{"id":"ca3f777f-926a-4152-92f7-84d08ae016a5","arxiv_id":"2607.29123","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Electrically controlled flip-flop qubit pairs can implement sqrt(iSWAP) and iSWAP gates with optimized local Rz corrections, but multi-qubit layouts require co-design because spectator qubits shift the entangling landscape.","lead":"This paper simulates two-qubit entangling gates in silicon flip-flop qubit arrays and optimizes electrical pulses to reach infidelities below 0.1%. It shows spectator qubits distort the gates, so designs should optimize pulse settings, local phase corrections, and qubit layout together.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Multi-qubit fidelities in Sec. 6 may be conditional on spectators staying in |0⟩; spectator excitation is never quantified, so the co-design conclusion could rest on an optimistic projection (the 4^N truncation itself is safe due to spin conservation).","rationale":"The reader's weakest assumption concerned the 4^N reduced basis in Appendix A. I checked the model Hamiltonian and do not think that concern lands: all terms conserve total S_z+I_z, so parallel-spin states are in a different symmetry sector and cannot mix with the logical flip-flop subspace; the truncation is exact for the ideal closed dynamics. However, a different and more load-bearing gap exists in Sec. 6. The paper explicitly projects spectator qubits onto |0⟩ to define the effective two-qubit operator and labels this a 'conditional trajectories' restriction, but the final infidelity metric used in Figs. 5-6 is not specified as full-register. If it inherits the projection, any spectator excitation is invisible, so the multi-qubit 'co-design' conclusion and the symmetric-vs-asymmetric advantage could be artifacts of post-selection. The absence of any reported spectator-excitation probability or full-N-qubit fidelity makes this impossible to rule out. This is an addressable missing check, not a demonstrated failure, so the conditional verdict remains appropriate. Credit is due for the open code and the internally consistent Makhlin/local-correction workflow, and the two-qubit isolated results (Sec. 5) are not affected by this concern.","tokens_in":10820,"tokens_out":16351,"duration_ms":178986,"concrete_test":"Re-run the 3-qubit linear-array case with the active pair at one end (the asymmetric layout of Fig. 5) using the reported optimized parameters, and compute: (1) the full 3-qubit infidelity 1−F with U_ideal = sqrt(iSWAP) ⊗ I_3 from Eq. (14) with d=8, and (2) the probability that the spectator qubit is found in |0⟩ at final time. If the full-register infidelity is substantially above 1e−3 or the spectator-excitation probability exceeds ~1e−4, the reported conditional fidelities are not representative and the co-design conclusion requires qualification. The same check on a symmetric layout would show whether the symmetry advantage survives.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Sec. 6 builds the multi-qubit analysis on an effective two-qubit operator obtained by projecting idle spectators onto their |0⟩ state, and the paper itself notes this 'implicitly restricts our analysis to conditional trajectories where the spectator qubits remain unexcited at the final gate time.' It is never stated whether the final infidelities in Figs. 5 and 6 are computed from the full N-qubit propagator or from this projected operator. If the latter, spectator excitations are discarded by construction, making the reported 'minimum achievable infidelity' a conditional success probability rather than the unconditional gate infidelity relevant for fault tolerance. The central claim that symmetric layouts are beneficial and that gates cannot be optimized in isolation would be weakened if spectator leakage is non-negligible, and no spectator-excitation probability or full-register infidelity is reported anywhere. This is distinct from the Appendix A 4^N reduced-basis concern: in Eqs. (5)-(12) every term (Zeeman, hyperfine S·I, orbital, dipole-dipole) conserves the total electron+nuclear spin projection S_z+I_z, so the excluded parallel-spin states are not coupled to the logical dynamics and the truncation is effectively exact for closed evolution. The load-bearing gap is the unquantified spectator excitation channel in the multi-qubit extension.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents a numerical study of two-qubit entangling gates in silicon flip-flop qubit arrays. Using a MATLAB framework (FlipFlopQSim), the authors simulate coherent spin-orbital dynamics under electrical control, classify two-qubit operations via Makhlin invariants, and identify hold-time windows locally equivalent to sqrt(iSWAP) and iSWAP. They then design electrically driven Rz rotations to compensate local phases, reporting full physical infidelities of 4.72e-4 for sqrt(iSWAP) and 6.14e-4 for iSWAP. The multi-qubit section analyzes linear, square, and star arrays and concludes that spectator couplings shift the entangling landscape, so gate optimization must co-optimize pulses, local phases, and geometry; symmetric layouts are consistently better than asymmetric ones. All simulations use the reduced 4^N anti-parallel spin basis.","tokens_in":11155,"tokens_out":6073,"duration_ms":61531,"significance":"The paper's contribution is a concrete computational pipeline and a qualitative design principle: isolated two-qubit optimization is insufficient in multi-qubit flip-flop arrays, and symmetric qubit environments mitigate spectator-induced distortions. Strengths include the openly available code, the use of Makhlin invariants to separate local from nonlocal content, and the comparison of ideal versus physically simulated Rz corrections, which goes beyond bare optimization. If the spectator-excitation and basis-truncation concerns are resolved, the framework will be useful for device design. The spin-conservation structure of the Hamiltonian suggests the 4^N reduction may be exact, but this is not demonstrated in the manuscript; and the multi-qubit fidelity numbers are not yet shown to be unconditional.","major_comments":[{"comment":"The analysis projects idle spectators onto |0> to define a reduced two-qubit operator and notes that this 'implicitly restricts our analysis to conditional trajectories where the spectator qubits remain unexcited.' The manuscript never states whether the reported infidelities are computed from the full N-qubit propagator or from this projected operator. If the latter, spectator excitations are discarded by construction, and the quoted values are conditional success probabilities rather than unconditional gate fidelities. Please report spectator excitation probabilities or full-register infidelity for the optimal operating points.","section":"Sec. 6, Figs. 5-6"},{"comment":"All quantitative results, including the headline values 4.72e-4 and 6.14e-4, are obtained in the reduced 4^N basis with no 8^N cross-check. The excluded parallel-spin configurations are asserted to be 'energetically detuned and weakly coupled,' but no truncation error is quantified. Given the spin-projection conservation of the Hamiltonian in Eqs. (5)-(12), the reduction may be exactly decoupled; this should be stated and ideally verified numerically at the two-qubit level. Without this, the sub-threshold claims are not fully justified.","section":"Appendix A; Sec. 5"},{"comment":"The reported infidelities are minima obtained by optimizing correction angle and hold time with the same fidelity metric used to evaluate performance. This does not invalidate the optimization, but reporting only the minimum is fragile; the paper itself notes strong sensitivity to theta. To support the fault-tolerance claim, the authors should report the infidelity over a finite operating range (e.g., a small neighborhood of the optimal theta and thold) or provide an out-of-sample check for the optimized parameters.","section":"Sec. 5, Eq. (14)"}],"minor_comments":[{"comment":"The expression 'ˆI+' appears garbled; please clarify the identity and formatting throughout the Hamiltonian definitions.","section":"Sec. 2, Eq. (5)"},{"comment":"The relative threshold 0.1 is arbitrary; please state how the classification changes with the threshold or justify the choice.","section":"Eq. (16)"},{"comment":"The left panel's axis label 'piSWAP' should be '√iSWAP'; also consider using consistent notation for the square-root gate in the text and figures.","section":"Fig. 4"},{"comment":"The statement that the two local-correction forms give a 'negligible difference' should be quantified (e.g., maximum infidelity difference) rather than left qualitative.","section":"Sec. 6"},{"comment":"'Energetically detuned and weakly coupled' is imprecise; if the excluded states are decoupled by spin conservation, say so explicitly.","section":"Appendix A"}],"recommendation":"major_revision","confidential_remarks":"The main risk is the unquantified spectator excitation in Sec. 6; this is fixable with additional simulations. The 4^N truncation is likely exact due to spin conservation, so the Appendix wording should be corrected rather than requiring a large 8^N computation, though a small cross-check is desirable. I would be willing to review a revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Take on arXiv:2607.29123: this is a solid numerical study. The new thing is a full simulation pipeline for flip-flop qubit gates that combines Makhlin classification, electrically driven Rz calibration, and a multi-qubit spectator analysis. The authors show convincingly that the entangling interaction's local phases and the spectator environment must be handled together; the symmetric-vs-asymmetric layout result is the most useful design principle here. The code is openly available, and the Hamiltonian truncation to the 4^N subspace is actually safe because every term conserves S_z+I_z, so the excluded parallel-spin states are never populated under closed evolution. That concern can be set aside.\n\nThe real soft spot is the multi-qubit section. The paper explicitly projects spectator qubits onto |0⟩ and says this restricts the analysis to conditional trajectories, but it never reports whether spectator excitations are negligible. If the quoted fidelities in Figs. 5–6 come from the projected two-qubit operator rather than the full N-qubit propagator, they are conditional success probabilities, not unconditional gate infidelities. The qualitative conclusion that symmetric layouts are better may survive even with leakage, but the reader cannot tell. That is a gap; the fix is to report the spectator excitation probability or the full-register infidelity. It is not fatal, but it needs to be addressed.\n\nThe headline infidelities (4.72e-4 and 6.14e-4) are minima from optimization, so they are not independent predictions. The paper is transparent about this, but it would be good to see a validation with a different pulse shape or an out-of-sample point. The Makhlin distance threshold is also somewhat arbitrary; minor.\n\nWho should read this: anyone working on flip-flop or donor-based silicon qubits, and people doing pulse-level optimization for spin qubits. It does not break new theoretical ground, but it is a useful design tool and a reasonable template for co-optimization studies.\n\nRecommendation: send it out. The method is sound, the code is available, and the open questions (spectator leakage, out-of-sample fidelities) are addressable in revision. A good referee should look at it; I'd probably take it after moderate revision.","headline":"Solid, self-contained numerical study of flip-flop qubit gates; the co-design conclusion is plausible, but the multi-qubit fidelity numbers need an explicit statement of whether spectator leakage is included.","tokens_in":11619,"tokens_out":2952,"would_cite":true,"duration_ms":28624,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81P68","81V70"],"pacs":["03.67.Lx"],"model":"deepseek-v4-flash","headline":"The paper claims that high-fidelity entangling gates in silicon flip-flop qubit arrays require co-optimizing control pulses, local phase corrections, and device geometry rather than tuning the interaction in isolation.","keywords":["flip-flop qubits","silicon quantum computing","two-qubit gate optimization","Makhlin invariants","dipole-dipole coupling","quantum control","Rz rotation calibration","spectator qubits"],"falsifier":"Recompute the optimized sqrt(iSWAP) and iSWAP pulse sequences in the full 8^N basis, or experimentally measure population in parallel-spin states at the optimal operating points. If the resulting infidelity exceeds 1e-3, or if observed leakage is non-negligible, the quantitative support for the central co-design claim fails.","tokens_in":10717,"feed_emoji":"⚛️","tokens_out":4339,"duration_ms":39856,"temperature":0.7,"pith_summary":"The paper argues that two-qubit entangling gates in silicon donor flip-flop qubit arrays cannot be made high-fidelity by optimizing the interaction alone; the control pulse, the local single-qubit phase corrections, and the surrounding qubit geometry must be tuned together. Using a full spin-orbital simulator, it maps the electrically induced dipole-dipole interaction onto canonical gates such as sqrt(iSWAP) and iSWAP via Makhlin invariants, then shows that electrically driven Rz rotations can compensate accumulated local phases. With this co-design, composite-gate infidelities reach 4.72e-4 for sqrt(iSWAP) and 6.14e-4 for iSWAP, below the 1e-3 fault-tolerance threshold. Extending to multi-qubit arrays, spectator qubits shift the interaction landscape and the optimal correction angle, and symmetric environments consistently outperform asymmetric ones. A careful reader would care because this outlines a concrete optimization route toward scalable, all-electrical silicon quantum processors.","feed_headline":"Co-design unlocks sub-threshold two-qubit gates in silicon","feed_subtitle":"Co-optimizing pulses, phase corrections, and layout yields infidelity below 1e-3 in flip-flop qubit arrays.","key_machinery":"The argument is carried by three interlocking tools: FlipFlopQSim, a simulator that propagates the coupled electron-nuclear spin-orbital Hamiltonian under realistic time-dependent electric fields; Makhlin invariants, which classify the two-qubit entangling content up to local operations and identify hold-time windows where the dipole-dipole interaction is locally equivalent to sqrt(iSWAP) or iSWAP; and the electrically driven Rz rotation calibrated at the clock-transition bias point, whose hold time maps linearly onto rotation angle with a tunable fine structure. The Makhlin invariants do the work of separating the nonlocal gate from local phases, while the calibrated Rz dataset supplies the","core_discovery":"The central claim is that isolated two-qubit optimization is insufficient: residual dipolar couplings from spectator qubits alter both the nonlocal entangling evolution and the local phases, so the correction angle calibrated for an isolated pair is no longer optimal in larger arrays. The paper demonstrates this by simulating the full spin-orbital dynamics, classifying the entangling landscape with Makhlin invariants, calibrating electrically driven Rz rotations over a full 2π phase range, and reconstructing complete pulse sequences. The resulting sqrt(iSWAP) and iSWAP gate infidelities, 4.72e-4 and 6.14e-4, fall below the 1e-3 threshold, and symmetric active-idle configurations in linear, s","pith_inferences":["The reduced 4^N subspace used throughout could be checked against the full 8^N basis at the optimized operating points; if leakage into parallel-spin states is non-negligible, the quoted infidelities would rise and possibly cross the threshold.","The strong sensitivity of the correction angle (deviations of 1e-4 rad degrade fidelity) suggests that closed-loop experimental calibration will be needed, not just open-loop simulations.","The co-design principle likely transfers to other dipolarly coupled spin-orbital qubit platforms, where spectator-induced phase shifts are similarly layout-dependent.","Dynamically tunable metallic couplers, mentioned as future work, would make the geometry dependence controllable and could turn asymmetric layouts from liabilities into assets."],"forward_implications":["Two-qubit gate calibration must be performed in the actual array environment, not inferred from isolated-pair simulations.","Symmetric qubit layouts are preferable; they reduce spectator-induced distortion and can outperform asymmetric layouts even when the latter assume ideal local corrections.","Electrically driven Rz rotations calibrated over a full 2π range are sufficient to implement the necessary local corrections at the required accuracy.","iSWAP gates are far more sensitive to geometry and system size than sqrt(iSWAP), making sqrt(iSWAP) the more practical target for early multi-qubit demonstrations.","The absence of any CNOT-equivalent region confirms that the XY-type dipole-dipole interaction cannot produce CNOT in a single entangling step, guiding gate set selection."],"fun_headline_variants":["Co-optimizing layout and pulses beats isolated qubit tuning","Spectator qubits force co-design for silicon gate fidelity","Flip-flop qubit gates hit 99.95% via co-design","Co-design tames spectator qubits for sub-1e-3 gates","Numerical co-design quashes spectator interference in qubit arrays"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The quoted fidelities assume a reduced four-dimensional-per-qubit subspace that discards parallel electron-nuclear spin configurations; if those discarded states participate in the driven dynamics at the optimized operating points, the infidelity numbers are optimistic.","fun_headline_variants_meta":{"raw":{"variants":["Co-optimizing layout and pulses beats isolated qubit tuning","Spectator qubits force co-design for silicon gate fidelity","Flip-flop qubit gates hit 99.95% via co-design","Co-design tames spectator qubits for sub-1e-3 gates","Numerical co-design quashes spectator interference in qubit arrays"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001122,"raw_usage":{"total_tokens":4557,"prompt_tokens":851,"completion_tokens":3706,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":595,"completion_tokens_details":{"reasoning_tokens":3615}},"tokens_in":595,"tokens_out":3706,"duration_ms":25624,"temperature":1.0,"reasoning_tokens":3615,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T13:16:52.565358+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the optimized sqrt(iSWAP) and iSWAP pulse sequences in the full 8^N basis, or experimentally measure population in parallel-spin states at the optimal operating points. If the resulting infidelity exceeds 1e-3, or if observed leakage is non-negligible, the quantitative support for the central co-design claim fails.","supporting_citations":[],"review_version":1}