Using near-field second-harmonic scanning-probe microscopy of dislocation-rich III-V semiconductor films, Shafiei and collaborators reported that Anderson localization of light is suppressed once the optical mean free path (MFP) between dislocation scattering sites falls below roughly 13 nm: the smallest localized hotspot observed (about 55 nm) corresponds to a 14.5 nm MFP, and a film with a 13 nm MFP shows no localization at all, while idealized hard-wall wedge simulations of Maxwell's equations reproduce suppression at a comparable gap width. The authors interpret this as a fundamental lower size limit of light localization, but the observation is confined to one material system and one probe wavelength pair (780 nm excitation, 390 nm second-harmonic collection). The open problem is to establish whether such a lower limit genuinely exists as a general property of electromagnetic waves in disordered media, to determine how the critical MFP depends on wavelength and material system, to identify the microscopic mechanism that suppresses localization at deep sub-wavelength confinement (going beyond the charge-free hard-wall model to realistic charged, leaky dislocation walls and three-dimensional disorder), and to characterize the statistics of localization near the threshold.

Background

Anderson localization is the arrest of wave diffusion in a disordered medium by wave interference, originally proposed by P. W. Anderson in 1958 for electrons in dirty lattices and later extended to light: in a strongly scattering medium, multiply scattered waves interfere constructively at random positions, trapping electromagnetic energy in sub-wavelength 'hotspots'. The strength of disorder is measured by the optical mean free path (MFP), the average distance a wave travels between scattering events; the standard Ioffe-Regel criterion expects localization to set in when the product of wave number and MFP becomes of order unity, and in the conventional scaling picture localization becomes stronger, not weaker, as the MFP shrinks. Laboratory light localization has been reported in disordered semiconductors, random lasers, and photonic lattices, with MFPs typically of a few hundred nanometers, and the very existence of Anderson localization of electromagnetic waves in three dimensions remains an active theoretical debate (Yamilov, Skipetrov, Hughes, Minkov, Yu and Cao, Nature Physics 19, 1308 (2023) found that vector-wave localization is absent in idealized point-scatterer ensembles). Suppression of Anderson localization through deliberately correlated or structured disorder has been described in one-dimensional photonic superlattices (Reyes-Gómez et al., Physical Review B 85, 195110 (2012); Journal of Physics: Condensed Matter 25, 075901 (2013)), but there the suppression follows from the engineered multilayer band structure rather than from disorder strength; none of that literature anticipates an upper bound on disorder, i.e., a smallest MFP below which localization fails in an ordinary disordered medium.

The present question arises from a specific experimental platform. Metamorphic III-V films (e.g., InGaAs on GaAs, GaAs on Si) contain a controllable density of threading dislocations — line defects penetrating the film — which act as nanoscale light-scattering sites whose average spacing can be tuned down to about ten nanometers. Because GaAs lacks inversion symmetry, it has a strong second-order nonlinear susceptibility: intense concentrations of fundamental light at 780 nm are selectively converted into second-harmonic (frequency-doubled) light at 390 nm. A near-field scanning optical microscope with an uncoated 50-nm-aperture fiber probe, held tens of nanometers above the sample and collecting only the second-harmonic signal, can therefore map the deepest sub-wavelength light concentrations with about 19 nm spatial resolution without the metallic-tip field enhancement that would perturb them. Shafiei, Orzali, Vert, Miri, Hung, Wong, Alù, Bersuker and Downer (Advanced Optical Materials 9, 2002252 (2021)) previously used this technique to detect subsurface dislocations via nonlinear light scattering and localization, and to relate the observed hotspot size to the underlying MFP using the localization theory of disordered photonic systems (Lee et al., Science Advances 4, e1602796 (2018)).

In the source work (arXiv:2211.02808 (2022), published as Shafiei, Masir, Orzali, Vert, Wong, Bersuker and Downer, 'Observation of suppression of Anderson localization of light', Applied Physics Letters 129, 041109 (2026)), the group scanned a series of dislocated films with different dislocation densities. They observed localized hotspots whose size tracks the MFP, the smallest being about 55 nm, corresponding to a 14.5 nm MFP; a film whose dislocation spacing gives a 13 nm MFP showed no localized hotspots at all. Since the probe resolution (about 19 nm) corresponds to a MFP of about 8.8 nm, the authors argue the absence of smaller hotspots is not a resolution artifact. A supplementary analytic calculation solves Maxwell's equations between the arms of a 'V-shape' (wedge) scattering structure under hard-wall, charge-free boundary conditions and finds that localization is suppressed when the wedge opening falls to roughly 11-12.6 nm, consistent with the experimental threshold; finite-element simulations with electric charges assigned to the dislocation arms reproduce the suppression for the modeled geometries. The authors conclude that localization of light is suppressed when the mean free path falls below roughly 13 nm at the studied wavelength, and the published abstract describes the limit as 'general, albeit material and lambda-dependent'. Three gaps are thereby opened and acknowledged in the source itself. First, the analytic model is idealized: real dislocation arms carry charge and leak field, acting as soft walls, so how realistic boundary conditions and charge distributions shift the predicted suppression threshold is unquantified. Second, everything is measured at one wavelength pair on one materials family; whether the critical MFP is universal, or how it depends on wavelength and material system, is unknown. Third, the smallest hotspots become very rare near the threshold, and the probability distribution of hotspot sizes versus MFP is explicitly not shown, so the statistical character of the transition is uncharacterized. Notably, a suppression of localization at small MFP is not predicted by any standard theory of Anderson localization for ordinary disordered media, which is what makes identifying its mechanism a well-posed open physics question rather than a routine extension.

Problem Statement

Establish the physical origin and generality of the observed lower size limit of Anderson localization of light in disordered media. Concretely: (1) determine whether the suppression of localization below a critical optical mean free path, reported at ~13 nm for second-harmonic collection at 390 nm in dislocation-rich InGaAs/GaAs and GaAs/Si films (smallest observed hotspot ~55 nm at a 14.5 nm mean free path; no hotspots at a 13 nm mean free path), is a general property of electromagnetic waves in disordered media, and derive or measure how the critical mean free path depends on wavelength and material system — the dependencies the source itself concedes as open ('general, albeit material and lambda-dependent'); (2) identify the microscopic mechanism responsible for suppression at deep sub-wavelength confinement, replacing the source's idealized charge-free, hard-wall V-shape (wedge) calculation with a treatment of realistic charged and field-leaking 'soft' dislocation walls, material inhomogeneity, and three-dimensional disorder ensembles, and showing quantitatively that this mechanism reproduces the observed threshold (~11-13 nm gap/mean-free-path scale) and predicts its scaling; and (3) characterize the statistics of localization approaching the limit — the distribution and frequency of localized hotspot sizes as a function of mean free path near the threshold, which the source explicitly could not resolve because the smallest hotspots become very rare. Items (1)-(3) are concrete facets of this single question rather than independently sufficient subgoals; whether an answer is complete is governed by the criterion below. A complete answer either (A) provides a quantitative account of the suppression threshold, its wavelength and material-system dependence, and the statistics of localized hotspot sizes near the threshold — addressing facets (1)-(3) — backed by a derivation, numerical simulation, or measurement, internally consistent with the reported ~13 nm threshold and ~55 nm hotspot at a 14.5 nm mean free path, and covering at least one wavelength or material system beyond the reported 390 nm second-harmonic case, or (B) demonstrates that no such fundamental limit exists (localization surviving at arbitrarily small mean free path) and identifies the true origin of the reported suppression — resolving facet (1)'s generality question and facet (2)'s mechanism negatively and, for facet (3), establishing that hotspot-size statistics show no cutoff as the mean free path shrinks rather than reporting a measured near-threshold distribution.

References

  1. Shafiei, F., Masir, M.R., Orzali, T., Vert, A., Wong, M.H., Bersuker, G., Downer, M.C. 'Observation of suppression of Anderson localization of light', Applied Physics Letters 129, 041109 (2026). DOI: 10.1063/5.0314762, https://doi.org/10.1063/5.0314762
  2. Shafiei, F., Masir, M.R., Orzali, T., Vert, A., Wong, M.H., Bersuker, G., Downer, M.C. 'Observation of Fundamental Limit of Light Localization', arXiv:2211.02808 (2022). DOI: 10.48550/arXiv.2211.02808, https://arxiv.org/abs/2211.02808
  3. Shafiei, F., Orzali, T., Vert, A., Miri, M.-A., Hung, P.Y., Wong, M.H., Alù, A., Bersuker, G., Downer, M.C. 'Detection of Subsurface, Nanometer-Scale Crystallographic Defects by Nonlinear Light Scattering and Localization', Advanced Optical Materials 9, 2002252 (2021). DOI: 10.1002/adom.202002252, https://doi.org/10.1002/adom.202002252
  4. Shafiei, F., Downer, M.C. 'Collection of propagating electromagnetic fields by uncoated probe', Ultramicroscopy 240, 113597 (2022). DOI: 10.1016/j.ultramic.2022.113597, https://doi.org/10.1016/j.ultramic.2022.113597
  5. Anderson, P.W. 'Absence of Diffusion in Certain Random Lattices', Physical Review 109, 1492 (1958). DOI: 10.1103/PhysRev.109.1492, https://doi.org/10.1103/PhysRev.109.1492
  6. John, S. 'Localization of Light', Physics Today 44, 32 (1991). DOI: 10.1063/1.881300, https://doi.org/10.1063/1.881300
  7. Wiersma, D.S., Bartolini, P., Lagendijk, A., Righini, R. 'Localization of light in a disordered medium', Nature 390, 671 (1997). DOI: 10.1038/37757, https://doi.org/10.1038/37757
  8. Lee, M., Lee, J., Kim, S., Callard, S., Seassal, C., Jeon, H. 'Anderson localizations and photonic band-tail states observed in compositionally disordered platform', Science Advances 4, e1602796 (2018). DOI: 10.1126/sciadv.1602796, https://doi.org/10.1126/sciadv.1602796
  9. Yamilov, A., Skipetrov, S.E., Hughes, T.W., Minkov, M., Yu, Z., Cao, H. 'Anderson localization of electromagnetic waves in three dimensions', Nature Physics 19, 1308 (2023). DOI: 10.1038/s41567-023-02091-7, https://doi.org/10.1038/s41567-023-02091-7
  10. Reyes-Gómez, E., Bruno-Alfonso, A., Cavalcanti, S.B., Oliveira, L.E. 'Suppression of Anderson localization of light in one-dimensional disordered photonic superlattices', Physical Review B 85, 195110 (2012). DOI: 10.1103/physrevb.85.195110, https://doi.org/10.1103/physrevb.85.195110
  11. Reyes-Gómez, E., Cavalcanti, S.B., Oliveira, L.E. 'Omnidirectional suppression of Anderson localization of light in disordered one-dimensional photonic superlattices', Journal of Physics: Condensed Matter 25, 075901 (2013). DOI: 10.1088/0953-8984/25/7/075901, https://doi.org/10.1088/0953-8984/25/7/075901