{"id":"5a2cdbf5-12e1-4d05-8cb8-cd8406c15a2f","arxiv_id":"2507.16783","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A silicon nanophotonic CNOT gate was teleported between remote photonic qubits with an average gate fidelity of 86.5±2.2%.","lead":"Researchers demonstrated teleportation of a two-qubit CNOT gate on a silicon photonic chip, meaning a quantum logic operation was transferred between remote qubits using shared entanglement instead of direct interaction. This is the first reported chip-scale gate teleportation and a step toward modular, distributed quantum computers built from silicon photonics.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Gate-teleportation identity requires a specific control-target pairing for the two local CNOTs; the paper never states it, and the chip is only referenced to a prior SWAP-gate paper.","rationale":"The reader's weakest assumption is the correct one and is the most load-bearing. I re-derived the protocol identity under all four control-target conventions. Only one pairing (C12: control polarization/target path; C34: control path/target polarization) yields a teleported CNOT between qubits 1 and 4 with Pauli corrections σ1Z, σ4X, and their product. Any reversal destroys the CNOT character, so the headline claim is true only if the physical gates realize that exact orientation. The paper does not state it: the chip is cited to a SWAP-gate paper, and a 'balanced MZI' is not a CNOT without specifying which arm contains the HWP and how the path basis is labeled. The reported teleported truth table and QPT are consistent with the claim, so this is not an internal inconsistency; it is an unverified precondition. Other issues (truth-table fidelity called gate fidelity, post-selection without active feed-forward, supplementary missing) are secondary and do not by themselves falsify the claim. A path-resolved local truth table or a statement of the control-target convention would settle it. I therefore keep the CONDITIONAL verdict.","tokens_in":15687,"tokens_out":21002,"duration_ms":224408,"concrete_test":"Request the path-resolved truth tables for the two local gates (Fig. 2A/B) and the full teleportation setup: verify that for the chip, the output path flips when input polarization is |V> while |H> stays straight (C12 = control polarization, target path), and for the free-space MZI, the output polarization flips only when the path control is |1> (C34 = control path, target polarization). Independently re-derive the Results identity with these assignments; if either target is the opposite degree of freedom, the measured process in Fig. 4B cannot be C14.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The teleportation identity C34C12(|Ψ>14⊗|Φ>23)=... is not an identity for arbitrary CNOT gates: it holds only for the specific pairing control-1/target-2 for C12 and control-3/target-4 for C34. With C12 reversed (path controls polarization), the conditional output becomes identity for one measurement outcome and X1X4 for the other; with C34 reversed, the output acquires Z1^s Z4^s factors instead of a CNOT. The paper never states which degree of freedom is the control on either local gate. The chip gate is referred to Ref. [60], whose title is a polarization-spatial-momentum SWAP gate, not a CNOT, and the free-space gate is described only as a 'balanced Mach-Zehnder interferometer' with HWP5 at 45°, which is not by itself a CNOT unless the HWP sits in one arm and the path basis is defined accordingly. The local truth-table measurements in Fig. 2A/B are described via output polarization projections; without path-resolved data they do not fix the target. Since all four qubit labels and the claimed C14 process depend on this orientation, the central claim rests on an unstated experimental detail. The missing Methods and source data in the arXiv version make this impossible to check from the preprint alone.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports the experimental implementation of quantum gate teleportation of a CNOT gate realized on a silicon photonic chip. Two remote polarization qubits (1 and 4) are entangled via path qubits (2 and 3) prepared in a shared EPR state; local CNOT gates C12 (on-chip) and C34 (free-space) are applied, and measurement of the path qubits in the X basis teleports the local CNOT to a non-local CNOT acting on the polarization qubits 1 and 4. The authors characterize the teleported gate through a truth-table fidelity of 93.1 ± 0.3%, an average quantum state fidelity of 87.0 ± 2.2%, a Bell-state fidelity of 86.2 ± 0.8%, and a process fidelity of 83.1 ± 2.0%, corresponding to an average gate fidelity of 86.5 ± 2.2%. They claim that this is the first quantum teleportation of a chip-scale CNOT gate.","tokens_in":15916,"tokens_out":7784,"duration_ms":75979,"significance":"If the control-target orientation and the post-selection/feed-forward details are as the protocol requires, the result is a genuine milestone: it combines a CMOS-compatible silicon chip, high-fidelity local two-qubit gates, and teleportation-based non-local gate operation on photonic qubits. The measured fidelities are mutually consistent (local gates at 97.9% and 98.1%, entanglement fringe visibility at 93.9% before background subtraction, and teleported gate fidelity at 86.5%), and the central teleportation identity is correct when the CNOT orientations are specified as control-1/target-2 and control-3/target-4. No fitted parameters appear in the reported fidelities, which are measured quantities. The work is of clear interest to modular photonic quantum computing and quantum networking.","major_comments":[{"comment":"The gate-teleportation identity C34C12(|Ψ>14⊗|Φ>23) = ... is valid only for a specific control-target pairing: C12 must be the CNOT with control on qubit 1 (polarization) and target on qubit 2 (path), and C34 must be the CNOT with control on qubit 3 (path) and target on qubit 4 (polarization). The manuscript never states these assignments, and the descriptions of the on-chip gate (referenced to Ref. [60], a SWAP-gate paper) and the free-space gate (a 'balanced Mach-Zehnder interferometer' with HWP5 at 45°) do not by themselves fix the orientation. Because the entire claim that the teleported operation is C14 depends on this orientation, please state the convention explicitly (e.g., CNOT_{control→target}) and provide path-resolved truth-table data or an equivalent calibration that establishes the orientation of each local gate.","section":"Results, 'Silicon chip CNOT gate teleportation realization' (gate-teleportation identity)"},{"comment":"The protocol has four measurement outcomes on qubits 2 and 3, each requiring a distinct Pauli correction (I, Z1, X4, and Z1X4). The paper does not state whether active feed-forward was applied, whether the data were post-selected to a subset of outcomes, or whether the Pauli corrections were applied in post-processing. Since all quoted fidelities are based on coincidence counts without subtracting accidentals, the outcome-selection rule is essential for interpreting the 93.1% truth-table fidelity and the 83.1% process fidelity. Please specify which outcomes were accepted and how the corrections were implemented.","section":"Results, 'Teleported chip-scale CNOT gate truth able...' and Fig. 1C/2C"},{"comment":"Ref. [60] is titled 'A chip-scale polarization-spatial-momentum quantum SWAP gate in silicon nanophotonics,' yet the text says that the design optimized map of the on-chip CNOT gate is given in that prior study. A SWAP operation is not a CNOT, and the paper does not explain how the same device (or the same chipset) is configured as a CNOT with the orientation required by the teleportation identity. Please clarify the relationship between the SWAP gate in Ref. [60] and the CNOT gate used here, including the operating point or device parameters that distinguish the two operations.","section":"Results, 'Silicon chip CNOT gate teleportation realization' (device description)"},{"comment":"The main text repeatedly refers to 'Methods in Supplementary Materials' and to Figures S.1–S.7 for source characterization, quantum state tomography, and quantum process tomography details, but the arXiv version contains none of these. In particular, the path-resolved local gate measurements and the exact set of tomographic settings are not in the main text. Without this material the experimental claims cannot be reproduced or fully checked. Please include the supplementary information with the revised submission and ensure that the source-data statement is consistent with what is actually provided.","section":"Supplementary Materials and Data Availability"}],"minor_comments":[{"comment":"The heading contains a typo: 'truth able' should be 'truth table'.","section":"Section heading, 'Teleported chip-scale CNOT gate truth able...'"},{"comment":"Several mathematical expressions are missing fraction bars or contain garbled notation, such as '(00|23 + |11>23) √2⁄', which should be written as (|00>23 + |11>23)/√2. Please ensure all equations are typeset correctly.","section":"Throughout the equations"},{"comment":"The text says 14 input polarization states are used in the second analysis and that the remaining two states (|+0> and |+1>) appear in the third analysis; this is confusing because the full set for two-qubit state tomography is 16 states. Please state explicitly that 16 input states are covered across the two sections.","section":"Results, quantum state tomography input states"},{"comment":"The term 'average gate fidelity' is used both for the truth-table fidelity (93.1%) and for the process-derived average gate fidelity (86.5%). These are different quantities with different definitions; please use 'truth-table fidelity' for the former and reserve 'average gate fidelity' for the process-derived quantity.","section":"Results, fidelity terminology"}],"recommendation":"major_revision","confidential_remarks":"The load-bearing concerns are fixable in revision: the missing control-target orientation, the outcome-selection rule, and the relation to the prior SWAP-gate device. The current arXiv version is not self-contained because the Methods and Supplementary Materials are absent, which makes the experimental details impossible to verify. The incremental novelty over bulk-optical gate teleportation is modest but real for chip-scale integration, and the measured fidelities are credible if the implied orientation is correct. The authors should also be asked to clarify the connection to Ref. [60] to avoid any impression that a SWAP device is being relabeled as a CNOT without explanation."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First on-chip teleportation of a CNOT gate is a real milestone, and the experiment looks sound; but the paper never specifies the control-target orientation of the two local CNOTs, and the chip design is credited to a prior SWAP-gate paper, so the central identity can't be checked from the preprint.\n\nThe combination is new: gate teleportation has been done in free space and with trapped ions and superconducting qubits, and state teleportation has been on chips, but a teleported CNOT on a silicon chip is a first. The local gates are well characterized (97.9% and 98.1% truth table fidelities), and the teleported gate is cross-checked four ways: truth table (93.1%), state tomography (87.0%), Bell-state generation (86.2%), and process tomography (83.1% process, 86.5% average gate). These numbers are mutually consistent, and the experiment is done without background subtraction, which is honest.\n\nThe soft spots are real but fixable. The teleportation identity is only valid if C12 controls on qubit 1 with target qubit 2, and C34 controls on qubit 3 with target qubit 4. The paper never says which degree of freedom is the control. The chip is described as the device from Ref. [60], whose title says 'SWAP gate,' not CNOT. The free-space C34 is just a 'balanced Mach-Zehnder interferometer' with HWP5 at 45°. A reader cannot verify the orientation from the text. The measured teleported truth table matching a CNOT suggests the orientation is in fact correct, but the paper needs to state it explicitly and ideally show the path-resolved local truth tables. The 'truth table fidelity' is also loosely called 'gate fidelity' in a few places; the process tomography number is the better one. Finally, the arXiv version lacks the supplementary methods and source data, which are needed to reproduce this.\n\nThis is a paper for a serious experimental journal, and it should get full peer review. The referee should ask for the orientation clarification, the prior device's characterization as a CNOT, and the supplementary data. If those are provided, I'd be comfortable with the claim.","headline":"First on-chip teleportation of a CNOT gate is a plausible experimental milestone, but the paper's unstated local-gate orientations and loose link to a prior SWAP-gate device need referee scrutiny before the claim is fully reproducible.","tokens_in":16519,"tokens_out":5049,"would_cite":true,"duration_ms":50173,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81P68","81V80"],"pacs":["03.67.Lx","42.50.Ex"],"model":"deepseek-v4-flash","headline":"The paper reports the first experimental teleportation of a controlled-NOT gate on a silicon nanophotonic chip, using a shared EPR pair, two local CNOT operations, and classical communication, with an average gate fidelity of 86.5 ± 2.2%.","keywords":["quantum teleportation","CNOT gate","silicon nanophotonics","distributed quantum computing","photonic qubits","quantum gate teleportation","Bell states","quantum process tomography"],"falsifier":"Reconstruct the teleported process matrix with the free-space CNOT $C_{34}$ configured so that its control and target are swapped relative to the assumed orientation; if the teleported truth table no longer matches the ideal CNOT's 2×2 block structure (fidelity dropping well below 93%), the reported non-local CNOT depends critically on that unstated orientation.","tokens_in":15477,"feed_emoji":"⚛️","tokens_out":4552,"duration_ms":44889,"temperature":0.7,"pith_summary":"The paper reports the first experimental teleportation of a controlled-NOT gate on a photonic chip. Using a shared polarization-entangled pair, two local CNOT operations—one on a silicon nanophotonic chip, the other in free space—and classical communication, the authors turn a local chip CNOT into a non-local CNOT acting on two remote polarization qubits. They characterize the teleported gate by truth-table, state-tomography, Bell-state, and process-tomography measurements, reporting an average gate fidelity of 86.5 ± 2.2% and a process fidelity of 83.1 ± 2.0%. The result matters because it shows that chip-scale silicon photonics can supply the non-local two-qubit gates that modular distributed quantum computers need.","feed_headline":"Silicon chip teleports a CNOT gate to remote qubits","feed_subtitle":"Two remote photonic qubits get a CNOT gate via shared entanglement, at 86.5% fidelity on a chip.","key_machinery":"The protocol rests on the gate-teleportation identity $C_{34}C_{12}(|\\psi\\rangle_{14}\\otimes|\\Phi\\rangle_{23}) = |0+\\rangle_{23}\\otimes C_{14}|\\psi\\rangle_{14} + |0-\\rangle_{23}\\otimes \\sigma_1^Z C_{14}|\\psi\\rangle_{14} + |1+\\rangle_{23}\\otimes \\sigma_4^X C_{14}|\\psi\\rangle_{14} + |1-\\rangle_{23}\\otimes (-\\sigma_1^Z\\sigma_4^X) C_{14}|\\psi\\rangle_{14}$, where $C_{12}$ is the on-chip CNOT (control qubit 1, target qubit 2), $C_{34}$ is the free-space CNOT (control qubit 3, target qubit 4), and $|\\Phi\\rangle_{23}$ is the shared EPR pair. Measuring qubits 2 and 3 in the $|\\pm\\rangle$ basis and applying the corresponding Pauli corrections teleports the local gate $C_{12}$ to a non-local CNOT $C_{14}$ between qubit 1 and qubit 4. The identity carries the argument because it shows the teleported operation is exactly a CNOT, not just an entangling operation.","core_discovery":"The central claim is that quantum gate teleportation of a CNOT gate can be executed on an integrated silicon platform: a local on-chip CNOT acting on a polarization qubit and a path qubit is converted, through a shared EPR pair and a second local CNOT on Bob's side, into a CNOT acting on the two remote polarization qubits. The paper argues this is the first chip-scale demonstration, and supports it with four independent characterizations: a truth-table fidelity of 93.1 ± 0.3%, an average state fidelity of 87.0 ± 2.2% over fourteen input states, an average Bell-state fidelity of 86.2 ± 0.8% for the four generated Bell states, and a full quantum process tomography with process fidelity 83.1 ± 2.0%, corresponding to an average gate fidelity of 86.5 ± 2.2%.","pith_inferences":["The same gate-teleportation identity, with $C_{12}$ and $C_{34}$ replaced by controlled-Z gates, would teleport a CZ gate with the same classical communication cost, suggesting the result generalizes beyond CNOT.","The gap between truth-table fidelity (93.1%) and process fidelity (83.1%) suggests that much of the error comes from state preparation and measurement rather than the teleported logical map, so improving chip polarization-extinction could raise both quantities.","The protocol's reliance on post-selected entanglement means the teleported gate succeeds only in coincidence; substituting a deterministic entanglement source would be a natural next step that the authors explicitly mention."],"forward_implications":["A chip-scale teleported CNOT gate provides a building block for modular distributed photonic quantum computers, where remote modules are linked by entanglement rather than direct interactions.","The demonstrated fidelities, though below fault-tolerance thresholds, are compatible with entanglement distillation or purification, which the authors note as a route to higher-quality non-local gates.","The 50×50 µm² chip footprint suggests that many such gates can be integrated, extending the scheme to multiple qubits and multi-chip networks.","Because the gate teleportation consumes fewer classical bits than state-teleportation-based gate implementations, it lowers the communication overhead for a distributed CNOT.","The approach can be adapted from post-selected to heralded or deterministic entanglement sources, making the teleported gate compatible with future deterministic photonic quantum computing."],"supporting_citations":[{"why":"Supplies the silicon nanophotonic CNOT gate device, its design, and its previously characterized high-fidelity operation that this work teleports onto remote qubits.","marker":"[60]"},{"why":"Provides the prior experimental demonstration of teleporting a quantum CNOT gate in free space, establishing the protocol that this paper extends to a chip-scale implementation.","marker":"[22]"},{"why":"Shows a teleportation-based realization of an optical two-qubit entangling gate, a key reference for the non-local gate mechanism used here.","marker":"[21]"},{"why":"Demonstrates quantum state teleportation on a photonic chip, highlighting the gap (no on-chip gate teleportation) that this paper fills.","marker":"[58]"},{"why":"Demonstrates chip-to-chip quantum state teleportation in silicon, providing the integrated-photonics context and baseline for the present gate-teleportation advance.","marker":"[59]"},{"why":"Gives the relation between process fidelity and average gate fidelity used to convert the measured 83.1% process fidelity into the reported 86.5% average gate fidelity.","marker":"[34]"}],"fun_headline_variants":["First chip-scale teleported CNOT gate","Silicon chip teleports CNOT gate to remote qubits","On-chip CNOT gate teleported with 86.5% fidelity","CNOT gate teleported on silicon","Teleported CNOT gate on a silicon chip"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The teleported operation is a CNOT only if the two local CNOT gates have a specific control-target orientation (qubit 1 controls qubit 2, and qubit 3 controls qubit 4), an orientation the paper does not explicitly state.","fun_headline_variants_meta":{"raw":{"variants":["First chip-scale teleported CNOT gate","Silicon chip teleports CNOT gate to remote qubits","On-chip CNOT gate teleported with 86.5% fidelity","CNOT gate teleported on silicon","Teleported CNOT gate on a silicon chip"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000674,"raw_usage":{"total_tokens":3133,"prompt_tokens":1074,"completion_tokens":2059,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":690,"completion_tokens_details":{"reasoning_tokens":1983}},"tokens_in":690,"tokens_out":2059,"duration_ms":17533,"temperature":1.0,"reasoning_tokens":1983,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T15:02:47.963704+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Reconstruct the teleported process matrix with the free-space CNOT $C_{34}$ configured so that its control and target are swapped relative to the assumed orientation; if the teleported truth table no longer matches the ideal CNOT's 2×2 block structure (fidelity dropping well below 93%), the reported non-local CNOT depends critically on that unstated orientation.","supporting_citations":[{"cited_title":"Cheng, K.- C","cited_arxiv_id":null,"evidence_quote":"Supplies the silicon nanophotonic CNOT gate device, its design, and its previously characterized high-fidelity operation that this work teleports onto remote qubits."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the prior experimental demonstration of teleporting a quantum CNOT gate in free space, establishing the protocol that this paper extends to a chip-scale implementation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows a teleportation-based realization of an optical two-qubit entangling gate, a key reference for the non-local gate mechanism used here."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrates quantum state teleportation on a photonic chip, highlighting the gap (no on-chip gate teleportation) that this paper fills."},{"cited_title":"Llewellyn, Y","cited_arxiv_id":null,"evidence_quote":"Demonstrates chip-to-chip quantum state teleportation in silicon, providing the integrated-photonics context and baseline for the present gate-teleportation advance."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the relation between process fidelity and average gate fidelity used to convert the measured 83.1% process fidelity into the reported 86.5% average gate fidelity."}],"review_version":1}