{"id":"325d5c3e-b2e4-435c-97b0-88431efec456","arxiv_id":"2509.06373","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":6,"one_line_summary":"Laser-induced, state-resolved loss in interacting Rydberg pairs shifts the Zeno/anti-Zeno exceptional point and freezes selected two-body spin states via a collective Zeno effect.","lead":"Researchers added a tunable, laser-driven loss channel to individual Rydberg atoms and showed that atomic interactions shift the boundary between quantum Zeno and anti-Zeno behavior. The work gives quantum simulators a practical knob for steering open many-body systems and protecting selected spin configurations.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The Markovian loss model may fail at V/γ=86 because Rydberg shifts of |0> (≈V≈6.9 MHz) far exceed the 5P linewidth Γ≈1 MHz, making the 975 nm loss strongly detuned in interacting pair states; the inferred EP shift and Zeno protection rest on this unverified elimination.","rationale":"The reader's weakest_assumption already anticipates off-resonant excitation, interaction-dependent γ, and non-Markovian 5P memory at V/γ=86, so there is substantial overlap. However, the reader's final rationale emphasizes the self-admitted lack of error propagation in the renormalization factors as the main issue. I find that caveat real but less load-bearing: Pu and Pl have ~1% uncertainties, so the resulting shifts are a few percent and would not overturn the qualitative trends. The more serious problem is that the dissipative channel itself is modeled as a state-independent Markovian collapse, while the physical 5P intermediate state has a linewidth Γ≈1 MHz that is much smaller than the interaction shifts V≈7 MHz and V↑≈4 MHz used in the experiments. If the 5P elimination is re-examined, the effective loss from interaction-shifted |0> states is likely suppressed, which would modify the exceptional-point location and the strength of the two-body Zeno effect. This is a load-bearing concern for the central claim because the paper's headline results are quantitative comparisons to Eqs. (1)–(3), not merely qualitative observations. The proposed explicit-5P simulation is straightforward for two atoms and would settle whether the reduced model is quantitatively valid. I still credit the experiment for its clean calibration of γ, the parameter-forward nature of the Lindblad fits, and the clear single-atom versus pair-atom contrast in the Zeno regime; the concern is about the theoretical mapping, not the raw data. The appropriate verdict remains conditional: the paper should be accepted only if the Markovian elimination is shown to hold in the strong-interaction regime or if the reported observables are recomputed with the interaction-dependent loss.","tokens_in":20263,"tokens_out":23528,"duration_ms":226813,"concrete_test":"Simulate the two-atom observables with the 5P state kept explicitly: take H = (Ω/2)Σ_i(|e>_i<0|+h.c.) + wΣ_i(|1>_i<0|+h.c.) + V(|0>_A<1|⊗|1>_B<0|+h.c.) for Fig. 2 (and the analogous V↑/V↓ terms for Fig. 3), collapse operators √Γ|g>_i<e|, choose Ω=√(γΓ) with γ and Γ from the paper, and compare the population loss after 1 μs and P↑ at t=h/2w with the reduced Lindblad model of Eqs. (1)–(3). If the curves differ by more than the quoted error bars at V/γ=86 or for w_c/w0≈3–5, the constant-γ Markovian elimination is invalid in the regime claimed. A cheaper analytical check is to compute the 975 nm scattering rate from |+>↑0 including the interaction shift δ≈V↑ and verify whether it remains close to γ.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central quantitative claims—the interaction-shifted exceptional point in Fig. 2 and the two-body Zeno freeze in Fig. 3—are interpreted through Eq. (1), which assigns a fixed Markovian loss rate γ to every |0>-containing pair state. Physically, the loss is a resonant 975 nm drive on |0>↔|5P> followed by 5P decay at Γ/h≈1 MHz. In the strongly interacting regime the Rydberg exchange energy is V/γ=86, i.e. V/h≈6.9 MHz, and in the spin-protection experiment V↑/h≈4 MHz; both are far larger than Γ. When a pair occupies a state such as |+>↑0, the interaction shifts the Rydberg |0> level by ≈V, while the low-lying 5P state is essentially unshifted, so the 975 nm transition becomes detuned by several linewidths. Adiabatic elimination of 5P then produces a state-dependent effective loss rate roughly γ/[1+(2δ/Γ)^2] rather than the constant γ used in Eqs. (1)–(3). The fixed imaginary potentials -iγ/2 on |0> states therefore overestimate dissipation from interaction-shifted configurations, which can shift the exceptional point and alter the Zeno condition. The omission is not a minor calibration issue: the same strong interactions that produce the claimed enhancement also detune the dissipative channel. Agreement of a constant-γ Lindblad simulation with the data does not by itself validate the elimination when V≫Γ, and no parameter in the paper quantifies this effect.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript reports experiments with 39K Rydberg atoms in optical tweezers, in which a resonant 975 nm laser induces a tunable, state-selective loss channel from the Rydberg state |0⟩=|42S1/2⟩ via coupling to 5P1/2. For atom pairs, the authors study the interplay of this dissipation with microwave coupling w between |0⟩ and |1⟩=|42P3/2⟩ and dipolar exchange V. They observe that V shifts the exceptional point of the effective non-Hermitian Hamiltonian (Eq. 2), leading to interaction-enhanced decay for parameters that would be in the Zeno regime at V=0. They further demonstrate a configuration-selective two-body Zeno effect: with an auxiliary microwave coupling to a lossy state, atom pairs in |↑↑⟩ are protected from spin flips, while the corresponding single-atom dynamics remain unaffected. The paper closes with a theoretical proposal to use the same mechanism in triangular and chain geometries for W-state distillation and dissipative purification.","tokens_in":20675,"tokens_out":18212,"duration_ms":165618,"significance":"The platform is well-suited to the emerging field of engineered dissipation in Rydberg arrays, and the experimental data show clear, reproducible effects. The modeling is careful in several respects: γ, w, and V are calibrated in separate measurements or from known C3 coefficients, the V=0 exceptional point is benchmarked against previous work, and the solid curves in the figures are full Lindblad simulations rather than fits to the target data. If the central interpretational assumption—a fixed Markovian loss rate γ for every |0⟩-containing pair state—is valid in the strongly interacting regime, the results constitute a clean demonstration of interaction-controlled dissipative phase transitions and two-body Zeno physics. However, the validity of that assumption is the main open question, as detailed below; the quantitative claims therefore require additional support before the paper can be accepted.","major_comments":[{"comment":"The Lindblad model assigns a single Markovian loss rate γ to every pair state containing |0⟩, including the interaction-shifted states |+⟩ and |+⟩↑0. However, the 975 nm transition |0⟩→|5P1/2⟩ is resonant only for an isolated atom; in a pair state such as |+⟩, the |0⟩ level is shifted by the dipolar exchange energy V, while the 5P state is essentially unshifted. With V/h=6.88 MHz (Fig. 2f) and V↑/h=4.0 MHz (Fig. 3d) versus the 5P linewidth Γ/h∼1 MHz, the effective loss rate from those shifted states is suppressed by a factor of order [1+(2V/Γ)^2]^{-1}, i.e., by roughly two orders of magnitude under the stated parameters. Since the exceptional point in Fig. 2(e) and the Zeno condition in Fig. 3(d,e) depend directly on the loss rate of these shifted states, the constant-γ model is not obviously valid in the V≫Γ regime. The authors neither measure nor bound this effect, and agreement between a constant-γ Lindblad simulation and the data does not by itself validate the elimination. Please provide a quantitative estimate of the interaction-detuned loss rate, ideally with a control measurement (e.g., loss from |+⟩ with the 975 nm laser detuned by V), and revise the model in Eqs. (2)-(4) to include a state-dependent γeff(δ) if needed.","section":"Main text, Eq. (1) and Figs. 2(e,f,h), 3(d,e)"},{"comment":"The Supplement explicitly states that the statistical variations of the renormalization factors P_u and P_l are not propagated into the renormalized data. Since the renormalized loss probabilities in Figs. 1-3 are the quantities compared with the Lindblad simulations, the shown error bars underestimate the total uncertainty, and the claimed quantitative agreement (e.g., Fig. 2(f,h) and Fig. 3(d,e)) is not fully supported. Please propagate these uncertainties or provide a sensitivity analysis showing that their effect is negligible relative to the reported statistical errors.","section":"Supplement, 'Renormalization of the experimental measurements'"}],"minor_comments":[{"comment":"Typo in the caption: 'Comaprison' should be 'Comparison'.","section":"Supplement Fig. S3"},{"comment":"The validity conditions for the reduction to Eq. (3), stated as |V↑−V↓|≫w,γ and |Δ−V↓|≫w0,γ, are only marginally satisfied for the largest w0 values used in Fig. 3(e) (w0/h=1.25 MHz, |Δ−V↓|≈2 MHz); please comment on whether the reduced Hamiltonian remains quantitatively accurate for those parameters, since the interpretive discussion around Eq. (4) relies on it.","section":"Supplement, derivation of Eq. (3)"},{"comment":"The phrase 'opens possible new routines for dissipative preparation' is awkward; consider rephrasing, for example, 'opens up new possibilities for dissipative preparation'.","section":"Abstract"}],"recommendation":"major_revision","confidential_remarks":"The experimental work appears solid, and the central qualitative observations are likely correct. My main reservation is the unaddressed interaction detuning of the 975 nm loss channel; if the authors can rule it out or incorporate it into the model, the paper would be a strong candidate for publication. The renormalization uncertainty is also worth addressing. I recommend major revision rather than rejection because the issue seems fixable within the manuscript's scope."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"What you should know: this is a solid experimental paper from a capable group. They add laser-induced, state-resolved dissipation to Rydberg tweezer arrays and show that dipolar exchange shifts the Zeno/anti-Zeno crossover, and that a two-body Zeno effect can freeze a chosen spin configuration. The data are not fitted to the model; gamma, w, and V come from independent calibrations, and the V=0 benchmark matches the known exceptional-point location. The theory extensions in the Supplement, including the W-state distillation, are clearly labeled as proposals, not results. I found no circular fitting or invented parameters.\n\nThe soft spot is the Markovian treatment of the loss. The paper writes a constant gamma in Eq. (1) for every state containing |0>, including the pair state |+>= (|01>+|10>)/√2, which carries the dipolar exchange energy V. But the loss goes through the low-lying 5P state, which is essentially unshifted, so in an interacting pair the 975 nm transition is detuned by roughly V. With V/h around 6.9 MHz and the 5P linewidth about 1 MHz, the effective loss from |+> should be strongly suppressed, not constant gamma. The stress-test note pushes this hard, and I think it lands: the quoted exceptional-point shift at V/gamma = 86 and the Zeno condition in the two-body freeze both rely on the constant-gamma model. The qualitative point probably survives--the enhanced loss at small w_c/gamma is mostly the loss of Zeno protection as V detunes the coherent coupling--but the quantitative location of the EP and the effective Zeno threshold could shift.\n\nThat is a real gap, but not a fatal one. The obvious fix is to redo the adiabatic elimination of 5P with the interaction shift included, or to add an explicit state-dependent gamma to the Lindblad simulation and see whether the data still track. The paper's own admission that the renormalization factors' uncertainties are not propagated is minor but worth cleaning up.\n\nThis deserves a serious referee. I would send it out, but I'd ask the referee to push on the state-dependent loss and to require either a quantitative estimate of the detuning effect or an experimental check at smaller V/gamma where the constant-gamma approximation is safer. For a reader working on open Rydberg systems, this is directly useful; I'd cite it.","headline":"A genuinely useful tunable loss channel for Rydberg arrays, with an interaction-shifted exceptional point and a two-body Zeno freeze that are worth a careful look; the main soft spot is a constant-gamma model that likely overestimates loss from interaction-shifted states.","tokens_in":21165,"tokens_out":9285,"would_cite":true,"duration_ms":88531,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81Q12","81V80","81P40"],"pacs":["03.65.Yz","32.80.Ee"],"model":"deepseek-v4-flash","headline":"In Rydberg atom pairs, dipolar exchange shifts the exceptional point between the quantum Zeno and anti-Zeno regimes, turning engineered loss into a selective spin-protection tool.","keywords":["Rydberg atoms","engineered dissipation","quantum Zeno effect","exceptional point","open quantum systems","optical tweezers","dissipative state preparation","spin chain distillation"],"falsifier":"Scan the surviving population after a fixed evolution time as a function of $w_c/\\gamma$ for an interacting pair at $V=86\\gamma$ and compare it to the single-atom benchmark: the claim predicts the loss minimum moves from $w_c/\\gamma=\\sqrt{2}/4$ to about 7, with larger loss than the non-interacting case between the two. A measurement showing identical minima, or a loss curve that is independent of interatomic spacing at fixed $w_c/\\gamma$, would falsify the interaction-shifted exceptional point. In the spin-protection scheme, observing that a single atom freezes just like the pair when $\\Delta\\gg w_0$ would also contradict the claim that the Zeno effect is two-body-collective.","tokens_in":20088,"feed_emoji":"⚛️","tokens_out":7422,"duration_ms":62269,"temperature":0.7,"pith_summary":"This paper establishes that in pairs of Rydberg atoms, the dipolar exchange interaction between Rydberg states can move the exceptional point that separates the quantum Zeno regime, where strong coupling suppresses decay, from the anti-Zeno regime, where decay is enhanced. The authors build a tunable, state-resolved loss channel by coupling the $42S_{1/2}$ Rydberg state to a short-lived $5P_{1/2}$ state with a 975 nm laser, giving a controlled decay rate $\\gamma$. With two atoms, the same microwave coupling that protects a single atom can instead let an interacting pair decay faster, because the interaction shifts the exceptional point. By choosing pseudospin states whose dipolar exchange connects them to a lossy auxiliary state, the authors realize a two-body quantum Zeno effect: an atom pair in the $|{\\uparrow\\uparrow}\\rangle$ configuration is frozen, while a single atom in the same state Rabi-oscillates normally. They argue theoretically that this configuration-selective loss generalizes to spin chains, where it can distil or purify unwanted spin configurations.","feed_headline":"Dipolar exchange shifts the Zeno-to-anti-Zeno boundary","feed_subtitle":"A tunable loss laser plus Rydberg interactions can freeze chosen two-atom spin states, a route to dissipative state preparation.","key_machinery":"The load-bearing object is the state-resolved loss channel: a weak 975 nm laser couples $|0\\rangle=|42S_{1/2}\\rangle$ to the fast-decaying $|5P_{1/2}\\rangle$, producing an effective Markovian decay at rate $\\gamma$ via the collapse operator $L_0=\\sqrt{\\gamma}(|g\\rangle_A\\langle0|+|g\\rangle_B\\langle0|)$. The argument runs through the non-Hermitian Hamiltonian obtained after eliminating quantum jumps, in the pair basis $\\{|00\\rangle,|+\\rangle,|11\\rangle\\}$, whose exceptional points are set by the ratio of collective microwave coupling $w_c=\\sqrt{2}w$ to $\\gamma$ and shifted by the dipolar exchange $V$. For the spin-protection scheme, the central mechanism is the effective correlated loss rate $\\gamma_{\\rm eff}\\approx 4w_0^2/\\gamma$ acting on the triplet pair state $|+\\rangle_{\\uparrow\\downarrow}$, which turns strong dissipation into a projective measurement that confines the pair to the $|{\\uparrow\\uparrow}\\rangle,|{\\downarrow\\downarrow}\\rangle$ Zeno subspace.","core_discovery":"The central claim is that dipolar exchange between Rydberg atoms changes the dissipation dynamics at the level of two-body configurations, not just through density shifts. On the non-interacting benchmark, the exceptional point sits at $w_c/\\gamma=\\sqrt{2}/4$, matching earlier results; with dipolar exchange $V=86\\gamma$, the exceptional point moves to $w_c/\\gamma\\approx 7$, and the experiment sees the corresponding loss minimum shift. The authors also show a configuration-selective two-body quantum Zeno effect in a pseudospin encoding: when the microwave coupling to the lossy auxiliary state $w_0$ dominates over the spin-flip coupling $w$, strong dissipation of the $|+\\rangle_{\\uparrow\\downarrow}$ pair state projects the pair into the Zeno subspace $\\{|{\\uparrow\\uparrow}\\rangle,|{\\downarrow\\downarrow}\\rangle\\}$, freezing $\\langle P_\\uparrow\\rangle$ near one at the $\\pi$ time while a single atom still Rabi-oscillates. They support this with a reduced non-Hermitian Hamiltonian and a simplified effective model $H_{\\rm eff}=w(\\sigma^x_A+\\sigma^x_B)-i\\gamma_{\\rm eff}|+\\rangle_{\\uparrow\\downarrow}\\langle+|$ with $\\gamma_{\\rm eff}\\approx 4w_0^2/\\gamma$, and they show numerically that the same physics in triangular three-atom systems and five-atom chains gives $|\\mathbf{k}|$-selective dissipative distillation of $W$ states.","pith_inferences":["If the mechanism survives at larger fillings, arrays with site-resolved loss could act as autonomous pumps that push arbitrary initial product states toward ferromagnetic or staggered order without coherent feedback.","Since the two-body Zeno freeze relies on the energy matching $\\Delta=V_\\uparrow$, a direct experimental probe of the frozen fraction versus detuning would test the microscopic picture beyond the two extreme limits shown in the paper.","The same level scheme could be read as a driven-dissipative realization of a non-Hermitian tight-binding model by placing the lossy auxiliary states on one sublattice, predicting chiral edge currents in the steady state.","An immediate testable extension is to apply the pair-protection protocol to two atoms with $V_\\uparrow\\neq V_\\downarrow$, which should create direction-dependent Zeno shadows in the loss pattern."],"forward_implications":["At fixed microwave coupling, increasing dipolar exchange $V$ pushes the pair across the exceptional point into the anti-Zeno regime, so interaction strength alone can switch loss on.","The same dissipation channel that leaves a single atom freely oscillating freezes an interacting pair in its initial spin configuration, so correlated Zeno protection is a genuinely two-body effect.","In spin chains, the inferred imaginary spin-exchange term $-i\\gamma_{\\rm eff}(\\sigma^x_j\\sigma^x_{j+1}+\\sigma^y_j\\sigma^y_{j+1}+I_jI_{j+1}-\\sigma^z_j\\sigma^z_{j+1})/4$ makes anti-aligned neighboring spins decay while preserving aligned configurations, enabling dissipative purification of ferromagnetic states.","In triangular three-atom and five-atom ring settings, tuning the detuning $\\Delta$ to $-E_k=-2V\\cos k$ selectively damps specific Bloch $W_k$ states, allowing distillation of chosen single-excitation states.","The tunable 975 nm loss channel provides a building block for open-system quantum simulation with Rydberg arrays, including dissipative phase transitions and non-Hermitian many-body physics."],"supporting_citations":[{"why":"Supplies the benchmark exceptional point $w/\\gamma=1/4$ for a non-interacting dissipative two-level system, against which the interaction shift is measured.","marker":"[17]"},{"why":"Establishes the PT-symmetry-breaking phase transition at an exceptional point in a qubit, providing the non-interacting reference.","marker":"[18]"},{"why":"Provides the theoretical link between the exceptional point and the quantum Zeno to anti-Zeno crossover.","marker":"[22]"},{"why":"Gives the effective correlated loss rate $\\gamma_{\\rm eff}\\approx 4w_0^2/\\gamma$ used to reduce the two-body dynamics to a Zeno subspace.","marker":"[24]"},{"why":"Demonstrates quantum Zeno dynamics in single Rydberg atoms, the single-particle precursor of the pair-level freeze.","marker":"[26]"},{"why":"Formalizes Zeno subspaces, the structure used to describe the protected $\\{|{\\uparrow\\uparrow}\\rangle,|{\\downarrow\\downarrow}\\rangle\\}$ manifold.","marker":"[29]"},{"why":"Shows two-body loss suppression via the quantum Zeno effect in reactive molecules, the conceptual precedent for configuration-selective loss.","marker":"[42]"},{"why":"Provides the Rydberg tweezer platform methods and measurement renormalization used in the experiments.","marker":"[61]"},{"why":"Justifies eliminating quantum jumps to obtain a non-Hermitian Hamiltonian that shares the exceptional point of the full Lindblad equation.","marker":"[65]"}],"fun_headline_variants":["Rydberg interactions shift the Zeno boundary","Tunable loss laser freezes selected Rydberg spin pairs","Dissipative distillation of spin states in Rydberg arrays","Engineered loss reveals interaction-driven Zeno shift","Laser loss tunes Rydberg Zeno to anti-Zeno regime"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper assumes the 975 nm laser acts as a pure, state-resolved Markovian loss on the $|0\\rangle$ state with a fixed rate $\\gamma$, and that all other Rydberg states, including the $42P$ states, have negligible decay; if off-resonant excitation or interaction-dependent $\\gamma$ becomes significant at the small spacings used ($V/\\gamma=86$), the inferred exceptional-point shift and the Zeno protection would need revision.","fun_headline_variants_meta":{"raw":{"variants":["Rydberg interactions shift the Zeno boundary","Tunable loss laser freezes selected Rydberg spin pairs","Dissipative distillation of spin states in Rydberg arrays","Engineered loss reveals interaction-driven Zeno shift","Laser loss tunes Rydberg Zeno to anti-Zeno regime"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000201,"raw_usage":{"total_tokens":1412,"prompt_tokens":1010,"completion_tokens":402,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":626,"completion_tokens_details":{"reasoning_tokens":319}},"tokens_in":626,"tokens_out":402,"duration_ms":3985,"temperature":1.0,"reasoning_tokens":319,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T16:17:51.827751+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Scan the surviving population after a fixed evolution time as a function of $w_c/\\gamma$ for an interacting pair at $V=86\\gamma$ and compare it to the single-atom benchmark: the claim predicts the loss minimum moves from $w_c/\\gamma=\\sqrt{2}/4$ to about 7, with larger loss than the non-interacting case between the two. A measurement showing identical minima, or a loss curve that is independent of interatomic spacing at fixed $w_c/\\gamma$, would falsify the interaction-shifted exceptional point. In the spin-protection scheme, observing that a single atom freezes just like the pair when $\\Delta\\gg w_0$ would also contradict the claim that the Zeno effect is two-body-collective.","supporting_citations":[{"cited_title":"Sun and W","cited_arxiv_id":null,"evidence_quote":"Provides the theoretical link between the exceptional point and the quantum Zeno to anti-Zeno crossover."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the effective correlated loss rate $\\gamma_{\\rm eff}\\approx 4w_0^2/\\gamma$ used to reduce the two-body dynamics to a Zeno subspace."},{"cited_title":"Signoles, A","cited_arxiv_id":null,"evidence_quote":"Demonstrates quantum Zeno dynamics in single Rydberg atoms, the single-particle precursor of the pair-level freeze."},{"cited_title":"Facchi and S","cited_arxiv_id":null,"evidence_quote":"Formalizes Zeno subspaces, the structure used to describe the protected $\\{|{\\uparrow\\uparrow}\\rangle,|{\\downarrow\\downarrow}\\rangle\\}$ manifold."},{"cited_title":"Syassen, D","cited_arxiv_id":null,"evidence_quote":"Shows two-body loss suppression via the quantum Zeno effect in reactive molecules, the conceptual precedent for configuration-selective loss."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Rydberg tweezer platform methods and measurement renormalization used in the experiments."}],"review_version":2}