{"id":"5d6fff3d-cd6f-4768-8b02-94c9b479492d","arxiv_id":"2501.19238","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":8,"one_line_summary":"The first time-resolved measurement of Hubbard U renormalization in NiO shows the on-site interaction is quenched by about 100 meV within a few femtoseconds.","lead":"Researchers measured how nickel oxide responds to a few-femtosecond infrared laser pulse using attosecond XUV probe pulses. They find the material's electron-electron interaction drops within about 7 femtoseconds, a new timescale for controlling correlated materials.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Frozen-U control may not isolate Hubbard U dynamics because the frozen-U trajectory absorbs less pump energy; the attribution of the M-edge redshift to ΔU(t) is therefore not yet uniquely established.","rationale":"I read the paper in good faith. The experiment is careful: MgO control, Kramers-Kronig analysis, pure redshift extraction, and a 7 fs response measurement. The dynamic-U versus frozen-U comparison is the key piece of evidence for the central claim. My concern is that the frozen-U control may be confounded by unequal excitation, because U controls the gap and hence how strongly the pump is absorbed. This is testable by matching absorbed energy or carrier density between the two calculations. The multiplet result requiring a hybridization reduction further supports model dependence in the conversion from spectral shift to ΔU(t). These issues do not refute the paper; they make the 'direct measurement' wording conditional on an additional control. The reader's CONDITIONAL verdict remains appropriate; my concern is more specific than the reader's weakest assumption, hence partial agreement.","tokens_in":14222,"tokens_out":7581,"duration_ms":79473,"concrete_test":"Recompute the TDDFT+U transient reflectivity for NiO with the frozen-U approximation at a higher pump intensity chosen so that the time-integrated absorbed energy or excited-carrier density matches the dynamic-U run over the first 20 fs. If the M2,3-edge redshift remains absent, the frozen-U control is valid and the U-attribution is supported. If a similar redshift appears, the dynamic-U/frozen-U comparison is confounded by excitation density, and the claim that the shift measures ΔU(t) would need re-evaluation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central inference rests on the difference between TDDFT+U runs with dynamic and frozen U (Fig. 3b,c). The frozen-U calculation is presented as a control that removes U dynamics while retaining 'all population transfer and local-field effects.' But the control does not match the excitation conditions: because the gap depends on U, a frozen-U trajectory experiences different pump absorption and carrier density than the dynamic-U trajectory. The near-zero signal in the frozen-U run could therefore reflect a weaker pump-induced excitation rather than the absence of U renormalization. This concern is reinforced by the authors' own multiplet analysis, which finds that a pure ΔU does not reproduce the measured rigid shift unless a 3.3% hybridization reduction is added: the observable is not uniquely tied to U within the parameter model. Consequently, extracting ΔU(t) as simply -2×(edge shift) (Fig. 4c) and calling it a direct measurement of Hubbard U renormalization is not yet uniquely established.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports attosecond transient reflectivity measurements on NiO and MgO under identical 5 fs, 1.58 eV pump pulses. MgO shows the oscillatory response of the dynamical Franz-Keldysh effect, whereas NiO shows a weaker, slower, non-oscillatory response that is well described by a rigid -51 meV redshift of the Ni M2,3 edge. The authors interpret this redshift as a light-driven quench of the on-site Hubbard U by roughly 100 meV, with a material response time of 7.0 +/- 1.5 fs, and support this interpretation with real-time TDDFT+U calculations comparing dynamic and frozen U, plus multiplet calculations. They also follow the signal to 50 ps and identify two decay timescales consistent with earlier electron-diffraction work on a nonthermal, structurally distorted state.","tokens_in":14415,"tokens_out":4350,"duration_ms":41923,"significance":"If the identification is correct, this would be the first time-resolved observation of light-driven Hubbard U renormalization in a strongly correlated insulator at its intrinsic few-femtosecond timescale, an important result for attosecond spectroscopy and for light-driven control in correlated materials. The paper has clear strengths: a clean experimental characterization of the rigid edge shift, a state-of-the-art TDDFT+U framework, a dynamic-versus-frozen-U comparison, a CEP-dependence check, and a thoughtful connection to longer timescale structural dynamics. However, the claim of a direct measurement of U(t) is currently stronger than the evidence, because the mapping from the measured spectral shift to Delta U(t) is model-dependent and the frozen-U control does not match the excitation conditions of the dynamic-U trajectory. These issues are correctable, but they affect the central claim.","major_comments":[{"comment":"The frozen-U control is presented as keeping all population transfer and local-field effects while removing U dynamics, but the two trajectories are not excited identically: with U frozen, the gap remains closer to its ground-state value, so the pump absorption and the resulting carrier density differ from the dynamic-U run. The near-zero signal in Fig. 3c could therefore reflect weaker pump excitation rather than the absence of U renormalization. Please report the absorbed energy or the excited carrier density in the two runs, or repeat the frozen-U calculation at an increased pump strength that matches the dynamic-U excitation, and discuss how this affects the conclusion that the experimental redshift is almost entirely due to the renormalization of U.","section":"Fig. 3c and Methods, 'TDDFT+U simulations'"},{"comment":"The multiplet analysis shows that a pure Delta U = -100 meV does not reproduce the measured rigid shift; a simultaneous 3.3% reduction of the hybridization is required, and the authors explicitly state that the parameter choice is not unique and only linearly constrained. This means the observed observable is not uniquely tied to U within the model. The central claim that the redshift directly gives Delta U(t) therefore needs to be qualified. The authors should either provide an independent estimate of the hybridization contribution or present Delta U(t) in Fig. 4c as an effective quantity inferred within a specific model, not as a direct measurement.","section":"Fig. 3e and Methods, 'Multiplet calculations'"},{"comment":"The theory reproduces the magnitude of the effect but not its time constant: TDDFT+U predicts a faster rise than the measured electronic response time of 7.0 +/- 1.5 fs. Since one of the central claims is that the experiment accesses the true timescale of U renormalization, the disagreement in the rise time is more than a cosmetic discrepancy. The paper invokes non-local screening in EDMFT models, but no quantitative comparison is provided. Please clarify what the measured 7 fs timescale represents if the ab initio theory cannot reproduce it, and whether the extracted Delta U(t) in Fig. 4c is robust to alternative deconvolution assumptions and to the instrument response function.","section":"Fig. 4c and 'Few-femtosecond screening dynamics'"}],"minor_comments":[{"comment":"The sentence 'The CEP of the laser is not stabilized. However, this does not prevent the measurement of sub-cycle features as demonstrated by our MgO measurements and previous experiments' should cite the previous experiments or state explicitly how the MgO measurement demonstrates this.","section":"Experimental setup"},{"comment":"There is a typo: 'ultrafast phemonena' should read 'ultrafast phenomena'.","section":"Final paragraph"},{"comment":"The sentence 'We checked, see SI, that the CEP as no impact on the light-induced of change of the Hubbard U' should read 'has no impact on the light-induced change of the Hubbard U'.","section":"Methods, 'TDDFT+U simulations'"},{"comment":"The time-axis label appears redundantly in the first panel of Fig. 3; please unify the axis labels across panels for clarity.","section":"Figures 3a-c"}],"recommendation":"major_revision","confidential_remarks":"The paper reports an interesting and potentially important experiment, and the theoretical framework is substantial. However, the central claim of a direct measurement of Hubbard U renormalization currently rests on a model-dependent conversion and on a control calculation that does not match excitation conditions. These issues are fixable within the scope of a revision, so I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Know this: the paper reports the first attosecond-resolved transient reflectivity on a canonical correlated insulator (NiO) and attributes the observed rigid redshift of the Ni M2,3 edge to a light-driven reduction of Hubbard U. The data are original, the MgO control is well chosen, and the dynamic-U vs frozen-U TDDFT comparison is a genuinely informative control. If the interpretation holds, the 7 fs response time is a new observable.\n\nThe experimental work is careful: delay stability of 100 as, Kramers-Kronig handled properly, rigid shift extracted cleanly at -51 meV. The contrast with MgO, which shows clear DFKE oscillations, while NiO shows none, is striking. The TDDFT+U calculations match the shape and magnitude of the transient reflectivity, and the frozen-U run nearly eliminates the signal. That is real evidence.\n\nThe soft spots are real but not fatal. The frozen-U control is not perfectly clean: because the gap depends on U, the frozen-U trajectory absorbs less pump energy, so part of the vanishing signal could be weaker excitation rather than absence of U dynamics. The multiplet analysis needs a 3.3% hybridization reduction to reproduce the shift, so the observable is not uniquely tied to U within that model. The theory underestimates the rise time, and the authors acknowledge this. Data and code are only available on request, which is a limitation for reproducibility. The phrase \"direct measurement\" overstates what is actually a model-dependent extraction.\n\nThis paper deserves a serious referee. The central claim is plausible and important, but the identification of the edge shift with ΔU(t) is not airtight. The authors should address the frozen-U excitation issue, clarify the non-uniqueness of multiplet parameters, and soften the direct-measurement language. I would bring it to reading group to discuss the interpretation, and I would cite the experimental data in my own work.","headline":"Fresh attosecond data on NiO with a plausible but not airtight claim that the M-edge redshift tracks Hubbard U renormalization; worth a careful referee.","tokens_in":15060,"tokens_out":1752,"would_cite":true,"duration_ms":17811,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["71.27.+a","78.47.J-"],"model":"deepseek-v4-flash","headline":"Attosecond pump-probe reflectivity on nickel oxide shows that intense laser light quenches the on-site Hubbard interaction U by about 100 meV within a few femtoseconds, marking the first direct time-resolved measurement of Hubbard U…","keywords":["attosecond transient reflectivity","Hubbard U renormalization","strongly correlated insulators","nickel oxide (NiO)","time-dependent DFT + U","charge-transfer insulator","few-femtosecond dynamics","dynamical Franz-Keldysh effect"],"falsifier":"An attosecond pump-probe measurement that tracks a transition whose final state is not the upper Hubbard band (for example, a Ni 2p to 3d resonant transition) would settle the claim: if its shift is not half the M-edge redshift, the interpretation of the redshift as ΔU/2 is incorrect.","tokens_in":13961,"feed_emoji":"⚡","tokens_out":7989,"duration_ms":68715,"temperature":0.7,"pith_summary":"This paper reports attosecond pump-probe reflectivity measurements on the strongly correlated insulator nickel oxide (NiO) and the band insulator magnesium oxide (MgO) under identical excitations. It finds that the two insulators respond differently: MgO shows field-driven oscillations typical of the dynamical Franz-Keldysh effect, while NiO shows a pure rigid redshift of its M2,3 absorption edge with no sub-cycle oscillations. The authors argue that this redshift is caused by a laser-driven reduction of the effective on-site Coulomb repulsion (Hubbard U) of about 100 meV, which shifts the upper Hubbard band downward. Because the spectral change is a pure rigid shift, the measurement directly tracks ΔU(t), showing that the interaction quench develops with a material response time of 7.0 ± 1.5 fs. If correct, this establishes that electron-electron interactions in a correlated material can be manipulated on their natural femtosecond timescale.","feed_headline":"Laser quenches Hubbard U in NiO within 7 fs","feed_subtitle":"The 51 meV redshift of NiO's M-edge shows the on-site Coulomb repulsion dropping ~100 meV in a few femtoseconds.","key_machinery":"The central object is the effective on-site Coulomb repulsion U of the Hubbard model, here the Ueff for Ni 3d orbitals in NiO. The measurement exploits the fact that the XUV probe excites a Ni 3p core electron into the empty upper Hubbard band, whose energy lies U above the lower Hubbard band; a change ΔU(t) therefore shifts the 3p-to-upper-Hubbard-band transition by ΔU(t)/2. The argument is carried by the contrast between time-dependent DFT+U simulations with a dynamically renormalized U and a control calculation with U frozen to its ground-state value: the frozen-U calculation suppresses the predicted transient reflectivity, identifying U renormalization rather than carrier injection, local-field effects, or heating as the dominant mechanism. A second object is the multiplet ligand-field model used to show that the observed rigid shift can be reproduced by a parameter set with ΔU = −100 meV and a 3.3% hybridization reduction, providing an independent route to the same conclusion.","core_discovery":"The central claim is that the attosecond response of NiO is governed not by single-particle band-structure effects but by a light-driven quench of on-site electronic correlations. Under a 5 fs, 1.58 eV pump at 6 TW cm−2, the Ni M2,3 edge of NiO undergoes a pure rigid redshift of (−51 ± 2) meV with no resolvable 2ω oscillations, in contrast to MgO measured under identical conditions. Using real-time TDDFT+U calculations that allow the effective U to evolve dynamically, the authors show that this redshift is almost entirely due to a reduction of Hubbard U by about 100 meV, with the transition energy shifting by ΔU/2; freezing U to its ground-state value removes nearly the whole signal. Multiplet ligand-field calculations reproduce the rigid shift only if a 3.3% reduction of hybridization accompanies the U change. The measured ΔU(t) rises with a response time of 7.0 ± 1.5 fs, and the signal persists to picosecond timescales, matching earlier electron-diffraction evidence for a non-thermal structurally distorted state. The paper claims the first direct time-resolved observation of Hubbard U renormalization at its natural timescale.","pith_inferences":["If the 7 fs timescale reflects the hopping or screening time of the Hubbard model, the same experiment on other charge-transfer or Mott insulators with different hopping integrals should show a correspondingly different response time, providing a testable scaling relation.","The requirement of a 3.3% hybridization reduction in the multiplet fit suggests that the laser field also modifies Ni–O hybridization; future measurements with O K-edge or Ni L-edge probes could discriminate the hybridization change from the pure U change.","The success of this interpretation implies that attosecond transient reflectivity can be used as a general tool to watch interaction quenches in strongly correlated systems, potentially extending to photo-induced phase transitions where U collapse is the trigger.","The paper's model-dependence could be sharpened by deriving the redshift-to-ΔU conversion from a sum rule or from an independent experimental observable, rather than from a theoretical fit."],"forward_implications":["The measured response time of 7.0 ± 1.5 fs constitutes a new observable: the speed at which the electron-electron interaction, or its screening, can be modified by light in a correlated insulator.","Since the redshift is exactly half the U change, attosecond transient reflectivity at a core-to-upper-Hubbard-band transition can serve as a quantitative, time-resolved probe of ΔU(t) in correlated materials.","The absence of 2ω oscillations and of CEP dependence in NiO implies that the U renormalization is governed by laser intensity rather than the instantaneous electric field, so lightwave control schemes based on field-driven band shifts may not transfer to strongly correlated systems.","The few-femtosecond electronic quench precedes a slower (picosecond) exchange-striction-driven lattice distortion, unifying the attosecond and long-lived response of NiO in a single sequence: interaction quench, antiferromagnetic-order reduction, then structural relaxation."],"supporting_citations":[{"why":"Predicted that strong laser fields reduce Hubbard U in NiO, providing the theoretical hypothesis tested here.","marker":"[19]"},{"why":"Developed the TDDFT+U approach for transient absorption in NiO and predicted the importance of U renormalization.","marker":"[20]"},{"why":"Established the self-consistent real-time DFT+U method used for the dynamical-U simulations.","marker":"[35]"},{"why":"Showed via extended dynamical mean-field theory that screening dynamics in photoexcited Mott insulators are delayed on the hopping timescale, supporting the 7 fs response interpretation.","marker":"[40]"},{"why":"Measured ultrafast exchange-striction-driven lattice dynamics in NiO, providing the picosecond timescales that match the long-lived response observed here.","marker":"[41]"},{"why":"Reported attosecond core-level transient reflectivity on MgO, serving as the reference band-insulator measurement and instrument calibration.","marker":"[24]"},{"why":"Provided the static Ni M-edge X-ray absorption spectrum used to benchmark the multiplet parameters.","marker":"[27]"},{"why":"Demonstrated the dynamical Franz-Keldysh effect in a band insulator, the reference response that NiO is shown to lack.","marker":"[15]"}],"fun_headline_variants":["NiO's Hubbard U quenches in 7 fs under laser","Attosecond probe reveals Hubbard U collapse in NiO","Laser-driven quench of Hubbard U seen in NiO in 7 fs","Hubbard U renormalization timescale measured: 7 fs in NiO","NiO's Hubbard U drops 100 meV in 7 fs, quenched by laser"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The attribution of the 51 meV M-edge redshift to a pure reduction of Hubbard U assumes that no other mechanism produces an identical rigid shift at these timescales; this identification rests on a TDDFT+U comparison and on multiplet fits that require an additional fitted hybridization reduction, while the theory gives a faster rise time than the measured 7.0 fs.","fun_headline_variants_meta":{"raw":{"variants":["NiO's Hubbard U quenches in 7 fs under laser","Attosecond probe reveals Hubbard U collapse in NiO","Laser-driven quench of Hubbard U seen in NiO in 7 fs","Hubbard U renormalization timescale measured: 7 fs in NiO","NiO's Hubbard U drops 100 meV in 7 fs, quenched by laser"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000588,"raw_usage":{"total_tokens":2790,"prompt_tokens":1005,"completion_tokens":1785,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":621,"completion_tokens_details":{"reasoning_tokens":1683}},"tokens_in":621,"tokens_out":1785,"duration_ms":11024,"temperature":1.0,"reasoning_tokens":1683,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T20:52:53.059960+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"An attosecond pump-probe measurement that tracks a transition whose final state is not the upper Hubbard band (for example, a Ni 2p to 3d resonant transition) would settle the claim: if its shift is not half the M-edge redshift, the interpretation of the redshift as ΔU/2 is incorrect.","supporting_citations":[{"cited_title":"Tancogne-Dejean, M","cited_arxiv_id":null,"evidence_quote":"Predicted that strong laser fields reduce Hubbard U in NiO, providing the theoretical hypothesis tested here."},{"cited_title":"Tancogne-Dejean, M","cited_arxiv_id":null,"evidence_quote":"Developed the TDDFT+U approach for transient absorption in NiO and predicted the importance of U renormalization."},{"cited_title":"Tancogne-Dejean, M","cited_arxiv_id":null,"evidence_quote":"Established the self-consistent real-time DFT+U method used for the dynamical-U simulations."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Measured ultrafast exchange-striction-driven lattice dynamics in NiO, providing the picosecond timescales that match the long-lived response observed here."},{"cited_title":"G´ eneaux, C","cited_arxiv_id":null,"evidence_quote":"Reported attosecond core-level transient reflectivity on MgO, serving as the reference band-insulator measurement and instrument calibration."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provided the static Ni M-edge X-ray absorption spectrum used to benchmark the multiplet parameters."},{"cited_title":"Lucchini, S","cited_arxiv_id":null,"evidence_quote":"Demonstrated the dynamical Franz-Keldysh effect in a band insulator, the reference response that NiO is shown to lack."}],"review_version":1}