{"id":"0bfad5ca-7155-4db6-9aff-9dceaa898af5","arxiv_id":"2505.07204","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Cr2Br6 on Au(111) shows coexisting vibrational and magnetic inelastic tunneling excitations, with a spin-degenerate S=3 ferromagnetic ground state.","lead":"Using a scanning tunneling microscope, researchers studied a new two-chromium molecule, Cr2Br6, on a gold surface, spotting a blend of vibrational and magnetic quantum excitations. The work points to a compact, well-defined system for studying how two coupled magnetic atoms behave, with possible uses in quantum devices.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The spin-step assignment requires the Au(111) substrate to compress calculated SOiCI(2) gaps from 8.5-13.5 meV to 0.2-2.4 meV, but no substrate-included calculation tests this.","rationale":"The paper's strongest claim has two parts: (i) the molecule has an S=3 ground state with a 2-2-2-1 SOC splitting, and (ii) the multi-step dI/dV spectrum is explained by coexisting vibrations and spin transitions. Part (i) is supported by consistent DMRG and iCI results for the isolated molecule and is not the main weakness. Part (ii) is where the load-bearing assumption sits. The fitted positions of the alleged spin transitions are far below the calculated SOC gaps: 0.14-0.23 meV versus 8.5 meV for the first transition, and 2.39 meV versus 13-13.5 meV for the second. The manuscript itself flags this deviation and offers two speculative explanations, but neither is tested. A substrate-induced compression large enough to close an 8.5 meV gap to below 0.3 meV is a strong effect; it would plausibly also change the vibrational frequencies, which the DFT-on-substrate calculation already reproduces reasonably well, and it could alter the degeneracy pattern. Without a substrate-included calculation, the assignment of E1 and E4 is not uniquely constrained: many other mechanisms can produce low-energy, field-dependent steps in IETS. This is the same concern identified by the reader. It does not invalidate the experimental data or the isolated-molecule quantum chemistry, but it makes the central spectroscopic assignment conditional on an untested mechanism. The proposed check would settle it, and the verdict should therefore remain CONDITIONAL rather than being upgraded.","tokens_in":10767,"tokens_out":4053,"duration_ms":40474,"concrete_test":"Compute the lowest SOiCI levels of Cr2Br6 embedded in an Au(111) cluster or with explicit charge transfer (e.g., an anionic species plus image-charge corrections), using the same CAS(26,20)/SOiCI(2) protocol. If the first excited doublet remains above about 2 meV, or if the 2-2-2-1 degeneracy pattern is not preserved, the assignment of E1 and E4 to spin transitions fails. As a minimal version, repeat the SOiCI(2) calculation on the DFT-optimized adsorbed geometry and on the anionic molecule; if neither compresses the 8.5 meV gap to below about 1 meV, the central spin assignment needs revision.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that E1, E4, E0, and E5 are spin excitations out of a spin-degenerate S=3 ground state depends on matching the observed zero-field IETS energies to the SOC-split levels of the septet. Table 2 gives the first and second excited doublets at 8.5-8.6 meV and 13.0-13.5 meV (SOiCI(2)), whereas the fits assign E1 ~0.14-0.23 meV and E4 ~2.39 meV. This is a factor of roughly 40 and 5 discrepancy, respectively, and the paper's only response is to invoke substrate charge transfer or band compression (citing Refs. 38-39) without any calculation that includes the Au(111) substrate or charge transfer. The same substrate would also need to preserve the 2-2-2-1 degeneracy pattern almost exactly, since the field-dependent splittings are interpreted as Zeeman lifts of those degeneracies. Until a substrate-included or charge-transferred calculation shows that the SOC splitting can be compressed by this amount while keeping the degeneracy pattern, the spin assignment is underdetermined: the six-mode fit is partly justified by the very theory whose gaps disagree, and the field-dependent E0/E5 steps alone do not uniquely identify which pair of states is being split.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports an STM/S study of individual Cr2Br6 molecules on Au(111), prepared by evaporating CrBr3. The authors observe a Fano resonance and several inelastic tunneling spectroscopy (IETS) steps, which they decompose into four zero-field modes and, under magnetic field, six modes. DFT calculations are used to assign two of the zero-field modes (E2 and E3) to hindered rotational and translational vibrations of the molecule, with calculated energies of 1.41 and 2.18 meV compared to fitted values of about 1.14 and 1.90 meV. DMRG and iCI calculations with a CAS(26,20) active space predict a septet (S = 3) ground state with ferromagnetic coupling between the two Cr ions, and the SOiCI and SOiCI(2) calculations split the septet into four levels with degeneracies described as 2-2-2-1. The remaining observed steps (E1, E4, E0, E5) are assigned to transitions within this SOC-split manifold, with E0 and E5 showing linear magnetic-field dependence with slopes near 2 μB. The central claim is that the molecule possesses a spin-degenerate ground state and that the complex spectrum arises from coexisting vibrational and magnetic excitations.","tokens_in":11090,"tokens_out":9793,"duration_ms":86633,"significance":"If the spin assignment is correct, this work is a significant advance: a well-defined dual-core trihalide molecule containing a single near-90° superexchange bridge is characterized by a combination of atomically resolved STM/S and state-of-the-art multi-reference calculations. The DMRG/iCI prediction of an S = 3 ground state is a nontrivial ab initio result, and the field-dependent slopes of E0 and E5 close to 2 μB provide a striking experimental fingerprint. The DFT-based vibrational assignment is also plausible. However, the credibility of the spin-excitation assignment currently rests on bridging a large quantitative discrepancy between the calculated and observed excitation energies, a bridge supported only by an untested substrate-compression assumption. The work is therefore of high interest but requires additional quantitative support before the spin-assignment claim can be accepted as established.","major_comments":[{"comment":"The central assignment of E1, E4, E0, and E5 to transitions within the SOC-split septet is difficult to reconcile with the calculated excitation energies in Table 2. The SOiCI(2) calculation places the first and second excited doublets at 8.5-8.6 and 13.0-13.5 meV, whereas the fits in Fig. 3i give c1 ≈ 0.23 meV and c4 = c5 ≈ 2.39 meV, a discrepancy of about a factor of 37 for E1 and a factor of 5 for E4/E5. The manuscript's only response, in the paragraph beginning 'It is worth noting...', invokes charge transfer from the Au(111) substrate and substrate-induced band compression, citing Refs. 38 and 39, but no calculation that includes the substrate or charge transfer is presented. Moreover, a simple uniform compression cannot reconcile the relative spacings: the calculated gap between the ground and first excited doublet (8.5 meV) is larger than the gap from the first to the second excited doublet (4.5 meV), while the fitted values imply the opposite ratio (0.23 meV versus about 2.16 meV). The authors need to provide a quantitative model or a substrate-included calculation (for example, an embedded multi-reference calculation or a DFT+U calculation with explicit charge transfer) demonstrating that the substrate can compress the gaps by these factors and alter their relative sizes while preserving the degeneracy pattern. Until such a calculation is provided, the spin assignment is underdetermined.","section":"Table 2 and Fig. 3i"},{"comment":"The text states that the transition from ground state 1 to state 4 is 'neglected in this analysis and included in E0, as the energy difference between state 2 and state 4 is too small', while elsewhere E0 is described as the Zeeman splitting of the doubly degenerate ground state (states 1 and 2). These two descriptions are mutually inconsistent. If E0 is the 1→2 transition, its zero-field energy should be zero; if it is the 1→4 transition, its zero-field energy should equal the calculated first-excited gap of about 8.5 meV, not the fitted c0 ≈ 0.05 meV. The schematic in Fig. 3j should be revised to show unambiguously which pairs of eigenstates contribute to each fitted step, including whether the 1→4 transition is actually resolved in the data.","section":"Fig. 3j"},{"comment":"The six-mode fit is justified in the text by the statement that 'the theoretical calculations suggest that analyzing with six excitation modes is more reasonable.' However, the authors also state that E4 and E5 are 'too close to be clearly distinguished', and the fitted E0 has zero or near-zero energy at zero field (c0 ≈ 0.05 meV). The number of IETS steps is therefore not independently determined by the data. A statistical comparison between fits with five and six steps, without imposing the theoretical level pattern, should be reported (for example, an F-test or AIC comparison) to demonstrate that the additional E0 and E5 steps are real spectral features rather than artifacts of overfitting. This is directly relevant to the Zeeman slopes that constitute the main experimental evidence for spin excitations.","section":"Fig. 3a and Supplementary Fig. 4"},{"comment":"The claim of a 2-2-2-1 degeneracy pattern is only approximate. In the SOiCI(2) results, states 3 and 4 are split by 0.1 meV and states 5 and 6 by 0.5 meV. The latter splitting is larger than the entire fitted E1 energy (c1 ≈ 0.23 meV) and an order of magnitude larger than the fitted c0. The authors should state whether these intra-doublet splittings are physical or numerical, and how the field-dependent assignment can treat these states as degenerate doublets when the calculated splittings are comparable to or larger than the observed IETS energies.","section":"Table 2"}],"minor_comments":[{"comment":"In the abstract, 'the Cr2Br6, exhibits' contains a stray comma before the verb; it should read 'Cr2Br6 exhibits'.","section":"Abstract"},{"comment":"The sentence 'the energy resolution of the calculation method does reach the meV scale' appears to be a typo; presumably it should read 'does not reach the sub-meV scale'. As written, it does not explain the discrepancy it is meant to address.","section":"Results, paragraph beginning 'It is worth noting...'"},{"comment":"The zero-field energies reported in the Fig. 2 caption (Δ1 ≈ 0.14 meV, Δ4 ≈ 2.5 meV) differ from the c1 and c4 values in Fig. 3i (0.23 meV and 2.39 meV) by amounts (0.09 meV and 0.11 meV) larger than the stated fitting error for ci (±0.05 meV). Please report a single consistent set of final zero-field energies and explain how the two fitting procedures were reconciled.","section":"Fig. 2 caption and Fig. 3i"},{"comment":"The DFT vibrational calculations are reported to give three lowest-energy modes at 1.413, 2.180, and 2.984 meV, but only the first two are discussed in relation to observed steps. The 2.984 meV mode is not assigned to any spectral feature; the authors should clarify whether this mode is expected to be IETS-inactive or falls outside the measured energy window.","section":"Results, vibrational assignment"},{"comment":"The caption states 'The dashed curve in h corresponds to the 1 T curve from g', but the dashed curve is difficult to identify in the figure; consider plotting it in a different color or using a more visible line style.","section":"Fig. 3h caption"}],"recommendation":"major_revision","confidential_remarks":"The paper reports a substantial experimental and theoretical effort, and the multi-reference calculations are state-of-the-art. My concern is strictly about the quantitative gap between the calculated spin-excitation energies and the fitted IETS steps. The reliance on substrate-induced compression is plausible in principle but is not supported by a calculation in the manuscript, and the cited references (38 and 39) concern different systems and do not directly demonstrate compression of spin-excitation gaps in an adsorbed magnetic molecule. I recommend major revision rather than rejection because a feasible substrate-included or charge-transferred calculation could resolve the discrepancy and would fit within the scope of a revision. I also encourage the editor to verify that the supplementary material contains the detailed fitting statistics and the sixth excitation energy extraction, as these are needed to assess the six-mode decomposition."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe paper reports a clean STM/S study of a new on-surface molecule, Cr2Br6 on Au(111), with a dual-core structure and a near-90° Cr-Br-Cr superexchange bridge. The experiments are careful and the data look credible. The molecule is genuinely new, and the combination of high-resolution IETS with large-active-space DMRG/iCI calculations on an adsorbed species is a step beyond what is usually done.\n\nThe strongest part is the structural work. The STM images, the DFT-optimized adsorption geometry, and the simulated images agree well, and the vibrational assignment for E2/E3 (1.41 and 2.18 meV calculated vs 1.14 and 1.9 meV measured) is quantitatively reasonable. That part of the story I buy. The spin-state degeneracy pattern (2-2-2-1) from SOiCI(2) is also a genuine ab initio prediction, not a fit.\n\nThe sore spot is exactly where the reader put it. The zero-field spin gaps from SOiCI(2) are 8.5–13.5 meV, while the E1/E4 features assigned to spin excitations sit at 0.14–2.39 meV. That is a factor of 40 and 5. The paper acknowledges the mismatch and gestures at substrate charge transfer and band compression, but no calculation with the substrate or with charge transfer is shown. The same substrate would also have to compress the gaps while preserving the 2-2-2-1 degeneracy nearly exactly, because the field-dependent E0/E5 splitting is interpreted as a Zeeman lift. That is a lot to assume untested. The six-mode fit is partly theory-driven (E4/E5 are nearly degenerate in the data), so the fit is not independent confirmation of the spin assignment. The field-dependent slopes near 2 μB are suggestive, but they do not pin down which states are involved.\n\nThis is not a fatal flaw: the experimental observations remain interesting, and the mismatch is a specific, addressable problem. But as written, the central spin assignment is underdetermined. A substrate-included or charge-transferred calculation with the SOC gaps would go a long way.\n\nWho should read it: anyone working on on-surface molecular magnets, vibrational vs magnetic IETS separation, or dual-core superexchange systems. It deserves a serious referee and likely revision, not a desk reject. I would not cite the spin assignment as established, but I would cite the structural work and the method combination.\n\nRecommendation: send to peer review; the referee should push for a quantitative test of the substrate compression or a repositioning of the claim.","headline":"Solid experimental STM/STS study of a new dual-core Cr2Br6 molecule with a credible vibrational assignment, but the spin-excitation interpretation leans on an unexplained order-of-magnitude gap between theory and experiment.","tokens_in":11699,"tokens_out":667,"would_cite":true,"duration_ms":7448,"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":"This paper establishes that the low-energy tunneling spectrum of a single Cr2Br6 molecule on Au(111) comes from coexisting molecular vibrations and spin excitations of a spin-degenerate S=3 ground state formed by two ferromagnetically…","keywords":["scanning tunneling microscopy","spin excitations","inelastic electron tunneling spectroscopy","Kondo effect","magnetic molecule","spin-orbit coupling","superexchange","chromium bromide"],"falsifier":"Run the same 26-electron, 20-orbital multireference spin-orbit calculation on a model that explicitly includes the Au(111) substrate (or a gold cluster) and check whether the S=3 spin gaps compress to the observed ~0.2–2.5 meV while maintaining the 2-2-2-1 degeneracy. Alternatively, measure the dI/dV spectrum of Cr2Br6 on an insulating decoupling film such as NaCl on Cu(111): if the spin steps do not shift to the predicted 8.5–13.5 meV range, the substrate-compression assumption is falsified.","tokens_in":10554,"feed_emoji":"🧲","tokens_out":12269,"duration_ms":112859,"temperature":0.7,"pith_summary":"An individual Cr2Br6 molecule, two chromium ions bridged by one near-90° Cr–Br–Cr scissors bond, is shown to have a septet (S=3) ground state in which the two Cr spins couple ferromagnetically. Tunneling spectra on the molecule show an asymmetric Fano-shaped Kondo resonance at zero bias, flanked by mirror-symmetric steps, and magnetic-field measurements split only some of those steps while leaving others fixed. The paper assigns two steps to molecular vibrations (hindered rotation and translation, computed at 1.4 and 2.2 meV) and the remaining steps to spin excitations of a spin-orbit-split manifold whose lowest levels have degeneracies 2-2-2-1 with a doubly degenerate ground state. This provides a minimal, well-defined two-core superexchange-coupled molecule for studying quantum spin excitations at the atomic scale.","feed_headline":"Spin and vibration steps in Cr2Br6 split to expose an S=3 ground state","feed_subtitle":"Two Cr spins couple into a magnetic ground state, giving a minimal two-core platform for molecular quantum spin functions.","key_machinery":"The load-bearing object is the spin-orbit-coupled septet manifold of Cr2Br6: a 26-electron, 20-orbital active space treated with density matrix renormalization group and iterative configuration interaction, then with spin-orbit coupling, yields a ground S=3 state that splits into four levels with degeneracies 2-2-2-1. The identity that carries the argument is the Zeeman branching of that manifold: under a magnetic field the doubly degenerate ground and excited states split into minority and majority states, so transitions to majority states stay field-independent while transitions to minority states rise linearly with field. That pattern, together with the two computed vibrational modes at 1.4 and 2.2 meV, is what lets the authors separate the observed steps into vibrations and spin excitations.","core_discovery":"The central discovery is that the inelastic tunneling spectrum of a single Cr2Br6 molecule combines two coexisting excitation channels, and that the magnetic channel is a ferromagnetically coupled S=3 septet. At zero field the dI/dV spectrum shows a Fano peak with four inelastic tunneling spectroscopy (IETS) steps (E1–E4); under an out-of-plane magnetic field two additional steps (E0, E5) appear, and their energies grow linearly with field (slopes about 0.10 and 0.09 meV/T) while E1–E4 stay nearly constant. DFT identifies the 1.4 and 2.2 meV modes as hindered rotation and translation, matching E2 and E3. Multireference calculations with a 26-electron, 20-orbital active space place the S=3 state as ground state, and including spin-orbit coupling splits it into four levels with degeneracies 2-2-2-1, the lowest doubly degenerate. The paper attributes E1, E4, E0, and E5 to transitions within this spin manifold, noting that the calculated gaps (8.5–13.5 meV) are larger than the observed sub-meV-to-few-meV splittings and invoking a substrate-induced compression to reconcile the difference.","pith_inferences":["A decisive test would be to measure Cr2Br6 on a decoupling insulating film such as NaCl on Cu(111): if the spin steps move toward the calculated 8.5–13.5 meV energies, the Au(111) substrate compression is confirmed; if they remain sub-meV, the calculated gaps themselves are likely overestimated.","The observed field slopes near 0.10 meV/T are close to the value expected for simple g-factor Zeeman shifts, so a quantitative g-factor analysis of E0 and E5 could check the S=3 multiplet assignment against the field-dependent data.","The near-90° Cr–Br–Cr bridge is a two-center realization of the standard superexchange rules; varying the bridge angle or the halide could tune the sign and strength of the coupling, making Cr2Br6 a testbed for exchange control in molecule-based spin units.","If the doubly degenerate ground manifold is magnetically addressable, the molecule is a natural candidate for a surface-supported molecular spin qubit pair, although decoherence on a metal substrate remains an open question."],"forward_implications":["The 2-2-2-1 degeneracy pattern with a doubly degenerate ground state means the two Cr ions are ferromagnetically superexchange-coupled through a single near-90° Cr–Br–Cr bond, giving a minimal molecular realization of the coupling path in layered CrBr3.","In this class of molecules, the low-energy dI/dV spectrum cannot be read as purely magnetic; vibrational modes must be identified before assigning spin excitation energies.","Field-dependent IETS can distinguish spin excitations from non-magnetic ones by the slope of the step energy versus magnetic field, and the slope pattern maps the degeneracy structure of the ground manifold.","If the substrate-compression explanation is correct, quantitative predictions of spin excitations in adsorbed molecules require wavefunction calculations that include the metal surface, not just the free molecule."],"supporting_citations":[{"why":"Supplies the standard lineshape and interpretation for spin excitations as step-like features in STM dI/dV spectra.","marker":"[1]"},{"why":"Establishes single-atom spin-flip spectroscopy, the experimental signature the paper extends to a two-core molecule.","marker":"[10]"},{"why":"Provides the Fano resonance function and the linewidth-to-temperature relation used to fit the Kondo peak and extract Tk.","marker":"[11]"},{"why":"Introduces the density matrix renormalization group method used for the 26-electron, 20-orbital spin-state calculation.","marker":"[25]"},{"why":"Supplies the iterative configuration interaction (iCI) method used as the second independent multireference approach.","marker":"[29]"},{"why":"Cited for the observation that substrate charge transfer lowers molecular excitation energies, supporting the proposed reconciliation of computed and measured gaps.","marker":"[38]"},{"why":"Cited for substrate-induced compression of CrBr3-derived electronic bands, further supporting the same reconciliation.","marker":"[39]"}],"fun_headline_variants":["Dual-core Cr2Br6 exposes S=3 via spin-vibration steps","Cr2Br6's magnetic ground state S=3 from spin and vibration steps","Two Cr spins couple to S=3 in Cr2Br6's vibrational steps","Spin and vibration mix to yield S=3 in dual-core Cr2Br6"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the Au(111) substrate compresses the calculated spin gaps (8.5–13.5 meV) down to the observed sub-meV-to-few-meV energies while preserving the 2-2-2-1 degeneracy pattern; the paper states this as a possibility but does not test it with a calculation that includes the substrate.","fun_headline_variants_meta":{"raw":{"variants":["Dual-core Cr2Br6 exposes S=3 via spin-vibration steps","Cr2Br6's magnetic ground state S=3 from spin and vibration steps","Two Cr spins couple to S=3 in Cr2Br6's vibrational steps","Spin and vibration mix to yield S=3 in dual-core Cr2Br6"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000606,"raw_usage":{"total_tokens":2865,"prompt_tokens":1023,"completion_tokens":1842,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":639,"completion_tokens_details":{"reasoning_tokens":1755}},"tokens_in":639,"tokens_out":1842,"duration_ms":12105,"temperature":1.0,"reasoning_tokens":1755,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T22:22:47.395936+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same 26-electron, 20-orbital multireference spin-orbit calculation on a model that explicitly includes the Au(111) substrate (or a gold cluster) and check whether the S=3 spin gaps compress to the observed ~0.2–2.5 meV while maintaining the 2-2-2-1 degeneracy. Alternatively, measure the dI/dV spectrum of Cr2Br6 on an insulating decoupling film such as NaCl on Cu(111): if the spin steps do not shift to the predicted 8.5–13.5 meV range, the substrate-compression assumption is falsified.","supporting_citations":[{"cited_title":"J., Gupta, J","cited_arxiv_id":null,"evidence_quote":"Establishes single-atom spin-flip spectroscopy, the experimental signature the paper extends to a two-core molecule."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Fano resonance function and the linewidth-to-temperature relation used to fit the Kondo peak and extract Tk."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the density matrix renormalization group method used for the 26-electron, 20-orbital spin-state calculation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the iterative configuration interaction (iCI) method used as the second independent multireference approach."},{"cited_title":"& Morgenstern, K","cited_arxiv_id":null,"evidence_quote":"Cited for the observation that substrate charge transfer lowers molecular excitation energies, supporting the proposed reconciliation of computed and measured gaps."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Cited for substrate-induced compression of CrBr3-derived electronic bands, further supporting the same reconciliation."}],"review_version":1}