{"id":"f1a5c0cd-e9da-4c71-8d14-4a28bbdee3ab","arxiv_id":"2411.18031","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"Pressure-dependent pump-probe data on Ta2NiSe5 track phase transitions at about 1 and 3 GPa, and a Rothwarf-Taylor analysis gives a bandgap pressure coefficient of 65 meV/GPa, close to a DFT value of 70 meV/GPa.","lead":"Ultrafast pump-probe spectroscopy under pressure maps Ta2NiSe5 from an excitonic insulator to a semiconductor at about 1 GPa and then to a semimetal near 3 GPa. The authors infer a bandgap closing rate of 65 meV/GPa from a model-based fit, matching a first-principles value of 70 meV/GPa.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 65 meV/GPa slope is not robustly determined: Eq. 2 is inverted without specifying how A0, J, K, L are constrained, and the extracted Delta(1 GPa)=63 meV contradicts the paper's own cited 2Delta=160 meV at P_C1.","rationale":"The reader's weakest assumption correctly identifies that the Rothwarf-Taylor parameters A0, J, K, L are assumed pressure-independent, which is a plausible source of bias. My stress-test goes further: even under that assumption, the paper does not specify how these parameters are fixed before Delta(P) is inverted, making the extraction underdetermined. The clearest evidence that something is wrong is the internal inconsistency: the paper's own cited 2Delta=160 meV gap at P_C1=1 GPa conflicts with the extracted 2Delta(1 GPa)=126 meV. This is not a disagreement with external consensus but an inconsistency within the manuscript itself, and it directly undermines the central quantitative claim. The qualitative phase boundaries are supported by prior transport, Raman, and XRD work, so a conditional verdict with a request for raw data, parameter values, and uncertainty propagation remains appropriate. I therefore recommend no change to the reader's CONDITIONAL verdict, while emphasizing that the quantitative centerpiece requires the proposed test before acceptance.","tokens_in":9909,"tokens_out":4033,"duration_ms":36408,"concrete_test":"Re-derive Delta(P) from the A1 and tau1 data using a global fit in which A0, J, K, L are allowed to vary linearly with pressure (or are fixed by an independent constraint, e.g., from the ambient-pressure values and the cited 2Delta=160 meV at 1 GPa). If the resulting slope differs from 65 meV/GPa by more than 10%, or if the fit forces Delta(1 GPa) to move by more than 10 meV from the extracted value, the reported slope is not robust. Additionally, report the best-fit values and covariance of A0, J, K, L so the inversion is reproducible.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim rests on extracting the pressure-dependent semi-bandgap Delta(P) from the Rothwarf-Taylor expressions in Eq. 2 using the measured fast amplitude A1 and relaxation time tau1. The paper states that A0, J, K, L and Delta are fitting parameters, but it never reports how A0, J, K, L are fixed before Delta(P) is inverted. Two readings are possible: (i) a point-by-point inversion at each pressure using arbitrarily chosen constants, which would make the resulting slope a fitting artifact; or (ii) a global fit with Delta(P)=Delta0 - a*P and constant A0, J, K, L, in which case the reported 'evaluated' Delta values are not independent of the model. In either case, no error bars or parameter values are given, so the 65 meV/GPa slope has no stated uncertainty. More seriously, the extracted Delta(1 GPa) = 128 - 65 = 63 meV gives 2Delta = 126 meV, whereas the paper itself cites a bandgap of 2Delta = 160 meV at P=1 GPa from Ref. [48]. The model curve in Figure 3a,b therefore does not match the independently known gap at the EI-to-semiconductor transition. This internal inconsistency suggests that either the RT parameters are not actually pressure-independent, or the inversion procedure is biased. Because the DFT value of 70 meV/GPa is derived from a different constrained-occupation calculation and may share systematic errors, the claimed agreement cannot rescue the experimental slope without resolving this discrepancy. The qualitative phase boundaries at ~1 and ~3 GPa are credible, but the quantitative 65 meV/GPa pressure coefficient is not sufficiently supported.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports pressure-dependent optical pump-optical probe reflectivity measurements on the candidate excitonic insulator Ta2NiSe5 inside a diamond anvil cell. A biexponential fit separates fast and slow relaxation components; changes in the fitting parameters near 1 and 3 GPa are assigned to an excitonic-insulator-to-semiconductor transition and a semiconductor-to-semimetal transition. In the EI phase, the fast amplitude and relaxation time are interpreted with the Rothwarf-Taylor model, Eq. (2), to yield a linear semi-bandgap variation Delta(P) = 128 - 65 P meV, and the 65 meV/GPa slope is compared with a constrained-occupation DFT value of 70 meV/GPa. The authors also report fluence-dependent relaxation times and discuss the pressure dependence of the slow phonon-related channel.","tokens_in":10237,"tokens_out":5734,"duration_ms":51538,"significance":"If the quantitative claim were robust, the paper would provide a direct ultrafast-spectroscopy measurement of the pressure coefficient of the excitonic gap in Ta2NiSe5, a cross-check for constrained DFT, and a useful demonstration of in situ high-pressure pump-probe spectroscopy. The qualitative phase assignments are supported by clear changes in independently fitted relaxation parameters and are consistent with prior resistivity and Raman work. The DFT calculation is a genuine external benchmark rather than a refit of the experimental data, which is a strength. However, the load-bearing Rothwarf-Taylor inversion is underdetermined as presented, the extracted slope has no stated uncertainty, and one of the paper's own cited gap values is inconsistent with the fitted Delta(P). For these reasons the central quantitative claim is not yet established.","major_comments":[{"comment":"The inversion procedure that yields the central 65 meV/GPa slope is not specified. Equation (2) is written with A0, J, K, L and Delta as fitting parameters; since A1 and tau1 provide only two measured numbers per pressure, a point-by-point inversion is underdetermined, and a global fit would require stating how A0, J, K, L are constrained. No parameter values, no number of fitted points, and no error bars on Delta0 or the slope are given. The authors should report the full fitting protocol, the optimized RT constants, and the confidence interval of the slope, or the quantitative claim should be explicitly downgraded to a model-dependent estimate.","section":"Results and discussion, after Eq. (2)"},{"comment":"The black RT curves in Fig. 3(a,b) are not an independent validation of the model. The text evaluates Delta(P) from the experimental A1 and tau1 through Eq. (2) and then plots Eq. (2) with that same Delta(P); the visual agreement is therefore a consistency check imposed by the inversion. The DFT slope of 70 meV/GPa remains the only external benchmark, and the discussion should be revised so that the fitted curves are not presented as confirmation of the extracted slope.","section":"Figure 3(a,b) and derivation of Delta(P)"},{"comment":"There is an internal numerical inconsistency with the cited literature gap. The linear relation Delta(P) = 128 - 65 P meV gives 2Delta(1 GPa) = 126 meV, while the same paragraph states that at P = 1 GPa the semiconductor bandgap is 2Delta = 160 meV (Ref. [48]). If the 160 meV value is correct, the extracted slope is incompatible with it; if it is not meant to be compared, the reason must be stated. This should be resolved either by imposing the 160 meV point as a constraint in the fit or by providing uncertainties large enough to cover the difference.","section":"Results and discussion, paragraph on extraction of Delta"},{"comment":"The assignment of all pressure dependence of A1 and tau1 to Delta(P) assumes that A0, J, K, and L are independent of pressure up to P_C1. The manuscript itself notes that Raman linewidths increase under pressure (Ref. [27]) and discusses pressure-dependent electron-phonon coupling, so the phonon-bottleneck parameters J and L, and possibly K and A0, may vary. A concrete check is to fit the EI-phase dataset with Delta(P) = Delta0 - a P while allowing one or more RT parameters to vary and to compare the quality of fit, or to constrain Delta(P) independently.","section":"Eq. (2) and pressure dependence of RT parameters"}],"minor_comments":[{"comment":"The phrase 'The pressure coefficient of the bandgap decreases, 65 meV/GPa' reads awkwardly; suggest 'decreases with a pressure coefficient of 65 meV/GPa'.","section":"Abstract"},{"comment":"The symbol Delta is used both for the excitonic binding energy (Delta_EI = 2Delta) and for the semi-bandgap Delta; this double use should be clarified with distinct notation.","section":"Equation (2)"},{"comment":"The caption calls the black lines a fit, while the text says the same curves are produced from the extracted Delta(P); the wording should be made consistent.","section":"Figure 3 caption"},{"comment":"The sentence 'A direction correlation is observed...' contains a typo ('direction' should be 'direct').","section":"Conclusions"},{"comment":"The values P_C1 ~ 1 GPa and P_C2 ~ 3 GPa are reported but no criterion is defined for how they were read off from the pressure-dependent fitting parameters; an explicit procedure would help.","section":"Results and discussion, Figure 3"}],"recommendation":"major_revision","confidential_remarks":"Given that the RT inversion is underdetermined and one of the cited gap values is inconsistent with the reported slope, the quantitative central claim needs substantial revision. I recommend major revision rather than rejection, because the qualitative phase assignments and the DFT benchmark provide a viable path forward if the fitting procedure, parameter values, and uncertainty analysis are supplied. The paper is within the journal's scope."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear X,\n\nThe paper gives the first pressure-resolved optical pump-probe data on Ta2NiSe5 up to 4.7 GPa and uses it to locate the known EI-to-semiconductor (~1 GPa) and semiconductor-to-semimetal (~3 GPa) transitions. The qualitative phase assignments are credible: A1 and tau1 show systematic changes that line up with earlier transport, Raman and XRD. The new quantitative claim is a pressure coefficient of the semi-bandgap, 65 meV/GPa, backed by a constrained-DFT estimate of 70 meV/GPa. That agreement is the most interesting result in the paper.\n\nThe soft spot is the extraction. Delta(P) is inverted from A1 and tau1 through Eq. 2, and the same equation is then used to draw the black curves in Fig. 3. The authors never say how A0, J, K, L are pinned down, and no error bars are given for Delta. So the 65 meV/GPa slope is not robustly determined. There's also a real inconsistency: their Delta(1 GPa)=63 meV gives 2Delta=126 meV, while the paper itself cites 160 meV for the gap at P_C1 from Ref. [48]. Either the RT parameters are not pressure independent or the inversion is biased. This needs to be fixed before the number is usable. The DFT calculation is an external benchmark and does help, but it cannot fully rescue the experimental slope because the DFT constrained-occupation method may share systematic errors.\n\nThe rest is fine. The fluence dependence in Fig. 5 is a sensible check, and the slow relaxation time discussion is appropriately qualitative.\n\nWho is this for? The EI and high-pressure ultrafast communities. A referee should ask for the raw data, the fitted RT parameters with uncertainties, and a reconciliation of the 1 GPa gap. But the dataset is new and the phase boundary confirmation is useful. It deserves peer review, not a desk reject.","headline":"New pressure-resolved ultrafast data on Ta2NiSe5 with credible phase boundaries, but the 65 meV/GPa gap slope is weakened by circular RT inversion and an internal gap inconsistency.","tokens_in":10864,"tokens_out":1934,"would_cite":false,"duration_ms":15695,"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":"Pressure closes the excitonic gap in Ta2NiSe5 at 65 meV/GPa","keywords":["Excitonic insulator","Ta2NiSe5","High pressure","Optical pump-optical probe spectroscopy","Ultrafast carrier dynamics","Rothwarf-Taylor model","Pressure-dependent bandgap","Diamond anvil cell"],"falsifier":"A direct measurement of the optical or electronic gap in Ta2NiSe5 under hydrostatic pressure in the 0–1 GPa range, for example by pressure-dependent absorption, reflectivity-edge, or angle-resolved photoemission, would settle the claim: if the gap does not follow a roughly linear 65 meV/GPa decrease, or if phonon linewidths change enough to shift the Rothwarf–Taylor parameters, the extracted pressure coefficient is an artifact of the model.","tokens_in":9632,"feed_emoji":"⚡","tokens_out":5701,"duration_ms":45957,"temperature":0.7,"pith_summary":"The paper claims that in the excitonic insulator Ta2NiSe5, applied pressure linearly closes the excitonic semi-bandgap at a rate of 65 meV/GPa between ambient pressure and about 1 GPa. The claim is extracted from the pressure dependence of the fast relaxation amplitude and time in optical pump–probe measurements, interpreted through the Rothwarf–Taylor model, and it agrees with a first-principles value of 70 meV/GPa. If correct, this identifies pressure as a continuous tuning knob for the excitonic ground state and locates the insulator-to-semiconductor and semiconductor-to-semimetal boundaries at roughly 1 and 3 GPa. The result matters because it ties a collective many-body phenomenon, exciton condensation, to a simple single-parameter description of the gap.","feed_headline":"Pressure closes the excitonic gap in Ta2NiSe5 at 65 meV/GPa","feed_subtitle":"Ultrafast pump-probe and first-principles calculations agree on how compression drives the excitonic insulator toward a semimetal.","key_machinery":"The load-bearing object is the Rothwarf–Taylor model of photoexcited carriers in a gapped state, originally formulated for superconductors: photoexcitation breaks excitons into free electrons and holes, recombination releases high-frequency phonons, and those phonons either re-break excitons (bottleneck) or decay into low-frequency phonons. Its closed-form expressions for the fast relaxation amplitude $A_1$ and time $\\tau_1$, Eq. (2), depend on the semi-bandgap $\\Delta$ through a Boltzmann factor $\\exp(-\\Delta/k_B T)$; because the experiment runs at fixed temperature, pressure enters only through $\\Delta(P)$. On the theory side, the supporting machinery is a constrained-occupation density-functional calculation in which an electron is promoted to a conduction state at the $\\Gamma$ point and the total-energy difference defines the gap at each pressure. Together the two pieces convert measured relaxation traces into a pressure coefficient for the gap.","core_discovery":"The central discovery is that the fast relaxation channel observed in time-resolved differential reflectivity of Ta2NiSe5 carries the signature of gap closure: both the amplitude $A_1$ and the relaxation time $\\tau_1$ fall as pressure rises through the excitonic-insulator region. Using the Rothwarf–Taylor expressions $A_1 = \\frac{A_0/\\Delta(P)}{1+J\\exp(-\\Delta(P)/k_B T)}$ and $\\tau_1 = \\frac{1}{K + L\\sqrt{\\Delta(P)\\,T}\\exp(-\\Delta(P)/k_B T)}$, the authors assign the entire pressure dependence to the semi-bandgap $\\Delta$ and obtain $\\Delta(P) = 128 - 65P$ meV. A constrained-occupation DFT calculation in a $2\\times2\\times1$ supercell gives a slope of 70 meV/GPa for the same quantity, which the paper takes as close agreement. The same data set shows abrupt changes in $A_1$ and $\\tau_1$ near 1 GPa and again near 3 GPa, which the authors identify with the transition from the excitonic insulator to a semiconductor and then to a semimetal.","pith_inferences":["If the linear gap closure holds, pressure offers a clean route through the BEC–BCS crossover region of the excitonic insulator without chemical doping, since it tunes the exciton binding energy relative to the band gap.","The assumption of pressure-independent Rothwarf–Taylor parameters could be tested directly by combining the same pump–probe traces with pressure-dependent Raman linewidths: a correlated change in $K$ or $L$ would mean the extracted $\\Delta(P)$ partly absorbs electron–phonon coupling changes.","Cooling the same experiment toward the 328 K transition temperature would separate the Boltzmann factor $\\exp(-\\Delta/k_B T)$ from the gap itself, giving a second, independent estimate of $\\Delta(P)$.","A natural extension is to repeat the analysis at higher pump fluences; the paper's fluence dependence of $\\tau_1$ suggests the bottleneck efficiency changes with carrier density, which would modify the extracted $\\Delta$ if $L$ depends on fluence."],"forward_implications":["Within the excitonic-insulator phase, the semi-bandgap obeys $\\Delta(P) = 128 - 65P$ meV, meaning compression of about 1 GPa nearly halves the gap and destabilizes the EI phase.","The phase boundaries at roughly 1 GPa (EI to semiconductor) and 3 GPa (semiconductor to semimetal), previously inferred from transport and Raman data, appear directly in the ultrafast relaxation amplitudes and times.","Above 3 GPa the fast relaxation time shortens again, consistent with a semimetal whose growing carrier density opens extra relaxation channels.","The slow relaxation channel, assigned to hot-phonon decay, also speeds up with pressure in the EI phase and correlates with pressure-broadened Raman linewidths.","Agreement between the experimental 65 meV/GPa and calculated 70 meV/GPa supports using constrained-occupation DFT as a predictor for pressure-driven gap closure in other candidate excitonic insulators."],"supporting_citations":[{"why":"Applies the Rothwarf–Taylor analysis to Ta2NiSe5 and assigns the slow relaxation component to hot-phonon thermalization, giving the model basis for the fast amplitude.","marker":"[21]"},{"why":"Original Rothwarf–Taylor bottleneck model for photoexcited quasiparticles in a gapped state, from which the amplitude and relaxation expressions descend.","marker":"[28]"},{"why":"Derives the gap- and temperature-dependent expression for the relaxation amplitude $A_1$ used in Eq. (2).","marker":"[29]"},{"why":"Derives the relaxation time $\\tau_1$ expression with the square-root and exponential gap dependence used in Eq. (2).","marker":"[30]"},{"why":"High-pressure XRD and resistivity study that established the ~1 GPa and ~3 GPa phase boundaries and the structural sliding at 3 GPa.","marker":"[25]"},{"why":"High-pressure Raman study of Ta2NiSe5 whose phonon linewidth trends are compared with $\\tau_2$ and which supplies the DFT Hubbard-$U$ setting.","marker":"[27]"},{"why":"Provides the ~160 meV semiconductor bandgap near 1 GPa used as an anchor for the pressure-dependent gap.","marker":"[48]"},{"why":"Gives ambient-pressure estimates of the excitonic gap and valence-band flatness used to justify the initial $\\Delta_0$ value.","marker":"[16]"}],"fun_headline_variants":["Pressure closes Ta2NiSe5 excitonic gap at 65 meV/GPa","Experiment and theory agree: pressure closes Ta2NiSe5 gap at 65 meV/GPa","Ta2NiSe5 gap shrinks 65 meV per GPa under pressure","Pressure drives Ta2NiSe5 from excitonic insulator to semimetal","Two phase transitions in Ta2NiSe5 as pressure closes the gap"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The weakest point is the assumption that the Rothwarf–Taylor parameters $A_0$, $J$, $K$, and $L$ are independent of pressure between 0 and 1 GPa, so that every change in $A_1$ and $\\tau_1$ is attributed to the gap $\\Delta(P)$.","fun_headline_variants_meta":{"raw":{"variants":["Pressure closes Ta2NiSe5 excitonic gap at 65 meV/GPa","Experiment and theory agree: pressure closes Ta2NiSe5 gap at 65 meV/GPa","Ta2NiSe5 gap shrinks 65 meV per GPa under pressure","Pressure drives Ta2NiSe5 from excitonic insulator to semimetal","Two phase transitions in Ta2NiSe5 as pressure closes the gap"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001138,"raw_usage":{"total_tokens":4729,"prompt_tokens":955,"completion_tokens":3774,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":571,"completion_tokens_details":{"reasoning_tokens":3665}},"tokens_in":571,"tokens_out":3774,"duration_ms":22598,"temperature":1.0,"reasoning_tokens":3665,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T11:36:02.670560+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct measurement of the optical or electronic gap in Ta2NiSe5 under hydrostatic pressure in the 0–1 GPa range, for example by pressure-dependent absorption, reflectivity-edge, or angle-resolved photoemission, would settle the claim: if the gap does not follow a roughly linear 65 meV/GPa decrease, or if phonon linewidths change enough to shift the Rothwarf–Taylor parameters, the extracted pressure coefficient is an artifact of the model.","supporting_citations":[{"cited_title":"Journal of Physics: Condensed Matter , 30(30):305602, 2018","cited_arxiv_id":null,"evidence_quote":"Applies the Rothwarf–Taylor analysis to Ta2NiSe5 and assigns the slow relaxation component to hot-phonon thermalization, giving the model basis for the fast amplitude."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Original Rothwarf–Taylor bottleneck model for photoexcited quasiparticles in a gapped state, from which the amplitude and relaxation expressions descend."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Derives the gap- and temperature-dependent expression for the relaxation amplitude $A_1$ used in Eq. (2)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Derives the relaxation time $\\tau_1$ expression with the square-root and exponential gap dependence used in Eq. (2)."},{"cited_title":"IU- CrJ, 5(2):158–165, 2018","cited_arxiv_id":null,"evidence_quote":"High-pressure XRD and resistivity study that established the ~1 GPa and ~3 GPa phase boundaries and the structural sliding at 3 GPa."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"High-pressure Raman study of Ta2NiSe5 whose phonon linewidth trends are compared with $\\tau_2$ and which supplies the DFT Hubbard-$U$ setting."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the ~160 meV semiconductor bandgap near 1 GPa used as an anchor for the pressure-dependent gap."}],"review_version":1}