{"id":"3cb59ba0-839b-4f12-a938-f5867c1acbd9","arxiv_id":"2608.02106","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"XTANT-3 simulations predict transient superionic states in SrTiO3 and KTaO3 under femtosecond XUV irradiation, with a possible picosecond ferromagnetic instability in SrTiO3 but not KTaO3.","lead":"Simulations predict that intense femtosecond XUV pulses drive SrTiO3 and KTaO3 into a transient superionic state, and that in SrTiO3 the excited electrons may briefly trigger ferromagnetism. The work suggests the spatial extent of B-site d orbitals controls both structural and magnetic responses of perovskites under extreme irradiation.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Equations (I)–(II) may double-count spin degeneracy: if N_d(E) is the total d-DOS, S = I*N_eff overestimates the Stoner number by a factor of 2, potentially erasing the STO S>1 instability.","rationale":"The reader's weakest assumption focuses on spin fluctuations and the sensitivity of the Stoner exchange parameter I to excitation. Those are legitimate concerns about whether S>1 is sufficient for ferromagnetism, but they presuppose that the reported S values are correctly computed. A more fundamental and concrete issue is the normalization of the density of states entering Eq. (II). In a non-spin-polarized tight-binding calculation, the DOS is conventionally given per unit cell and includes both spin channels; the Stoner criterion, however, requires the DOS per spin. If the XTANT-3 d-DOS is not already divided by 2, the reported S is a factor of 2 too large. For STO, where the claim rests on S exceeding 1 by a modest margin, halving could remove the instability entirely. This can be checked directly from the code output, and it does not depend on subtleties of many-body theory. My recommendation is therefore conditional: publication should require the authors to state the spin normalization explicitly and to demonstrate that S(t)>1 survives when the DOS per spin is used. If it does not, the central magnetic claim fails, even if the structural superionic and melting results remain valid.","tokens_in":17114,"tokens_out":7945,"duration_ms":76601,"concrete_test":"Obtain the d-projected DOS used in Figure 11 from the XTANT-3 output for pristine STO at Te=0 and integrate Eq. (II) to compute S = I*N_eff; compare with the known per-spin DOS at the Fermi energy from a DFT calculation (e.g., refs. 61–62). Equivalently, rerun the analysis of Fig. 11 with N_eff divided by 2 (per-spin normalization) for the 1.2 and 2.0 eV/atom cases. If S(t) no longer exceeds 1 at any time, the transient ferromagnetism conclusion is an artifact of double-counting spin and must be retracted or re-derived.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central ferromagnetic claim rests on the generalized Stoner criterion S = I*N_eff (Eqs. I–II), where N_eff is obtained by integrating the d-projected DOS with a Fermi-function weight. In the standard Stoner model, the relevant DOS is per spin: the criterion is I*N(E_F)>1 with N per spin direction. The text describes the result as 'per spin channel' (Sec. III.C), but Eq. (II) contains no factor 1/2. If the XTANT-3 d-DOS is the usual spin-degenerate total DOS (standard for non-spin-polarized tight-binding calculations), then the plotted S(t) values are too large by exactly a factor of 2. For STO, where S is claimed to exceed 1 on a ~1 ps timescale, halving would place S below 1 (or near the threshold), invalidating the ferromagnetic instability claim. The paper gives no explicit normalization statement for N_d(E), and Figure 11 lacks error bars that might reveal the magnitude. This is a more direct threat to the numerical result than the important question of whether S>1 is sufficient (spin fluctuations), because it concerns whether S>1 is even obtained.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript presents XTANT-3 multiscale simulations of the ultrafast response of SrTiO3 (STO) and KTaO3 (KTO) to intense femtosecond XUV irradiation. It reports dose thresholds for superionic state formation, melting, and band-gap collapse, and attributes the different thresholds to the spatial extent of the B-site d orbitals. The headline claim is that STO, with its narrow Ti 3d conduction band and larger exchange parameter, passes a generalized Stoner criterion S = I N_eff and develops a transient ferromagnetic instability on ~1 ps timescales, whereas KTO remains paramagnetic. A Landau-Devonshire analysis is used to show that irradiation transiently deepens the ferroelectric potential well, with strain amplifying the effect. The structural thresholds are supported by Born-Oppenheimer versus non-BO comparisons and prior validation of the XTANT-3 code, but the ferromagnetic conclusion rests on a mean-field Stoner criterion applied to an approximate tight-binding DOS without spin-polarized validation.","tokens_in":17343,"tokens_out":5165,"duration_ms":45944,"significance":"If the ferromagnetic instability claim is correct, the paper establishes d-orbital spatial extent as a materials-design parameter for transient magnetic and ferroelectric response under ultrafast excitation, which would be of interest to the ultrafast and perovskite device communities. The structural and thermodynamic results are grounded in a multiscale code with documented prior validation, and the comparison between two materials is a clean way to motivate the orbital-bandwidth mechanism. The paper makes its data and code availability explicit. However, the central magnetic claim is not yet supported: the Stoner analysis may contain a spin-degeneracy normalization error, and the sufficiency of S > 1 for ferromagnetism is assumed without testing spin fluctuations or the excitation dependence of the exchange parameter. These issues must be resolved before the abstract and conclusions can be taken at face value.","major_comments":[{"comment":"The Stoner criterion is usually written as I N(E_F) > 1 with N(E_F) the per-spin density of states. Equation (II) defines N_eff as the Fermi-function-weighted integral of the d-projected DOS N_d(E) and contains no factor of 1/2. If N_d(E) is the spin-degenerate total d-DOS, as is standard for non-spin-polarized tight-binding codes, the computed S is too large by a factor of 2, and the STO S>1 result may vanish after correction. The text's statement that N_eff is 'per spin channel' does not resolve the ambiguity. Please state the normalization of N_d(E) explicitly and, if it is the total DOS, revise Eqs. (I)-(II), Figure 11, and the ferromagnetic conclusion accordingly.","section":"Section III.C, Eqs. (I)-(II)"},{"comment":"The paper treats S>1 as sufficient for a ferromagnetic instability, but the Stoner criterion is a mean-field condition; spin fluctuations and the dependence of the exchange parameter I on the excited electronic configuration are not addressed. The assumption that I (0.76 eV for STO, 0.45 eV for KTO) is unchanged under intense XUV excitation and superionic disorder is stated but not tested. To support the headline claim, the authors should provide a spin-polarized DFT or constrained DFT calculation for representative excited or oxygen-deficient configurations, or soften the abstract and conclusions to indicate a possible rather than established ferromagnetic instability.","section":"Section III.C"},{"comment":"The assertion that the band-gap overestimation makes the S>1 prediction a conservative lower bound is not self-evident; a larger gap changes where the chemical potential sits relative to the d-band peak and could either increase or decrease N_eff. Please provide the quantitative argument (or the numerical experiment) behind the 22-63% estimate and the lower-bound claim, or remove this statement.","section":"Section III.C, bandgap-overestimation paragraph"}],"minor_comments":[{"comment":"The notation in Eq. (I) is malformed: 'N_eff, d(µ, Te)' should be written as N_eff,d(μ,T_e) with the subscript 'd' and the argument 'μ,T_e' clearly separated.","section":"Eq. (I)"},{"comment":"The text states that electronic heat conductivity reaches a plateau at 'temperatures of ~1 eV'; please specify that this is the electron temperature and, if desired, also give the equivalent in kelvin for readers.","section":"Section III.A / Figure 2"},{"comment":"The abbreviation 'BO' appears in Table 1 before it is defined in Section III.B; please define 'Born-Oppenheimer (BO)' and 'non-BO' at first use in the text or in the table caption.","section":"Table 1 and Section III.B"},{"comment":"Figure 11 would benefit from error bars or at least a discussion of statistical uncertainty, since the S(t) values are the basis of the ferromagnetic claim.","section":"Figure 11"},{"comment":"The abstract uses '1 ps timescales' without a tilde; please use '~1 ps' for consistency with the text.","section":"Abstract"}],"recommendation":"major_revision","confidential_remarks":"The factor-of-two spin-degeneracy question is the most concrete threat to the central claim; if the authors can confirm that N_d(E) is per-spin (e.g., the TB code outputs spin-resolved DOS), the ferromagnetic conclusion may survive. Otherwise, halving S likely erases the STO instability. The novelty relative to known oxygen-vacancy ferromagnetism in STO is modest, but the ultrafast context and comparative KTO study are valuable. The manuscript would also benefit from softening the abstract if the spin-fluctuation issue remains unresolved."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know about arXiv:2608.02106. The structural damage part is solid and worth keeping: element-resolved thresholds for superionic oxygen diffusion and melting in STO and KTO, with BO/non-BO comparisons showing thermal effects dominate near threshold, plus a careful Landau–Devonshire analysis of the polar well under strain. That work is reproducible in principle and the numbers are specific enough to test. The headline claim, transient ferromagnetism in STO, is much weaker and may be wrong by a factor of two. The paper defines S = I*N_eff in Eq. (I) and N_eff via Eq. (II) with no factor 1/2, while the text twice says the result is 'per spin channel.' If N_d(E) is the total d-DOS, which is the usual output of a non-spin-polarized tight-binding code, then S is overestimated by exactly 2. STO's S reportedly exceeds 1, but the margin is not shown in Figure 11; halving would likely drop it below 1. The authors need to state the normalization of N_d explicitly. If they actually used per-spin DOS, fine, but then the equation is misleading. This is the first thing a referee should check. There is a second issue: even if the factor of two is resolved, the Stoner criterion is mean-field and I is taken from ground-state DFT and assumed unchanged by excitation. The authors themselves estimate band-gap-related uncertainty of 22–63% in N_eff, so S>1 is not robust. No spin-polarized calculation or experimental confirmation is provided. These are not fatal to the whole paper, but they are fatal to the abstract as written. The structural thresholds and the d-band-width mechanism differentiating STO and KTO are genuinely useful and likely correct. The Landau analysis, the charge-transfer timing around 350 fs, and the recrystallization arguments are all well presented. The self-citations are appropriate given the XTANT-3 development. In short: send to peer review, but expect major revision. The referee should ask for the DOS normalization, uncertainty bands on S(t), and a softer discussion of the magnetic prediction. If the factor-of-two concern holds, the magnetism should be reframed as a speculative possibility, not a result. I would not cite the ferromagnetism in my own work until it is resolved, but I might cite the structural thresholds.","headline":"Solid structural damage thresholds with a plausible orbital-bandwidth mechanism, but the headline ferromagnetism may rest on a factor-of-two spin degeneracy error and should not be trusted as published.","tokens_in":17885,"tokens_out":3456,"would_cite":false,"duration_ms":32062,"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":"Simulations indicate that femtosecond XUV irradiation can drive SrTiO3 into a transient ferromagnetic instability lasting roughly a picosecond, while KTaO3 remains paramagnetic.","keywords":["transient ferromagnetism","Stoner criterion","ultrafast phase transitions","perovskite oxides","SrTiO3","KTaO3","superionic states","femtosecond XUV irradiation"],"falsifier":"A pump–probe magneto-optical Kerr or X-ray magnetic circular dichroism measurement on SrTiO3 after a femtosecond 30 eV pulse depositing roughly 1.2 eV/atom: if no magnetization appears within the first few picoseconds and disappears by 5 ps, the predicted transient ferromagnetic instability is not present. Equivalently, recomputing $I$ for the excited, disordered supercell and finding it drops below the value needed for $S > 1$ would undermine the mechanism.","tokens_in":16910,"feed_emoji":"🧲","tokens_out":9804,"duration_ms":79100,"temperature":0.7,"pith_summary":"This paper asks what happens to two closely related perovskite oxides, SrTiO3 and KTaO3, when they absorb an intense femtosecond XUV pulse. The authors find that both materials first pass through a superionic state, in which the oxygen sublattice melts while the metal sublattices stay ordered, and then melt completely at higher doses. The magnetic response, however, diverges: in SrTiO3 the generalized Stoner number rises above one on a picosecond timescale, signalling a transient ferromagnetic instability, while in KTaO3 it never does. The paper attributes this difference to the spatial extent of the B-site d orbitals—compact Ti 3d states in SrTiO3 produce a narrow conduction band and a large exchange parameter, whereas the extended Ta 5d states in KTaO3 produce a broad band and a smaller exchange parameter. If correct, this makes d-orbital extent a single structural parameter that controls both the phase-transition sequence and the magnetic response of perovskite oxides under extreme excitation.","feed_headline":"XUV flash may make SrTiO3 magnetic for a picosecond","feed_subtitle":"A compact Ti 3d band tips SrTiO3 past the Stoner threshold; KTaO3 does not.","key_machinery":"The load-bearing object is the generalized Stoner criterion $S = I N_{\\mathrm{eff}}$, applied to transient, irradiation-modified tight-binding densities of states. Here $I$ is the intra-atomic Stoner exchange parameter (0.76 eV for STO, 0.45 eV for KTO) and $N_{\\mathrm{eff}}$ is the Fermi–Dirac-weighted average of the d-projected DOS around the chemical potential; $S > 1$ means the exchange energy gain of spin polarization outweighs the kinetic-energy cost, so the paramagnetic state becomes unstable. A second piece of machinery is the Landau–Devonshire free-energy fit $F(Q) = a_2 Q^2 + a_4 Q^4$, used to track how irradiation deepens the ferroelectric double well through the sign of $a_2$.","core_discovery":"The central discovery is a mechanism that ties the transient magnetic response of a perovskite to the spatial extent of its B-site d orbitals. Using a coupled simulation of electron cascades, tight-binding electronic structure, and molecular dynamics, the authors compute the generalized Stoner number $S = I N_{\\mathrm{eff}}$, where $I$ is the intra-atomic exchange parameter and $N_{\\mathrm{eff}}$ is the thermally averaged d-projected density of states near the chemical potential. After irradiation at doses around 1.2–2.0 eV/atom, $S$ exceeds unity in SrTiO3 on the ~1 ps timescale, indicating a ferromagnetic instability; in KTaO3 it remains below unity for all simulated doses. The physical origin is the narrow Ti 3d conduction band, which gives a large $N_{\\mathrm{eff}}$ combined with $I = 0.76$ eV in STO, versus the broad Ta 5d band with smaller $N_{\\mathrm{eff}}$ and $I = 0.45$ eV in KTO. The authors also note that because their tight-binding band gaps overestimate the true gaps, the $S > 1$ prediction for STO is a conservative lower bound.","pith_inferences":["If the transient ferromagnetic instability in STO is real, a single femtosecond XUV pulse could act as an all-optical switch for magnetization in an initially non-magnetic oxide, with the magnetized state lasting only a few picoseconds; the dose would control the lifetime.","The proposed mechanism predicts a systematic trend across other perovskites: compounds with narrow 3d conduction bands, such as other titanates, should show transient ferromagnetic instabilities, while 4d/5d tantalates, niobates, and similar broad-band systems should not. This is a testable extension the paper does not itself simulate.","Because the paper argues that band-gap overestimation makes $S > 1$ a lower bound, the ferromagnetic window in STO should widen as the dose increases up to the melting threshold; an experiment could check whether the onset of magnetization tracks the predicted 1 ps timescale.","The strain enhancement seen in the Landau–Devonshire analysis suggests that strained films, not just bulk crystals, would be the best platforms to observe the transient ferroelectric and possibly coupled magnetic response."],"forward_implications":["SrTiO3 irradiated with femtosecond XUV pulses at doses between 0.7 and 1.6 eV/atom first forms a superionic state with a diffusing oxygen sublattice; above 1.6 eV/atom it melts completely.","KTaO3 shows the same sequence at slightly higher thresholds: superionic onset at 0.9 eV/atom and complete melting above 1.5 eV/atom.","At doses near 1.2–2.0 eV/atom, the Stoner number in SrTiO3 exceeds one for about a picosecond, implying a dose-tunable transient ferromagnetic window before the system melts or relaxes.","The same B-site d-orbital mechanism that controls the magnetic instability also controls band-gap collapse thresholds, so the two materials' damage and electronic response can be understood from one structural parameter.","Irradiation at 0.3 eV/atom transiently deepens the polar ferroelectric potential well, and strain amplifies this deepening by about 2 times in STO and 6 times in KTO."],"supporting_citations":[{"why":"Supplies the multiscale simulation framework used for all results, coupling electron cascades, tight-binding electronic structure, and molecular dynamics.","marker":"32"},{"why":"Provides the generalized Stoner criterion definition and the condition $S > 1$ for ferromagnetic instability.","marker":"74,75"},{"why":"Supplies the Stoner exchange parameter $I = 0.45$ eV used for KTaO3.","marker":"76,77"},{"why":"Supplies the Stoner exchange parameter $I = 0.76$ eV used for SrTiO3.","marker":"77,78"},{"why":"Establishes the original Stoner criterion for collective electron ferromagnetism that the generalized form extends to finite electronic temperature.","marker":"79"},{"why":"Shows prior application of the Stoner criterion to doped wide-band-gap oxides, supporting its use for these perovskite oxides.","marker":"82,83"},{"why":"Reports defect-induced ferromagnetism in SrTiO3 from oxygen vacancies, the equilibrium analogue the paper argues XUV irradiation can create transiently.","marker":"84–86"},{"why":"Documents how greater spatial extent of 5d orbitals broadens bands, explaining KTO's lower effective density of states.","marker":"87"}],"fun_headline_variants":["SrTiO3 flips ferromagnetic for 1 ps under XUV","XUV pulse magnetizes SrTiO3 for a picosecond","Ti 3d width controls ultrafast magnetism in SrTiO3","Perovskite magnetism: SrTiO3 vs KTaO3 under XUV","Transient ferromagnetism in SrTiO3 induced by XUV"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire ferromagnetic prediction depends on the assumption that the ground-state Stoner exchange parameter $I$ stays valid under intense excitation and that exceeding the Stoner number one, computed from a tight-binding density of states, is enough to produce transient ferromagnetic order in a disordered, hot lattice.","fun_headline_variants_meta":{"raw":{"variants":["SrTiO3 flips ferromagnetic for 1 ps under XUV","XUV pulse magnetizes SrTiO3 for a picosecond","Ti 3d width controls ultrafast magnetism in SrTiO3","Perovskite magnetism: SrTiO3 vs KTaO3 under XUV","Transient ferromagnetism in SrTiO3 induced by XUV"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000627,"raw_usage":{"total_tokens":2979,"prompt_tokens":1103,"completion_tokens":1876,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":719,"completion_tokens_details":{"reasoning_tokens":1782}},"tokens_in":719,"tokens_out":1876,"duration_ms":13187,"temperature":1.0,"reasoning_tokens":1782,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T15:02:26.061537+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A pump–probe magneto-optical Kerr or X-ray magnetic circular dichroism measurement on SrTiO3 after a femtosecond 30 eV pulse depositing roughly 1.2 eV/atom: if no magnetization appears within the first few picoseconds and disappears by 5 ps, the predicted transient ferromagnetic instability is not present. Equivalently, recomputing $I$ for the excited, disordered supercell and finding it drops below the value needed for $S > 1$ would undermine the mechanism.","supporting_citations":[],"review_version":2}