{"id":"33cea57a-9416-45ec-875b-d6fb75feee4e","arxiv_id":"2506.20632","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"A coherently controlled superposition of two operation orders on an orbital angular momentum photon measures rotation angle with precision scaling 1/(4ml), demonstrated up to an enhancement factor of 2317.","lead":"This experiment uses a photonic quantum switch, where the polarization of light controls the order of two optical operations, to measure tiny rotation angles with precision that improves as the product of two operation sizes. The setup achieved an enhancement factor over 2,000 and measured a 0.025 degree rotation to 0.0105 arcseconds with about 72 million photons.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed super-Heisenberg scaling hinges on counting the l-th order Q-plate as l unit gates; with the query-count or generator-spread resources of Refs [47-49], the scheme saturates but does not surpass the Heisenberg limit.","rationale":"The central derivation of W_QS and the phase 4mlθ is mathematically sound; the experimental fringes and RMSE are consistent with the predicted CRB. The soft spot is the interpretation of this precision as 'surpassing the Heisenberg limit.' The Heisenberg limit is not 1/Ng for any arbitrarily defined Ng; it is 1/(2Δh) for the probe's generator. Here Δh_QS = 2ml (for ΔL_z≈0), giving δθ ≥ 1/(4ml), which is exactly the demonstrated precision. The 'nonlinear enhancement' is simply that the Q-plate creates a large generator spread at the cost of a high-order optical element. Whether this counts as a linear resource is a convention; Refs [47-49] show that when resources are counted properly (energy, number of applications of elementary generators), nonlinear schemes do not beat the Heisenberg limit. The paper's Discussion cites these references but does not apply them to its own resource count. The reader's weakest_assumption identified resource accounting as the key issue, but the specific examples given (counting the Q-plate as one gate) would not reduce the apparent advantage; the more precise statement is that the query count and the state-preparation lever are conflated. The conditional verdict is appropriate: the result is likely correct as a demonstration of the gear effect, but the 'super-Heisenberg' framing is not established.","tokens_in":10918,"tokens_out":30379,"duration_ms":314288,"concrete_test":"Analytical check: compute the quantum Fisher information for the equivalent standard sequence without a switch — prepare the OAM superposition (|+l⟩+|−l⟩)/√2, apply the rotation D_{2mθ}, and perform the optimal projective measurement on the OAM state. If the QFI equals 16m^2l^2, then the hybrid switch is metrologically equivalent to a standard OAM interferometer with the same lever l; the only advantage is that the polarization readout is simpler. Then recompute the 'surpassing Heisenberg' comparison using the number of query applications m (not m+l) as the resource: δθ = 1/(4ml) matches the Heisenberg limit 1/(2m ΔL_z) with ΔL_z = l, so the experiment demonstrates the Heisenberg limit rather than going beyond it.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In the Results section, after Eq. (1), the paper defines Ng = 2(m+l) and uses it to claim that the scaling 1/(4ml) 'dramatically surpasses the linear Heisenberg limit as 1/Ng.' This is the load-bearing step. Yet the same section's parameter-based HUP, δθ·Δh ≥ 1/2, yields Δh_QS = 2mΔL_z/ℏ + 2ml; for the experimental input (ΔL_z≈0), the bound is δθ ≥ 1/(4ml), which the scheme saturates. Thus the scheme does not exceed the Heisenberg limit for its own generator; it realizes a generator of size 2ml and measures it optimally. The 'super-Heisenberg' factor relative to 1/Ng is an artifact of combining the number of unknown-gate queries (m) with the OAM-shift lever (l) into a single gate count. The paper does not justify why the l-th order Q-plate should be counted as l unit gates for the purpose of the Heisenberg comparison, nor does it engage with the resource measures of Refs [47-49], which address exactly this issue. The central claim of a nonlinear, resource-linear precision enhancement is therefore conditional on this accounting convention.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript reports a photonic experiment realizing a hybrid quantum switch in which a rotation gate on orbital angular momentum is combined with OAM displacement gates in a coherently controlled order, using photon polarization as the control qubit. The authors derive the controlled-order evolution W, show that the control qubit acquires a phase 4mlθ, and estimate θ from the projective probability P(θ) = 1/2[1−cos(4mlθ+φ0)]. They report the quantum Fisher information 16νm²l² and the Cramér-Rao bound δθ ≥ 1/(4√ν ml), and they present six experimental configurations whose measured RMSE tracks this bound within a factor of about 2.5, including a reported absolute precision of 0.0105 arcsec with 7.16×10⁷ photons. The central claim is that this constitutes a nonlinear precision enhancement that surpasses the Heisenberg limit while consuming only linearly increasing resources, based on the gate count N_g = 2(m+l).","tokens_in":11132,"tokens_out":13690,"duration_ms":146453,"significance":"If the main claim is upheld, this is a valuable experimental demonstration that coherently controlled order of operations can encode a product m×l into a single-qubit phase, and the common-path, round-trip interferometric realization with polarization-deflection compensation is technically impressive. The derivation of P(θ) and of the Fisher information is essentially parameter-free except for the phase offset φ0, and the experimental data track the predicted curve across six configurations, with the factor-of-2.5 gap plausibly attributed to mechanical noise. The paper also engages, at least superficially, with the literature on nonlinear metrology and resource counts [47-49]. However, the headline 'super-Heisenberg' claim is conditional on the resource convention N_g = 2(m+l), and the manuscript's own generator-spread calculation shows that the scheme saturates, rather than violates, the parameter-based Heisenberg bound for its own generator. The significance of the result therefore depends on whether the O(l) counting of the high-order Q-plate can be physically justified, or whether the claims are appropriately softened.","major_comments":[{"comment":"The claim that the precision 'dramatically surpasses the linear Heisenberg limit as 1/N_g' is load-bearing and depends entirely on the resource convention N_g = 2(m+l), in which the l-th order Q-plate is counted as l unit gates. The manuscript does not justify why an l-th order Q-plate should be counted as l elementary gates, nor does it engage with the resource measures in Refs. [47-49], which are precisely about whether nonlinear metrology can beat the Heisenberg limit under fair accounting. Under the generator-spread measure used in those references, the authors' own calculation gives Δh_QS = 2mΔL_z/ℏ + 2ml, and the parameter-based uncertainty relation δθ·Δh ≥ 1/2 then yields δθ ≥ 1/(4ml), which is exactly the Cramér-Rao bound that the scheme saturates. Thus the scheme does not surpass the Heisenberg limit for its own generator; the claimed super-Heisenberg factor is an artifact of measuring the resource by m+l while measuring precision by the generator size 2ml. The authors should either derive the O(l) cost of an l-th order Q-plate from a concrete physical model, or soften the claim to a nonlinear enhancement under a stated gate-counting convention.","section":"Results (Eq. (1) and following paragraph); Discussion"},{"comment":"The experimental evidence for the 1/(4ml) scaling rests on six configurations, and five of them have m = 2l, so that 4ml = 8l² over the initial portion of the scaling curve. A fit over those points cannot distinguish 1/(ml) from 1/l², and the sixth point (m = 8, l = 128) is the only one that breaks the m = 2l relation. To support the claimed two-parameter scaling, the authors should add configurations with fixed l and varying m (or vice versa), or at least report the fitted slope and its uncertainty when the (8,128) point is removed. Without this, the statement that the RMSE 'vanishes with the speed 1/4ml' is only weakly constrained by the data.","section":"Results, Fig. 4"},{"comment":"The headline numbers 'enhancement factor as high as 2317' and 'ultimate precision 0.0105 arcsec' are presented without a clear definition of the baseline against which the enhancement is measured. It is not specified whether 4ml is compared with a single rotation pass, with the multi-pass generator Δh_MP = 2mΔL_z/ℏ, or with the earlier schemes of Refs. [34,35], and the normalization (per photon, per gate, or per unit energy) is not stated. Since these numbers appear in the abstract as the main quantitative claims, the baseline and normalization must be defined when the enhancement factor is introduced.","section":"Abstract; Introduction; Results"}],"minor_comments":[{"comment":"There are several typos and formatting errors, including 'parallels the the commutation relation' in Results, 'twifold rotating operations' in the Fig. 2 caption, and 'DA TA A V AILABILITY' in the data availability heading.","section":"Throughout"},{"comment":"The abstract states a normalized precision of approximately 10⁻⁴ rad per photon, while the main text and the Fig. 4 discussion give 4.3×10⁻⁴ rad per photon; use one value consistently.","section":"Abstract; Results; Fig. 4"},{"comment":"The text says the experimental fitting curve is worse than the Cramér-Rao bound by a factor of approximately 2.5, while the Fig. 4 caption says 2.4; make the two numbers consistent.","section":"Results; Fig. 4 caption"},{"comment":"The sentence that a Mach-Zehnder-based scheme [36] 'renders QFI only one quarter one-sixteenth of that in our scheme' is ambiguous; specify whether the factor is 1/4, 1/16, or a range, and state the assumptions under which the comparison is made.","section":"Discussion"},{"comment":"The manuscript repeatedly refers to Supplementary Material Sections I-IV for derivations of Δh_QS, the Fisher information, the HRP behavior, and the Jones-matrix compensation, but the supplementary material is not included with the arXiv submission; the published version must include it and ensure all referenced equation numbers match.","section":"Supplementary Material references"},{"comment":"The term 'hybrid quantum SWITCH' is introduced with 'a little abuse of terminology'; a sentence distinguishing the common-path polarization-controlled superposition from genuine indefinite causal order in spacetime would prevent misinterpretation.","section":"Introduction; Results"}],"recommendation":"major_revision","confidential_remarks":"The experimental work appears solid and the derivation of the 4ml phase is internally consistent, but the super-Heisenberg claim is the main risk and is not yet established under a fair resource measure. I would support publication after the authors either justify the O(l) counting of the high-order Q-plate from a physical model or revise the abstract and conclusions to claim a nonlinear enhancement under a stated gate-counting convention rather than a surpassing of the Heisenberg limit. The limited scaling coverage in Fig. 4 and the undefined enhancement baseline should also be addressed before resubmission."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The genuinely new thing here is the experiment: a co-linear, round-trip hybrid quantum switch that puts a 128th-order OAM shift in coherent superposition with an unknown rotation, plus a clever polarization-deflection compensation to keep the control qubit alive. The six measured configurations track the predicted 1/(4ml) scaling within a factor of about 2.5, and the theory is parameter-free — the only fitted quantity is the phase offset. That part is solid and worth taking seriously.\n\nThe soft spot is the resource accounting that underlies the 'dramatically surpasses the Heisenberg limit' claim. The paper defines Ng = 2(m+l) and compares 1/(4ml) to 1/Ng. But the same section shows the generator of their scheme has size Δh_QS ≈ 2ml, and the parameter-based HUP δθ·Δh ≥ 1/2 gives exactly δθ ≥ 1/(4ml), which they saturate. So relative to their own generator, they are at the Heisenberg limit, not beyond it. The nonlinearity is real only if the l-th order Q-plate is counted as l unit gates. If you count it as one gate, the scaling in m is still 1/m and the advantage reduces to a constant factor. The Discussion cites Refs [47-49] but then just asserts 'considering the quantum resources are defined as the number of used gates in the evolution.' That is precisely the point in dispute. A revision should justify this convention or soften the super-Heisenberg language.\n\nThe data availability statement says datasets are available upon reasonable request, which is a bit dated but not a fatal issue. The polarization-deflection trick, the HRP round-trip design, and the measured fringes are the real contributions. The paper is for people working on quantum metrology with structured light or indefinite causal order; they will get concrete value from the experimental methods and the crisp derivation of P(θ).\n\nMy take: this deserves a serious referee. The core demonstration is reproducible in principle, the scaling behavior is measured rather than fitted, and the resource-counting question is worth an explicit discussion rather than a desk rejection. I would send it to peer review and ask for a revised framing of the Heisenberg comparison.","headline":"A real experimental advance in rotation metrology via coherently controlled order, with clean theory but a convention-dependent 'super-Heisenberg' claim.","tokens_in":11763,"tokens_out":2366,"would_cite":true,"duration_ms":29266,"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":"Coherently controlling the order of an unknown rotation and known OAM shifts transfers a 4mlθ phase to a control qubit, giving per-photon rotation precision that scales as 1/(4ml), beyond the Heisenberg limit.","keywords":["quantum metrology","indefinite causal order","quantum switch","orbital angular momentum","geometric phase","rotation measurement","Heisenberg limit","photonic experiment"],"falsifier":"Operate the same protocol with l first-order Q-plates in series instead of one l-th order Q-plate: if the precision scaling changes from 1/(4ml) to a weaker law as l grows, or if the per-gate precision with one high-order plate is lower than with l single plates, the resource-counting that underlies the claimed nonlinear advantage is falsified. Equivalently, a calculation of the Fisher information with the physical cost of the OAM shift counted as O(l) energy or O(l) optical elements should reveal whether 1/(ml) is the correct asymptotic scaling per resource.","tokens_in":10652,"feed_emoji":"🌀","tokens_out":6106,"duration_ms":61474,"temperature":0.7,"pith_summary":"The paper demonstrates that putting an unknown rotation and a known orbital angular momentum shift in a coherently controlled order—a 'quantum switch'—transfers a geometric phase 4mlθ to a control qubit, where m counts repeated rotations and l counts the OAM shift in units of ℏ. The rotation angle θ can then be estimated with a per-photon precision that scales as 1/(4ml), a nonlinear improvement over the Heisenberg limit that normally allows only 1/N with N independent queries. The authors build a collinear round-trip photonic setup with Q-plates and Dove prisms that realizes this for rotations, measure the interference fringes predicted by the phase, and obtain an enhancement factor of 2317 in practice, with absolute precision 0.0105 arcseconds using 7.16×$10^{7}$ photons. If correct, this shows that indefinite causal order offers a practical metrological advantage—no entangled probes or nonlinear interactions are needed, only a control qubit and ordinary linear optics.","feed_headline":"Quantum switch lifts rotation precision 2317-fold","feed_subtitle":"Coherently controlling the order of OAM shifts and rotations imprints a 4mlθ phase on a qubit, beating the Heisenberg limit.","key_machinery":"The central mechanism is the hybrid quantum switch, defined by the controlled-ordered evolution W = D^†_{lℏ}D_{2mθ}D_{lℏ}⊗|0⟩⟨0| + D_{lℏ}D_{2mθ}D^†_{lℏ}⊗|1⟩⟨1|, unitarily equivalent to W_QS = D_{2mθ}D_{2lℏ}⊗|0⟩⟨0| + D_{2lℏ}D_{2mθ}⊗|1⟩⟨1|. The key identity is the commutation relation D_{2lℏ}D_{2mθ} = $e^{{i4mlθ}}$D_{2mθ}D_{2lℏ}, which follows from the canonical commutation [θ̂, L̂_z]=iℏ; it transfers the phase 4mlθ to the control qubit. The Q-plate (spin–orbit coupling) implements the OAM shifts conditioned on polarization, and the Dove-prism pairs implement the unknown rotation; the round-trip configuration with a hollow roof prism preserves the rotation and cancels polarization deflection.","core_discovery":"The paper claims that the precision of estimating an unknown rotation angle θ can scale as 1/(4ml) per photon, where m is the number of rotation operations (in units of pairs of Dove prisms) and l is the OAM shift in units of ℏ, by placing the rotation D_{2mθ} between two opposite OAM shifts D_{±lℏ} in a superposition of orders controlled by a polarization qubit. Because the displacement operators satisfy D_{2lℏ}D_{2mθ}=$e^{{i4mlθ}}$D_{2mθ}D_{2lℏ}, the controlled-order evolution is unitarily equivalent to a quantum switch whose net effect is to imprint a relative phase 4mlθ on the control qubit, leaving the spatial mode untouched. The phase is read out by a simple polarizer, and the Fisher information is $16νm^{2}$$l^{2}$, which yields the Cramér–Rao bound δθ ≥ 1/(4√ν ml). The experiment realizes this with a collinear round-trip interferometer that avoids the phase drift of separated arms, uses a high-order Q-plate to generate the OAM shift, and achieves a measured RMSE that tracks the 1/(4ml) scaling within a factor of about 2.4.","pith_inferences":["If the gate-counting is instead taken as the number of queries to the unknown rotation only (O(m) rather than O(m+l)), the claimed nonlinear advantage over the Heisenberg limit may be less dramatic, suggesting the comparison should be made against the physical cost of generating a high-order OAM shift.","The fact that the phase is read out on the control qubit and the OAM mode is discarded suggests a general recipe: any unknown unitary can be converted to a measurable phase by sandwiching it between conjugate displacements in a coherently controlled order, provided the generator has a discrete spectrum.","The robustness to input state suggests a variant where the same enhancement is obtained with thermal or mixed states, which would strengthen the practical case.","Testable extension: apply the same switch to phase estimation in a fiber loop, or to gyroscope rotation sensing, and measure whether the 1/(ml) scaling persists under loss."],"forward_implications":["A rotation angle can be measured with precision scaling 1/(4ml) per photon using only linearly many physical gates, beating the Heisenberg limit 1/N_g with N_g = 2(m+l).","The scheme requires no specially tailored input state—even a scrambled speckle works—and no photon–photon interaction, because the enhancement comes from the coherent control of order.","The same mechanism transfers to any shift parameter whose generator has a discrete spectrum, including phase estimation.","The measured fringe period shrinks as 1/(ml), confirming the nonlinear enhancement for six (m,l) combinations.","The experimental precision (0.0105 arcsec with 7.16×10^7 photons, normalized 4.3×10^-4 rad per photon) surpasses the indefinite-evolution OAM scheme and the classical photonic gear."],"supporting_citations":[{"why":"Supplies the theory that quantum metrology with indefinite causal order can enhance geometric phase measurement.","marker":"[28]"},{"why":"Prior experiment demonstrating super-Heisenberg scaling with indefinite gate order, which the present work extends to a hybrid discrete-continuous switch.","marker":"[31]"},{"why":"Baseline indefinite-quantum-dynamics rotation measurement that the paper's experimentally achieved precision surpasses.","marker":"[34]"},{"why":"Classical photonic-gear scheme for rotation measurement, used as a comparison for normalized precision.","marker":"[35]"},{"why":"Establishes the parameter-based Heisenberg uncertainty relation δθ·Δh ≥ 1/2 that frames the paper's scaling argument.","marker":"[41]"},{"why":"Provides the Q-plate method for optical spin-to-orbital angular momentum conversion used to implement the OAM shifts.","marker":"[42]"},{"why":"Shows that high-order OAM states up to 10010ℏ are experimentally available, supporting the feasibility of large l.","marker":"[32]"}],"fun_headline_variants":["Quantum switch boosts rotation precision nonlinearly","Nonlinear quantum switch beats Heisenberg limit","Quantum switch achieves 1/4ml precision scaling","2317-fold precision boost from quantum switch","Quantum switch drives nonlinear metrology gain"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The claimed nonlinear enhancement over the Heisenberg limit relies on counting the total number of gates as N_g = 2(m+l), with an l-quantum OAM shift charged as l unit gates; if the physical cost of a single l-th order Q-plate instead scales like one gate independent of l, or like a resource proportional to the maximum OAM transferred, the advantage over the Heisenberg limit reduces or disappears.","fun_headline_variants_meta":{"raw":{"variants":["Quantum switch boosts rotation precision nonlinearly","Nonlinear quantum switch beats Heisenberg limit","Quantum switch achieves 1/4ml precision scaling","2317-fold precision boost from quantum switch","Quantum switch drives nonlinear metrology gain"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000388,"raw_usage":{"total_tokens":2114,"prompt_tokens":1081,"completion_tokens":1033,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":697,"completion_tokens_details":{"reasoning_tokens":966}},"tokens_in":697,"tokens_out":1033,"duration_ms":8205,"temperature":1.0,"reasoning_tokens":966,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T22:46:05.459340+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Operate the same protocol with l first-order Q-plates in series instead of one l-th order Q-plate: if the precision scaling changes from 1/(4ml) to a weaker law as l grows, or if the per-gate precision with one high-order plate is lower than with l single plates, the resource-counting that underlies the claimed nonlinear advantage is falsified. Equivalently, a calculation of the Fisher information with the physical cost of the OAM shift counted as O(l) energy or O(l) optical elements should reveal whether 1/(ml) is the correct asymptotic scaling per resource.","supporting_citations":[{"cited_title":"& Chiribella, G","cited_arxiv_id":null,"evidence_quote":"Supplies the theory that quantum metrology with indefinite causal order can enhance geometric phase measurement."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Prior experiment demonstrating super-Heisenberg scaling with indefinite gate order, which the present work extends to a hybrid discrete-continuous switch."},{"cited_title":"& Zeng, G","cited_arxiv_id":null,"evidence_quote":"Baseline indefinite-quantum-dynamics rotation measurement that the paper's experimentally achieved precision surpasses."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Classical photonic-gear scheme for rotation measurement, used as a comparison for normalized precision."},{"cited_title":"super-heisenberg","cited_arxiv_id":null,"evidence_quote":"Establishes the parameter-based Heisenberg uncertainty relation δθ·Δh ≥ 1/2 that frames the paper's scaling argument."},{"cited_title":"& Paparo, D","cited_arxiv_id":null,"evidence_quote":"Provides the Q-plate method for optical spin-to-orbital angular momentum conversion used to implement the OAM shifts."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows that high-order OAM states up to 10010ℏ are experimentally available, supporting the feasibility of large l."}],"review_version":1}