{"id":"a3838886-daf4-474e-b9a6-b835f7c73e5f","arxiv_id":"2411.15383","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Resonant x-ray scattering and high-temperature susceptibility establish that the Ru2O9 dimers in Ba3ZnRu2O9 have a conventional S=3/2 ground state.","lead":"Single-crystal measurements show that the magnetic dimers in the perovskite Ba3ZnRu2O9 are in a conventional S=3/2 spin state, not the exotic orbital-selective S=1 state proposed theoretically. The paper demonstrates that resonant inelastic x-ray scattering can identify the magnetic ground state of a correlated material directly.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"RIXS peaks may be contaminated by the 9% Ru-on-Zn site mixing; the 0.80/1.22 eV fingerprint is not shown to originate from the Ru2O9 dimer sites.","rationale":"The reader's conditional verdict is sound, but the most load-bearing weakness is not the energy-ratio mismatch alone. The paper itself establishes 9% Ru/Zn site mixing (Table I) and EDX shows excess Ru on the surface probed by RIXS (Ba3.1Zn0.9Ru2.2O8.9). These defect Ru ions are isolated rather than dimerized, are known to contribute free spins at low temperature, and can generate Hund multiplet transitions at similar energies in RIXS. Since RIXS sums over all Ru sites, a 9% defect population could dominate or significantly contaminate the 0.80/1.22 eV features even if the dimer ground state is the orbital-selective S=1 state. The authors address the metallic 4H-BaRuO3 impurity but not this site-mixing channel. The concern is concrete and falsifiable: a resonant-energy or site-ordered sample comparison would settle whether the peaks are dimer-specific. This does not overturn the paper; it reinforces the need for the quantitative, site-aware analysis that the conditional verdict already demands. The susceptibility analysis provides independent support for an S=3/2 dimer, so the conclusion is plausible, but the RIXS 'fingerprinting' claim should be conditioned on excluding defect-site scattering.","tokens_in":11606,"tokens_out":8223,"duration_ms":83351,"concrete_test":"Perform Ru L3-edge RIXS with fine incident-energy steps across the absorption edge for the 0.80 and 1.22 eV loss features and extract their resonance excitation profiles. Compare these profiles with the Ru L3 XAS of the stoichiometric dimer sites and with a single-ion Ru5+ reference (e.g., Ru substituted at the Zn site in a nonmagnetic analog). If the 0.80/1.22 eV peaks resonate at a distinct incident energy corresponding to the 9% Ru-on-Zn defect sites, or if their intensity scales with the defect fraction, the RIXS fingerprint is not attributable to the dimer ground state. If they share the main dimer resonance and survive in a site-ordered sample, the S=3/2 assignment stands.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step in the RIXS argument is not the 3JH/5JH energy ratio alone; it is the implicit assumption that the 0.80 and 1.22 eV peaks originate from the stoichiometric Ru2O9 dimer sites. The single-crystal XRD refinement in Table I reports 9% Ru/Zn site mixing (Zn1 occupancy 0.91, Ru1 0.09), and the EDX surface composition Ba3.1(1)Zn0.9(1)Ru2.2(1)O8.9(2) shows the same excess Ru on the RIXS-probed surface. These Ru ions occupy Zn sites, are not part of Ru2O9 dimers, and are known to behave as free spins at low temperature. Ru L3 RIXS is a local, site-summed probe; a 9% population of isolated Ru5+ ions in a different ligand environment can produce its own intra-t2g Hund multiplets at comparable energies. Its S=3/2 single-ion 3JH/5JH transitions would look exactly like the features the paper uses to exclude the S=1 dimer state, without establishing the ground state of the dimers. The authors carefully subtract the 4H-BaRuO3 impurity contribution to susceptibility and note that the RIXS surface is free of this metallic phase, but they never address the known site-mixing contribution to the RIXS spectra. Unless defect-site scattering is shown to be negligible, the central 'spectroscopic fingerprint' is not uniquely tied to the dimer S=3/2 ground state.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports magnetic susceptibility and Ru L3-edge RIXS measurements on single crystals of the dimer-based hexagonal perovskite Ba3ZnRu2O9. The RIXS spectra show two sharp peaks at approximately 0.80 and 1.22 eV, which the authors assign to intra-t2g Hund's-multiplet transitions (3JH and 5JH) from an S = 3/2 4A2 ground state, and a low-energy excitation at 37(13) meV assigned to the intradimer spin-triplet transition. High-temperature susceptibility data are fit to an antiferromagnetic S = 3/2 dimer model, yielding J = 277(37) K, which the authors state is consistent with the RIXS triplet energy within error. The authors conclude that the conventional S = 3/2 dimer state is realized, excluding the previously proposed orbital-selective S = 1 dimer state.","tokens_in":11951,"tokens_out":7560,"duration_ms":70177,"significance":"If the central claim holds, the paper resolves an open question about the ground state of Ba3ZnRu2O9 and provides a nice demonstration of RIXS-based spectroscopic fingerprinting for dimer-based magnets. The complementary use of bulk susceptibility and local RIXS, the explicit comparison with the S = 1 dimer scenario, and the deposition of experimental data are strengths. However, the decisive RIXS fingerprint currently rests on an energy ratio that deviates measurably from the ideal multiplet ratio, and the potential contribution of the known 9% Ru-on-Zn site mixing to the RIXS spectra is not addressed. These issues affect the load-bearing identification of the ground state and warrant additional quantitative analysis before the conclusions can be considered fully established.","major_comments":[{"comment":"The assignment of the 0.80 and 1.22 eV peaks to the 3JH and 5JH transitions of the S = 3/2 multiplet scheme relies on the ratio 1.22/0.80 = 1.53, whereas the ideal ratio for this scheme is 5/3 ≈ 1.67; the observed ratio deviates by about 8%. The alternative S = 1 scheme predicts a ratio of 2, so the observed value does exclude that simple assignment, but the identification with the S = 3/2 scheme requires a quantitative multiplet calculation including the monoclinic distortion, spin-orbit coupling, and hybridization. Please provide such a calculation for the actual C2/c crystal structure, or quantify the uncertainty in the peak positions and show that the deviation from 5/3 is within that uncertainty.","section":"Section III A, Fig. 2(c)"},{"comment":"The single-crystal refinement reports 9% Ru occupancy on the Zn site (Zn1 = 0.91, Ru1 = 0.09), and the EDX surface composition shows excess Ru (Ba3.1(1)Zn0.9(1)Ru2.2(1)O8.9(2)). Since Ru L3 RIXS is a local, site-summed probe, these defect-site Ru ions, which are not part of the Ru2O9 dimers, could contribute their own intra-t2g Hund multiplets at comparable energies. The paper attributes the low-temperature susceptibility upturn to Ru spins on Zn sites but does not address whether the 9% defect population affects the 0.80 and 1.22 eV RIXS features. The authors should either estimate the RIXS intensity expected from defect sites or demonstrate that their multiplet energies are clearly separated from the energies of the dimer sites.","section":"Table I and Appendix (EDX)"}],"minor_comments":[{"comment":"There is a grammar error: \"These results highlights\" should be \"These results highlight\".","section":"Abstract"},{"comment":"The peak positions used to compute the ratio 1.53 are not quoted with uncertainties; given the 75 meV experimental resolution, an error estimate on the ratio would help assess how significant the deviation from 5/3 is.","section":"Section III A, Fig. 2(c)"},{"comment":"The reported average triplet energy of 37(13) meV is derived from peaks that appear to vary between 30 and 45 meV; the procedure for obtaining the average and the error should be stated explicitly.","section":"Section III A, Fig. 3"},{"comment":"The susceptibility fit assumes the free-spin (site-mixing) contribution is T-independent above 400 K, but a Curie-like contribution from 9% Ru spins is not strictly T-independent; the authors should quantify its magnitude in the fitting range and state why it does not bias the extracted J.","section":"Section III B, Eq. (1)"},{"comment":"The assignment of the low-energy peak to a spin-triplet excitation is based on the temperature shift and the t2g resonance condition, but a phonon origin is not conclusively excluded; a comparison with phonon calculations or a polarization analysis would strengthen this assignment.","section":"Section III A"}],"recommendation":"major_revision","confidential_remarks":"The site-mixing concern raised in the major comments is, in my view, the more serious issue because it potentially undermines the uniqueness of the RIXS fingerprint. That said, it is addressable within the scope of a revision, for example by modeling the defect-site contribution or by measuring a Zn-free Ru5+ reference. The energy-ratio issue also requires a multiplet calculation, but the qualitative conclusion (S=3/2 versus S=1) is unlikely to change. The paper is well within the scope of the journal and the experimental data appear to be of good quality."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a solid experimental paper that probably has the right answer: Ba3ZnRu2O9 is a conventional S=3/2 dimer system, not the orbital-selective S=1 dimer proposed in Ref. [20]. The new content is real—first single-crystal RIXS and high-T susceptibility on this compound—and the two probes are genuinely independent. They also did careful impurity work: identified the 4H-BaRuO3 phase, subtracted its susceptibility, and checked the RIXS surface chemistry with EDX. Data are deposited. That part deserves credit.\n\nThe soft spot is one the paper itself hands you. The single-crystal XRD gives 9% Ru on the Zn site, and the EDX surface composition is consistent with that. RIXS is site-summed: it sees every Ru ion, including those isolated defects. Those defect Ru ions are also 4d3 in an octahedral environment, so they will have their own intra-t2g Hund multiplets with the same 3JH/5JH spacing that the paper uses to fingerprint S=3/2. The 0.80 and 1.22 eV peaks could then be dominated by the defect population rather than the dimers, and the conclusion about the dimer ground state would not follow from the RIXS alone. The authors mention isolated Ru spins on Zn sites when discussing the susceptibility, but never touch this for RIXS. That is a real hole, not a nitpick.\n\nThe energy-ratio point is secondary. 1.22/0.80 = 1.53 is not 5/3, but it is also not 2, and corrections from the monoclinic distortion and hybridization could plausibly account for the deviation. Still, a quantitative multiplet calculation would have settled it. The J mismatch between susceptibility (24 meV) and RIXS triplet (37 meV) is actually within about one combined standard deviation, so I don't see that as a problem.\n\nBottom line: the S=3/2 conclusion is probably correct—the susceptibility alone, with its clear distinction between S=1 and S=3/2 dimer fits, is fairly convincing. But the RIXS fingerprint is not uniquely tied to the dimers unless the defect contribution is shown to be negligible. The authors should either calculate or argue why the 9% defect Ru does not dominate the 0.8 and 1.2 eV peaks, or measure a purer sample. This is refereeable and worth engaging—it just needs a serious revision addressing the site-mixing confound.","headline":"A competent RIXS + susceptibility study that likely nails the S=3/2 dimer ground state, but the 9% Ru-on-Zn site mixing is an unaddressed confound for the RIXS fingerprint.","tokens_in":12536,"tokens_out":3572,"would_cite":false,"duration_ms":33480,"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 Ru2O9 dimers in Ba3ZnRu2O9 carry a conventional S=3/2 antiferromagnetic ground state, not an orbital-selective S=1 dimer.","keywords":["hexagonal perovskite","Ba3ZnRu2O9","Ru2O9 dimers","resonant inelastic x-ray scattering","Hund's multiplets","S = 3/2 spin dimer","orbital-selective Mott state","magnetic susceptibility"],"falsifier":"A ligand-field multiplet calculation that includes the monoclinic crystal-field distortion, spin-orbit coupling, and Ru-O hybridization for the dimer and finds that an $S = 1$ ($3A_2$) ground state also places two intra-$t_{2g}$ transitions near 0.80 and 1.22 eV would break the central assignment. Experimentally, the claim would be settled by resolving more than the two expected intra-$t_{2g}$ peaks, or by neutron or high-resolution RIXS measurements showing a singlet-triplet excitation spectrum incompatible with the $S = 3/2$ dimer.","tokens_in":11409,"feed_emoji":"🧲","tokens_out":13688,"duration_ms":104208,"temperature":0.7,"pith_summary":"Ba3ZnRu2O9 is a hexagonal perovskite whose magnetism lives on Ru2O9 face-sharing dimers, and this paper asks which spin state occupies each dimer: the conventional high-spin $S = 3/2$ dimer, with three unpaired $t_{2g}$ electrons on each Ru$^{5+}$ ion, or the theoretically proposed orbital-selective $S = 1$ dimer, in which strong orbital hybridization first pairs two spins in a bonding orbital. The authors use resonant inelastic x-ray scattering at the Ru $L_3$ edge on single crystals, together with high-temperature magnetic susceptibility, to distinguish the two scenarios. They report two sharp intra-$t_{2g}$ excitations at 0.80 and 1.22 eV whose energy ratio matches the Hund multiplet ladder of an $S = 3/2$ ground state, and a susceptibility that follows an antiferromagnetic $S = 3/2$ dimer model with intradimer coupling $J = 277(37)$ K, consistent with the 37 meV spin-triplet excitation observed by RIXS. If correct, the ground state of Ba3ZnRu2O9 is a conventional $S = 3/2$ dimer, and the proposed orbital-selective $S = 1$ dimer state is excluded.","feed_headline":"Ruthenium dimer magnet settles on S=3/2 ground state","feed_subtitle":"Two RIXS peaks at 0.80 and 1.22 eV fingerprint S=3/2, excluding the orbital-selective S=1 dimer proposal","key_machinery":"The load-bearing object is the Hund intraionic multiplet ladder of the $t_{2g}^3$ configuration. In a cubic field with trigonal distortion, the three $t_{2g}$ electrons of Ru$^{5+}$ can realize either a high-spin $4A_2$ ($S = 3/2$) ground state with excited $S = 1/2$ multiplets at about $3J_H$ and $5J_H$, or an orbital-selective $3A_2$ ($S = 1$) ground state with excited $S = 0$ multiplets at about $2J_H$ and $4J_H$. The ratio of the two observed RIXS peak energies, 1.53, is close to $5/3$ and far from 2, which is what lets two numbers carry the spin-state assignment. The intradimer coupling is independently extracted from the position of the spin-triplet excitation near 37 meV and from the Heisenberg antiferromagnetic dimer susceptibility formula, which fixes $J = 277(37)$ K; the agreement of these independent estimates is what closes the argument.","core_discovery":"The central claim is that the magnetic ground state of Ba3ZnRu2O9 is a conventional antiferromagnetic $S = 3/2$ dimer of two Ru$^{5+}$ ($4d^3$, $t_{2g}^3$) ions, and not the orbital-selective $S = 1$ dimer proposed in [20]. The evidence is a spectroscopic fingerprint: the RIXS spectra at the Ru $L_3$ edge show two intra-$t_{2g}$ excitations at 0.80 and 1.22 eV with ratio 1.53, which the authors assign to the Hund intraionic multiplet transitions at about $3J_H$ and about $5J_H$ out of the $4A_2$ ($S = 3/2$) ground state; the $S = 1$ scenario would instead give about $2J_H$ and about $4J_H$ transitions with ratio 2. A magnetic excitation near 37 meV is assigned to the intradimer spin-triplet transition of the $S = 3/2$ dimer, and high-temperature susceptibility is fit by an isolated Heisenberg antiferromagnetic dimer with $J = 277(37)$ K ($24(3)$ meV), consistent with the triplet energy. The same susceptibility is not reproduced by the antiferromagnetic $S = 1$ dimer model. The paper concludes that spectroscopic fingerprinting by RIXS can determine the magnetic ground state of such dimer-based systems.","pith_inferences":["An implication the paper leaves implicit is that the same two-peak ratio test could be applied to related Ba3M Ru2O9 compounds (M = Mg, Ca, Cd, Sr, Y, In, La) to map where conventional $S = 3/2$ dimers give way to singlet or orbital-selective states.","Because the observed ratio 1.53 deviates from the ideal $5/3$, a quantitative multiplet calculation including the monoclinic distortion, spin-orbit coupling, and hybridization would test whether the fingerprint remains decisive; the paper does not provide such a calculation.","The 9% Ru/Zn site mixing and the 4H-BaRuO3 impurity mean the low-temperature susceptibility is dominated by disorder; a cleaner crystal could reveal whether intrinsic $S = 3/2$ dimer correlations sustain the absence of magnetic order.","If the $S = 3/2$ assignment holds, the effective single-ion physics is governed by a single Hund coupling $J_H \\approx 0.25$ eV, which gives future electronic-structure work a simple starting point for this compound."],"forward_implications":["Ba3ZnRu2O9 should be understood as a system of weakly coupled antiferromagnetic $S = 3/2$ dimers, so its low-temperature spin-liquid-like behaviour arises within the $S = 3/2$ dimer framework rather than from an orbital-selective $S = 1$ state.","Because the extracted intradimer coupling ($24$–$37$ meV) is large compared with the undetected interdimer dispersion, the absence of long-range magnetic order down to 37 mK is compatible with this ground state once disorder is accounted for.","The orbital-selective $S = 1$ mechanism proposed for this compound is not realized in Ba3ZnRu2O9, focusing future theoretical work on the $S = 3/2$ dimer manifold.","RIXS multiplet ratios can serve as a general fingerprint for spin states in dimer-based $4d$ perovskites, complementing bulk thermodynamic measurements."],"supporting_citations":[{"why":"Supplies the previous powder-sample study whose unanswered spin-state question motivates this work.","marker":"[16]"},{"why":"The theoretical proposal of the orbital-selective S=1 dimer state that the RIXS and susceptibility data must exclude.","marker":"[20]"},{"why":"The ligand-field multiplet scheme used to assign the 0.80 and 1.22 eV transitions to Hund intraionic excitations.","marker":"[19]"},{"why":"The family of Ru2O9 compounds with singlet dimer ground states that provide the contrasting magnetic background.","marker":"[13]"},{"why":"Relates the central energy of a spin-triplet excitation to the intradimer coupling, the basis for extracting J from the 37 meV peak.","marker":"[30]"},{"why":"Provides the reference Ru5+ Hund coupling scale (about 0.25 eV) used to validate the assignment.","marker":"[28]"},{"why":"Supplies the 4H-BaRuO3 susceptibility used to subtract the impurity contribution from the measured susceptibility.","marker":"[27]"},{"why":"Demonstrates a realized orbital-selective molecular Mott state in a related compound, showing the competing scenario is physically plausible.","marker":"[15]"}],"fun_headline_variants":["RIXS fingerprint pins Ru dimer to S=3/2 state","Spectroscopic fingerprint excludes S=1 Ru dimer","Two RIXS peaks settle Ru dimer spin state","Dimer magnet's S=3/2 confirmed by RIXS","Ba3ZnRu2O9 dimers found S=3/2, not S=1"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole identification rests on the two RIXS peaks at 0.80 and 1.22 eV being the approximately $3J_H$ and approximately $5J_H$ Hund multiplet transitions out of an $S = 3/2$ ground state; monoclinic distortion, spin-orbit coupling, and hybridization must shift these energies only mildly, since no quantitative multiplet calculation including those corrections is given.","fun_headline_variants_meta":{"raw":{"variants":["RIXS fingerprint pins Ru dimer to S=3/2 state","Spectroscopic fingerprint excludes S=1 Ru dimer","Two RIXS peaks settle Ru dimer spin state","Dimer magnet's S=3/2 confirmed by RIXS","Ba3ZnRu2O9 dimers found S=3/2, not S=1"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000218,"raw_usage":{"total_tokens":1515,"prompt_tokens":1097,"completion_tokens":418,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":713,"completion_tokens_details":{"reasoning_tokens":326}},"tokens_in":713,"tokens_out":418,"duration_ms":4083,"temperature":1.0,"reasoning_tokens":326,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T14:21:39.094030+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A ligand-field multiplet calculation that includes the monoclinic crystal-field distortion, spin-orbit coupling, and Ru-O hybridization for the dimer and finds that an $S = 1$ ($3A_2$) ground state also places two intra-$t_{2g}$ transitions near 0.80 and 1.22 eV would break the central assignment. Experimentally, the claim would be settled by resolving more than the two expected intra-$t_{2g}$ peaks, or by neutron or high-resolution RIXS measurements showing a singlet-triplet excitation spectrum incompatible with the $S = 3/2$ dimer.","supporting_citations":[{"cited_title":"Terasaki, T","cited_arxiv_id":null,"evidence_quote":"Supplies the previous powder-sample study whose unanswered spin-state question motivates this work."},{"cited_title":"Tanaka and C","cited_arxiv_id":null,"evidence_quote":"The theoretical proposal of the orbital-selective S=1 dimer state that the RIXS and susceptibility data must exclude."},{"cited_title":"Sugano, Y","cited_arxiv_id":null,"evidence_quote":"The ligand-field multiplet scheme used to assign the 0.80 and 1.22 eV transitions to Hund intraionic excitations."},{"cited_title":"Darriet, M","cited_arxiv_id":null,"evidence_quote":"The family of Ru2O9 compounds with singlet dimer ground states that provide the contrasting magnetic background."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the 4H-BaRuO3 susceptibility used to subtract the impurity contribution from the measured susceptibility."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrates a realized orbital-selective molecular Mott state in a related compound, showing the competing scenario is physically plausible."}],"review_version":1}