{"id":"da2bd6ad-f964-4105-aa31-b38f9224608d","arxiv_id":"2506.01124","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Counter-propagating photons in a Rydberg medium show record-long anti-correlation times that grow linearly with optical depth, enabling complete blockade of overlapping one-microsecond pulses.","lead":"This experiment sent pairs of light pulses toward each other through a cloud of ultracold rubidium atoms, making individual photons block each other over an unusually long time window. The result is a record-long photon anti-correlation range of over one microsecond, a step toward photon-by-photon logic gates.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Under the reported pulsed parameters (OD=88, γE=8.5·2π MHz), the paper's own Eq. (4) predicts ~21% residual pulse coincidences for the 1.1 μs pulses, so 'complete photon blockade of entire pulses' is not supported.","rationale":"The paper's core physics—counter-propagating Rydberg polaritons yielding long anti-correlation times that scale linearly with OD—is credible, parameter-free, and supported by full numerical simulations. The most load-bearing weakness is in the operational claim of 'complete photon blockade of entire pulses.' The pulsed experiment's reported parameters (OD=88, γE=8.5·2π MHz) imply τcross≈0.82 μs under the paper's own definition, shorter than the 1.1 μs pulse. The paper's Eqs. (3)-(4) then predict g2pulse(0)≈0.21, meaning a substantial fraction of photon pairs never overlap in the medium. This is a quantitative inconsistency within the manuscript, not a disagreement with external consensus. The reader's focus on spectator photons captures the same general concern about residual coincidences, but the parameter-based calculation is more direct and does not rely on an unverified attribution. A simple recomputation of Eq. (4) with the stated numbers settles the issue. The correct response is to keep the paper conditional: the demonstration of extended interaction range and the scaling law stand, but the full-pulse blockade wording must be reconciled with the stated parameters or revised.","tokens_in":14244,"tokens_out":19910,"duration_ms":195678,"concrete_test":"Recompute g2pulse(0) from Methods B Eq. (4) as 1−erf(√2 ln2 τcross/Tw) using the pulsed parameters stated in the Fig. 3 caption (OD=88, γE=8.5·2π MHz, Tw=1.1 μs). If the result is ≈0.21 (or ≈0.10 if τcross is instead taken from the CW run), compare with the plotted data and with the abstract's 'complete blockade' wording; a value ≥0.1 requires revising the claim and stating which τcross was used in the prediction.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central operational claim is full-pulse blockade ('complete photon blockade of entire pulses'). The pulsed experiment uses OD=88 and γE=8.5·2π MHz. With the paper's definition 2γE as the EIT linewidth, the counter-propagating anti-correlation time is τcross=OD/(2γE)=88/(17·2π MHz)≈0.82 μs. The chosen pulse width is Tw=1.1 μs. Inserting these numbers into the paper's own analytic pulse-blockade formula (Methods B, Eq. 4: 1−g2pulse(0)=erf(τcross/Tσ), with Tσ=Tw/√(2ln2)) gives τcross/Tσ=0.88 and g2pulse(0)≈0.21. This 21% residual is not the spectator floor; it is the fraction of input pairs whose entry times differ by more than the medium transit time, so they never co-propagate through the medium and cannot be blocked. The paper instead justifies the 1.1 μs choice by overlaying the CW anti-correlation curve (OD=72, 2γE=10·2π MHz, τcross≈1.08 μs) in Fig. 3e, but that is not the parameter set of the pulsed measurement. Thus, under the stated parameters, the 'complete' blockade claim is inconsistent with the paper's own model; only ~79% suppression is expected. The residual-floor attribution to spectator photons (Methods A) cannot rescue this discrepancy, and the γE notation in the Fig. 3 caption obscures it.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports an experimental study of photon-photon interactions between counter-propagating Rydberg polaritons in a cold 87Rb ensemble. The authors observe a record anti-correlation time of 1.08 microsecond in continuous wave, scaling linearly with optical depth, in contrast to the square-root scaling for co-propagating photons. They then demonstrate pulse-level blockade for synchronized 1.1-microsecond pulses, report three-photon suppression consistent with pairwise blockade, and support the observations with analytic diffusion-type models and parameter-free numerical simulations. The central claims are the linear-OD interaction range, full-pulse photon blockade, and enhanced three-photon interactions.","tokens_in":14583,"tokens_out":6011,"duration_ms":59352,"significance":"If the central claims hold, the work establishes counter-propagating Rydberg polaritons as a qualitatively new regime of quantum nonlinear optics, with interaction ranges set by the medium length rather than the EIT bandwidth and with tunable timing control over pulse-level interactions. The paper is commendable for presenting analytic expressions with stated limitations, for using no fitted parameters in the numerical comparisons, and for reporting scaling predictions (tau_cross proportional to OD, tau_self proportional to sqrt(OD)) that are directly falsifiable. However, the 'complete blockade' claim needs quantitative reconciliation with the authors' own model, as detailed below.","major_comments":[{"comment":"The claim of 'complete photon blockade of entire pulses' is not supported by the paper's own analytic model under the stated pulsed parameters. With OD=88 and 2 gamma_E = 17 x 2pi MHz (Section II), Eq. (2) gives tau_cross = OD/(2 gamma_E) = 0.82 microsecond; for T_w = 1.1 microsecond, Eq. (4) yields g2_pulse(0) = 1 - erf(tau_cross/T_sigma) = 1 - erf(0.88) = 0.21, i.e., a 21% residual coincidence rate, not the near-complete suppression claimed in the abstract and in Fig. 3g. The authors should report the measured g2_pulse(0) value and either reconcile the model with the data (e.g., by using the true anti-correlation range and including finite-bandwidth effects) or revise the 'complete' wording.","section":"Sec. IV and Methods B, Eq. (4)"},{"comment":"The choice of 1.1 microsecond pulse width is justified by overlaying the CW anti-correlation curve measured at OD=72, 2 gamma_E = 10 x 2pi MHz (tau_cross = 1.08 microsecond), but the pulsed experiment is performed at OD=88, 2 gamma_E = 17 x 2pi MHz, for which tau_cross = 0.82 microsecond. The design analysis should use the pulsed parameter set; as written, the overlay overstates the interaction range relevant to the pulse measurements.","section":"Sec. IV and Fig. 3e"},{"comment":"The residual coincidence floor s_cross = 3.5% is attributed entirely to non-interacting 'spectator' photons, but this attribution is not independently verified. If the residual coincidences contain partially interacting photons, then the extracted blockade fidelity and the 'complete' claim are overestimated. The authors should either provide a direct characterization of the spectator population (e.g., by measuring correlations as a function of transverse mode or polarization) or weaken the claim accordingly.","section":"Methods A"}],"minor_comments":[{"comment":"The counter-propagating diffusion equation is stated as 'we obtain' from Eq. (7) without showing the transformation; since this equation backs the linear-OD scaling and the tau_cross formula, the derivation should be provided or explicitly referenced in the Methods.","section":"Sec. III, Eq. (2)"},{"comment":"The notation for gamma_E is inconsistent: the text defines 2 gamma_E as the EIT linewidth, while the Fig. 3 caption reports gamma_E values without clarifying whether these are half-linewidths. Please define the convention once and use it consistently.","section":"Fig. 3 caption and Sec. II"},{"comment":"The paper does not quote numerical values for g2_pulse(0) or its uncertainty at T_w = 1.1 microsecond. Reporting these values, together with the parameters used in the color-coded analytic line, would make the 'complete' claim quantitatively checkable.","section":"Sec. IV, Fig. 3g,h"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is technically strong and the simulations are parameter-free, but the headline claim of complete photon blockade is contradicted by the authors' own analytic estimate under the stated pulsed parameters. This is fixable by reporting the measured pulse-level correlation and revising the claim; I see no citation or novelty problems."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The headline is simple: this is the first experimental realization of counter-propagating Rydberg polariton interactions, and the core scaling result is genuinely nice. The anti-correlation time for counter-propagating photons grows linearly with OD, reaching 1.08 μs at OD=72, versus the sqrt(OD) scaling for co-propagating photons. That separation of scales is real, it is backed by no-parameter simulations, and it opens a practical route to pulse-level photon blockade that co-propagating geometries struggle to reach. The three-photon data, showing enhanced suppression when a single photon meets a counter-propagating pair, is a solid bonus and matches the pairwise-blockade expectation. I have no serious doubts about the central physics or the experiments.\n\nThe soft spot is the \"complete photon blockade of entire pulses\" claim. The pulsed experiments use OD=88 and γE=8.5·2π MHz, which gives τcross=OD/(2γE)≈0.82 μs. With a 1.1 μs pulse, the paper's own Eq. (4) predicts a residual coincidence fraction of about 21%, not \"near-complete\" suppression. The paper justifies the pulse choice by overlaying the CW anti-correlation curve from OD=72, where τcross≈1.08 μs, but that is not the parameter set of the pulsed measurement. The spectator-floor explanation in Methods A covers part of the residual, but it cannot erase the 21% geometric leakage from pulse pairs that never co-propagate. The text softens to \"near-complete\" in places, but the abstract and discussion say \"complete,\" which is not supported by the paper's own analytic model. This is a wording/interpretation problem rather than a fatal flaw, and it is fixable by reporting the measured g2pulse(0) value and comparing it to the model prediction.\n\nMinor issues: the \"record\" anti-correlation duration is not compared to any prior literature, so the claim is unverified; Eq. (2) is said to be \"obtained\" from Eq. (7) without a derivation, though the step is plausible; the Fig. 3 caption inconsistently writes γE as 8.5·2π MHz while the text uses γE=8.5·2π MHz, and the extracted τcross for the pulsed case is never stated explicitly.\n\nThis paper deserves a serious referee. The experimental work is careful, the simulations are honest, and the scaling result is an important step for the field. The overclaim about complete blockade should be corrected, but that is a revision, not a rejection. I would bring it to a reading group and would cite it for the counter-propagating geometry and OD scaling.","headline":"First counter-propagating Rydberg-polariton experiment with a clean OD scaling result, but the 'complete pulse blockade' claim is weaker than the paper's own Eq. (4) allows under the stated pulsed parameters.","tokens_in":15107,"tokens_out":3082,"would_cite":true,"duration_ms":29642,"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":"Counter-propagating photons block whole pulses for over 1 microsecond","keywords":["quantum nonlinear optics","Rydberg polaritons","photon blockade","counter-propagating photons","electromagnetically induced transparency","photon-photon correlations","three-photon interactions","Rydberg blockade"],"falsifier":"Measure the coincidence floor while deliberately changing the fraction of spectator photons, for example by cleaning the spatial mode or polarization of the probe beams, or by varying the beam waist. If floor coincidences survive after such cleaning, the spectator attribution is wrong; if they vanish in proportion to the spectator fraction, the attribution is confirmed. A complementary check is whether $g^{(2)}_{\\mathrm{cross}}(0)$ follows the predicted $e^{-2\\mathrm{OD}_b}$ plus a constant floor as a function of optical depth.","tokens_in":14072,"feed_emoji":"⚛️","tokens_out":6323,"duration_ms":57447,"temperature":0.7,"pith_summary":"The paper shows that two photons meeting head-on inside a Rydberg-polariton medium interact much more lastingly than photons moving in the same direction, because the counter-propagating pair stays inside the interaction region for the entire traversal of the medium. The measured anti-correlation time grows linearly with optical depth, reaching 1.08 microseconds at optical depth 72, and the same-time coincidence drops to about four percent. Because this range is so long, a 1.1-microsecond pulse fits both inside the interaction window and inside the transmission bandwidth, so synchronized counter-propagating pulses display near-complete suppression of coincident transmission. The authors take this as evidence that timing-controlled, deterministic photon-photon interactions are achievable, and they extend the result to three photons, where a single photon colliding with a counter-propagating pair is suppressed even when pairwise blockade is incomplete.","feed_headline":"Counter-propagating photons block whole pulses for over 1 microsecond","feed_subtitle":"In a Rydberg-atom medium, opposing photon pulses suppress each other's transmission, and timing controls the interaction.","key_machinery":"The load-bearing object is the stationary two-polariton wavefunction in coordinate space $(x_1, x_2)$ and the diffusion-like equation for its symmetric component. For counter-propagation, the relative coordinate $r = x_2 - x_1$ plays the role of propagation coordinate, so the dissipative Rydberg potential $V(r) = r_b^6/(r_b^6 - i r^6)$ depletes the wavefunction over the full medium length, giving $\\tau_{\\mathrm{cross}} = L/v_g = \\mathrm{OD}/(2\\gamma_E)$; for co-propagation the center-of-mass coordinate is the propagation direction and bandwidth diffusion broadens the dip as $\\sqrt{\\mathrm{OD}}$. The same minimal Hamiltonian, extended to a three-component time-dependent equation and a stationary three-photon equation, reproduces the pulsed maps, the V-shaped spatiotemporal dispersion, and the three-photon anti-correlations without fit parameters.","core_discovery":"The central claim is that dissipative Rydberg blockade in a counter-propagating geometry produces an anti-correlation time equal to the full group delay through the medium, $\\tau_{\\mathrm{cross}} = \\mathrm{OD}/(2\\gamma_E)$, in contrast to co-propagating photons whose anti-correlation is set by the EIT bandwidth and grows only as the square root of optical depth. At $\\mathrm{OD} = 72$ the counter-propagating half-width is $1.08(1)\\,\\mu\\mathrm{s}$, more than double the co-propagating $0.48(2)\\,\\mu\\mathrm{s}$, and the same-time correlation is $g^{(2)}_{\\mathrm{cross}}(0) = 0.041(5)$. With $1.1\\,\\mu\\mathrm{s}$ pulses that satisfy both the bandwidth and interaction-range constraints, synchronized pulses show near-complete extinction of coincident transmission, and the pulse-level suppression disappears when the pulses are separated by 1.5 or 3 microseconds. In the three-photon channel, $g^{(3)}(0,0)$ drops to $0.06(4)$ at low optical depth where the pairwise blockade is only partial, showing that the combined cross and self interactions enhance suppression beyond what two-photon correlations would suggest.","pith_inferences":["If the same geometry is run in the dispersive (off-resonant EIT) regime, the long mutual overlap should accumulate a coherent phase over the entire pulse, suggesting a natural extension to deterministic cross-phase gates that the paper does not yet demonstrate.","The linear scaling implies the 1-microsecond range is not fundamental: at higher optical depth or with Rydberg states of larger blockade radius, multi-microsecond interaction windows should be reachable, limited mainly by atomic coherence and the spectator-photon floor.","The V-shaped dispersion visible in the pulsed correlation map is a time-dependent few-body signature; a pulsed three-photon experiment could probe whether the enhanced suppression seen for continuous waves persists when all three pulses are synchronized."],"forward_implications":["A 1.1-microsecond counter-propagating pulse can be almost completely extinguished by a synchronized partner, making pulse timing a deterministic control knob for photon-photon interactions.","Because the interaction range grows linearly with optical depth rather than as its square root, longer or denser media extend the usable interaction time without the bandwidth penalty that limits co-propagating setups.","Three-photon suppression exceeds the independent-pairwise expectation, indicating a resource for multi-photon nonlinear operations beyond two-photon gates.","The measured ratio $\\tau_{\\mathrm{cross}}/\\tau_{\\mathrm{self}} \\approx \\sqrt{\\mathrm{OD}/9}$ defines a concrete window, $\\mathrm{OD} > 9$, in which counter-propagation outperforms co-propagation.","The analytic formulas for pulse transmission and pulse-level correlation provide a design curve for choosing pulse width and optical depth in future gate experiments."],"supporting_citations":[{"why":"Establishes the dissipative Rydberg-polariton blockade measurement, the co-propagating diffusion model, and the correlation-analysis methods this work extends.","marker":"[19]"},{"why":"Theoretically proposed counter-propagating Rydberg photon-photon interactions and gates, supplying the dissipative interaction Hamiltonian used here.","marker":"[26]"},{"why":"Predicted long-range interactions and entanglement of counter-propagating slow-light pulses, motivating the geometry.","marker":"[25]"},{"why":"Supplies the experimental platform, including the time-modulated dipole trap and Rydberg-polariton measurement methods, plus the simulation approach.","marker":"[24]"},{"why":"Provides the multiband few-polariton framework and time-dependent outward-propagation treatment used for the three-photon correlations.","marker":"[17]"},{"why":"Models time-dependent counter-propagating pulse dynamics and the bandwidth limits that set the optimal pulse duration.","marker":"[29]"}],"fun_headline_variants":["Opposing photons block entire pulses for over a microsecond","Counter-propagating photons: full-pulse blockade, timing-controlled","Microsecond-long photon blockade via counter-propagating Rydberg polaritons","Three-photon interactions enhanced by counter-propagating photons","Pulse-wide photon blockade with counter-propagating light"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The residual coincidence floor in $g^{(2)}(0)$, quoted as $s_{\\mathrm{cross}} = 3.5\\%$, is attributed entirely to non-interacting spectator photons; if those residual coincidences instead come from photons that interact only partially, the complete-blockade claim would be weakened and the blockade fidelity extracted from the data would be too high.","fun_headline_variants_meta":{"raw":{"variants":["Opposing photons block entire pulses for over a microsecond","Counter-propagating photons: full-pulse blockade, timing-controlled","Microsecond-long photon blockade via counter-propagating Rydberg polaritons","Three-photon interactions enhanced by counter-propagating photons","Pulse-wide photon blockade with counter-propagating light"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000894,"raw_usage":{"total_tokens":3858,"prompt_tokens":955,"completion_tokens":2903,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":571,"completion_tokens_details":{"reasoning_tokens":2815}},"tokens_in":571,"tokens_out":2903,"duration_ms":21945,"temperature":1.0,"reasoning_tokens":2815,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T11:51:40.391413+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the coincidence floor while deliberately changing the fraction of spectator photons, for example by cleaning the spatial mode or polarization of the probe beams, or by varying the beam waist. If floor coincidences survive after such cleaning, the spectator attribution is wrong; if they vanish in proportion to the spectator fraction, the attribution is confirmed. A complementary check is whether $g^{(2)}_{\\mathrm{cross}}(0)$ follows the predicted $e^{-2\\mathrm{OD}_b}$ plus a constant floor as a function of optical depth.","supporting_citations":[{"cited_title":"Drori, B","cited_arxiv_id":null,"evidence_quote":"Theoretically proposed counter-propagating Rydberg photon-photon interactions and gates, supplying the dissipative interaction Hamiltonian used here."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Predicted long-range interactions and entanglement of counter-propagating slow-light pulses, motivating the geometry."},{"cited_title":"Tiarks, S","cited_arxiv_id":null,"evidence_quote":"Supplies the experimental platform, including the time-modulated dipole trap and Rydberg-polariton measurement methods, plus the simulation approach."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the multiband few-polariton framework and time-dependent outward-propagation treatment used for the three-photon correlations."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Models time-dependent counter-propagating pulse dynamics and the bandwidth limits that set the optimal pulse duration."}],"review_version":1}