{"id":"aa347502-78a2-43ad-86f2-9f6f1f2e43ba","arxiv_id":"2411.15726","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Demonstration of a deterministic two-phonon (N=2 N00N) entangled state between mechanical resonators on separate substrates, with fidelities 0.87 (Bell) and 0.75 (N00N).","lead":"Two microscopic resonators on separate chips were quantum-entangled with two shared vibrations, achieving measured fidelities of 0.87 for a single-phonon Bell state and 0.75 for a two-phonon N00N state. This advances mechanical systems as a modular platform for quantum information.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The N00N fidelity may be inflated by zero-padding the density matrix beyond two phonons per resonator; the authors should test reconstruction without this truncation.","rationale":"The reader's weakest assumption identifies exactly the N00N reconstruction truncation. This is the load-bearing concern because the paper's main novelty is multi-phonon (N=2 N00N) entanglement, and the reported fidelity is the quantitative evidence for that claim. If the truncation masks leakage into higher Fock states, the fidelity would overestimate the true overlap with the target state. The paper does provide independent support: the protocol is deterministic, parameters are independently measured, and simulations agree with the reported fidelity. These give confidence, but they do not directly establish that the physical state has no support beyond two phonons per resonator. The concrete test—running the simulation to check leakage and possibly redoing tomography without truncation—would settle whether this concern lands. The Bell-state fidelity inconsistency (0.872±0.002 in abstract vs 0.872±0.003 in text) is minor and does not affect the main argument. Therefore, the existing CONDITIONAL verdict remains appropriate, and no change to the reader's verdict is needed.","tokens_in":17,"tokens_out":4269,"duration_ms":140046,"concrete_test":"Run the master-equation simulation (QuTiP) with the independently measured parameters in Extended Data Table 1 for the full N00N generation protocol, and compute the joint probability that either resonator ends in a Fock state with n>2. If this probability is above ~1%, redo the tomographic reconstruction without zero-padding above n=2, allowing up to six levels per resonator as used in the time-trace fits, and report the resulting fidelity and the population in the n>2 subspace. If the leakage is negligible, the truncation is justified and the reported fidelity stands.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim of multi-phonon entanglement rests on the N=2 N00N state fidelity F=0.748±0.008. The supplementary states that for the N00N reconstruction they assume a maximum of two excitations in each resonator and zero-pad ρm for higher phonon indices. However, the time-trace fits themselves use up to six resonator levels per resonator (Extended Data Fig. 3). If the physical state has any population in Fock states |3⟩ or higher, the convex-optimization inversion will force that population to zero, potentially making the reconstructed state artificially close to the |20⟩+|02⟩ target. The Bell-state reconstruction is not subject to this particular truncation, so the concern is specific to the headline multi-phonon claim. The reported agreement with simulation (F=0.748 vs 0.745) suggests the truncation may be benign, but that simulation can directly test whether >2-phonon populations are in fact negligible. Without this check, the N00N fidelity is an upper bound under an unverified subspace assumption, not a direct measurement of the created state.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports deterministic generation of entanglement between two surface acoustic wave (SAW) resonators fabricated on separate lithium niobate substrates, each coupled to a superconducting transmon qubit. A two-qubit Bell state is generated and swapped into the resonators, yielding a mechanical Bell state (|10>+|01>)/√2 with reported fidelity F = 0.872±0.002/0.003. Extending the protocol to qubit f-states, the authors generate a two-phonon N00N state (|20>+|02>)/√2 and report fidelity F = 0.748±0.008 from joint Wigner tomography with convex optimization. The paper also presents device characterization, simulations based on independently measured parameters, and a discussion of scalability to multi-mode and multi-chip configurations.","tokens_in":13078,"tokens_out":4995,"duration_ms":44136,"significance":"If the reported fidelities hold, this work demonstrates multi-phonon (N=2) entanglement between mechanical resonators on separate substrates, going beyond previous single-phonon entanglement demonstrations and strengthening the case for modular SAW-based quantum information processing. The protocol is deterministic and produces entangled bosonic states that can be analyzed with direct qubit-resonator swaps. Notable strengths include the use of independently measured T1, T2, and coupling parameters in simulations, the consistency of experimental signatures (e.g., the √2 faster oscillations for the N00N state), and the transparent disclosure of the truncation assumption in the N00N tomography. However, the central N00N fidelity claim rests on a subspace assumption that is not yet verified; this must be addressed before the multi-phonon claim is fully supported.","major_comments":[{"comment":"The N00N state reconstruction imposes a maximum of two excitations per resonator and zero-pads higher phonon indices, as stated explicitly in the supplementary. Yet the time-trace fits used for the tomography (Extended Data Fig. 3) include up to six resonator levels per resonator, and the generation protocol involves two swap steps in which imperfect swaps and non-resonant phonon emission could populate higher Fock states. If the actual joint state has any population in |3> or above, the convex-optimization inversion will force it to zero, making the reported fidelity to (|20>+|02>)/√2 an upper bound rather than a measured value. This is load-bearing for the multi-phonon claim. Please either reconstruct ρm without the two-phonon truncation (e.g., allowing up to six phonons per resonator, matching the fit space) and report the fidelity and n>2 populations, or provide direct simulation evidence that n>2 populations are negligible. The existing simulation (F=0.745 vs 0.748 measured) can settle this, so the check appears straightforward.","section":"Supplementary, 'Joint state tomography of two mechanical resonators'"},{"comment":"The Bell-state fidelity is quoted as 0.872±0.002 in the abstract and in the Fig. 3e caption, but as 0.872±0.003 in the main text (paragraph after Fig. 3). The authors should correct this inconsistency and state the exact method for error propagation.","section":"Main text, Fig. 3 and abstract"}],"minor_comments":[{"comment":"The bootstrap procedure uses only 10 resamples. This is small for 225–261 pulse combinations; please report the distribution and consider more resamples or a separate uncertainty estimate.","section":"Supplementary, 'Joint state tomography of two mechanical resonators'"},{"comment":"The fidelity expression F = sqrt(Tr(ρ_target|ρ|)) is unusual; please clarify whether |ρ| denotes the absolute value of the density matrix and how the square root is applied. Standard definitions are F = ⟨ψ|ρ|ψ⟩ or the square root of that; please define explicitly.","section":"Main text, Fig. 3e and Fig. 4e"},{"comment":"The figure caption says 'For Fig. 3 in the main text, we only show three energy levels...' while the text says five levels are used in the fits; please clarify what is displayed versus what is used in the reconstruction.","section":"Supplementary, Extended Data Fig. 3"}],"recommendation":"major_revision","confidential_remarks":"The experimental platform is impressive and the reported agreement with independent simulation is a strong point. The main concern is the N00N tomography truncation, which is disclosed but not verified. I would support publication if the authors provide the requested check on higher-phonon populations, either by reconstructing without truncation or by showing from simulation that n>2 populations are negligible. The fidelity error inconsistency in the Bell state is a minor but easily fixed issue."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The headline result is genuine: the authors generate and tomographically verify a mechanical Bell state and, for the first time, a two-phonon N00N state shared between two SAW resonators on separate substrates. Fidelities (0.872 and 0.748) are well above classical thresholds, and the N00N value agrees with an independent simulation (0.745). The modular flip-chip architecture, with two LN dies bonded to a sapphire qubit chip, is new and plausibly scalable.\n\nWhat the paper does well: the device characterization is careful. Resonator lifetimes, dephasing, and qubit-resonator couplings are all measured independently, and the master-equation simulations use those parameters rather than fitting the target state. The tomography itself is also reasonably thorough, using displacement pulses with multiple amplitudes and phases, and the supplementary discloses the key modeling assumption. The Bell-state result is not affected by the main concern below, and its fidelity is credible.\n\nThe soft spot is real but not fatal. For the N00N reconstruction, the density matrix is constrained to at most two excitations per resonator, and higher phonon numbers are zero-padded. Meanwhile, the time-trace fits use up to six resonator levels. If the physical state has any |3> or higher population, the convex-optimization inversion will force it to zero, potentially inflating the reported N00N fidelity. The agreement with simulation suggests the effect is modest, but the simulation can directly check whether >2-phonon populations are negligible; the authors should run that check and, if possible, reconstruct without the truncation. Minor issues: the Bell fidelity error is quoted as ±0.002 in the abstract and ±0.003 in the main text, and the bootstrap uses only 10 resamples, making the reported uncertainties rough.\n\nThis paper is for the quantum acoustics and bosonic quantum computing communities. It deserves a serious referee: the experimental claim is new, the execution looks competent, and the limitations are discussed honestly. My recommendation is to send it to peer review, asking the authors to address the N00N truncation explicitly, fix the error-bar inconsistency, and ideally release the tomography data and analysis code.","headline":"Real step forward: first two-phonon N00N state between separate-substrate mechanical resonators, with a disclosed but untested tomography truncation that should be checked in revision.","tokens_in":13639,"tokens_out":3057,"would_cite":true,"duration_ms":28245,"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":"Two mechanical resonators on separate chips are deterministically entangled, with measured Bell and two-phonon N00N state fidelities of 0.872 and 0.748.","keywords":["mechanical entanglement","surface acoustic wave resonator","superconducting qubit","N00N state","multi-phonon entanglement","Wigner tomography","modular quantum architecture","deterministic entanglement"],"falsifier":"Re-run the N00N generation protocol but reconstruct the joint resonator density matrix allowing up to three or more phonons per resonator in the convex optimization; alternatively, directly measure the population in the $|30\\rangle$ and $|03\\rangle$ states by extending the swap-time traces to longer interactions. If these populations are non-negligible, the fidelity to $(|20\\rangle+|02\\rangle)/\\sqrt{2}$ is inflated relative to the true state.","tokens_in":12629,"feed_emoji":"⚛️","tokens_out":9949,"duration_ms":75300,"temperature":0.7,"pith_summary":"This paper reports deterministic generation of mechanical entanglement between two surface acoustic wave resonators that are not on the same chip. The two resonators sit on separate lithium niobate dies, each coupled to its own superconducting qubit on a third sapphire substrate. The authors prepare a two-qubit Bell state and then swap it into the resonators, obtaining a mechanical Bell state with fidelity $\\mathcal{F}=0.872\\pm0.002$. They then use qutrit states and a two-step swap to create a two-phonon N00N state, $(|20\\rangle+|02\\rangle)/\\sqrt{2}$, with fidelity $\\mathcal{F}=0.748\\pm0.008$, certified by joint Wigner tomography. A sympathetic reader would care because this shows multi-phonon entanglement can be distributed across physically separated mechanical modules, a step toward scalable bosonic quantum computing with mechanics.","feed_headline":"Resonators on separate chips share Bell and two-phonon N00N states","feed_subtitle":"Qubit-mediated swaps tie two surface-acoustic-wave resonators together with measured fidelities of 0.872 and 0.748.","key_machinery":"The central mechanism is the qubit-resonator coherent swap, in which each tunable transmon is tuned into resonance with its surface acoustic wave mode and its coupling strength is set by a variable coupler, so that an excitation deterministically moves between the qubit and the resonator. A two-qubit Bell state is generated by a controlled swap and then transferred in parallel to the two resonators. For the N00N state, each qubit is first raised to its $|f\\rangle$ state and the two qubits are entangled into $(|fg\\rangle+|gf\\rangle)/\\sqrt{2}$; the $f\\leftrightarrow e$ swap transfers one phonon per side, and the subsequent $e\\leftrightarrow g$ swap transfers a second phonon, ideally leaving $(|20\\rangle+|02\\rangle)/\\sqrt{2}$. Tomography is performed by applying coherent displacement pulses $\\hat D_j(-\\alpha_j)$ to each resonator and measuring the qubits after a fixed interaction time, then reconstructing the two-mode density matrix via convex optimization with bootstrap error bars.","core_discovery":"The central claim is that a modular platform, in which two surface acoustic wave resonators are fabricated on separate substrates and inductively coupled to two superconducting qubits on a third chip, can deterministically produce and fully characterize entangled phonon states spanning both mechanical modes. By first entangling the two qubits and then performing simultaneous resonant swaps, the system yields the mechanical Bell state $(|10\\rangle+|01\\rangle)/\\sqrt{2}$ with fidelity $0.872\\pm0.002$. By promoting the qubits to their second excited state and using sequential $f\\leftrightarrow e$ and $e\\leftrightarrow g$ swaps, the system yields the two-phonon N00N state $(|20\\rangle+|02\\rangle)/\\sqrt{2}$ with fidelity $0.748\\pm0.008$. The states are reconstructed from joint Wigner tomography using coherent displacement pulses and convex optimization. If correct, this demonstrates that the bosonic nature of phonons can be exploited across distant mechanical resonators, going beyond single-phonon entanglement.","pith_inferences":["If the two-phonon subspace truncation in the tomography were relaxed, the N00N state fidelity could fall; an independent measurement of higher Fock state populations would tell whether the reported 0.748 is robust.","The modular flip-chip design could be adapted to link more than two mechanical nodes, and the same qubit-mediated swap technique could entangle mechanical resonators at different frequencies without needing direct mechanical coupling.","Narrowing the IDT bandwidth to keep the $e\\leftrightarrow g$ transition outside the transducer emission window should suppress the unwanted phonon emission during the $f\\leftrightarrow e$ swap, potentially raising N00N fidelity and enabling efficient generation of higher-$N$ states."],"forward_implications":["The same swap ladder can be iterated to produce N00N states with $N>2$, limited by decoherence and unwanted non-resonant phonon emission during the $f\\leftrightarrow e$ swap.","Because the resonators are on separate dies with no direct mechanical contact, the architecture supports distributing entanglement across larger multi-chip networks, e.g., GHZ and W states.","Joint Wigner tomography with convex optimization provides a general readout for multi-mode phonon states, applicable to other circuit quantum acoustics platforms.","The measured fidelities are limited by short resonator lifetimes (roughly 0.27–0.38 µs) and qubit $T_1$ during swaps; improving these lifetimes would directly increase the achievable state fidelities."],"supporting_citations":[{"why":"The prior demonstration of entangled mechanical resonators that this work extends to multi-phonon states on separate substrates.","marker":"[8]"},{"why":"The protocol for deterministically entangling two microwave resonators via qubits, which the Bell-state generation follows.","marker":"[51]"},{"why":"The qubit-resonator swap and variable-coupler control that underlies the phonon state transfer.","marker":"[3]"},{"why":"The non-galvanic flip-chip integration that allows the two mechanical resonators to sit on separate substrates.","marker":"[47]"},{"why":"The bootstrap resampling method used to assign uncertainties to the reconstructed density matrices.","marker":"[57]"}],"fun_headline_variants":["Phonons entangle across two separate chips","Bell and N00N states connect separate resonators","Multi-phonon entanglement on separate substrates","Deterministic phonon links: Bell and N00N states","Two resonators share multi-phonon entanglement"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The fidelity of the N00N state is computed from a density matrix that was reconstructed under the assumption that neither resonator ever holds more than two phonons, so any leaked population in higher Fock states would be invisible to the tomography and the reported fidelity could be overstated.","fun_headline_variants_meta":{"raw":{"variants":["Phonons entangle across two separate chips","Bell and N00N states connect separate resonators","Multi-phonon entanglement on separate substrates","Deterministic phonon links: Bell and N00N states","Two resonators share multi-phonon entanglement"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000481,"raw_usage":{"total_tokens":2410,"prompt_tokens":1008,"completion_tokens":1402,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":624,"completion_tokens_details":{"reasoning_tokens":1327}},"tokens_in":624,"tokens_out":1402,"duration_ms":9752,"temperature":1.0,"reasoning_tokens":1327,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T13:58:33.150593+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-run the N00N generation protocol but reconstruct the joint resonator density matrix allowing up to three or more phonons per resonator in the convex optimization; alternatively, directly measure the population in the $|30\\rangle$ and $|03\\rangle$ states by extending the swap-time traces to longer interactions. If these populations are non-negligible, the fidelity to $(|20\\rangle+|02\\rangle)/\\sqrt{2}$ is inflated relative to the true state.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The prior demonstration of entangled mechanical resonators that this work extends to multi-phonon states on separate substrates."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The protocol for deterministically entangling two microwave resonators via qubits, which the Bell-state generation follows."},{"cited_title":"( ) ( ) QB QA RB RA b Figure 3: Deterministic mechanical Bell state generation and tomography","cited_arxiv_id":null,"evidence_quote":"The qubit-resonator swap and variable-coupler control that underlies the phonon state transfer."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The non-galvanic flip-chip integration that allows the two mechanical resonators to sit on separate substrates."}],"review_version":1}