{"id":"7657cce9-1268-44f1-9b87-b03c4830322b","arxiv_id":"2504.12572","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"Time-resolved magnetoelectric measurements reveal a coherent ~44 GHz oscillation of the axion angle θ in 6-layer MnBi2Te4, driven by out-of-phase antiferromagnetic magnons, providing the first direct observation of a dynamical axion quasiparticle.","lead":"A 6-layer magnetic crystal, MnBi2Te4, shows a fast (44 GHz) oscillation of its magnetoelectric coupling after a laser pulse, which the authors identify as the long-predicted dynamical axion quasiparticle. The result is a first direct time-domain look at this condensed-matter analog of the QCD axion, with possible uses in ultrafast magnetism and dark matter detection.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 44 GHz E-linear Kerr signal could be a pump-induced modulation of the Kerr conversion factor γ rather than an oscillation of α; a wavelength-dependent test is needed.","rationale":"Good-faith reading: the experiment is careful and the internal controls are extensive. The static E-linear Kerr is convincingly shown to measure E-induced M_z (SI II.4: simultaneous Kerr/Faraday, wavelength dependence, DFT σxy). The 7-of-8 null check (SI I.4.1) argues against gate modification of J and K. However, none of these controls addresses the time-dependent conversion factor. The derivation above shows that Δγ(t) at the out-of-phase magnon frequency enters the E-linear Kerr signal directly, multiplied by the static α, with no symmetry suppression. The plausibility of Δγ/γ ≈ 12% is not merely hypothetical: the paper's own mechanism is a magnon-induced modulation of the Berry curvature, which should also modulate the optical conductivity that determines γ at 515 nm, and the static γ is strongly wavelength-dependent, so a 12% modulation is within the range of the material's magneto-optical response. The proposed multi-wavelength check is decisive because α(t) is wavelength-independent by definition, while the artifact inherits the wavelength dependence of γ. The reader identified the same weak point (listing γ among the alternatives); my focus narrows it to the γ term as the only one not already excluded by the authors' symmetry and gate-dependence arguments. The verdict remains conditional: accept only if the time-resolved wavelength (or equivalent time-resolved Faraday) check is provided.","tokens_in":38580,"tokens_out":14817,"duration_ms":171471,"concrete_test":"Repeat the double lock-in Δα(t) measurement at B∥ = 6 T and the same pump conditions (1030 nm pump, 160 μJ/cm²) at probe wavelengths 515 nm, 600 nm and 700 nm, converting the Kerr slope to α using the static γ(λ) calibration of Methods 3 and Extended Data Fig. 4. The true Δα(t) is a DC magnetoelectric quantity and must be wavelength-independent; the γ-modulation artifact is proportional to Δγ(λ)/γ(λ), which should track the strongly wavelength-dependent static E-induced Kerr (Fig. 16). If the 44 GHz Δα amplitude is the same at all three wavelengths within error, the artifact is ruled out; if it varies with λ (especially if it changes sign where the static E-induced Kerr changes sign), the observed signal is dominated by Δγ and the DAQ claim fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The measured quantity is the pump-induced, E-linear Kerr slope S(t) = dΔθ_K/dE. Writing θ_K = M_z/γ + θ_AFM and keeping only the linear-in-E part, S(t) = Δα(t)/γ − α Δγ(t)/γ² + dΔθ_AFM/dE, where α is the static magnetoelectric coefficient and Δγ(t) is the pump-induced change of the Kerr conversion factor. The AFM-Kerr term vanishes by the authors' PT argument (SI II.3), and an E-induced change of the out-of-phase magnon amplitude cannot appear in the regular Kerr channel because that mode has M_z = 0. But the second term is not forbidden: the out-of-phase magnon modulates L_z, and γ can depend on L_z, so Δγ(t) oscillates at 44 GHz. With M_z = αE, this yields a spurious Δα_spurious = −α Δγ/γ. The claimed Δα amplitude is 12% of α at B∥ = 6 T, so a 12% modulation of γ at 44 GHz would fully explain the observation. Such a modulation is plausible because the same magnon is argued to modulate the Berry curvature, and therefore the off-diagonal optical conductivity at 515 nm; the static E-induced Kerr itself varies by more than a factor of two and changes sign across 500–1000 nm (Fig. 16), showing γ is highly dispersive. The static wavelength, Kerr/Faraday and 7-of-8 gate checks (SI I.4, II.4) establish that the static E-linear Kerr measures E-induced M_z, but they do not constrain the time-dependent γ at the magnon frequency. Hence the central identification of a coherent θ(t) oscillation is not yet established.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports time-resolved measurements of the magnetoelectric coefficient α in 6-layer MnBi2Te4 using a dual-gated device combined with ultrafast pump-probe Kerr spectroscopy. The authors observe a coherent oscillation of the pump-induced E-linear Kerr slope at ~44 GHz with amplitude about 12% of the static α at B∥=6 T, matching the out-of-phase antiferromagnetic magnon frequency. They attribute this to a coherent oscillation of the axion angle θ(t) (the dynamical axion quasiparticle, DAQ), and support the interpretation with static magnetoelectric characterization, symmetry-based separation of AFM and Mz Kerr contributions, gate dependence checks, temperature and field dependence, DFT calculations of the Berry-curvature dipole, and a comparison with layer-Hall transport. The paper also extrapolates the DAQ to a dark-matter axion detection scheme in the meV regime.","tokens_in":39008,"tokens_out":11463,"duration_ms":127228,"significance":"If correct, this would be the first time-domain observation of a dynamical axion quasiparticle in a condensed-matter system, with implications for ultrafast control of Berry curvature and for axion detection proposals. The paper is commendable for its extensive cross-checks: DC and AC E-field extraction give consistent results (SI I.4.2); the 44 GHz mode tracks the out-of-phase magnon frequency versus B∥ and temperature (Fig. 3f,g; SI I.4.5); spatial reproducibility is shown (SI I.4.4); and the static E-linear Kerr signal is demonstrated to arise from E-induced magnetization through simultaneous Kerr/Faraday measurements (SI II.4). The comparison of dα/dn with the layer-Hall dD/dn (Extended Data Fig. 3) is an independent check on the Berry-curvature origin. The dark-matter sensitivity analysis is clearly an extrapolation. The main weakness is that the time-resolved E-linear Kerr slope could in principle reflect pump-induced modulation of the Kerr conversion factor rather than of α itself; this is the primary reason the present version does not fully establish the central claim.","major_comments":[{"comment":"The central identification of the 44 GHz oscillation with a coherent α(t) oscillation is not yet established because the measurement does not exclude a pump-induced modulation of the Kerr conversion factor γ. With θ_K = M_z/γ + θ_AFM and M_z = αE, the E-linear slope extracted by the double lock-in is S(t) = dΔθ_K/dE = Δα(t)/γ − α Δγ(t)/γ² + dΔθ_AFM/dE. The PT argument in SI II.3 eliminates the AFM term, but the Δγ term is allowed: the out-of-phase magnon modulates L_z, and γ can depend on L_z through the same Berry-curvature/σ_xy changes that the paper itself argues produce the axion response. Since the static E-induced Kerr is strongly wavelength dispersive (SI Figs. 16 and 40), a ~12% modulation of γ at 44 GHz would fully account for the claimed Δα amplitude. The DC-versus-AC comparison (SI I.4.2) and the 7-of-8 gate checks (SI I.4.1) do not distinguish Δα from Δγ, because both procedures extract dΔθ_K/dE. A wavelength-dependent measurement of the 44 GHz Δα (α should be wavelength-independent, whereas γ(λ) is not), or a time-resolved Faraday measurement, is required to rule out this alternative.","section":"Fig. 3c; Methods.4; SI II.3-II.4"},{"comment":"The absolute calibration of α in units of e²/2h relies on the assumption that the Kerr-to-magnetization conversion factor γ is identical in the spin-flop state at B⊥=6 T and in the antiferromagnetic state at B⊥=0 T, an approximation the authors acknowledge but do not quantify. Because the claimed static α and its comparison with first-principles calculations (Fig. 4b,c) and the layer-Hall dα/dn comparison (Extended Data Fig. 3) all inherit this calibration, the absence of an uncertainty estimate weakens the quantitative support for α=D and for the 'large DAQ' claim. The ratio Δα/α is independent of γ, so the qualitative observation is unaffected; nevertheless, the authors should provide an error estimate or an independent in-situ calibration of γ.","section":"Methods.3; Extended Data Fig. 4"},{"comment":"The symmetry argument that the AFM Kerr effect cannot have a linear-in-E component is applied to a pump-probe experiment, where the pump pulse itself breaks PT symmetry. The authors should justify that the pump acts as a scalar perturbation (e.g., heating), consistent with the pump-polarization independence shown in Extended Data Fig. 2d, before using this argument to exclude linear-in-E AFM-Kerr contributions to the time-resolved signal. Without this justification, the conclusion that the E-linear slope contains only regular-Kerr terms is incomplete.","section":"SI II.3"}],"minor_comments":[{"comment":"The phrase 'Wilczek and Weinberg theoretically discovered a new boson' should be 'predicted a new boson', since the axion has not been experimentally discovered.","section":"Abstract"},{"comment":"The sentence 'our E-field induced Kerr rotation measures the E-field induced Mz black' contains a stray word 'black' and should be corrected.","section":"Main text, Static θ measurements"},{"comment":"Two panels in the Fig. 2 caption are labeled 'b'; the second 'b' should be 'c' to match the panels.","section":"Fig. 2 caption"},{"comment":"The FFT inset in Fig. 3c lacks axis labels and units; the frequency axis should be labeled with GHz, and the amplitude axis with the corresponding units.","section":"Fig. 3c inset"},{"comment":"The dark-matter sensitivity estimate depends on several parameters not yet demonstrated (THz single-photon detector efficiency, sample area 0.16 m², thickness 0.4 mm, loss parameters Γm and Γρ). The text should more clearly separate this projection from the experimental results, perhaps by marking it as a proposal rather than a measured sensitivity.","section":"Methods.6; Fig. 4g"},{"comment":"The 7-of-8 null-result summary would be more convincing if the magnitude and noise floor of the null results were stated (e.g., an upper bound on the unobserved n/E dependences), so that the reader can assess whether the null results are statistically meaningful.","section":"SI I.4.1; Extended Data Fig. 45"}],"recommendation":"major_revision","confidential_remarks":"The central observation is potentially very important, but the γ-modulation alternative is not yet excluded and is, if anything, suggested by the authors' own Berry-curvature modulation mechanism. The required additional control experiments (wavelength dependence of the 44 GHz signal or time-resolved Faraday) are feasible and within the scope of a revision. The dark-matter sensitivity section is speculative and could be condensed; it is not the basis of my recommendation. The paper's numerous cross-checks and independent transport comparison are strengths that should be acknowledged in any decision letter."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, the important thing: this is a real step forward experimentally. Combining a dual-gated 6-layer MnBi2Te4 device with two lock-ins to measure the time-resolved slope dΔKerr/dE is clever, and the paper is full of careful cross-checks—DC vs AC E-field, temperature, field, spatial reproducibility, the Kerr/Faraday symmetry separation, the 7-of-8 null results, and the independent layer-Hall comparison. If the interpretation is right, this is the first time-domain observation of a dynamical axion quasiparticle, and the technique will be useful beyond this material.\n\nBut the stress-test concern is real, and it lands on the central claim. The measured quantity is the E-linear part of the pump-induced Kerr slope. The out-of-phase magnon modulates Lz, and the Kerr conversion factor γ (the factor turning Mz into Kerr rotation) very plausibly depends on Lz—the static wavelength dependence shows γ changes sign and factor-of-two across 500–1000 nm. So a 12% modulation of γ at 44 GHz would produce exactly the observed 12% \"Δα\". The static PT-symmetry argument that the AFM Kerr has no linear-in-E part applies to the equilibrium state, not to the pump-excited, time-dependent one. The static Kerr/Faraday proof establishes that the linear-in-E Kerr is due to Mz, but it says nothing about whether γ itself oscillates. The 7-of-8 checks are consistent with the DAQ interpretation, but they don't exclude γ modulation, because the in-phase mode doesn't modulate Lz and thus wouldn't modulate γ.\n\nThe decisive missing control is a time-resolved Faraday measurement (or at least the wavelength dependence of the Δα(t) at the magnon frequency). If the E-linear signal appears in both Kerr and Faraday at 44 GHz, that would nail Mz(t). If it's Kerr-only, it's a γ artifact.\n\nOther, smaller issues: the absolute α calibration relies on assuming the same γ in the spin-flop state at 6 T and the AFM state at 0 T—explicitly approximate—and there are no error bars on the key figures. The dark-matter sensitivity projection is clearly flagged as an extrapolation and depends on many assumed parameters; treat it as a roadmap, not a result. And there are citation-numbering errors in the reference list that need fixing.\n\nWho's it for: researchers in topological magnetism, ultrafast optics, and axion physics. It deserves serious refereeing; I'd accept it provisionally and send it to referees with a strong request for the time-resolved Faraday or multi-wavelength control.","headline":"A real experimental step towards time-resolved magnetoelectric detection, but the 44 GHz E-linear Kerr signal could be a Kerr-conversion modulation rather than a true α(t) oscillation.","tokens_in":39704,"tokens_out":6070,"would_cite":true,"duration_ms":62760,"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":"This paper reports the first real-time observation of the dynamical axion quasiparticle: a ~44 GHz oscillation of the magnetoelectric coefficient (∝θ) in six-layer MnBi2Te4, driven by the out-of-phase antiferromagnetic magnon.","keywords":["dynamical axion quasiparticle","magnetoelectric coupling","MnBi2Te4","antiferromagnetic magnon","Berry curvature","ultrafast pump-probe spectroscopy","time-resolved Kerr rotation","dark matter axion detection"],"falsifier":"Perform a time-resolved Faraday rotation measurement on the same six-layer device while modulating $E_z$: because Faraday rotation is sensitive to out-of-plane magnetization $M_z$ but not to the antiferromagnetic order $L_z$, a 44 GHz oscillation in the E-linear Faraday signal would confirm $M(t) = \\alpha(t)E$; if the Faraday signal stays flat while the Kerr signal oscillates at 44 GHz, the observed $\\Delta\\alpha(t)$ must instead originate from a Kerr-only artifact such as a dynamical conversion factor or an AFM-Kerr background.","tokens_in":38404,"feed_emoji":"🧲","tokens_out":14099,"duration_ms":118519,"temperature":0.7,"pith_summary":"The paper reports the first real-time observation of a dynamical axion quasiparticle (DAQ): a coherent oscillation of the axion angle $\\theta(t)$ in a condensed-matter system, directly analogous to the high-energy axion particle. The material is six-layer MnBi2Te4, where the magnetoelectric coefficient $\\alpha(t) \\propto \\theta(t)$ is measured with femtosecond time resolution by combining a dual-gated device with pump–probe Kerr rotation and double lock-in detection. A coherent oscillation of $\\alpha(t)$ at $\\sim$44 GHz is observed, with amplitude about 12% of the static magnetoelectric coefficient, and its frequency tracks the out-of-phase antiferromagnetic magnon mode. The authors show that this oscillation arises because the magnon coherently modulates the Berry-curvature real-space dipole that constitutes $\\alpha$. If correct, the DAQ becomes a measurable, tunable object with proposed applications in ultrafast spintronics and in dark-matter axion detection in the meV regime.","feed_headline":"Axion quasiparticle seen oscillating in a 2D magnet at 44 GHz","feed_subtitle":"In six-layer MnBi2Te4, the magnetoelectric coefficient α(t) swings by 12% as the out-of-phase magnon precesses.","key_machinery":"The load-bearing object is the time-resolved magnetoelectric coefficient $\\alpha(t)$, extracted from the slope of pump-induced Kerr rotation versus the applied out-of-plane electric field $E_z$. The experimental machinery is a double lock-in scheme: the pump is chopped at 1 kHz while $E_z$ is AC-modulated at 0.7 Hz, and the product-frequency signal isolates the E-linear magnetic response; a separately determined conversion factor $\\gamma$ maps Kerr angle to magnetization. The second central piece is the out-of-phase antiferromagnetic magnon, the one mode that couples to $\\theta$: its spin precession modulates the top and bottom surface Berry curvatures $\\Omega_T$ and $\\Omega_B$ in opposite phase, so the Berry-curvature real-space dipole $D = \\frac{e^2}{4\\pi h}\\int_k(\\Omega_T - \\Omega_B)$ oscillates, and because $\\alpha = D$, this constitutes the dynamical axion quasiparticle. The identity $\\theta = \\pi(2h/e^2)\\alpha$ ties the optical measurement to the axion angle.","core_discovery":"The paper claims the direct observation of the dynamical axion quasiparticle in a six-layer MnBi2Te4 device. Since $\\theta \\propto \\alpha$ (specifically $\\theta = \\pi (2h/e^2)\\alpha$), an oscillation of the axion angle is detected as an oscillation of the magnetoelectric coefficient. A pump pulse launches coherent magnons; a probe pulse, combined with an AC-modulated gate electric field, yields the pump-induced change in the E-linear Kerr slope, $\\Delta\\alpha(t)$, at femtosecond time resolution. At $B_{\\parallel}=6$ T the data show a coherent oscillation of $\\Delta\\alpha(t)$ at $\\sim$44 GHz with amplitude about $0.05\\,e^2/2h$, roughly 12% of the static $\\alpha$; the FFT peak matches the independently measured out-of-phase antiferromagnetic magnon mode, and the oscillation follows that mode's dependence on $B_{\\parallel}$ and temperature. First-principles calculations indicate that the out-of-phase magnon makes the top and bottom surface Berry curvatures $\\Omega_T$ and $\\Omega_B$ oscillate in opposite phase, so their difference—the Berry-curvature real-space dipole $D$, which equals $\\alpha$—oscillates in time, whereas the in-phase magnon leaves $D$ unchanged. The paper further proposes that this material, with DAQ mass $\\sim$44 GHz ($\\sim$0.18 meV), could serve as a magnetically tunable dark-matter axion detector in the meV regime.","pith_inferences":["The same double lock-in measurement applied to other antiferromagnetic topological insulators (for example Mn2Bi2Te5 or heterostructures with different hybridization gaps) would test whether the DAQ amplitude follows the predicted $\\delta\\theta/\\delta L_z$ trend with thickness—a direct check of the paper's Berry-curvature mechanism.","A time-resolved Faraday rotation experiment would settle whether the observed 44 GHz E-linear signal is truly magnetoelectric: Faraday rotation is blind to the antiferromagnetic order $L_z$ but sensitive to $M_z$, so an oscillating E-linear Faraday signal would confirm $M(t) = \\alpha(t)E$, whereas its absence would point to a Kerr-only artifact.","The claimed dark-matter sensitivity extrapolates from a microscopic six-layer device to a macroscopic detector of area 0.16 square meters and thickness 0.4 millimeters; a testable intermediate step is measuring the DAQ resonance in a thick superlattice or bulk Mn2Bi2Te5 crystal at the same 44 GHz frequency to verify that the resonant enhancement survives scaling."],"forward_implications":["The DAQ provides a tabletop, tunable realization of axion electrodynamics: $\\theta(t)$ oscillates coherently at $\\sim$44 GHz with an amplitude set by the magnetoelectric coupling, and the frequency is tunable by the in-plane magnetic field.","A DC electric field applied to the device produces a time-varying magnetization $M(t) = \\alpha(t)E$ at the magnon frequency, enabling electric control of ultrafast spin polarization—a route toward coherent antiferromagnetic spintronics.","The measured DAQ mass of $\\sim$44 GHz ($\\sim$0.18 meV) and its magnetic-field tunability set the operating frequency of a proposed dark-matter axion detector in the meV regime, with estimated sensitivity crossing below the astrophysical bound and, in part, reaching the QCD axion band.","The demonstration that magnons coherently modulate the Berry-curvature real-space dipole establishes a general mechanism for ultrafast manipulation of quantum geometry, extendable to the quantum metric and to magnetic Weyl semimetals where the spin direction controls Weyl-node positions.","The double lock-in technique for time-resolved magnetoelectric measurement is itself a new probe for ultrafast axion and magnetoelectric dynamics in other PT-symmetric antiferromagnets."],"supporting_citations":[{"why":"Defines the axion as the coherent oscillation of the θ field, giving the high-energy phenomenon that the condensed-matter quasiparticle is meant to simulate.","marker":"[1, 2]"},{"why":"Establishes the magnetoelectric polarizability and axion electrodynamics description of crystalline insulators, grounding the identification θ ∝ α.","marker":"[3]"},{"why":"Cited as the initial theoretical prediction that a coherent oscillation of θ in a magnetic insulator constitutes a dynamical axion quasiparticle.","marker":"[13]"},{"why":"Established the theory that a dynamical axion quasiparticle can resonantly convert dark-matter axions into THz photons, the detector scheme the paper extends with its measured parameters.","marker":"[8, 21, 23]"},{"why":"Provided the experimental method (WSe2-stacked pump–probe Kerr) the paper follows to identify the in-phase and out-of-phase magnon frequencies.","marker":"[7]"},{"why":"Reported the layer Hall effect in even-layer MnBi2Te4, which the paper uses as an independent transport probe that the magnetoelectric coefficient equals the Berry-curvature real-space dipole D.","marker":"[25]"}],"fun_headline_variants":["Axion quasiparticle's 44-GHz oscillation captured in 2D magnet","2D MnBi2Te4 reveals axion quasiparticle oscillating at 44 GHz","Dynamical axion quasiparticle directly observed in six-layer magnet","Coherent oscillation of axion angle seen in 2D antiferromagnet","Magnon-induced axion quasiparticle oscillates at 44 GHz"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the double lock-in signal assigned to $\\Delta\\alpha(t)$ is truly the pump-induced change of the magnetoelectric coefficient—that the only E-linear contribution to the pump-induced Kerr rotation at the out-of-phase magnon frequency is the electric-field-induced magnetization through $\\alpha(t)$, and not an electric-field-induced modulation of the magnon amplitude, of the Kerr conversion factor, or of the antiferromagnetic Kerr background.","fun_headline_variants_meta":{"raw":{"variants":["Axion quasiparticle's 44-GHz oscillation captured in 2D magnet","2D MnBi2Te4 reveals axion quasiparticle oscillating at 44 GHz","Dynamical axion quasiparticle directly observed in six-layer magnet","Coherent oscillation of axion angle seen in 2D antiferromagnet","Magnon-induced axion quasiparticle oscillates at 44 GHz"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000935,"raw_usage":{"total_tokens":4155,"prompt_tokens":1253,"completion_tokens":2902,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":869,"completion_tokens_details":{"reasoning_tokens":2796}},"tokens_in":869,"tokens_out":2902,"duration_ms":21291,"temperature":1.0,"reasoning_tokens":2796,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T12:28:54.274713+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Perform a time-resolved Faraday rotation measurement on the same six-layer device while modulating $E_z$: because Faraday rotation is sensitive to out-of-plane magnetization $M_z$ but not to the antiferromagnetic order $L_z$, a 44 GHz oscillation in the E-linear Faraday signal would confirm $M(t) = \\alpha(t)E$; if the Faraday signal stays flat while the Kerr signal oscillates at 44 GHz, the observed $\\Delta\\alpha(t)$ must instead originate from a Kerr-only artifact such as a dynamical conversion factor or an AFM-Kerr background.","supporting_citations":[{"cited_title":"As shown in Fig","cited_arxiv_id":null,"evidence_quote":"Establishes the magnetoelectric polarizability and axion electrodynamics description of crystalline insulators, grounding the identification θ ∝ α."},{"cited_title":"& Zhang, S.-C","cited_arxiv_id":null,"evidence_quote":"Cited as the initial theoretical prediction that a coherent oscillation of θ in a magnetic insulator constitutes a dynamical axion quasiparticle."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provided the experimental method (WSe2-stacked pump–probe Kerr) the paper follows to identify the in-phase and out-of-phase magnon frequencies."},{"cited_title":"We estimate our sensitivity based on a THz single-photon detector with efficiency∼ 95% and dark-count rate of 10−5 Hz","cited_arxiv_id":null,"evidence_quote":"Reported the layer Hall effect in even-layer MnBi2Te4, which the paper uses as an independent transport probe that the magnetoelectric coefficient equals the Berry-curvature real-space dipole D."}],"review_version":1}