{"id":"34be414e-1320-4777-9a51-9f624a557727","arxiv_id":"2501.04176","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"The (4,0)-multichannel Dyson equation, coupling two- and four-body Green's functions, describes single and double neutral excitations in a two-level model, unlike static BSE.","lead":"This paper derives a new set of equations, the multichannel Dyson equations, that couple the two-body and four-body Green's functions to capture double excitations in neutral spectra. A proof-of-principle on a two-level model shows that the method finds a double-excitation state that standard Bethe-Salpeter approaches miss.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The effective Hamiltonian in Eq. (43) has occupation-factor off-diagonal blocks inconsistent with the factorization in Eq. (42), which would break Hermiticity and the eigenvalue mapping for general occupations.","rationale":"The reader's weakest assumption concerns the sufficiency of the static first-order multichannel self-energy, and the baseline MCDE@HF double excitation (75.73 eV vs 58.02 eV exact) does indeed contradict the abstract's 'good agreement' wording. That is a real concern about quantitative accuracy and presentation. However, the most load-bearing condition for the central claim is that the effective Hamiltonian used to produce all numerical results is correctly derived. The occupation-factor asymmetry in Eq. (43) is not merely stylistic: if the off-diagonal blocks lack the correct occupation factors, H_eff ceases to be Hermitian for general fillings, and the eigenvalues no longer represent poles of L4. The paper's own model is immune because all relevant occupation differences equal 1, so the inconsistency is not detected numerically. This is an internal consistency problem that must be fixed before the method can be applied beyond the toy model. In other respects the paper has independent support: a parameter-free derivation, a clear diagrammatic analysis, and a demonstration that the double excitation appears where static BSE fails. The formal issue is therefore the single load-bearing concern we would press; the verdict remains conditional pending correction of Eq. (43).","tokens_in":21055,"tokens_out":14911,"duration_ms":130001,"concrete_test":"Symbolically re-derive Eq. (43) from Eq. (41) and check whether the off-diagonal blocks of H_eff are Hermitian conjugates for a system with non-binary occupations, e.g., a three-level model with one partially occupied level. Equivalently, for such a model, diagonalize H_eff as printed and compare the eigenvalues with the poles of L4 obtained by direct inversion of Eq. (28); any mismatch would confirm the error. If the printed Eq. (43) is a typo, the correct factors should be -f_jl Σ̃^{2p/4p}_{jl;mokp} and -f_in f_il f_jn Σ̃^{4p/2p}_{i>jl>n;ok}.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that (4,0)-MCDE eigenvalues are the neutral excitation energies rests on the equivalence between the Dyson equation (28) and the effective Hamiltonian problem (44). Starting from Eq. (41), L4 = [L0^{-1} - Σ̃]^{-1}, and using [AB]^{-1} = B^{-1}A^{-1} with D = diag(f_jl, ..., f_in f_il f_jn, ...), one obtains L4 = (H_eff - ω)^{-1} D with H_eff = ε - D Σ̃. This gives off-diagonal blocks -f_jl Σ̃^{2p/4p}_{jl;mokp} and -f_in f_il f_jn Σ̃^{4p/2p}_{i>jl>n;ok}. Instead, Eq. (43) prints -Σ̃^{2p/4p}_{jn;m>ok>p} f_mp f_mk f_op and -Σ̃^{4p/2p}_{i>jl>n;ok} f_ok. These occupation factors are not equal for general filling, so the printed H_eff is not Hermitian even though the paper states that it is. Consequently, the spectral representation in Eq. (45) and the use of real eigenvalues in Table I are not mathematically guaranteed for general systems. In the two-level model all relevant occupation differences equal 1, so the numerical results are unaffected; however, the method as a general formalism is not well-defined until this inconsistency is resolved. This is a load-bearing internal consistency issue for the paper's general claim.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper extends the multichannel Dyson equation (MCDE) formalism from the previously studied coupling of the 1-GF and 3-GF to the general coupling of even- and odd-order Green's functions, and then focuses on the (4,0)-MCDE that couples the 1e-1h channel of the 2-GF with the 2e-2h channel of the 4-GF. The multichannel self-energy is approximated at first order in the interaction, the Dyson equation is mapped to an effective eigenvalue problem, and the method is illustrated on a two-level, two-electron helium-like model. The central qualitative claim is that this static formulation captures both single and double neutral excitations, including a double-excitation pole that is absent in static BSE@GW. The paper also provides a diagrammatic analysis, explicit real-space self-energy expressions, and a discussion of how dressing the independent-particle 4-GF improves the double-excitation energy.","tokens_in":21376,"tokens_out":16867,"duration_ms":149180,"significance":"If the general formalism is correct, the paper offers a systematic route to treat single and double neutral excitations on equal footing without introducing a frequency-dependent BSE kernel, which addresses a well-known limitation of static BSE-based approaches. The derivation is genuinely diagrammatic and contains no fitted parameters; the double-excitation pole emerges from the coupling to the 4-GF, and the two-level model provides a clean demonstration of this mechanism. The explicit formulas for the multichannel self-energy and the real-space diagrams are useful and checkable. However, the general validity of the eigenvalue mapping in Section II.F is compromised by an occupation-factor inconsistency in the effective Hamiltonian, and the abstract overstates the agreement obtained at the baseline MCDE@HF level. The numerical demonstration is limited to a single model system, so the quantitative generality is not yet established.","major_comments":[{"comment":"The printed effective Hamiltonian does not follow from the factorization claimed in Eq. (42). With L0_4 = (ε − ω)^{-1} D, where D = diag(f_jl, f_in f_il f_jn), the inverse in Eq. (40) gives L4 = (ε − DΣ̃ − ω)^{-1} D, so H_eff = ε − DΣ̃ with the occupation factor on the left of every self-energy block. The (2p,4p) block should therefore be −f_jl Σ̃^{2p/4p}_{jl;m>ok>p} and the (4p,2p) block should be −f_in f_il f_jn Σ̃^{4p/2p}_{i>jl>n;ok}. Equation (43) instead places the factors f_mp f_mk f_op and f_ok on the right for the off-diagonal blocks while using left factors on the diagonal. These two prescriptions are not equivalent for general fillings, because the off-diagonal blocks of Σ̃ do not commute with the block-diagonal occupation matrix D. As a result, H_eff as printed is not Hermitian in general, contrary to the statement preceding Eq. (40), and the spectral representation in Eq. (45) is not justified. The two-level model is insensitive to this problem because all relevant occupation factors equal 1, but the general formalism is not well defined until this is corrected, for example by using the symmetrized form H_eff = ε − D^{1/2} Σ̃ D^{1/2} with L4 = D^{1/2} (H_eff − ω)^{-1} D^{1/2}.","section":"Section II.F, Eqs. (42)-(43)"},{"comment":"The abstract's claim of 'good agreement with the exact results' is not supported for the baseline MCDE@HF method. In Table I, the double-excitation energy is 75.73 eV versus the exact 58.02 eV, an error of about 30%, and the text itself states that the double excitation is overestimated by about 30%. Good agreement is reached only after dressing L0,4p with the GW or experimental quasiparticle gap (MCDE@GW: 66.80 eV; MCDE@Exp: 60.26 eV). The abstract should either attribute the good agreement explicitly to the dressed variants or soften the claim for the undressed method.","section":"Abstract and Section III, Table I"}],"minor_comments":[{"comment":"The 2p block of D is printed as δ_jo δ_lp f_jl, but Eq. (26) gives L0,2p_{jl;ok} with δ_jo δ_lk; the printed index appears to be a typo.","section":"Eq. (42)"},{"comment":"The acronym 'BSW@GW' should read 'BSE@GW'.","section":"Section III, last paragraph"},{"comment":"There are several typos: 'such biexcitons' in the abstract, 'exitation' and 'illutsration' in the Conclusions, 'diagolanizes' in Appendix C, and 'indroducing' in Section III should be corrected.","section":"Throughout"},{"comment":"In the first off-diagonal block, the subscript 'jn' of Σ̃^{2p/4p} appears to be a typo for 'jl'.","section":"Eq. (43)"}],"recommendation":"major_revision","confidential_remarks":"The occupation-factor inconsistency in Section II.F is the main technical obstacle; it is local and fixable, so I do not recommend rejection. The numerical illustration is a single two-level model, so the quantitative generality of the method remains to be demonstrated beyond this benchmark. The paper would benefit from a careful rewrite of the effective-Hamiltonian section and a more cautious abstract."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, the new thing: this paper extends the multichannel Dyson equation idea from the (3,1) photoemission case to even-order Green's functions, specifically coupling the 1e-1h channel of the 2-GF to the 2e-2h channel of the 4-GF. That gives a static, first-order self-energy that puts single and double neutral excitations on the same footing, which standard static BSE cannot do. The two-level helium model demonstrates the central qualitative claim: a double-excitation pole appears in the MCDE spectrum. The diagrammatic analysis is coherent, the derivations are spelled out, and there are no fitted parameters (the experimental-gap variant is an external input, not a fit to the target energy).\n\nCredit where due: the formalism is systematic, the self-energy approximants are explicit, and the paper is honest about the 30% error in the baseline double-excitation energy. The improvement from dressing L0,4p is sensible and points to a clear path forward.\n\nNow the soft spots. The abstract's \"good agreement\" is generous for MCDE@HF: the double excitation sits at 75.7 eV against an exact 58.0 eV. The authors acknowledge this in the text, so it is mainly an abstract-wording problem. More important, the stress-test on Eq. (43) holds up. Starting from Eq. (41) and factoring out the occupation-number matrix D, the off-diagonal blocks of H_eff should carry the same occupation structure on both sides. The printed H_eff instead uses f_mp f_mk f_op on one off-diagonal and f_ok on the other; these are not equal for general occupations, so H_eff is not Hermitian as claimed and the spectral representation in Eq. (45) is not guaranteed. In the two-level model every occupation difference is ±1, so the numerical results are unaffected, but the general formalism is not well-defined as written. That needs correction. Also, the demonstration is a single toy model; that is fine for a proof of principle, but the quantitative value in realistic systems is unproven.\n\nWho this is for: people working on BSE, double excitations, and multichannel Green's-function methods. It is a serious contribution to method development and deserves a serious referee. My recommendation: send it to review, but the referee should push for a corrected derivation of H_eff and a toned-down abstract.","headline":"A useful extension of MCDE to double excitations, with a correctable occupation-factor bug in the effective Hamiltonian and an overstating abstract.","tokens_in":21956,"tokens_out":9591,"would_cite":true,"duration_ms":84837,"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":"A multichannel Dyson equation coupling the two-body and four-body Green's functions describes single and double neutral excitations on equal footing, catching the double excitation that static BSE misses.","keywords":["multichannel Dyson equation","Green's functions","double excitations","Bethe-Salpeter equation","four-body Green's function","neutral excitations","two-level model","many-body perturbation theory"],"falsifier":"Compute the neutral excitation spectrum of a molecule with a known optically dark double excitation, such as the lowest $2^1A_g$ singlet of a polyene, using the (4,0)-MCDE with both Hartree-Fock and GW-dressed four-body propagators. If the double-excitation energy cannot be reproduced to the few-percent accuracy seen in the helium-like model, or if spurious additional states appear, then the truncation at the four-body Green's function with a first-order static self-energy is not sufficient for quantitative spectroscopy.","tokens_in":20859,"feed_emoji":"⚛️","tokens_out":9384,"duration_ms":74800,"temperature":0.7,"pith_summary":"The paper extends the multichannel Dyson equation (MCDE) framework, previously used to couple one- and three-body Green's functions for photoemission, to even-order Green's functions and neutral excitations. The central claim is that coupling the electron-hole channel of the two-body Green's function with the two-electron-two-hole channel of the four-body Green's function, through a static multichannel self-energy truncated at first order in the interaction, yields neutral excitation spectra that contain both single and double excitations. In a two-level helium-like model the method reproduces the triplet (18.74 vs 19.22 eV) and the single singlet (24.05 vs 23.77 eV) in good agreement, and produces the double excitation (75.73 vs 58.02 eV) that the static BSE@GW misses entirely. The paper also shows that dressing the independent-particle four-body propagator with the GW or experimental quasiparticle gap improves the double-excitation energy to 66.80 or 60.26 eV, respectively. A sympathetic reader would take the contribution to be a general, systematically improvable framework for multi-excitation spectra that standard static approaches cannot access.","feed_headline":"Coupling 2- and 4-body Green's functions catches double excitations","feed_subtitle":"Static multichannel Dyson equation puts single and double neutral excitations on equal footing, unlike BSE.","key_machinery":"The central object is the multichannel Dyson equation, a block-matrix Dyson equation in which independent-particle n-body Green's functions of different order are coupled through a multichannel self-energy. For neutral excitations the (4,0)-MCDE couples the 1e-1h block $L^{0,2p}(\\omega)$, whose poles are single-particle energy differences $\\Delta\\epsilon_{jl}$, with the 2e-2h block $L^{0,4p}(\\omega)$, whose poles are sums of two such differences, via static self-energy blocks $\\tilde\\Sigma^{2p}$, $\\tilde\\Sigma^{4p}$, and coupling blocks $\\tilde\\Sigma^{2p/4p}$ and $\\tilde\\Sigma^{4p/2p}$. All self-energy terms are truncated at first order in the Coulomb interaction (the RPAx level), yet iterating the equation generates diagrams of all orders in the interaction, including screening and ladder diagrams, because the coupling blocks dress individual propagators and interactions. The equation is solved as an eigenvalue problem for an effective four-particle Hamiltonian $\\bar{H}^{\\rm eff}_4$, whose eigenvalues are the excitation energies; the construction of this Hamiltonian scales as $N_v^3 N_c^3$, i.e., $N^6$ in the number of electrons.","core_discovery":"The central claim of the paper is that the (4,0)-multichannel Dyson equation, which couples the 1e-1h channel of the two-body Green's function with the 2e-2h channel of the four-body Green's function through a static multichannel self-energy containing only first-order-in-the-interaction terms, describes single and double neutral excitations on equal footing. In the two-level helium-like model, diagonalization of the resulting effective Hamiltonian yields a triplet and two singlet excited states, one of single and one of double excitation character; the MCDE@HF energies are 18.74, 24.05, and 75.73 eV against exact values 19.22, 23.77, and 58.02 eV, while standard static BSE@GW produces only the two single-excitation states. Replacing the Hartree-Fock gap in the four-body propagator with the GW or experimental quasiparticle gap brings the double excitation to 66.80 or 60.26 eV, respectively, and the paper reports that the singlet single excitation also improves to 23.77 or 23.48 eV. The paper concludes that the MCDE provides a natural framework for multi-excitation effects such as biexcitons, and notes that the same recipe can be applied to odd-order Green's functions and to other channels and spectroscopies.","pith_inferences":["If the accuracy seen in the helium-like model carries over to realistic systems, static first-order multichannel self-energies could replace frequency-dependent kernels for absorption spectra with multi-exciton character, substantially lowering computational cost.","The strong starting-point dependence of the double excitation (75.73 to 60.26 eV against an exact 58.02 eV) suggests that a fully predictive implementation will need self-consistent or GW-quality dressing of $L^{0,4p}$; the bare RPAx truncation alone does not carry the quantitative accuracy.","The even/odd decoupling is a structural prediction that could be tested independently: a (3,1)-MCDE for charged excitations and a (4,0)-MCDE for neutral excitations should each close without leakage into the other class, at any truncation order.","A natural next test is the particle-particle channel of the 2-GF, where the same multichannel construction could be compared against the anomalous-propagator BSE for pairing problems."],"forward_implications":["The (4,0)-MCDE produces a double-excitation state that static BSE@GW misses entirely, while avoiding the spurious unphysical energies that dynamical BSE@GW produced for the same two-level model.","Dressing the independent-particle four-body propagator with GW or experimental quasiparticle gaps markedly improves the double-excitation energy, pointing to a systematic improvement path for the framework.","The MCDE with a static first-order self-energy is exact through second order in the interaction for the two-body block and naturally includes screening and ladder diagrams despite using only the bare Coulomb interaction.","The even/odd decoupling of Green's functions means neutral and charged excitation spectra can be treated by separate multichannel equations, and the same construction extends to higher-order Green's functions and to other spectroscopies such as trions and RIXS."],"supporting_citations":[{"why":"the (3,1)-MCDE strategy for photoemission that this work extends to even-order Green's functions","marker":"[11]"},{"why":"introduces the two-level helium-like model, its exact energies, and the dynamical BSE@GW results used as benchmark","marker":"[30]"},{"why":"shows that channels of higher-order Green's functions share the poles of lower-order ones, the theoretical basis for coupling them","marker":"[35]"},{"why":"the idea that a static self-energy in a large space captures dynamical properties of a smaller space, underpinning the static MCDE","marker":"[39]"},{"why":"standard BSE review defining the baseline that misses double excitations","marker":"[42]"},{"why":"previous extension of BSE to the 2e-2h space for biexcitons, the approach the paper situates itself against","marker":"[44]"}],"fun_headline_variants":["Multichannel Dyson equations bring double excitations into reach","Coupling 2- and 4-body Green's functions goes beyond BSE","A static multichannel Dyson equation puts double excitations on equal footing"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that truncating the multichannel Dyson equation at the four-body Green's function, with a static self-energy containing only first-order-in-interaction terms, captures the essential physics of double excitations; the strong sensitivity of the double-excitation energy to the dressing of the four-body propagator (75.73 to 60.26 eV against an exact 58.02 eV) shows how much this assumption depends on the starting point.","fun_headline_variants_meta":{"raw":{"variants":["Multichannel Dyson equations bring double excitations into reach","Coupling 2- and 4-body Green's functions goes beyond BSE","A static multichannel Dyson equation puts double excitations on equal footing"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00111,"raw_usage":{"total_tokens":4650,"prompt_tokens":992,"completion_tokens":3658,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":608,"completion_tokens_details":{"reasoning_tokens":3595}},"tokens_in":608,"tokens_out":3658,"duration_ms":23496,"temperature":1.0,"reasoning_tokens":3595,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T21:39:46.481051+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the neutral excitation spectrum of a molecule with a known optically dark double excitation, such as the lowest $2^1A_g$ singlet of a polyene, using the (4,0)-MCDE with both Hartree-Fock and GW-dressed four-body propagators. If the double-excitation energy cannot be reproduced to the few-percent accuracy seen in the helium-like model, or if spurious additional states appear, then the truncation at the four-body Green's function with a first-order static self-energy is not sufficient for quantitative spectroscopy.","supporting_citations":[{"cited_title":"Torche \\ and\\ author G","cited_arxiv_id":null,"evidence_quote":"previous extension of BSE to the 2e-2h space for biexcitons, the approach the paper situates itself against"}],"review_version":1}