{"id":"2e45505b-2ba7-4d20-90e5-fc4b648bfde6","arxiv_id":"1908.02546","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"A reconfigurable inline spin-wave majority gate in a metallic ferromagnetic waveguide is demonstrated electrically at sub-micrometer wavelengths and at two resonant frequencies.","lead":"Researchers built a sub-micrometer spin-wave majority gate inside a magnetic waveguide and showed it computes the majority of three input signals from wave interference. The all-electrical device can swap input and output roles and operate at multiple frequencies, with single-channel demonstrations for each.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"All-electrical truth table assumes equal spin-wave amplitudes at the output, but only input microwave amplitudes are equalized; unequal per-path transmission could make the measured phase follow a minority input.","rationale":"The paper's central claim is the all-electrical majority truth table in an 850-nm CoFeB waveguide and in the 4.7-µm device at two harmonics. The majority function is defined by equal weighting of the three inputs, and the measurement premise of equal output amplitudes is exactly what the Methods do not establish. This is the weakest load-bearing point, not because the authors made an obvious error but because the amplitude-equalization statement concerns the electrical input ports while the propagation paths to the output have different lengths. A concrete per-path S21 measurement would settle it without redoing fabrication. The STXM phase maps are real independent evidence for majority operation in the 2-µm permalloy waveguide, which is why this concern supports a conditional acceptance rather than a rejection. The reader's weakest assumption is the same concern, so I agree with the CONDITIONAL verdict and leave it unchanged.","tokens_in":12994,"tokens_out":8488,"duration_ms":97200,"concrete_test":"Measure the complex spin-wave transmission of each input path individually at the relevant operating points (e.g., 13.86 GHz and 90 mT for Fig. 3c; 12.22 and 14.92 GHz at 42 mT for Fig. 5c,d; 17.72 GHz at 12 mT for Fig. 5e), with the other two input antennas terminated and with the same attenuator and phase-shifter settings. From the three measured vectors T1, T2, and T3, predict the output for each of the eight phase patterns as T1 e^{iφ1} + T2 e^{iφ2} + T3 e^{iφ3}, and overlay these predictions on the measured polar plots. If all eight predicted output quadrants and strong-versus-weak classifications match the measured data, the amplitude-balance assumption is vindicated; if not, the truth table must be re-evaluated after adding per-path amplitude compensation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The all-electrical demonstration of the majority truth table rests on an unstated premise: the three spin-wave contributions arriving at the output port have equal amplitudes. The Methods say only that 'variable attenuators to equalise the amplitudes of the microwave signals at each input port' were used; nothing calibrates the spin-wave transmission of each input-output path. In the inline geometry, P1, P2, and P3 are at distances 6F, 4F, and 2F from O, so even with identical input currents the output amplitudes A1, A2, and A3 differ because of propagation loss and any antenna or mode-conversion asymmetry. For a phase-coded gate the output is A1 e^{iφ1} + A2 e^{iφ2} + A3 e^{iφ3}. If A3 > A1 + A2, the input (0,0,π) yields an output phase of π instead of 0; if A1 > A2 + A3, (0,π,π) yields 0 instead of π. The measured separated clusters would then be a weighted vote, not MAJ. The STXM measurements give independent support for the 2-µm permalloy device, but the headline sub-µm all-electrical claim (Figs. 3 and 5) has no per-path amplitude calibration, and baseline or crosstalk subtraction for the S21 data is not described.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports an inline spin-wave majority gate implemented in CoFeB and permalloy waveguides with inductive microwave antennas. Logic is encoded in the phase (0 or pi) of spin waves launched by three input ports, and the output phase is read at a fourth port. The authors present time-resolved scanning transmission x-ray microscopy (STXM) imaging of the spin-wave interference in a 2.0-um-wide permalloy waveguide, showing the full majority truth table, and all-electrical vector-network-analyzer measurements in 850-nm-wide and 4.7-um-wide CoFeB waveguides. The all-electrical experiments demonstrate majority operation, reconfigurable output ports, frequency-division multiplexing at N=1 and N=2 resonances, and a minority function at a detuned frequency. The paper also discusses area scaling (16F^2), fan-out, and the potential for a spin-wave-based computing platform.","tokens_in":13185,"tokens_out":9655,"duration_ms":99889,"significance":"If the claims are fully substantiated, this is a significant advance over earlier millimeter-scale YIG trident gates: it moves to metallic waveguides with sub-micron width and wavelength, all-electrical readout, and demonstrates two logic functions (majority and minority) in a single device. The minority-function prediction at 11.89 GHz is a concrete falsifiable test that is confirmed experimentally, and the STXM images directly reveal the spin-wave interference pattern behind the majority operation. The main caveat is that the all-electrical sub-micron truth table depends on an unstated premise of equal spin-wave amplitudes from the three input paths, which is not explicitly calibrated in the manuscript.","major_comments":[{"comment":"The equalization of microwave amplitudes at the input ports does not guarantee equal spin-wave amplitudes at the output. The three paths from P1, P2, and P3 to the output port O have lengths 6F, 4F, and 2F, so propagation losses and any antenna or mode-conversion asymmetry make the output amplitudes A1, A2, and A3 differ. For phase-coded interference, the output is A1 exp(i*phi1) + A2 exp(i*phi2) + A3 exp(i*phi3); if one amplitude exceeds the sum of the other two, the phase of the sum can follow a minority input rather than the majority (for example, if A3 > A1+A2, the input (0,0,pi) yields an output phase of pi instead of 0, and if A1 > A2+A3, the input (0,pi,pi) yields 0 instead of pi). The manuscript does not report per-path spin-wave transmission calibrations (such as single-input measurements at each port) or a robustness check against amplitude mismatch. Since the sub-micron all-electrical claims in Figs. 3 and 5 rest on this premise, this is a load-bearing point that needs to be addressed.","section":"Methods - All-electrical microwave measurements"},{"comment":"The all-electrical truth tables are extracted from S21 measurements, but the paper does not describe any subtraction of the electromagnetic crosstalk baseline between the input and output antennas. The manuscript mentions that U-shaped antennas were chosen for their low parasitic crosstalk, but no control measurement without propagating spin waves (e.g., at a magnetic field above saturation or with no excitation) is shown. A field-independent crosstalk contribution would add a constant phasor to the coherent output and could shift the extracted phase for weak-majority states, potentially making the truth-table separation less meaningful. Please provide a control experiment or clearly describe how crosstalk was excluded from the phase analysis.","section":"Electrical operation of nanoscale spin-wave majority gates (Figs. 3c,d) and Frequency-division multiplexing (Figs. 5c-e)"},{"comment":"The abstract lists fan-out as a key feature of the device, but the fan-out-of-two claim is inferred from the bidirectional spin-wave propagation visible in the STXM images (Fig. 2) and from the symmetry of the geometry, not from a direct demonstration of simultaneous readout at two output ports. To substantiate the fan-out claim, the authors should either show an all-electrical or STXM measurement with two output ports read concurrently, or soften the wording to 'bidirectional propagation, which enables fan-out in principle.' This distinction matters because the abstract explicitly promises a demonstrated fan-out.","section":"Abstract and Time-resolved imaging of spin-wave majority gate operation"}],"minor_comments":[{"comment":"The sentence 'contributions from backward-volume spin wave modes are also are also visible' contains a duplicated phrase; please correct it.","section":"Main text, page 8"},{"comment":"The claim that the majority gate 'works equally well' in the two port configurations is based on visual inspection of the frequency traces; a quantitative comparison of the separation between the '0' and 'pi' clusters (e.g., the minimum distance between the two sets of curves) would be more convincing.","section":"Fig. 4"},{"comment":"Please specify the calibration procedure and accuracy of the delay-based phase shifters, and state how the phase at each port was set and verified at the target frequencies.","section":"Methods - All-electrical microwave measurements"},{"comment":"Reference [41] contains a punctuation error: 'Liu, X., Zhang, W. Carter, M.J. & Xiao, G.' should read 'Liu, X., Zhang, W., Carter, M.J. & Xiao, G.'","section":"References"},{"comment":"The stated gate area of 16F^2 (20F^2 with fan-out) counts only the active waveguide region, not the transducers, feedlines, and contact pads; please clarify this in the area comparison with CMOS majority gates.","section":"Operation principle and Implementation sections"},{"comment":"The term 'reconfigurable input and output ports' is used to describe experiments that select among fixed physical ports, not runtime reconfiguration; please adjust the wording to avoid implying dynamic reprogrammability.","section":"Reconfigurability section"}],"recommendation":"major_revision","confidential_remarks":"The manuscript should clearly distinguish its advance over the authors' earlier IEDM paper (ref. [26]) on the linear majority gate concept; the present work adds sub-micron dimensions, frequency-division multiplexing, and a broader set of experiments, but the novelty relative to [26] is not spelled out. The central issue for the referee is the uncalibrated amplitude balance in the all-electrical truth-table measurement, which should be addressed with additional data before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a real advance in spin-wave logic, not a repackaging. The inline sub-micron majority gate works, with a full truth table measured two independent ways (STXM imaging and all-electrical spectroscopy), reconfigurable input/output ports, operation at two resonant harmonics in the same device, a minority function at a detuned frequency, and wavelengths down to 600 nm. The minority result is a genuine prediction from the dispersion model that the measurement confirms, which is the kind of evidence that separates a demonstration from a calibrated curve.\n\nThe STXM imaging of the 2-micron permalloy device is the most direct evidence and is convincing on its own. The all-electrical data in the 850-nm CoFeB device are internally consistent and the polar plots show clear separation of logic levels, including strong vs weak majority.\n\nThe soft spot is the one the stress test flags. The Methods equalize microwave amplitudes at the input ports, but the truth-table analysis assumes the three spin-wave contributions arriving at the output have balanced amplitudes. The ports sit at different distances from O, so propagation loss and antenna asymmetries set A1, A2, A3 differently. If one path dominates, the output phase can be a weighted vote rather than majority. The paper would be materially stronger with a supplemental figure showing each single-input transmission path and a statement of the amplitude margin, or an explicit calibration. This matters most for the 600-nm, low-bias measurement at 17.72 GHz, where the attenuation length is not quoted and the margin could easily be narrower.\n\nTwo smaller asks. First, the frequency-division multiplexing claim should be tightened: what is shown is sequential multi-frequency operation of the same device, not simultaneous independent channels. The interferometric principle should support true FDM, but that is not yet demonstrated. Second, no error bars or repeated-measurement statistics appear anywhere; a few repeated traces would help.\n\nThe relation to IEDM 2018 (ref 26) should also be explicit: the basic inline gate is from that earlier work, and the new contribution here is reconfigurability, the second harmonic, the minority function, sub-micron wavelengths, and fan-out. That is enough to make the paper a solid journal contribution once the amplitude question is answered.\n\nFor anyone working in magnonic logic or beyond-CMOS devices, this is a citable and useful result. It deserves serious peer review, not desk rejection. I would send it out with a request for the per-path amplitude calibration and the FDM wording fix.","headline":"A genuine experimental step forward in spin-wave majority logic, but the all-electrical truth table needs an explicit spin-wave amplitude-balance check to rule out a weighted vote.","tokens_in":13815,"tokens_out":4928,"would_cite":true,"duration_ms":52190,"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":"An inline spin-wave majority gate with three equally spaced electrical antennas computes the majority of three phase-encoded inputs by wave interference, demonstrated in an 850-nm-wide CoFeB waveguide.","keywords":["spin waves","magnonic majority gate","phase-encoded logic","interference-based computing","CoFeB waveguides","frequency-division multiplexing","scanning transmission X-ray microscopy","reconfigurable logic"],"falsifier":"Repeat the eight input-phase combinations while attenuating one input antenna by a known amount, for example 3 dB; if the output phase ever follows the weakened input's phase instead of the majority of the three phases, the equal-amplitude premise is violated and the gate is not robust to realistic path imbalance.","tokens_in":12755,"feed_emoji":"🧲","tokens_out":11700,"duration_ms":119649,"temperature":0.7,"pith_summary":"This paper reports an inline spin-wave majority gate: three equally spaced microwave antennas on a ferromagnetic waveguide inject spin waves whose phases ($0$ or $\\pi$) encode logic zero or one, and the phase of the interfering wave at a fourth antenna gives the majority vote. The authors reconstruct the full majority truth table all-electrically in an 850-nm-wide Co40Fe40B20 waveguide and in a 4.7-µm-wide device, and they image the spin-wave dynamics with time-resolved X-ray microscopy. They also show that the same gate lets any port serve as the output, that it works at two different resonant harmonics so one device carries multiple frequency channels, and that a detuned frequency turns it into a minority (inverted majority) gate. If these results hold, this supplies a nanoscale, electrically interfaced building block for magnonic logic that could be integrated with conventional semiconductor electronics.","feed_headline":"Spin-wave majority gate shrinks to 850 nm, runs all-electrically","feed_subtitle":"Full truth table by wave interference; any port can be the output and minority logic runs at a detuned frequency.","key_machinery":"The device is the inline spin-wave majority gate: input and output transducers placed at equally spaced positions $n\\times 2F$ along a waveguide, with $F$ smaller than the spin-wave attenuation length. Logic 0 and 1 are spin-wave phases $0$ and $\\pi$, and the output is the phase of the coherent sum of the three waves. The load-bearing condition is resonant operation, $2F = N\\lambda$, which makes identical microwave input phases arrive in phase at the output; the phase-sensitive readout extracts the output phase from the bias-field derivative of the microwave transmission parameter.","core_discovery":"The central discovery is that a straight ferromagnetic waveguide with three equally spaced inductive input antennas and an output antenna acts as a majority gate when information is encoded in spin-wave phase. Because each input-to-output distance is a multiple of the spin-wave wavelength under resonant operation, the three coherent spin waves arrive at the output with aligned phases, so the phase of their superposition is the majority phase. The paper verifies this by reconstructing all eight rows of the majority truth table from phase-sensitive all-electrical measurements at sub-micrometre scale, and by time-resolved X-ray imaging of a wider permalloy device. The same inline geometry also yields reconfigurable inputs and outputs, fan-out of two, frequency-division multiplexing, and a minority function at a detuned frequency.","pith_inferences":["The equal-amplitude caveat points to a direct stress test: calibrate each input-to-output spin-wave transmission separately and check whether the truth table survives a deliberate 3 dB imbalance.","The same resonant-interference argument should extend to more than three inputs by adding transducers at further multiples of $2F$; the paper mentions this possibility but does not test it.","If frequency-division multiplexing holds at higher power, two simultaneous tone sets on the same waveguide should yield two independent majority outputs that can be demodulated separately; that experiment would confirm true parallel computation.","The nonreciprocity of inductive excitation in the transverse-field geometry leaves the backward fan-out arm with a weaker signal; a natural follow-up is to test whether forward-volume spin waves or nonchiral transducers make fan-out symmetric."],"forward_implications":["The inline gate occupies $16F^2$ ($20F^2$ with a fan-out of two), far below the roughly $330F^2$ area of a CMOS majority gate, so phase-based magnonic logic becomes area-competitive if the scaling holds.","Because any port can be selected as the output, one fabricated device can be reprogrammed for different logic functions without changing the layout, demonstrated on a 4.7-µm-wide waveguide.","Resonant operation at two harmonics means the same gate can compute independently in two frequency channels at once, raising throughput without extra area.","A detuned frequency gives the minority function in the same device, so combining the majority gate with a $\\pi$ phase shift (inverter) yields a functionally complete logic set.","The truth table is also obtained at a 600-nm spin-wave wavelength, showing the interference mechanism survives at the sub-micrometre length scales needed for dense integration."],"supporting_citations":[{"why":"Introduces the spin-wave majority-gate concept that this work scales down to nanoscale dimensions.","marker":"[20]"},{"why":"Reports the earlier millimetre-scale YIG trident prototype that this inline design replaces.","marker":"[24]"},{"why":"Defines the equally spaced inline port arrangement used for the device geometry.","marker":"[26]"},{"why":"Provides the inductive-antenna transduction approach for launching and detecting sub-micrometre spin waves.","marker":"[32]"},{"why":"Supplies the all-electrical propagating spin-wave spectroscopy method used for phase-sensitive readout.","marker":"[33]"},{"why":"Provides the dipole-exchange dispersion relation used to model spin-wave transmission in the waveguides.","marker":"[36]"},{"why":"Defines the transverse magnetisation spin-wave geometry used in experiments and simulations.","marker":"[34]"}],"fun_headline_variants":["Spin-wave majority gate shrinks to sub-micron, runs all-electrically","All-electrical spin-wave majority gate now sub-micron","Sub-micron spin-wave majority gate with fan-out","Reconfigurable sub-micron spin-wave majority gate demonstrated","Spin-wave interference enables sub-micron majority gate"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The all-electrical truth table assumes the three spin-wave contributions arriving at the output have nearly equal amplitudes; the measurement equalises only the microwave amplitudes sent to the three input antennas, not the spin-wave transmission amplitude of each input-to-output path.","fun_headline_variants_meta":{"raw":{"variants":["Spin-wave majority gate shrinks to sub-micron, runs all-electrically","All-electrical spin-wave majority gate now sub-micron","Sub-micron spin-wave majority gate with fan-out","Reconfigurable sub-micron spin-wave majority gate demonstrated","Spin-wave interference enables sub-micron majority gate"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.002,"raw_usage":{"total_tokens":7746,"prompt_tokens":831,"completion_tokens":6915,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":447,"completion_tokens_details":{"reasoning_tokens":6828}},"tokens_in":447,"tokens_out":6915,"duration_ms":45660,"temperature":1.0,"reasoning_tokens":6828,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:41:24.731129+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Repeat the eight input-phase combinations while attenuating one input antenna by a known amount, for example 3 dB; if the output phase ever follows the weakened input's phase instead of the majority of the three phases, the equal-amplitude premise is violated and the gate is not robust to realistic path imbalance.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the spin-wave majority-gate concept that this work scales down to nanoscale dimensions."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports the earlier millimetre-scale YIG trident prototype that this inline design replaces."},{"cited_title":"SPIE10962, 1096202 (2019)","cited_arxiv_id":null,"evidence_quote":"Defines the equally spaced inline port arrangement used for the device geometry."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the inductive-antenna transduction approach for launching and detecting sub-micrometre spin waves."},{"cited_title":"& Bailleul, M","cited_arxiv_id":null,"evidence_quote":"Supplies the all-electrical propagating spin-wave spectroscopy method used for phase-sensitive readout."},{"cited_title":"Micromagnetic simulations of magnetoelastic spin wave excitation in scaled magnetic waveguides.Appl","cited_arxiv_id":null,"evidence_quote":"Provides the dipole-exchange dispersion relation used to model spin-wave transmission in the waveguides."},{"cited_title":"& Radu, I.P","cited_arxiv_id":null,"evidence_quote":"Defines the transverse magnetisation spin-wave geometry used in experiments and simulations."}],"review_version":1}