{"id":"cfa758a3-cd6d-4a6b-800c-7c3a700f0583","arxiv_id":"2509.02209","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A theory paper shows indefinite causal order in a two-cavity Jaynes-Cummings setup can entangle non-interacting fields and swap a photon between cavities without changing the atom.","lead":"This paper models a single atom passing through two cavities in a superposition of two orders, an indefinite causal order setup. It shows this can entangle two cavity fields that never interact and can swap a photon between cavities while leaving the atom unchanged.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Central 'photon interchange with probability one' claim is only a postselected branch: unconditional probability is <1, and Sec. V's series claim of probability 1 is mathematically impossible for integer n.","rationale":"The reader's verdict (CONDITIONAL) is appropriate, but for a reason that goes beyond the stated weakest assumption about path decoherence. The most load-bearing problem is internal to the ideal calculation: the paper's signature 'probability equal to one' result is not actually derived. The ICO state in Eq. (30) is a postselected branch, and a direct calculation shows the unconditional success probability is below 1 and approaches 1/2 for large n, not 1. Moreover, the series comparison in Sec. V contains a mathematically false statement: the plotted probability can never equal 1 for integer n because it requires simultaneous resonant passage at two incommensurate Rabi frequencies. This does not necessarily invalidate the conclusion that ICO outperforms fixed order, but it means the manuscript's own derivation does not support the abstract's deterministic claim. The entanglement results in Sec. VI appear largely correct and are valuable, so a conditional acceptance with required revisions (clarify postselection, correct Sec. V, and rephrase the abstract) is the right outcome.","tokens_in":13635,"tokens_out":12567,"duration_ms":140196,"concrete_test":"Compute the two quantities from the paper's own formulas. (1) Series: plot/maximize P_series(n,T)=sin^2(g√(n+1)T) sin^2(g√n T) for n=1 and n=5; verify max<1, disproving the Sec. V claim that the probability reaches 1. (2) ICO: for n=m=1, γ_nT=π/2, evaluate the absolute probability P = ||(|0>_c<0|⊗|e><e|)|Ψ(t)>||^2 using Eq. (28) and the Appendix coefficients; it is ≈0.40, not 1. If the authors instead intend the conditional probability P/(P_control=0), state that explicitly and revise the abstract.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Load-bearing concern: The headline claim (abstract) that ICO can 'interchange one photon between both cavities with a total probability equal to one, without changing the quantum state of the atom' is not established by the calculations. The only ICO expression for such a process is Eq. (30), which is conditional on the control qubit being found in |0>_c and the atom in |e>. For the simplest nontrivial case n=m=1 and γ_nT=π/2, the absolute probability of this branch from Eq. (28) is P = (1/2) sin^2(π/2 √(n/(n+1))) ≈ 0.40, not 1. It tends to 1/2 for large n, never reaching 1. Unless 'total probability' means probability within the postselected subspace, the abstract overstates the result. Separately, Sec. V states that in series the |e,n+1,n-1> probability 'can even reach the value of 1 only when n=m>0.' From Eq. (23) that probability is sin^2(γ_nT) sin^2(γ_{n-1}T); equality to 1 would require γ_nT and γ_{n-1}T to be odd multiples of π/2 simultaneously, i.e. √((n+1)/n) rational, which is false for all integers n. Thus the claimed contrast between ICO and fixed order is not correctly supported in the manuscript, even though the contrast may be true. The central claim therefore needs an explicit statement of postselection and a corrected series calculation before it can be trusted.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies a quantum switch built from a two-level atom and two single-mode cavities. A path control qubit superposes the order in which the atom crosses the cavities; each crossing is governed by the resonant Jaynes-Cummings interaction. The authors derive closed-form states for the two fixed orders (Sec. V) and for the switch after a Hadamard on the control and postselection on the control outcome (Sec. VI). They report two main results: (i) ICO can entangle two noninteracting cavity fields, with a higher linear entropy than any fixed order in some cases, and (ii) ICO can interchange a photon between the cavities with probability one while leaving the atom unchanged, a process claimed impossible for cavities in series. The technical core (Eqs. (10), (18), (23), (28)-(29), (44)-(45)) is an explicit algebraic calculation with no free parameters.","tokens_in":14043,"tokens_out":15147,"duration_ms":162721,"significance":"The derivations are transparent and parameter-free; the entanglement bounds are concrete and could be tested. If the photon-interchange claim were correct in the advertised form, it would be a notable resource. However, as stated the claim is not supported: the 'probability one' is conditional on postselection, and the asserted impossibility in series is false for certain Fock-state occupancies. The entanglement results, especially the always-maximal (for the ground branch) linear entropy for n=m, appear sound and are a useful contribution, but the paper needs correction of the overclaimed statements before acceptance.","major_comments":[{"comment":"The central claim that ICO can interchange one photon 'with a total probability equal to one' is not supported by the equations. Eq. (28) describes the state after the control qubit has been projected onto |0>_c, and Eq. (30) is obtained after additionally postselecting on the atom in |e>; the unconditional probability of this outcome is N_0^2 times the atomic excitation probability, which is less than one. For example, for n=m=1 and γ_nT=π/2 the branch has probability about 0.40. The manuscript should explicitly distinguish conditional from unconditional probability and should not describe the process as deterministic.","section":"Abstract and Sec. VI.A (Eqs. (28)-(31))"},{"comment":"The text states that the probability of |e,n+1,m-1> 'can even reach the value of 1 only when n=m>0'. From Eq. (23) this probability is sin^2(γ_nT)sin^2(γ_{m-1}T). For n=m>0, equality to 1 would require g√(n+1)T and g√nT to be simultaneous odd multiples of π/2, which is impossible because √(n+1)/√n is irrational. The condition for unit probability is instead that √(n+1)/√m is a ratio of odd integers; the simple case n+1=m already works (e.g., n=4, m=5). Hence the asserted contrast with fixed order is wrong as stated, and the fixed-order counterexample also invalidates the abstract's claim that such transfer is impossible in series.","section":"Sec. V, after Eq. (23)"},{"comment":"The assertion that for n=m≥0 and atom detected in |g> one always has SL(ρ0g)=1/2 is used to conclude an advantage over fixed order, but the proof is omitted. Substituting ξ=0 and n=m into Eq. (29) gives the needed simplification (c7=s7=c4=s4=0 and c3+s8=s3+c8), so the claim is true, but a reader should not have to reconstruct it. Please add the two-line derivation or a reference to it.","section":"Sec. VI.A, linear entropy discussion (Eqs. (39)-(41))"}],"minor_comments":[{"comment":"In the off-diagonal term for |n-1><n|, '(c2+c5)|n-1><n|' should be '(c2+s5)|n-1><n|', consistent with the conjugate term.","section":"Eq. (38)"},{"comment":"The color label 'magneta' should be 'magenta'.","section":"Fig. 2 caption"},{"comment":"The notation 'S⟨C1C0|σz|C1C0⟩S' is confusing; use a standard subscript label such as '⟨C1C0|σz|C1C0⟩' with a sentence explaining the state.","section":"Eq. (42)"},{"comment":"The idealized assumptions (no which-path information leakage, perfectly balanced beam splitter, identical traversal times, negligible decoherence) are stated but their experimental impact is not discussed. A short limitations paragraph would make the proposal more complete and is especially relevant for the 'probability one' claims.","section":"Secs. II and VI"},{"comment":"For m=0 or n=0, γ_{-1} appears implicitly in coefficients such as c6 and s6; the convention γ_{-1}=0± should be spelled out explicitly.","section":"Appendix, Eq. (44)"}],"recommendation":"major_revision","confidential_remarks":"To the editor: This is a theory paper with clean algebra, but the headline claim overstates postselection and is contradicted by the series case for n+1=m. I think these are correctable with a revision, not fatal to the core entanglement results. There is also a novelty concern: the comparison 'impossible in fixed order' should be checked against examples such as n=4,m=5 before the authors claim an ICO advantage for photon interchange."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The core quantum-switch/Jaynes-Cummings calculation is basically right, and the platform idea is worth taking seriously. This is the first explicit treatment of ICO in a light-matter JC model, and the closed-form state evolution in Eqs. (28)–(29) checks out. The genuine result is that postselecting on the control and atomic spin can give cavity-field entanglement up to linear entropy 2/3, fixed order caps at 1/2, and the vacuum-state branch gives a constant 1/2. That is a real advantage, and the photon interchange effect is also real: at the special interaction times, the initial |n,n> component cancels, leaving only a coherent superposition of the two one-photon-swap directions with the atom still excited. No series configuration does that.\n\nThe soft spots are in the framing, not the algebra. The abstract says 'always' and 'probability equal to one' without saying these are conditional on the control qubit being measured in |0> and, in the vacuum case, the atom found in |g>. Unconditional probabilities are at most 1/2 for large n. That should be stated upfront. Second, Sec. V claims the series probability for |e,n+1,n−1> 'can even reach the value of 1 only when n=m>0.' That is false: the probability is sin^2(γ_n T) sin^2(γ_{n−1} T), and reaching 1 would require √((n+1)/n) rational, impossible for integer n. It can approach 1 arbitrarily closely, but never hit 1. That's a fixable wording error, but it matters because the advertised contrast leans on it. Third, the comparison is slightly apples-to-oranges: series is a definite Fock-state swap, ICO gives a superposition of the two swap directions. The advantage is real, but the paper should say so explicitly.\n\nThe intended audience is quantum-switch theorists and cQED experimentalists interested in interferometric atom-cavity control. The paper deserves a serious referee: the main derivation is sound, the errors are localized and correctable, and the new capabilities are concrete enough to be worth a revision.\n\nSend to peer review.","headline":"Solid quantum-switch analysis for cQED, but the abstract's headline 'probability-one' and 'always' claims depend on postselection, and Sec. V contains a mathematically false series claim.","tokens_in":14484,"tokens_out":6190,"would_cite":true,"duration_ms":69806,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["42.50.Pq","03.65.Ud"],"model":"deepseek-v4-flash","headline":"The paper claims that a two-level atom traversing two cavities in a quantum superposition of the two temporal orders can entangle the two cavity fields and, at specially chosen interaction times, transfer one photon between them with unit p","keywords":["indefinite causal order","quantum switch","cavity quantum electrodynamics","Jaynes-Cummings model","entanglement generation","photon exchange","linear entropy","atomic interferometry"],"falsifier":"Prepare the atom in the excited state and both cavities in the vacuum state, run the ICO sequence, then postselect on the control qubit in |0> and the atom in the ground state; tomograph the two cavity fields. The paper predicts a Bell state with linear entropy 1/2 for every interaction time gT. Observing linear entropy below 1/2, or any dependence of the postselected state on which-path timing, would falsify the central claim. A second test: with one photon in each cavity, choose γ_nT=(2N−1)π/2 and verify that the excited-atom postselected field has unit total probability of one-photon transf","tokens_in":13568,"feed_emoji":"🔀","tokens_out":12474,"duration_ms":138291,"temperature":0.7,"pith_summary":"Indefinite causal order—having a single two-level atom traverse two cavities in a coherent superposition of \"C0 then C1\" and \"C1 then C0\"—is shown to be a working resource in cavity quantum electrodynamics. The cavity fields never interact directly; the atom is the only mediator. The authors compute the full Jaynes-Cummings evolution and show that after recombining and postselecting the atomic path, the two fields can be left in Bell-type entangled states: for vacuum cavities the postselected atom-ground outcome gives a Bell state with linear entropy 1/2 for essentially all interaction times, while fixed-order arrangements reach at most 1/2 and often less. For one photon per cavity, ICO reaches linear entropy 2/3, beyond the fixed-order ceiling. The paper also claims that, at special interaction times, one photon can be interchanged between the cavities with unit probability while the atom starts and ends in its excited state, something they argue a fixed-order cQED sequence cannot do.","feed_headline":"Indefinite causal order swaps a photon between two cavities","feed_subtitle":"A single atom passing through two cavities in a superposition of orders entangles fields that never interact directly.","key_machinery":"The control qubit is the atom's path degree of freedom: |0>_c means the atom traverses cavity C0 then C1, while |1>_c means C1 then C0. The target operations are resonant Jaynes-Cummings interactions with each single-mode cavity field. After both traversals the evolution is |0>_c⊗U1(T)U0(T)+|1>_c⊗U0(T)U1(T); applying a Hadamard to the path qubit and postselecting on |0>_c turns which-path information into the interference term ⟨C1C0|C0C1⟩ (Eq. 27), which is what generates the entangled field states and the photon-exchange effect.","core_discovery":"The central discovery is that the quantum switch changes the atom-field state qualitatively, not just by superposing two evolutions but through their interference. After the atom leaves the cavities, a Hadamard on the path qubit followed by postselecting on |0>_c produces the coherent sum |C1C0>+|C0C1>; the overlap ⟨C1C0|C0C1⟩ (Eq. 27) is the term that carries the new physics. For equal initial photon numbers n=m≥0, choosing γ_n T=(2N−1)π/2 and detecting the atom in the excited state leaves the cavity fields in a Bell-like superposition of \"one photon moved left\" and \"one photon moved right\", so the atom acts as a shuttle that begins and ends in |e>. For n=m=0, postselecting the atom in the","pith_inferences":["If the unit-probability photon exchange survives a more realistic treatment, the quantum switch could act as a noiseless quantum interconnect between distant cavities, with the control atom's internal state untouched and available for reuse.","The vacuum Bell-state result may extend to coherent or squeezed cavity states because of the linearity of the Jaynes-Cummings interaction, but the paper does not prove this; a numerical test would be straightforward.","A Sagnac-type atomic interferometer with cavities on the loop is one concrete implementation route; ion-trap or circuit-QED analogues with synthetic path degrees of freedom could test the same interference term without atomic beam splitters.","Because the effect rests on Eq. (27), any which-path information or beam-splitter imbalance converts the predicted interference into a classical mixture, so an experimental demonstration must independently verify the coherence of the control qubit."],"forward_implications":["Distributed quantum nodes could become entangled via a passing atom even though the nodes never couple to each other, with the atom's path controlling the order.","Vacuum fields suffice: every successful postselection with the atom in the ground state yields the same Bell state, so entanglement generation from vacuum is robust against the exact interaction time.","The atom can mediate a one-photon transfer between cavities without a final atomic flip, offering a new way to shuttle quantum information between bosonic modes.","ICO changes the atomic inversion curve, creating plateaus in the Rabi oscillations, so the timing of the atom-cavity interaction acts as a coherent control knob.","The same construction extends naturally beyond the Jaynes-Cummings model to other light-matter interaction models, as the authors note."],"supporting_citations":[{"why":"Supplies the Jaynes-Cummings dressed states and Rabi frequencies on which the full evolution and all coefficients are built.","marker":"[23]"},{"why":"Provides the standard Jaynes-Cummings dynamics, Rabi oscillations, and the atomic inversion used in Sec. VI B.","marker":"[24]"},{"why":"The atomic beam splitter that creates the coherent superposition of the two paths, i.e., the control qubit.","marker":"[31]"},{"why":"Defines the quantum-switch formalism in which operations can be applied in a coherent superposition of orders.","marker":"[1]"},{"why":"Introduces correlations with no causal order, the conceptual basis for treating the order as indefinite.","marker":"[2]"},{"why":"Justifies why, without the final measurement on the control qubit, the atom-field state is a statistical mixture of the two orders.","marker":"[32]"},{"why":"Establishes the cQED platform of single atoms interacting with cavity fields that the proposal targets.","marker":"[22]"}],"fun_headline_variants":["Causal superposition lets one atom move a photon between cavities","Vacuum cavities get entangled via indefinite causal order","Quantum switch: photon changes cavity, atom stays unchanged","Bell-like entanglement from two empty cavities"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The atom's two paths must stay coherent and meet at a perfectly balanced beam splitter, with no which-path information leaking into the atom-field interaction; if the paths become distinguishable, the interference term that creates the entanglement and photon exchange disappears.","fun_headline_variants_meta":{"raw":{"variants":["Causal superposition lets one atom move a photon between cavities","Vacuum cavities get entangled via indefinite causal order","Quantum switch: photon changes cavity, atom stays unchanged","Bell-like entanglement from two empty cavities"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000197,"raw_usage":{"total_tokens":1190,"prompt_tokens":723,"completion_tokens":467,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":467,"completion_tokens_details":{"reasoning_tokens":407}},"tokens_in":467,"tokens_out":467,"duration_ms":6092,"temperature":1.0,"reasoning_tokens":407,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T11:46:25.200955+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Prepare the atom in the excited state and both cavities in the vacuum state, run the ICO sequence, then postselect on the control qubit in |0> and the atom in the ground state; tomograph the two cavity fields. The paper predicts a Bell state with linear entropy 1/2 for every interaction time gT. Observing linear entropy below 1/2, or any dependence of the postselected state on which-path timing, would falsify the central claim. A second test: with one photon in each cavity, choose γ_nT=(2N−1)π/2 and verify that the excited-atom postselected field has unit total probability of one-photon transf","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Jaynes-Cummings dressed states and Rabi frequencies on which the full evolution and all coefficients are built."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the standard Jaynes-Cummings dynamics, Rabi oscillations, and the atomic inversion used in Sec. VI B."},{"cited_title":"Cassettari, B","cited_arxiv_id":null,"evidence_quote":"The atomic beam splitter that creates the coherent superposition of the two paths, i.e., the control qubit."},{"cited_title":"Chiribella, G","cited_arxiv_id":null,"evidence_quote":"Defines the quantum-switch formalism in which operations can be applied in a coherent superposition of orders."},{"cited_title":"Oreshkov, F","cited_arxiv_id":null,"evidence_quote":"Introduces correlations with no causal order, the conceptual basis for treating the order as indefinite."},{"cited_title":"Ban, Decoherence of a two-level system in a coherent superposition of two dephasing environments, Quantum Infor- mation Processing 19, 409 (2020)","cited_arxiv_id":null,"evidence_quote":"Justifies why, without the final measurement on the control qubit, the atom-field state is a statistical mixture of the two orders."},{"cited_title":"Gleyzes, S","cited_arxiv_id":null,"evidence_quote":"Establishes the cQED platform of single atoms interacting with cavity fields that the proposal targets."}],"review_version":1}