)$00 is the hard-wall billiard potential, and $R(T)=R_0+A_b\ln(T/T_0)$01 is the effective two-dimensional interaction strength (Hartmann et al., 2011). For a monochromatic source $R(T)=R_0+A_b\ln(T/T_0)$02 and stationary ansatz $R(T)=R_0+A_b\ln(T/T_0)$03, one obtains the stationary nonlinear equation

$R(T)=R_0+A_b\ln(T/T_0)$04

with $R(T)=R_0+A_b\ln(T/T_0)$05 (Hartmann et al., 2011).

The linear retarded Green function admits the Van Vleck form

$R(T)=R_0+A_b\ln(T/T_0)$06

where $R(T)=R_0+A_b\ln(T/T_0)$07 labels classical trajectories and $R(T)=R_0+A_b\ln(T/T_0)$08 is the action (Hartmann et al., 2011). A weak vector potential contributes a dephasing phase $R(T)=R_0+A_b\ln(T/T_0)$09, while weak nonlinearity leads, after self-consistent iteration restricted to ladder pairings, to a dressed Green function with an additional phase

$R(T)=R_0+A_b\ln(T/T_0)$10

In the diagrammatic interpretation, ladder pairings shift the density uniformly and do not produce weak localization, whereas crossed pairings of $R(T)=R_0+A_b\ln(T/T_0)$11 and its time reverse $R(T)=R_0+A_b\ln(T/T_0)$12 generate the coherent-backscattering correction (Hartmann et al., 2011).

The resulting weak-localization correction for retroreflection in the incident channel $R(T)=R_0+A_b\ln(T/T_0)$13 is

$R(T)=R_0+A_b\ln(T/T_0)$14

with

$R(T)=R_0+A_b\ln(T/T_0)$15

Equivalently, introducing the nonlinearity-induced dephasing rate $R(T)=R_0+A_b\ln(T/T_0)$16, one may write

$R(T)=R_0+A_b\ln(T/T_0)$17

For total reflection and transmission,

$R(T)=R_0+A_b\ln(T/T_0)$18

or, in Lorentzian form,

$R(T)=R_0+A_b\ln(T/T_0)$19

$R(T)=R_0+A_b\ln(T/T_0)$20

where $R(T)=R_0+A_b\ln(T/T_0)$21 sets the magnetic width and $R(T)=R_0+A_b\ln(T/T_0)$22 sets the nonlinear dephasing scale (Hartmann et al., 2011).

Numerically, for hard-wall chaotic billiards with $R(T)=R_0+A_b\ln(T/T_0)$23 at $R(T)=R_0+A_b\ln(T/T_0)$24, finite-difference discretization at roughly $R(T)=R_0+A_b\ln(T/T_0)$25 points per de Broglie wavelength, and nonlinearity confined to the cavity, the computed $R(T)=R_0+A_b\ln(T/T_0)$26 and $R(T)=R_0+A_b\ln(T/T_0)$27 for $R(T)=R_0+A_b\ln(T/T_0)$28 show Lorentzian weak-localization peaks whose height scales as $R(T)=R_0+A_b\ln(T/T_0)$29 (Hartmann et al., 2011). A further nuance is that an apparent weak antilocalization regime can arise from short self-retracing paths, which form a non-universal reflection background and reduce the zero-field peak. In this formulation, the effect is attributed not to a universal sign reversal of bosonic interference, but to omitted short-path contributions in the Cooperon sum (Hartmann et al., 2011).

5. Quasiperiodic lattices, fragmentation, and localization diagnostics

In one-dimensional quasiperiodic lattices, the weakly interacting Bose gas is often modeled by the Bose-Hubbard Hamiltonian with Aubry-André onsite modulation,

$R(T)=R_0+A_b\ln(T/T_0)$30

with $R(T)=R_0+A_b\ln(T/T_0)$31 and, in the weakly interacting high-filling regime, a condensate order parameter $R(T)=R_0+A_b\ln(T/T_0)$32 governed by the discrete nonlinear Schrödinger equation (Ray et al., 2015). Two standard diagnostics are the inverse participation ratio,

$R(T)=R_0+A_b\ln(T/T_0)$33

and the superfluid fraction

$R(T)=R_0+A_b\ln(T/T_0)$34

For $R(T)=R_0+A_b\ln(T/T_0)$35, the model is self-dual at $R(T)=R_0+A_b\ln(T/T_0)$36: all states are extended for $R(T)=R_0+A_b\ln(T/T_0)$37 and exponentially localized for $R(T)=R_0+A_b\ln(T/T_0)$38. With weak repulsion, the IPR becomes appreciable at $R(T)=R_0+A_b\ln(T/T_0)$39 but rises only gradually because the condensate fragments over several quasi-degenerate minima, a phenomenon described as multisite localization; at $R(T)=R_0+A_b\ln(T/T_0)$40, $R(T)=R_0+A_b\ln(T/T_0)$41 drops to zero only for $R(T)=R_0+A_b\ln(T/T_0)$42 (Ray et al., 2015). In the associated classical dynamical-map formulation, the largest Lyapunov exponent equals the inverse localization length, and in the non-interacting Aubry-André model $R(T)=R_0+A_b\ln(T/T_0)$43 (Ray et al., 2015). Bogoliubov-de Gennes analysis further shows that the quasiparticle spectrum develops gaps and a devil’s-staircase structure, the non-condensate density itself localizes in a multisite pattern, and the condensate fraction drops sharply around $R(T)=R_0+A_b\ln(T/T_0)$44 (Ray et al., 2015).

A related few-body treatment of a finite one-dimensional quasiperiodic lattice uses

$R(T)=R_0+A_b\ln(T/T_0)$45

with $R(T)=R_0+A_b\ln(T/T_0)$46, $R(T)=R_0+A_b\ln(T/T_0)$47, $R(T)=R_0+A_b\ln(T/T_0)$48, and weak interaction $R(T)=R_0+A_b\ln(T/T_0)$49 (Chakrabarti et al., 20 Feb 2025). Localization is characterized through the reduced one-body density matrix $R(T)=R_0+A_b\ln(T/T_0)$50, its natural orbitals, the fragmentation parameter $R(T)=R_0+A_b\ln(T/T_0)$51, the order parameter $R(T)=R_0+A_b\ln(T/T_0)$52, the participation ratio $R(T)=R_0+A_b\ln(T/T_0)$53, and the inverse localization length extracted from the asymptotic decay of $R(T)=R_0+A_b\ln(T/T_0)$54 (Chakrabarti et al., 20 Feb 2025). In the $R(T)=R_0+A_b\ln(T/T_0)$55 limit, the sharp Aubry-André transition occurs at $R(T)=R_0+A_b\ln(T/T_0)$56, with

$R(T)=R_0+A_b\ln(T/T_0)$57

Weak repulsion slightly increases the localization length through $R(T)=R_0+A_b\ln(T/T_0)$58, $R(T)=R_0+A_b\ln(T/T_0)$59, and for $R(T)=R_0+A_b\ln(T/T_0)$60 the leading occupation $R(T)=R_0+A_b\ln(T/T_0)$61 remains above $R(T)=R_0+A_b\ln(T/T_0)$62 up to $R(T)=R_0+A_b\ln(T/T_0)$63, while for $R(T)=R_0+A_b\ln(T/T_0)$64 the localization length agrees with $R(T)=R_0+A_b\ln(T/T_0)$65 within a few percent (Chakrabarti et al., 20 Feb 2025).

In two-dimensional optical lattices with uncorrelated on-site disorder $R(T)=R_0+A_b\ln(T/T_0)$66, mean-field computational experiments quantify localization through the inverse participation ratio,

$R(T)=R_0+A_b\ln(T/T_0)$67

the density variance $R(T)=R_0+A_b\ln(T/T_0)$68, and the peak-fraction $R(T)=R_0+A_b\ln(T/T_0)$69 defined by a density threshold $R(T)=R_0+A_b\ln(T/T_0)$70 (Gonzalez-Garcia et al., 2016). For square and triangular lattices with $R(T)=R_0+A_b\ln(T/T_0)$71 and $R(T)=R_0+A_b\ln(T/T_0)$72, weak disorder $R(T)=R_0+A_b\ln(T/T_0)$73 leaves the condensate nearly uniform, while above $R(T)=R_0+A_b\ln(T/T_0)$74 a smooth fragmentation begins: IPR grows exponentially in $R(T)=R_0+A_b\ln(T/T_0)$75, $R(T)=R_0+A_b\ln(T/T_0)$76 decreases, and $R(T)=R_0+A_b\ln(T/T_0)$77 rises. The interval $R(T)=R_0+A_b\ln(T/T_0)$78–$R(T)=R_0+A_b\ln(T/T_0)$79 is described as “weak-localisation,” while for $R(T)=R_0+A_b\ln(T/T_0)$80 the condensate has effectively collapsed onto a few isolated islands (Gonzalez-Garcia et al., 2016). Honeycomb lattices behave differently because the effective tunnelling is larger at the same lattice depth: for $R(T)=R_0+A_b\ln(T/T_0)$81, no significant fragmentation is found up to $R(T)=R_0+A_b\ln(T/T_0)$82, whereas increasing $R(T)=R_0+A_b\ln(T/T_0)$83 produces a crossover at $R(T)=R_0+A_b\ln(T/T_0)$84 (Gonzalez-Garcia et al., 2016). The disorder-averaged momentum-space energy spectrum simultaneously evolves from sharp Bragg peaks at $R(T)=R_0+A_b\ln(T/T_0)$85 to a dense, almost flat spectrum at strong disorder (Gonzalez-Garcia et al., 2016).

System	Control parameter	Localization signature
1D Aubry-André condensate	$R(T)=R_0+A_b\ln(T/T_0)$86	IPR rise, $R(T)=R_0+A_b\ln(T/T_0)$87, multisite localization (Ray et al., 2015)
1D finite few-boson quasiperiodic lattice	$R(T)=R_0+A_b\ln(T/T_0)$88	$R(T)=R_0+A_b\ln(T/T_0)$89 reduction, exponential decay of $R(T)=R_0+A_b\ln(T/T_0)$90, $R(T)=R_0+A_b\ln(T/T_0)$91 (Chakrabarti et al., 20 Feb 2025)
2D disordered optical lattices	$R(T)=R_0+A_b\ln(T/T_0)$92 for square/triangular	IPR growth, $R(T)=R_0+A_b\ln(T/T_0)$93 increase, $R(T)=R_0+A_b\ln(T/T_0)$94 decrease (Gonzalez-Garcia et al., 2016)

These lattice studies do not use the same transport observable as the quasi-two-dimensional film work. Instead, they characterize weak localization through condensate fragmentation, loss of superfluid stiffness, finite-size crossover scales, and the spatial structure of the dominant orbital. A plausible implication is that, in bosonic systems, the operational meaning of weak localization is platform-specific even when the underlying interference physics is common.

6. Superconductor-insulator transitions, percolation, and interpretive issues

The quasi-two-dimensional film results place weak localization of bosons directly in the broader problem of magnetic-field-tuned superconductor-insulator transitions. In the intermediate regime, finite-size scaling with $R(T)=R_0+A_b\ln(T/T_0)$95, $R(T)=R_0+A_b\ln(T/T_0)$96, and $R(T)=R_0+A_b\ln(T/T_0)$97 holds only for $R(T)=R_0+A_b\ln(T/T_0)$98; below $R(T)=R_0+A_b\ln(T/T_0)$99, the strongly nonmonotonic $G(T)=G_0-\frac{e^2}{2\pi^2\hbar}\ln(T/T_0),$00 indicates a mixed regime rather than a direct continuous quantum critical point (Lin et al., 2010). The interpretation given is that, at finite temperature, the film separates into superconducting regions embedded in an insulating matrix, forming a percolating network. In the $G(T)=G_0-\frac{e^2}{2\pi^2\hbar}\ln(T/T_0),$01 limit, the system ends in one of two pure phases—superconducting or bosonic-localized insulator—depending on field, which is consistent with a percolative quantum phase transition rather than a direct $G(T)=G_0-\frac{e^2}{2\pi^2\hbar}\ln(T/T_0),$02-dimensional XY critical point (Lin et al., 2010).

The small effective exponent $G(T)=G_0-\frac{e^2}{2\pi^2\hbar}\ln(T/T_0),$03 was explicitly compared with Monte Carlo studies of unfrustrated Josephson-junction arrays at rational frustration $G(T)=G_0-\frac{e^2}{2\pi^2\hbar}\ln(T/T_0),$04, where a first-order-like jump to a phase mixture is predicted for some parameters (Lin et al., 2010). This comparison does not equate the systems, but it places the observed indirect transition route within a broader class of bosonic models where coexistence and percolation intervene between idealized critical regimes.

Several recurrent misconceptions are addressed by the cited literature. First, weak localization of bosons is not equivalent to any insulating bosonic phase: in the film work, the absence of a large magnetoresistance peak was used to distinguish weakly localized bosons from simple boson-pair hopping or a true Bose glass (Lin et al., 2010). Second, interaction effects do not simply suppress interference monotonically: in three-dimensional disorder, inelastic collisions reduce coherence globally yet produce a residual enhancement of coherent backscattering in a narrow spectral band around the initial energy (Geiger et al., 2013). Third, apparent weak antilocalization in nonlinear bosonic matter waves need not signal a universal bosonic reversal of the interference correction; in the semiclassical billiard problem it is attributed to non-universal short self-retracing paths that modify the zero-field reflection background (Hartmann et al., 2011).

Taken together, these results define weak localization of bosons as a coherence-sensitive interference phenomenon whose observables include $G(T)=G_0-\frac{e^2}{2\pi^2\hbar}\ln(T/T_0),$05, coherent backscattering lineshapes, condensate fragmentation, and decay of one-body correlations. The most direct condensed-matter signature remains the logarithmic temperature dependence of the resistance in the high-field insulating regime of ultrathin films, while cold-atom realizations provide microscopic control over collisions, spectral redistribution, geometry, and quasiperiodicity (Lin et al., 2010, Geiger et al., 2013, Hartmann et al., 2011, Ray et al., 2015, Gonzalez-Garcia et al., 2016, Chakrabarti et al., 20 Feb 2025).