{"id":"a21fec83-8782-43a1-9ef2-b7c52de4e91e","arxiv_id":"2502.04800","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Spin-free orbital entropy, pair entropy, and mutual information are defined, shown to be invariant under M_s, and applied to CH2 dimers and iron-sulfur complexes.","lead":"This paper defines new spin-free versions of orbital entropy, pair entropy, and mutual information that ignore whether an orbital electron has up or down spin, making the analysis independent of the spin projection M_s. The goal is to help chemists separate correlation caused by spin coupling from genuine strong correlation in molecules such as iron-sulfur clusters.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed inequality S̃_ij ≤ S_ij is false: a one-electron two-orbital superposition gives S̃_12 = ln 2 > S_12 = 0, so the spin-free pair entropy is not always a reduction of the original.","rationale":"The reader's primary weakest assumption concerned the factorization in Eq. (30) used to prove M_s invariance. That is a genuine proof gap: the factorization holds for a single spin-coupled configuration but not for a general multiconfigurational spin eigenfunction. However, the M_s invariance itself is likely true because each spin-free orbital/pair probability is the expectation of a spin-scalar operator, so it should be identical for all components of a multiplet; the numerical examples support this. The more load-bearing defect is the false pair-entropy inequality. The paper's strongest claim explicitly includes the statement that spin-free counterparts cannot be bigger than the original ones. This is demonstrably false for S̃_ij, and the error is not a matter of convention: it follows directly from Eqs. (19)-(20) applied to a simple spin-pure one-electron state. This does not destroy the practical value of the proposed spin-free analysis, and the M_s invariance and DPI-based mutual-information bound may survive, so the correct disposition is still conditional rather than rejection. The concrete counterexample is the decisive check; if implemented, it settles the issue immediately.","tokens_in":17253,"tokens_out":9586,"duration_ms":105227,"concrete_test":"Perform an analytical two-orbital, one-electron test: take |Ψ⟩ = (|↑0⟩ + |0↑⟩)/√2, form ρ_12 = |Ψ⟩⟨Ψ|, diagonalize it, and apply Eqs. (19)-(20) to obtain S̃_12. The state is a valid spin doublet (S = 1/2, M_s = 1/2). Since ρ_12 has a single eigenvalue 1, S_12 = 0; the spin-free weights are 1/2 in sectors |10⟩ and |01⟩, so S̃_12 = ln 2. If the computation returns this value, the claimed inequality S̃_ij ≤ S_ij is definitively refuted. No quantum chemistry package is needed: the 16×16 density matrix has support only on the |↑0⟩, |0↑⟩ subspace.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The theorem that the spin-free pair entropy cannot exceed the original pair entropy (Section 2, discussion after Eq. (23)) is false. The map in Eq. (20), ω̃_q = Σ_p ω_p Σ_{k∈B_q} c²_pk, is a column-stochastic map from eigenbasis probabilities to spin-free occupation probabilities. Unlike the single-orbital case in Eq. (18), whose diagonal coarse-graining is a deterministic merging and cannot increase entropy, this pair map is not a merging: an eigenvector can have weight in several different spin-free sectors, and a column-stochastic map can increase Shannon entropy. A concrete counterexample is the one-electron two-orbital spin-doublet state |Ψ⟩ = (|↑0⟩ + |0↑⟩)/√2. The pair density matrix ρ_12 is pure, so S_12 = 0; Eq. (20) yields ω̃_{|10⟩} = 1/2 and ω̃_{|01⟩} = 1/2, hence S̃_12 = ln 2 > 0 = S_12. This is a valid Fock-space wave function of the form (4) and a spin eigenfunction. Thus the paper's claim that spin-free counterparts are always no larger than the spin-including ones is incorrect for pair entropy, and the interpretation of any positive difference as spin-coupling correlation is not generally justified. The mutual-information inequality Ĩ ≤ I may still follow from the data processing inequality, but the pair-entropy inequality needs to be withdrawn or re-stated under restricted conditions.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces spin-free analogues of the standard orbital entropy, pair entropy, and mutual information for quantum-chemical wave functions. The spin-free quantities are defined by coarse-graining the four-dimensional orbital basis {empty, spin-up, spin-down, doubly occupied} to a three-dimensional occupation-number basis, and the spin-free pair entropy is defined through spin-summed eigenvalues of the pair density matrix. The central claims are that (i) the spin-free entropy, pair entropy, and mutual information cannot exceed their spin-including counterparts; (ii) the spin-free quantities are invariant with respect to the spin projection M_s of a given spin multiplet; and (iii) comparing spin-free with spin-including quantities separates static correlation due to spin couplings from genuine strong correlation. The approach is illustrated on a non-interacting dimer of CH2 diradicals and on [Fe(SCH3)4]^- and [Fe2S2(SCH3)4]^2- complexes computed with DMRG. Numerical results show that the spin-free quantities are indeed invariant across M_s in the studied examples and that the comparison with spin-including values can simplify correlation diagrams.","tokens_in":17605,"tokens_out":4049,"duration_ms":47452,"significance":"If the formal claims were fully correct, the proposed spin-free analysis would be a useful, cheap supplement to standard DMRG-based entanglement analysis, particularly for open-shell transition-metal complexes where low-M_s components of high-spin states contain many spin-coupling-dominated determinants. The paper's main positive contribution is the clear definition of spin-free orbital and pair entropies and the extensive numerical demonstration on realistic iron-sulfur systems. The observation that spin-free quantities filter out much of the spin-coupling-dominated correlation in these systems is interesting and potentially valuable. However, the theoretical foundation contains a demonstrably false inequality and an incomplete invariance proof, both of which are load-bearing for the advertised interpretation. The numerical work itself is honest about convergence limitations, and the single-orbital entropy inequality is correctly proved by concavity; the same cannot be said for the pair entropy claim.","major_comments":[{"comment":"The claim that the spin-free pair entropy is never larger than the original pair entropy, S̃_ij ≤ S_ij, is false. The argument 'by the same argument as for orbital entropy' does not apply, because Eq. (20) defines a column-stochastic map, not a deterministic merging of eigenvalues. A concrete counterexample is the one-electron two-orbital spin doublet |Ψ> = (|↑0> + |0↑>)/√2. Its pair density matrix is pure, so S_ij = 0. Applying Eq. (20) with the single eigenvalue ω = 1 and eigenvector (1/√2)(|↑0> + |0↑>) gives ω̃_{|10>} = 1/2 and ω̃_{|01>} = 1/2, hence S̃_ij = ln 2 > S_ij. This is a valid wave function of the form (4) and a spin eigenfunction, so the counterexample cannot be dismissed as unphysical. The inequality in Eq. (23) must therefore be withdrawn or restated under additional restrictions, and the interpretation that a positive difference S_ij - S̃_ij always quantifies spin-coupling static correlation is not generally justified.","section":"Section 2, after Eq. (23)"},{"comment":"The proof of M_s invariance assumes that the total N-electron wave function factorizes as Ψ = Ψ_P(N,S) Ψ_S(N,S,M_s), as stated in Eq. (30). This factorization holds for a single spin-adapted configuration but not for a general multiconfigurational spin eigenfunction, which is a linear combination of such products with different spatial functions. The derivation that follows, leading to Eq. (39), therefore does not establish the advertised invariance for general CAS/DMRG wave functions. The numerical evidence in Tables 2, 3, and 5 is supportive but approximate, and Table 5 itself shows small deviations for higher DMRG roots. The invariance claim needs either a proof that avoids the product assumption, or a clear restriction of the theorem to the cases where the factorization is valid.","section":"Section 2, Eqs. (30)-(39)"},{"comment":"The derivation of Ĩ_ij ≤ I_ij from the classical data processing inequality is not rigorous as written. The argument states that the spin summation (20) is a function mapping between two probability distributions, but Eq. (20) is not a deterministic function of the original eigenvalue index p: it mixes contributions from different eigenvectors through the coefficients c_pk. A stochastic map can also obey a data-processing inequality under suitable conditional-independence conditions, but that additional structure is not demonstrated in the manuscript. The mutual-information bound may be true, and it may follow from a more careful argument, but the proof given here is incomplete.","section":"Section 2, after Eq. (29)"}],"minor_comments":[{"comment":"There is a typo: 'measures' is written as 'mesures'.","section":"Section 2, paragraph after Eq. (3)"},{"comment":"The text says 'an equilateral triangular geometry with bond lengths 1.121 Å and angle 152.7°'; an equilateral triangle has all angles 60°, so the description is inconsistent and should be clarified.","section":"Section 3, first paragraph"},{"comment":"For the (CH2)2 rows, some entries in parentheses contain two values (e.g., 0.040,0.693); the caption says the original entropies may differ for equivalent orbitals, but the table would be clearer if columns for the two monomer orbitals were shown explicitly.","section":"Table 1 and Section 4.1"},{"comment":"The phrases 'The bar graphs shows' and 'the color-coded size of the mutual information matrix elements is displayed' contain grammatical errors; please rephrase.","section":"Figure captions 3-7"},{"comment":"The statement that the spin-free mutual information 'vanishes' in the open-shell 4x4 subspace is interesting but would be easier to evaluate if the actual numerical values or a representative matrix were given, since the total Itot values in Table 5 are nonzero.","section":"Section 4.1"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within the scope of the journal and addresses a topic of practical interest for DMRG users. The main reason for major revision is not the numerical study, which is careful, but the false pair-entropy inequality and the incomplete M_s-invariance proof, both of which are central to the paper's interpretation. I see a clear path to a publishable revision: restate or restrict the pair-entropy inequality, provide a more general invariance argument or clearly label the invariance as an observed numerical property, and tighten the data-processing argument. I do not recommend rejection because the tool itself appears useful and the numerical demonstrations are substantial."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague — this paper is worth knowing about, but read the theory section with suspicion. The spin-free orbital entropy is a simple and sensible coarse-graining, and the pair-level construction in Eqs. (20)-(25) is genuinely new. The numerical demonstration of Ms-invariance on the CH2 dimer and iron-sulfur complexes is convincing in the examples, and the comparison of spin-free vs spin-including quantities is a useful pragmatic diagnostic for separating spin-coupling static correlation from other strong correlation. The trend in total spin-free entropy with decreasing spin is clearly presented.\n\nThe soft spots are real and they are in the formal claims. The assertion that the spin-free pair entropy is never larger than the original pair entropy is false. The map in Eq. (20) is not the same kind of merging as the single-orbital case; it is column-stochastic, and column-stochastic maps can increase Shannon entropy. A one-electron two-orbital doublet state |Ψ⟩=(|↑0⟩+|0↑⟩)/√2 gives S_12=0 (the pair RDM is pure) while ω̃_{10}=ω̃_{01}=1/2, so S̃_12=ln2. That directly contradicts the claim after Eq. (23) and the corresponding statement in the conclusions. The inequality might be repairable under restricted conditions, but as stated it is wrong, and the interpretive claim that any positive difference reflects spin coupling needs qualification.\n\nSecond, the proof of Ms-invariance rests on Eq. (30), the factorization of the total wave function into a spatial part and a spin part. That holds for a single spin-adapted configuration, not for a general multiconfigurational spin eigenfunction, which is a sum of such products. The numerical Ms-invariance in the examples is suggestive, and I suspect the statement is true in reasonable generality, but the proof as written does not cover it. This is a fixable gap.\n\nThe paper is honest about its scope and the numerical work is adequate for qualitative analysis. The authors use DMRG with moderate bond dimensions and acknowledge convergence limitations. I would like to see this in the literature after the formal issues are fixed. The tool is cheap and will likely be used.\n\nFor peer review: send it out, but the referee should insist on correcting the pair-entropy inequality and reworking the Ms-invariance proof. The central concept and applications are worth a round of revision.","headline":"Useful spin-free orbital entanglement diagnostics, but the central pair-entropy inequality is false and the Ms-invariance proof over-reaches; the idea survives, the theorems need rewriting.","tokens_in":18065,"tokens_out":3350,"would_cite":false,"duration_ms":35845,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper introduces spin-free orbital entropy, pair entropy, and mutual information that are invariant with respect to the spin projection component and can separate static correlation due to spin couplings from genuine strong correlation.","keywords":["orbital entropy","pair entropy","mutual information","spin-free","DMRG","multiconfigurational","spin projection invariance","iron-sulfur complexes"],"falsifier":"Compute the spin-free orbital entropy and mutual information for a small multiconfigurational state that is a linear combination of two spin-adapted configuration state functions with different spatial parts but the same total spin, for example a complete-active-space wave function for a diradical with two active orbitals. If the spin-free quantities differ between the maximum and lower spin-projection components, the invariance claim fails outside the single-factorization regime.","tokens_in":17030,"feed_emoji":"🧪","tokens_out":4394,"duration_ms":42609,"temperature":0.7,"pith_summary":"The paper introduces spin-free versions of orbital entropy, pair entropy, and mutual information, in which a singly occupied orbital is treated as a single state regardless of whether the electron has up or down spin. The central claim is that these spin-free quantities are invariant under the choice of the spin projection component for a given spin multiplet, while the standard spin-including quantities are not. Comparing spin-free with spin-including measures then separates static correlation caused by spin coupling from genuine strong correlation caused by multiconfigurational character. The paper proves that the spin-free measures can never exceed their spin-including counterparts and demonstrates the diagnostic on a model non-interacting dimer of triplet diradicals and on iron-sulfur complexes.","feed_headline":"Spin-free entropies separate spin coupling from true strong correlation","feed_subtitle":"By lumping up and down spin states, orbital entropies become independent of spin projection and expose genuine multiconfigurational…","key_machinery":"The central object is the spin-free orbital basis, which identifies the two singly occupied spin states of an orbital as a single microstate, and the corresponding spin-summed eigenvalues of the pair density matrix, defined by projecting the pair reduced density matrix's eigenstates onto the nine spin-free pair basis states. This construction gives a valid reduced density matrix whose partial traces recover the spin-free one-orbital reduced density matrices, and it is what carries the spin-projection invariance argument through the separation of spatial and spin degrees of freedom.","core_discovery":"The paper claims that by collapsing the four-dimensional on-site orbital basis to the three-dimensional spin-free basis, where the singly occupied state treats alpha and beta spins as one microstate, one obtains orbital and pair entropies and mutual information that are invariant to the spin projection component of a spin multiplet. The spin-free pair entropy is constructed from spin-summed eigenvalues of the pair density matrix in the nine-dimensional product basis, which preserves additivity for uncorrelated orbitals. The paper further proves the inequalities that the spin-free quantities cannot exceed the original spin-including ones, so the spin-free measures are a coarsening of the original ones. In test calculations, spin-free total quantum information is constant across spin projection components while spin-including values grow sharply as the component decreases, and comparison of the two identifies how much apparent correlation is just spin coupling.","pith_inferences":["The same spin-free projection could be applied to other one- and two-orbital correlation measures, such as negativity or discord, raising the question of which entanglement diagnostics remain invariant after spin summation.","For wave functions that are superpositions of several spin-adapted configurations, the formal spin-projection invariance proof would need to be generalized; numerical tests suggest invariance may still hold, but this is not established by the paper.","The spin-free pair information could serve as an orbital-ordering cost function for DMRG that is less sensitive to spin contamination in non-spin-adapted calculations.","One testable extension is to apply the analysis to spin-adapted wave functions where the spatial and spin product factorization is exact per configuration, and check whether across-configuration interference breaks the invariance."],"forward_implications":["Spin-free total quantum information and total mutual information can be tabulated for a spin multiplet without repeating calculations for every spin projection component.","The gap between spin-free and spin-including entropy for a low-component state quantifies how much of the apparent orbital correlation is an artifact of spin coupling rather than genuine strong correlation.","Because spin-free measures are invariant to the spin projection, they expose trends such as increasing total spin-free entropy with decreasing total spin, which spin-including values obscure.","The method adds negligible cost to an existing DMRG or full-CI entropy analysis, so it can be reported routinely as a complement to spin-including values."],"supporting_citations":[{"why":"Defines the orbital mutual information and pair entropy measures that the paper generalizes to spin-free form.","marker":"[13]"},{"why":"Establishes the standard orbital entanglement and pair entropy analysis that this work modifies.","marker":"[16]"},{"why":"Supplies the iron-sulfur dimer complexes and benchmark data used for the realistic test.","marker":"[29]"},{"why":"Provides the restricted open-shell Hartree-Fock description of the iron complexes and the geometries used in the DMRG calculations.","marker":"[35]"},{"why":"Gives the von Neumann entropy and mutual information definitions and subadditivity used throughout.","marker":"[39]"},{"why":"Provides the data processing inequality used to prove that spin-free mutual information cannot exceed the spin-including one.","marker":"[44]"}],"fun_headline_variants":["Spin-free entropies separate spin coupling from true correlation","Spin-projection-free entropies reveal genuine correlation","Spin-free method distinguishes spin coupling from genuine correlation"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof of spin-projection invariance assumes the N-electron wave function factorizes as a single product of a spatial function and a spin function, an assumption that holds for one spin-adapted configuration but not for a general multiconfigurational spin eigenfunction.","fun_headline_variants_meta":{"raw":{"variants":["Spin-free entropies separate spin coupling from true correlation","Spin-projection-free entropies reveal genuine correlation","Spin-free method distinguishes spin coupling from genuine correlation"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001005,"raw_usage":{"total_tokens":4227,"prompt_tokens":901,"completion_tokens":3326,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":517,"completion_tokens_details":{"reasoning_tokens":3276}},"tokens_in":517,"tokens_out":3326,"duration_ms":26929,"temperature":1.0,"reasoning_tokens":3276,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-08T21:26:04.647259+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the spin-free orbital entropy and mutual information for a small multiconfigurational state that is a linear combination of two spin-adapted configuration state functions with different spatial parts but the same total spin, for example a complete-active-space wave function for a diradical with two active orbitals. If the spin-free quantities differ between the maximum and lower spin-projection components, the invariance claim fails outside the single-factorization regime.","supporting_citations":[{"cited_title":"M.; White, S","cited_arxiv_id":null,"evidence_quote":"Defines the orbital mutual information and pair entropy measures that the paper generalizes to spin-free form."},{"cited_title":"Orbital Entanglement in Bond-Formation Processes","cited_arxiv_id":null,"evidence_quote":"Establishes the standard orbital entanglement and pair entropy analysis that this work modifies."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the iron-sulfur dimer complexes and benchmark data used for the realistic test."},{"cited_title":"L.; Maganas, D.; Neese, F","cited_arxiv_id":null,"evidence_quote":"Provides the restricted open-shell Hartree-Fock description of the iron complexes and the geometries used in the DMRG calculations."},{"cited_title":"A.; Chuang, I","cited_arxiv_id":null,"evidence_quote":"Gives the von Neumann entropy and mutual information definitions and subadditivity used throughout."},{"cited_title":"M.; Thomas, J","cited_arxiv_id":null,"evidence_quote":"Provides the data processing inequality used to prove that spin-free mutual information cannot exceed the spin-including one."}],"review_version":1}