{"id":"1cbe1e30-333a-417a-9fd6-fe1ff9b776f7","arxiv_id":"2608.08454","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Finite-temperature spin-adapted ROKS and TDDFT are defined by Boltzmann-averaging integer high-spin component responses, and applied to assign the 425 nm thermally activated band of TTM-TTM to T1->T6.","lead":"Scientists have combined two ideas in quantum chemistry, finite-temperature thermal averaging and spin-preserving calculations, into one framework for open-shell molecules. The authors use it to argue that a temperature-activated absorption band in a diradicaloid comes from excited triplet states.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eqs. 39-41 assert by 'linearity of the trace' that the thermal response is a weighted average of component SA-TDDFT matrix pencils, but M,N are not traces over a common Hamiltonian and component Fock matrices differ; the ensemble response is therefore underived.","rationale":"The reader's weakest_assumption identifies the same load-bearing gap: the claim that the thermal linear response equals the weighted average of component SA-TDDFT matrix pencils, justified only by trace linearity. My reading of Eqs. (8), (39)-(41) and the surrounding text confirms that no derivation from ensemble linear-response theory is given. The ground-state FT-SA-ROKS part is internally coherent and reduces properly at T→0; the problem is specifically the response section. This is not an internal inconsistency, but an unproven approximation on which the numerical interpretation depends. The concern is testable with the provided IQC binary and SI input files, so a conditional verdict remains appropriate rather than rejection. If the concrete test passes, the method gains support; if it fails, the claimed consistency and the 425 nm assignment would need substantial revision.","tokens_in":9078,"tokens_out":12547,"duration_ms":151142,"concrete_test":"Recompute the 290 K TTM-TTM spectrum two ways: (i) solve the averaged pencil Eq. (41) as reported; (ii) retain the same weights, diagonalize each component's M_I and N_I separately, and form the Boltzmann-weighted sum of oscillator strengths as a function of energy. If the bright peak near 2.98 eV is not reproduced as a weighted component peak, or if a weighted component peak is missing from the averaged spectrum, the trace-linearity claim is falsified. A further check with the noninteracting K=0 limit would show whether the difference arises from the occupation averaging or from the component-dependent Fock/kernel terms.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central response step is Eqs. (39)-(41), where the finite-temperature SA-TDDFT pencil is the weighted average of component pencils. The paper justifies this by 'linearity of the trace (eq. 8)', but Eq. (8) only says Γθ is a weighted sum of projectors. For this to imply the matrix-pencil average, the component matrices M_I and N_I would have to be the expectation values, in each component, of one common operator with a common Hamiltonian. They are not: in the ROKS reference each component has its own spin-Fock matrices F^{I,σ} and its own density D_I, and the SA-TDDFT M_I and N_I depend on these quantities. They also contain the XC kernel, which is a density functional; the kernel at the ensemble density is not the weighted average of kernels at component densities. Moreover, for a fixed incoherent mixture the exact linear response is the Boltzmann-weighted sum of component response functions, so the observable spectrum is a sum of component spectra. The paper explicitly says the roots of Eq. (41) are not averages of component roots; without a derivation this means the bright poles of the averaged pencil need not coincide with the thermal ensemble's absorption peaks. The T1→T6 assignment at 2.983 eV therefore rests on an unproven approximation to the linear response.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes finite-temperature extensions of restricted open-shell Kohn-Sham (ROKS) theory and spin-adapted time-dependent density functional theory (TDDFT). For FT-SA-ROKS, a canonical ensemble is built from integer high-spin ROKS Slater determinants that share the same spin quantum numbers, with Boltzmann weights over component energies; the orbital gradient is the weighted sum of component gradients, and the zero-temperature single-component limit recovers ordinary high-spin ROKS. For FT-SA-TDDFT, each integer component contributes a zero-temperature spin-adapted TDDFT matrix pair (M_I, N_I), which is embedded into a common response basis and averaged with the same Boltzmann weights; the resulting matrix pencil is diagonalized to yield finite-temperature excitation energies, transition densities, and oscillator strengths. The method is applied to the diradicaloid TTM-TTM, and the calculated T1->T6 transition at 2.983 eV (about 416 nm, oscillator strength 0.00995) is used to interpret the experimentally observed temperature-activated absorption near 425 nm. The ground-state free-energy construction is clean, but the central response step is asserted rather than derived from ensemble linear-response theory, and the numerical application relies on a separately supplied singlet-triplet gap.","tokens_in":9423,"tokens_out":8587,"duration_ms":103708,"significance":"If the central response equations were properly derived, this would fill a genuine gap: a spin-adapted finite-temperature TDDFT for open-shell systems with a correct zero-temperature limit. The FT-SA-ROKS part of the paper is transparent and internally consistent, and the authors are to be credited for shipping the IQC package, reporting raw data, and providing explicit, falsifiable predictions (e.g., the T1->T6 transition at 2.983 eV and its oscillator strength). However, the ensemble linear-response construction is not established: Eqs. (39)-(41) average component response matrices without deriving the response of the thermal ensemble, and the eigenvalues of an averaged matrix pencil need not coincide with the poles of the exact thermal response. Since the numerical assignment of the 425 nm absorption depends directly on this step, the central claim is currently supported only conditionally. The paper is well written and the zero-temperature reduction is a strength, but the response formalism requires a derivation or an explicit validation before the method can be accepted.","major_comments":[{"comment":"The finite-temperature response matrices are asserted to be the weighted averages of component matrices by 'the linearity of the trace (eq. 8)'. That step is not justified, because Eq. (8) concerns only the density operator Gamma_theta, whereas M_I and N_I are response matrices, not expectation values of a common operator on Gamma_theta; they depend on component Fock matrices, component densities, and the exchange-correlation kernel evaluated at each component density. The exact linear response of an incoherent ensemble is the Boltzmann-weighted sum of the component response functions, whose poles are the union of the component excitation energies, whereas the roots of Eq. (41) are neither the component roots nor necessarily the poles of the ensemble response. Please provide a derivation of Eqs. (39)-(40) from ensemble linear-response theory, or explicitly present the averaging as an approximation and benchmark it against the component-resolved response.","section":"§2.2, Eqs. (39)-(41)"},{"comment":"The paper reports Delta_EST = 0.194 eV and uses it in the two-multiplet population model of Eq. (72), but it does not state how the singlet energy S0 was obtained. Since the FT-SA-ROKS ensemble is restricted to high-spin components with S = S_i, it cannot by itself produce an open-shell singlet. Please specify the level of theory, functional, basis, and geometry used for S0 and for the lowest triplet, and confirm that both are computed consistently; without this, the temperature dependence of the absorption interpretation is incomplete.","section":"§3, Table 1 and Eq. (72)"},{"comment":"The numerical results depend on the active-window composition (highest C, all O, and two lowest V orbitals) and on the metric truncation threshold lambda_cut = 10^-10 max(1, max_j |lambda_j|), but no convergence or sensitivity tests are reported. Since the finite-temperature roots in Table 1 are obtained from an averaged pencil after embedding and truncation, the sensitivity of the T1->T6 energy and oscillator strength to these choices should be quantified before the 425 nm assignment can be considered robust.","section":"§2.2, Eqs. (42)-(49) and §3"},{"comment":"The ensemble is restricted to components with a fixed spin quantum number S, so the theory does not describe a thermal mixture across different spin multiplets. The application compensates with the ad hoc two-multiplet model of Eq. (72), which is outside the presented formalism. This limitation should be stated explicitly in the abstract or introduction, since the current wording suggests a unified finite-temperature framework for both ground and excited states.","section":"§2.1, Eq. (9)"}],"minor_comments":[{"comment":"The caption states that the table reports energies 'at 77 and 290 K', but only one set of energies is displayed; please report the 77 K and 290 K values separately or state explicitly that they are identical to the precision shown.","section":"Table 1"},{"comment":"The notation for the transition coupling in Eq. (61) and the subsequent vectors in Eqs. (62)-(63) is not self-contained; a reference to the original definitions in SA-TDDFT would improve readability for readers not familiar with the author's prior work.","section":"§2.2, Eqs. (61)-(63)"},{"comment":"The geometry was optimized at U-M06-2X/def2-SVP+D3 while the response calculations use BHHLYP/cc-pVDZ; since the singlet-triplet gap and excited-state energies can be sensitive to geometry, the choice of a fixed geometry from a different functional should be justified.","section":"§3"},{"comment":"There are minor typographical issues, including the stray 'P APER' at the beginning of the text and missing spacing in 'Keywords:density'; the manuscript should be copyedited.","section":"Manuscript text"}],"recommendation":"major_revision","confidential_remarks":"The critical question for acceptance is whether Eqs. (39)-(41) can be derived from an ensemble linear-response framework or validated as a controlled approximation. If the authors cannot supply such a derivation or a benchmark showing that the averaged pencil reproduces the poles of the thermal response, the central numerical assignment will not stand. The paper relies heavily on the author's own prior SA-TDDFT work and the IQC package, which makes independent verification harder; however, this is not a reason for rejection by itself."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read it, and read the stress test. The stress test is right. The paper does something new—finite-temperature spin-adapted ROKS and TDDFT, with a clean ground-state theory and useful new response machinery. The ensemble free energy (Eq. 12), the envelope-theorem orbital gradient (Eq. 14), and the zero-temperature single-component limit are all properly built. The finite-temperature transition densities and oscillator strengths (Eqs. 51–71) are new working equations. The numerical application is not fitted to the experimental band: Δ_EST = 0.194 eV and the T1→T6 energy are computed, not adjusted. The code and input files are available, which is helpful and makes the failure modes testable.\n\nThe load-bearing problem is in Sec. 2.2. Equations 39–41 construct the thermal response by Boltzmann-weighting the component matrix pairs and justify it by the linearity of the trace. But M and N are not expectation values of a common operator. Each component has its own Fock matrix and density, and the XC kernel is density-dependent. Linearity of the trace gives the density operator, not the response matrices. The exact linear response of a fixed incoherent mixture is a sum of component response functions; an averaged pencil has roots that are not averages of component roots and need not align with the mixture's absorption peaks. So the T1→T6 assignment at 2.983 eV rests on an approximation that is not derived.\n\nSmaller issues: the active-window composition and weight cutoff are practical choices, but sensitivity to them is not tested. The comparison to the observed 425 nm band is wavelength-only, with no predicted intensity against experiment; with PT1 ≈ 1.3×10^-3 at 290 K, the absorption may be weak. The two-multiplet population estimate (Eq. 72) is crude, though that's peripheral. Citations are fine; leaning on the author's SA-TDDFT work is appropriate since it is prior published material.\n\nWho gets value: researchers working on finite-temperature open-shell DFT or on temperature-dependent diradicaloid spectroscopy. The paper deserves a serious referee. The central query should be: derive the thermal response properly (e.g., from the finite-temperature TDDFT Kubo formula for a fractional-occupation KS system, which will either validate the averaged-pencil form or replace it), or explicitly scope it as an approximation and test it against a more direct construction. I would send it to review.","headline":"A solid ground-state extension with a genuine gap in the response derivation; the T1→T6 assignment rests on an unproven averaged-pencil approximation.","tokens_in":9908,"tokens_out":7171,"would_cite":true,"duration_ms":82008,"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":"Boltzmann-averaging high-spin components puts temperature into spin-adapted DFT.","keywords":["finite-temperature DFT","spin-adapted TDDFT","ROKS","thermal ensemble","diradicaloid","triplet-triplet transitions","oscillator strength","spin purity"],"falsifier":"Compute the TTM–TTM triplet absorption spectrum using a formally exact finite-temperature linear-response TDDFT formulation that builds the retarded response function from explicitly Boltzmann-weighted many-body states, and compare the resulting excitation energies and oscillator strengths with the Boltzmann-averaged matrix-pencil roots; if they differ beyond numerical precision, the central averaging identity is false.","tokens_in":8839,"feed_emoji":"⚛️","tokens_out":10962,"duration_ms":97204,"temperature":0.7,"pith_summary":"This paper proposes a way to include temperature in spin-adapted density functional theory for open-shell molecules without breaking spin symmetry. It builds a canonical ensemble from integer high-spin restricted open-shell Kohn–Sham (ROKS) components, each weighted by a Boltzmann factor, and then averages the spin-adapted TDDFT response matrix pairs of those components in a common orbital basis. In the zero-temperature single-component limit the theory reduces to ordinary high-spin ROKS and the corresponding spin-adapted TDDFT. Applied to the diradicaloid TTM–TTM, the calculation assigns the thermally activated absorption near 425 nm to T1→T6 and other allowed triplet–triplet transitions, whose intensities grow as the triplet population rises with temperature.","feed_headline":"Boltzmann-averaged spin states bring temperature into open-shell DFT","feed_subtitle":"New finite-temperature ROKS/TDDFT assigns the 425 nm band of a diradicaloid to thermally populated triplet transitions.","key_machinery":"The central object is the thermal ensemble built from integer high-spin ROKS components $|\\Phi^{S,S}_I\\rangle$, each assigning hard closed/open/virtual labels to spatial orbitals and each carrying Boltzmann weight $w_I \\propto \\exp(-\\beta E_I)$. Because every component has the same $\\hat{S}^2$ eigenvalue, the ensemble density operator is spin-pure. The argument is carried by embedding the zero-temperature spin-adapted TDDFT response matrix pair $(M_I, N_I)$ of every retained component into a common global response basis and weighting them by $w_I$; the averaged matrix pencil then yields thermal excitation energies and, via the derived transition density matrices and oscillator strengths, thermal absorption intensities.","core_discovery":"The central claim is that a finite-temperature spin-adapted description of ground and excited states follows from Boltzmann-averaging ordinary zero-temperature high-spin ROKS determinants and their spin-adapted response matrices. The paper defines the thermal ensemble density operator as a weighted sum of highest-weight $|SS\\rangle$ determinants with the same spin quantum number, so the ensemble commutes with $\\hat{S}^2$ and remains spin-pure. For excited states, each component supplies an ordinary spin-adapted TDDFT matrix pair $(M_I, N_I)$; these are embedded into one global spatial-orbital basis and averaged to form the thermal matrix pencil $M^\\theta Z = \\omega N^\\theta Z$. The roots are eigenvalues of the averaged pencil, not averages of component roots. In the zero-temperature single-component limit, the method returns to standard high-spin ROKS and the corresponding spin-adapted TDDFT.","pith_inferences":["A testable extension would apply the same Boltzmann-averaged matrix-pencil construction to other open-shell chromophores with measured temperature-dependent absorption, comparing predicted thermally activated band positions and intensities with experiment.","If the linear-response averaging is not exact, a formal derivation starting from the finite-temperature linear-response function would clarify whether this construction is an approximation or an identity; comparing its roots against an exact thermal response calculation on a small molecule would settle the question.","The two-multiplet estimate of triplet population implies that the 425 nm band intensity should track the triplet population with roughly the same activation energy as the 0.194 eV singlet–triplet gap, a scaling that intensity measurements over temperature could test.","The active-window restriction (highest-energy closed orbital, all open orbitals, two lowest virtual orbitals) means accuracy depends on the window choice; expanding the window on small systems would probe convergence of the thermal excitation energies."],"forward_implications":["If the averaging is correct, finite-temperature excitation spectra of open-shell systems can be computed at ordinary zero-temperature SA-TDDFT cost, one component at a time, without enumerating all ensemble states beyond a small active window.","Spin purity is preserved by construction, so thermal spectra of diradicals and other open-shell species avoid the spin contamination that plagues finite-temperature broken-symmetry approaches.","The theory supplies thermal transition density matrices and oscillator strengths, so temperature-dependent absorption intensities can be interpreted quantitatively through the Boltzmann population of the reference spin state.","For TTM–TTM specifically, the paper identifies T1→T6 at 2.983 eV (416 nm), with oscillator strength 0.00995, as the main carrier of the thermally activated 425 nm band, with T1→T3 and T1→T4 contributing in the blue-visible region.","The zero-temperature limit recovers ordinary ROKS and spin-adapted TDDFT, making the method a controlled extension rather than a separate formalism."],"supporting_citations":[{"why":"Establishes the thermal density-functional description of equilibrium ensembles that motivates Boltzmann-weighted components.","marker":"[8]"},{"why":"Provides the ensemble-DFT formalism for unequally weighted states used to justify the Boltzmann component weights.","marker":"[10]"},{"why":"Supplies the restricted open-shell SCF equations that each integer high-spin component satisfies.","marker":"[15]"},{"why":"Defines the spin-adapted TDDFT response matrix pair applied to each component.","marker":"[16]"},{"why":"Gives the spin-adapted open-shell RPA that underlies the SA-TDDFT response construction.","marker":"[17]"},{"why":"Provides the spin-flip tensor equation-of-motion CIS used in the spin-adaptation of the response.","marker":"[18]"},{"why":"Supplies the tensor equation-of-motion formalism for spin-adapted excitation operators.","marker":"[19]"},{"why":"Reports the variable-temperature UV/Vis 425 nm band and the thermally accessible triplet of TTM–TTM that the calculation interprets.","marker":"[22]"},{"why":"Supplies the optimized TTM–TTM geometry on which the finite-temperature calculations are performed.","marker":"[23]"}],"fun_headline_variants":["Boltzmann spin states heat open-shell DFT","Spin-pure thermal DFT: Boltzmann-averaged open-shell states","Finite-T spin-adapted ROKS/TDDFT: thermal ensemble","Heated open-shell DFT: spin-adapted ROKS and TDDFT","Thermal spin states for open-shell DFT: 425 nm explained"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument assumes that a warm molecule's excited-state behavior is exactly the sum of the behaviors of its cold high-spin components, each weighted by thermal probability, with no additional mixing between components; if that simple averaging is wrong, the predicted 425 nm explanation falls apart.","fun_headline_variants_meta":{"raw":{"variants":["Boltzmann spin states heat open-shell DFT","Spin-pure thermal DFT: Boltzmann-averaged open-shell states","Finite-T spin-adapted ROKS/TDDFT: thermal ensemble","Heated open-shell DFT: spin-adapted ROKS and TDDFT","Thermal spin states for open-shell DFT: 425 nm explained"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001874,"raw_usage":{"total_tokens":7379,"prompt_tokens":994,"completion_tokens":6385,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":610,"completion_tokens_details":{"reasoning_tokens":6293}},"tokens_in":610,"tokens_out":6385,"duration_ms":45239,"temperature":1.0,"reasoning_tokens":6293,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T04:35:46.393087+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the TTM–TTM triplet absorption spectrum using a formally exact finite-temperature linear-response TDDFT formulation that builds the retarded response function from explicitly Boltzmann-weighted many-body states, and compare the resulting excitation energies and oscillator strengths with the Boltzmann-averaged matrix-pencil roots; if they differ beyond numerical precision, the central averaging identity is false.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the restricted open-shell SCF equations that each integer high-spin component satisfies."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the spin-adapted open-shell RPA that underlies the SA-TDDFT response construction."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the spin-flip tensor equation-of-motion CIS used in the spin-adaptation of the response."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the tensor equation-of-motion formalism for spin-adapted excitation operators."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports the variable-temperature UV/Vis 425 nm band and the thermally accessible triplet of TTM–TTM that the calculation interprets."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the optimized TTM–TTM geometry on which the finite-temperature calculations are performed."}],"review_version":1}