{"id":"eab30151-9f02-4170-a478-453358339ece","arxiv_id":"2607.18059","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A commercial CMOS foundry process is repurposed as a multimodal platform, achieving low-power NV ODMR and cryogenic magnetic susceptibility measurements of Fe3GeTe2 without sample-specific fabrication.","lead":"Researchers turned a standard 65-nm computer chip process into a measurement platform that can drive quantum sensors and probe magnetic materials. The chip delivers microwaves to a diamond sensor with roughly 100x less power than a commercial antenna and measured a known magnetic transition in a 2D material at 1.75 K.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The quantitative susceptibility χ≈12.2 rests on an unreported simulated inductor field and a bulk-literature remanence calibration; the qualitative Tc peak is robust, but the numerical claim is not yet supported.","rationale":"The paper's central claim has multiple parts; I considered the ODMR power-reduction benchmark and the 'multimodal' framing, but neither is as decisive. The power-reduction comparison is setup-specific and could be debated, but the reported sensitivities and contrasts are directly measured; the multimodal framing overstates maturity but is clearly identified as subsystem demonstrations. The susceptibility number is a quantitative result presented without error bars and based on an invisible calibration. It is the single most load-bearing because the abstract explicitly claims 'cryogenic magnetic susceptibility measurements' as a key capability, and the quantitative χ anchors that claim. The reader's weakest_assumption identified this same issue; I agree. Recommendation: keep the CONDITIONAL verdict; the authors should report the field calibration and uncertainties. No change to reader's verdict.","tokens_in":13790,"tokens_out":6418,"duration_ms":69811,"concrete_test":"Request the authors to (1) report the simulated on-chip B-field amplitude and position used in the χ=12.2 estimate, and (2) calibrate it experimentally with an independent probe — e.g., measure NV Rabi frequency at the same standoff height above the inductor and compute B1 = 2πΩ/γ, then compare with the simulation. Recompute χ using the calibrated field. If the calibrated field differs by >20% from the value used, the quoted χ is unsupported and should be revised or removed.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing concern is the calibration of the absolute susceptibility reported in the 'Magnetic susceptibility measurements' section. The paper states: 'Knowing the remanent magnetization value of Fe3GeTe2 from previous literature [23] and the maximum field value generated by the CMOS chip, we conclude that at 220 K we measure a susceptibility of χ ≃ ΔM/ΔH ≃ 12.2.' Neither the simulated field amplitude at the flake nor its calibration is reported. The conversion from Kerr rotation to ΔM uses a bulk literature remanence, which may not represent the transferred thin flake at 220 K, and no uncertainty is propagated into χ or into the stated detection limit (~0.75). Since χ scales linearly with the assumed field amplitude, an uncalibrated or inaccurate EM simulation directly changes the headline numerical result, even though the qualitative peak at Tc is robust. This is the weakest point in the otherwise well-supported demonstration.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports a CMOS chip fabricated in a commercial 65 nm process that integrates an RF octagonal inductor, a resistive meander heater, and interdigitated electrodes within a 1 mm2 footprint, with the metal stack partitioned by function. The authors demonstrate three applications: Kerr-rotation magnetic susceptibility measurements on hBN/Fe3GeTe2/hBN heterostructures transferred directly onto the chip at cryogenic temperatures, room-temperature ODMR on NV centers with >20% contrast at 4–9 dBm, and SEM imaging during device operation with no resolvable image degradation. The paper claims this establishes a scalable, foundry-manufacturable platform for multimodal quantum materials characterization.","tokens_in":14013,"tokens_out":7673,"duration_ms":77590,"significance":"If the results hold, the platform addresses a real need for reproducible, scalable hardware in quantum materials research. The foundry CMOS approach is attractive because it uses a standard commercial process, and the functional partitioning of the metal stack is a sensible design. The experimental demonstrations are relevant: the susceptibility measurement on an exfoliated air-sensitive flake without sample-specific lithography is a useful capability, and the direct benchmarking against a commercial antenna on the same microscope is a strength. The paper also provides open-source analysis code for the knife-edge SEM analysis, which supports reproducibility. The main quantitative weakness is the calibration of the absolute susceptibility, which is not documented.","major_comments":[{"comment":"The absolute susceptibility χ≈12.2 is derived from an unreported simulated field value ('the maximum field value generated by the CMOS chip') and a bulk-literature remanent magnetization [23]. The conversion from Kerr rotation to ΔM is not described, no drive current is specified, and no uncertainty is propagated into χ or the detection limit (~0.75). Since χ scales linearly with the assumed field amplitude, the headline numerical value is not reproducible from the manuscript. Provide the simulated field value at the flake position, the applied inductor current, the calibration procedure linking Kerr rotation to magnetization units, and a full uncertainty budget.","section":"Magnetic susceptibility measurements"},{"comment":"The claimed '20–25 dB reduction in required power' relative to the commercial antenna is derived from a range comparison (24.6–30 dBm vs 4–9 dBm) that yields a spread of ~15.6 to 26.5 dB depending on the operating points chosen. Specify the exact pair of power levels used for the comparison (e.g., minimum power achieving ≥20% contrast in each case, or matched contrast) and show the corresponding data from the power sweep (Supplementary Fig. S4). Without this, the quantitative power-reduction claim is not uniquely determined.","section":"Quantum control enabled by the CMOS RF architecture"}],"minor_comments":[{"comment":"References [22] and [23] are identical (May et al., PRB 93, 014411, 2016); merge them and update the in-text citations accordingly.","section":"References"},{"comment":"The magnetic sensitivity values (2.2–2.9 µT/√Hz) are reported without specifying the measurement parameters (e.g., integration time, photon collection rate, linewidth, and the formula used). Provide these details so the values are reproducible.","section":"Quantum control enabled by the CMOS RF architecture"},{"comment":"The data availability statement says data are available 'upon request'. For a platform paper claiming reproducibility and scalability, depositing raw data and analysis scripts (beyond the knife-edge code) in a public repository would strengthen the contribution.","section":"Data availability"},{"comment":"The phrase 'maximum field value generated by the CMOS chip' is ambiguous. Specify the drive current, frequency, and any matching conditions under which the field is defined; this is closely related to the major comment above but should be clarified even in a revised manuscript.","section":"Magnetic susceptibility measurements"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is promising and the core experimental demonstrations appear sound. The principal revision needed is to document the susceptibility calibration and to tighten the ODMR power-reduction comparison. With those changes, I would likely support acceptance. The duplicate reference is a minor editorial issue."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"What's actually new: a single commercial 65-nm CMOS chip repurposed as a passive multimodal platform, with an on-chip inductor, resistive heater, and interdigitated electrodes in one footprint, then used for NV ODMR and cryogenic magnetic susceptibility of Fe3GeTe2 without sample-specific lithography. That integration is real, and the direct power comparison against a commercial antenna is a genuine benchmark: >20% ODMR contrast at 4–9 dBm versus 24.6–30 dBm for the antenna, with sensitivities in the few µT/√Hz range. The SEM-compatibility section is also carefully done; the knife-edge analysis with sigmoid fits is methodical and the code is on GitHub, which deserves credit.\n\nThe susceptibility measurement has a solid qualitative core: the peak near 220 K matches Fe3GeTe2's known Curie temperature, and the linear-bias and frequency-flatness checks support the interpretation. The soft spot is the one the stress-test flags: the absolute susceptibility χ ≈ 12.2 is computed using an undisclosed simulated inductor field and a bulk literature remanence. No calibration of the field at the flake, no uncertainty propagation, no error bars. The claim scales linearly with that field, so the quantitative headline is not supported as written. That is a load-bearing missing number, not a stylistic quibble. The detection limit of ~0.75 is likewise unsupported because it inherits the same calibration gap.\n\nMinor issues: the heater and IDE are characterized electrically but never demonstrated on a material, so the \"multimodal\" framing runs ahead of the evidence; the data availability statement is \"upon request\" rather than a public archive; and Table 1's sensitivity comparison is explicitly not apples-to-apples, which is honest but weakens the framing. None of these are fatal, and the ODMR result plus the qualitative susceptibility peak would survive even if the absolute number changes.\n\nWho should read this: people developing CMOS-integrated quantum sensing or looking for reproducible cryogenic measurement platforms. It deserves a serious referee, not a desk reject, but referee attention should focus on the field calibration and the quantitative susceptibility propagation. If those are supplied—or the quantitative claim is downgraded to a qualitative demonstration—the paper becomes a solid contribution.\n\nRecommendation: send to peer review, with a concrete request for the simulated field amplitude, calibration procedure, and uncertainty analysis on χ and the stated sensitivity.","headline":"A genuine integration result—RF, heater, and electrodes on one foundry CMOS chip—with a strong ODMR power benchmark and a robust qualitative susceptibility peak, undercut by an uncalibrated absolute susceptibility number that should be fixed before publication.","tokens_in":14545,"tokens_out":1186,"would_cite":true,"duration_ms":15428,"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":"A commercial 65-nm CMOS chip, repurposed as a passive platform, can perform cryogenic magnetic susceptibility measurements and low-power NV-center magnetometry without custom fabrication.","keywords":["CMOS integration","foundry platform","quantum materials characterization","nitrogen-vacancy centers","ODMR","magnetic susceptibility","Fe3GeTe2","cryogenic"],"falsifier":"Measure the RF magnetic field at the sample position (e.g., with a calibrated pick-up coil, a known paramagnet, or a spectroscopy-based field probe) and compare it with the electromagnetic simulation used to compute χ. If the measured field differs substantially, the absolute susceptibility claim fails even though the phase-transition peak remains.","tokens_in":13688,"feed_emoji":"🧲","tokens_out":3827,"duration_ms":38503,"temperature":0.7,"pith_summary":"This paper claims that a standard commercial 65-nm CMOS chip, built for digital and radio-frequency circuits, can be repurposed as a passive, foundry-manufacturable platform for characterizing quantum materials. The authors partition the chip's metal stack into an on-chip microwave inductor, a resistive heater, and interdigitated electrodes—all within a 1 mm² footprint—and demonstrate three functions: cryogenic magnetic susceptibility measurements of the van der Waals magnet Fe3GeTe2 at temperatures down to 1.75 K, room-temperature NV-center optically detected magnetic resonance with greater than 20% contrast at only 4–9 dBm microwave power, and compatible operation inside a scanning electron microscope. The significance lies in replacing experiment-specific hardware with a reproducible, mass-producible chip that delivers localized microwave fields, making quantum sensing and materials characterization cheaper, smaller, and more standardized.","feed_headline":"Foundry chip cuts quantum-sensing microwave power by 100×","feed_subtitle":"Standard 65-nm CMOS stack is repurposed into a multimodal platform for cryogenic magnetic susceptibility and NV-center magnetometry.","key_machinery":"The load-bearing object is the on-chip octagonal loop inductor (four turns, 212 µm outer span, 10 µm conductor width, 2 µm spacing) fabricated in the top metal layers of the CMOS stack. Electromagnetic simulations and S-parameter measurements show that this inductor delivers a localized RF magnetic field within the central region, and the near-field localization is what allows high ODMR contrast (>20%) at only 4–9 dBm, a 20–25 dB power reduction compared to a commercial antenna. The platform also partitions lower metal layers into a resistive meander heater (M5–M6) and interdigitated electrodes (M7), giving thermal control and electrical interfacing in the same 1 mm² footprint.","core_discovery":"The paper's central discovery is that the existing metal stack of a commercial 65-nm CMOS process can be 'functionally partitioned' into the passive subsystems needed for quantum materials experiments. The integrated RF architecture—specifically a four-turn octagonal loop inductor on the top metal layers—concentrates microwave magnetic fields near the chip surface, enabling AC magnetic excitation and spin control. The authors validate the platform through three demonstrations: an hBN-encapsulated Fe3GeTe2 flake transferred directly onto the chip shows a magnetic susceptibility peak near 220 K (the known Curie temperature) with an absolute susceptibility of about 12.2; NV centers in diamond s","pith_inferences":["A natural extension is to calibrate the inductor's absolute field strength using a standard paramagnetic or superconducting sample, which would turn the susceptibility measurement from a relative peak into a quantitative magnetometer.","The near-field RF architecture could be adapted to other spin qubits and to pulsed or AC-modulated magnetometry, since the inductor supports kilohertz-to-megahertz field modulation that superconducting magnets cannot.","The 1 mm² platform suggests a route to dense arrays of independent measurement sites, each with its own heater and electrodes, enabling combinatorial screening of phase diagrams.","Combining this passive platform with existing cryo-CMOS active circuits (microwave generation, readout) could yield a fully integrated, autonomous materials-characterization system."],"forward_implications":["If correct, quantum materials labs can replace custom-built antennas and sample-specific excitation structures with a standard foundry chip, improving reproducibility and reducing development time.","The 20–25 dB reduction in microwave power needed for NV-center ODMR would allow compact, low-power quantum sensors, possibly battery-powered or integrated into portable devices.","Because the chip is foundry-manufacturable, arrays of identical platforms could enable high-throughput, automated materials characterization across multiple laboratories.","The demonstrated compatibility with electron-beam imaging means the same chip can support combined structural, electrical, magnetic, and optical probes in one workflow.","The integrated heater and IDE open the door to temperature- and field-dependent studies of air-sensitive van der Waals materials without lithographic processing."],"fun_headline_variants":["Foundry chip becomes a quantum materials testing lab","CMOS chip slashes quantum sensing microwave power 100×","Chip probes quantum materials at 1.75 K","Standard CMOS runs multimodal quantum sensing","NV sensing on CMOS chip uses 20 dB less power"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The quantitative susceptibility value (χ ≈ 12.2) is computed using a simulated maximum inductor field that is not reported or experimentally calibrated; the qualitative peak at the known Curie temperature (220 K) is robust, but the absolute number depends on that uncalibrated simulation.","fun_headline_variants_meta":{"raw":{"variants":["Foundry chip becomes a quantum materials testing lab","CMOS chip slashes quantum sensing microwave power 100×","Chip probes quantum materials at 1.75 K","Standard CMOS runs multimodal quantum sensing","NV sensing on CMOS chip uses 20 dB less power"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001609,"raw_usage":{"total_tokens":6224,"prompt_tokens":707,"completion_tokens":5517,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":451,"completion_tokens_details":{"reasoning_tokens":5456}},"tokens_in":451,"tokens_out":5517,"duration_ms":43030,"temperature":1.0,"reasoning_tokens":5456,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T16:12:21.354905+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the RF magnetic field at the sample position (e.g., with a calibrated pick-up coil, a known paramagnet, or a spectroscopy-based field probe) and compare it with the electromagnetic simulation used to compute χ. If the measured field differs substantially, the absolute susceptibility claim fails even though the phase-transition peak remains.","supporting_citations":[],"review_version":1}