{"id":"28f7c248-1286-4bae-8d09-f02ec809e6c1","arxiv_id":"2505.17007","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"high","formal_verification":"none","parameter_count":6,"one_line_summary":"A proposed axion haloscope using shift-current difference frequency generation in the Weyl semimetal TaAs could probe QCD axions at 10-100 meV, assuming a 10^8 V/m, 100 ps THz drive field is achievable.","lead":"This paper proposes detecting axion dark matter by using a crystal called TaAs to turn a faint axion-induced electric field plus a strong applied laser field into a measurable current. If the required strong, sustained laser field can be built, the method could search a mass range most current experiments cannot reach.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed reach hinges on E_exp = 10^8 V/m at ~100 meV being effectively monochromatic over the full integration time; the paper concedes only short-pulse sources exist and never addresses duty cycle or spectral density, so Eq. (14) may overstate SNR by many orders of magnitude.","rationale":"The reader's weakest assumption identifies the same load-bearing premise: the experimental drive field E_exp = 10^8 V/m with ≥100 ps pulses and sufficient duty cycle. I agree with that identification, and I emphasize that the problem is not merely pulse duration but the combination of spectral density and live integration time. A short pulse of the same peak field deposits much less power in the 10 GHz resolution bin, and the duty cycle shortens the effective integration time; both effects enter Eq. (14) multiplicatively and can destroy the claimed QCD-axion reach. The paper itself flags the pulse-duration requirement in Section IV, so this is a genuine acknowledged gap rather than a manufactured objection. I considered other possible concerns, including the extrapolation of the measured single-frequency shift-current conductivity to the two-frequency cross term and the spatial coherence of the axion field over a 4x4 cm sample. These are real but secondary: a conductivity error of even one order of magnitude changes the required coupling by only sqrt(10), while a short-pulse duty-cycle deficit of 10^-8 changes it by 10^4. The theoretical derivation and the use of TaAs symmetry properties are not the weak point; the quantitative sensitivity projection is. The reader's CONDITIONAL verdict remains appropriate, hence UNCHANGED.","tokens_in":29760,"tokens_out":23626,"duration_ms":206485,"concrete_test":"Take the source parameters from the cited experiment [56] (pulse duration τ_p, repetition rate f_rep, peak field E_pk) and recompute Eq. (14) with E_exp^2 replaced by E_pk^2 (τ_p Δfbin) and τscan replaced by f_rep τ_p τscan, where Δfbin = 10 GHz. If the resulting SNR at the QCD axion line (e.g., ma = 100 meV and gaγγ from Eq. (2)) falls below 1, the claimed reach in Fig. 3 is unsupported by the cited source. If the authors instead identify a CW or high-duty-cycle source with effective E_exp = 10^8 V/m in a 10 GHz bin over the full τscan, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The sensitivity in Eq. (13)-(14) treats E_exp = 10^8 V/m as a monochromatic amplitude available continuously for τscan = 1 day per frequency setting. Section IV explicitly concedes that this field strength is currently achievable only with short-pulse lasers, and that pulses of at least 100 ps are needed to reach the 10 GHz resolution bandwidth. The paper does not provide a source with that pulse duration, a repetition rate, or a duty cycle. This matters in two ways. First, a transform-limited pulse of duration τ_p has its spectral power spread over ~1/τ_p, so the power in one 10 GHz bin is suppressed by a factor ~τ_p × 10 GHz relative to a 100 ps pulse of the same peak field. Second, the live integration time is f_rep τ_p τscan, not τscan. For a 100 fs, 1 kHz source, the SNR in Eq. (14) is reduced by roughly (10^-3) × sqrt(10^-10) = 10^-8 relative to the CW assumption. Since the required coupling scales as 1/sqrt(SNR), the SNR=1 contour in Fig. 3 would move upward by about four orders of magnitude in gaγγ, far above the QCD axion band. The central claim therefore rests on an unverified and self-admittedly unresolved experimental premise.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a new axion dark matter detection scheme based on the shift current, a second-order nonlinear optical response in non-centrosymmetric materials. The axion under a static magnetic field produces an oscillating electric field E_DM; applying a strong oscillating field E_exp at frequencies in the 10–100 meV range generates a shift-current cross-term at the difference frequency, which can be read out at low frequency. Using literature values for the shift conductivity of TaAs, the authors derive a sensitivity estimate (Eqs. (13)–(14), Fig. 3) that claims to reach the QCD axion band for masses 10–100 meV and couplings g_{aγγ} ~ 10^–12 to 10^–11 GeV^–1 with B0 = 1 T, E_exp = 10^8 V/m, T = 1 mK, and 100 days of observation. The appendices contain detailed derivations of the nonlinear response, and the broadband nature of the response is illustrated with a Rice-Mele toy model.","tokens_in":30055,"tokens_out":20630,"duration_ms":178611,"significance":"If the sensitivity estimate were correct, this would open a new solid-state route to axion dark matter detection in a mass range that is challenging for cavity haloscopes. The formal derivation of the shift-current response is detailed and self-consistent, and the idea of using a two-field cross-term to downconvert axion signals is original. The Rice-Mele example usefully demonstrates the broad frequency response. However, the quantitative claim depends on several experimental and electrodynamic assumptions that are not justified in the manuscript, and at least two of these can suppress the signal by many orders of magnitude.","major_comments":[{"comment":"The sensitivity estimate assumes that the cross-term current J ∝ σ E_DM E_exp integrates coherently over the sample volume, yielding a total current scaling as L_exp^2 and power as L_exp^4. The axion field is spatially uniform over the de Broglie wavelength (~1 cm), but the experimental field at ω_exp ~ 100 meV has a vacuum wavelength of ~12 µm. The product E_DM E_exp therefore acquires a spatial phase e^{i k_exp·r} with period ~12 µm, so for a 1 cm sample the current cancels upon integration unless the sample is a thin film (thickness ≪ 12 µm) in the propagation direction. If such a thin film is intended, the current-collection cross-section is then reduced by a factor of order λ_exp/L_exp relative to the L_exp^2 used in Eq. (12), suppressing the signal power by ~10^-8. The manuscript does not specify a geometry that avoids this cancellation, nor does it discuss phase matching; as written, Eq. (13) appears to overestimate the signal by many orders of magnitude.","section":"Section III, Eqs. (12)–(14)"},{"comment":"The SNR calculation assumes the applied field E_exp = 10^8 V/m is available continuously for τscan = 1 day per frequency setting. The manuscript concedes in Section IV that such fields are currently achievable only with short-pulse lasers and that pulses of at least 100 ps are needed for the 10 GHz resolution bandwidth, but it does not identify a source with that pulse duration and a high repetition rate, nor does it evaluate the duty cycle. For a pulsed source with duty cycle D, the time-averaged signal power entering the radiometer equation is reduced by D (for a periodic pulse train whose spectrum overlaps the signal bin), and the required coupling scales as D^{-1/2}. Even for a duty cycle of 10^-5, the reach in g_{aγγ} degrades by a factor of ~300, moving the SNR = 1 contour above the QCD axion band. The paper does not quantify the pulsed-source penalty at all.","section":"Section IV and Eq. (14)"},{"comment":"The value σ_xzx = 200 µA/V^2 is taken from single-frequency shift-current measurements in Ref. [47] and is assumed without justification to apply to the two-frequency cross-term σ_xzx(Δω; m_DM, ω_exp) in Eq. (11) over the entire 10–100 meV range. The sensitivity scales as σ^2, so a factor-of-10 error in σ changes the reach by a factor of 10 in g_{aγγ}. The Rice-Mele demonstration in Fig. 1 shows that the conductivity has significant frequency dependence, especially near band edges, so applying a single measured value to the full mass range requires either a frequency-dependent calculation for TaAs or a clear argument for why the approximation is valid.","section":"Section II, conductivity assumption"},{"comment":"The axion-induced electric field E_DM in Eq. (4) is the free-space expression. The manuscript does not address how this field is modified inside the TaAs sample, which is a semimetal with free carriers and a complex dielectric response at THz frequencies. The internal field may be screened or attenuated depending on the conductivity, permittivity, and sample boundary conditions. Since the signal power scales as E_DM^2, an order-of-magnitude reduction in the internal field would shift the reach by a factor of 10 in g_{aγγ}. The authors should model the axion-to-photon conversion in the presence of the detector material or justify why free-space boundary conditions apply.","section":"Section II, Eq. (4)"}],"minor_comments":[{"comment":"The sentence 'we adopt the ferroelectric material as a sample' is inconsistent with the rest of the paper, which uses TaAs (a Weyl semimetal), and appears to be a leftover from an earlier version.","section":"Section IV"},{"comment":"The caption states 'area Lexp × Lexp ∼ 4 × 4 cm2', but the text in Section III uses L_exp = 1 cm; this discrepancy should be resolved, since Eq. (13) depends sensitively on L_exp^4.","section":"Fig. 3 caption"},{"comment":"The symbol σ is used both for the shift-current conductivity and for the signal linewidth σ_sig; this dual use may confuse readers and should be disambiguated.","section":"Section II"},{"comment":"The relationship between the coherent-segment duration τ = 0.1 ns and the total integration time τscan = 1 day in the radiometer equation is not fully explained; a sentence describing how the Fourier-transformed segments are combined would improve clarity.","section":"Section III"}],"recommendation":"reject","confidential_remarks":"The paper's central claim is not supported by the sensitivity estimate as written. Two load-bearing issues, the spatial phase cancellation of the pump field over the sample and the pulsed nature of available 10^8 V/m sources, each suppress the signal by several orders of magnitude and are not addressed in the manuscript. The formal derivation of the shift-current response is solid and could be of interest to a specialized audience, but the proposed detector concept and its quantitative reach would require a substantially different design to be viable. I recommend rejection in its current form."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a proposal paper, and the proposal is genuinely new: using shift-current difference frequency generation in a non-centrosymmetric crystal as an axion haloscope, with an applied pump field to amplify the axion-induced field and downconvert the signal to a measurable band. The background suppression via σ_xxx = 0 by C4v symmetry is a neat and concrete trick. The analytic derivation of the shift-current response in the appendices is careful and matches the standard form; the Rice-Mele toy model in Fig. 1 illustrates why the response is broadband.\n\nThe soft spot is not the derivation, it's the pump. The sensitivity in Eqs. (12)–(14) treats E_exp = 10^8 V/m as a monochromatic continuous field over τ_scan = 1 day per frequency setting. Section IV concedes that such fields today come from short-pulse lasers and that at least 100 ps pulses are needed to reach the 10 GHz resolution bandwidth. That is not a technical detail; it is load-bearing. With a 100 fs, 1 kHz source—typical of the cited technology—the spectral power in a 10 GHz bin is down by ~10^-3, and the live integration time is down by ~10^-10, so the SNR is suppressed by ~10^-8 relative to the CW assumption. That moves the SNR=1 contour in Fig. 3 upward by about four orders of magnitude in g_aγγ, far above the QCD axion band. The paper essentially concedes this by saying 'we still need a long enough pulse'. Until a source with high peak field and long pulse or high repetition rate exists, the claimed reach is a future projection, not a current capability.\n\nSecond, the conductivity extrapolation is optimistic: a measured single-frequency σ_xzx ≈ 200 µA/V^2 at ~100 meV is used for the two-field cross-term over the whole 25–350 meV band. The cross-term can have different frequency and temperature dependence; the paper does not compute it from band structure, only assumes it. Worth a caveat, but not fatal because the order of magnitude is plausible.\n\nThird, a smaller point: the axion-induced electric field inside a material is not the free-space value; boundary conditions and dielectric screening should be discussed. Probably a minor correction.\n\nWho is this for? People thinking about new axion detection schemes, particularly at the 10–100 meV gap where cavities fail. It is a theory proposal with a clean idea and an honest acknowledgment of the main obstacle. It deserves peer review, but the referee should require the sensitivity estimate to be recast with duty cycle and spectral density of realistic sources, or clearly labeled as a target for source development. I would not cite the numerical reach as is; I would cite the concept.","headline":"A genuinely new axion detection idea with an elegant background suppression, but the claimed reach depends on a pulsed 10^8 V/m source whose duty cycle is never addressed—so the sensitivity forecast is not yet supported.","tokens_in":30628,"tokens_out":3118,"would_cite":true,"duration_ms":26563,"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 Weyl-semimetal shift current could detect QCD axion dark matter in the 10-100 meV mass window.","keywords":["axion dark matter","shift current","Weyl semimetal","difference frequency generation","nonlinear optical response","QCD axion","bulk photovoltaic effect","dark matter detection"],"falsifier":"Measure the shift-current conductivity $\\sigma^{xzx}_{\\rm shift}$ of TaAs with two input fields whose frequencies differ by $\\Delta\\omega\\lesssim1$ THz across the 25-350 meV range; if the conductivity is far below the assumed $\\sim200\\,\\mu$A/V$^2$ or the response is not flat in $\\Delta\\omega$, the SNR=1 curve in Fig. 3 moves to larger $g_{a\\gamma\\gamma}$ and may no longer cover the QCD axion band.","tokens_in":29518,"feed_emoji":"⚛️","tokens_out":6468,"duration_ms":57776,"temperature":0.7,"pith_summary":"The paper proposes using the shift current, a nonlinear photocurrent generated in crystals without inversion symmetry, to detect axion dark matter. A static magnetic field converts axions into a feeble electric field, and a strong applied oscillating electric field mixes with it to produce a current at a low difference frequency that is easy to read out. Using the Weyl semimetal TaAs, the authors estimate that QCD axions with masses of roughly 10-100 meV and photon couplings of order $10^{-12}$ to $10^{-11}$ GeV$^{-1}$ could be seen at signal-to-noise ratio one with a 1 T magnet, a $10^8$ V/m drive field, a 1 mK receiver, and 100 days of observation. If correct, this would open a new broad-band laboratory window on axion dark matter without requiring a large resonant cavity.","feed_headline":"Weyl-semimetal shift current could hear QCD axions at 10-100 meV","feed_subtitle":"A 1-tesla magnet, a strong laser field, and a TaAs crystal could cover the QCD axion window in 100 days.","key_machinery":"The load-bearing object is the shift current: a non-dissipative second-order photocurrent arising from the quantum-geometric shift of electron wave functions during optical transitions in crystals lacking inversion symmetry. The conductivity formula, Eqs. (5) and (11), is built from Berry connections and band energies, and the broadband operation comes from the momentum-space integral that keeps the response finite for many input frequencies. In the proposed detector, the cross term $\\sigma^{xzx}_{\\rm shift}E_{\\rm DM}E_{\\rm exp}$ is selected while the symmetry of TaAs forbids the background term $\\sigma^{xxx}E_{\\rm exp}E_{\\rm exp}$, leaving Johnson-Nyquist noise as the main limitation.","core_discovery":"The central claim is that the axion-induced electric field $E_{\\rm DM}=g_{a\\gamma\\gamma}B_0\\sqrt{2\\rho_{\\rm DM}}/m_{\\rm DM}$, obtained by applying a static magnetic field $B_0$, can be amplified and down-converted by mixing it with a strong applied field $E_{\\rm exp}$ inside a noncentrosymmetric crystal. The shift-current conductivity $\\sigma_{\\rm shift}$ generates an output current $J=\\sigma_{\\rm shift}E_{\\rm DM}E_{\\rm exp}$ at the difference frequency $\\Delta\\omega = \\omega_{\\rm exp}-m_{\\rm DM}$. Because the shift current is a non-dissipative second-order response whose resonance condition is automatically met by integrating over momentum, the signal remains strong over a broad frequency range; a single setting of $\\omega_{\\rm exp}$ covers about a 1 THz window. Using the measured shift-current conductivity of TaAs, the paper derives a sensitivity estimate that reaches the QCD axion band in the 10-100 meV mass range.","pith_inferences":["The quoted reach assumes effectively continuous integration; if the $10^8$ V/m field can only be delivered in low-duty-cycle short pulses, the effective observation time drops and the SNR, which scales as $\\sqrt{\\tau_{\\rm scan}}$, would push the sensitivity curve to larger couplings.","A direct material check is available: measuring the shift-current conductivity of TaAs with two input fields separated by a small $\\Delta\\omega$ would test the prediction that the response is flat for $\\Delta\\omega\\ll\\omega_{\\rm exp}$, which is what makes the 1 THz instantaneous scan window possible.","The method could in principle be aimed at other feebly interacting high-frequency fields by choosing crystals whose band gaps match the source frequency, though the practical limitation is finding materials with large, non-dissipative second-order conductivities."],"forward_implications":["A single experimental setup with a modest 1 T magnet and a square-centimeter TaAs sample could scan roughly 10-300 meV of axion masses in about 100 days, without the frequency tuning needed in cavity haloscopes.","The signal appears at DC to about 1 THz, so readout can use standard low-frequency electronics rather than detectors operating at the axion frequency itself.","The background from the drive field is suppressed by crystal symmetry, so the search is limited mainly by thermal noise in the readout circuit.","The same difference-frequency mechanism could be applied to other noncentrosymmetric materials, and the paper notes that magnon and phonon shift currents might extend the method to smaller axion masses and to hidden-photon dark matter."],"supporting_citations":[{"why":"Supplies the QCD axion mass-coupling relation that defines the target parameter band.","marker":"[26]"},{"why":"Provides the classical-wave description of axion dark matter and the amplitude formula used for $E_{\\rm DM}$.","marker":"[40]"},{"why":"Establishes difference frequency generation in topological semimetals, the low-frequency readout mechanism.","marker":"[42]"},{"why":"Provides the measured TaAs shift-current conductivity $\\sigma^{xzx}_{\\rm shift}\\sim200\\,\\mu$A/V$^2$ used in the sensitivity estimate.","marker":"[47]"},{"why":"Gives the symmetry argument that makes $\\sigma^{xxx}=0$ in TaAs, eliminating the dominant background.","marker":"[46]"},{"why":"The radiometer equation used to convert the predicted signal power into a signal-to-noise ratio.","marker":"[50]"},{"why":"Cited as the experimental source reaching $10^8$ V/m oscillating electric fields, which sets the drive-field assumption.","marker":"[56]"},{"why":"Gives the axion-photon conversion in external fields that underlies the whole detection concept.","marker":"[17]"}],"fun_headline_variants":["Shift current in Weyl semimetal could reveal QCD axions","Broadband axion detection via shift current in TaAs","Weyl semimetal shift current: a new ear for axions","Axion search gets a shift-current boost in Weyl semimetals","TaAs shift current may cover QCD axion window"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The reach assumes that an oscillating electric field near $10^8$ V/m can be supplied with pulses at least 100 ps long and with enough duty cycle to accumulate about one day of integration per frequency step; the paper states that such a long-enough pulse is still required.","fun_headline_variants_meta":{"raw":{"variants":["Shift current in Weyl semimetal could reveal QCD axions","Broadband axion detection via shift current in TaAs","Weyl semimetal shift current: a new ear for axions","Axion search gets a shift-current boost in Weyl semimetals","TaAs shift current may cover QCD axion window"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000276,"raw_usage":{"total_tokens":1636,"prompt_tokens":927,"completion_tokens":709,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":543,"completion_tokens_details":{"reasoning_tokens":618}},"tokens_in":543,"tokens_out":709,"duration_ms":5868,"temperature":1.0,"reasoning_tokens":618,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T14:52:49.957139+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the shift-current conductivity $\\sigma^{xzx}_{\\rm shift}$ of TaAs with two input fields whose frequencies differ by $\\Delta\\omega\\lesssim1$ THz across the 25-350 meV range; if the conductivity is far below the assumed $\\sim200\\,\\mu$A/V$^2$ or the response is not flat in $\\Delta\\omega$, the SNR=1 curve in Fig. 3 moves to larger $g_{a\\gamma\\gamma}$ and may no longer cover the QCD axion band.","supporting_citations":[{"cited_title":"Ogawa, M","cited_arxiv_id":null,"evidence_quote":"Supplies the QCD axion mass-coupling relation that defines the target parameter band."},{"cited_title":"DFSZ-Type Axions and Where to Find Them","cited_arxiv_id":"2302.04667","evidence_quote":"Establishes difference frequency generation in topological semimetals, the low-frequency readout mechanism."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the measured TaAs shift-current conductivity $\\sigma^{xzx}_{\\rm shift}\\sim200\\,\\mu$A/V$^2$ used in the sensitivity estimate."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The radiometer equation used to convert the predicted signal power into a signal-to-noise ratio."},{"cited_title":"Supernova-scope for the Direct Search of Supernova Axions","cited_arxiv_id":"2008.03924","evidence_quote":"Cited as the experimental source reaching $10^8$ V/m oscillating electric fields, which sets the drive-field assumption."}],"review_version":1}