{"id":"a2ddde99-43d6-4ed4-b0fe-7c4a82f11000","arxiv_id":"2506.18010","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Staggered pulse timing makes standard dynamical-decoupling sequences robust to static ZZ crosstalk on two-colorable qubit topologies, with several-fold slower fidelity decay on fixed-coupler IBM processors.","lead":"Quantum computers lose information while their qubits sit idle, and neighboring superconducting qubits can disturb each other through a static coupling called ZZ crosstalk. This paper shows that running decoupling pulses at staggered times on neighboring qubits suppresses that crosstalk for a broad family of standard pulse sequences, and that the staggered versions protect multi-qubit states several times longer on IBM's fixed-coupler chips.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'any equidistant pi-pulse sequence' claim for Eq. (32) lacks proof: cancellation of pulse-induced cross terms requires an additional alternating-phase condition that the paper never states or verifies.","rationale":"The reader's formal weakest_assumption concerns the error model (1-local plus static ZZ, no generic correlated noise or time-dependent ZZ). That is a legitimate scope limitation, but the protocol is explicitly advertised as robust to static ZZ crosstalk, so non-ZZ noise is future work rather than a flaw in the central claim. The more load-bearing gap is the unproven 'any sequence' generalization. The paper's novelty over Ref. [9] is precisely the extension from XY4 to all equidistant pi-pulse DD sequences; if that extension is only demonstrated for specific sequences, the headline claim is not established as stated. The experimental demonstrations on IBM devices are substantial and provide real evidence that the protocol works for the tested sequences, which is why the appropriate response is to add a condition or proof rather than reject the work. The concrete test either finds a counterexample, forcing a qualified claim, or verifies that the alternating-phase condition is redundant, which would strengthen the paper. Thus the reader's CONDITIONAL verdict remains the right one, and no adjustment is needed.","tokens_in":27895,"tokens_out":15769,"duration_ms":157548,"concrete_test":"Search for a valid single-qubit DD sequence with L=6 or L=8 equidistant pi-pulses that satisfies the 1-local suppression condition Eq. (15) but has nonzero alternating sums sum_j (-1)^{j-1} sin(phi_j) and/or sum_j (-1)^{j-1} cos(phi_j). Using the same DRAG pulse profile as the paper, numerically evaluate the 2-local error matrix Eq. (14) for the CR implementation Eq. (32). If any chi^{ZZ alpha beta}_2 component is nonzero, the 'any sequence' claim fails and the protocol needs an explicit phase-ordering condition. If no such sequence can be found, prove that Eq. (15) plus the equal-delay pi-pulse structure implies the alternating sums vanish, which would close the gap.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that Eq. (32) makes any single-qubit DD sequence of equidistant pi-pulses satisfy the first-order ZZ suppression condition Eq. (16) on a two-colorable topology. The paper supports this with numerical examples (XY4, KDD, UR10/12, RGA64c) and symmetry arguments, but no general proof is given. The cancellation is not automatic from equal delays and pi-pulse composition. For the pulse-induced cross terms, e.g., chi^{ZX}_2 = integral of R^{ZX}_R(t) R^{ZZ}_B(t) dt, the integrand receives contributions only while the red qubit is pulsing. Writing the pulse contribution for a phase phi as I_alpha(phi), cancellation requires a discrete orthogonality condition of the form sum_j (-1)^{j-1} I_alpha(phi_j) = 0 for alpha = X,Y. For symmetric pulses this reduces to alternating sums of sin(phi_j) and cos(phi_j) vanishing. This condition is not implied by the 1-local suppression condition Eq. (15) and is not stated in the paper. The text itself flags the gap: Section I says analytical understanding of pulse staggering is extremely limited and it is unclear whether the approach extends to broader classes of DD protocols, while Section VI B concludes that 'examples above substantiate our claim.' That is induction, not proof. If a valid universal DD sequence violates the alternating-phase condition, Eq. (32) leaves residual first-order ZZ errors and the 'any sequence' claim is false.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript introduces a pulse-timing protocol, CR-DD, that targets static ZZ crosstalk on two-colorable qubit topologies. Given a single-qubit DD sequence composed of equally spaced π-pulses, the proposal is to implement the sequence on the two color classes with a relative time shift: one color uses pulses followed by a delay of one pulse duration, while the other uses the same phase list with the delay preceding each pulse (Eq. (32)). The authors claim that, for any such sequence, this staggered implementation satisfies the first-order ZZ suppression condition Eq. (16) while preserving the sequence's single-qubit robustness, with no increase in pulse number and only a factor-2 cycle-time overhead. The theoretical framework is based on first-order Magnus/control-matrix conditions (Eqs. (15)-(16)), and the claim is supported by numerical studies of the control matrices for XY4, KDD, and a heterogeneous XY4/UR12 composition, as well as by hardware demonstrations on IBM fixed-coupler (Eagle r3) and tunable-coupler (Heron r2) devices with up to 20 qubits. The fixed-coupler experiments show consistently higher survival probabilities for the staggered variants than for simultaneous DD, and the tunable-coupler experiments show little SIM/CR gap, which the authors interpret as evidence that ZZ crosstalk is suppressed at the hardware level.","tokens_in":28205,"tokens_out":33767,"duration_ms":370866,"significance":"If the central universality claim were established, this would be a broadly useful low-overhead method for making a large family of practical DD sequences robust to static ZZ crosstalk. The first-order perturbative framework and the explicit suppression conditions are clean, and the multi-device, multi-embedding, multi-sequence hardware study is a substantial experimental contribution. The padding study and the fixed-versus-tunable-coupler comparison are also valuable and, for the tested sequences, the reported improvements in characteristic fidelity-decay time are striking. However, the paper's headline theoretical assertion — that Eq. (32) works for every equidistant π-pulse sequence — is not proven; the text offers examples and symmetry arguments but no general derivation, and the paper itself concedes the analytical understanding is limited. The experimental conclusions for the tested sequences are credible, but the universality claim, as written, goes beyond what the manuscript establishes.","major_comments":[{"comment":"The central claim that Eq. (32) satisfies the first-order ZZ suppression condition Eq. (16) for every single-qubit DD sequence of equally spaced π-pulses is asserted without proof. A direct first-order calculation using the control-matrix convention of Eq. (6) shows that cancellation is not automatic: the pulse-induced cross terms such as χ^{ZX}_2 and χ^{XZ}_2 reduce to discrete sums of single-pulse integrals, and the condition for their vanishing is not implied by the 1-local suppression condition Eq. (15). The two conditions involve different sign weightings of the pulse integrals; an additional discrete orthogonality or alternating-phase condition on the phase list is required, and this condition is neither stated nor verified in the manuscript. The paper's own statements in Sec. I ('Analytical understanding of pulse staggering ... extremely limited') and Sec. VI B ('The examples above substantiate our claim') confirm that the general claim is supported by induction from examples rather than by a proof. Please either supply a general proof, state and verify the missing condition, or narrow the claim to the class of sequences for which the condition provably holds.","section":"Sec. V B, Eq. (32)"},{"comment":"The abstract states that on fixed-coupler devices the authors 'observe at least a 3× improvement in the fidelity decay rate' from CR-DD, and the introduction repeats an 'approximately 3× to 10× improvement' claim. This is contradicted by Table II: on ibm kyiv, the CR/SIM characteristic-time ratios for UR10 are 2.32 for n=10 and 2.58 for n=20. The quantitative summary should be revised to reflect the device- and sequence-dependent range actually reported in the tables, for example by saying that most tested sequences show improvements of roughly 3× to 10×, with smaller gains for UR10 on ibm kyiv.","section":"Abstract and Table II"},{"comment":"The diagnostic conclusion that tunable-coupler devices have substantially suppressed ZZ crosstalk is inferred from the absence of a large SIM/CR gap on ibm marrakesh. This inference presupposes that, for the specific sequences used, CR-DD removes all first-order ZZ-induced errors while leaving 1-local errors unchanged. Because the general cancellation in Eq. (16) is not proven (see the first major comment), the diagnostic is only as reliable as the unproven cancellation. Please quantify the predicted residual first-order error for the actual sequences and DRAG pulses used on ibm marrakesh — for example, by reporting the numerical value of the left-hand side of Eq. (16) at the end of the CR cycle — and include that alongside the hardware comparison.","section":"Sec. VII F"}],"minor_comments":[{"comment":"The XY4 phase list in Eq. (19) is Φ=(0,π/2,0,π/2), which corresponds to X,Y,X,Y, while Eq. (18) writes the sequence as Y-f-X-f-Y-f-X-f. Please clarify the time-ordering convention used in the product notation so that the two expressions are unambiguously consistent.","section":"Sec. IV B, Eqs. (18)-(19)"},{"comment":"The caption says the padded sequences have cycle duration 'τc = 80τ γ'; this should read 'τc = 80τp', consistent with the main text.","section":"Table III caption"},{"comment":"The characteristic times in Tables I-III are obtained from three-parameter exponential fits (A, γ, c) to survival-probability traces, but no fit uncertainties or goodness-of-fit statistics are reported. Please provide uncertainties on τγ or a goodness-of-fit measure so that the ratios in the tables can be assessed.","section":"Sec. VII A 2, Eq. (38)"},{"comment":"The caption describes the heterogeneous composition with 'red-orange' and 'blue' sequences, but the figure itself does not show a legend or a key matching these colors to the two qubit colors; please add an explicit legend and also state the full UR12 phase list in the caption.","section":"Fig. 3"}],"recommendation":"major_revision","confidential_remarks":"The hardware study is solid and the protocol is clearly promising for the tested sequences, but the headline universality claim is not established in the current text; the missing proof (or a missing stated condition) is load-bearing. I would not accept the paper in its present form. If the authors can supply a proof for the claimed class, or honestly narrow the claim and re-verify the numerical figures and the abstract's quantitative statements, the paper would be a strong candidate for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a legitimate generalization of the pulse-staggering trick from Zhou et al. — from XY4 to the class of equidistant pi-pulse sequences, including heterogeneous pairings and padded variants — and the hardware data are the strongest part. The weak point is the scope claim: Eq. (32) is asserted to work for any such sequence, but the paper supports it with examples and symmetry arguments, not a general proof. The abstract's \"at least 3x\" claim is also not supported by Table II, where UR10 gives 2.32x and 2.58x on ibm_kyiv.\n\nWhat is new: Ref. [9] had the XY4 staggering case. This paper generalizes to KDD, UR10/12, RGA64c, heterogeneous composition (XY4 with UR12), and symmetric/asymmetric padding, and validates on three IBM devices with up to 20 qubits. That is a real extension. The first-order suppression condition derived from the Magnus/control-matrix framework is clean and parameter-free; the protocol is not fit to the data. The control-matrix plots make the cancellation mechanism visible. The tunable-coupler result is also worth having: SIM and CR are nearly equivalent on ibm_marrakesh, consistent with hardware-level ZZ suppression, and the time-averaged comparison makes the fixed-coupler-plus-CR advantage concrete rather than hand-wavy.\n\nSoft spots. The \"any equidistant pi-pulse sequence\" theorem is not proven. The paper itself says in the introduction that analytical understanding of pulse staggering is limited, and Sec. VI.B concludes \"examples above substantiate our claim.\" That is induction. The stress-test note's alternating-phase concern is exactly the right question: cancellation of pulse-induced cross terms such as chi^{ZX} requires more than equal delays and pi-pulse composition, and the paper never states or verifies the needed orthogonality condition. I did not find a counterexample among the sequences tested, and the hardware data are directionally convincing, but the claim as stated is too broad. Fix it by proving the condition or narrowing the claim to sequences satisfying an explicit phase condition. The abstract should be corrected. Also, no code or raw data are provided; for a paper whose value is partly in the hardware protocol, that should be fixed. The citation pattern is fine: Ref. [9] is credited, and the related syncopated-DD work is cited.\n\nWho it is for: experimentalists and compiler people using DD on fixed-coupler hardware, and anyone doing hardware-level crosstalk characterization. The practical impact is contained but real. A serious referee should see it; I would not desk reject. I would ask for the proof or scope fix and data release, then likely accept.","headline":"Legitimate generalization of pulse-staggered DD from XY4 to equidistant pi-pulse sequences, with strong hardware evidence but an unproven 'any sequence' claim and an overstated abstract.","tokens_in":28717,"tokens_out":3012,"would_cite":true,"duration_ms":33530,"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":"Staggering the timing of equidistant π-pulse DD sequences cancels first-order ZZ crosstalk on any two-colorable qubit layout, with no extra pulses and twice the cycle time.","keywords":["dynamical decoupling","ZZ crosstalk","pulse staggering","bipartite qubit topology","bounded control","control matrix","superconducting qubits","state preservation"],"falsifier":"Set up two qubits on a fixed-coupler device with known ZZ coupling, run CR-DD versus SIM-DD with an engineered non-ZZ crosstalk term or a time-modulated ZZ coupling, and measure survival probability or the first-order Magnus term; if CR-DD does not outperform SIM-DD or $\\chi^{ZZ\\alpha\\beta}_{2,e}(\\tau_c)$ fails to vanish, the claimed universality of Eq. (32) is refuted. A simpler laboratory check is to violate one premise—unequal inter-pulse delays or unequal pulse durations—and observe whether the staggered pair still cancels the ZZ error matrix to first order.","tokens_in":27670,"feed_emoji":"⚛️","tokens_out":8786,"duration_ms":85037,"temperature":0.7,"pith_summary":"This paper claims that any dynamical decoupling (DD) sequence built from equally spaced π-pulses can be made robust to static ZZ crosstalk on a two-colorable qubit topology by a simple change in pulse timing: run the sequence on one color with each pulse followed by the inter-pulse delay, and on the other color with each pulse preceded by the same delay. The claim is proven through first-order time-dependent perturbation theory: the staggered pair forces products of control-matrix elements to integrate to zero, satisfying the two-local suppression condition while keeping each sequence's single-qubit decoupling intact. The practical consequence is a software-only upgrade that costs no extra pulses and only doubles the cycle time, and it permits different sequences on the two colors. State-preservation experiments on fixed-coupler superconducting processors with up to 20 qubits show rough 3×–10× increases in fidelity-decay time for staggered variants. On tunable-coupler hardware the staggered and unstaggered versions perform nearly equally, which the paper uses as evidence that hardware-level crosstalk suppression works, while fixed-coupler processors with the staggered protocol still maintain higher survival probabilities in the tests.","feed_headline":"Staggered pulses cancel ZZ crosstalk with no extra pulses","feed_subtitle":"Any equidistant π-pulse DD sequence becomes crosstalk-robust on two-color qubit layouts, at twice the cycle time.","key_machinery":"The machinery is the staggered two-color pulse schedule of Eq. (32), in which one color applies each π-pulse before the free-evolution delay $f_p$ and the other color places the delay before the pulse. This schedule makes the products of single-qubit control-matrix elements $R^{Z\\alpha}_R(t)R^{Z\\beta}_B(t)$ either mutually orthogonal or displacement-antisymmetric over the cycle, so their time integral—the first-order two-local error matrix $\\chi^{ZZ\\alpha\\beta}_{2,e}(\\tau_c)$—vanishes. The equal-delay, equal-duration π-pulse structure preserves each base sequence's own first-order single-qubit suppression, because the staggering only permutes pulse and delay order without changing the local sequence.","core_discovery":"Eq. (32) is the paper's central claim: for any single-qubit DD sequence composed of π-pulses with equal delays, the two-color implementation $D_R = \\prod_{\\phi\\in\\Phi_R}(\\pi)_\\phi f_p$ and $D_B = \\prod_{\\phi\\in\\Phi_B} f_p(\\pi)_\\phi$ suppresses all effective first-order ZZ error terms on a two-colorable graph, while preserving the sequence's original single-qubit robustness. The suppression is shown through the control-matrix condition $\\chi^{ZZ\\alpha\\beta}_{2,e}(\\tau_c)=\\int_0^{\\tau_c} R^{Z\\alpha}_R(t)R^{Z\\beta}_B(t)\\,dt=0$, which the staggered timing produces through orthogonality and displacement (anti)symmetry between the two colors' control-matrix elements. The result holds for homogeneous pairs of sequences (XY4, KDD, UR10, RGA64c) and for heterogeneous pairs (such as XY4 with UR12), and it survives symmetric or asymmetric idle padding around each pulse. The paper identifies the π-pulse-driven toggling of $R^{ZZ}(t)$ between $\\pm 1$ as the feature that makes the cancellation robust for this whole family.","pith_inferences":["If Eq. (32) is as general as claimed, the same timing recipe should transfer to other platforms with bipartite nearest-neighbor coupling and equal-duration pulses, such as trapped-ion chains or neutral-atom arrays, provided their crosstalk is static ZZ.","A testable extension would be to run the SIM-versus-CR comparison on hardware where the dominant correlated noise is not ZZ (e.g., XX or YY crosstalk); the first-order suppression condition is specific to ZZ terms, so the CR advantage should disappear or shrink.","The protocol's diagnostic use is a cheap reverse-engineering tool: a pulse-count-matched pair of SIM and CR experiments gives a direct estimate of whether ZZ crosstalk limits memory fidelity on an unfamiliar device, without a full noise-spectroscopy analysis.","Because the method allows arbitrary equal padding and heterogeneous sequence pairs, DD-aware compilers could insert idle gaps or choose per-color sequences to fit algorithmic scheduling constraints while keeping crosstalk suppression; this scheduling flexibility is implicit in the paper's Eqs. (33)–(34) but is not developed into an algorithm there."],"forward_implications":["Every equidistant-π-pulse DD sequence can be converted into a crosstalk-robust variant at the cost of a factor-2 increase in cycle time and no increase in pulse count.","The two colors need not run the same sequence, so DD schedules can be composed from different base sequences while retaining ZZ suppression.","Equal padding added symmetrically or asymmetrically around pulses preserves the suppression, letting practitioners trade pulse count against wall time when protecting idling qubits.","Comparing SIM and CR variants of the same sequence serves as an indirect ZZ-crosstalk diagnostic: a large performance gap indicates significant static ZZ crosstalk, and near-equal performance indicates hardware-level suppression.","On the tested fixed-coupler devices, CR-DD raised the median characteristic time of survival-probability decay by factors of roughly 3 to 10 relative to SIM-DD, across system sizes from 5 to 20 qubits."],"supporting_citations":[{"why":"introduced pulse staggering for XY4 and the crosstalk-suppression condition that this paper generalizes to all equidistant π-pulse sequences.","marker":"[9]"},{"why":"supplied the high-performing DD sequences and the state-preservation experimental methodology used on multi-qubit superconducting processors.","marker":"[8]"},{"why":"grounds the time-dependent perturbation theory and first-order suppression analysis in the strong-control regime.","marker":"[61]"},{"why":"provides the Magnus expansion used to define and compute the error dynamics that the suppression condition targets.","marker":"[62]"},{"why":"defined Eulerian DD in the bounded-control setting, one of the sequence families whose staggered variants are analyzed.","marker":"[46]"},{"why":"defined the universally robust DD family (URDD), whose UR10 and UR12 sequences are tested as crosstalk-robust variants.","marker":"[44]"},{"why":"supplied the RGA64c sequence discovered by genetic-algorithm search, another tested base sequence.","marker":"[49]"},{"why":"documents parasitic ZZ crosstalk in fixed-coupler superconducting qubit architectures, motivating the error model.","marker":"[56]"},{"why":"describes tunable-coupler hardware designed to suppress ZZ errors, the architecture used for the diagnostic comparison.","marker":"[60]"}],"fun_headline_variants":["Staggered pi-pulses make DD crosstalk-proof on two-color chips","3x better DD fidelity by just retiming pulses on two-color qubits","No extra pulses: staggered timing neutralizes ZZ crosstalk","Any equidistant pi-pulse DD sequence now crosstalk-robust"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The derivation assumes the error Hamiltonian contains only single-qubit terms and static ZZ couplings between adjacent qubits, with equal-duration pulses and equal inter-pulse delays; any significant non-ZZ correlated noise, time-dependent ZZ coupling, or unequal pulse timing can break the first-order suppression condition.","fun_headline_variants_meta":{"raw":{"variants":["Staggered pi-pulses make DD crosstalk-proof on two-color chips","3x better DD fidelity by just retiming pulses on two-color qubits","No extra pulses: staggered timing neutralizes ZZ crosstalk","Any equidistant pi-pulse DD sequence now crosstalk-robust"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000933,"raw_usage":{"total_tokens":4048,"prompt_tokens":1057,"completion_tokens":2991,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":673,"completion_tokens_details":{"reasoning_tokens":2909}},"tokens_in":673,"tokens_out":2991,"duration_ms":21849,"temperature":1.0,"reasoning_tokens":2909,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T18:57:42.519130+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Set up two qubits on a fixed-coupler device with known ZZ coupling, run CR-DD versus SIM-DD with an engineered non-ZZ crosstalk term or a time-modulated ZZ coupling, and measure survival probability or the first-order Magnus term; if CR-DD does not outperform SIM-DD or $\\chi^{ZZ\\alpha\\beta}_{2,e}(\\tau_c)$ fails to vanish, the claimed universality of Eq. (32) is refuted. A simpler laboratory check is to violate one premise—unequal inter-pulse delays or unequal pulse durations—and observe whether the staggered pair still cancels the ZZ error matrix to first order.","supporting_citations":[{"cited_title":"Haeberlen,High Resolution NMR in solids selective averaging: supplement 1 advances in magnetic reso- nance, V ol","cited_arxiv_id":null,"evidence_quote":"supplied the high-performing DD sequences and the state-preservation experimental methodology used on multi-qubit superconducting processors."},{"cited_title":"Efficient Chromatic-Number-Based Multi-Qubit Decoherence and Crosstalk Suppression","cited_arxiv_id":"2406.13901","evidence_quote":"grounds the time-dependent perturbation theory and first-order suppression analysis in the strong-control regime."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"defined Eulerian DD in the bounded-control setting, one of the sequence families whose staggered variants are analyzed."},{"cited_title":"Khodjasteh and D","cited_arxiv_id":null,"evidence_quote":"defined the universally robust DD family (URDD), whose UR10 and UR12 sequences are tested as crosstalk-robust variants."},{"cited_title":"Viola and E","cited_arxiv_id":null,"evidence_quote":"supplied the RGA64c sequence discovered by genetic-algorithm search, another tested base sequence."},{"cited_title":"Rotteler and P","cited_arxiv_id":null,"evidence_quote":"documents parasitic ZZ crosstalk in fixed-coupler superconducting qubit architectures, motivating the error model."}],"review_version":2}