{"id":"911cf6b9-fcda-4889-bd5d-35d787f7390c","arxiv_id":"2412.14147","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Polariton condensate dimers on a hyperbolic dispersion show a continuous, angle-controlled transition between evanescent and ballistic coupling.","lead":"This paper demonstrates that pairs of exciton-polariton condensates in a photonic crystal waveguide can switch between two distinct coupling regimes by rotating the angle of the pair relative to the grating. The same platform can act as a molecule-like evanescently coupled system or as a ballistically coupled system with interference fringes, offering a new control knob for polariton simulators.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The evanescent splitting is read as the full spectral span, so trapped inter-condensate modes may be mistaken for molecular levels; the quoted angular and distance decay could be an artifact of the min-max rule.","rationale":"I read the paper as establishing a qualitatively new capability: a single hyperbolic-dispersion photonic-crystal waveguide in which the angle of a two-spot pump independently selects evanescent or ballistic coupling. That qualitative claim is supported by several independent observations: the single-condensate propagation cone in Fig. 2b, the fringe-count scaling with projected distance r_x in Fig. 4c, the phase and velocity maps in Figs. 2f-g, and the mean-field simulations. The quantitative claim, however, rests on the spectral gaps in Fig. 4a, and those gaps are extracted with a rule that explicitly does not identify which spectral features form the molecular pair. SI Section VI admits that multiple trapped modes appear in the gap region, and the analysis procedure then takes the lowest and highest observed peaks as the evanescent splitting. Since the trapped modes are the spectral fingerprint of the ballistic coupling channel, the extracted 'evanescent' gap in the crossover region is exactly where contamination is most likely. A parity-based or node-counting mode assignment would settle this cleanly, and the mean-field model already used in the paper can provide the same check without new experimental data. I therefore agree with the reader's weakest-assumption identification. The concern does not overturn the qualitative narrative, but it does mean the quantitative transition curve should be regarded as conditional until the mode assignment is demonstrated. Since the reader's verdict was already CONDITIONAL, my recommendation is UNCHANGED.","tokens_in":14228,"tokens_out":5008,"duration_ms":51404,"concrete_test":"Re-analyze the energy-resolved PL datasets behind Figs. 3b-g and 4a by classifying each spectral peak according to its spatial parity (or nodal count) about the midpoint between the two condensates. The true molecular bonding and antibonding states should be the pair with zero and one node in the inter-condensate region, whereas trapped modes have additional nodes. Recompute the evanescent splitting using only that parity-identified pair, and compare with the min-max rule. A complementary check is to solve Eq. S5 below threshold for the same (r, θ) grid as Fig. 4a, decompose the spectral density into eigenmodes, and verify whether the lowest/highest peaks coincide with the zero-node/one-node pair; if the two splittings differ by more than the ~100 μeV linewidth at any point used in the transition curve, the quoted decay is contaminated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The quantitative core of the paper is the angle- and distance-dependent spectral splitting in Figs. 3b-g and 4a, which underlies the claimed continuous evanescent-to-ballistic transition. SI Section VI states that when the spectrum contains more than two peaks, the authors 'always considered the energy difference between the lowest and highest energy state observed in the spectrum' as the evanescent mode splitting. This rule is only valid if those extremal peaks are the bonding/antibonding pair of the dimer. Yet the same section acknowledges that additional peaks are trapped modes of the inter-condensate potential. In the ballistic regime those trapped states are the dominant spectral features, so at intermediate and small angles the min-max span will include a molecular state plus one or more ballistic/trapped resonances rather than the two molecular partners. The reported monotonic decay from about 1.4 meV to 0.1 meV and the e^{-r^2} / e^{-sin^2 θ} behavior in Fig. 4a may therefore measure the shrinking width of a multi-mode manifold rather than the splitting of the two coupled condensate states. Because the transition curve is the central quantitative claim of the paper, this extraction rule is load-bearing: if it is wrong, the qualitative geometric-control claim may survive, but the specific decay laws and the location of the crossover are not supported by the presented data.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports experiments and mean-field simulations on pairs of exciton-polariton condensates in a photonic-crystal waveguide with a saddle-like dispersion. Two non-resonant Gaussian pump spots form a condensate dimer, and the angle θ of the inter-spot axis relative to the positive-mass (ballistic) direction is varied. At large θ the condensates are trapped along the negative-mass direction and form an evanescently coupled molecule with a large spectral splitting; at small θ they are ballistically coupled by propagating waves, with interference fringes and phase locking that switches between in-phase and anti-phase as the projected distance changes. The authors support this picture with energy-resolved spectra, real-space images, phase and velocity maps, and mean-field spectral-density calculations, and they propose a compact overlap model for the coupling strength.","tokens_in":14576,"tokens_out":8415,"duration_ms":74025,"significance":"If the central claim is accepted, this is a significant result for polaritonics and analogue simulation: it introduces a single geometric parameter that continuously interpolates between tight-binding-like evanescent coupling and ballistic, phase-locked coupling in a platform with very large effective-mass anisotropy (m_y/m_x ≈ 100) and bound-in-the-continuum polaritons. The paper contains a large body of direct experimental evidence—dispersion measurements, real-space spectra, fringe counting, phase/velocity imaging—and the qualitative angle dependence is shown both experimentally and in simulations. The supplemental material is also transparent enough to state the mode-extraction rule that creates the main weakness, which allows the issue to be identified and fixed. However, the quantitative decay laws for the evanescent splitting are not yet reliably established because of the mode-identification problem and the absence of uncertainties, and the overlap model contains fitted parameters, so it is illustrative rather than a parameter-free validation.","major_comments":[{"comment":"SI §VI states that when the spectrum contains more than two peaks, the authors 'always considered the energy difference between the lowest and highest energy state observed in the spectrum' as the evanescent mode splitting. This rule is only correct if those extremal states are in fact the two partners of the molecular bonding/antibonding doublet. The same section, however, acknowledges additional 'trapped' modes originating from the inter-condensate potential and notes that such modes appear in the ballistic configuration; Fig. 3e–g and Fig. 4a include parameter ranges with more than two peaks. The min-max span will then mix a molecular state with a ballistic or trapped resonance, so the reported decay from about 1.4 meV to 0.1 meV and the claimed e^{-r^2}/e^{-sin^2 θ} dependence in Fig. 4a are not necessarily the evanescent dimer splitting. Because Fig. 4a is the quantitative core of the evanescent-to-ballistic transition, this extraction rule is load-bearing. The authors should re-extract the gaps by tracking the two molecular branches across r and θ, or by fitting the full multi-peak spectrum with an explicit mode model, and should verify in each reported point that the extremal peaks indeed correspond to the bonded/antibonded pair. Error bars and the number of independent acquisitions should accompany the re-analysis.","section":"SI §VI; main Fig. 4a; Fig. 3b–g"},{"comment":"Eq. (4) is presented in the main text as an overlap-integral result with a closed form, but it is not a parameter-free derivation of the measured splitting. The transverse size σ_y is fitted from solutions of Eq. (S5), the outflow wavenumber k_c is taken from the simulated momentum-space photoluminescence, and κ is set by the lifetime. Therefore the statements that the evanescent splitting 'decreases as e^{-r^2}' and 'decreases as e^{-sin^2 θ}' are at least partly a consistency check with a model containing adjustable inputs, not an independent prediction. In addition, SI §VII assumes a factorized Gaussian/plane-wave ansatz and explicitly defers the BiC π-phase-jump topology that the Discussion itself says can affect dimer coupling. The manuscript should state which parameters are measured, which are fitted, and should justify the neglect of the BiC lobe structure quantitatively, or soften the claims based on Eq. (4) to 'consistent with a model' rather than presenting them as measured laws.","section":"Eq. (4); SI §VII"},{"comment":"No error bars, confidence intervals, or numbers of independent measurements are given for the energy gaps in Fig. 4a, nor are peak-position uncertainties reported for the fringe counts in Fig. 4b–c. Without this information, deviations from the monotonic trends and from Eq. (4) cannot be assessed, for example the scatter in Fig. 4c or the spread of splitting values at fixed separation. At minimum, representative error bars from repeated acquisitions or from peak-fitting uncertainties should be added, together with a clear description of the extraction method for each data point.","section":"Fig. 4; data-analysis"}],"minor_comments":[{"comment":"The parameter σ_y appears in Eq. (4) but is defined only in SI §VII; it should be defined in the main text at the point of first use.","section":"Eq. (4)"},{"comment":"The caption of Fig. 3 begins 'aCalculated spectrum...' with a missing space, and there are several similar typographical slips elsewhere; a careful proofreading pass is needed.","section":"Fig. 3 caption"},{"comment":"SI §VI states that the data in Fig. 3g of the main text have been renormalized to the power recorded during acquisition, but the main text does not mention this correction; please specify the panel(s) to which the renormalization applies and how it was performed.","section":"SI §VI"},{"comment":"The new element with respect to Ref. [14], which already reports evanescent coupling of BiC condensates along the grating direction, should be stated more explicitly as the angular control of the evanescent/ballistic crossover; this would help the reader identify the standalone contribution of the present work.","section":"Introduction / Discussion"},{"comment":"The text says the spectral gap at θ = 90° is 'one order of magnitude larger than the polariton linewidth of 100 μeV'; given that the peaks in Fig. 3b–g appear broader than this, the manuscript should specify whether 100 μeV is the single-condensate linewidth or the instrumental resolution, and how the peak positions were determined.","section":"Section II"}],"recommendation":"major_revision","confidential_remarks":"The stress-test concern from the reader's report lands precisely on SI §VI, where the authors openly describe the min-max extraction rule for the evanescent splitting. This is a load-bearing issue because Fig. 4a is the quantitative basis for the claimed transition curve. I do not view the disclosure as evidence of bad faith; on the contrary, it makes the issue fixable. The qualitative story is likely correct and is supported by many independent observations, so I recommend major revision rather than rejection. The main burden on the authors is to re-analyze the spectral gaps with explicit mode assignment and uncertainties and to reframe Eq. (4) as an illustrative model rather than a parameter-free prediction."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a genuinely new experimental result—a single polariton platform where the coupling between two condensates can be steered between evanescent and ballistic by rotating the dimer axis on a hyperbolic dispersion. The qualitative picture is convincing: the spectra, real-space fringes, phase and velocity maps, and mean-field simulations all hang together. The angular dependence of the fringe count and its projection onto r_x is nicely demonstrated in Fig. 4c. Credit where due: this extends the same group's BiC waveguide work from evanescent molecules to a continuously tunable dimer, and the overlap integral Eq. (4) is a useful simple description.\n\nThe soft spot is exactly where the stress-test lands. SI Section VI says that when more than two peaks appear, the evanescent splitting is read as the energy difference between the lowest and highest state in the spectrum. The same section acknowledges that some of those peaks are trapped modes of the inter-condensate potential. In the crossover and ballistic regimes those trapped modes are the dominant features, so the min-max span is likely a manifold width, not the bonding/antibonding splitting. If that is right, the quoted decay laws—e^{-r^2}, e^{-sin^2 theta}—and the location of the crossover are not actually established by the data. The qualitative transition (large splitting at 90°, small features at small angles) survives, but the quantitative curve in Fig. 4a is contaminated.\n\nAlso minor: no error bars on the gaps, no public data or code, and the mean-field model is fitted to the same system, so it functions as a consistency check rather than a prediction. Those are fixable but worth noting.\n\nNet: I'd send this to peer review. The central claim is important enough for the polariton/photonic lattice subfield and the experiments look carefully done. But I'd ask for a reanalysis of the evanescent splitting using a mode-assignment that distinguishes molecular levels from trapped modes, and for error bars. If they can do that, the paper becomes a strong benchmark rather than a suggestive one.","headline":"Nice angle-controlled evanescent-to-ballistic polariton dimer, but the central splitting curve likely mixes trapped modes with molecular levels.","tokens_in":15098,"tokens_out":2026,"would_cite":true,"duration_ms":19314,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"By rotating the axis connecting two exciton-polariton condensates in a photonic crystal waveguide, this paper demonstrates continuous control over whether the condensates couple by tunneling (evanescent) or by propagating waves (ballistic).","keywords":["hyperbolic dispersion","exciton-polariton","photonic crystal waveguide","bound-in-the-continuum","evanescent coupling","ballistic coupling","polariton dimer","saddle dispersion"],"falsifier":"A decisive check would be to spatially resolve the near-field intensity pattern of each spectral peak in the inter-condensate region and confirm that the highest- and lowest-energy states have the symmetric/antisymmetric spatial structure expected for the bonding and antibonding modes; alternatively, one could show that the extracted splitting follows the predicted exponential form only when exactly two modes are present, and deviates when additional trapped modes appear.","tokens_in":14027,"feed_emoji":"🔬","tokens_out":7564,"duration_ms":62593,"temperature":0.7,"pith_summary":"By rotating the axis connecting two exciton-polariton condensates in a photonic crystal waveguide, this paper demonstrates continuous control over whether the condensates couple by quantum tunneling (evanescent coupling) or by exchanging propagating waves (ballistic coupling). The underlying hyperbolic, saddle-shaped dispersion gives opposite effective masses along the two in-plane directions, so a dimer oriented near the evanescent axis behaves like a molecule with a large bonding/antibonding splitting, while a dimer oriented near the ballistic axis shows phase-locked standing-wave resonances. The authors measure the spectral splitting, interference fringe count, and flow patterns as functions of angle and distance, and match them to a mean-field model and a simple overlap-integral coupling parameter. If correct, the result turns a single photonic-crystal chip into a reconfigurable testbed for coupled quantum fluids, bridging tight-binding and delay-coupled oscillator physics.","feed_headline":"Rotating a polariton dimer tunes coupling from tunneling to waves","feed_subtitle":"One photonic crystal waveguide gives continuous geometric control over evanescent and ballistic coupling between two condensates.","key_machinery":"The central machinery is the hyperbolic (saddle) dispersion of the lower polariton branch, ε(k) = (ħ²/2)(k_x²/m_x − k_y²/m_y), whose opposite-sign effective masses along the two in-plane directions determine the coupling anisotropy. This dispersion produces a real-space 'propagation cone' of half-angle θ_T ≈ arctan(√(m_y/m_x)) ≈ 5.7°–10°, outside which ballistic outflow is forbidden; along the evanescent direction polaritons are trapped by the pump-induced potential. The coupling between two condensates is captured by the overlap integral J(r,θ) = cos(k_c r cosθ) $e^{{−κ r cosθ}}$ $e^{{−r² sin²θ/(8σ_y²)}}$, which combines phase matching of propagating waves (cos term), damping from finite polariton lifetime (exponential decay), and confinement of the trapped condensate (Gaussian in the perpendicular direction). Mean-field simulations of the condensate and reservoir (Eq. S5) provide the quantitative spectral densities and real-space intensity and phase maps that match the measured dimer behaviour.","core_discovery":"The paper's central claim is that the inter-condensate coupling mechanism in a polariton dimer is set by the dimer's orientation relative to the grating, because the lowest-energy polariton branch has a hyperbolic dispersion ε(k) = (ħ²/2)(k_x²/m_x − k_y²/m_y) with positive effective mass along x and negative effective mass along y. Negative-mass polaritons are attracted to the pump-induced potential wells and couple evanescently, producing bonding and antibonding modes split by up to ~1.1 meV for small separations; positive-mass polaritons see the pumps as barriers and propagate ballistically between them, forming interference fringes whose number changes by one whenever the projected separation r_x = r cosθ grows by half the condensate wavelength. The authors map the evolution from evanescent to ballistic behaviour as θ decreases from 90° to below 45°, with both mechanisms coexisting near the critical angle set by the propagation cone of the hyperbolic dispersion. They quantify the coupling with the overlap integral J(r,θ) = cos(k_c r cosθ) $e^{{−κ r cosθ}}$ $e^{{−r² sin²θ/(8σ_y²)}}$, which reproduces the observed exponential distance decay of the evanescent splitting and the linear-in-r_x fringe count of the ballistic regime, and they support the interpretation with mean-field simulations of the condensate order parameter.","pith_inferences":["If the geometric control demonstrated here transfers to two-dimensional gratings, the same angle knob could be used to build tighter polariton networks with half the angular range, a possibility the authors sketch in their discussion.","The coupling integral J(r,θ) assumes a factorized, isotropic condensate profile; a more complete treatment that accounts for the observed anisotropic density might predict small angle-dependent corrections to the extracted coupling that could be tested by precision spectroscopy.","Because the bound-in-the-continuum states carry a topological π phase between lobes, placing pumps at specific positions could create phase dislocations in the dimer region; this is an untested consequence that would show up as a discontinuity in the interference pattern."],"forward_implications":["A single photonic-crystal waveguide can act as a reconfigurable polariton lattice where the coupling type and sign are set by the angle between lattice sites, without altering the sample.","The measured exponential distance and angle dependence of the evanescent splitting gives design rules for placing condensates so that they either hybridize strongly or remain essentially independent.","The linear dependence of the ballistic fringe count on projected separation r_x provides a way to encode in-phase/anti-phase synchronization states in the geometry of the pump spots.","Near the critical angle where both coupling mechanisms coexist, the dimer offers a controllable setting to study the competition between mode hybridization and phase locking in driven-dissipative quantum fluids."],"supporting_citations":[{"why":"Supplies the BiC polariton condensate platform and the evanescent dimer behaviour that this paper extends to the full angular range.","marker":"[14]"},{"why":"Defines the photonic crystal waveguide BiC condensate, including the π-phase nodal structure used in the dimer analysis.","marker":"[21]"},{"why":"Provides the ballistic polariton dimer model with interference fringes and standing-wave resonances that the ballistic regime is compared against.","marker":"[19]"},{"why":"Gives the theoretical framework for phase locking and in-phase/anti-phase synchronization in ballistic condensate dimers.","marker":"[18]"},{"why":"Demonstrates the mode-flipping behaviour in ballistic microcavity condensates that the hyperbolic system reproduces.","marker":"[29]"},{"why":"Supplies the Hamiltonian and dispersion parameters for the grated photonic crystal waveguide used in the mean-field calculations.","marker":"[28]"}],"fun_headline_variants":["Angle tunes polariton dimer coupling: tunnel or wave","One rotation flips polariton coupling from evanescent to ballistic","Polariton dimer orientation selects coupling mechanism","Twist the dimer: coupling goes from tunneling to wave flow","Geometric control of polariton coupling: evanescent to ballistic"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The evanescent mode splitting is extracted as the energy difference between the lowest and highest spectral features even when the spectrum contains multiple trapped modes; if those features are not the bonding/antibonding pair of the dimer, the quoted exponential distance and angle decay of the evanescent coupling is contaminated by ballistic resonances.","fun_headline_variants_meta":{"raw":{"variants":["Angle tunes polariton dimer coupling: tunnel or wave","One rotation flips polariton coupling from evanescent to ballistic","Polariton dimer orientation selects coupling mechanism","Twist the dimer: coupling goes from tunneling to wave flow","Geometric control of polariton coupling: evanescent to ballistic"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000811,"raw_usage":{"total_tokens":3622,"prompt_tokens":1073,"completion_tokens":2549,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":689,"completion_tokens_details":{"reasoning_tokens":2467}},"tokens_in":689,"tokens_out":2549,"duration_ms":16020,"temperature":1.0,"reasoning_tokens":2467,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T12:25:31.091800+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive check would be to spatially resolve the near-field intensity pattern of each spectral peak in the inter-condensate region and confirm that the highest- and lowest-energy states have the symmetric/antisymmetric spatial structure expected for the bonding and antibonding modes; alternatively, one could show that the extracted splitting follows the predicted exponential form only when exactly two modes are present, and deviates when additional trapped modes appear.","supporting_citations":[{"cited_title":"Gianfrate, H","cited_arxiv_id":null,"evidence_quote":"Supplies the BiC polariton condensate platform and the evanescent dimer behaviour that this paper extends to the full angular range."},{"cited_title":"Ardizzone, F","cited_arxiv_id":null,"evidence_quote":"Defines the photonic crystal waveguide BiC condensate, including the π-phase nodal structure used in the dimer analysis."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the ballistic polariton dimer model with interference fringes and standing-wave resonances that the ballistic regime is compared against."},{"cited_title":"Ohadi, R","cited_arxiv_id":null,"evidence_quote":"Gives the theoretical framework for phase locking and in-phase/anti-phase synchronization in ballistic condensate dimers."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrates the mode-flipping behaviour in ballistic microcavity condensates that the hyperbolic system reproduces."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Hamiltonian and dispersion parameters for the grated photonic crystal waveguide used in the mean-field calculations."}],"review_version":1}