{"id":"d059eaa9-c806-4d4b-98ef-4f6d8579d979","arxiv_id":"2507.00977","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Terahertz spin-current dynamics in ferromagnet/heavy-metal stacks are unchanged when pump photon energy is varied from 1.5 to 3 eV, indicating ultrafast electron heating, not hot-electron injection, dominates.","lead":"Researchers tested whether changing laser color from red to ultraviolet affects the ultrafast spin currents that generate terahertz pulses, and found the shape and timing of the spin current stay the same across many magnetic materials. The result suggests these currents are driven mainly by rapid heating of electrons rather than by the specific excited electronic states, which simplifies how terahertz spintronic sources can be designed.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Amplitude-invariance claim is weakened by a normalization factor that may include spin-current generation efficiency, making the 'only slightly' result partly an input rather than an outcome.","rationale":"The paper's central experimental finding is a systematic null result: THz spin-current waveforms are nearly identical for 1.5 eV and 3 eV excitation across diverse F|HM systems. The waveform-dynamics part of this claim is robust to the normalization procedure because only a multiplicative amplitude factor is applied, not a time-scale or shape transformation. The reader's identification of the normalization factor as the weakest assumption is correct and is the most load-bearing concern: since the factor explicitly includes potential changes in spin-current generation efficiency (Section 3, Results), the amplitude comparison cannot by itself distinguish between 'intrinsically similar amplitudes' and 'amplitudes made similar by the normalization.' The mechanistic conclusion that ultrafast electron heating dominates is plausible but rests on estimates relegated to the Supplementary Materials, and those estimates are not independently verified. A concrete test that fixes the amplitude scale with independent optical measurements and detection calibration would settle whether the null result extends to amplitudes. Given these considerations, the CONDITIONAL verdict is appropriate; no adjustment is needed.","tokens_in":13710,"tokens_out":3022,"duration_ms":124755,"concrete_test":"Perform a parameter-free amplitude comparison: independently determine the absorbed pump fluence in the F and HM layers at 1.5 eV and 3 eV using spectroscopic ellipsometry, and calibrate the THz detection sensitivity at both wavelengths using a reference emitter (e.g., ZnTe or GaP) whose emission efficiency is wavelength-independent. Then compute the amplitude ratio of the raw 3 eV and 1.5 eV THz waveforms for Fe|Pt and γ-Fe2O3|Pt without allowing any free scaling factor. If the ratio deviates from the value expected from absorbed-fluence differences by more than the noise level, the amplitude claim needs revision; if it does not, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3 states that all 3 eV traces are multiplied by a normalization factor that explicitly accounts for (i) absorption differences, (ii) setup sensitivity, and (iii) potential changes in the efficiency of spin-current generation and injection. Since factor (iii) is exactly the quantity the amplitude comparison is meant to test, allowing it in the normalization permits the scaling to absorb any photon-energy-dependent amplitude difference. The dynamics comparison is unaffected because no time-scaling or waveform reshaping is applied, and the waveform overlap is convincing; however, the quantitative claim that amplitudes change only slightly is not independently evidenced. The paper reports, for instance, that the maghemite signal at 3 eV is roughly a factor of two smaller and attributes this to Pt absorptance, but with the free scaling factor this attribution cannot be falsified. No error bars or statistics are given for the normalization factors, so it remains unclear whether the chosen factors are determined by independent calibrations or chosen to make traces overlap. This is the single most load-bearing concern because the headline mechanistic conclusion about electron-heating dominance depends on the amplitude null result as much as on the dynamics null result.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports terahertz-emission spectroscopy measurements of laser-induced spin currents in a broad set of ferromagnet/heavy-metal (F/HM) heterostructures, including metallic ferromagnets (Fe, CoFeB), ferrimagnetic insulators (maghemite, YIG, GIG), a half-metal (magnetite), a rare-earth alloy (Tb30Fe70), and CoFeB|MgO|Pt tunnel-barrier samples. For each system, the authors compare the THz emission excited at 1.5 eV and 3 eV photon energies. They find that the time-domain waveforms are nearly identical at the two photon energies, and, after applying a normalization factor to the 3 eV traces, that the signal amplitudes differ only moderately. The paper interprets these observations as evidence that ultrafast laser-induced electron heating is the dominant driving force of THz spin-current generation, while highly excited primary photoelectrons play a minor role.","tokens_in":13979,"tokens_out":3323,"duration_ms":42414,"significance":"If the result holds, the paper establishes a useful experimental fact: the dynamics of THz spin currents in a wide range of F/HM stacks are robust against changes in pump photon energy, which strengthens the case for electron-heating-mediated generation and simplifies practical use of spintronic THz sources across wavelengths. The systematic material coverage is a clear strength, as it explicitly addresses previously proposed photon-energy-sensitive scenarios: band-gap excitation in oxides, resonant 4f excitation in rare-earth alloys, and energy-dependent tunneling through MgO barriers. The dynamics comparison is clean because the waveforms are compared without any time scaling or reshaping, so the null result for the temporal shape carries real evidential weight. The main weakness is that the amplitude comparison is partly circular because the normalization factor allowed for the 3 eV traces includes the very spin-current-generation efficiency that is under test; the paper needs to address this for the amplitude part of the central claim to be credible.","major_comments":[{"comment":"The amplitude claim that the TSC amplitude is 'only slightly' different at 3 eV is weakened by the normalization procedure. The text states that all 3 eV traces are multiplied by a normalization factor that accounts for '(iii) potential changes in the efficiency of spin-current generation and injection.' Since this efficiency is exactly the quantity the amplitude comparison is meant to test, allowing it into the normalization permits the scaling to absorb any photon-energy-dependent amplitude difference. For the amplitude part of the central claim to be falsifiable, the paper must either determine the factor independently (e.g., from measured absorptance and a separate calibration of the setup transfer function) and report its value with uncertainty, or explicitly restrict the claim to the dynamics null result. As written, the report that the maghemite signal is a factor of two smaller at 3 eV and attributed to Pt absorptance (Section 4.1) cannot be verified because the scaling factor is free and no error bars or statistics are given.","section":"Section 3 (Results)"},{"comment":"The quantitative estimates used to support the central interpretation—that the fraction of primary high-energy electrons is less than 1%, and that the number of spins generated by electron heating is 2-3 orders of magnitude larger than the number of primary excited carriers—are referenced to the Supplementary Materials but are not presented or derived in the manuscript text. These numbers are load-bearing because they are used to explain why primary-photoelectron contributions are below the noise level and why OISTR signatures are undetectable. The derivation should be included in the submission (or clearly stated in the main text with the key assumptions), otherwise the main text cannot be fully evaluated by a reader. If the Supplementary Material already contains these details, please confirm that it is part of the submitted version.","section":"Sections 4.4 and 5 (and Supplementary Materials)"}],"minor_comments":[{"comment":"The section numbering is inconsistent: after Section 4.1, the text introduces '3.3 Excitation of spin-polarized 4f-type electrons...' and '4.4 Energy-dependent tunneling...', which appear to be misnumbered and should be sequential (e.g., 4.2, 4.3, 4.4).","section":"Section numbering"},{"comment":"The caption of Figure 4 repeats the panel label '(a)' twice ('(a) THz-emission signal ... (a) THz signals for various MgO thicknesses'); the second should presumably be '(b)' or a different label to match the figure panels.","section":"Figure 4 caption"},{"comment":"The phrase 'all THz signals are normalized by the absorbed pump fluence' is potentially misleading because the additional normalization factor applied to the 3 eV traces also includes effects (i)-(iii); please clarify which normalization is applied first and how the absorbed pump fluence is determined for each photon energy.","section":"Section 3 (Results), normalization wording"},{"comment":"The abstract and conclusions state that amplitudes are 'only slightly' different, but for maghemite the difference is about a factor of two. Please specify whether 'slightly' refers to the amplitude after the normalization factor or to the raw measured amplitudes, and quantify the range of amplitude differences observed across all samples.","section":"Abstract and Conclusions"},{"comment":"References 15 and 47 are arXiv preprints; for a formal publication, please provide peer-reviewed versions if available, or clearly mark them as preprints.","section":"References 15 and 47"}],"recommendation":"major_revision","confidential_remarks":"The central experimental finding—that the THz spin-current dynamics are essentially identical for 1.5 eV and 3 eV excitation across the studied systems—is a solid and useful result. The circularity in the amplitude comparison is the main technical issue; it is fixable by reporting the normalization factors and their independent determination, or by reframing the paper's claim to focus on the dynamics. I would encourage the authors to provide a table of the normalization factors and their uncertainties, since this would directly address the concern. The paper is within the scope of a condensed-matter physics journal and would be of interest to the ultrafast spintronics community."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth a look, but read the normalization section before quoting the amplitude claim. The paper reports a systematic null result: for a wide range of F|HM systems (metals, insulators, half-metals, rare-earth alloys, tunnel barriers), switching pump photon energy from 1.5 to 3 eV leaves THz spin-current dynamics essentially identical, with only modest amplitude changes. The dynamics comparison is clean—no time scaling or waveform reshaping, just a constant multiplicative factor—and the waveform overlap across all samples is convincing. That part deserves serious attention.\n\nThe new contribution is real. Prior refs 44-46 already reported wavelength insensitivity for metallic FM|HM, but this extends it to ferrimagnetic insulators (maghemite, YIG, GIG), half-metals (magnetite), 4f-resonant TbFe, and CoFeB/MgO tunnel barriers. The design directly targets the three scenarios where photon-energy dependence was most plausible: across-bandgap excitation, 4f resonance, and defect-mediated tunneling. The null result in those cases is informative.\n\nThe soft spot is the amplitude comparison. The 3 eV traces are scaled by a factor that explicitly includes '(iii) potential changes in the efficiency of spin-current generation and injection'—the very quantity under test. With that freedom, the conclusion that amplitudes are 'only slightly' different is partly an input. The paper does report, e.g., a factor of two smaller maghemite signal and attributes it to Pt absorptance, but because the normalization factor is free per sample, that attribution cannot be falsified. No error bars or statistics accompany the factors. This does not kill the dynamics result, and the mechanistic conclusion is still plausible given the estimates that primary hot electrons are <1% of the population, but the amplitude-invariance claim needs independent calibration of absorption and setup response, with the residual treated as the measured amplitude effect.\n\nThe supporting estimates (spin transfer numbers, primary electron fraction) are in the Supplementary, so the main text leans on claims not shown. That is a minor structural issue; the estimates look reasonable.\n\nBottom line: the central null result for dynamics is probably correct and the material coverage is valuable. The amplitude claim is under-evidenced. A serious referee should ask for the normalization to be redone with independently measured absorption and setup sensitivity. This is a conditional accept in my view, not a reject. I'd bring it to our reading group to discuss the normalization, and I'd cite the dynamics result.","headline":"The dynamics null result is solid and broadens the wavelength-insensitivity claim to new material classes, but the amplitude part of the claim is weakened by a normalization factor that may absorb the very efficiency change under test.","tokens_in":14465,"tokens_out":1988,"would_cite":true,"duration_ms":23204,"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":"Photon energy does not change terahertz spin current dynamics across a wide range of magnetic stacks.","keywords":["terahertz spin currents","ultrafast spintronics","photon-energy dependence","electron heating","spin Seebeck effect","pyrospintronic effect","terahertz emission spectroscopy","ferrimagnetic insulators"],"falsifier":"Measure the absolute THz amplitude per absorbed photon at 1.5 eV and 3 eV with the detection transfer function calibrated independently for each wavelength. If the ratio at equal absorbed energy differs from unity by more than the experiment's ~1% noise floor, then at least part of the photon-energy dependence was being absorbed into the normalization factor, and the amplitude part of the claim fails. A second falsifier: observe any wavelength-dependent dynamics, such as an extra slow component when 3 eV excites 4f states in Tb30Fe70, in a higher signal-to-noise version of the same measurement.","tokens_in":13488,"feed_emoji":"⚡","tokens_out":5367,"duration_ms":57898,"temperature":0.7,"pith_summary":"The paper sets out to settle whether the photon energy of the excitation pulse controls the initial step of terahertz spin current generation in ferromagnet/heavy-metal stacks. By comparing 1.5 eV and 3 eV pump pulses on a deliberately broad family of systems (metallic ferromagnets, ferrimagnetic insulators, half-metals, rare-earth alloys, and MgO tunnel barriers), it finds that the emitted THz waveforms are essentially identical in shape and only mildly different in amplitude. The authors conclude that the dominant driving force is ultrafast heating of the electron system by the pump, while highly excited primary photoelectrons contribute less than the noise floor. If true, the microscopic generation step is controlled by how much energy is deposited into the electrons, not by the energy of each photon.","feed_headline":"Terahertz spin currents don't care about pump photon energy","feed_subtitle":"Across ferromagnets, insulators, half-metals and tunnel barriers, only deposited heat matters, not where photons start.","key_machinery":"The central object is the terahertz spin current emitted by an F|HM bilayer and detected by electro-optic sampling as the magnetization-odd THz signal $S(t) = [S(t,+M_0)-S(t,-M_0)]/2$. The comparison loop is: 1.5 eV pulses from a Ti:sapphire oscillator, 3 eV pulses by second-harmonic generation in a BBO crystal with identical beam geometry, signals normalized by absorbed pump fluence, and 3 eV traces scaled by a factor stated to absorb absorption differences, setup sensitivity, and possibly injection efficiency. The two generation laws under test are the spin Seebeck relation $j_s^{SSE} \\propto (T_e^{HM}-T_{mag}^{FI})$ and the pyrospintronic relation $j_s^{PSE} \\propto (\\mu_s^{HM}-\\mu_s^{FM})$; the invariance across photon energy is what pins the driving force to electron heating.","core_discovery":"The central claim is that, in all F|HM stacks studied, switching the pump photon energy from 1.5 eV to 3 eV leaves the dynamics of the generated terahertz spin currents unchanged and changes only the amplitude slightly, once traces are normalized by absorbed fluence and a trace-specific factor. This is presented as evidence that the spin Seebeck effect (driven by the electronic temperature difference between HM and magnetic layer) and the pyrospintronic effect (driven by a transient spin-voltage difference) are set by the hot-electron population near the Fermi level, not by the initially excited photoelectrons. Cases engineered to be sensitive to photon energy—above-gap excitation of maghemite and magnetite, direct 4f excitation in Tb30Fe70, and resonant tunneling through MgO defect states—all show the same insensitivity.","pith_inferences":["A direct test of the heating picture would compare one-photon 3 eV excitation with two-photon 1.5 eV excitation at equal absorbed energy: the heating model predicts similar TSC amplitudes, while a model with primary-electron sensitivity predicts differences.","The normalization freedom means the amplitude statement is weaker than the dynamics statement; a calibrated absolute-amplitude measurement would cleanly separate photon-energy-dependent generation efficiency from absorption.","The result suggests that optically induced spin transfer and superdiffusive hot-electron spin currents will be hard to isolate in these stacks; experiments aiming at them should use detection below the current ~1% noise floor or use stacks designed to suppress thermal currents.","Extending the pump range to lower or higher photon energies (for example, 0.95 eV or 4.7 eV) would test whether the thermal bottleneck holds across a wider bandwidth."],"forward_implications":["Spintronic THz emitters do not need a specific pump wavelength, so cheaper and simpler laser sources can be used without changing emission dynamics.","Models of TSC generation can treat the pump as heating the electron system; resonant features in the band structure within one pump photon of the Fermi level do not need to be included to reproduce the observed dynamics.","Direct optical excitation of 4f states in rare-earth ferrimagnets, or of carriers across the band gap in ferrimagnetic insulators, does not produce a measurable signal at the sensitivity of the present experiment.","Energy-dependent tunneling through MgO defect states is not the rate-limiting step: the THz spin-current waveform is the same at 1.5 eV and 3 eV.","The population of primary hot electrons is estimated at about one percent or less, so pushing sensitivity beyond the current noise floor may reveal their signatures."],"supporting_citations":[{"why":"Provides the SSE reference and the finding of SSE/PSE coexistence in iron oxides that the above-gap test builds on.","marker":"[30]"},{"why":"Supplies the pyrospintronic-effect model and the spin-voltage formalism that the photon-energy-independence test is designed to check.","marker":"[31]"},{"why":"Establishes the coherent and incoherent tunneling channels in MgO barriers that the barrier samples are meant to probe.","marker":"[13]"},{"why":"Locates Tb 4f states about 2.3 eV below the Fermi level, which is what makes 3 eV excitation of these states possible.","marker":"[33]"},{"why":"Supplies atomic-scale control of spin current transmission at interfaces, informing the MgO barrier interpretation.","marker":"[36]"},{"why":"Previous report of negligible pump-wavelength impact in metallic trilayers that this study extends to a wider material set.","marker":"[44]"},{"why":"Supports the electron-heating picture by comparing optical and THz excitation of demagnetization in iron.","marker":"[45]"},{"why":"Reports SSE without PSE in Gd|Pt, underpinning the choice of rare-earth samples to test 4f excitation.","marker":"[47]"},{"why":"Provides the hot-electron thermalization picture used to argue that the primary-electron fraction is below one percent.","marker":"[54]"}],"fun_headline_variants":["Spin currents ignore pump photon energy, only heat matters","Photon energy can't tune THz spin currents; heat does","Pump photon energy doesn't govern THz spin currents","For THz spin currents, heat beats photon energy","Electron heating dominates THz spin currents, not photon energy"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The conclusion rests on the assumption that the normalization factor applied to the 3 eV traces accounts only for absorption and setup sensitivity, not for genuine photon-energy-dependent changes in spin-current generation efficiency; if it quietly absorbs such changes, the amplitude comparison is circular.","fun_headline_variants_meta":{"raw":{"variants":["Spin currents ignore pump photon energy, only heat matters","Photon energy can't tune THz spin currents; heat does","Pump photon energy doesn't govern THz spin currents","For THz spin currents, heat beats photon energy","Electron heating dominates THz spin currents, not photon energy"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000939,"raw_usage":{"total_tokens":4016,"prompt_tokens":946,"completion_tokens":3070,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":562,"completion_tokens_details":{"reasoning_tokens":2988}},"tokens_in":562,"tokens_out":3070,"duration_ms":25610,"temperature":1.0,"reasoning_tokens":2988,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T21:00:43.146661+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the absolute THz amplitude per absorbed photon at 1.5 eV and 3 eV with the detection transfer function calibrated independently for each wavelength. If the ratio at equal absorbed energy differs from unity by more than the experiment's ~1% noise floor, then at least part of the photon-energy dependence was being absorbed into the normalization factor, and the amplitude part of the claim fails. A second falsifier: observe any wavelength-dependent dynamics, such as an extra slow component when 3 eV excites 4f states in Tb30Fe70, in a higher signal-to-noise version of the same measurement.","supporting_citations":[{"cited_title":"S.; Wolf, M.; Rouzegar, R.; Brouwer, P","cited_arxiv_id":null,"evidence_quote":"Provides the SSE reference and the finding of SSE/PSE coexistence in iron oxides that the above-gap test builds on."},{"cited_title":"A.; Chekhov, A","cited_arxiv_id":null,"evidence_quote":"Supplies the pyrospintronic-effect model and the spin-voltage formalism that the photon-energy-independence test is designed to check."},{"cited_title":"A.; Chekhov, A","cited_arxiv_id":null,"evidence_quote":"Establishes the coherent and incoherent tunneling channels in MgO barriers that the barrier samples are meant to probe."},{"cited_title":"K.; Baer, Y .; Cox, P","cited_arxiv_id":null,"evidence_quote":"Locates Tb 4f states about 2.3 eV below the Fermi level, which is what makes 3 eV excitation of these states possible."},{"cited_title":"A.; Şaşıoğlu, E.; Hoppe, W.; Zhou, X.; Deniz, H.; Rouzegar, R.; Kampfrath, T.; Mertig, I.; Parkin, S","cited_arxiv_id":null,"evidence_quote":"Supplies atomic-scale control of spin current transmission at interfaces, informing the MgO barrier interpretation."},{"cited_title":"I.; Hornett, S","cited_arxiv_id":null,"evidence_quote":"Previous report of negligible pump-wavelength impact in metallic trilayers that this study extends to a wider material set."},{"cited_title":"L.; Behovits, Y .; Heitz, J","cited_arxiv_id":null,"evidence_quote":"Supports the electron-heating picture by comparing optical and THz excitation of demagnetization in iron."},{"cited_title":"Magnon-mediated terahertz spin transport in metallic Gd|Pt stacks","cited_arxiv_id":"2503.22483","evidence_quote":"Reports SSE without PSE in Gd|Pt, underpinning the choice of rare-earth samples to test 4f excitation."},{"cited_title":"S.; Güdde, J.; Conrad, U.; Jähnke, V.; Matthias, E","cited_arxiv_id":null,"evidence_quote":"Provides the hot-electron thermalization picture used to argue that the primary-electron fraction is below one percent."}],"review_version":1}