{"id":"09c1ce45-7e72-4e33-aa49-7deb03e286f3","arxiv_id":"2411.12059","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Electrically tunable partial photon blockade is observed in dipolar waveguide polaritons, with an extracted blockade radius above 4 micrometers.","lead":"This paper demonstrates a partial photon blockade in electrically biased waveguide polaritons on a semiconductor chip, showing photon anti-bunching that can be tuned with a gate voltage. It reports a blockade radius of several micrometers, comparable to Rydberg systems, which could be a stepping stone to chip-based photonic quantum gates.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Blockade-radius extraction relies on a contact-interaction single-mode model whose applicability to a 5 μm waveguide with micrometer-range dipolar interactions is not established.","rationale":"The reader's weakest_assumption correctly identifies the single-mode blockade model and the calibrated κ as load-bearing. My concern is more specific: even if κ is correct for a single mode, the subsequent step from U to g_dd via Eq. 3 presupposes a contact interaction inside a well-defined mode area. Because the inferred R_b (3.4–4.4 μm) is not small compared to the transverse mode width (≈5 μm), the contact approximation is suspect for a finite-range dipolar potential. The supplementary explicitly acknowledges that the finite laser spot would make the effective pulse longer and the density lower, which biases the extraction toward underestimation of g_dd and R_b; this is a conservative effect and not the central risk. The proposed test is a direct calculation using already-available mode profiles and the interaction form from the group's prior work (Ref. 18). It would settle whether the contact-area conversion is quantitatively valid. The qualitative claim of a partial, electrically tunable photon blockade is supported by the detuning-dependent g2 data and the V=0 control, so the paper's core observation stands. The uncertainty concerns the magnitude of the nonlinearity and the comparison to Rydberg polaritons. Thus the reader's CONDITIONAL verdict remains appropriate; my read does not change it.","tokens_in":14325,"tokens_out":14307,"duration_ms":149511,"concrete_test":"Compute U_eff explicitly: take the waveguide mode profile ψ(y) from Fig. S4 (or the FDE solver), the longitudinal pulse envelope, and the dipolar potential V(r) used in Ref. 18 for the same sample, then evaluate the two-body matrix element U_eff = ∫∫ |ψ(r1)|^2 |ψ(r2)|^2 V(|r1−r2|) d²r1 d²r2 for two polaritons in the fundamental mode. Compare U_eff with the value 2 g_dd/A obtained from Eq. 3 using the reported g_dd and A. If they differ by more than 30%, the extracted g_dd and R_b must be revised; if they agree, the contact-area conversion is validated for this system.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claims—the two-orders-of-magnitude enhancement of g_dd and the blockade radius R_b ≈ 3.4–4.4 μm—are obtained by converting the measured g^(2)_0 dip into an interaction strength U using the single-mode anharmonic Hamiltonian (Eq. S8) and the calibrated factor κ (Eq. S12). This U is then mapped to g_dd via U_dd = 2 g_dd / A, with A = w τ_p v_g (Eq. 3). That mapping assumes a contact (zero-range) interaction: every pair of polaritons inside the mode area experiences the same energy shift U. However, the extracted blockade radius is comparable to the transverse mode width: R_b ≈ 4 μm while w ≈ 5 μm (mode FWHM 4.9 μm). For a genuine dipole-dipole interaction, V(r) ∝ 1/r^3, the mode-averaged two-body matrix element is U_eff = ∫∫ |ψ(r1)|^2 |ψ(r2)|^2 V(|r1−r2|) d²r1 d²r2, which is not generally equal to 2 g_dd/A. The paper does not provide the explicit form of V or the mode functions used to justify Eq. 3 in this regime. The same contact-area approximation enters the blockade condition Eq. 4 and the derived R_b in Eq. S15. Consequently, the extracted g_dd and R_b carry an unquantified, possibly order-one systematic error. The qualitative observation of a partial blockade is not in question, but the quantitative magnitude of the interaction and the comparison to Rydberg polaritons rest on this unvalidated assumption. The supplementary's note that a larger pulse width would reduce the density is a separate, conservative bias and does not address this issue.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a photon-correlation study of propagating lower-polariton pulses in an electrically gated 5 μm-wide, 200 μm-long GaAs/AlGaAs strip waveguide. At V=2.5 V and an exciton fraction of 68%, the authors observe g^(2)(0)=0.94±0.02 at Δ=−0.36 meV and g^(2)(0)=1.030±0.015 at Δ=+0.23 meV, which they interpret as partial blockade and anti-blockade. The effect weakens when the exciton fraction is reduced to 31% and is absent at V=0. Using κU/γ≈1−g^(2)_0,min (Eq. 2), the mode area A=wτ_p v_g (Eq. 3), and κ=0.61 from a single-mode Lindblad model, they extract g_dd≈4.0±1.4 and 3.6±2.2 meV μm², a blockade radius R_b=3.4–4.4 μm (Eq. S15), and a condition (Eq. 5) predicting full blockade in a narrower waveguide.","tokens_in":182,"tokens_out":9185,"duration_ms":117135,"significance":"If valid, the result is significant: it would place chip-integrated, voltage-reconfigurable polariton nonlinearities in a regime previously limited to Rydberg-atom systems, with a two-order-of-magnitude enhancement over unpolarized polaritons. Strengths include the controlled comparison across voltages and exciton fractions, the flat V=0 baseline, explicit treatment of HBT cross-talk, and independent order-of-magnitude support from earlier energy-shift and transmission experiments on the same sample. The main caveat is that all quantitative figures—the enhancement factor, g_dd, and R_b—depend on reducing a finite-range dipole interaction to a single-mode contact shift without a demonstrated validity condition. This concern is directly raised by the stress-test note, and I find it justified. A properly supported quantitative claim would strengthen the paper substantially.","major_comments":[{"comment":"The conversion of the measured g^(2)_0 dip into g_dd and R_b assumes that two polaritons in the mode area experience a single, spatially uniform energy shift U_dd=2g_dd/A. This is equivalent to a contact interaction and is not the same as the mode average of a physical dipole-dipole potential V(|r1-r2|) over the two-particle wavefunction; the two quantities agree only if V is constant across the mode. The claimed R_b of 3.4–4.4 μm is comparable to the 4.9 μm transverse mode FWHM (Fig. S4A), so the interaction can vary appreciably over the mode, yet the manuscript does not give V(r), the mode functions, or a validity condition for Eq. 3. Because Eq. 4 and Eq. S15 inherit this replacement, the extracted g_dd, the two-orders-of-magnitude enhancement, and the comparison with Rydberg polaritons all carry an unquantified systematic error. The qualitative partial-blockade observation is not affected by this issue.","section":"Eq. 3; supplementary Eqs. S14-S15"},{"comment":"The calibration of κ and the pulse-integrated g^(2)_0 use a single-mode anharmonic Hamiltonian with a Fourier-limited pulse. The supplementary itself notes that the effective pulse is broadened by the finite laser spot, giving τ̃_p = sqrt(τ_p² + (δ/v_g)²), and the 200 μm propagation channel can introduce additional velocity dispersion or multimode effects. The paper does not quantify how κ and the extracted g_dd change with the actual pulse shape, spot size, or residual mode structure. Since κ enters linearly in Eq. 2, this is a load-bearing uncertainty for the quantitative claims, even though the authors show that the worst-case spot-size effect would lower the inferred density in a conservative direction.","section":"Supplementary Text, 'Blockade model', Eqs. S11-S12"},{"comment":"The prediction of full blockade in a narrower waveguide rests on the scaling g_dd ∝ d|χ_X|⁴ and v_g ∝ (1−|χ_X|²) taken from Ref. 18, combined with a simulated mode width for a 0.5 μm etched waveguide. This is a plausible extrapolation, but it is not experimentally tested in the narrower geometry, and it inherits the same contact-area mapping used in Eq. 3. The abstract and summary should present this as a target for future tests rather than as a demonstrated consequence, or the authors should include a propagation-of-errors estimate for Eq. 5.","section":"Eq. 5 and supplementary Eq. S18"}],"minor_comments":[{"comment":"The measured output pulse length is quoted as ∼5.4 ps, while the model uses τ_p=ℏ/γ (3.1 or 5.7 ps); the relation between these numbers and the choice of τ_p should be stated explicitly.","section":"Caption, Fig. 2C; Materials and Methods"},{"comment":"The quantities γ, v_g, and d are introduced without error bars; the error bars quoted for g_dd should state which inputs were propagated and which were treated as exact.","section":"Main text, around Eq. 2"},{"comment":"There are typos ('meachnsim', 'poalriton') in the supplementary text, and the sentence 'The fit is robust and independent of γ' should be accompanied by the fitting range and the value of b used in Eq. S12.","section":"Supplementary Text, 'Blockade model'"},{"comment":"The bunching value 1.030±0.015 is only about 2σ above unity; the text calls this a 'clear' anti-blockade. I suggest either showing the full detuning scan with error bars in the same figure or softening the language to 'consistent with anti-blockade'.","section":"Fig. 4"},{"comment":"The discarded white peaks and their origin are well explained, but the reader cannot distinguish the number of scans or total integration time; stating the acquisition time per detuning point would help assess the statistical precision.","section":"Methods, HBT cross-talk"}],"recommendation":"major_revision","confidential_remarks":"For the editor: the paper is well within the journal's scope and the qualitative demonstration is believable. The main risk is not fraud but over-interpretation of a reduced model. I would welcome a revision that adds a mode-averaged interaction calculation or, failing that, explicitly demotes the two-orders-of-magnitude and R_b claims to model-dependent estimates. The strongest path to acceptance is to make the finite-range correction quantitative."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nRead the Ordan et al. paper. The genuinely new thing here is the g^(2)(0) measurement: first time anyone has seen anti-bunching/bunching in electrically gated dipolar waveguide polaritons, with a voltage-tunable effect that tracks the exciton fraction and disappears at V=0. That is a solid experimental result and a natural extension of the group's earlier energy-shift and transmission work (Refs 16-18). The qualitative case for a partial blockade is decent: the dip at negative detuning, the bunching at positive detuning, the flat baseline at V=0, and the roughly factor-of-two reduction when the exciton fraction drops from 68% to 31% all hang together.\n\nThe soft spot is the quantitative extraction. The interaction strength g_dd, the two-orders-of-magnitude enhancement claim, and the blockade radius R_b all come from converting the small measured dip (g^(2)_0 ≈ 0.94) into U_dd through a single-mode anharmonic model with a calibrated numerical factor κ, plus inputs (γ, v_g, mode area) that are reported without error bars. Worse, Eq. 3 maps that U to g_dd via a contact-interaction formula U = 2 g_dd / A, while the paper simultaneously defines the blockade radius through a spatially dependent U(R_b) = γ. If the interaction is genuinely dipolar (1/r^3), the mode-averaged two-body matrix element is not 2 g_dd / A when the blockade radius (~4 μm) is comparable to the transverse mode width (~5 μm). The stress-test note has this right: the extracted g_dd and R_b carry an unquantified, possibly order-one systematic error from that step. The full-blockade prediction (Eq. 5) inherits that uncertainty and is a forward extrapolation anyway, not a demonstration.\n\nNone of that invalidates the central observation. But the headline numbers should be treated as model-dependent estimates, and the paper would be stronger if it showed the mode functions and explicitly justified or replaced the contact-area approximation.\n\nWho is this for? People working on polariton nonlinearities and integrated quantum photonics. It deserves a serious referee: the experiment is nontrivial, the control measurements are the right ones, and the modeling question is exactly what a referee should push on. I'd engage with it.\n\nRecommendation: send to review. Require the authors to address the contact-interaction assumption and to give error bars on the inputs. My own verdict would be conditional acceptance after those points are handled.","headline":"First photon-correlation evidence of partial blockade in dipolar waveguide polaritons, with a real caveat on how the blockade radius is extracted.","tokens_in":15242,"tokens_out":1883,"would_cite":true,"duration_ms":17422,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"An electrically gated semiconductor waveguide hosting dipolar polaritons shows a partial photon blockade with a blockade radius up to 4.4 μm, tunable by voltage, and a predicted path to full blockade.","keywords":["photon blockade","dipolar polaritons","waveguide polaritons","quantum correlations","electrical tunability","second-order correlation function","semiconductor quantum wells","integrated quantum photonics"],"falsifier":"A time-resolved Hanbury Brown-Twiss measurement with sub-picosecond resolution, or a measurement in which the excitation spot size and pulse duration are varied independently, would test the pulse-integration assumption: if the extracted $g_{dd}$ changes with spot size or pulse width, the single-mode model is incomplete. Independently, a direct measurement of the density-dependent polariton blueshift at the same gate voltage should give the same $g_{dd}$; disagreement would falsify the blockade-radius extraction.","tokens_in":14140,"feed_emoji":"⚛️","tokens_out":7852,"duration_ms":72367,"temperature":0.7,"pith_summary":"This paper reports an electrically tunable partial photon blockade in dipolar waveguide polaritons on a semiconductor chip. The authors measure the zero-delay second-order correlation function $g^{(2)}(0)$ and find anti-bunching ($0.94 \\pm 0.02$) at negative detuning, bunching ($1.030 \\pm 0.015$) at positive detuning, and a flat correlation at zero voltage, attributing the effect to strong dipole-dipole interactions between field-induced dipoles. From these data they extract a polariton interaction strength $g_{dd} \\sim 4$ meV$\\,\\mu$m$^2$, about two orders of magnitude larger than that of unpolarized polaritons, and a blockade radius of 3.4–4.4 $\\mu$m, exceeding the optical wavelength in the waveguide and approaching values found with Rydberg atoms. If correct, the result would establish a scalable, voltage-reconfigurable solid-state route to two-photon nonlinearities, with a full blockade predicted when the waveguide is narrowed to a mode width of 0.28 $\\mu$m.","feed_headline":"Chip polaritons block photons over 4-micrometer range","feed_subtitle":"Voltage-tunable polaritons in a waveguide show two-photon correlations — a step toward chip quantum circuits.","key_machinery":"The load-bearing object is the electrically induced dipole moment of the polaritons, which turns a weak contact interaction into a long-range dipole-dipole interaction $U_{dd} = 2g_{dd}(|\\chi_X|^2, V)/A$. The argument combines a single-mode anharmonic blockade model—a Lindblad master equation for a Kerr-like polariton ladder—with pulse-integrated $g^{(2)}$ evaluation, relating the depth of the $g^{(2)}(0)$ dip to $U_{dd}/\\gamma$ through a fitted numerical factor $\\kappa$. The blockade radius follows from the condition $U_{dd}(R_b) = \\gamma$, and the electrically tuned exciton fraction $|\\chi_X|^2$ and dipole length $d$ set the strength and range.","core_discovery":"The central claim is that electrically polarizing exciton-polaritons in a planar waveguide produces a photon blockade that is partial in the present device but already visible in photon-correlation statistics, and whose strength and spatial range are controlled by a gate voltage. The signature is the detuning-dependent $g^{(2)}(0)$: anti-bunching below the one-polariton resonance and bunching above it, appearing only when the voltage is applied, with the effect scaling with the excitonic fraction. The paper converts the measured $g^{(2)}_0$ minimum into an interaction strength $U_{dd}$ through the calibrated relation $\\kappa U_{dd}/\\gamma \\simeq 1 - g^{(2)}_{0,\\mathrm{min}}$, and then into a blockade radius $R_b = \\sqrt{2g_{dd}/(\\pi\\gamma)}$ of several micrometers. It further argues that the same parameters imply a full blockade in a half-micrometer-wide waveguide, with a 0.28 $\\mu$m mode and an 8 nm dipole length, for exciton fractions above 0.56.","pith_inferences":["If the blockade radius is genuinely set by the dipole length and can be pushed into the full-blockade regime, the same device family could enable deterministic two-photon gates without ultrahigh-Q cavities, because the interaction range would exceed the wavelength inside the waveguide.","A direct testable extension is to measure $g^{(3)}(0)$: the anharmonic ladder model predicts a specific hierarchy of multiphoton suppression that would distinguish blockade from other nonlinear mechanisms.","The voltage dependence suggests a programmable architecture in which local gate voltages configure two-photon nonlinearities on demand, an extrapolation the paper does not itself demonstrate.","Mapping $g^{(2)}(0)$ as a function of excitation position along the waveguide could directly image the blockade radius and reveal whether it is uniform or modified by the finite grating and ITO geometry."],"forward_implications":["A voltage-tunable partial blockade with radius around 4 $\\mu$m means on-chip photon-photon interactions can be reconfigured electrically rather than fixed at fabrication.","The two-order-of-magnitude enhancement over unpolarized polaritons means the $g^{(2)}$ dip appears at average two-polariton densities about two orders of magnitude lower than in earlier polariton blockade experiments.","Narrowing the waveguide to a mode width of 0.28 $\\mu$m, with an 8 nm dipole length and exciton fraction above 0.56, is predicted to satisfy the full-blockade condition $n > n_b$.","The devices can be integrated with waveguide couplers and switched electrically with GHz bandwidth at below 3 fJ per gate operation, providing a scalable platform for electrically tuned quantum photonic circuits."],"supporting_citations":[{"why":"Supplies the quantum blockade model and Lindblad master-equation framework used to extract $U_{dd}$ from $g^{(2)}(0)$.","marker":"(26)"},{"why":"Provides the minimal-$g^{(2)}$ to interaction-strength relation and a prior polariton blockade measurement that this extraction builds on.","marker":"(14)"},{"why":"Reports quantum correlations from interacting fibre-cavity polaritons and the pulse-integrated $g^{(2)}$ procedure adapted here.","marker":"(15)"},{"why":"Introduces the electrically gated waveguide dipolariton system and its dispersion, the platform used in the experiment.","marker":"(16)"},{"why":"Measures density-dependent energy shifts of dipolar polaritons, the earlier evidence for strongly enhanced interactions.","marker":"(17)"},{"why":"Provides the scaling $g_{dd} \\propto d |\\chi_X|^4$, the device design, and the huge nonlinearity parameters the paper compares with.","marker":"(18)"},{"why":"Provides the weak contact-like interaction baseline for unpolarized polaritons that makes the dipolar enhancement meaningful.","marker":"(13)"},{"why":"Establishes the photon-blockade benchmark in a single trapped atom-cavity system, the effect this paper aims to bring to a chip.","marker":"(2)"},{"why":"Demonstrates strongly interacting Rydberg excitations whose multi-micrometer blockade radii the paper compares with its extracted radius.","marker":"(8)"},{"why":"Shows quantum nonlinear optics with strongly interacting atoms and serves as the Rydberg polariton comparison point.","marker":"(9)"}],"fun_headline_variants":["Chip polaritons show micrometer photon blockade","Voltage-tuned photon blockade in dipolar polaritons","Micrometer blockade radius on a semiconductor chip","Electric field switches photon blockade in waveguides","Dipolar polaritons block photons over microns"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The extraction assumes that a single-mode, Kerr-like blockade model with pulse-averaged correlations correctly converts the small measured $g^{(2)}(0)$ deviations into the interaction strength, so if the waveguide actually carries additional spatial or momentum modes, or if the finite laser spot and velocity spread are not captured by the pulse integration, the reported $g_{dd}$ and blockade radius could be overestimated.","fun_headline_variants_meta":{"raw":{"variants":["Chip polaritons show micrometer photon blockade","Voltage-tuned photon blockade in dipolar polaritons","Micrometer blockade radius on a semiconductor chip","Electric field switches photon blockade in waveguides","Dipolar polaritons block photons over microns"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000643,"raw_usage":{"total_tokens":2973,"prompt_tokens":977,"completion_tokens":1996,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":593,"completion_tokens_details":{"reasoning_tokens":1924}},"tokens_in":593,"tokens_out":1996,"duration_ms":14066,"temperature":1.0,"reasoning_tokens":1924,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T17:57:09.261623+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A time-resolved Hanbury Brown-Twiss measurement with sub-picosecond resolution, or a measurement in which the excitation spot size and pulse duration are varied independently, would test the pulse-integration assumption: if the extracted $g_{dd}$ changes with spot size or pulse width, the single-mode model is incomplete. Independently, a direct measurement of the density-dependent polariton blueshift at the same gate voltage should give the same $g_{dd}$; disagreement would falsify the blockade-radius extraction.","supporting_citations":[],"review_version":1}