Pith. sign in

REVIEW 3 major objections 5 minor 74 references

Non-Hermitian global synchronization

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Non-reciprocally coupled Stuart-Landau oscillators globally synchronize through the non-Hermitian skin effect, independent of initial conditions and weak disorder.

desk verdict Plausible new mechanism for global oscillator synchronization via skin-effect mode convergence, with real circuit evidence, but the min-IPR selection rule that carries the theory is a heuristic, not a derivation. read the letter →

arxiv 2501.14169 v1 pith:QGRGIIAW submitted 2025-01-24 nlin.PS cond-mat.mes-hall

Weixuan Zhang , Fengxiao Di , Xiangdong Zhang This is my paper · ORCID
classification nlin.PScond-mat.mes-hall
keywords globalsynchronizationnon-HermitianskineffectStuart-Landauoscillatorsnon-reciprocalcouplingtopologicalzeromodestopoelectricalcircuitsinverseparticipationratioKuramotoorderparameter
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper claims that the non-Hermitian skin effect—the boundary pile-up of eigenstates in chains with unequal left/right couplings—can be converted into a synchronization mechanism. In a Hatano-Nelson chain (a one-dimensional lattice with unequal left/right couplings) of Stuart-Landau oscillators (self-sustained nonlinear oscillators), the dynamics is reported to converge, for one thousand random initial states, to one of two single-frequency oscillations whose frequencies match the eigenenergies of the two most concentrated linear eigenstates. The proposed selection rule is that the nonlinear effective potential is minimized by the eigenstate with the smallest inverse participation ratio, $\mathrm{IPR} = \sum_l |\varphi_l(\varepsilon)|^4$, and the strong non-orthogonality of skin-localized modes lets that state absorb all the others. The same rule is shown to extend to the zero-energy midgap state of a non-Hermitian SSH chain (a two-band lattice with alternating couplings), producing topological global synchronization, and to nonlinear eigenstates in longer chains. The authors validate the mechanism in fabricated nonlinear topoelectrical circuits, with a 9-site circuit locking at 23.5 kHz to a skin mode and a 15-site circuit locking at 92.6 kHz to a topological zero mode.

What carries the argument

The load-bearing object is the minimum-IPR selection rule: the eigenstate with the smallest inverse participation ratio, $\mathrm{IPR} = \sum_l |\varphi_l|^4$, is the state that minimizes the nonlinear part of the effective Hamiltonian and therefore becomes the global synchronized state. The machinery has three parts. First, non-reciprocal couplings ($J_+\neq J_-$) in a Hatano-Nelson chain produce the non-Hermitian skin effect, localizing every eigenstate at one boundary and breaking the orthogonality of the eigenbasis; the biorthogonal right/left eigenstate expansion then shows that initially excited modes couple strongly. Second, the effective Hamiltonian $H = \sum_l 0.5(-\alpha|Z_l|^2 + 0.5\beta|Z_l|^4) - i\sum_{kl}J_{kl}Z_l^*Z_k$ assigns a lower potential to more concentrated states, so the dynamics should transfer population to the minimum-IPR mode. Third, in the topological extension, the non-Hermitian SSH chain's zero-energy midgap state can become the minimum-IPR state for suitable intercell coupling, making the topological mode the global attractor; when the linear expansion breaks down, nonlinear eigenstates play the same role.

What would settle it

Start a chain from an initial condition whose overlap with the two most concentrated patterns is exactly zero, for example by exciting only a weakly concentrated pattern. The paper's initial-state-independence claim predicts that this trajectory must still lock to the frequency and spatial profile of the most concentrated pattern; observing a multi-frequency steady state, or locking to a different pattern, would refute the central mechanism.

Watch

Extended reading notes

Core claim

The central discovery is that non-reciprocal coupling makes collective synchronization a property of the lattice's most concentrated eigenstate rather than of carefully prepared initial conditions. For a Hatano-Nelson chain (a one-dimensional tight-binding chain with unequal hopping amplitudes $J_+\neq J_-$), all eigenstates are localized at one boundary, and the authors find that the long-time dynamics always settles into the linear eigenstate with the minimal inverse participation ratio, $\mathrm{IPR} = \sum_l |\varphi_l(\varepsilon)|^4$, with the common oscillation frequency given by $|\varepsilon+\omega_0|$. The explanation is that the effective Hamiltonian $H = \sum_l 0.5(-\alpha|Z_l|^2 + 0.5\beta|Z_l|^4) - i\sum_{kl} J_{kl}Z_l^*Z_k$ has a nonlinear potential term $0.25\beta\sum_l|Z_l|^4$ that is smallest for the most concentrated state, while the skin effect's breakdown of eigenstate orthogonality couples all initially excited modes so the system drains into that minimum. The same minimum-IPR logic is then applied to a non-Hermitian SSH chain, where tuning the intercell coupling can make the midgap topological zero mode the minimum-IPR state and thereby synchronize the whole lattice; in longer chains the role passes to nonlinear eigenstates obtained numerically.

Load-bearing premise

The argument assumes, without proof, that the nonlinear terms create a genuine energy landscape whose lowest point is always the most concentrated vibration pattern, and that the interactions between patterns reliably push every starting condition into that lowest point.

Editorial extensions

If this is right

  • With the paper's parameters, any of one thousand random initial states of a 15-site Hatano-Nelson chain converges to one of two single-frequency oscillations, so global synchronization is achieved without preparing a particular initial condition.
  • Weak disorder in the onsite frequencies (up to $W=0.1\omega_0$) leaves the synchronization probability at 100 percent in the linear-skin region and above 90 percent in the nonlinear-skin region, so the effect tolerates structural perturbations.
  • System size acts as a control knob: longer chains first turn on linear skin-state synchronization, then pass through an unsynchronized transition region, and then recover synchronization through nonlinear skin states.
  • In the non-Hermitian SSH chain, tuning the intercell coupling selects among linear skin-state, nonlinear skin-state, topological, and skin-topological cluster synchronization, with the midgap zero mode synchronizing the full lattice when it has the minimum IPR.
  • Fabricated circuits confirm the two headline cases: a 9-site circuit shows single-frequency voltage oscillations at 23.5 kHz matching the minimum-IPR skin mode, and a 15-site circuit synchronizes at 92.6 kHz matching the topological zero mode.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Beyond the paper's examples, the minimum-IPR rule suggests a general design recipe: to make an array synchronize at a target frequency and spatial profile, engineer the non-reciprocal lattice so that the target eigenstate has the smallest IPR; mode competition should then select it automatically.
  • The same logic should transfer to other non-Hermitian lattices with strongly non-orthogonal eigenmodes, such as disordered or quasiperiodic chains and two-dimensional skin-effect geometries; direct simulations of those systems would test the generality of the mechanism.
  • The transition regions where synchronization fails, and the observation that disorder can partially rescue synchronization there, imply that the basin of attraction is controlled by effective mode-coupling strength rather than by nonlinearity alone, which could be tested by tuning coupling asymmetry at fixed gain.
  • For laser arrays, the paper's picture points to a concrete engineering target: non-reciprocal coupling alone could enforce single-mode, single-frequency operation in a large array without external injection locking or careful pump shaping.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper proposes and experimentally demonstrates that in chains of Stuart-Landau oscillators with non-reciprocal couplings, the dynamics converge to a single-frequency synchronized state whose spatial profile matches the minimum-IPR eigenstate of the linear Hatano-Nelson or non-Hermitian SSH model, for essentially all initial conditions over a range of parameters and weak disorder. The authors call this non-Hermitian global synchronization, identify linear skin-state, nonlinear skin-state, topological, and skin-topological synchronized regimes, and support the claims with ODE simulations, disorder statistics, phase diagrams, and circuit experiments for N=5 versus N=9 (skin) and N=9 versus N=15 (topological).

Significance. If the proposed mechanism is correct, the work offers a novel and appealing route to robust global synchronization that is insensitive to initial conditions and scalable in size, and the experimental realization in nonlinear topoelectrical circuits is a substantial contribution. The paper contains nontrivial, falsifiable predictions: the final oscillation frequency and spatial profile are compared with independently computed minimum-IPR eigenstates, and the size-dependent onset of synchronization (N=5 vs N=9 and N=9 vs N=15) was predicted before the circuit measurements. These are genuine strengths. The central limitation is that the theoretical selection rule identifying the min-IPR eigenstate as the global attractor is asserted rather than derived, and the experimental evidence, while qualitatively consistent, covers only a small number of samples.

major comments (3)
  1. [Section 2 (effective Hamiltonian paragraph)] The statement that the system evolves into the linear eigenstate with the minimal effective potential, namely the minimal-IPR eigenstate, is not supported by the equations as written. The effective Hamiltonian H = sum 0.5(-alpha|Z_l|^2 + 0.5 beta|Z_l|^4) - i sum J_kl Z_l^* Z_k is complex, so it is not a real potential or Lyapunov function for the flow Z_dot = -dH/dZ*. Moreover, with standard Wirtinger derivatives this H does not reproduce Eq. (1): the gain and nonlinearity terms come out with coefficients alpha/2 and beta/2 rather than alpha and beta. Thus the 'minimal effective potential' argument, which is the only link between non-Hermitian mode non-orthogonality and the predicted final state, is a heuristic at this point. This matters because the paper's central predictions, namely which eigenstate wins and therefore which frequency and spatial profile appear, rest entirely on this rule.
  2. [Section 2 (convergence to min-IPR)] Even if the effective Hamiltonian were corrected, the claim that the quartic term sum 0.25 beta |Z_l|^4 favors the minimal-IPR state requires an implicit fixed-norm assumption. During the evolution, the Stuart-Landau gain term alpha|Z_l|^2 changes the total norm, so the comparison of IPR at fixed norm is not valid. The paper does not provide a derivation of gradient-descent-like dynamics on the mode coefficients C_n(t), nor a numerical test showing that these coefficients follow such a landscape. I would like to see either an analytical argument (for example, a Lyapunov function or an adiabatic elimination of the mode amplitudes) or direct numerical evidence that the expansion coefficients evolve toward the minimal-IPR mode in a way that depends on an effective potential. Without this, the mechanism remains a conjecture even though the simulations and experiments are consistent with it.
  3. [Section 2 (order parameter R_o)] The order parameter R_o is defined as max[R(t)] - min[R(t)] for t>t0, which was chosen after observing that the two attractors are in-phase and anti-phase. Because R_o is calibrated to the two observed synchronized states, the phase diagrams in Figs. 1f, 1g, 3j, and 3k partly codify the classification rather than independently detecting synchronization. This does not invalidate the reported profile and frequency matches, but it weakens the claim that the diagrams demonstrate a synchronization transition without prior knowledge of the attractors. Please either justify R_o on independent grounds (for example, by showing it is equivalent to a standard measure after a sublattice rotation) or present an additional order parameter that detects anti-phase synchronization in a principled way.
minor comments (5)
  1. [Section 2] There is a typo in the sentence 'the system is excepted to evolve' which should read 'expected to evolve'.
  2. [Section 2 (IPR definition)] The definition IPR(epsilon) = sum |phi_l(epsilon)|^4 should specify the normalization convention for the eigenstates, for example sum |phi_l|^2 = 1, since the IPR value depends on the normalization.
  3. [Figure 1 caption] The caption states 'at the time marked by blue lines in (b1)-b(2)' but the blue lines are dashed vertical lines in the main panels; the notation is inconsistent and should be cleaned up.
  4. [Equations (3)-(4)] The imaginary unit is denoted j in the circuit equations but i in Eq. (1); while this is a common convention, it should be stated explicitly to avoid confusion.
  5. [Section 2 (effective Hamiltonian)] The factor-of-two mismatch between the effective Hamiltonian and Eq. (1) may be a typographical error, but it should be corrected and the derivation rechecked so that the gradient-descent statement is at least dimensionally consistent.

Circularity Check

1 steps flagged · score 4.0 of 10

The min-IPR mode-selection rule is a definitional restatement in the explanatory chain, but the synchronized final states are independently compared with computed eigenmodes and circuit measurements, so the paper is only partially circular.

  1. self definitional [Section 2, paragraph beginning "Finally, we explain why all initially excited eigenstates eventually converge to a single eigenmode" (after Eq. (1))]
    "The first term in the Hamiltonian represents the effective potential energy. In this framework, different eigenstates of the system experience different effective potentials due to the nonlinear term ∑ 0.25𝛽|𝑍𝑙|4𝑙 . The eigenstate that minimizes the potential energy corresponds to the state with the minimal inverse participation ratio (IPR), which is defined as 𝐼𝑃𝑅(𝜀) = ∑ |𝝋𝑙(𝜀)|4𝑙 . ... the system is excepted to evolve into the linear eigenstate with the minimal effective potential — namely, the eigenstate with the minimal IPR"

    The quartic term 0.25βΣ|Z_l|^4 is, up to the fixed total intensity, exactly 0.25β times the IPR Σ|Z_l|^4. Therefore "minimizes the effective potential energy" and "minimal IPR" are the same statement by construction. This paragraph is offered as the explanation of convergence, but it only restates the mode-selection rule in potential-language; it does not derive from Eq. (1) that trajectories actually approach that state. The effective Hamiltonian is complex and not a real Lyapunov function, and the asserted identity Ż = −∂H/∂Z^* does not reproduce Eq. (1) (the gain/nonlinearity terms are off by a factor 1/2 and the coupling/onsite imaginary parts have the wrong signs), so the 'expected' convergence to the minimum-IPR eigenstate is an input assumption rather than a derived consequence.

full rationale

The paper's main predictive claims—that the steady-state oscillation frequency and spatial profile match the minimum-IPR eigenstate of the linear Hatano-Nelson or SSH chain—are tested against direct numerical integration and circuit measurements without fitting parameters. The size dependence (N=5 vs N=9 for skin synchronization; N=9 vs N=15 for topological synchronization) is predicted from the model before the circuit measurements. Thus the central empirical content is not circular. The circularity is confined to the theoretical 'explanation' in Section 2: the effective-potential argument is constructed so that the minimum of its nonlinear term is, by definition, the minimum-IPR state, so it cannot independently justify the mode-selection rule. This is a genuine definitional reduction in the explanatory chain, but it does not make the observation of synchronization fitted or tautological. The additional mismatch between the stated effective Hamiltonian and Eq. (1) is a correctness/rigor concern rather than a further circularity. Overall, the derivation is partially circular in its mechanism explanation while retaining independent empirical content.

Assumptions & free parameters 5 free parameters · 6 assumptions · 1 invented entities

No parameters are fitted to an external target; the Stuart-Landau and coupling parameters are standard model inputs chosen by hand, and the circuit component values implement the same effective parameters. The main non-standard input is the heuristic minimum-IPR selection rule, and the voltage pseudospin is a basis change rather than a new physical entity.

free parameters (5)
  • Stuart-Landau gain coefficient α = 5e-3 (skin model), 5e-4 (topological model)
    Chosen by hand to set self-oscillation amplitude; no external calibration. Regime boundaries shift with α but the phenomenon is not fitted to a target.
  • Nonlinearity coefficient β = 5e-4 (skin), 5e-5 (topological)
    Sets amplitude scale and the threshold for linear vs nonlinear eigenstate expansion.
  • Natural frequency ω0 = 0.1
    Sets the oscillation frequency offset and timescale.
  • Non-reciprocal couplings J+ and J- = J+=1.5, J-=1 (skin); J+=0.56, J-=0.1 (SSH)
    Chosen to place the system in the skin-effect regime; the phase diagram is scanned over J- and J.
  • Intercell coupling J (SSH) = scanned 0.2, 0.35, 0.5, 1.0; J=0.4 for size scan
    Scanned to select different synchronization regimes; not fitted.
assumptions (6)
  • domain assumption Each oscillator is a Stuart-Landau oscillator with cubic nonlinearity.
    Used throughout Eq. (1) and Eq. (2); standard model for self-oscillating units.
  • domain assumption The lattice is described by a tight-binding Hatano-Nelson or non-Hermitian SSH Hamiltonian with non-reciprocal nearest-neighbor couplings.
    Defines the model in Secs. 2 and 3; the skin effect and topological modes come from this Hamiltonian.
  • domain assumption When β|Z_l|^2 is much smaller than the couplings, the waveform can be expanded in linear right eigenstates using biorthogonal left and right basis.
    Stated in Sec. 2 before Fig. 1c; validity condition is checked post hoc by amplitudes.
  • ad hoc to paper The nonlinear term ∑0.25β|Z_l|^4 acts as an effective potential that drives the system to the minimal-IPR eigenstate.
    Introduced in Sec. 2 in the paragraph beginning 'Finally, we explain why all initially excited eigenstates eventually converge to a single eigenmode'; not derived from the dynamical equations, and the coupling term is not shown to be a descent direction.
  • ad hoc to paper Two attractors (in-phase and anti-phase) coexist, and convergence to either counts as global synchronization.
    The R_o order parameter in Sec. 2 is defined so that both states give R_o=0; this broadens the meaning of 'global synchronization'.
  • domain assumption Circuit voltage pseudospins V↑ and V↓ follow the same equations as the lattice model.
    Derived in SI 11 under the condition 1/(C R_a)-2/(C R_2)=0; used to claim experimental realization.
invented entities (1)
  • Voltage pseudospins V↑ and V↓ at each circuit site independent evidence
    purpose: Maps the four-node nonlinear circuit to the two-species lattice model with real-valued non-reciprocal couplings, allowing experimental simulation of Eq. (1) and Eq. (2).
    Defined in Section 4 as linear combinations of node voltages; the mapping is derived from Kirchhoff's laws in SI 11 and the resulting synchronized voltage profiles match the predicted skin and topological modes.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Non-Hermitian global synchronization." pith.science (2026). https://pith.science/paper/QGRGIIAW

@misc{pith2026250114169,
  author       = {Pith},
  title        = {Pith review of: Non-Hermitian global synchronization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QGRGIIAW}},
  note         = {Machine review of arXiv:2501.14169}
}
read the original abstract

Synchronization of coupled nonlinear oscillators is a prevalent phenomenon in natural systems and can play important roles in various fields of modern science, such as laser arrays and electric networks. However, achieving robust global synchronization has always been a significant challenge due to its extreme susceptibility to initial conditions and structural perturbations. Here, we present a novel approach to achieve robust global synchronization by manipulating the interplay between non-Hermitian physics and nonlinear dynamics. Remarkably, the initial-state-independent non-Hermitian skin and topological global synchronization are proposed, exhibiting diverse anomalous effects such as the enlarged-size triggered non-Hermitian global synchronization and nonlinear skin states-dominated global synchronization. To validate our findings, we design and fabricate nonlinear topoelectrical circuits for experimental observation of non-Hermitian global synchronization. Our work opens up a promising avenue for establishing resilient global synchronization with potential applications in constructing high-radiance laser arrays and topologically synchronized networks.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

74 extracted references · 74 canonical work pages

  1. [1]

    Strogatz, Sync: The Emerging Science of Spontaneous Order

    S. Strogatz, Sync: The Emerging Science of Spontaneous Order. Physics Today, 2003

    work page 2003

  2. [2]

    S.-B. Shim, M. Imboden, P. Mohanty, Science 2007, 316, 95–99

    work page 2007

  3. [3]

    J. Yan, M. Bloom, S. C. Bae, E. Luijten, S. Granick, Nature 2012, 491, 578–581

    work page 2012

  4. [4]

    Prindle et al., Nature 2012, 481, 39–44

    A. Prindle et al., Nature 2012, 481, 39–44

    work page 2012

  5. [5]

    Rohden, A

    M. Rohden, A. Sorge, M. Timme, D. Witthaut, Phys. Rev. Lett. 2012, 109, 064101

    work page 2012

  6. [6]

    Jenkins, Physics Reports

    A. Jenkins, Physics Reports. 2013, 525, 167-222

    work page 2013

  7. [7]

    Nitsan, S

    I. Nitsan, S. Drori, Y. E. Lewis, S. Cohen, S. Tzlil, Nat. Phys. 2016, 12, 472-477

    work page 2016

  8. [8]

    E. Post, M. C. Forchhammer, Nature 2002, 420, 168-171. 20

    work page 2002

Show all 74 references
  1. [9]

    J. A. Acebró n, L. L. Bonilla, C. J. Pé rez Vicente, F. Ritort, R. Spigler, Rev. Mod. Phys. 2005, 77, 137-185

    2005

  2. [10]

    M. Z. Hasan, C. L. Kane, Rev. Mod. Phys. 2010, 82, 3045-3067

    2010

  3. [11]

    Qi, S.-C

    X.-L. Qi, S.-C. Zhang, Rev. Mod. Phys. 2011, 83, 1057-1110

    2011

  4. [12]

    Ozawa et al., Rev

    T. Ozawa et al., Rev. Mod. Phys. 2019, 91, 015006

    2019

  5. [13]

    G. Ma, M. Xiao, C. T. Chan, Nat. Rev. Phys. 2019, 1, 281-294

    2019

  6. [14]

    Smirnova, D

    D. Smirnova, D. Leykam, Y. Chong, Y. Kivshar, Appl. Phys. Rev. 2020, 7

    2020

  7. [15]

    Ota et al., Active topological photonics

    Y. Ota et al., Active topological photonics. Nanophotonics 2020, 9, 547-567

    2020

  8. [16]

    Kotwal et al., Proceedings of the National Academy of Sciences 2021,118, e2106411118

    T. Kotwal et al., Proceedings of the National Academy of Sciences 2021,118, e2106411118

    2021

  9. [17]

    Zhang et al., Nat

    Z. Zhang et al., Nat. Commun, 2020, 11, 1902

    2020

  10. [18]

    B. G.-g. Chen, N. Upadhyaya, V. Vitelli, Proceedings of the National Academy of Sciences 2014,111, 13004-13009

    2014

  11. [19]

    Leykam, Y

    D. Leykam, Y. D. Chong, Phys. Rev. Lett. 2016, 117, 143901

    2016

  12. [20]

    Lumer, Y

    Y. Lumer, Y. Plotnik, M. C. Phys. Rev. Lett. 2013, 111, 243905

    2013

  13. [21]

    Hadad, A

    Y. Hadad, A. B. Khanikaev, A. Alù , Phys. Rev. B 2016, 93, 155112

    2016

  14. [22]

    Hadad, J

    Y. Hadad, J. C. Soric, A. B. Khanikaev, A. Alù , Nat. Electron. 2018, 1, 178-182

    2018

  15. [23]

    Jü rgensen, S

    M. Jü rgensen, S. Mukherjee, M. C. Rechtsman, Nature 2021, 596, 63-67

    2021

  16. [24]

    Jü rgensen, M

    M. Jü rgensen, M. C. Rechtsman, Phys. Rev. Lett. 2022, 128, 113901

    2022

  17. [25]

    C. W. Wä chtler, G. Platero, Phys. Rev. Res. 2023, 5, 023021

    2023

  18. [26]

    K. Sone, Y. Ashida, T. Sagawa, Phys. Rev. Res. 2022, 4, 023211

    2022

  19. [27]

    Fruchart et al., Nature 2021, 592, 363-369

    M. Fruchart et al., Nature 2021, 592, 363-369

    2021

  20. [28]

    J. Rohn, K. P. Schmidt, C. Genes, Phys. Rev. A 2023, 108, 023721

    2023

  21. [29]

    C. M. Bender, S. Boettcher, Phys. Rev. Lett. 1998, 80, 5243-5246

    1998

  22. [30]

    S. Yao, Z. Wang, Phys. Rev. Lett. 2018, 121, 086803

    2018

  23. [31]

    Yokomizo, S

    K. Yokomizo, S. Murakami, Phys. Rev. Lett. 2019, 123, 066404

    2019

  24. [32]

    Kawabata, K

    K. Kawabata, K. Shiozaki, M. Ueda, M. Sato, Phys. Rev. X 2019, 9, 041015

    2019

  25. [33]

    Weidemann et al., Science 2020, 368, 311-314

    S. Weidemann et al., Science 2020, 368, 311-314

    2020

  26. [35]

    Hatano and D

    N. Hatano and D. R. Nelson, Phys. Rev. Lett. 1996, 77, 570

    1996

  27. [36]

    L. Li, C. H. Lee, S. Mu, J. Gong, Nat. Commun. 2020, 11, 5491

    2020

  28. [37]

    Ghatak, M

    A. Ghatak, M. Brandenbourger, J. van Wezel, C. Coulais, Proceedings of the National Academy of Sciences 2020, 117, 29561-29568

    2020

  29. [38]

    E. J. Bergholtz, J. C. Budich, F. K. Kunst, Rev. Mod. Phys. 2021, 93, 015005. 21

    2021

  30. [39]

    Zhang, Z

    K. Zhang, Z. Yang, C. Fang, Nat. Commun. 2022, 13, 2496

    2022

  31. [40]

    P. Gao, M. Willatzen, J. Christensen, Phys. Rev. Lett. 2020, 125, 206402

    2020

  32. [41]

    W. Wang, X. Wang, G. Ma, Nature 2022, 608, 50-55

    2022

  33. [42]

    K. Wang, A. Dutt, K. Yang, C. C. Wojcik, J. Vučković and S. Fan, Science 2021, 371, 1240- 1245

    2021

  34. [43]

    Longhi, Annalen der Physik 2018, 530, 1800023

    S. Longhi, Annalen der Physik 2018, 530, 1800023

    2018

  35. [44]

    S. Wong, S. S. Oh, Phys. Rev. Res. 2021, 3, 033042

    2021

  36. [45]

    Gao et al., Nature 2015, 526, 554-558

    T. Gao et al., Nature 2015, 526, 554-558

    2015

  37. [46]

    Xiao et al., Nat

    L. Xiao et al., Nat. Phys. 2020, 16, 761-766

    2020

  38. [47]

    Longhi, Phys

    S. Longhi, Phys. Rev. Lett. 2020, 124, 066602

    2020

  39. [48]

    Miri and A

    M.-A. Miri and A. Alù , Science 2019, 363, eaar7709

    2019

  40. [49]

    L. Feng, R. El-Ganainy, and L. Ge, Nat. Photonics. 2017, 11, 752

    2017

  41. [50]

    El -Ganainy, K

    R. El -Ganainy, K. G. Makris, M. Khajavikhan, Z. H. Musslimani, S. Rotter, and D. N. Christodoulides, Nat. Phys. 2018, 14, 11

    2018

  42. [51]

    Bouganne, M

    R. Bouganne, M. Bosch Aguilera, A. Ghermaoui, J. Beugnon, and F. Gerbier, Nat. Phys. 2020, 16, 21

    2020

  43. [52]

    Syassen, D

    N. Syassen, D. M. Bauer, M. Lettner, T. Volz, D. Dietze, J. J. Garcí a-Ripoll, J. I. Cirac, G. Rempe, and S. Dü rr, Science 2008, 320, 1329

    2008

  44. [53]

    McDonald, R

    A. McDonald, R. Hanai, A. A. Clerk, Phys. Rev. B 2022, 105, 064302

    2022

  45. [54]

    S. Begg, R. Hanai, Phys. Rev. Lett. 2024, 132, 120401

    2024

  46. [55]

    C. H. Lee et al., Commun. Phys. 2018, 1, 39

    2018

  47. [56]

    Ningyuan, C

    J. Ningyuan, C. Owens, A. Sommer, D. Schuster, J. Simon, Phys. Rev. X 2015, 5, 021031

    2015

  48. [57]

    V. V. Albert, L. I. Glazman, L. Jiang, Phys. Rev. Lett. 2015, 114, 173902

    2015

  49. [58]

    Imhof et al., Nat

    S. Imhof et al., Nat. Phys. 2018, 14, 925-929

    2018

  50. [59]

    Helbig et al., Nat

    T. Helbig et al., Nat. Phys. 2020, 16, 747-750

    2020

  51. [60]

    N. A. Olekhno et al., Nat. Commun. 2020, 11, 1436

    2020

  52. [61]

    Chen et al., Nat

    P.-Y. Chen et al., Nat. Electron. 2018, 1, 297-304

    2018

  53. [62]

    Wu et al., Nat

    J. Wu et al., Nat. Electron. 2022, 5, 635-642

    2022

  54. [63]

    Wang, X.-T

    Z. Wang, X.-T. Zeng, Y. Biao, Z. Yan, R. Yu, Phys. Rev. Lett. 2023, 130, 057201

    2023

  55. [64]

    L. Song, H. Yang, Y. Cao, P. Yan, Nat. Commun. 2022, 13, 5601

    2022

  56. [65]

    H. Yang, L. Song, Y. Cao, P. Yan, Nano Lett. 2022, 22, 3125-3132

    2022

  57. [66]

    Li et al., Natl

    R. Li et al., Natl. Sci. Rev. 2020, 8

    2020

  58. [67]

    Liu et al., Research 2021

    S. Liu et al., Research 2021. 22

    2021

  59. [68]

    Lv et al., Commun

    B. Lv et al., Commun. Phys. 2021, 4, 108

    2021

  60. [69]

    Y. Yang, D. Zhu, Z. Hang, Y. Chong, Sci. China Phys. Mech. Astron. 2021, 64, 257011

    2021

  61. [70]

    Zhang et al., Commun

    X. Zhang et al., Commun. Phys. 2023, 6, 151

    2023

  62. [71]

    Wang, L.-J

    Y. Wang, L.-J. Lang, C. H. Lee, B. Zhang, Y. D. Chong, Nat. Commun. 2019, 10, 1102

    2019

  63. [72]

    Hohmann et al., Phys

    H. Hohmann et al., Phys. Rev. Res. 2023, 5, L012041

    2023

  64. [73]

    Di Ventra, Y

    M. Di Ventra, Y. V. Pershin, C.-C. Chien, Phys. Rev. Lett. 2022, 128, 097701

    2022

  65. [74]

    Liu, Y.G.N., Wei, Y., Hemmatyar, O. et al. Light Sci Appl 2022,11, 336

    2022

  66. [75]

    Z. Guo, Y. Wang, S. Ke, X. Su, J. Ren, H. Chen, Adv. Phys. Res. 2023, 3, 2300125

    2023

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.