{"id":"4bbb9629-5ab8-46cb-97db-6bb5ec253533","arxiv_id":"2506.16361","paper_version":1,"verdict":"REJECT","confidence":"LOW","novelty_score":4.0,"correctness_risk":"high","formal_verification":"none","parameter_count":3,"one_line_summary":"A proposed JPA gain-profile co-design aims to remove Purcell filters from superconducting readout, together with a low-power CMOS SoC receiver, both supported only by simulation.","lead":"The paper reports a simulated co-design of a four-qubit superconducting chip with a modified Josephson parametric amplifier whose gain peaks and valleys are used both to amplify readout signals and to suppress qubit leakage, replacing Purcell filters. It also presents a 45 nm CMOS analog receiver, simulated after layout, that amplifies and down-converts the microwave signals with 72 dB gain and a noise figure below 1 dB.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Even if Fig. 2b is stable, the paper never shows that gain troughs suppress Purcell decay; 0 dB gain is not a filter. A Gamma_P calculation for the chip+feedline+JPA chain is missing.","rationale":"The reader's weakest assumption is real, but it is about whether Fig. 2b survives contact with the chip; my concern is that Fig. 2b, taken at face value, does not establish Purcell suppression. In a standard circuit-QED readout, the qubit relaxes through the resonator-feedline channel at a rate controlled by kappa_ext and detuning; a downstream amplifier's gain magnitude at the qubit frequency does not set that rate. The paper explicitly computes no Gamma_P, no Q_eff for qubits, and no comparison with a no-JPA baseline. If the JPA is a two-port with 0 dB transmission at qubit frequencies, the leakage channel is open. If it is a reflective one-port, 0 dB reflection gain alone does not imply an open-circuit boundary at the qubit; the complex impedance matters. So the central claim has a missing physical link, not merely a missing measurement. I therefore keep the reader's REJECT verdict. Independent support in the Qiskit Metal EPR table and CMOS post-layout simulations is real but does not touch this link. The only reason for LOW confidence is that the manuscript text is partly garbled and no primary data are available to check; that does not change the verdict.","tokens_in":16007,"tokens_out":8310,"duration_ms":98909,"concrete_test":"Simulate the full chain of Fig. 3 with a 50-ohm load at the feedline output and compute each qubit's Purcell rate Gamma_P^0 (standard input-output result Gamma_P ~ kappa_ext (g/Delta)^2). Then replace the 50-ohm load by the JPA described by the Fig. 2a S-parameters and recompute Gamma_P at each qubit frequency. If the JPA-loaded rates are not substantially lower than the no-JPA rates (equivalently, if Q_eff at qubit frequencies is not substantially increased), the central claim fails. This is a simulation-only check using components and formulas already present in the paper.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The abstract's central claim rests on two premises: (i) the comb-like gain profile of Fig. 2b is real and stable under chip loading, and (ii) placing qubit frequencies at gain minima suppresses Purcell decay. The reader's weakest assumption targets (i); I see the more load-bearing gap in (ii), because (ii) is required even if (i) is granted. In this circuit each qubit is coupled to a readout resonator that is coupled to a feedline; the Purcell rate is set by the qubit-resonator coupling g, the detuning Delta, and the resonator's external coupling to the feedline, Gamma_P ~ kappa_ext (g/Delta)^2. The JPA is downstream of that feedline. A gain minimum of 0 dB at the qubit frequency means unity transmission through the JPA, not isolation of the resonator from the feedline. If the quoted gain is instead a reflection gain, the paper still gives no impedance or S-parameter data at qubit frequencies from which isolation could be inferred. The theory section defines kappa_eff = omega_eff/Q_eff, but no kappa_eff or Q_eff value is computed for the qubit modes, and no Purcell decay rate with and without the JPA is reported. In the absence of such a calculation, the central design concept 'gain trough = Purcell filter' is unsupported even with a perfectly stable Fig. 2b. The loading/instability concern from the reader then compounds the problem, but it is not the primary failure.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports a co-designed quantum readout architecture in which a modified Blochnium Josephson parametric amplifier (JPA) with a simulated comb-like gain profile (peaks ~24 dB, troughs ~0 dB) is used both to amplify readout resonator signals and to suppress qubit Purcell decay by placing qubit frequencies at gain minima, thereby eliminating the need for Purcell filters. A four-transmon chip is modeled in Qiskit Metal; a quantum master-equation analysis of two coupled qubits is used to claim entanglement signatures in current probabilities; and a 45 nm CMOS SoC receiver (LNA, VCO, mixers, IF chain) is designed with post-layout simulations showing ~72 dB gain, <1 dB noise figure, ~70 mW power, and 4.1-8.2 GHz bandwidth. The paper concludes that the architecture is scalable to 20-160 qubits.","tokens_in":16322,"tokens_out":6678,"duration_ms":77291,"significance":"If the central Purcell-suppression claim were correct, combining amplification and filtering in a single device would be a useful simplification of superconducting readout chains. The paper also contains concrete engineering content: a detailed JPA topology, an EPR table from Qiskit Metal, and post-layout SoC simulation results with explicit performance numbers. However, the key physical claim is unsupported and appears inconsistent with standard Purcell physics, the entanglement analysis lacks a quantitative witness, and the JPA gain profile is simulation-only with no loading or sensitivity analysis. These issues affect the abstract's central claims rather than merely the presentation.","major_comments":[{"comment":"The central claim that placing qubit frequencies at gain troughs suppresses Purcell decay is not established and, as stated, is physically inconsistent. A gain minimum of 0 dB means unity transmission through the JPA, not attenuation or isolation at the qubit frequency. The Purcell rate is set by the qubit-resonator coupling g, the detuning Delta, and the resonator's external coupling to the feedline (Gamma_P ~ kappa_ext (g/Delta)^2); the JPA is downstream of that feedline and cannot reduce this rate unless it presents a high-reflection or absorbing termination at the qubit frequency. The manuscript defines kappa_eff = omega_eff/Q_eff but never computes kappa_eff or Q_eff for the qubit modes, nor reports Gamma_P with and without the JPA. This is the load-bearing claim of the abstract and must be supported by an explicit Purcell-rate calculation or by S-parameter/impedance data showing isolation at f_Q.","section":"JPA design and Engineering / Fig. 2b"},{"comment":"The tailored gain profile in Fig. 2b is a simulation result only, and the entire frequency plan is derived from it; no measured S-parameters, no fabrication details, and no loading study are presented. The text states that qubit and resonator frequencies are selected based on the JPA gain spectrum and aligned with gain minima, but if the actual device gain profile shifts under the loading of the four-qubit chip, the feedline, and the following amplifier chain, the Purcell-protection property disappears. At minimum, the authors must provide a sensitivity analysis or a measured gain profile under representative operating conditions before the architecture's central benefit can be claimed.","section":"JPA design and Engineering / Fig. 2"},{"comment":"The entanglement prediction is not demonstrated. Equation (5) is a linear system of equations whose coefficients contain n1, n2, and nb, which are state-dependent expectation values; no self-consistent solution is given, and the coupling g in Eq. (3) is introduced as a free parameter without a circuit-level derivation. The claim that a new frequency component fc in the FFT of Fig. 4d indicates the onset of entangled states is not backed by any entanglement witness, concurrence, or density-matrix measure, and a new frequency component can also arise from nonlinear mixing in a coupled oscillator system. Using Qiskit Metal's confirmation of a mode fc to validate the Hamiltonian term that produces fc is circular; a quantitative entanglement measure or a two-qubit state tomography calculation is needed.","section":"Equation (5) and Fig. 4"},{"comment":"The JPA analysis is not self-contained: the Lagrangian in Eq. (1) and the C and L matrices in Eq. (2) are given, but the total Hamiltonian, the quantum Langevin equations, and the derivation of the gain profile are not presented; the text refers the reader to the authors' prior work [22] for the detailed treatment. Because the central design depends on the simulated gain profile of Fig. 2b, the absence of a reproducible derivation makes it impossible to assess whether the new C and L matrices actually produce the claimed comb-like gain. At least the eigenvalue problem, the resulting mode frequencies, and the simulation parameter values should be reported.","section":"Theoretical background / Eqs. (1)-(2)"},{"comment":"The scalability numbers are contradictory: the text first states that with a conservative 10 MHz per resonator and 15 MHz spacing the receiver supports 160 qubits at 0.44 mW per qubit, then states that according to the present design, where each coupled pair requires approximately 400 MHz, the SoC supports only 20 qubits at 3.5 mW per qubit. The conclusion reports only the 20-qubit figure. Since the abstract and conclusion advertise scalability, the authors should state which bandwidth allocation is relevant to the proposed four-qubit chip and justify it with the actual resonator and qubit frequency separations.","section":"Quantum Analog chip / scalability paragraph"}],"minor_comments":[{"comment":"The caption and the main text disagree about which curve is S11 and which is S21; please correct this inconsistency.","section":"Fig. 2a caption"},{"comment":"There are numerous typos and inconsistent notations, including 'Abstarct' in the abstract, 'Quaron' versus 'Quarton', and 'imagnary' in Eq. (5); a careful proofread is needed.","section":"Abstract and throughout"},{"comment":"Equations (3) and (5) are rendered with garbled symbols; the definitions of n1, n2, and nb as expectation values, and the meaning of the 'current probabilities' I2 and I4, should be stated precisely.","section":"Eqs. (3) and (5)"},{"comment":"The reference list contains incomplete entries, for example refs [1]-[3] are arXiv preprints with version suffixes and no journal data, and the capitalization of 'Blochnium' is inconsistent.","section":"References"},{"comment":"The caption of Fig. 5 says 'Qiskit Metal simulation' but no simulation parameters or mode profile are shown; please clarify what exactly is plotted and how it confirms the g coupling.","section":"Fig. 5 caption"},{"comment":"The post-layout simulations are described as confirming robustness, but no corner analysis, temperature variation, or parasitic extraction details are reported; a brief description of the simulation setup would improve reproducibility.","section":"Post-layout simulation description"}],"recommendation":"reject","confidential_remarks":"This is a design-study manuscript whose central claim conflicts with the standard understanding of Purcell decay: a 0 dB transmission minimum at the qubit frequency does not block qubit emission to the feedline. The paper would need either a corrected physical mechanism with a quantitative Gamma_P calculation or a substantial reformulation of the claimed Purcell-filter replacement. Given the scope of the paper, I do not see a path to acceptance without changing the central claim, so I recommend rejection rather than a revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take. The paper has a genuinely interesting idea: use the natural frequency selectivity of a modified JPA to place readout resonators at gain peaks and qubits at gain troughs, thereby eliminating a separate Purcell filter. The SoC design in 45nm CMOS is substantial, with post-layout simulations showing 72 dB gain, NF<1 dB, 70 mW, and a compact footprint. That part looks like real engineering, though it is all simulation.\n\nThe problem is the headline claim. The JPA sits downstream of the feedline, after the readout resonators. Purcell decay is set by the qubit-resonator coupling and the resonator's external coupling to the feedline; the JPA's gain at the qubit frequency does not change that. A 0 dB trough means the amplifier doesn't amplify at that frequency, not that it presents high impedance isolation from the feedline. The paper never computes a Purcell rate with and without the JPA. So the central mechanism is unsupported, and not just due to missing measurements. The loading/instability concern you flagged is real, but secondary; even an ideal, stable gain profile wouldn't do what the authors claim.\n\nThe entanglement result is also weak: an extra FFT peak in the current probabilities is not an entanglement witness. The Hamiltonian in Eq. 3 looks hand-assembled, with a four-body coupling term g added without derivation. The derivation of the gain profile is deferred to the authors' previous paper, and the circularity of choosing frequencies from the simulated profile is present.\n\nWhat credit is due: the paper is honest about what is simulated and what is analytic, the Qiskit Metal cross-Kerr table gives concrete numbers, and the frequency-allocation co-design is a reasonable system-level exercise. The SoC block is a separate, potentially useful contribution, but it is not enough to carry the paper's quantum claims.\n\nMy recommendation: this should go to peer review because the architecture idea is worth a serious expert look and the SoC work is detailed, but a reviewer should insist on a Purcell calculation and, ideally, a measurement of the gain profile under chip loading. As is, it is a clear reject.","headline":"The JPA gain-trough-as-Purcell-filter claim is physically unsupported—0 dB gain is not isolation—though the SoC design work is substantial.","tokens_in":16875,"tokens_out":2870,"would_cite":false,"duration_ms":33018,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper claims that one engineered Josephson parametric amplifier, with a comb-shaped gain profile, can simultaneously provide quantum-limited amplification of readout signals and suppress qubit energy leakage, eliminating the Purcell…","keywords":["quantum chip","Josephson parametric amplifier","Purcell filter","superconducting qubits","readout","system-on-chip","45 nm CMOS","entanglement"],"falsifier":"Fabricate the modified BJPA, bias it at the simulated flux and pump settings, and measure its gain while it is connected to the four-qubit chip and feedline; if the gain at any qubit frequency is not within a few dB of 0 dB, or if the 24 dB peaks shift by more than a resonator linewidth, the Purcell-filter function disappears. A second direct check is to measure qubit relaxation through the readout channel with the JPA on: if it is not suppressed relative to the no-filter case, the claim fails.","tokens_in":15785,"feed_emoji":"⚛️","tokens_out":10905,"duration_ms":114592,"temperature":0.7,"pith_summary":"The paper proposes a readout architecture in which a single Josephson parametric amplifier does the work of two components: it amplifies the weak readout signals from superconducting qubits and, at the same time, blocks qubit energy from leaking into the readout line. The amplifier is engineered to have a comb-like gain spectrum across the 4–8 GHz band, with roughly 24 dB peaks and 0 dB valleys; the four qubits are placed at the valleys and their readout resonators at the peaks. A four-qubit chip and a 45 nm CMOS receiver chip (72 dB gain, below 1 dB noise figure, about 70 mW) are designed around this frequency plan. If the gain profile is real and stable, the design removes the need for separate Purcell filters and points toward readout electronics that scale to many qubits with sub-milliwatt power per qubit.","feed_headline":"One amplifier can replace the Purcell filter in qubit readout","feed_subtitle":"Gain peaks amplify readout signals while gain dips block qubit energy leakage, simplifying the chip.","key_machinery":"The mechanism that carries the argument is the engineered gain profile of the modified Blochnium Josephson parametric amplifier. Each cell is a Quarton—a nonlinear Josephson-junction block with a primary SQUID for tuning and secondary junctions for fine control—and two such cells coupled through a resonator create a spectrum with multiple separated resonances. The paper takes the simulated 24 dB peaks and 0 dB valleys as a fixed frequency map, then assigns qubit transition frequencies to the valleys and readout-resonator frequencies to the peaks. That frequency allocation is what turns a single nonlinear amplifier into a combined amplifier-plus-Purcell-filter, because the same device that boosts the resonator signal also suppresses qubit emission into the readout line.","core_discovery":"The central discovery is that the gain profile of a modified Blochnium Josephson parametric amplifier can be shaped into a frequency comb, and that this comb can be used as the organizing principle for the whole readout chain. The amplifier consists of two Quarton cells—Josephson-junction blocks with primary and secondary SQUIDs—coupled through a resonator, giving extra degrees of freedom compared with a standard BJPA. Its simulated gain has sharp peaks of about 24 dB at the resonator frequencies and deep minima near 0 dB at the qubit frequencies. The paper argues that this is enough to make the JPA act as its own Purcell filter: qubits sitting in the minima cannot efficiently emit into the readout channel, while resonator photons at the peaks are amplified near the quantum limit. In addition, the coupled-qubit analysis shows a new spectral component appearing under stronger coupling, which the paper reads as a signature of entanglement and uses to set the readout resonator frequency. On top of this sits a 45 nm CMOS system-on-chip that amplifies the resulting signal by 72 dB with noise below 1 dB and down-converts it to a 200 MHz intermediate frequency, with post-layout simulations showing the full chain working from a roughly 100 nV input.","pith_inferences":["A generalization the paper does not pursue: any parametric amplifier whose gain spectrum has deep minima could be used the same way, so Purcell protection could become a standard frequency-planning step rather than a separate component.","Because the comb is created by the primary SQUID flux bias, one could actively track the peaks and valleys after fabrication; that would turn the Purcell-filter function into a reconfigurable feature and relax the fabrication-precision requirement.","The entanglement readout argument yields a calibration recipe beyond the paper's scope: monitor the amplitude of the fc component while tuning the coupling strength and use it as a live indicator for entangling-gate activation."],"forward_implications":["A quantum chip can be read out without a Purcell filter, removing extra components, their insertion loss, and their design constraints from the cryogenic chain.","Qubit relaxation through the readout line is suppressed by placing qubit frequencies at the JPA gain minima, protecting coherence without extra filtering circuitry.","Readout resonator signals receive up to about 24 dB of quantum-limited amplification before leaving the cryostat, improving signal-to-noise and readout fidelity.","The 45 nm CMOS SoC's sub-1 dB noise figure and 72 dB gain could make a separate HEMT amplifier unnecessary, while its bandwidth math supports up to 160 qubits at roughly 0.44 mW per qubit in a conventional frequency plan.","The coupled-qubit analysis predicts a distinct spectral component fc under stronger coupling; tuning the readout resonator to fc would let the same chain sense entangled states rather than only single-qubit populations."],"supporting_citations":[{"why":"Defines the Blochnium JPA that this design modifies; the paper takes its Hamiltonian, eigenmode, and gain-calculation framework from here.","marker":"[22]"},{"why":"Establishes the dispersive circuit-QED readout setting where qubit leakage into the measurement line is a known problem.","marker":"[10]"},{"why":"Provides the quantitative treatment of Purcell filtering that the paper claims its gain-minima placement can replace.","marker":"[12]"},{"why":"Demonstrates Purcell-limited qubit relaxation in circuit QED, the decoherence mechanism the JPA valleys are meant to suppress.","marker":"[13]"},{"why":"Gives the effective-LC eigenvalue method used to map the JPA cells to equivalent circuits and compute their modes.","marker":"[18]"},{"why":"Provides the experimental and theoretical JPA treatment whose quantum-Langevin machinery is used to compute signal and idler gain.","marker":"[21]"},{"why":"Shows time-resolved current readout of a Cooper-pair box, the experimental pattern behind the probe-current formalism.","marker":"[33]"},{"why":"Demonstrates entanglement of two charge qubits, used as the experimental analogue for the entangled current-probability signatures.","marker":"[34]"},{"why":"Benchmarks cryogenic low-noise amplifier performance and supports the paper's claim that a CMOS LNA can replace the HEMT stage.","marker":"[26]"}],"fun_headline_variants":["Gain comb lets one amplifier replace Purcell filter","Single JPA provides quantum-limited gain and Purcell protection","Frequency comb gain profile simplifies qubit readout chip","Amplifier's gain notch and peak replace Purcell filter","Shaped JPA gain acts as its own Purcell filter"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole design rests on the assumption that the simulated comb-like gain profile of the modified amplifier—24 dB peaks at resonator frequencies and 0 dB valleys at qubit frequencies—is a real, stable property of the fabricated device and remains unchanged when the amplifier is connected to the four-qubit chip and the rest of the readout chain.","fun_headline_variants_meta":{"raw":{"variants":["Gain comb lets one amplifier replace Purcell filter","Single JPA provides quantum-limited gain and Purcell protection","Frequency comb gain profile simplifies qubit readout chip","Amplifier's gain notch and peak replace Purcell filter","Shaped JPA gain acts as its own Purcell filter"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00043,"raw_usage":{"total_tokens":2167,"prompt_tokens":884,"completion_tokens":1283,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":500,"completion_tokens_details":{"reasoning_tokens":1212}},"tokens_in":500,"tokens_out":1283,"duration_ms":11829,"temperature":1.0,"reasoning_tokens":1212,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T23:44:50.678605+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Fabricate the modified BJPA, bias it at the simulated flux and pump settings, and measure its gain while it is connected to the four-qubit chip and feedline; if the gain at any qubit frequency is not within a few dB of 0 dB, or if the 24 dB peaks shift by more than a resonator linewidth, the Purcell-filter function disappears. A second direct check is to measure qubit relaxation through the readout channel with the JPA on: if it is not suppressed relative to the no-filter case, the claim fails.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the quantitative treatment of Purcell filtering that the paper claims its gain-minima placement can replace."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrates Purcell-limited qubit relaxation in circuit QED, the decoherence mechanism the JPA valleys are meant to suppress."},{"cited_title":"Planat, Ph.D","cited_arxiv_id":null,"evidence_quote":"Gives the effective-LC eigenvalue method used to map the JPA cells to equivalent circuits and compute their modes."},{"cited_title":"Planat, R","cited_arxiv_id":null,"evidence_quote":"Provides the experimental and theoretical JPA treatment whose quantum-Langevin machinery is used to compute signal and idler gain."},{"cited_title":"Nakamura, Y","cited_arxiv_id":null,"evidence_quote":"Shows time-resolved current readout of a Cooper-pair box, the experimental pattern behind the probe-current formalism."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrates entanglement of two charge qubits, used as the experimental analogue for the entangled current-probability signatures."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Benchmarks cryogenic low-noise amplifier performance and supports the paper's claim that a CMOS LNA can replace the HEMT stage."},{"cited_title":"Salmanogli, H","cited_arxiv_id":null,"evidence_quote":"Defines the Blochnium JPA that this design modifies; the paper takes its Hamiltonian, eigenmode, and gain-calculation framework from here."},{"cited_title":"Blais, R","cited_arxiv_id":null,"evidence_quote":"Establishes the dispersive circuit-QED readout setting where qubit leakage into the measurement line is a known problem."}],"review_version":1}